World Cancer Report 2020

World
Cancer
Report
Cancer research
for cancer prevention
Edited by CHRISTOPHER P. WILD,

ELISABETE WEIDERPASS, and BERNARD W. STEWART
World Cancer Report
Cancer research
for cancer prevention
Edited by CHRISTOPHER P. WILD,

ELISABETE WEIDERPASS, and BERNARD W. STEWART
LYON, 2020
Private copy of Marlon Patino <mpatinob@unal.edu.co>. Unauthorized reproduction of this work is prohibited.
© International Agency for Research on Cancer 2020
Some rights reserved. This work is available under the Creative Commons Attribution-NonCommercial-NoDerivs 3.0 IGO
licence (CC BY-NC-ND 3.0 IGO; https://creativecommons.org/licenses/by-nc-nd/3.0/igo/).
Under the terms of this licence, you may copy and redistribute the work for non-commercial purposes, provided the work is
appropriately cited, as indicated below. In any use of this work, there should be no suggestion that WHO endorses any specific
organization, products or services. The use of the WHO logo is not permitted.
Any mediation relating to disputes arising under the licence shall be conducted in accordance with the mediation rules of the
World Intellectual Property Organization.
Suggested citation.
Wild CP, Weiderpass E, Stewart BW, editors (2020). World Cancer Report: Cancer Research for Cancer Prevention. Lyon,
France: International Agency for Research on Cancer. Available from: http://publications.iarc.fr/586. Licence: CC BY-NC-ND
3.0 IGO.
Sales, rights and permissions.

To purchase print copies distributed by WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland,
see http://apps.who.int/bookorders. Tel.: +41 22 791 3264; Fax: +41 22 791 4857; email: bookorders@who.int.
To purchase IARC publications in electronic format, see the IARC Publications website (http://publications.iarc.fr).
To submit requests for adaptations or commercial use and queries on rights and licensing, see the IARC Publications website
(http://publications.iarc.fr/Rights-And-Permissions).
Third-party materials.
If you wish to reuse material from this work that is attributed to a third party, such as tables, figures or images, it is your respon-
sibility to determine whether permission is needed for that reuse and to obtain permission from the copyright holder. The risk of
claims resulting from infringement of any third-party-owned component in the work rests solely with the user.
General disclaimers.
The designations employed and the presentation of the material in this publication do not imply the expression of any opinion
whatsoever on the part of WHO or contributing agencies concerning the legal status of any country, territory, city or area or of
its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent approxi-
mate border lines for which there may not yet be full agreement.
The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommend-
ed by WHO or contributing agencies in preference to others of a similar nature that are not mentioned. Errors and omissions
excepted, the names of proprietary products are distinguished by initial capital letters.
All reasonable precautions have been taken by WHO to verify the information contained in this publication. However, the pub-
lished material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpreta-
tion and use of the material lies with the reader. In no event shall WHO or contributing agencies be liable for damages arising
from its use.
Cover images, from top to bottom: Photomicrograph of retinoblastoma (Credit: Ralph C. Eagle, Jr, MD, Department of Pathology,
Wills Eye Hospital, Philadelphia, Pennsylvania, USA). A woman undergoes a computed tomography scan (Credit: T2 Images/
Getty Images). A waiter in a hoteli (tearoom) pours scalding hot milky tea for customers in Nyaru, a settlement at the edge of the
Rift Valley in Kenya; IARC and Moi University are investigating the high rates of oesophageal cancer in Kenya and have found
that consumption of hot tea is implicated (Credit: Daniel Middleton/IARC). Storage of biological samples in the IARC Biobank
(Credit: Morena Sarzo/IARC). A girl receives a human papillomavirus (HPV) vaccination; in Pune district in Maharashtra, India,
IARC is working closely with national institutions to evaluate the efficacy of fewer than three doses of HPV vaccine in protect-
ing women against cervical cancer (Credit: Morena Sarzo/IARC). Background image: The Blue Marble: Next Generation is a
mosaic of satellite data taken mostly from a NASA sensor called the Moderate Resolution Imaging Spectroradiometer (MODIS)
that flies aboard NASA’s Terra and Aqua satellites (Credit: NASA/Goddard Space Flight Center/Reto Stöckli).
IARC Library Cataloguing in Publication Data
Names: Wild, Christopher P., editor. | Weiderpass, Elisabete, editor. | Stewart, Bernard W., editor.
Title: World cancer report: cancer research for cancer prevention / edited by Christopher P. Wild, Elisabete Weiderpass, Bernard
W. Stewart.
Other titles: World cancer report 2020.
Description: Lyon: International Agency for Research on Cancer, 2020. | Includes bibliographical references and index.
Identifiers: ISBN 978-92-832-0447-3 (pbk.) | ISBN 978-92-832-0448-0 (ebook)
Subjects: MESH: Neoplasms. | Neoplastic Processes. | Global Health.
Classification: NLM QZ 220
World Cancer Report

Cancer research for cancer prevention
Editors Christopher P. Wild

Elisabete Weiderpass
Bernard W. Stewart

Associate Editors Ian A. Cree
Jacques Ferlay
Kurt Straif

Managing Editors Nicolas Gaudin
Teresa Lee
Programme Officer Véronique Chajès

English Editor Karen Müller
Project Manager Sylvia Lesage
Production Assistant Freya Damrell

Layout http://messaggio.ch
Printing Imprimerie Faurite, France
Contributors
Christian C. Abnet Christopher Bullen Ronny Drapkin
Clement A. Adebamowo Gloria M. Calaf Eric J. Duell
Demetrius Albanes Karen Canfell Karin Ekström Smedby
Laia Alemany Vilches Bochen Cao A. Heather Eliassen
Maribel Almonte Franco Cavalli Steffen Emmert
Devasena Anantharaman Stephen J. Chanock Karen M. Emmons
Annie S. Anderson Isabelle Chemin Carolina Espina
Benjamin O. Anderson Chien-Jen Chen Jessica N. Everett
Bruce K. Armstrong Wanqing Chen Veronika Fedirko
Patricia Ashton-Prolla Zhengming Chen Ian S. Fentiman
Dagfinn Aune Zvavahera Mike Chirenje Jacques Ferlay
Anssi Auvinen Vincent J. Cogliano Pietro Ferrari
Anna Babayan Aaron J. Cohen Miranda M. Fidler-Benaoudia
Chunxue Bai Graham A. Colditz James Flanagan
Rosamonde E. Banks Pietro Comba Silvia Franceschi
Partha Basu David I. Conway David O. Francis
Linda Bauld Ian A. Cree Neal D. Freedman
Iacopo Baussano Jack Cuzick Christine M. Friedenreich
Laura E. Beane Freeman Luigino Dal Maso Peter P. Fu
Sonja I. Berndt Diona L. Damian Koraljka Gall Trošelj
Margherita Bignami Robert D. Daniels Judy E. Garber
Maria Blettner George Davey Smith Gail Garvey
Ron Borland Louise Davies Gemma Gatta
Freddie Bray Sanford M. Dawsey Cindy L. Gauvreau
Paul Brennan Harry J. de Koning Adi F. Gazdar (deceased)
Louise A. Brinton Catherine de Martel Ophira Ginsburg
Jennifer D. Brooks Lynette Denny Edward L. Giovannucci
Julia Brotherton Carol E. DeSantis Rüdiger Greinert
Karen Brown Joanna Didkowska John D. Groopman
Laia Bruni Eugenia Dogliotti Giuseppe Grosso
Nele Brusselaers Laure Dossus Marc Gunter
Jason Gurney Béatrice Lauby-Secretan Josiah Ochieng

Kathryn Z. Guyton Dominique Laurier Hiroko Ohgaki
Bothwell Takaingofa Guzha C. René Leemans Klaus Pantel
Janet Hall Michael Leitzmann Alexander Parker
Susan E. Hankinson Sarah Lewis Electra D. Paskett
Zdenko Herceg Donghui Li Julietta Patnick
Rolando Herrero He Li Graham Pawelec
Rayjean J. Hung Terry Lichtor Neil Pearce
David J. Hunter Martha S. Linet David H. Phillips
Ivano Iavarone Johan P. Mackenbach Sydney E. Philpott-Streiff
André M. Ilbawi Núria Malats Martyn Plummer
Lisa Iversen Reza Malekzadeh Igor Pogribny
Charles W. Jameson Mohandas K. Mallath Kornelia Polyak

Nagarajan Rajendra Prasad
Dorota Jarosińska Alberto Mantovani
Liang Qiao
Mazda Jenab Richard M. Martin
You-Lin Qiao
Mattias Johansson John D. Mathews
Ewa Rajpert-De Meyts
Michael E. Jones Valerie McCormack
Kunnambath Ramadas
Shaoqing Ju Marjorie L. McCullough
Timothy R. Rebbeck
Rudolf Kaaks James McKay
K. Srinath Reddy
Sakari Karjalainen Francis Mégraud
Jürgen Rehm
Ausrele Kesminiene Filip Meheus
Natalie Reimers
Timothy J. Key Ronald L. Melnick
Leandro Fórnias Machado
Malcolm King Wenbo Meng de Rezende
Manolis Kogevinas Dominique S. Michaud Sabina Rinaldi
Anita Koushik J. David Miller Bridget H. Robson
James R. Krycer Steven C. Moore Eve Roman
Alan Prem Kumar Colin R. Muirhead Martin Röösli
Kunjan Kunjan Raúl Murillo Thierry Roumeguère
Carlo La Vecchia Robert Newton Esther Roura Fornells
Dirk W. Lachenmeier Chikako Nishigori Anja Rudolph
Marc Ladanyi Joëlle L. Nortier Lesley Rushton
Aoife Ryan Alexandra G. Smith Christine Varon

Rengaswamy Sankaranarayanan Martyn T. Smith Paolo Vineis
Diana Sarfati Robert A. Smith Elizabeth Ward
Catherine Sauvaget Isabelle Soerjomataram Penelope M. Webb
Augustin Scalbert Aswathy Sreedevi Elisabete Weiderpass
Ghislaine Scelo Bernard W. Stewart Jeffrey N. Weitzel
David Schottenfeld Kurt Straif Elizabeth A. Whelan
Mary K. Schubauer-Berigan Michael J. Thun David Whiteman
Wolfgang A. Schulz Herbert Tilg Christopher P. Wild
Joachim Schüz Massimo Tommasino Walter C. Willett
Nereo Segnan Steinar Tretli Martin J. Wiseman
Carlo Senore Ioannis P. Trougakos Diana R. Withrow
Gautam Sethi Michelle C. Turner Zhixun Yang
Muthu K. Shanmugam Renée Turzanski Fortner Jiri Zavadil
Tatsuhiro Shibata Giske Ursin Georg Zeller
Kevin D. Shield Toshikazu Ushijima Ariana Znaor
Jack Siemiatycki Salvatore Vaccarella
Diane M. Simeone Piet van den Brandt
For a complete list of contributors
Colinda Simons Mieke Van Hemelrijck
and their affiliations, see pages
Niels E. Skakkebæk Katherine Van Loon 573–581.
Contents
Foreword ix 3 Biological processes in cancer 147
Preface x development
Introduction xii 3.1 Sporadic cancer 148
Tumorigenesis in the absence of an established
or avoidable cause
1 The global cancer burden 15
1.1 The burden and prevention of premature 16 3.2 Genomics 154
Susceptibility and somatic patterns
deaths from noncommunicable diseases,
including cancer: a global perspective 3.3 Gene–environment interactions 164
1.2 Global trends in cancer incidence 23 The preventive implications are still not clear
and mortality 3.4 DNA repair and genetic instability 170
1.3 Transitions in human development 34 Endogenous and exogenous sources of damage
and the global cancer burden and hereditary syndromes
Known causes of human cancer by organ site 45 3.5 Inflammation 181

Playing a pivotal role in cancer pathogenesis
2 Causes of cancer, including 49 3.6 Reproductive and hormonal factors 189

hazardous circumstances Important contributors to several cancer sites
2.1 Tobacco products 50 3.7 Metabolic change and metabolomics 200

Massive and still growing causes of cancer Emerging approaches and new insights
worldwide 3.8 Epigenetics 206
2.2 Infectious agents 61 Potential in diagnostics, therapy, and prevention
Missed opportunities for prevention 3.9 Immune function 215
2.3 Alcohol consumption 68 From the tumour microenvironment to therapeutic
A leading risk factor for cancer targeting
2.4 Sunlight and ultraviolet radiation 77 3.10 The microbiome 221
Affecting skin cancer incidence in many countries Its influence on tumorigenesis and therapy
2.5 Ionizing radiation and radiofrequency 84 3.11 Identifying carcinogens from 10 key 229
electromagnetic fields characteristics
Further clarification of particular risks A new approach based on mechanisms
2.6 Diet and nutrition 92 The IARC Handbooks of Cancer Prevention 236
Understanding which factors are critical
2.7 Physical activity, sedentary behaviour, 101 4 Inequalities that affect cancer 237
and obesity prevention
Established and emerging modifiable risk factors 4.1 Inequalities between and within countries 238
2.8 Dietary carcinogens 109 Impact on cancer prevention
A continuing concern in various contexts
4.2 Socioeconomic factors and cancer 246
2.9 Contamination of air, water, soil, and food 115 prevention in Africa
The challenge is to characterize specific risks Cervical cancer as an example
2.10 Occupation 127 4.3 Cancer in urban and rural communities 252
The need for continuing vigilance in China
2.11 Pharmaceutical drugs 137 Patterns reflect social dynamics
A current focus on hormones 4.4 Socioeconomic factors and cancer 258
World Cancer Research Fund International/ 144 prevention in India
American Institute for Cancer Research Diverse interventions are needed
4.5 Variations in implementation of cancer 266 5.16 Kidney cancer 447

screening in European countries Multiple risk factors, but currently limited
Striving for best practice preventive strategies
4.6 Disparities in cancer prevention services 276 5.17 Brain cancer 454
in the USA Increasing attention on the immune response
A long-standing, persistent cause of inequity 5.18 Thyroid cancer 461
4.7 Cancer in Indigenous populations 288 The challenge of overdiagnosis
Focusing on inequalities that are sometimes 5.19 Non-Hodgkin lymphoma 468
invisible Complex etiology, including the role of immune
Towards the World Code Against Cancer 295 function
5.20 Leukaemias 477
5 Preventing particular tumour types 297 Understanding pathogenesis through similarities
A guide to the epidemiology data in Section 5: 298 and differences
Preventing particular tumour types WHO Report on Cancer: Setting priorities, 485
5.1 Lung cancer 299 investing wisely and providing care for all
Continues to be the leading cause of cancer death
5.2 Head and neck cancers 310 6 The basis for, and outcomes 487
New etiological insights from, prevention strategies
5.3 Oesophageal cancer 323 Tobacco cessation: the WHO perspective 488
A tale of two malignancies 6.1 Changing behaviour 499
5.4 Stomach cancer 333 The need for sustainable implementation
Still one of the main cancer types worldwide 6.2 Improving diet and nutrition, physical 506
5.5 Colorectal cancer 344 activity, and body weight
Decreasing disparities and promoting prevention From evidence to practice
are policy priorities 6.3 Vaccination 513
5.6 Liver cancer 355 The prospect of eliminating some cancer types
An infectious disease for many communities 6.4 Preventive therapy 522
5.7 Pancreatic cancer 367 Certain interventions clearly established
Many risk factors too poorly characterized 6.5 Managing people with high and moderate 530
to enable prevention genetic risk
5.8 Skin cancer 374 Genomic tools to promote effective cancer risk
A focus on primary prevention reduction
5.9 Breast cancer 382 6.6 Screening 540
Multiple, often complex, risk factors From biology to public health
5.10 Cervical cancer 394 6.7 Circulating DNA and other biomarkers for 550
Successes in some communities to be extended early diagnosis
worldwide Great potential, but challenges recognized
5.11 Endometrial cancer 403 6.8 Governmental action to control carcinogen 557
Prevention through control of obesity exposure
5.12 Ovarian cancer 411 Multiple options covering diverse scenarios
Complicated etiology and very few preventive 6.9 Prevention strategies common to 565
options noncommunicable diseases
5.13 Prostate cancer 421 A focus on tobacco, alcohol, obesity, and physical
Challenges for prevention, detection, and inactivity
treatment
Contributors 573
5.14 Testicular cancer 430
New inroads into early diagnosis Disclosures of interests 582
5.15 Bladder cancer 439 Sources 584
A genotoxic causal agent recognized Subject index 595
Foreword
Cancer is the second most common cause of death globally, accounting for an estimated 9.6 million deaths in 2018.
At the United Nations General Assembly in 2018, world leaders agreed to take responsibility for preventing and
treating cancer and other noncommunicable diseases, including fiscal measures to protect people from cancer-
causing products, to promote evidence-based treatment, and to work towards universal health coverage.
We have no time to lose. The cancer burden is rising globally – but not equally. The greatest impact of cancer and
the fastest increase in the cancer burden over the coming decades is projected to be in low- and middle-income
countries, many of which already face difficulties coping with the current burden. There are massive social in-
equalities in cancer, with large variations in incidence, survival, and mortality between social groups.
We have learned that many cancer cases can be prevented, and even when prevention is not possible, early
diagnosis saves lives. By using evidence-based and feasible interventions and adapting them to low- and middle-
income countries where most new cancer cases will occur, a large proportion of those cases can be prevented.
There is much that can be done to reduce social inequalities in cancer globally.
Robust, independent scientific evidence is critical, focused on the scale and patterns of cancer and its causes, pre-
vention, and early detection. The high-quality research produced by the International Agency for Research on Cancer
(IARC), working with researchers around the world, is essential for the development of evidence-based guidelines
and policy by WHO, and for regulatory decisions by national institutions to protect the health of their populations.
This new IARC World Cancer Report presents the most comprehensive, up-to-date science on cancer preven-
tion, including statistics, causes, and mechanisms, and how this can be used to implement effective, resource-
appropriate strategies for cancer prevention and early detection. It also includes examples of successful preven-
tion strategies. This book is a useful reference for researchers, cancer professionals, public health workers, and
policy-makers.
The 2017 World Health Assembly requested WHO, in collaboration with IARC, to provide a global perspective
on all measures that are recognized to limit the burden of cancer. The outcome of this charge – the WHO Report
on Cancer: Setting priorities, investing wisely and providing care for all – complements the IARC World Cancer
Report by synthesizing evidence to translate the latest knowledge into actionable policies to support govern-
ments. These two publications provide a solid foundation for effective cancer policies, and bring us closer to our
goal of changing the trajectory of cancer for communities around the world.
Dr Tedros Adhanom Ghebreyesus
Director-General
World Health Organization
World Cancer Report ix

Preface
The objective of the International Agency for Research on Cancer (IARC) is to promote international collaboration
in cancer research. The Agency is interdisciplinary, bringing together skills in epidemiology, laboratory sciences,
and biostatistics to identify the causes of cancer so that preventive measures may be adopted and the burden of
disease and associated suffering reduced. A significant feature of IARC is its expertise in coordinating research
across countries and organizations; its independent role as an international organization facilitates this activity.
As part of its wide dissemination of information about cancer, the Agency produces World Cancer Report.
The previous World Cancer Report, published in 2014, identified a foreseeable increase in the global burden of
cancer, with a particularly heavy burden projected to fall on low- and middle-income countries. This new World
Cancer Report is focused on the only consideration that will credibly decrease that burden: prevention. This vol-
ume addresses cancer research for cancer prevention.
IARC routinely coordinates specialist assessments in which multiple individual research studies – typically
hundreds or thousands of articles – are assessed by international groups of expert scientists. The results are
published as volumes of publications series, and each series is widely recognized as providing authoritative
determinations. These series include the IARC Monographs on the Identification of Carcinogenic Hazards to
Humans, which address the causes of cancer; the volumes of Cancer Incidence in Five Continents, which present
population-based data on cancer occurrence; the IARC Handbooks of Cancer Prevention, which evaluate cancer
prevention strategies; and the WHO Classification of Tumours series (also known as the WHO Blue Books), for
the histological and genetic classification of human tumours. Typically, a particular volume in each of these series
is focused on some aspect of cancer causation, prevention, pathology, and so on. This approach is not amenable
to the provision of broad perspectives.
For broad perspectives, World Cancer Report is the relevant publication. World Cancer Report is not a digest
of assessments made by IARC or any other authority. World Cancer Report is based on purpose-made assess-
ments, prepared by recognized investigators worldwide and published after undergoing peer review.
A broad perspective – and, where possible, a “bottom line” – is crucial in several respects. First, it ensures that all
relevant findings are taken into account. For example, for ultraviolet radiation in sunlight, evidence of tissue injury
from low-level exposure must be considered together with known biological benefits, including production of vita-
min D. Second, although knowledge of biological mechanisms provides valuable insights, it may not necessarily
account for human circumstances. For example, in preventing exposures to known human carcinogens, inequal-
ities between populations may contribute to marked variations in health outcomes. Third, although investigative
design may be constrained to parameters that can be readily determined, human behaviour is never restricted in
such a way. For example, the incidence of obesity-related cancers is critically affected by dietary composition,
physical activity, and sedentary practices, because these vary between communities. Finally, factors that influ-
ence cancer causation and prevention may have markedly different outcomes when implemented across com-
munities or countries that differ environmentally, sociologically, climatically, and economically.
x
IARC is uniquely placed to encompass a broad spectrum of knowledge while presenting the results in manage-
able terms. The production of World Cancer Report is achieved by engaging the Agency’s scientific staff to col-
laborate in the development of the publication at every level. This includes ensuring that the planned contents
address all relevant knowledge; identifying distinguished authors and reviewers from across the globe; ensuring
that differing perspectives are offered in a balanced, evidence-based manner; and considering circumstances
that my restrict implementation of cancer-preventive interventions.
Cancer can be prevented by avoiding exposure to a known carcinogen. However, this fundamental concept can-
not always encompass why different tumour types are particularly prevalent in some populations and not others,
or how genomics and related technologies may reveal individual susceptibility and new methods of early diagno-
sis. Nor can a simplistic understanding of cancer prevention explain why health service-related and other inequal-
ities differentially determine the degree of success of preventive initiatives in different communities. Smoking
cessation remains the most widely established means of cancer prevention, and new insights are offered in this
World Cancer Report. However, efforts to reduce the burden of cancer cover a broad range, from contending
with tumour types that essentially have no known causative agents all the way through to the prospect of cervical
cancer being eliminated by the use of vaccines, which have been developed because of research on particular
cancer-causing viruses.
Accordingly, this new World Cancer Report provides investigators with detailed information across a multidisci-
plinary spectrum. Equally, World Cancer Report provides people in the wider community, no matter where they
are located worldwide, with insights into how the cancer types that have for so long affected their communities
may now have a lesser impact than was previously thought.
Dr Elisabete Weiderpass
Director
International Agency for Research on Cancer
World Cancer Report xi

Introduction
World Cancer Report is an initiative of the International Agency for Research on Cancer (IARC) and is published
about every 5 years. Since the inception of World Cancer Report, in 2003, the editorial policy has been to provide
a concise, multidisciplinary assessment of current research, made as accessible as possible through a high il-
lustrative content and a minimum of scientific jargon. For every chapter included, authority is achieved in the first
instance by engaging experts worldwide, who then face the challenge of presenting information covering broad
fields in a few thousand words. All chapters are subject to peer review.
The scope of this World Cancer Report

The breadth of knowledge addressed in each World Cancer Report has varied to meet the needs of the time. In
2003, when the availability of concise overviews across all aspects of cancer causation, prevention, and treat-
ment in a single volume was unprecedented, a comprehensive approach was taken. Although a section on can-
cer treatment was included in the first World Cancer Report, since then there has been an explosive increase in
research on precision therapy, and coverage of this proved to be impracticable if World Cancer Report were to
remain of manageable size. The fact that World Cancer Report is concise is a central consideration and one that
readers collectively value. This may be one reason why World Cancer Report 2014 has been downloaded more
than 35 000 times.
As explained in the Preface, this World Cancer Report is focused on cancer research for cancer prevention. This
focus has necessitated the inclusion of a new section, so that the scope of available research can be adequately
recognized: a section on inequalities that affect cancer prevention. This section has not appeared in any previous
World Cancer Report.
Section 4, on inequalities that affect cancer prevention, is the antithesis of, for example, Section 3, on biological
processes in cancer development. The chapters in Section 3 concern human biology, largely without reference
to geography or community, whereas the chapters that discuss inequalities must involve references to particular
communities and their circumstances. The need to address what is particular to various communities also under-
pins the content of Section 1, about the global cancer burden.
Another first for this World Cancer Report is the inclusion of a chapter on sporadic cancer. On the basis of current
research, an attainable reduction in the incidence of cancer worldwide depends primarily on reducing exposure
to known carcinogens. However, currently available research on several cancer types, including prostate cancer,
brain cancer, and leukaemias, does not allow a clear proportion of these malignancies to be attributed to particu-
lar exogenous factors. So, in such cases, is the development of sporadic cancer due to “bad luck”, and is preven-
tion no longer a consideration? Not at all! Indeed, in such situations particularly, genomics and other technologies
are key to further investigations of etiology and to delivering new or improved procedures for early diagnosis and
screening; these matters are covered in Section 6.
xii
What information is provided in World Cancer Report?

World Cancer Report is designed to provide cancer researchers, health-care professionals, regulators, and pol-
icy-makers with current findings about the causes of cancer, its prevention, and other matters tending to reduce
the burden of cancer. In particular, this volume provides insights into fields of investigation that may be adjacent
to those with which a particular reader may be familiar. Broader professional engagement with cancer control and
a need for information by journalists, governments, and community-based cancer-oriented authorities and the
teaching profession is also recognized.
As cancer research scientists, we, the editors of this World Cancer Report, readily acknowledge the need to pro-
vide information about cancer causes and prevention to the wider community with as few barriers as are compat-
ible with an accurate understanding. In the past, such a commitment to immediate comprehension has involved
providing explanations for technical terms and/or including a glossary. We have not adopted such options, for
several reasons: to avoid interrupting the flow of information, because most of the text is immediately accessible,
and considering that search engines are available to provide access to specifics.
In providing insight to those who are not necessarily undertaking research in a particular field, some background
information must be specified. This is an important but secondary consideration. Indeed, this World Cancer Report
is not intended to be a textbook that provides a comprehensive overview of well-established key knowledge.
Therefore, given the overall constraints on length, the authors of each chapter have provided a separate set of state-
ments covering the Fundamentals (presented in a shaded sidebar). The information provided in the Fundamentals
is axiomatic to the field of research covered in the chapter, but, unlike the points given in the chapter’s Summary, is
not necessarily addressed in the main text of the chapter.
To meet the immediate needs of professionals for contemporary data, the authors of each chapter were asked
to focus on research results achieved during the past 5 years. This determinant of content is not the same as
summarizing current knowledge. For example, the chapters in Section 2, on the causes of cancer, are not neces-
sarily comprehensive. Tobacco smoking continues to be the major preventable cause of death from cancer, and
indeed from multiple other diseases, but this long-held knowledge does not, in our view, require reiteration at the
expense of describing the latest research findings, including the latest approaches to smoking cessation.
A feature of this volume, as in all previous World Cancer Reports, is that the largest single section (Section 5) is
that devoted to particular cancer types: 20 chapters. In numerical terms, 20 is small compared with the hundreds
of tumour types as documented in the WHO Classification of Tumours series (also known as the WHO Blue
Books; http://whobluebooks.iarc.fr). However, the 20 types of cancer that are covered here, when taken together,
account for the overwhelming majority of cancer cases worldwide and, of greater importance, account for almost
all initiatives aimed at cancer prevention.
World Cancer Report xiii

The volumes of Cancer Incidence in Five Continents (http://ci5.iarc.fr/) and the associated GLOBOCAN data-
base document data on incidence, prevalence, mortality, and trends for multiple cancer types across hundreds
of communities. These findings are summarized and made readily accessible online through the IARC Global
Cancer Observatory (https://gco.iarc.fr). Therefore, the epidemiological information in chapters in Section 5 is
not documented systematically. Rather, authors were invited to give priority to recent epidemiological findings
that have contributed to an increased understanding of etiology or, in some rare cases, prevention. As a result,
there are marked differences between the chapters with respect to the amount of epidemiological data presented.
Similarly, information about exogenous causes or population-based screening varies markedly between cancer
types, from comprehensive data to nothing relevant, and such circumstances account in large part for differences
between chapters in Section 5.
Where to from here?

All the research described in this World Cancer Report is calculated, directly or indirectly, to reduce the burden
of cancer, whether globally or in particular communities or for certain categories of people at risk. Typically, such
outcomes occur as a result of the adoption of certain policies, either by governments or by other competent au-
thorities. Then, many cancer-preventive options depend on individual decision-making and commitment. All such
matters are themselves amenable to research.
There is no generally operative procedure that determines the transition from cancer research findings to cancer
prevention policies. When such a pathway is charted for a particular innovation, the ease of its implementation will
be determined by many factors as they operate in particular countries or communities. In this context, World Cancer
Report is not designed as a vehicle for advocacy: research needs are not listed as such, nor are priorities specified.
The key role of cancer research in cancer prevention, as a record of achievement, is clear and unequivocal on
a global scale. Since the publication of World Cancer Report 2014, the burden of cancer attributable to obesity
and – separately – to pollution has been made clearer than ever before. More immediately in terms of the ultimate
goal of prevention, there is global progress in reducing tobacco-attributable cancers, at least in some countries.
And where once there was the goal of increased screening for cervical cancer, there is now, through vaccination,
the prospect of eliminating cervical cancer as a public health concern.
In short, “cancer research for cancer prevention” is not simply a way to describe a particular field of investigation.
Far more importantly, these words identify a pathway that may materially reduce the acknowledged burden of
cancer faced by humanity. There is, in fact, no other way.
Christopher P. Wild, Elisabete Weiderpass, and Bernard W. Stewart (Editors)
xiv
1 The global
cancer burden
As far as we know, cancer has always afflicted to high-income countries, has long been rec-
humans, although for centuries its relative im- ognized as a global scourge. Previously, low-
pact was overshadowed by early death from income countries primarily had a relatively
infectious diseases. Until recently, informa- high incidence of stomach, liver, and cervical
tion on the global distribution of cancer was cancer, but changes in incidence over time
limited for certain communities and countries. for these and other cancer types illustrate
We now have a reasonable basis for estimat- variation between countries. Finally, there are
ing the global cancer burden. For several tu- marked differences between countries or re-
mour types – colorectal, prostate, and breast gions in cancer mortality, with an increasing
cancer – high incidence rates were once re- burden in low- and middle-income countries,
stricted to North America, western Europe, attributable both to less-than-optimal imple-
and Australia, but now incidence rates are ris- mentation of preventive measures and to di-
ing in many other countries. Lung cancer, for agnosis at a later stage, rather than an early
which high incidence was initially restricted stage, of cancer development.
1.1
The burden and prevention of premature

deaths from noncommunicable diseases,
including cancer: a global perspective
Bochen Cao Bernard W. Stewart (reviewer)
Isabelle Soerjomataram Elisabete Weiderpass (reviewer)
Freddie Bray Christopher P. Wild (reviewer)
or medium levels of the Human NCDs have become the lead-

SUMMARY Development Index (HDI). ing cause of death worldwide and
pose a major threat to healthy
●● Cancer is the first or second ●● Feasible, affordable, and cost-
ageing, accounting for 72% of all
leading cause of premature effective interventions that re-
deaths globally in 2016 [1]. The
death (i.e. at ages 30–69 years) duce exposure to the key caus-
total of 40.5 million deaths from
in 134 of 183 countries, and it es and other risk factors for
NCDs globally in 2016 is a sharp in-
ranks third or fourth in an ad- cancer and for other noncom-
crease from the corresponding fig-
ditional 45 countries. municable diseases, increase
access to essential health-care ure of 31.6 million deaths in 2000.
●● Of the 15.2 million premature services, and ensure the avail- In 2016, about one third (15.2 mil-
deaths from noncommunicable ability of effective and afford- lion) of all deaths from NCDs oc-
diseases worldwide in 2016, able essential medicines and curred at ages 30–69 years. Of
4.5 million (29.8%) were due to vaccines are crucial for disease these premature deaths, 6.2 million
cancer. control globally. (40.8%) were due to cardiovascu-
lar diseases, 4.5 million (29.8%) to
●● The United Nations, within the cancer, 1.1 million (7.0%) to chronic
Sustainable Development Goals respiratory diseases, and 0.7 mil-
This chapter reviews the burden
agenda, has set a target to re- lion (4.5%) to diabetes [1].
and trends of premature mortality
duce the total premature mortal- These increasing trends in mor-
(i.e. deaths at ages 30–69 years)
ity from noncommunicable dis- tality from NCDs accompany the
from noncommunicable diseases
eases by one third by 2030. decline in mortality from infectious
(NCDs), with a focus on cancer,
●● Mortality rates from noncom- based on the WHO Global Health diseases, but they also result from
municable diseases, and can- Estimates that are available nation- the demographic and epidemiologi-
cer in particular, are declining ally by cause and year of death [1]. cal transitions that are taking place.
in most higher-income coun- When studying cancer patterns Demographic transition refers
tries, but such progress is lack- and trends, it is important to consid- to population-level shifts from a
er what constitutes human develop- pattern of high birth (fertility) rates
ing in lower-income countries,
ment, and how it may be measured. and high death (mortality) rates to
posing challenges in meeting
The Human Development Index one of low birth rates and low death
the Sustainable Development
(HDI) is a composite index of three rates. This shift increases the num-
Goals target.
basic dimensions of human de- ber of older adults, who are more
●● Attaining the goal of a reduc- velopment: a long and healthy life susceptible to ageing-related dis-
tion by one third in premature (based on life expectancy at birth), eases, including cancer, particular-
mortality from the four ma- education (based on average and ly in countries in transition [2].
jor types of noncommunica- expected years of schooling), and a Epidemiological transition re-
ble diseases would increase decent standard of living (based on fers to changes in mortality rates
the average expected years gross national income per capita). and causes of death that reflect
lived in the target age group The development levels of coun- underlying changes in exposure to
(30–69 years) by 0.64 years tries can be considered according risk factors. During the past centu-
worldwide, with larger gains to four tiers of HDI: low, medium, ry, a pattern of dominance of infec-
foreseen in countries with low high, and very high HDI. tious diseases has gradually been
16
SECTION 1
CHAPTER 1.1
replaced with one in which chronic tional 45 countries (Fig. 1.1.2) [1]. thetic changes to the physical envi-
or degenerative diseases, such as Specifically, cancer is currently the ronment, including structural condi-
NCDs, predominate. Within this leading cause of premature death tions that have impacts on mobility
diverse group of NCDs, the rela- in most of the countries with high and recreation, diet, and exposure
tive contribution to overall deaths or very high HDI, including Canada to environmental pollutants). The
has evolved with trends in mortality and the USA in North America, inherent disparities and widening
rates. For example, there have been Argentina and Chile in South gaps between and within countries
greater reductions in mortality rates America, most countries in Europe in levels of medical practice and
for cardiovascular diseases than in (including France, Germany, and health infrastructure also influence
those for cancer in many popula- the United Kingdom), Australia the diverging patterns and trends in
tions with medium or high HDI, and and New Zealand in Oceania, and cancer mortality [5–10].
the absolute and relative reductions Japan, the Republic of Korea, and In most countries with high HDI,
in cancer mortality rates have been Singapore in Asia. Cancer also cancer mortality rates are declining,
considerably larger in populations ranks first in Thailand and Viet primarily as a result of recent suc-
with very high HDI (Fig. 1.1.1) [3,4]. Nam. Cancer is the second lead- cesses in combating common can-
ing cause of premature death, after cer types through effective interven-
cardiovascular diseases, in Brazil, tions for prevention, early detection,
Cancer as a leading and treatment. In contrast, in coun-
China, and many countries in east-
cause of death ern Europe (including the Russian tries in transition, mortality rates are
worldwide Federation and Ukraine), as well as still increasing, or at best stabilizing,
In the past 60 years, better sanita- Algeria and Egypt. In most coun- for many cancer types, including
tion and the development of vac- tries in sub-Saharan Africa, cancer breast cancer, prostate cancer, and
cines and antibiotics have brought ranks third or fourth, and there are colorectal cancer [5,9,10].
about dramatic declines in mortal- only a few countries in this region in
ity from infectious diseases. With which cancer ranks fifth or sixth [1]. The Sustainable Develop-
improving primary and secondary Cancer is a complex disease, ment Goals target for
prevention for cardiovascular dis- for which the patterns and trends combating noncommuni-
eases, changing demographic and in mortality vary markedly between
risk factors have led to today’s ob- countries and across specific can-
cable diseases
servation that cancer is the first or cer types. These variations are due In response to the major threat that
second leading cause of premature to differences in changing lifestyles NCDs pose to sustainable human
death (i.e. at ages 30–69 years) and in local exposures to known or development, and to curb the rapid
in 134 of 183 countries, and it putative determinants, as well as an rise in NCDs worldwide, the United
ranks third or fourth in an addi- altering built environment (e.g. syn- Nations, within the Sustainable De
velopment Goals agenda, has set
an overarching target (Target 3.4)
to reduce the total premature mor-
Fig. 1.1.1. Changes between 1981–1985 and 2006–2010 in age-standardized mortal-
ity rates per 100 000 people, for ages 40–84 years in men and women combined, in tality from NCDs by one third by
populations with very high Human Development Index (HDI) and medium or high HDI. 2030 [11,12]. For the successful
realization of Target 3.4, a set of
health targets have been proposed
to reduce the exposure to risk fac-
tors for NCDs and to improve the
prevention and treatment of NCDs.
A subsequent reduction in prema-
ture deaths from NCDs would have
a profound effect on population lon-
gevity and an economic impact (see
Chapter 6.9).
If the goal of a reduction by one
third in premature mortality from the
four major types of NCDs is attained
in 2015–2030, the average expected
years lived in the target age group
(30–69 years) could potentially in-
crease by 0.64 years worldwide [13].
This figure ranges from 0.44 years in
countries with very high HDI to about
Chapter 1.1 • The burden and prevention of premature deaths from noncommunicable diseases, including cancer: a global perspective 17
0.70 years in countries with low or of the 20th century: 2.5–3.7 years deed meet this target. Using the his-
medium HDI (Fig. 1.1.3). Extending in countries with very high HDI and torical trends in premature mortality
the one third reduction in premature 1.1–1.4 years in countries with me- from the four major types of NCDs
mortality to all NCDs would lead to dium or high HDI. in the 15-year period between 2000
a further gain of 20% in average ex- Although attaining Target 3.4 of and 2015 as a reference, one ob-
pected years lived [13]. These are the Sustainable Development Goals serves that higher-income coun-
significant gains when considered in is a promising prospect for popula- tries are well on track to meeting
light of the increases in life expec- tion longevity in the long run, it is the target between 2015 and 2030,
tancy over the last three decades debatable whether countries will in- whereas lower-income countries
Fig. 1.1.2. Global map of cancer as a leading cause of premature death (i.e. at ages 30–69 years), indicating the rankings, with the
numbers of countries in parentheses.
No data Not applicable
The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever Data source: GHE 2016
on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, Map production: CSU
or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent approximate border lines World Health Organization
for which there may not yet be full agreement. © WHO 2018. All rights reserved
Fig. 1.1.3. Global map of estimated gains in average expected years lived (LE) between ages 30 years and 69 years if the Sustain
able Development Goals target of a reduction by one third in premature mortality from the four major types of noncommunicable
diseases is attained in 2015–2030.
No data Not applicable
The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever Data source: IARC
on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, Map production: CSU
or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent approximate border lines World Health Organization
18
for which there may not yet be full agreement. © WHO 2018. All rights reserved
SECTION 1
CHAPTER 1.1
still face considerable challenges. for cardiovascular diseases. There and other countries in central and
A similar picture is seen for can- is a clear need to prioritize preven- eastern Europe [12,15–17], leading
cer. In higher-income countries, tion strategies at the national level to high rates of premature mortality
a large part of the targeted reduc- and to structure health systems ac- from NCDs, including cancer.
tion has generally been attained. cordingly to manage the imminent However, lower-income coun-
In contrast, in low- and middle-in- epidemic of NCDs worldwide. tries face the additional burden
come countries the achievements A key and effective measure in of poverty-related NCDs, such as
are more limited (Fig. 1.1.4) [13]. the prevention of cancer and other infection-related cancers (including
It should be noted that the lack of NCDs is to reduce the exposure stomach cancer [see Chapter 5.4],
progress in lower-income countries to modifiable causes of NCDs, in- liver cancer [see Chapter 5.6], and
in 2000–2015 does not necessarily cluding several risk factors that cervical cancer [see Chapter 5.10]),
predict future failings in attaining contribute significantly to the oc- cardiovascular diseases due to fetal
the target in such populations in the currence of these diseases, such and childhood malnutrition, and res-
longer term, given that many NCDs as behavioural factors (e.g. tobacco piratory diseases that are correlated
can still be successfully prevented, use [see Chapter 2.1], harmful al- with a poor living environment [18,19].
treated, and managed. cohol consumption [see Chapter As countries progress societally and
The distinct patterns of causes 2.3], unhealthy diet, and physical economically, and epidemiological
of death help to prioritize approach- inactivity [see Chapter 2.7]), meta- transitions continue, the reduction
es to reduce mortality from specific bolic factors (e.g. high blood pres- in NCDs linked to poverty-related
major causes in a given country. sure, overweight and obesity, and factors is expected to be offset by
Specifically, cancer has surpassed high cholesterol level), and envi- increasing exposure to many behav-
cardiovascular diseases as the ronmental factors (e.g. air pollution ioural, environmental, and occupa-
leading cause of death in countries [see Chapter 2.9]), [12,14]. In many tional risk factors linked with indus-
with high or very high HDI. In con- middle-income countries, risk fac- trialized settings, including tobacco
trast, cardiovascular diseases re- tors for NCDs continue to prevail. use, harmful alcohol consumption,
main the leading cause of death in For example, the highest levels of and physical inactivity [20–26]. The
lower-income countries, largely be- smoking prevalence, harmful alco- path towards attaining Target 3.4 of
cause of inadequate and ineffective hol consumption, and high blood the Sustainable Development Goals
implementation of the available pre- pressure globally are observed in will be particularly challenging for
vention and treatment modalities countries of the former Soviet Union resource-constrained countries if
Fig. 1.1.4. Changes between 2000 and 2015 in the risk of dying from cancer at ages 30–69 years, for selected countries with low
or medium Human Development Index (HDI) and high or very high HDI.
Low / Medium HDI High / Very High HDI
14 13.9 14
13
-21%
12 -16% 12
Change in Risk of dying (%),30-69 years old
Change in Risk of dying (%),30-69 years old

11 11
10 10
9.2 9.3
8.7
-17% -23%
8 7.9 8
7.6 7.6 -21%
-16% 7.2
-16% 7.1
6.7 6.8
6.4 -23%
6 6
5.7 -3%
5.5 -1% 5.5
4 4
2 2
0 0
India Morocco Philippines South Uganda Sweden Brazil USA China Russian
Africa Federation
2000 2015 Reducing the risk

in 2000 by 1/3
Fig. 1.1.5. Dancers in Ayquina, Chile, illustrate the diversity of communities affected to the pace in higher-resource coun-
by cancer. In Chile, the incidence rates of cancer of the gall bladder are among the tries (Fig. 1.1.4) highlights the need
highest in the world. for accelerated actions to achieve
the Sustainable Development Goals
target in these countries. However,
inadequate access to affordable
primary care, early detection, and
treatment continues to be a bar-
rier to effective prevention and treat-
ment in these settings, leading to
poorer survival outcomes in patients
[12,17]. For example, whereas can-
cer surgery services are available in
95% of high-income countries, the
equivalent rate is only about 25%
in low-income countries [32], lead-
ing to substantially higher cancer
case fatality in lower-income coun-
tries (70%) than in higher-income
countries (45%) [33]. As part of the
Sustainable Development Goals tar-
gets, achieving universal health cov-
erage, including access to essential
health-care services and access to
their adoption of unhealthy lifestyles expected to contribute substantially effective and affordable essential
and activities with high environmen- to a reduction in premature deaths medicines and vaccines for NCDs
tal impact is not halted. Therefore, in from NCDs by 2030 [30,31]. Finally, for all, is crucial to ensure a narrow-
the coming decades it will be increas- establishing high-quality surveil- ing of the inequity gap and a reduc-
ingly critical to mitigate the rise in lance systems for cancer and oth- tion in mortality from NCDs globally.
NCDs in lower-income countries by er NCDs is imperative to plan and The potential for health improve-
preventing the adoption of unhealthy evaluate national responses to the ment is particularly striking in low-
behaviours (see Chapter 6.1) and en- Sustainable Development Goals tar- and middle-income countries, if the
suring that environmental actions are get [29]. prompt adoption of “best buys” in-
sustainable [27,28]. The slow pace of progress in re- terventions leads to the Sustainable
To curb the rising burden of source-limited countries that are un- Development Goals target being
NCDs, WHO proposed a “best dergoing major transitions, relative met, because in these countries
buys” package to facilitate inter-
ventions that are feasible, afford-
Fig. 1.1.6. The disparities that are evident within many countries are illustrated in this
able, and cost-effective [12,29]. An
view of Manila, Philippines.
extended list of options to reduce
the prevalence of tobacco smok-
ing, harmful alcohol consumption,
unhealthy diet, and physical in-
activity as well as environmental
action, for example to reduce air
pollution, are essential elements
to control NCDs, including cancer.
Furthermore, measures proposed
by the WHO “best buys” and by the
“essential package” of interventions
presented in the third edition of
Disease Control Priorities – includ-
ing implementing vaccination pro-
grammes, extending the preventive
and early detection measures for
cancer at the primary care level,
and improving access to services
for cancer and other NCDs – are
20
SECTION 1
CHAPTER 1.1
NCDs commonly rank higher as a comes, the additional societal and hence increasing workplace produc-
cause of death. A parallel impact economic potential of these interven- tivity and reducing costs of health
across the four major types of NCDs tions for NCDs is large, because the care and social care. Ultimately,
is expected, with a marked reduc- targeted decline in mortality would these potential benefits provide fur-
tion in cancer mortality rates, many bring about a substantial increase ther arguments for implementing ac-
of which have stagnated nationally. in the number of person-years lived tions aimed at reducing the global
In addition to improved health out- in the most productive age groups, burden of NCDs.
References
1. WHO (2018). Global health estimates 2016: 9. Siegel R, DeSantis C, Jemal A (2014). 16. Danaei G, Finucane MM, Lin JK, Singh
deaths by cause, age, sex, by country and by Colorectal cancer statistics, 2014. CA GM, Paciorek CJ, Cowan MJ, et al.;
region, 2000–2016. Geneva, Switzerland: Cancer J Clin. 64(2):104–17. https://doi. Global Burden of Metabolic Risk Factors
World Health Organization. Available from: org/10.3322/caac.21220 PMID:24639052 of Chronic Diseases Collaborating Group
https://www.who.int /healthinfo/global_ (Blood Pressure) (2011). National, re-
10. Center MM, Jemal A, Lortet-Tieulent J,
burden_disease/en/. gional, and global trends in systolic
Ward E, Ferlay J, Brawley O, et al. (2012).
blood pressure since 1980: systematic
2. Caldwell JC (1976). Toward a restate- International variation in prostate cancer
incidence and mortality rates. Eur Urol. analysis of health examination surveys
ment of demographic transition theory.
61(6):1079–92. https://doi.org/10.1016/j. and epidemiological studies with 786
Popul Dev Rev. 2(3/4):321–66. https://doi.
eururo.2012.02.054 PMID:22424666 country-years and 5.4 million partici-
org/10.2307/1971615
pants. Lancet. 377(9765):568–77. https://
3. Kanavos P (2006). The rising bur- 11. United Nations General Assembly (2015). doi.org/10.1016/S0140-6736(10)62036-3
den of cancer in the developing world. Transforming our world: the 2030 Agenda PMID:21295844
Ann Oncol. 17(Suppl 8):viii15–viii23. for Sustainable Development. Available
from: https://sustainabledevelopment.un. 17. Di Cesare M, Khang Y-H, Asaria P, Blakely
https://doi.org/10.1093/annonc/mdl983
org/post2015/transformingourworld. T, Cowan MJ, Farzadfar F, et al.; Lancet
PMID:16801335
NCD Action Group (2013). Inequalities in
12. WHO (2014). Global status report on non- non-communicable diseases and effec-
4. Omran AR (1971). The epidemiologic
communicable diseases 2014. Geneva, tive responses. Lancet. 381(9866):585–
transition. A theory of the epidemiolo-
Switzerland: World Health Organization. 97. https://doi.org/10.1016/S0140-6736
gy of population change. Milbank Mem
Available from: https://apps.who.int/iris/ (12)61851-0 PMID:23410608
Fund Q. 49(4):509–38. https://doi.
handle/10665/148114.
org/10.2307/3349375 PMID:5155251
18. Ezzati M, Riboli E (2012). Can noncom-
13. Cao B, Bray F, Ilbawi A, Soerjomataram
5. DeSantis CE, Bray F, Ferlay J, Lortet- municable diseases be prevented?
I (2018). Effect on longevity of one-third
Tieulent J, Anderson BO, Jemal A (2015). Lessons from studies of populations and
reduction in premature mortality from
International variation in female breast can- individuals. Science. 337(6101):1482–7.
non-communicable diseases by 2030:
cer incidence and mortality rates. Cancer https://doi.org/10.1126/science.1227001
a global analysis of the Sustainable
Epidemiol Biomarkers Prev. 24(10):1495– Development Goal health target. Lancet PMID:22997325
506. https://doi.org/10.1158/1055-9965. Glob Health. 6(12):e1288–e1296. https:// 19. de Martel C, Ferlay J, Franceschi S, Vignat
EPI-15-0535 PMID:26359465 doi.org/10.1016/S2214-109X(18)30411-X J, Bray F, Forman D, et al. (2012). Global
PMID:30420032 burden of cancers attributable to infec-
6. Torre LA, Bray F, Siegel RL, Ferlay J,
Lortet-Tieulent J, Jemal A (2015). Global 14. GBD 2015 Risk Factors Collaborators tions in 2008: a review and synthetic anal-
cancer statistics, 2012. CA Cancer J Clin. (2016). Global, regional, and national ysis. Lancet Oncol. 13(6):607–15. https://
65(2):87–108. https://doi.org/10.3322/caac. comparative risk assessment of 79 be- doi.org/10.1016/S1470-2045(12)70137-7
21262 PMID:25651787 havioural, environmental and occupation- PMID:22575588
al, and metabolic risks or clusters of risks,
7. Torre LA, Siegel RL, Ward EM, Jemal 20. WHO (2017). WHO report on the global
1990–2015: a systematic analysis for the
A (2014). International variation in lung Global Burden of Disease Study 2015. tobacco epidemic, 2017: monitoring tobac-
cancer mortality rates and trends among Lancet. 388(10053):1659–724. https:// co use and prevention policies. Geneva,
women. Cancer Epidemiol Biomarkers doi.org/10.1016/S0140-6736(16)31679-8 Switzerland: World Health Organization.
Prev. 23(6):1025–36. https://doi.org/10.1158/ PMID:27733284 Available from: https://apps.who.int/iris/
1055-9965.EPI-13-1220 PMID:24836468 handle/10665/255874.
15. Ezzati M, Lopez AD, Rodgers A, Vander
8. Ferro A, Peleteiro B, Malvezzi M, Bosetti Hoorn S, Murray CJ; Comparative 21. Singh T, Arrazola RA, Corey CG, Husten
C, Bertuccio P, Levi F, et al. (2014). Risk Assessment Collaborating Group CG, Neff LJ, Homa DM, et al. (2016).
Worldwide trends in gastric cancer mortal- (2002). Selected major risk factors and Tobacco use among middle and high
ity (1980-2011), with predictions to 2015, global and regional burden of disease. school students – United States, 2011–
and incidence by subtype. Eur J Cancer. Lancet. 360(9343):1347–60. https://doi. 2015. MMWR Morb Mortal Wkly Rep.
50(7):1330–44. https://doi.org/10.1016/j. org /10.1016/S 0140 - 673 6(02)11403 - 6 65(14):361–7. https://doi.org/10.15585/
ejca.2014.01.029 PMID:24650579 PMID:12423980 mmwr.mm6514a1 PMID:27077789
22. Samet JM, Yoon S-Y, editors (2001). 27. Preston SH (1976). Mortality patterns in 32. WHO (2016). Assessing national ca-
Women and the tobacco epidemic: chal- national populations: with special refer- pacity for the prevention and control
lenges for the 21st century. Geneva, ence to recorded causes of death. New of noncommunicable diseases: report
Switzerland: World Health Organization. York (NY), USA: Academic Press. of the 2015 global survey. Geneva,
Available from: https://apps.who.int/iris/ Switzerland: World Health Organization.
handle/10665/66799. 28. Salomon JA, Murray CJL (2002). The epi-
demiologic transition revisited: composi- Available from: https://apps.who.int/iris/
23. Guthold R, Louazani SA, Riley LM, Cowan tional models for causes of death by age and handle/10665/246223.
MJ, Bovet P, Damasceno A, et al. (2011). sex. Popul Dev Rev. 28(2):205–28. https:// 33. Ferlay J, Ervik M, Lam F, Colombet M,
Physical activity in 22 African countries: doi.org/10.1111/j.1728-4457.2002.00205.x
Mery L, Piñeros M, et al. (2018). Global
results from the World Health Organization
29. WHO (2013). Global action plan for the Cancer Observatory: Cancer Today. Lyon,
STEPwise approach to chronic disease
prevention and control of noncommu- France: International Agency for Research
risk factor surveillance. Am J Prev Med.
41(1):52–60. https://doi.org/10.1016/j. nicable diseases 2013–2020. Geneva, on Cancer. Available from: https://gco.iarc.
amepre.2011.03.008 PMID:21665063 Switzerland: World Health Organization. fr/today.
Available from: https://apps.who.int/iris/
24. Pratt M, Sarmiento OL, Montes F, Ogilvie handle/10665/94384.
D, Marcus BH, Perez LG, et al.; Lancet
Physical Activity Series Working Group 30. Gelband H, Jha P, Sankaranarayanan
(2012). The implications of megatrends R, Horton S, editors (2015). Disease
in information and communication tech- control priorities. 3rd ed. Vol. 3, Cancer.
nology and transportation for changes Washington (DC), USA: World Bank.
in global physical activity. Lancet. 380 31. Gelband H, Sankarayanarayanan S,
(9838):282–93. https://doi.org/10.1016/ Gauvreau CL, Horton S, Anderson BO,
S0140-6736(12)60736-3 PMID:22818940 Bray F, et al.; Disease Control Priorities-3
25. Bauman AE, Reis RS, Sallis JF, Wells JC, Cancer Author Group (2016). Costs,
Loos RJF, Martin BW; Lancet Physical affordability, and feasibility of an es-
Activity Series Working Group (2012). sential package of cancer control in-
Correlates of physical activity: why are terventions in low-income and middle-
some people physically active and others income countries: key messages from
not? Lancet. 380(9838):258–71. https:// Disease Control Priorities, 3rd edition.
doi.org/10.1016/S0140-6736(12)60735-1 Lancet. 387(10033):2133–44. https://
PMID:22818938 doi.org/10.1016/S0140-6736(15)00755-2
PMID:26578033
26. WHO (2014). Global status report on
alcohol and health – 2014. Geneva,
Switzerland: World Health Organization.
Available from: https://apps.who.int/iris/
handle/10665/112736.
22
SECTION 1
CHAPTER 1.2
1.2
Global trends in cancer incidence

and mortality
Isabelle Soerjomataram Bernard W. Stewart (reviewer)

Freddie Bray Elisabete Weiderpass (reviewer)
Christopher P. Wild (reviewer)
●● An increase in prostate can- This chapter reviews the incidence

SUMMARY cer incidence rates followed by and mortality trends for the six
a decline, as observed in the most common cancer types world-
●● In men, lung cancer incidence USA, is attributable to prostate- wide (lung cancer, breast cancer,
and mortality rates vary across specific antigen (PSA) testing. colorectal cancer, prostate cancer,
countries and are almost invari- In several countries in Asia and stomach cancer, and cervical can-
ably correlated with the preva- cer) and the main determinants of
Latin America, incidence rates
lence of tobacco smoking 20– these trends, including the role of
increased substantially and then
30 years earlier. In women, the the changing prevalence and distri-
stabilized. Mortality rates have
smoking epidemic typically be- bution of key risk factors as well as
been declining in most countries.
gan later, or – in some countries – the impact of preventive, screening,
not at all, and this is reflected in ●● Worldwide, stomach cancer and therapeutic interventions.
the corresponding rates. ranks fifth in terms of incidence IARC is responsible for the com-
and third in terms of mortality. pilation, estimation, and reporting of
●● Rising breast cancer incidence
Incidence and mortality rates cancer statistics generated through
rates are correlated with trends
of stomach cancer (predomi- flagship projects and databases,
towards earlier ages at menar-
nantly the non-cardia type) are including Cancer Incidence in Five
che, later ages at first birth, and
decreasing, whereas incidence Continents (http://ci5.iarc.fr) and
lower parity. In many countries
with high levels of the Human of cancer of the cardia region GLOBOCAN, for which the result-
Development Index (HDI), inci- of the stomach is increasing ing statistics are disseminated on the
dence rates have stabilized and in several populations. Most Global Cancer Observatory, an inter-
mortality rates are declining, cases of stomach cancer are active, user-friendly, and data-driven
whereas in countries in transi- attributable to infection with online interface (http://gco.iarc.fr).
tion towards higher HDI levels, Helicobacter pylori. The primary source for this chap-
mortality trends have tended ter is the cancer incidence trends
●● Cervical cancer incidence and from successive volumes of Cancer
to parallel the increasing inci- mortality rates have declined Incidence in Five Continents, the
dence trends. in most countries in recent de- compendium of data sets from na-
●● Incidence rates of colorectal can- cades, as a result of the detec- tional or subnational high-quality
cer have increased in countries tion of precancerous lesions by population-based cancer registries.
in transition, whereas in coun- screening, but increasing rates Equivalent data on cancer mortality
tries with high HDI, rates have have been observed in younger trends were obtained from the na-
either stabilized or decreased. generations of women in some tional statistics compiled in the WHO
However, incidence is increasing countries. Global elimination of Mortality Database (https://www.who.
in younger age groups and in re- the disease – in terms of cervi- int/healthinfo/mortality_data/en/).
cent generations in a diverse set cal cancer no longer being con- This chapter also makes refer-
of countries. Mortality rates have sidered a public health problem ence to the current global burden of
decreased in countries with high – is attainable during this centu- the six most common cancer types
HDI; mortality rates are increas- ry through HPV vaccination and using the GLOBOCAN 2018 esti-
ing in many low- and middle-in- screening programmes. mates of incidence and mortality,
come countries. which are provided for 185 countries
Chapter 1.2 • Global trends in cancer incidence and mortality 23

or territories worldwide on the Cancer lowed by Australia, New Zealand, Development Index (HDI), trends in
Today subsite of the Global Cancer and Canada), were also the first rates are largely stable over time,
Observatory (http://gco.iarc.fr/today). countries in which the prevalence of reflecting either that smoking is not
smoking decreased, followed about being taken up or that the smoking
20–30 years later by a decline in epidemic is at too early a stage to
Lung cancer
lung cancer incidence and mortali- be visible in the lung cancer trends.
Lung cancer is the most common ty rates (Fig. 1.2.1). In world regions The trends by histological sub-
cancer type worldwide in terms of where lung cancer rates have his- type present a somewhat different
both incidence (2.1 million new cas- torically been low (e.g. Costa Rica, picture. Incidence rates of squa-
es in 2018) and mortality (1.8 million Ecuador, and India) or intermedi- mous cell carcinoma of the lung
deaths in 2018). The key cause of ate (e.g. Japan and Turkey), lung are currently decreasing (at least
lung cancer is tobacco smoking (see cancer incidence in men appears in men), whereas rates of adeno-
Chapter 2.1), which is responsible for to have recently stabilized or in- carcinoma of the lung are rising in
63% of overall global deaths from creased (e.g. Thailand). some populations (particularly in
lung cancer and for more than 90% In women, the tobacco smok- women) [2]. In men, squamous cell
of lung cancer deaths in countries ing habit has commonly been ac- carcinoma was previously the most
where smoking is prevalent in both quired more recently, or – in some common lung cancer subtype, but
sexes [1]. Therefore, trends in lung countries – not at all. Therefore, by the end of the 1990s a shift had
cancer incidence and mortality are the most common trend is of ris- occurred and adenocarcinoma was
determined largely by past exposure ing lung cancer rates, as observed the most common subtype. In wom-
to tobacco smoking, reflecting the dif- in Australia, Japan, the United en, this effect is delayed, meaning
ferential evolution of the smoking epi- Kingdom, and the USA, with a peak that in many countries with high
demic by sex in individual countries. and a recent decline that are most HDI, incidence rates of adenocarci-
In men, the countries where the evident in the United Kingdom and noma of the lung are now decreas-
smoking epidemic first began (the the USA (Fig. 1.2.2). In many coun- ing in men and are still increasing in
United Kingdom and the USA, fol- tries with lower levels of the Human women (see Chapter 5.1).
Fig. 1.2.1. Age-standardized (World) (a) incidence rates and (b) mortality rates per 100 000 person-years by calendar year in
selected countries for lung cancer in men, circa 1975–2012. Asterisks indicate regional registries (other registries are national).
24
SECTION 1
CHAPTER 1.2
selected countries for lung cancer in women, circa 1975–2012. Asterisks indicate regional registries (other registries are national).
Lung cancer survival remains low reduce the prevalence of tobacco several countries in Asia (e.g. India,
globally. The fact that lung cancer is smoking remains a key pillar in dis- Japan, Thailand, and Turkey) and in
the leading cause of cancer death ease control. Latin America (e.g. Costa Rica and
has motivated the assessment of Ecuador) (Fig. 1.2.3a).
the benefits of lung cancer screen- Artefactual factors may inflate
Breast cancer
ing, i.e. low-dose computed tomog- incidence. Breast cancer screening
raphy (CT), among heavy smokers. Breast cancer is the most com- captures prevalent cases for a few
A 16% reduction in lung cancer mor- monly diagnosed cancer in women years after implementation of screen-
tality among those screened in a (2.1 million new cases in 2018) and ing, and the reported increases in in-
large trial in the USA [3] has led to the leading cause of cancer death cidence in Brazil and Mexico of 2.9%
the recommendation of lung cancer in women globally (627 000 deaths and 5.9% per year, respectively, were
screening in the USA, followed by in 2018) (see Chapter 5.9) [5]. greatest among women aged 55–
similar recommendations in Europe The rising incidence rates ob- 64 years, the targeted screening age
[4]. However, controversy still exists, served in many higher-income group [6]. In contrast, in countries
because the current short-term tri- countries during the past five de- with high HDI (e.g. Australia, Canada,
als have not shown any beneficial cades – and in lower-income the United Kingdom, and the USA),
impact on deaths [3]; further re- countries more recently – can be incidence rates have stabilized after
sults and a complete assessment attributed partly to the changing a marked decline in incidence start-
of the long-term costs, benefits, and prevalence and distribution of sev- ing in about 2000, which is consid-
harms are needed before the imple- eral reproductive and hormonal ered to result from the publication of
mentation of national programmes factors (see Chapter 3.6), includ- two landmark studies that reported
(see Chapter 6.6). ing a trend towards earlier ages at on the harmful effects of menopau-
Given that tobacco smoking is menarche, later ages at first birth, sal hormone replacement therapy
a major contributor to the burden of and lower parity [6]. These changes on breast cancer risk (see Chapter
multiple cancer types and chronic may partly explain the rapid rises 2.11) [7]. Dietary factors (including
diseases, primary prevention to in breast cancer incidence rates in an increasing prevalence of alcohol

selected countries for breast cancer in women, circa 1975–2012. Asterisks indicate regional registries (other registries are national).
consumption in women), obesity, and the earlier detection of breast cancer part explain these favourable trends,
physical inactivity (see Chapter 2.7) through earlier diagnosis and effec- the marked decline of rates in non-
cannot be ruled out as potential con- tive screening programmes may in screened age groups indicates the
tributors to the previous rising trends
in these countries with high HDI, be-
Fig. 1.2.4. Women in Peru wearing traditional dress. In Peru and in many other
cause rates also increased in women countries in transition, breast cancer mortality trends have tended to parallel the
outside of the targeted screening age increasing incidence trends.
group [6].
In countries in transition towards
higher HDI levels, breast cancer
mortality trends have tended to par-
allel the increasing incidence trends;
rising mortality rates have consis-
tently been observed in countries in
Asia and Latin America (Fig. 1.2.3b)
[8], for all age groups and also for
women in the targeted screening age
group (which suggests an absence
of effective screening programmes).
In contrast, a steady decline in
breast cancer mortality has been ob-
served in numerous countries with
high HDI [8,9], including Australia,
Canada, and the USA, where breast
cancer mortality rates declined by
18–22% from 2002 to 2012. Although
26
SECTION 1
CHAPTER 1.2
importance of multiple improvements the United Kingdom, and the USA (characterized by low consumption
in the management and treatment of are observed predominantly in older of fruits, vegetables, and fibre and
the disease. age groups (55 years and older); high consumption of red meat and
these populations are subject to ear- processed meat [see Chapter 2.6]),
ly detection programmes that detect
Colorectal cancer a lack of physical activity, and an in-
and remove precancerous colorectal creasing prevalence of overweight
Colorectal cancer is the third most polyps, leading to a decline in malig- and obesity (see Chapter 2.7).
common cancer in both sexes world- nancies [11]. Other factors may have Consistent with the declines in
wide (1.8 million new cases in 2018). contributed, including the adoption incidence, colorectal cancer mor-
It ranks second in terms of mortality of preventive therapies such as reg-
(880 000 deaths in 2018). The fact tality rates have decreased in coun-
ular use of aspirin, postmenopausal tries with high HDI (e.g. Australia,
that mortality is considerably lower estrogen therapy, or – as a matter of
than incidence reflects the relatively Canada, the United Kingdom, and
greater speculation – an increasing
good prognosis for cases on aver- the USA) in both sexes (Fig. 1.2.5b
intake of vitamin D [12].
age (see Chapter 5.5). and Fig. 1.2.6b). These decreases
However, marked increases in
In general, in countries in transi- can be linked partly to improving
incidence in younger age groups
tion, where overall risk of colorectal have been observed in countries survival through the adoption of
cancer has typically been low, inci- with higher HDI and are now also best practices in cancer treatment
dence rates have increased, whereas observed in recent birth cohorts in and management, in addition to ear-
in countries with high HDI, where risk Asia (e.g. in Japan, Thailand, and lier detection of colorectal cancer in
of colorectal cancer tends to be rela- Turkey) and in Latin America (e.g. in these countries [10]. The contrasting
tively high, incidence rates have either Costa Rica and Ecuador). The ris- increases in mortality rates in several
stabilized or decreased in both sexes ing risk is seen in successive gen- countries in Asia and Latin America
(Fig. 1.2.5a and Fig. 1.2.6a) [10]. erations, implying the importance of may reflect the limited health infra-
As an example, the declining in- changing risk factors; these are still structure and poorer access to early
cidence trends in Australia, Canada, ill-defined but may include poor diet detection and treatment [10].
selected countries for colorectal cancer in men, circa 1975–2012. Asterisks indicate regional registries (other registries are national).

Fig. 1.2.6. Age-standardized (World) (a) incidence rates and (b) mortality rates per 100 000 person-years by calendar year in selected
countries for colorectal cancer in women, circa 1975–2012. Asterisks indicate regional registries (other registries are national).
Cancer survival is highly depen- ing for 360 000 deaths (6.7% of the publication of the results of two
dent on the stage of cancer at diag- cancer deaths in men) in 2018 (see large randomized trials [16,17] and
nosis, and the unfavourable stage Chapter 5.13). a broad consensus to cease the
distribution of colorectal cancer Until the mid-1990s, prostate testing of men older than 75 years.
partly explains the higher excess cancer incidence rates in the USA Where increases in incidence rates
mortality from this cancer in a given were increasing substantially, which have been observed, competing
region [13]. Furthermore, the com- was largely attributed to the intro- explanations may include greater
plexity of treatment, which requires duction of prostate-specific antigen population awareness of the dis-
a combination of chemotherapy and (PSA) testing as a diagnostic test ease, the diagnosis of small and
radiotherapy (for rectal cancers) for asymptomatic prostate cancers latent cancers through PSA test-
after major surgery, can further [14]. This increase was followed by ing, or a genuine increase in the
complicate adequate management a peak and a subsequent decline incidence rates of invasive pros-
of colorectal cancer. In the future, by 2000. Similar time trends were tate cancer. A changing lifestyle
improved access to earlier cancer observed in Australia and Canada, has been proposed as one of the
detection and treatment may de- with a later decline in incidence drivers of trends, including an in-
crease the evident inequalities in rates (Fig. 1.2.7a). Similar trends of creased prevalence of obesity and
colorectal cancer survival globally. incidence rates that increased sub- increased consumption of dairy
stantially and then stabilized were products and calcium, but these
Prostate cancer observed in several countries in Asia factors confer only a small or mini-
Prostate cancer is now the second (e.g. Turkey) and Latin America (e.g. mal increase in risk [14]. Prostate
most common cancer in men world- Costa Rica and Ecuador) [14,15]. cancer incidence rates are much
wide, with an estimated 1.3 million Where incidence rates have de- higher in Black populations, which
new cases in 2018, accounting for creased or stabilized, these trends points to a role of genetic factors,
13.5% of new cancer cases in men. may have resulted partly from a although it is unlikely that such fac-
It is a somewhat less important decline in PSA testing in general tors explain much of the time trends
cause of cancer mortality, account- practice and among urologists after observed in different populations.
28
SECTION 1
CHAPTER 1.2
selected countries for prostate cancer, circa 1975–2012. Asterisks indicate regional registries (other registries are national).
In contrast to incidence rates, cancer mortality rates. A better un- (Fig. 1.2.8). Trends in women (not
prostate cancer mortality rates derstanding of the causes and fac- shown) are similar to those in men,
have largely been declining in most tors that affect incidence is urgently but the rates are generally lower.
countries, with the exception of needed to inform future prevention The risk of non-cardia stomach
Thailand, where rates have consis- strategies. cancer is closely related to infection
tently been low (Fig. 1.2.7b). The with Helicobacter pylori; 75–90% of
two main factors causing the ob- Stomach cancer all stomach cancer cases can be
served decline in mortality rates are attributed to infection with this bac-
In the first systematic collation of
probably a stage shift in prostate terium (see Chapter 2.2) [20]. H.
global high-quality cancer incidence
cancer related to PSA testing (i.e. pylori infection is generally acquired
data, in the 1960s, stomach cancer
more cancers are detected at an was the most common cancer type at a young age. The risk of infection
earlier stage) and better manage- worldwide [19]. Stomach cancer is is increased by overcrowding, and
ment of patients diagnosed with the now the fifth most common cancer therefore stomach cancer is strongly
disease [18]. The rather short lead type globally, with an estimated 1 mil- associated with low socioeconomic
time from the observed decline in lion new cases in 2018 (5.7% of new status. The declining rates of stom-
incidence and mortality has brought cancer cases), but because survival ach cancer have been attributed
considerable controversy with reis poor, stomach cancer ranks third in partly to improved living conditions,
gard to the beneficial impact of PSA terms of mortality (783 000 deaths in in particular among young cohorts.
testing on prostate cancer mortal- 2018) (see Chapter 5.4) [5]. Furthermore, improved food pres-
ity. The causes of the decline are A key epidemiological finding is ervation practices and better nutri-
probably manifold, including earlier the steady decline in incidence and tion, including refrigeration for the
detection and improved treatment; mortality rates of stomach cancer transportation and storage of food,
also, greater specificity and less (predominantly the non-cardia type have been suggested as leading to a
misclassification of earlier deaths of stomach cancer) that has consis- declining trend (see Chapter 2.8) [7].
from prostate cancer may have tently been observed over more than In Japan and the Republic of
led to a slight downturn in prostate five decades across all world regions Korea – countries that have some of

selected countries for stomach cancer in men, circa 1975–2012. Asterisks indicate regional registries (other registries are national).
the highest stomach cancer rates – Within the next decade, results from an increasing incidence of cardia
part of the decline has been linked to these randomized trials may provide stomach cancer (which accounted
the national screening programmes further insights to decrease the cur- for 27% of all stomach cancer cases
that have been implemented over the rent uncertainties about H. pylori in 2012 [22]) in several populations
past few decades [21]. Randomized screening and treatment. [23]. This increase has been linked to
trials are under way to assess the In contrast to the overall de- the increased prevalence of Barrett
impact of H. pylori eradication on cline in rates of non-cardia stom- oesophagus and adenocarcinoma
non-cardia stomach cancer [21]. ach cancer, studies have indicated of the lower third of the oesophagus,
which are strongly associated with
Fig. 1.2.9. Shibuya Crossing in Tokyo, Japan. Although incidence rates of stomach overweight and obesity. This double
cancer are declining in almost all countries worldwide, Japan has one of the highest burden of infection-related and obe-
recorded rates. sity-related stomach cancer calls
for targeted public health actions
that tackle the emerging divergence
in the burden and trends observed
across the world.
Cervical cancer
Cervical cancer is the fourth most
common cancer type in women
worldwide in terms of both incidence
and mortality, with an estimated
570 000 new cases and 311 000
deaths in 2018 [5]. Infection with hu-
man papillomavirus (HPV) — nota-
bly HPV types 16, 18, 31, and 45 —
30
SECTION 1
CHAPTER 1.2
selected countries for cervical cancer, circa 1975–2012. Asterisks indicate regional registries (other registries are national).
is an established cause of the dis- applies to the Baltic countries, parts Conclusions
ease and is estimated to cause of eastern Europe and western Asia
This brief overview of global inci-
all cases of cervical cancer (see [29], and Japan, where the effect
dence and mortality trends for six
Chapter 5.10) [24]. has been occurring for an extended major cancer types in a subset of
Incidence and mortality rates of period [30], in the absence of effec- countries is based on the availabil-
cervical cancer have consistently de- tive screening programmes. Other ity of recent data from national or
clined in most countries in the past determinants have contributed to the subnational population-based can-
few decades (Fig. 1.2.10) [25,26], declines in cervical cancer rates in cer registries and/or national vital
and rates appear to have stabilized countries without effective screen- registration systems. Local high-
in many countries with high HDI ing programmes, including improved quality cancer surveillance systems
(e.g. Australia, Canada, the United genital hygiene and the impact of co- are needed to gain a reasonably
Kingdom, and the USA), where de- factors linked to progression of HPV accurate picture of how the cancer
clines have been ascribed to the infection to cervical cancer: parity, burden and risk are changing with
success of cytology-based screen- age at first birth, use of oral contra- time in different communities. The
ing programmes [26]. However, sev- ceptives, and tobacco use. focus on rates for all ages has pre-
eral studies have shown that, within A recent WHO call to action cluded a more detailed exposition
the overall decline in incidence and seeks to overcome the multiple chal- of trends by age and birth cohort,
mortality rates, increases have been lenges to global cervical cancer pre- which is needed to fully understand
observed in the younger generations vention by scaling up HPV vaccina- the underlying factors responsible
of women in some countries, such tion (see Chapter 6.3) and screening for these time trends.
as Finland [27] and the Netherlands programmes in countries to eliminate Evidently there are increasing
[28]. The general consensus is that cervical cancer as a public health global inequalities in cancer control
these trends relate to changes in concern during this century (https:// planning and outcomes. Although
sexual behaviour and increased www.who.int/reproductivehealth/ there have been many triumphs in
transmission of persistent HPV in- cervical-cancer-public-health-con the prevention, early diagnosis, and
fection among birth cohorts. This cern/en/). management of these major cancer

types in recent decades, those ben- from cancer becomes a reality in all action plan 2013–2020, in which in-
efits have occurred predominantly countries of the world within the first terventions are rated with reference
in countries with higher HDI, where half of this century, it is paramount to “best buys” – are implemented
health systems infrastructure and that the existing evidence-based and their success evaluated equita-
capacity are already in place. To en- and cost-effective interventions – bly in lower-resource settings.
sure that the potential for prevention, such as those listed in the updated
cure, and alleviation of suffering Appendix 3 [31] of the WHO global
References
1. GBD 2015 Risk Factors Collaborators 8. Carioli G, Malvezzi M, Rodriguez T, 16. Andriole GL, Crawford ED, Grubb RL 3rd,
(2016). Global, regional, and national Bertuccio P, Negri E, La Vecchia C (2018). Buys SS, Chia D, Church TR, et al.; PLCO
comparative risk assessment of 79 be- Trends and predictions to 2020 in breast Project Team (2009). Mortality results from
havioural, environmental and occupation- cancer mortality: Americas and Australasia. a randomized prostate-cancer screen-
al, and metabolic risks or clusters of risks, Breast. 37:163–9. https://doi.org/10.1016/j. ing trial. N Engl J Med. 360(13):1310–9.
1990–2015: a systematic analysis for the breast.2017.12.004 PMID:29246526 https://doi.org/10.1056/NEJMoa0810696
Global Burden of Disease Study 2015. PMID:19297565
Lancet. 388(10053):1659–724. https:// 9. Carioli G, Malvezzi M, Rodriguez T,
doi.org/10.1016/S0140-6736(16)31679-8 Bertuccio P, Negri E, La Vecchia C 17. Schröder FH, Hugosson J, Roobol MJ,
PMID:27733284 (2017). Trends and predictions to 2020 in Tammela TLJ, Ciatto S, Nelen V, et al.;
breast cancer mortality in Europe. Breast. ERSPC Investigators (2009). Screening
2. Lortet-Tieulent J, Soerjomataram I, Ferlay and prostate-cancer mortality in a ran-
36:89–95. https://doi.org/10.1016/j.breast.
J, Rutherford M, Weiderpass E, Bray F domized European study. N Engl J Med.
2017.06.003 PMID:28988610
(2014). International trends in lung cancer 360(13):1320–8. https://doi.org/10.1056/
incidence by histological subtype: adeno- 10. Arnold M, Sierra MS, Laversanne M, NEJMoa0810084 PMID:19297566
carcinoma stabilizing in men but still in- Soerjomataram I, Jemal A, Bray F (2017).
creasing in women. Lung Cancer. 84(1):13– Global patterns and trends in colorec- 18. Wong MC, Goggins WB, Wang HHX,
22. https://doi.org/10.1016/j.lungcan.2014. tal cancer incidence and mortality. Gut. Fung FDH, Leung C, Wong SYS, et al.
01.009 PMID:24524818 (2016). Global incidence and mortality for
66(4):683–91. https://doi.org/10.1136/gut
jnl-2015-310912 PMID:26818619 prostate cancer: analysis of temporal pat-
3. Coureau G, Salmi LR, Etard C, Sancho-
terns and trends in 36 countries. Eur Urol.
Garnier H, Sauvaget C, Mathoulin-Pélissier
11. Siegel RL, Miller KD, Fedewa SA, Ahnen 70(5):862–74. https://doi.org/10.1016/j.
S (2016). Low-dose computed tomography
DJ, Meester RGS, Barzi A, et al. (2017). eururo.2016.05.043 PMID:27289567
screening for lung cancer in populations
highly exposed to tobacco: a systematic Colorectal cancer statistics, 2017. CA
19. Doll R, Payne P, Waterhouse JAH, edi-
methodological appraisal of published ran- Cancer J Clin. 67(3):177–93. https://doi.
tors (1966). Cancer incidence in five
domised controlled trials. Eur J Cancer. org/10.3322/caac.21395 PMID:28248415
continents: a technical report. Berlin,
61:146–56. https://doi.org/10.1016/j.ejca. Germany: Springer.
12. Fidler MM, Bray F, Vaccarella S,
2016.04.006 PMID:27211572
Soerjomataram I (2017). Assessing glob-
20. Plummer M, Franceschi S, Vignat J,
4. Oudkerk M, Devaraj A, Vliegenthart R, al transitions in human development and
Forman D, de Martel C (2015). Global
Henzler T, Prosch H, Heussel CP, et al. colorectal cancer incidence. Int J Cancer.
burden of gastric cancer attributable
(2017). European position statement on 140(12):2709–15. https://doi.org/10.1002/
to Helicobacter pylori. Int J Cancer.
lung cancer screening. Lancet Oncol. ijc.30686 PMID:28281292
136(2):487–90. https://doi.org/10.1002/
18(12):e754–66. https://doi.org/10.1016/
13. Maringe C, Walters S, Rachet B, Butler J, ijc.28999 PMID:24889903
S1470-2045(17)30861-6 PMID:29208441
Fields T, Finan P, et al.; ICBP Module 1 21. IARC Helicobacter pylori Working Group
5. Bray F, Ferlay J, Soerjomataram I, Siegel Working Group (2013). Stage at diagnosis (2014). Helicobacter pylori eradication
RL, Torre LA, Jemal A (2018). Global and colorectal cancer survival in six high- as a strategy for preventing gastric can-
cancer statistics 2018: GLOBOCAN es- income countries: a population-based cer. Lyon, France: International Agency
timates of incidence and mortality world- study of patients diagnosed during 2000–
wide for 36 cancers in 185 countries. CA for Research on Cancer (IARC Working
2007. Acta Oncol. 52(5):919–32. https:// Group Reports, No. 8). Available from:
Cancer J Clin. 68(6):394–424. https://doi. doi.org/10.3109/0284186X.2013.764008
org/10.3322/caac.21492 PMID:30207593 http://publications.iarc.fr/391.
PMID:23581611
6. Torre LA, Islami F, Siegel RL, Ward EM, 22. Colquhoun A, Arnold M, Ferlay J,
14. Zhou CK, Check DP, Lortet-Tieulent J, Goodman KJ, Forman D, Soerjomataram
Jemal A (2017). Global cancer in women:
Laversanne M, Jemal A, Ferlay J, et al. I (2015). Global patterns of cardia and
burden and trends. Cancer Epidemiol
(2016). Prostate cancer incidence in 43 non-cardia gastric cancer incidence
Biomarkers Prev. 26(4):444–57. https://
doi.org/10.1158/1055-9965.EPI-16-0858 populations worldwide: an analysis of in 2012. Gut. 64(12):1881–8. https://
PMID:28223433 time trends overall and by age group. Int d o i .o r g /10 .113 6 /g u t j n l - 2 014 - 3 0 8 915
J Cancer. 138(6):1388–400. https://doi. PMID:25748648
7. Arnold M, Karim-Kos HE, Coebergh JW, org/10.1002/ijc.29894 PMID:26488767
Byrnes G, Antilla A, Ferlay J, et al. (2015). 23. Anderson WF, Rabkin CS, Turner N,
Recent trends in incidence of five common 15. Sierra MS, Soerjomataram I, Forman Fraumeni JF Jr, Rosenberg PS, Camargo
cancers in 26 European countries since D (2016). Prostate cancer burden in MC (2018). The changing face of noncar-
1988: analysis of the European Cancer Central and South America. Cancer dia gastric cancer incidence among US
Observatory. Eur J Cancer. 51(9):1164– Epidemiol. 44(Suppl 1):S131–40. https:// non-Hispanic Whites. J Natl Cancer Inst.
87. https://doi.org/10.1016/j.ejca.2013.09. doi.org /10.1016/ j.c anep. 2016.0 6.010 110(6):608–15. https://doi.org/10.1093/
002 PMID:24120180 PMID:27678315 jnci/djx262 PMID:29361173
32
SECTION 1
CHAPTER 1.2
24. Plummer M, de Martel C, Vignat J, Ferlay J, 27. Anttila A, Pukkala E, Söderman B, Kallio 29. Bray F, Lortet-Tieulent J, Znaor A, Brotons
Bray F, Franceschi S (2016). Global burden M, Nieminen P, Hakama M (1999). Effect M, Poljak M, Arbyn M (2013). Patterns and
of cancers attributable to infections in 2012: trends in human papillomavirus-related
of organised screening on cervical can-
a synthetic analysis. Lancet Glob Health. diseases in Central and Eastern Europe
cer incidence and mortality in Finland, and Central Asia. Vaccine. 31(Suppl 7):
4(9):e609–16. https://doi.org/10.1016/S22 1963–1995: recent increase in cervical H32–45. https://doi.org/10.1016/j.vaccine.
14-109X(16)30143-7 PMID:27470177
cancer incidence. Int J Cancer. 83(1):59– 2013.02.071 PMID:24332296
25. Vaccarella S, Laversanne M, Ferlay J, Bray 65. https://doi.org/10.1002/(SICI)1097-0215 30. Utada M, Chernyavskiy P, Lee WJ,
F (2017). Cervical cancer in Africa, Latin (19990924)83:1<59::AID-IJC12>3.0.CO; Franceschi S, Sauvaget C, de Gonzalez
America and the Caribbean and Asia: re- 2-N PMID:10449609 AB, et al. (2019). Increasing risk of uterine
gional inequalities and changing trends. Int cervical cancer among young Japanese
J Cancer. 141(10):1997–2001. https://doi. 28. de Kok IM, van der Aa MA, van women: comparison of incidence trends
org/10.1002/ijc.30901 PMID:28734013 Ballegooijen M, Siesling S, Karim-Kos in Japan, South Korea and Japanese-
HE, van Kemenade FJ, et al.; Working Americans between 1985 and 2012. Int
26. Vaccarella S, Lortet-Tieulent J, Plummer J Cancer. 144(9):2144–52. https://doi.
Group Output of the Netherlands Cancer org/10.1002/ijc.32014 PMID:30474210
M, Franceschi S, Bray F (2013). Worldwide
trends in cervical cancer incidence: Registry (2011). Trends in cervical can-
cer in the Netherlands until 2007: has 31. WHO (2017). Updating Appendix 3 of the
impact of screening against changes WHO global NCD action plan 2013–2020.
in disease risk factors. Eur J Cancer. the bottom been reached? Int J Cancer. Geneva, Switzerland: World Health Organi
49(15):3262–73. https://doi.org/10.1016/j. 128(9):2174–81. https://doi.org/10.1002/ zation. Available from: https://www.who.int/
ejca.2013.04.024 PMID:23751569 ijc.25553 PMID:20626043 ncds/governance/appendix3-update/en/.

1.3
Transitions in human development

and the global cancer burden
Miranda M. Fidler-Benaoudia Bernard W. Stewart (reviewer)

Freddie Bray Elisabete Weiderpass (reviewer)
Christopher P. Wild (reviewer)
year by 2040, a 50% increase on health and disease patterns are

SUMMARY the estimated 18.1 million can- influenced by demographic, eco-
cers in 2018. The estimated in- nomic, and societal factors [1]. In
●● The Human Development Index creases in the cancer incidence particular, Omran described how,
(HDI), with a four-tier categori- burden from 2018 to 2040 using in the third stage of the transition,
zation of countries as having demographic changes will occur infections become less important
low, medium, high, or very high in all countries, but the predict- and chronic diseases become more
HDI, provides a useful frame- ed increases will be proportion- important as the major causes of
work for assessing the global ately greatest in countries with morbidity and mortality as life ex-
cancer burden geographically low and medium HDI. pectancy increases to more than
and over time. 70 years and mortality – from “de-
●● Human development plays a
●● The average HDI values at the critical role in understanding the generative diseases” – is delayed.
country level can be linked to the shifting scale and profile of can- This late stage of the transition
corresponding scale and profile cer globally. However, using the corresponds with the current rising
of cancer to document the effect four-tier HDI to describe transi- prominence of noncommunicable
of transitions towards higher HDI tions has limitations, given that diseases, which in the past de-
levels, and this can serve as evi- it de-emphasizes the diversity cades have surpassed communica-
dence for national cancer con- of cancer occurrence and can ble diseases as the leading causes
trol priorities. Similar linkages to oversimplify the multifactorial of death worldwide [2].
risk factors and cancer-related influences, including sex, eth- Among noncommunicable dis-
outcomes can help to further ex- nicity, and cultural aspects, on a eases, cancer has emerged as a
plain transitions and inequalities complex set of diseases. particularly important health con-
in the cancer burden. cern. Cancer is the first or second
●● Although attention has been leading cause of premature mortali-
●● A high residual burden of infec- drawn to broad patterns of can- ty (i.e. deaths at ages 30–69 years)
tion-related cancers is observed cer incidence according to hu- in more than 90 countries world-
in countries with low HDI. Several man development level, there wide (see Chapter 1.1). An esti-
countries with medium and high are clear examples of national mated 18.1 million new cancer cas-
HDI, which are often undergo- and regional cancer diversity of es and 9.6 million cancer-related
ing major social and economic cancer occurrence that depart deaths occurred worldwide in 2018,
transitions, have experienced from this model. Also, because and 1 in 8 men and 1 in 10 women
marked declines in the burden of HDI indicates national aver- are likely to develop the disease
infection-related cancers. These ages, it does not reflect any during their lifetimes [3]. When
declines have subsequently inequalities in human develop- coupled with the estimated cost of
been offset by increasing rates ment within countries.
cancer care of US$ 1.16 trillion per
of cancer types that are more
year [4], this clearly makes cancer
frequently observed in industrial-
a public health priority. As a result,
ized countries. Epidemiological there has been a growing recogni-
●● The predicted global cancer transitions in cancer tion of the need for action to reduce
burden is expected to exceed Omran’s theory of epidemiological the cancer burden. This is exempli-
27 million new cancer cases per transition described how changing fied by the World Health Assembly
34
SECTION 1
CHAPTER 1.3
resolution on cancer prevention and the expected escalating numbers of reflect any inequalities in human de-
control, which was adopted unani- cancer patients in coming decades. velopment within countries.
mously by WHO Member States in The global map of countries ac-
May 2017 [5]. The Human Development cording to the HDI tiers is shown in
Fig. 1.3.1. The low HDI tier includes
Although cancer was once con- Index countries that are largely concen-
sidered to be a disease of rich peo-
Human development focuses on trated in sub-Saharan Africa, al-
ple and of the highest-income coun-
two core dimensions: (i) directly though several countries in this
tries, it is now a global problem that enhancing human abilities, and region have now transitioned to the
affects all countries. The increasing (ii) creating conditions for human medium HDI level. The countries
magnitude of the cancer burden is in development [6]. Like the previous in the high and very high HDI tiers
part a consequence of declining fer- two chapters, this chapter uses the are geographically diverse, span-
tility and increasing life expectancy, Human Development Index (HDI), ning across continents, although
but it is also the result of societal, a summary measure developed by the very high HDI tier remains clos-
economic, and lifestyle changes as- the United Nations Development est to the traditional view of “devel-
sociated with globalization. Programme. HDI is an indicator of oped” countries in that it includes
In this chapter, the impact of national achievement in attaining Europe and North America, Japan,
transitions in human development a long and healthy life (based on and Australia and New Zealand.
on cancer occurrence worldwide is life expectancy at birth), acquiring The very high HDI tier also includes
illustrated by the profound effects knowledge (based on average and several countries in Asia, the
on the patterns and trends of cancer expected years of schooling), and Eastern Mediterranean region, and
incidence, mortality, and prevalence achieving a decent standard of living South America. Most the world’s
at the national, regional, and global (based on gross national income per population live in countries in the
capita) [7]. HDI values range from 0 medium (36.2%) and high (32.3%)
levels. The predicted increases in
to 1; lower values indicate the least HDI tiers, followed by the very high
the cancer burden will be propor-
developed countries in terms of hu- (18.0%) and low (13.5%) HDI tiers.
tionately greatest in countries in
man development, and higher values
transition towards higher levels of indicate the most developed coun-
human development. Such findings tries. Values are commonly present- Cancer burden by HDI
have major implications for pub- ed, as in this chapter, according to level in 2018
lic health and cancer control plan- four tiers of HDI (low, medium, high, When the cancer burden in 2018 was
ning, and therefore should alert the and very high HDI), using the pre- assessed by the four-tier HDI, a step-
global community to the growing defined cut-off points of the United wise increase in the number of new
cancer burden and the need for ac- Nations Development Programme. cancer cases and in the age-stan-
tion, particularly in countries that are Because HDI is a composite indica- dardized incidence rates was evi-
currently ill-equipped to deal with tor of national averages, it does not dent with each increase in HDI level
Fig. 1.3.1. Global map of the development levels of individual countries according to the four-tier Human Development Index (HDI),
in 2015.
Chapter 1.3 • Transitions in human development and the global cancer burden 35
(Fig. 1.3.2). In 2018, 45% of the es- standardized incidence rates indicate cancer types in 2018 for each sex
timated new cancer cases occurred a slightly different pattern, in which are shown in Fig. 1.3.3, which com-
in countries with very high HDI, countries in the low and medium HDI pares the burden in countries with
compared with 36%, 16%, and 4% tiers have comparable burdens, al- high or very high HDI with that in
in countries with high, medium, and though the burden is slightly higher countries with low or medium HDI.
low HDI, respectively. In contrast, the in the low HDI tier. For age-standard- With the exception of rates of a few
greatest number of cancer deaths ized mortality rates, no correlation cancer types, the incidence rates
occurred in countries with high HDI, with HDI level is observed. were generally greater in coun-
driven by the 2.9 million cancer The age-standardized incidence tries with higher HDI; the age-stan-
deaths that occurred in China. Age- and mortality rates for the top 15 dardized incidence rates in many
of these countries were 2–3 times
those in countries in transition to-
Fig. 1.3.2. The total burden of new cancer cases and cancer deaths (above) and the wards higher HDI levels.
corresponding age-standardized (World) incidence and mortality rates per 100 000 In contrast, the mortality rates
person-years (below) for each Human Development Index (HDI) tier, in 2018. were broadly comparable between
the two groups of countries. For
some cancer types, such as breast
cancer and ovarian cancer, the mor-
tality burden was greater in countries
with low or medium HDI, although the
incidence rates in those countries
were lower than the rates in coun-
tries with high or very high HDI. The
proportionately higher case fatalities
in countries with low or medium HDI
relates to the poorer survival pros-
pects after diagnosis on average, for
reasons that include a lack of access
to timely diagnosis and treatment.
For example, when the mortality-to-
incidence ratio is used as a proxy of
survival, the case fatality for breast
cancer is 48% in countries with low
or medium HDI, 4 times that in coun-
tries with high or very high HDI.
Cancer profile by HDI

level in 2018
Cancer profiles by HDI level differ
when assessed by incidence, mor-
tality, and 5-year prevalence.
In women, the five major can-
cer types accounted for more than
50% of the burden in each of these
three indicators (Fig. 1.3.4). Uniquely,
breast cancer was the most com-
mon cancer type across all HDI tiers
in terms of incidence, followed by
cervical cancer in the low and me-
dium HDI tiers and colorectal cancer
in the high and very high HDI tiers.
Cervical cancer was the most
common cause of cancer mortality
in the low HDI tier and the second
most common in the medium HDI
tier, highlighting a residual burden of
infection-related cancers in countries
36
SECTION 1
CHAPTER 1.3
Fig. 1.3.3. Bar charts of age-standardized (World) incidence and mortality rates per 100 000 person-years for the top 15 cancer
types in 2018 in countries with high or very high Human Development Index (HDI) compared with countries with low or medium
HDI, in women (top) and men (bottom).
Fig. 1.3.4. The five leading cancer types in terms of incidence, mortality, and 5-year prevalence for each Human Development Index
(HDI) tier in women in 2018.
38
SECTION 1
CHAPTER 1.3
Fig. 1.3.5. The five leading cancer types in terms of incidence, mortality, and 5-year prevalence for each Human Development Index
(HDI) tier in men in 2018.
in these tiers. In contrast, in both the Future cancer burden by of carcinogenic hazards, including
high and very high HDI tiers, infec- HDI level tobacco use and alcohol consump-
tion-related cancers (see Chapter tion [9,10], infectious agents [11],
The predicted global cancer burden
2.2) have been displaced by lung obesity [12], diet [13–15], radiation
is expected to exceed 27 million
cancer, breast cancer, and colorectal [16], solar radiation [17,18], and air
new cancer cases per year by 2040,
cancer; these cancer types, which pollution [19,20]. Of these, obesity
a 50% increase on the estimated
are associated with behaviours and and infectious agents are particu-
18.1 million new cancer cases in
lifestyles that are more typical of in- larly interesting to examine accord-
2018. Although the predicted can-
dustrialized societies, have become ing to HDI, because of their relative
cer incidence burden is highest in
the leading causes of cancer mortali- importance in the cancer burden in
countries with high and very high
ty in the high and very high HDI tiers. countries with higher HDI (obesity)
HDI, the predicted increases will be
In women, the 5-year prevalence and lower HDI (infectious agents).
proportionately greatest in countries
burden in each HDI tier generally
with low and medium HDI: the esti- Obesity
had a similar profile of cancer types
mated increase from 2018 to 2040
to that observed for incidence. Excess body fatness (see Chapter
using demographic changes alone
In 2018, the cancer profile by 2.7) is considered to cause the fol-
is 100% for the low HDI tier and 75%
HDI level varied more substantially lowing cancer types: cancers of the
for the medium HDI tier (Fig. 1.3.6).
in men than in women. In men, the oesophagus (adenocarcinoma), gas-
Because countries with low and
top five cancer types were different tric cardia, colon and rectum, liver,
medium HDI levels are currently
in each HDI tier (Fig. 1.3.5). In terms gall bladder, pancreas, breast (in
the least equipped to deal with the
of incidence, lung cancer was the postmenopausal women), endome-
impending increase in the cancer
most common type in the medium trium, ovary, kidney (renal cell carci-
burden, these findings underscore
and high HDI tiers, whereas prostate noma), and thyroid, and meningioma
the necessity for investment in
cancer was the most common type and multiple myeloma [12]. When the
targeted, resource-dependent, ef-
in the low and very high HDI tiers; relationship between excess weight
fective, and cost-effective interven-
this pattern may relate to ethnic and – or obesity – and cancer was as-
tions that can reduce the burden of
underlying genetic predispositions sessed by HDI, the attributable frac-
the disease [5,8].
in the low HDI tier and to prostate- tions in countries with very high and
specific antigen (PSA)-related diag- high HDI (~5% each) were 2–3 times
nosis of latent cancers in the very Cancer risk factors by those in countries with medium HDI
high HDI tier. Although the burden HDI level (1.6%) or low HDI (1.0%) [21]. When
of infection-related cancers, such Despite the broad associations be- the relationship was assessed by
as liver cancer and Kaposi sarcoma, tween cancer and HDI described sex, the number of cancer cases at-
is higher in countries in transition, above, there remain a large number tributable to obesity was observed to
there remains a large burden of liver
cancer in the high HDI tier; this is
due to the nearly 393 000 new cases Fig. 1.3.6. The estimated number of new cancer cases in 2018 and the predicted
in China in 2018, which accounted increase in the number of new cancer cases from 2018 to 2040, assuming only a
for 84% of the liver cancer cases in demographic effect, by Human Development Index (HDI) tier.
the high HDI tier.
Prostate cancer was the lead-
ing cause of cancer mortality in the
low HDI tier, whereas lung cancer
was the leading cause in the me-
dium, high, and very high HDI tiers.
Liver cancer and colorectal cancer
were also among the most com-
mon causes of cancer mortality in
all four HDI tiers. The cancer types
contributing to the remaining mor-
tality burden varied by HDI level.
In men, the 5-year prevalence
burden in each HDI tier had a simi-
lar profile of cancer types to that
observed for incidence, except that
the ranking was higher for cancer
types associated with better surviv-
al prospects after diagnosis.
40
SECTION 1
CHAPTER 1.3
Fig. 1.3.7. The number of cancer cases attributable to high body mass index (BMI) and the number of preventable cancers if BMI
scores remained the same in 2012 as in 1982, by sex and Human Development Index (HDI) tier.
increase with HDI level in both men portion of infection-related cancers Soerjomataram et al. assessed
and women (Fig. 1.3.7). attributable to HPV decreased with DALYs globally by the four-tier HDI
When the number of preventable increasing HDI [22]. In contrast, in- and found the total DALYs to be
cancers was assessed, the number fection with Helicobacter pylori con- similar across HDI tiers (Fig. 1.3.8)
increased with HDI level in men. tributed substantially to the cancer [23]. However, the contribution of
This relationship was less consis- burden in countries in the high and YLL and YLD to the total DALYs
tent in women; the number of pre- very high HDI tiers [22]. varied substantially by HDI tier: in
ventable cancers was greatest in the Because two thirds of infection- general, the number of years lived
very high and medium HDI tiers [21]. attributable cancer cases occurred with disability (YLD) was greater in
Therefore, although prevention pro- in less-developed countries, effec- countries with higher HDI levels,
grammes that seek to control weight tive population-based vaccination and the burden of premature mor-
gain are clearly needed in the most and screen-and-treat programmes tality (YLL) was greater in countries
developed countries, these find- should be prioritized and imple- with lower HDI levels.
ings also emphasize the need for a mented in a cost-effective manner The relationship between DALYs
global effort to reduce the number of to combat the disproportionately and HDI level varied depending on
people with high body mass index, high burden in these countries. the cancer site being assessed. In
because the continuation of current particular, for cancer types more
patterns of population weight gain commonly attributable to obesity
will increase the future cancer bur-
Cancer outcomes by HDI (e.g. breast cancer and colorectal
den across all HDI tiers [21]. level cancer), DALYs were greater in coun-
Given that cancer contributes sub- tries with higher HDI levels, whereas
Infections stantially to morbidity and mortality for infection-related cancer types
In 2012, approximately 15% of new globally, it is important to assess the (e.g. cervical cancer and liver can-
cancer cases worldwide were attri- implications of cancer and the extent cer), DALYs were greater in countries
butable to infections (see Chapter of cancer-related sequelae. To deter- with lower HDI levels [23]. YLL was
2.2) [22]. When the proportion of mine the impact of fatal and non-fatal consistently the main contributor to
cancers attributable to infections cancer outcomes, disability-adjusted DALYs across HDI tiers, but the frac-
was assessed by HDI tier, a gradient life years (DALYs) are often used as a tion of DALYs due to YLL in the low-
was observed: the attributable frac- measure. DALYs combine the degree est HDI tier was generally the same
tions were 25%, 22%, 13%, and 8%, of illness and disability in patients and as or larger than the fraction in the
respectively, in the low, medium, long-term survivors (years of healthy highest HDI tier, reflecting the poorer
high, and very high HDI tiers [22]. life lost due to disability [YLD]) and average prognosis of patients with
Infection with human papilloma- the burden of cancer mortality (years cancer in low-resource settings.
virus (HPV) caused approximately of life lost due to premature mortal- In another study, the impact of
half of all infection-attributable can- ity [YLL]), to quantify the number of cancer on changes (increases or
cers in the low HDI tier, and the pro- years of healthy life lost. decreases) in life expectancy was
Fig. 1.3.8. Age-adjusted disability-adjusted life years (DALYs) per 100 000 population by Human Development Index (HDI) tier for
all cancer sites combined and selected cancer sites. YLD, years of healthy life lost due to disability; YLL, years of life lost due to
premature mortality.
assessed worldwide for the period and 1.5 years for women [24]. These ing prevalence and distribution of
1981–2010 [24]. The findings sug- results provide evidence of dispro- specific reproductive, dietary, and
gested that countries with very high portionate improvements in cancer metabolic factors.
HDI had larger gains in life expec- outcomes according to HDI level, Using the four-tier HDI to de-
tancy compared with countries with leading to widening gaps in life ex- scribe transitions has limitations,
medium or high HDI. In particular, pectancy between more-developed given that it de-emphasizes the
declines in cancer mortality were and less-developed countries. diversity of cancer occurrence
responsible for the increases in life worldwide and the extent to which
expectancy for individuals aged 40– it varies between and within coun-
Evidence of diversity
84 years by 0.8 years for men and tries. Although attention has been
0.5 years for women in countries with
within HDI levels drawn to broad patterns of cancer
very high HDI, whereas the corre- Evidently, the marked differences in incidence according to human de-
sponding gains were less in countries the scale and profile of cancer inci- velopment level, there are clear
with medium or high HDI: 0.2 years dence and mortality by HDI level re- examples of national and regional
for both men and women [24]. sult from a myriad of factors, which diversity of cancer occurrence that
Similar inequalities in life ex- will dictate whether, in the longer depart from this model.
pectancy gains were observed for term, gains in societal and econom- For example, although there
the hypothetical situation of elimi- ic development will reduce the wid- have been systematic declines in
nating all deaths from cancer. The ening gap between countries with cervical cancer incidence rates in
resulting increase in life expectancy low versus very high HDI in the risk countries with medium or high HDI,
for individuals aged 40–84 years of developing or dying from cancers the 40-year trends in incidence
for the period 2006–2010 was that are preventable or treatable. rates indicate recent increases in
2.5 years for men and 1.9 years for Some of the determinants are sys- countries with high or very high HDI
women in countries with very high tems-related, including the extent (e.g. Belarus and Japan) (Fig. 1.3.9).
HDI, whereas the increases were to which cancer control initiatives Such increases are likely to be due
only modest in countries with me- are implemented, and others link to changes in sexual behaviour that,
dium or high HDI: 1.6 years for men to risk directly, such as the chang- in the absence of effective screening
42
SECTION 1
CHAPTER 1.3
programmes, have led to an increas- Fig. 1.3.9. Age-standardized (World) incidence rates per 100 000 person-years for
ing risk of persistent infection with cervical cancer by calendar year in selected countries with high and very high Human
Development Index (HDI) levels, circa 1975–2012. Asterisks indicate regional registries
high-risk HPV subtypes and subse-
(other registries are national).
quent increases in the occurrence of
cervical cancer (see Chapter 1.2).
Conclusions
Despite inherent diversity in the can-
cer burden within a given HDI level,
HDI provides a useful framework
to map out continuing transitions in
cancer incidence, risk factors, and
outcomes. In particular, HDI serves
as an exploratory tool to monitor
shifts in the profile of cancer types,
as clearly demonstrated by the dis-
placement of infection-related can-
cers by cancers associated with be-
haviours and lifestyles that are more
typical of industrialized societies,
and with increasing societal and
economic development.
Although the cancer incidence
burden is currently highest in coun-
tries with very high HDI, the pre-
dicted increases in the cancer bur-
den will have the greatest impacts
on countries with low and medium
HDI. Because cancer outcomes
are already poorer in countries in
transition, appropriate scaling up
of resources for effective strategies
in primary and secondary preven-
tion in these countries is critical to
effectively control the prevalence
of adverse lifestyle factors, to ulti-
mately reduce the cancer burden.
References
1. Omran AR (1971). The epidemiologic 10. IARC (2004). Tobacco smoke and in- 18. Hrushesky WJM, Sothern RB, Rietveld
transition. A theory of the epidemiolo- voluntary smoking. IARC Monogr WJ, Du Quiton J, Boon ME (2005). Season,
gy of population change. Milbank Mem Eval Carcinog Risks Hum. 83:1–1438. sun, sex, and cervical cancer. Cancer
Fund Q. 49(4):509–38. https://doi. Available from: http://publications.iarc. Epidemiol Biomarkers Prev. 14(8):1940–
org/10.2307/3349375 PMID:5155251 fr/101 PMID:15285078 7. https://doi.org/10.1158/1055-9965.EPI-
2. WHO (2016). Global Health Observatory 04-0940 PMID:16103441
11. Bouvard V, Baan R, Straif K, Grosse Y,
(GHO) data repository. Geneva, Switzer
Secretan B, El Ghissassi F, et al.; WHO 19. Gharibvand L, Shavlik D, Ghamsary M,
land: World Health Organization. Available
International Agency for Research on Beeson WL, Soret S, Knutsen R, et al.
from: https://www.who.int/gho/en/.
Cancer Monograph Working Group (2017). The association between am-
3. Bray F, Ferlay J, Soerjomataram I, Siegel (2009). A review of human carcinogens – bient fine particulate air pollution and
RL, Torre LA, Jemal A (2018). Global Part B: biological agents. Lancet Oncol. lung cancer incidence: results from
cancer statistics 2018: GLOBOCAN es- 10(4):321–2. https://doi.org/10.1016/S1470- the AHSMOG-2 Study. Environ Health
timates of incidence and mortality world- 2045(09)70096-8 PMID:19350698 Perspect. 125(3):378–84. https://doi.
wide for 36 cancers in 185 countries. CA org/10.1289/EHP124 PMID:27519054
Cancer J Clin. 68(6):394–424. https://doi. 12. Lauby-Secretan B, Scoccianti C, Loomis
org/10.3322/caac.21492 PMID:30207593 D, Grosse Y, Bianchini F, Straif K; 20. Yang W-S, Zhao H, Wang X, Deng Q, Fan
International Agency for Research on W-Y, Wang L (2016). An evidence-based
4. Seventieth World Health Assembly
Cancer Handbook Working Group (2016). assessment for the association between
(‎ 2017)‎ . Resolution WHA70.12. Cancer
Body fatness and cancer – viewpoint of long-term exposure to outdoor air pol-
prevention and control in the context
the IARC Working Group. N Engl J Med. lution and the risk of lung cancer. Eur J
of an integrated approach. Geneva,
Switzerland: World Health Organization. 375(8):794–8. https://doi.org/10.1056/ Cancer Prev. 25(3):163–72. https://doi.
Available from: https://apps.who.int/iris/ NEJMsr1606602 PMID:27557308 org/10.1097/CEJ.0 0 0 0 0 0 0 0 0 0 0 0 0158
handle/10665/275676. PMID:25757194
13. Armstrong B, Doll R (1975). Environmental
5. WHO (2017). Seventieth World Health factors and cancer incidence and mortal- 21. Arnold M, Pandeya N, Byrnes G, Renehan
Assembly update, 30 May 2017. Geneva, ity in different countries, with special ref- PAG, Stevens GAG, Ezzati PM, et al.
Switzerland: World Health Organization. erence to dietary practices. Int J Cancer. (2015). Global burden of cancer attribu-
Available from: https://www.who.int/news- 15(4):617–31. https://doi.org/10.1002/ijc. table to high body-mass index in 2012: a
room/detail/30-05-2017-seventieth-world- 2910150411 PMID:1140864 population-based study. Lancet Oncol.
health-assembly-update-30-may-2017. 16(1):36–46. https://doi.org/10.1016/S1470-
14. Grant WB (2013). A multicountry ecologi- 2045(14)71123-4 PMID:25467404
6. UNDP Human Development Report
cal study of cancer incidence rates in 2008
Office (2015). What is human devel- 22. Plummer M, de Martel C, Vignat J, Ferlay
with respect to various risk-modifying fac-
opment? New York (NY), USA: United J, Bray F, Franceschi S (2016). Global
Nations Development Programme. tors. Nutrients. 6(1):163–89. https://doi.
org/10.3390/nu6010163 PMID:24379012 burden of cancers attributable to infec-
Available from: http://hdr.undp.org/en/
tions in 2012: a synthetic analysis. Lancet
content/what-human-development. 15. Bouvard V, Loomis D, Guyton KZ, Grosse Glob Health. 4(9):e609–16. https://doi.
7. UNDP (2015). Human development re- Y, Ghissassi FE, Benbrahim-Tallaa L, et org /10.1016/S2 214 -10 9X(16) 3 014 3 -7
port 2015: work for human development. al.; International Agency for Research PMID:27470177
New York (NY), USA: United Nations on Cancer Monograph Working Group
Development Programme. Available from: (2015). Carcinogenicity of consumption of 23. Soerjomataram I, Lortet-Tieulent J, Ferlay
http://hdr.undp.org/en/content /human- red and processed meat. Lancet Oncol. J, Forman D, Mathers C, Parkin DM, et al.
development-report-2015-work-human- 16(16):1599–600. https://doi.org/10.1016/ (2012). Estimating and validating disabili-
development. S1470-2045(15)00444-1 PMID:26514947 ty-adjusted life years at the global level:
a methodological framework for can-
8. WHO (2017). Updating Appendix 3 16. IARC (2012). Radiation. IARC Monogr cer. BMC Med Res Methodol. 12(1):125.
of the WHO global NCD action plan Eval Carcinog Risks Hum. 100D:1–437. https://doi.org/10.1186/1471-2288-12-125
2013–2020. Geneva, Switzerland: World
Available from: http://publications.iarc. PMID:22901001
Health Organization. Available from:
fr/121 PMID:23189752
https://w w w.who.int /ncds/governance/ 24. Cao B, Bray F, Beltrán-Sánchez H,
appendix3-update/en/. 17. Fleischer AB Jr, Fleischer SE (2016). Ginsburg O, Soneji S, Soerjomataram
9. IARC (2010). Alcohol consumption Solar radiation and the incidence and I (2017). Benchmarking life expectancy
and ethyl carbamate. IARC Monogr mortality of leading invasive cancers in and cancer mortality: global comparison
Eval Carcinog Risks Hum. 96:1–1428. the United States. Dermatoendocrinol. with cardiovascular disease 1981–2010.
Available from: http://publications.iarc. 8(1):e1162366. https://doi.org/10.1080/ BMJ. 357:j2765. https://doi.org/10.1136/
fr/114 PMID:21735939 19381980.2016.1162366 PMID:27195056 bmj.j2765 PMID:28637656
44
Known causes of human cancer by organ site
Agents classified as carcinogenic to humans (Group 1) by the IARC Monographs programme (IARC Monographs
on the Evaluation of Carcinogenic Risks to Humans, Volumes 1–125), listed by organ site with sufficient evidence.
Organ site Agent
All cancer sites (combined) 2,3,7,8-Tetrachlorodibenzo-para-dioxin
Anus Human immunodeficiency virus type 1

Human papillomavirus type 16
Biliary tract 1,2-Dichloropropane

Clonorchis sinensis
Opisthorchis viverrini
Bladder Aluminium production

4-Aminobiphenyl
Arsenic and inorganic arsenic compounds
Auramine production
Benzidine
Chlornaphazine
Cyclophosphamide
Magenta production
2-Naphthylamine
Painter (occupational exposure as)
Rubber production industry
Schistosoma haematobium
Tobacco smoking
ortho-Toluidine
X-radiation, γ-radiation
Bone Plutonium
Radium-224 and its decay products
Brain and central nervous system X-radiation, γ-radiation
Breast Alcoholic beverages

Diethylstilbestrol
Estrogen–progestogen contraceptives
Estrogen–progestogen menopausal therapy
Cervix Diethylstilbestrol (exposure in utero)

Human immunodeficiency virus type 1
Human papillomavirus types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59
Tobacco smoking
Colon and rectum Alcoholic beverages

Consumption of processed meat
Tobacco smoking
Known causes of human cancer by organ site 45

Organ site Agent
Corpus uteri (endometrium) Estrogen menopausal therapy

Estrogen–progestogen menopausal therapy
Tamoxifen
Endothelium (Kaposi sarcoma) Human immunodeficiency virus type 1

Kaposi sarcoma herpesvirus
Eye Human immunodeficiency virus type 1

Ultraviolet-emitting tanning devices
Ultraviolet radiation from welding
Gall bladder Thorium-232 and its decay products
Kidney Tobacco smoking

Trichloroethylene
Larynx Acid mists, strong inorganic

Alcoholic beverages
Asbestos (all forms)
Tobacco smoking
One or more subtypes of leukaemia Azathioprine

or lymphoma Benzene
Busulfan
1,3-Butadiene
Chlorambucil
Cyclophosphamide
Cyclosporine
Epstein–Barr virus
Etoposide with cisplatin and bleomycin
Fission products, including strontium-90
Formaldehyde
Helicobacter pylori
Hepatitis C virus
Human immunodeficiency virus type 1
Human T-cell lymphotropic virus type 1
Kaposi sarcoma herpesvirus
Lindane
Melphalan
MOPP combined chemotherapy (vincristine, prednisone, nitrogen mustard,
and procarbazine mixture)
Pentachlorophenol
Phosphorus-32, as phosphate
Semustine [1-(2-chloroethyl)-3-(4-methylcyclohexyl)-1-nitrosourea, or methyl-CCNU]
Thiotepa
Thorium-232 and its decay products
Tobacco smoking
Treosulfan
Liver (angiosarcoma) Vinyl chloride
Liver (hepatocellular carcinoma) Aflatoxins

Alcoholic beverages
Hepatitis B virus
Hepatitis C virus
Plutonium
Thorium-232 and its decay products
Tobacco smoking (in smokers and in smokers’ children)
46
Organ site Agent
Lung Acheson process (occupational exposures associated with)

Aluminium production
Arsenic and inorganic arsenic compounds
Asbestos (all forms)
Beryllium and beryllium compounds
Bis(chloromethyl)ether; chloromethyl methyl ether (technical grade)
Cadmium and cadmium compounds
Chromium(VI) compounds
Coal, indoor emissions from household combustion
Coal gasification
Coal-tar pitch
Coke production
Diesel engine exhaust
Haematite mining (underground)
Iron and steel founding
MOPP combined chemotherapy (vincristine, prednisone, nitrogen mustard,
and procarbazine mixture)
Nickel compounds
Outdoor air pollution
Outdoor air pollution, particulate matter in
Plutonium
Radon-222 and its decay products
Silica dust, crystalline
Soot
Sulfur mustard
Tobacco smoke, second-hand
Tobacco smoking
Welding fumes
Mesothelium (pleura and peritoneum) Asbestos (all forms)

Erionite
Fluoro-edenite
Multiple sites (unspecified) Cyclosporine

Fission products, including strontium-90
X-radiation, γ-radiation (exposure in utero)
Nasal cavity and paranasal sinus Isopropyl alcohol manufacture using strong acids
Leather dust
Nickel compounds
Tobacco smoking
Wood dust
Nasopharynx Epstein–Barr virus

Formaldehyde
Salted fish, Chinese-style
Tobacco smoking
Wood dust
Oesophagus Acetaldehyde associated with consumption of alcoholic beverages

Alcoholic beverages
Betel quid with tobacco
Betel quid without tobacco
Smokeless tobacco
Tobacco smoking
Known causes of human cancer by organ site 47

Organ site Agent
Oral cavity Alcoholic beverages

Betel quid with tobacco
Betel quid without tobacco
Smokeless tobacco
Tobacco smoking
Ovary Asbestos (all forms)

Estrogen menopausal therapy
Tobacco smoking
Pancreas Smokeless tobacco

Tobacco smoking
Penis Human papillomavirus type 16
Pharynx Alcoholic beverages

(oropharynx, hypopharynx, Betel quid with tobacco
and/or not otherwise specified) Human papillomavirus type 16
Tobacco smoking
Renal pelvis and ureter Aristolochic acid, plants containing

Phenacetin
Phenacetin, analgesic mixtures containing
Tobacco smoking
Salivary gland X-radiation, γ-radiation
Skin (melanoma) Polychlorinated biphenyls

Solar radiation
Ultraviolet-emitting tanning devices
Skin (other malignant neoplasms) Arsenic and inorganic arsenic compounds

Azathioprine
Coal-tar distillation
Coal-tar pitch
Cyclosporine
Methoxsalen plus ultraviolet A
Mineral oils, untreated or mildly treated
Shale oils
Solar radiation
Soot
Stomach Helicobacter pylori

Tobacco smoking
Thyroid Radioiodines, including iodine-131 (exposure during childhood and adolescence)

Tonsil Human papillomavirus type 16
Upper aerodigestive tract Acetaldehyde associated with consumption of alcoholic beverages
Vagina Diethylstilbestrol (exposure in utero)

Vulva Human papillomavirus type 16
Group 1 agents with less than sufficient evidence in humans: 2,3,4,7,8-pentachlorodibenzofuran; polychlorinated biphenyls, dioxin-like, with a Toxicity
Equivalency Factor (TEF) according to the World Health Organization (WHO); 4,4′-methylenebis(2-chloroaniline) (MOCA); α- and β-particle emitters;
areca nut; aristolochic acid; benzidine, dyes metabolized to; benzo[a]pyrene; ethanol in alcoholic beverages; ethylene oxide; etoposide; ionizing radiation
(all types); neutron radiation; N′-nitrosonornicotine (NNN) and 4-(N-nitroso-methylamino)-1-(3-pyridyl)-1-butanone (NNK); ultraviolet radiation.
48
2 Causes of cancer,
including hazardous
circumstances
At the community or national level, causes are types most common in low-income countries
established for a proportion of all cancers – were those caused by human papillomavirus
a proportion that differs markedly between (HPV) infection or mediated by chronic inflam-
tumour types. Tobacco smoking was once matory diseases caused by infectious agents.
prevalent mostly among men in high-income These patterns are changing, particularly with
countries but is now much more prevalent, in-
industrialization. The highest exposures are
volving women in many countries, and tobac-
co use is highest in Asia, Africa, and South often those of workers near industrial sources
America. Cancers attributable to unhealthy of pollution. Emissions from factories and ve-
diet and lack of exercise are often correlated hicles contribute to air pollution, a cause of
with the increasing prevalence of overweight lung cancer. Identifying the causes of cancer
and obesity worldwide. Previously, the cancer indicates a potential means of prevention.
2.1
Tobacco products
Massive and still growing causes
of cancer worldwide
Neal D. Freedman David H. Phillips (reviewer)
Michael J. Thun Catherine Sauvaget (reviewer)
●● The introduction of electronic Tobacco products

SUMMARY nicotine delivery systems, heat-
Commonly used tobacco products
ed tobacco products, and other
●● Tobacco products have been are listed in Box 2.1.1.
emerging nicotine and tobacco
studied for decades and are Electronic nicotine delivery sys-
products challenges regulatory
well known to cause cancer. tems (ENDS), of which e-cigarettes
approaches to tobacco control.
Nevertheless, with larger epi- are the most common, are not con-
Their long-term impact is un-
demiological studies, longer sidered as tobacco products by
known and is, rightly, the sub-
follow-up, and better control WHO. Some countries classify and
ject of considerable debate.
for confounding, the number regulate these products as tobacco
of types or subtypes of cancer products. According to the Report of
known to be caused by tobacco the Advisory Group to Recommend
Tobacco comes in many forms, and
products continues to increase. Priorities for the IARC Monographs
tobacco use has long been estab-
●● Worldwide, most tobacco is during 2020–2024, no data are avail-
lished to cause multiple types of
now consumed in low- and able so far pertaining to the carcino-
cancer and other major noncommu-
middle-income countries in the genicity of ENDS in humans. The
nicable diseases. Cigarette smoking
form of smoked products, chief- Advisory Group assigned ENDS a
causes at least 20 different types of
ly as manufactured or hand- high priority for evaluation by the
cancer [1,2]. An estimated 1.3 billion
rolled cigarettes. IARC Monographs programme with-
people use tobacco products world-
in 5 years.
●● Both smoked and smokeless wide [3]. Together, cigarettes and
products are widely used in other tobacco products are estimated Smoked/combustible
South-East Asia. to cause 2.4 million tobacco-related products
cancer deaths worldwide per year
●● In North America and Europe, Most of the tobacco consumed
[4]. Moreover, tobacco use causes
and increasingly elsewhere, worldwide is in the form of smoked
even more deaths from vascular con-
non-cigarette products such as products, chiefly as manufactured
ditions (3.1 million deaths per year)
electronic nicotine delivery sys-
tems, heated tobacco products, and respiratory conditions (606 000
water pipes, and cigars have deaths per year from respiratory in- Fig. 2.1.1. A woman in Rajasthan, India,
fections and tuberculosis; 1.5 million smoking a bidi.
become popular, particularly
among young people. deaths per year from chronic respi
ratory disease) than from cancer [4].
●● Progress in tobacco control Previous comprehensive reviews
is notable but far from suffi- have described the carcinogenicity
cient. Worldwide, an estimated of smoked and smokeless tobacco
2.4 million tobacco-related can- products [1,2], their impact on non-
cer deaths occur per year. malignant diseases, the evolution of
●● Without dramatic declines in use, the tobacco epidemic, and the harm-
tobacco products are projected ful effects of second-hand smoke
to cause 1 billion deaths world- [2,5]. This chapter focuses on select-
wide this century, mostly in lowed recent developments that affect
and middle-income countries. cancer risk and tobacco control.
50
Box 2.1.1. Commonly used tobacco products.
Smoked/combustible products
FUNDAMENTALS
• Cigarettes (manufactured and hand-rolled) ■■ Tobacco use is the leading
preventable cause of cancer
The most commonly used tobacco product worldwide
worldwide. Cigarettes are
• Cigars (large and small) the predominant form and
Tobacco that is wrapped in tobacco leaf have been determined to
cause at least 20 different
• Pipes types or subtypes of cancer.
Other forms of tobacco use
SECTION 2
CHAPTER 2.1
The oldest tobacco product, in which tobacco is placed in a bowl and
smoked through a stem are of growing importance
worldwide, but they have been
• Bidis less studied than cigarettes.
Hand-rolled tobacco made of shredded tobacco leaves wrapped in ■■ Although the prevalence of
dried temburni leaf and tied with a string smoking has decreased in
• Kreteks most regions of the world,
an estimated 1.3 billion
Clove- and coca-flavoured small cigarettes, used particularly in people use tobacco products
Indonesia worldwide, and an estimated
• Water pipes (hookah, shisha) 2.4 million tobacco-related
cancer deaths occur per year.
Users draw smoke through a water chamber by use of a long hose
■■ Reductions in smoking preva-
Other nicotine and tobacco products lence in high-income countries
• Smokeless tobacco have substantially reduced the
incidence rates of lung cancer
Used in many forms worldwide (see Fig. 2.1.2) and laryngeal cancer in men
• E-cigarettes and other electronic nicotine delivery systems (ENDS) and younger women.
An emerging product in which a nicotine-containing solution is heat- ■■ However, about 80% of
ed to produce an aerosol the world’s smokers live in
low- and middle-income
• Heated tobacco products countries, where the disease
An emerging product in which tobacco sticks are heated to produce burden from tobacco use
an aerosol continues to increase as a
result of population growth
and the ageing of long-term,
continuing smokers. Even if
or hand-rolled cigarettes but also as of shredded tobacco leaves wrapped the age-specific death rates
cigars, pipes, water pipes, kreteks, in dried temburni leaf and tied with a from tobacco-attributable
cancers remain the same, the
and bidis [6,7]. In high-income string. Kreteks are clove- and coca-
number of people affected by
countries, manufactured cigarettes flavoured small cigarettes, which are these cancers will increase
displaced other forms of tobacco by manufactured and used particularly dramatically because of these
the mid-20th century. Since then, in Indonesia. Both bidis and kreteks demographic changes.
products that were not previously are now marketed worldwide. ■■ The WHO Framework
of concern, such as ENDS [8] and Water pipes (also called hookah Convention on Tobacco
water pipes [9,10], have been intro- or shisha) were traditionally smoked Control is a public health
duced or more intensively marketed in the Middle East but are now also treaty that has been signed by
in high-income countries, and man- marketed worldwide [10]. Users draw 181 countries to protect their
ufactured cigarettes have gained smoke through a water chamber by populations from the dangers
market share in low- and middle- use of a long hose. The introduction of tobacco use. WHO Member
income countries. of mu‘assel (a molasses-soaked to- States have also pledged to
meet the target of a 30% rela-
Cigars, which consist of tobacco bacco mix) and fruit flavourings in
tive reduction in the preva-
that is wrapped in tobacco leaf, are the early 1990s increased the ap- lence of tobacco use by 2025.
available in many shapes and sizes, peal of water pipe smoking to youn-
■■ Without accelerated
including small, filtered cigars, which ger people [10]. progress, tobacco products
often appear indistinguishable from Cigars, pipes, and smokeless are projected to cause
cigarettes. Bidis, which are tradition- tobacco have all been determined 1 billion deaths this century,
ally smoked in India and Pakistan, are to cause cancer [1,2]. Tobacco many from cancer.
a form of hand-rolled tobacco made products other than cigarettes have
Chapter 2.1 • Tobacco products 51

Fig. 2.1.2. Examples of smokeless tobacco products, by country or region: South-East half of the ENDS market in the USA
Asia: kiwam, zarda, gutka; USA: moist snuff, dry snuff, moist snuff (caffeinated), plug, (Fig. 2.1.3). Heated tobacco prod-
twist tobaccos, dissolvables (orbs, strips, sticks, tobacco-coated toothpicks); Sweden: ucts, which heat tobacco [13] rather
snus (pouch); Venezuela: chimó; Uzbekistan: nasway; Sudan: toombak; India: red
toothpowder, mawa; Saudi Arabia: shammah; Brazil: rapé.
than a nicotine solution, are avail-
able in selected countries [14].
The eventual impact of e-ciga-
rettes and other putative harm-re-
duction products on health is not yet
known, but there are substantial con-
cerns. Although these products gen-
erally produce lower exposures to
toxic and carcinogenic compounds
than combusted tobacco does, us-
ers of these products may become
addicted to nicotine and transition
to more traditional forms of tobacco
use, including cigarettes and other
combustible products [15–18].
Biological impact of
tobacco products
Cigarettes
Cigarette smoke contains more
than 8000 compounds, includ-
ing more than 70 carcinogens [19].
Certain carcinogens are thought to
be particularly important, includ-
ing tobacco-specific nitrosamines,
polycyclic aromatic hydrocarbons,
and aromatic amines. The molecu-
lar mechanisms linking cigarettes to
cancer have been comprehensively
reviewed [2,19,20]. Nevertheless,
knowledge about the physiological
generally been much less studied the tobacco market in Europe, North and pathogenic consequences of
than cigarettes, despite their grow- America, and elsewhere. ENDS cigarette smoking continues to ex-
ing importance. heat a solution of nicotine without pand (see Chapter 3.11). For exam-
producing smoke [8]. ENDS were ple, over the past 5 years cigarette
Other nicotine and tobacco first sold by a pharmacist in China smoking has been linked to altered
products in 2003 and have been marketed patterns of circulating inflammatory
Other tobacco products also come in the USA since 2007. Although markers [21], altered DNA meth-
in many forms (Fig. 2.1.2) [11]. Some ENDS are supposedly marketed to ylation patterns [22], altered airway
traditional forms of smokeless to- adults, they often include flavours gene expression patterns [23], an al-
bacco include only tobacco, where- (such as strawberry and gummy tered oral microbiome [24], specific
as others include flavours and other bear) that are attractive to younger mutational signatures [25], and Y
constituents. In South-East Asia, people. ENDS products are diverse chromosome loss [26]. It is plausible
smokeless tobacco is widely used and are rapidly evolving. For exam- that non-cigarette tobacco products
with areca nut, lime, wood, and ash. ple, the Juul e-cigarette is a highly also cause many of these changes,
Another form of smokeless tobacco, engineered product that delivers a but fewer molecular studies on the
naswar, is commonly used in central high dose of nicotine and is a small, biological effects of these products
Asia. Naswar is frequently prepared discreet device. Its use was uncom- have been published.
by mixing lime and ground, pow- mon a few years ago, but as a result
dered tobacco. of marketing campaigns through Other combustible tobacco
During the past decade, novel social media [12] and the absence products
and emerging nicotine and tobacco of regulatory policies or under- Smokers of other combustible prod-
products have rapidly transformed regulation, it now makes up about ucts, including bidis, cigars, and
52
pipes, are exposed to many, if not ENDS Cancer types caused by

all, of the carcinogens found in Unlike other products described tobacco use
cigarette smoke [27]. Although wa- here, ENDS have emerged only
ter pipe smoking is less studied, it
during the past decade [8]. Typical Cigarettes
also generates high levels of car- With larger epidemiological stud-
ENDS products include nicotine,
cinogens and toxicants that are ies, longer follow-up, and better
glycerine, propylene glycol, and
not removed by passage through control for confounding, the num-
flavours in a liquid solution, which
water [28]. Water pipe smoking re- ber of sites or subsites of cancer
is then vaporized into an aerosol.
quires users to breathe very deeply known to be caused by cigarette
Laboratory studies indicate that
and, by doing so, replace much of smoking continues to increase.
SECTION 2
CHAPTER 2.1
ENDS devices generally heat to a
the air in the lungs with smoke, in The IARC Monographs [1] and the
lower temperature and have lower
contrast to the smaller puffs of ciga- United States Surgeon General [2]
rette smoke [9]. The charcoal used levels of most carcinogens than
designate causal relationships with
to ignite the tobacco in water pipe combusted cigarettes [29]. ENDS
at least 20 types of cancer, includ-
smoking seems to expose users to products also contain numerous
ing cancers of the lung, oral cav-
even higher levels of carbon mon- different flavourings, such as fruit
ity, nasal cavity and accessory si-
oxide and benzene compared with or caramel.
nuses, nasopharynx, oropharynx,
cigarette smokers [28]. hypopharynx, larynx, oesophagus
Heated tobacco products
(adenocarcinoma and squamous
Smokeless tobacco Heated tobacco products use a cell carcinoma), stomach, pancreas,
Smokeless tobacco is available in similar ignition system but use to- colorectum, liver, kidney (body and
many forms throughout the world [11]. bacco instead of a liquid [13,14]. pelvis), ureter, bladder, cervix, and
The levels of specific carcinogens Because of the rapidly changing ovary (mucinous), and acute mye-
vary across the different products, nature of these products [12–14], it loid leukaemia (Table 2.1.1). This list
but smokeless tobacco has been is important that their composition is conservative, because it does not
shown to contain at least 30 carcino- and carcinogen content be moni- include breast cancer or advanced
gens [11] and to release high levels of tored regularly by researchers inde- prostate cancer, two sites for which
tobacco-specific nitrosamines. pendent of the industry. the evidence for causality has been
Fig. 2.1.3. Sales (in millions of United States dollars) of e-cigarettes in Nielsen-tracked retail channels in the USA in 2011–2017,
by brand.
450
$ Millions
400
350
300
Total sales dollar
250
200
150
100
50
0
2011 2011 2011 2011 2012 2012 2012 2012 2013 2013 2013 2013 2014 2014 2014 2014 2015 2015 2015 2015 2016 2016 2016 2016 2017 2017 2017 2017
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
IMPERIAL TOBACCO (Blu) JAPAN Tobacco (LOGIC) ALTRIA (Markten & Green smoke) BAT (Vuse) Others NJOY JUUL

Table 2.1.1. Types of cancer caused by cigarette smoking
Cancer site or type Year formally classified by Year formally classified by Relative risk for current
the United States Surgeon the IARC Monographs versus never smoking
General
Men Women
Lip, oral cavity, pharynx 1964/1971a 1986 5.7 5.6
Oesophagus 1982 1986 3.9 5.1
Stomach 2004 2004 1.9 1.7
Colorectum 2014 2012 1.4 1.6
Liver 2014 2004 2.3 1.8
Pancreas 1982 1986 1.6 1.9
Larynx 1964 1986 13.9 103.8
Trachea, lung, bronchus 1964/1968 b
1986 25.3 22.9
Cervix 2004 2004 – 3.5
Bladder 1979 1986 3.9 3.9
Kidney, other urinary tract 1982 2004 1.8 1.2
Acute myeloid leukaemia 2004 2004 1.9 1.1
a
Lip cancer was classified as causal in 1964, and other oropharyngeal cancers in 1971.
b
Lung cancer was classified as causal in men in 1964 and in women in 1968.
labelled suggestive but not conclu- cally collected data on the preva- lion people (nearly two thirds of the
sive. Recent meta-analyses and lence and determinants of tobacco world’s population) [34].
pooled analyses have supported use. These data were largely limited
possible associations with these to smoked tobacco products. Since Patterns and trends in
sites [30,31]. then, population-based surveillance tobacco use
of tobacco use and tobacco control Descriptive studies of tobacco use
Non-cigarette tobacco has become a critical component of have often grouped all smoked to-
products and second-hand global tobacco control [34]. Several bacco products together and fo-
smoke multirisk-factor health surveys pro- cused on daily smoking, the most
The IARC Monographs have also vide nationally representative data common pattern [32,36,37]. In 2015,
concluded that cigar smoking and on schoolchildren and adults from an estimated 1.3 billion people
pipe smoking are strongly related an increasing number of countries; worldwide used tobacco products
to cancers of the lung and upper examples are the WHO STEPwise [3] and 1.1 billion people smoked, of
aerodigestive tract, including the approach to Surveillance (STEPS), which more than 80% smoked daily
oral cavity, oropharynx, hypophar- the Global Youth Tobacco Survey [7]. The prevalence of smoking is
ynx, larynx, and oesophagus [20]. (launched in 1999), and the Global higher in men than in women. About
Smokeless tobacco has been de- Adult Tobacco Survey (begun in 25% of men in the world are daily
termined to be causally related to 2007) [32]. smokers, compared with about 5%
cancers of the oesophagus, oral Six evidence-based measures of women [37]. Geographical pat-
cavity, and pancreas [1]. Exposure in line with the WHO Framework terns of smoking prevalence also
to second-hand smoke has been Convention on Tobacco Control differ by sex (Fig. 2.1.4). Among
determined to cause lung cancer have been identified or defined in men, the prevalence of daily smok-
[1,2]; associations with other can- the WHO MPOWER package for to- ing is highest in central and east-
cer types are less clear. bacco control [35]. These are moni- ern Europe and South-East Asia;
toring tobacco use and prevention among women, the prevalence is
policies (M), protecting people from highest in selected countries in east-
Surveillance of tobacco tobacco smoke (P), offering help to ern and western Europe (see the
use and tobacco control quit tobacco use (O), warning peo- interactive maps at the WHO Global
Population-based surveillance of to- ple about the harms of tobacco (W), Health Observatory; http://gamap
bacco use and tobacco control mea- enforcing bans on tobacco adver- server.who.int/gho/interactive_charts/
sures has expanded greatly in the tising, promotion, and sponsorship tobacco/use/atlas.html).
past decade [32]. When the WHO (E), and raising taxes on tobacco Overall, the age-standardized
Framework Convention on Tobacco (R). Since 2007, the number of peo- prevalence of daily smoking de-
Control [33] first entered into force ple protected by at least one best- creased from 1990 to 2015 in both
in 2005, only a few predominantly practice measure has more than men and women. An analysis of
high-income countries systemati- quadrupled, from 1 billion to 5 bil- 195 countries and territories by the
54
Fig. 2.1.4. Age-standardized prevalence of daily smoking for men (A) and women (B), in 2015.
SECTION 2
CHAPTER 2.1
Global Burden of Disease collabo- Global Burden of Disease project number of smokers in 2015, the larg-
ration estimated reductions of 28% had significant declines in the preva- est reduction in smoking prevalence
in men and 34% in women since lence of smoking in men, and 32 occurred in Brazil, where the preva-
1990 [37]. Similar reductions in had significant declines in the prevalence dropped by more than half be-
smoking prevalence have been re- lence in women [37]. The reductions tween 1990 and 2015 [37]. Pakistan,
ported in other studies [32,36]. were largest in high-income coun- Panama, and India are also notable
From 2005 to 2015, 53 of 195 tries and in Latin America [37]. Of for implementing numerous tobacco
countries and territories in the the 10 countries with the greatest control policies during the period

from 2005 to 2015 and having large 30% reduction in tobacco use in levels in the USA. According to the
declines in daily smoking prevalence men and women by 2025 [32]. National Youth Tobacco Survey, the
since 2005 [37]. Analyses of daily tobacco prevalence of any tobacco use in
Despite this encouraging prog smoking also have limitations. high school students fell from 24.2%
ress, the prevalence of tobacco use Combustible products other than in 2011 to 19.6% in 2017, and the
remains high worldwide, and prog cigarettes (pipes, cigars, bidis, etc.) prevalence of cigarette smoking fell
ress has been uneven. Indonesia predominate in some countries [38], from 15.8% in 2011 to 7.6% in 2017
has the highest recorded prevalence and nearly 20% of smokers world- (Fig. 2.1.5) [40]. However, the prev-
of smoking in men (46.7%) [37]. It is wide report occasional (non-daily) alence of smoking in adolescents
also the only country in South-East smoking [7], a pattern of exposure remains high in other countries, in-
Asia that has not signed the WHO to tobacco that itself appears to cluding in Europe. In 2015, 22 coun-
cause disease [39]. The need for tries had a smoking prevalence
Framework Convention on Tobacco
surveillance of dual use and use of above 15% in young women, and
Control. Four countries had signifi-
novel and emerging nicotine and to- 24 countries had a smoking preva-
cant increases in smoking preva-
bacco products (especially ENDS) lence above 20% in young men.
lence from 2005 to 2015: Congo and
is discussed below. Most of the countries with a high
Azerbaijan for men, and Kuwait and
prevalence in young women are in
Timor-Leste for women [37]. Smoking prevalence among Europe, whereas the countries with
There is concern about the fu- young people a high prevalence in young men are
ture impact of tobacco use in Africa. Current trends in smoking preva- in many world regions [37].
Although the prevalence of tobacco lence among young people are en-
smoking is currently relatively low in couraging. In the Global Burden of Number of smokers
most African countries, the impact Disease analysis, the prevalence Although there have been clear de-
of tobacco use is projected to rise of daily smoking among those clines in smoking prevalence world-
as a result of population growth, in- aged 15–19 years decreased be- wide, population growth has meant
creasing affluence, relatively weak tween 1990 and 2015, from 16.1% that trends in the absolute num-
tobacco control measures, and to 10.6% in males and from 4.8% ber of smokers worldwide are less
greater tobacco marketing [37]. to 3.0% in females [37]. The preva- clear. Conclusions about whether
Only one region, the Americas, is lence of cigarette smoking among the number of smokers is increas-
predicted to reach the target of a young people is at historically low ing, decreasing, or staying the same
Fig. 2.1.5. Estimated percentage of high school students who currently use any tobacco product, any combustible tobacco product,
two or more tobacco products, and selected tobacco products, from the National Youth Tobacco Survey, USA, in 2011–2017.
100
2011
2012
2013
30 2014
2015
2016
2017
25
Percentage of students
20
15
10
0
Any Any ≥2 types E-cigarettes Cigarettes Cigars Smokeless Hookah Pipe Bidis
combustible tobacco tobacco
Type of tobacco product
56
worldwide have differed in different less tobacco is more common in which accounts for almost 60% of
reports. A Global Burden of Disease men (237 million) than in women smokeless tobacco use in this age
analysis published in 2014 conclud- (129 million). Use of smokeless group worldwide [32].
ed that despite a decline in smoking tobacco was estimated to cause
prevalence from 1980 to 2012, the more than 101 000 cancer deaths Use of other nicotine and
number of daily smokers increased per year [41]. The Global Burden of tobacco products
from 721 million to 967 million [36]. Disease project published a com- Longitudinal information on the use
In contrast, the WHO global re- parable estimate, of 76 000 cancer of other nicotine and tobacco prod-
port on the prevalence of tobacco deaths per year from use of smoke- ucts is still limited. The available
use in 2000–2025, which included less tobacco [4]. Use of smokeless data indicate that water pipe smok-
SECTION 2
CHAPTER 2.1
both daily and occasional smoking, tobacco is common in every WHO ing is more common than cigarette
concluded that there was a modest region, each of which has at least smoking in many parts of the Middle
decrease in the number of smokers, 8 million users of smokeless to- East [10]; the highest reported prev-
from 1.14 billion in 2000 to 1.11 bil- bacco [32]. An estimated 82% of alence (almost 40%) is in adoles-
lion in 2015 [32]. users (301 million users) are in the
cent boys in Lebanon. Water pipe
WHO South-East Asia Region. The
use has also become commonplace
Tobacco smoking in low- and disease burden from smokeless to-
middle-income countries among young people worldwide. In
bacco use is substantial in that re-
the Eurobarometer survey, the prev-
About 80% of the world’s smok- gion. For example, it has been es-
alence of current water pipe smok-
ers live in low- and middle-income timated that 87% of cancer deaths
ing was 5% or higher in 11 European
countries. In addition, 64% of the from smokeless tobacco occur in
countries; the highest reported prev-
world’s daily smokers live in only 10 the South-East Asia Region [41].
Oral cancer is of particular concern alence (11.5%) was in Latvia [38].
countries [37], and more than 50%
in that region, reflecting the high Use of ENDS products has in-
of the world’s male smokers live in
prevalence of use of both smoke- creased rapidly over the past de
three countries: China, India, and
less tobacco and smoked tobacco cade in many countries, although
Indonesia [37]. Despite decreases
(cigarettes and bidis) [42]. surveillance data are largely re-
in smoking prevalence, the disease
In much of the world, children use stricted to high-income countries.
burden from tobacco use continues
to increase rapidly in low- and mid- smokeless tobacco. In every WHO In the USA, ENDS products have
dle-income countries, because of region except the European Region, become more popular than ciga-
the size and the growth of popula- there are at least 1 million young rettes among high school students
tions and the ageing of long-term, people aged 13–15 years who use aged 14–18 years: in 2017, 11.7%
continuing smokers [7]. smokeless tobacco [32]. The highest used ENDS and 7.6% used ciga-
prevalence in this age group is in the rettes (Fig. 2.1.5) [40]. It remains
Involuntary smoking South-East Asia Region (7.3% over- to be seen whether this pattern will
Involuntary smoking is the inhala- all; 9.5% in boys and 4.8% in girls), emerge in other countries. Rapid
tion of second-hand smoke by non-
smokers. In most countries, an es-
timated 15–50% of the population Fig. 2.1.6. Electronic nicotine delivery systems (ENDS) and the practice of vaping have
is exposed to second-hand smoke emerged only during the past decade.
(also called “environmental” tobacco
smoke); in some countries, expo-
sure to second-hand smoke affects
as much as 70% of the population
[7]. In China alone, an estimated
717 million people are exposed to
second-hand smoke at home [6].
Exposure to second-hand smoke
is estimated to cause more than
1.2 million deaths per year, of which
114 000 are deaths from cancer [4].
Use of smokeless tobacco

products
WHO has estimated that world-
wide there are more than 367 mil-
lion smokeless tobacco users aged
15 years or older [32]. Use of smoke-

changes in the design, flavours, using increased from 1.5 million per public health impact. For example,
age patterns, and names of these year in 1990 to 2.4 million per year a 50% increase in cigarette prices
products challenge surveillance in 2017 [4]. Nevertheless, effective in 13 middle-income countries in
efforts, particularly among young tobacco control could potentially Asia and Latin America with a to-
people [8,12]. To date, a range of prevent hundreds of millions of pre- tal of 2 billion men (500 million
regulatory approaches to these mature deaths [40]. male smokers) in their populations
products have been used in differ- As mentioned above, the ulti- would result in 450 million years of
ent countries [43]. mate impact of the shift towards life gained from smoking cessation
ENDS and other emerging products [48], with the largest gains among
Dual use and poly-use and dual use on cancer remains to lower-income individuals.
A growing proportion of tobacco be determined. Laboratory studies
users worldwide use more than can currently measure the carcino- Conclusions
one product. For example, in gen yield of novel products and bio-
Bangladesh, 22.5% of men who markers of exposure among users Tobacco products are well-estab-
use tobacco use both cigarettes [29] but cannot yet determine the lished causes of multiple types of
and smokeless tobacco. In India, potential long-term effects of these cancer. Tobacco control is, rightly,
19.4% of men who use tobacco are products on cancer risk or on the a poster child for public health inter-
dual users [6]. In a 2014 study in- use of more traditional tobacco ventions that use policy measures
cluding data from the 2008 to 2012 products. For example, cigarette and education to motivate behaviour
Global Adult Tobacco Survey and smokers may become dual users change. However, despite progress,
the Eurobarometer survey, at least of cigarettes and ENDS rather than the global health and economic bur-
20% of current smokers also used quitting smoking. Young people den of tobacco use remains enor-
another tobacco product in 28 of who become addicted to nicotine mous and is increasingly borne by
the 44 countries examined [38]. via ENDS may switch to cigarettes. low- and middle-income countries.
Among high school students in The United States Food and Drug Unfortunately, most countries are
the USA, dual use (9.2%) is now Administration Center for Tobacco not on track to achieve the global
more common than the use of ciga- Products is currently considering target of a 30% reduction in the
rettes alone (7.6%) (Fig. 2.1.5) [40]. reducing the nicotine content in cig- prevalence of tobacco use by 2025,
Among adults in the USA, most arettes, to encourage users to quit agreed to by WHO Member States.
ENDS users also use cigarettes cigarette smoking [46]. Such a pol- Furthermore, emerging tobacco
[44]. Determining the long-term icy would be expected to increase products challenge regulatory ap-
implications of these behavioural cessation of cigarette smoking but proaches to tobacco control and
changes on the burden of cancer would also be likely to encourage may undermine progress. Future
and other diseases is a critical re- users to switch to other products. research is needed to determine the
search and public health question. Global surveillance of the entire disease risks of emerging tobacco
range of tobacco products is critical products and to understand their ef-
Impact of continued smoking for understanding the future cancer fects on the use of established, and
on cancer burden and and public health impact of emerg- very harmful, traditional products.
smoking-attributable disease ing tobacco products. Continued tobacco and cancer sur-
Without dramatic global reductions veillance will also be needed to track
in cigarette use, the burden of to- the impact of public health interven-
bacco-related cancer and other dis-
Current and potential tions and to chart cancer rates.
eases will be substantially higher impact of tobacco control Without dramatic reductions in
in the future than it is now. In the Tobacco control policies have been tobacco use, the number of cancer
USA and other high-income coun- demonstrated to save lives. It has deaths per year caused by tobacco,
tries, declines in smoking preva- been estimated that tobacco con- which is already very large, is pro-
lence have resulted in substantial trol resulted in 8 million fewer jected to increase further, reflecting
decreases in incidence rates of premature deaths in 1964–2012 demographic factors and the global
lung cancer and laryngeal cancer in the USA [45]. Similarly, an esti- maturation of the tobacco epidemic,
[2,45]. Elsewhere, and especially in mated 22 million deaths were pre- and to cause 1 billion deaths world-
low- and middle-income countries, vented in 2007–2014 in 88 coun- wide this century [49]. Accelerated
the cancer burden from smoking tries that adopted at least one progress in tobacco control is ur-
continues to increase as a result of highest-level MPOWER policy [47]. gently needed. Monitoring of trends
population growth and the ageing Nevertheless, MPOWER and other in age-specific incidence or death
of smokers [37]. The Global Burden tobacco control interventions are rates from lung cancer at younger
of Disease collaboration has es- underutilized [7]. Accelerated images can be especially informative
timated that the number of cancer plementation of tobacco control in this regard.
deaths caused by tobacco smok- measures would have an enormous
58
References
1. IARC (2012). Personal habits and in- 8. Grana R, Benowitz N, Glantz SA (2014). 17. Siddiqui F, Mishu M, Marshall AM, Siddiqi
door combustions. IARC Monogr Eval E-cigarettes: a scientific review. Circ u K (2019). E-cigarette use and subsequent
Carcinog Risks Hum. 100E:1–575. lation. 129(19):1972–86. https://doi.org/ smoking in adolescents and young adults:
PMID:23193840. Available from: http:// 10.1161/CIRCULATIONAHA.114.007667 a perspective. Expert Rev Respir Med.
publications.iarc.fr/122. PMID:24821826 13(5):403–5. http://doi.org/10.1080/17476
348.2019.1589371 PMID:30822173
2. U.S. Department of Health and Human 9. Cobb C, Ward KD, Maziak W, Shihadeh
Services (2014). The health conse- AL, Eissenberg T (2010). Waterpipe to- 18. Jenssen BP, Walley SC; Section on
quences of smoking – 50 years of prog bacco smoking: an emerging health crisis Tobacco Control (2019). E-cigarettes
SECTION 2
CHAPTER 2.1
ress: a report of the Surgeon General. in the United States. Am J Health Behav. and similar devices. Pediatrics. 143(2):
Atlanta (GA), USA: U.S. Department of 34(3):275–85. https://doi.org/10.5993/ e20183652. http://doi.org/10.1542/peds.
Health and Human Services, Centers for AJHB.34.3.3 PMID:20001185 2018-3652 PMID:30835247
Disease Control and Prevention, National
Center for Chronic Disease Prevention 10. Maziak W, Taleb ZB, Bahelah R, Islam F, 19. Hecht SS, Szabo E (2014). Fifty years
and Health Promotion, Office on Smoking Jaber R, Auf R, et al. (2015). The global of tobacco carcinogenesis research:
and Health. Available from: https://www. epidemiology of waterpipe smoking. Tob from mechanisms to early detection
ncbi.nlm.nih.gov/books/NBK179276/. Control. 24(Suppl 1):i3–12. https://doi. and prevention of lung cancer. Cancer
org/10.1136/tobaccocontrol-2014-051903 Prev Res (Phila). 7(1):1–8. https://doi.
3. WHO (2019). WHO global report on trends PMID:25298368 org/10.1158/1940 - 6207.CAPR-13 - 0371
in prevalence of tobacco use 2000–2025. PMID:24403288
11. National Cancer Institute and Centers for
3rd ed. Geneva, Switzerland: World Health
Disease Control and Prevention (2014). 20. IARC (2004). Tobacco smoke and in-
Organization. Available from: https://
Smokeless tobacco and public health: voluntary smoking. IARC Monogr
w w w.who.int /publications- detail/who -
a global perspective (NIH Publication Eval Carcinog Risks Hum. 83:1–1438.
global-report-on-trends-in-prevalence-
No. 14-7983). Bethesda (MD), USA: PMID:15285078. Available from: http://
of-tobacco-use-2000-2025-third-edition.
U.S. Department of Health and Human publications.iarc.fr/101.
4. GBD 2017 Risk Factor Collaborators Services, Centers for Disease Control
and Prevention, and National Institutes of 21. Shiels MS, Katki HA, Freedman ND,
(2018). Global, regional, and national
Health, National Cancer Institute. Purdue MP, Wentzensen N, Trabert B,
comparative risk assessment of 84 be-
et al. (2014). Cigarette smoking and
havioural, environmental and occupa-
12. Huang J, Duan Z, Kwok J, Binns S, Vera variations in systemic immune and in-
tional, and metabolic risks or clusters of
LE, Kim Y, et al. (2019). Vaping ver- flammation markers. J Natl Cancer Inst.
risks for 195 countries and territories,
sus JUULing: how the extraordinary 106(11):dju294. https://doi.org/10.1093/
1990-2017: a systematic analysis for the
growth and marketing of JUUL trans- jnci/dju294 PMID:25274579
Global Burden of Disease Study 2017.
formed the US retail e-cigarette market.
Lancet. 392(10159):1923–94. https://doi. 22. Joehanes R, Just AC, Marioni RE, Pilling LC,
Tob Control. 28(2):146–51. https://doi.
org /10.1016/S0140 - 673 6(18) 32 2 25 - 6 Reynolds LM, Mandaviya PR, et al. (2016).
org/10.1136/tobaccocontrol-2018-054382
PMID:30496105 Epigenetic signatures of cigarette smok-
PMID:29853561
ing. Circ Cardiovasc Genet. 9(5):436–47.
5. U.S. Department of Health and Human
13. Simonavicius E, McNeill A, Shahab L, https://doi.org/10.1161/CIRCGENETICS.
Services (2006). The health consequences
Brose LS (2019). Heat-not-burn tobacco 116.001506 PMID:27651444
of involuntary exposure to tobacco smoke:
products: a systematic literature review.
a report of the Surgeon General. Atlanta 23. Billatos E, Faiz A, Gesthalter Y, LeClerc A,
Tob Control. 28(5):582–94. https://doi.
(GA), USA: U.S. Department of Health Alekseyev YO, Xiao X, et al. (2018). Impact
and Human Services, Centers for Disease of acute exposure to cigarette smoke on air-
PMID:30181382
Control and Prevention, Coordinating way gene expression. Physiol Genomics.
Center for Health Promotion, National 14. Tabuchi T, Gallus S, Shinozaki T, Nakaya 50(9):705–13. https://doi.org/10.1152/physi
Center for Chronic Disease Prevention T, Kunugita N, Colwell B (2018). Heat- olgenomics.00092.2017 PMID:29932825
and Health Promotion, Office on Smoking not-burn tobacco product use in Japan:
and Health. Available from: https://www. its prevalence, predictors and perceived 24. Wu J, Peters BA, Dominianni C, Zhang
cdc.gov/tobacco/data_statistics/sgr/2006/ symptoms from exposure to second- Y, Pei Z, Yang L, et al. (2016). Cigarette
index.htm. hand heat-not-burn tobacco aerosol. smoking and the oral microbiome in a
Tob Control. 27(e1):e25–33. https://doi. large study of American adults. ISME J.
6. Asma S, Mackay J, Song SY, Zhao L, 10(10):2435–46. https://doi.org/10.1038/
Morton J, Palipudi KM, et al. (2015). The ismej.2016.37 PMID:27015003
PMID:29248896
GATS atlas: Global Adult Tobacco Survey.
Atlanta (GA), USA: CDC Foundation. 15. Soneji S, Barrington-Trimis JL, Wills TA, 25. Alexandrov LB, Ju YS, Haase K, Van
Available from: http://gatsatlas.org/. Leventhal AM, Unger JB, Gibson LA, et Loo P, Martincorena I, Nik-Zainal S, et
al. (2017). Association between initial use al. (2016). Mutational signatures asso-
7. U.S. National Cancer Institute and World of e-cigarettes and subsequent cigarette ciated with tobacco smoking in human
Health Organization (2016). Patterns smoking among adolescents and young cancer. Science. 354(6312):618–22.
of tobacco use, exposure, and health adults: a systematic review and meta-anal- https://doi.org/10.1126/science.aag0299
consequences. In: The economics of to- ysis. JAMA Pediatr. 171(8):788–97. https:// PMID:27811275
bacco and tobacco control. NCI Tobacco doi.org/10.1001/jamapediatrics.2017.1488
Control Monograph 21. NIH Publication 26. Dumanski JP, Rasi C, Lönn M, Davies H,
PMID:28654986 Ingelsson M, Giedraitis V, et al. (2015).
No. 16-CA-8029A. Bethesda (MD), USA:
U.S. Department of Health and Human 16. Stanton CA, Bansal-Travers M, Johnson Mutagenesis. Smoking is associated with
Services, National Institutes of Health, AL, Sharma E, Katz L, Ambrose BK, et mosaic loss of chromosome Y. Science.
National Cancer Institute; and Geneva, al. (2019). Longitudinal e-cigarette and 347(6217):81–3. https://doi.org/10.1126/
Switzerland: World Health Organization; cigarette use among US youth in the PATH science.1262092 PMID:25477213
pp. 23–70. Available from: https://can study (2013–2015). J Natl Cancer Inst.
c erc ontrol.c anc er.gov/ br p/ tc r b/mono 111(10):1088–96. https://doi.org/10.1093/
graphs/21/index.html. jnci/djz006 PMID:30689915

27. Chen J, Kettermann A, Rostron BL, Day HR 35. WHO (2008). MPOWER: a policy package 43. Kennedy RD, Awopegba A, De León
(2014). Biomarkers of exposure among U.S. to reverse the tobacco epidemic. Geneva, E, Cohen JE (2017). Global approach-
cigar smokers: an analysis of 1999-2012 Switzerland: World Health Organization. es to regulating electronic cigarettes.
National Health and Nutrition Examination Available from: http://www.who.int/tobacco/ Tob Control. 26(4):440–5. https://doi.
Survey (NHANES) data. Cancer Epidemiol mpower/mpower_english.pdf. org/10.1136/tobaccocontrol-2016-053179
Biomarkers Prev. 23(12):2906–15. https:// PMID:27903958
36. Ng M, Freeman MK, Fleming TD,
doi.org/10.1158/1055-9965.EPI-14-0849 Robinson M, Dwyer-Lindgren L, Thomson 44. CDC (2016). QuickStats: Cigarette smok-
PMID:25380733 B, et al. (2014). Smoking prevalence and ing status* among current adult e-cigarette
cigarette consumption in 187 countries, users,† by age group – National Health
28. Jacob P 3rd, Abu Raddaha AH, Dempsey
1980-2012. JAMA. 311(2):183–92. https: Interview Survey,§ United States, 2015.
D, Havel C, Peng M, Yu L, et al. (2013).
//d o i.o r g /10.10 01/ j am a. 2 013 . 28 4 6 9 2 MMWR Morb Mortal Wkly Rep. 65(42):1177.
Comparison of nicotine and carcino-
PMID:24399557 https://doi.org/10.15585/mmwr.mm6542a7
gen exposure with water pipe and
cigarette smoking. Cancer Epidemiol 37. GBD 2015 Tobacco Collaborators (2017). PMID:27787495
Biomarkers Prev. 22(5):765–72. https:// Smoking prevalence and attributable dis- 45. Holford TR, Meza R, Warner KE, Meernik
doi.org/10.1158/1055-9965.EPI-12-1422 ease burden in 195 countries and territo- C, Jeon J, Moolgavkar SH, et al. (2014).
PMID:23462922 ries, 1990-2015: a systematic analysis from
Tobacco control and the reduction in smok-
the Global Burden of Disease Study 2015.
29. Shahab L, Goniewicz ML, Blount BC, ing-related premature deaths in the United
Lancet. 389(10082):1885–906. https://
Brown J, McNeill A, Alwis KU, et al. States, 1964-2012. JAMA. 311(2):164–71.
doi.org/10.1016/S0140-6736(17)30819-X
https://doi.org/10.1001/jama.2013.285112
(2017). Nicotine, carcinogen, and toxin PMID:28390697
PMID:24399555
exposure in long-term e-cigarette and
nicotine replacement therapy users: a 38. Agaku IT, Filippidis FT, Vardavas CI,
Odukoya OO, Awopegba AJ, Ayo-Yusuf 46. Apelberg BJ, Feirman SP, Salazar E,
cross-sectional study. Ann Intern Med. Corey CG, Ambrose BK, Paredes A,
OA, et al. (2014). Poly-tobacco use among
166(6):390–400. https://doi.org/10.7326/ et al. (2018). Potential public health ef-
adults in 44 countries during 2008-2012:
M16-1107 PMID:28166548 fects of reducing nicotine levels in ciga-
evidence for an integrative and compre-
hensive approach in tobacco control. Drug rettes in the United States. N Engl J Med.
30. Gaudet MM, Carter BD, Brinton LA, Falk
Alcohol Depend. 139:60–70. https://doi. 378(18):1725–33. https://doi.org/10.1056/
RT, Gram IT, Luo J, et al. (2017). Pooled
org/10.1016/j.drugalcdep.2014.03.003 NEJMsr1714617 PMID:29543114
analysis of active cigarette smoking and
PMID:24685560
invasive breast cancer risk in 14 cohort 47. Levy DT, Yuan Z, Luo Y, Mays D (2018).
studies. Int J Epidemiol. 46(3):881– 39. Inoue-Choi M, Liao LM, Reyes-Guzman Seven years of progress in tobacco con-
93. http://doi.org/10.1093/ije/dyw288 C, Hartge P, Caporaso N, Freedman ND trol: an evaluation of the effect of nations
PMID:28031315 (2017). Association of long-term, low-in- meeting the highest level MPOWER mea
tensity smoking with all-cause and cause- sures between 2007 and 2014. Tob Control.
31. Islami F, Moreira DM, Boffetta P, specific mortality in the National Institutes 27(1):50–7. https://doi.org/10.1136/tobac
Freedland SJ (2014). A systematic review of Health-AARP Diet and Health Study. cocontrol-2016-053381 PMID:27956650
and meta-analysis of tobacco use and JAMA Intern Med. 177(1):87–95. https://
prostate cancer mortality and incidence doi.org/10.1001/jamainternmed.2016.7511 48. Global Tobacco Economics Consortium
in prospective cohort studies. Eur Urol. PMID:27918784 (2018). The health, poverty, and finan-
66(6):1054–64. https://doi.org/10.1016/j. cial consequences of a cigarette price
eururo.2014.08.059 PMID:25242554 40. Wang TW, Gentzke A, Sharapova S, increase among 500 million male smok-
Cullen KA, Ambrose BK, Jamal A (2018). ers in 13 middle income countries: com-
32. WHO (2018). WHO global report on Tobacco product use among middle and partmental model study. BMJ. 361:k1162.
trends in prevalence of tobacco smok- high school students - United States, h t t p s : / / d o i . o r g / 1 0 .11 3 6 / b m j . k 11 6 2
ing 2000–2025. 2nd ed. Geneva, 2011-2017. MMWR Morb Mortal Wkly Rep. PMID:29643096
Switzerland: World Health Organization. 67(22):629–33. https://doi.org/10.15585/
Available from: https://www.who.int/ mmwr.mm6722a3 PMID:29879097 49. Jha P, Peto R (2014). Global effects
tobacco/publications/sur veillance/ of smoking, of quitting, and of taxing
41. Sinha DN, Suliankatchi RA, Gupta PC, tobacco. N Engl J Med. 370(1):60–8.
trends-tobacco-smoking-second-edition/ Thamarangsi T, Agarwal N, Parascandola
en/. https://doi.org/10.1056/NEJMra1308383
M, et al. (2018). Global burden of all-cause PMID:24382066
and cause-specific mortality due to smoke-
33. WHO (2003). WHO Framework Con
less tobacco use: systematic review and
vention on Tobacco Control. Geneva,
meta-analysis. Tob Control. 27(1):35–42.
https://doi.org/10.1136/tobaccocontrol-
Available from: https://www.who.int/fctc/ 2016-053302 PMID:27903956
text_download/en/.
42. Shield KD, Ferlay J, Jemal A,
34. WHO (2019). WHO report on the global Sankaranarayanan R, Chaturvedi AK,
tobacco epidemic, 2019: offer help to quit Bray F, et al. (2017). The global inci-
tobacco use. Geneva, Switzerland: World dence of lip, oral cavity, and pharyngeal
Health Organization. Licence: CC BY- cancers by subsite in 2012. CA Cancer J
NC-SA 3.0 IGO. Available from: https:// Clin. 67(1):51–64. https://doi.org/10.3322/
www.who.int/tobacco/global_report/en/. caac.21384 PMID:28076666
60
2.2
Infectious agents
Missed opportunities for prevention
Robert Newton Martyn Plummer (reviewer)
CHAPTER 2.2
SECTION 2
Catherine de Martel You-Lin Qiao (reviewer)
●● Preventive vaccines against hep- and in the lowest-income regions of

SUMMARY atitis B virus have been available the world [3]. In many high-income
since 1982, and direct-acting an- countries in Australasia, Europe,
●● Infectious agents are an impor- tiviral agents have the potential and North America, fewer than 5% of
tant cause of cancer, particularly to cure more than 95% of people cancer cases are attributable to in-
in low- and middle-income coun- with hepatitis C virus infection. fections. In countries in sub-Saharan
tries, which have limited ability to Africa, the proportion is at least one
manage the disease; therefore, third; this may be an underestimate,
prevention is a priority. The IARC Monographs programme
because there is limited cancer reg-
has classified 11 infectious agents,
●● The bacterium Helicobacter py- istration in many countries in this re-
or groups of related agents, as car-
lori was estimated to be respon- gion, and almost none in rural areas.
cinogenic to humans (Group 1) [1].
sible for about 810 000 new Four infectious agents – H. py-
These include one bacterium, seven
cancer cases in 2018, including lori, HPVs, HBV, and HCV – were to-
viruses, and three macroparasites.
89% of non-cardia gastric can- gether responsible for about 2 million
The bacterium is Helicobacter py-
cers (760 000 cases), 74% of cancer cases in 2018 (Table 2.2.1).
lori. The viruses are human papil-
gastric non-Hodgkin lymphoma More than one third of infection-relat-
lomaviruses (HPVs), 13 subtypes
(22 000 cases) and 29% of car- ed cancer cases occurred in China,
of which are classified as carcino-
dia gastric cancers in East Asia where 42% of all H. pylori-related
(36 000 cases). Treatment by a genic, hepatitis B virus (HBV), hep-
cancers and 69% of all HBV-related
combination of anti-microbial atitis C virus (HCV), Epstein–Barr
cancers occurred. Among the other
drugs is potentially preventive. virus (EBV), Kaposi sarcoma-as-
infectious agents, several, includ-
sociated herpesvirus (KSHV), hu-
ing HTLV-1 and the macroparasites,
●● Thirteen sexually transmitted man T-cell lymphotropic virus type contribute little to the global cancer
mucosal human papillomavirus 1 (HTLV-1), and HIV-1. The macro- burden but are significant causes of
subtypes are established hu- parasites are Schistosoma haema-
man carcinogens. Together, they cancer in endemic populations. (For
tobium, Opisthorchis viverrini, and a recent, extensive review of infec-
are responsible for all cervical Clonorchis sinensis. Each of these
cancer cases globally (570 000 tions and cancer, see [4]).
infectious agents causes at least
cases) and a variable proportion one type of cancer, and some cause
of cases of other anogenital and several cancer types (Table 2.2.1). Helicobacter pylori
oropharyngeal cancers (total- The burden of cancer associated The bacterium H. pylori was esti-
ling 120 000 cases). Vaccination with chronic infections is substan- mated to be responsible for about
against human papillomaviruses tial. It is estimated that in 2018, out 810 000 new cancer cases in 2018,
occurs in more than 80 countries. of a total of 18 million new cancer including 89% of non-cardia gas-
●● Chronic infection with hepatitis cases worldwide, 2.2 million – about tric cancers (760 000 cases), 74%
B virus and hepatitis C virus re- one eighth of all new cases – were of gastric non-Hodgkin lymphoma
sulted in about 360 000 cases caused by infection [2] (Table 2.2.1). cases (22 000 cases), and 29% of
and 140 000 cases, respective- However, the proportion of cancer cardia gastric cancers in East Asia
ly, of hepatocellular carcinoma cases caused by infection varies (36 000 cases) [2]. In addition, H.
in 2018, amounting to about markedly by geographical region pylori causes substantial morbidity
76% of all cases of hepatocel- and World Bank income group; it and mortality from peptic ulcer dis-
lular carcinoma. is substantially higher in East Asia ease. Millions of cases of duodenal
Chapter 2.2 • Infectious agents 61

Table 2.2.1. Estimated numbers of new cancer cases in 2018 attributable to infec-
tious agents
FUNDAMENTALS
Infectious agent Cancer types for which there is Number of new
sufficient evidence of causality cancer cases ■■ Eleven infectious agents, or groups
of related agents, are established
Helicobacter pylori Non-cardia gastric carcinoma, low- 810  000
grade B-cell mucosa-associated
human carcinogens, including
lymphoid tissue (MALT) gastric one bacterium, seven viruses, and
lymphoma three macroparasites.
Human papillomavirus Carcinomas of the cervix, vulva, vagina, 6 90  000 ■■ About 13% of cancers worldwide,
penis, anus, oral cavity, oropharynx, or 2.2 million cases per year, are
and tonsil caused by chronic infections. This
Hepatitis B virus Hepatocellular carcinoma 3 60  000 proportion varies by geographical
(chronic infection) region and World Bank income
group; it is highest in the lowest-
Hepatitis C virus Hepatocellular carcinoma, non-Hodgkin 160  000 income regions, especially for
lymphoma
cervical cancers caused by
Epstein–Barr virus Nasopharyngeal carcinoma, Burkitt 160  000 human papillomaviruses. In sub-
lymphoma, immunosuppression-related Saharan Africa, at least one third
non-Hodgkin lymphoma, extranodal of cancer cases are of infectious
NK/T-cell lymphoma (nasal type),
Hodgkin lymphoma
origin, and the proportion may
be significantly underestimated,
Kaposi sarcoma- Kaposi sarcoma, primary effusion 42 000 because there is limited cancer
associated herpesvirus lymphoma registration in many countries in
Human T-cell Adult T-cell leukaemia/lymphoma 3 600
this region.
lymphotropic virus type 1
■■ Four agents – Helicobacter pylori,
HIV-1 Kaposi sarcoma, non-Hodgkin – a human papillomaviruses, hepatitis
lymphoma, Hodgkin lymphoma, cervical B virus, and hepatitis C virus –
cancer, anal cancer, conjunctival cancer contribute most to the burden
of cancer caused by infections
Schistosoma Bladder cancer 6 000
haematobium globally. Several carcinogenic
infectious agents, including H.
Opisthorchis viverrini Cholangiocarcinoma 3 600 pylori, hepatitis B virus, hepatitis
Clonorchis sinensis C virus, Epstein–Barr virus, HIV,
a
Cancers attributable to HIV are included with the underlying causal infections. and macroparasites, also cause
substantial morbidity and mortality
from non-malignant diseases.
and gastric ulcer diseases are diag- H. pylori infection has been declin-
■■ Some cancer-causing infections,
nosed each year globally, although ing in tandem with the occurrence of
such as infections with
the proportion attributable to H. py- the diseases it causes, and is now macroparasites, contribute little
lori is unclear (see Chapter 5.4) [5]. rare in children and young adults. to the global cancer burden but
H. pylori is a highly adapted bac- However, gastric cancer tends to oc- are significant causes of cancer
terium that is able to live in the acidic cur at an advanced age (≥ 65 years) in endemic populations.
environment of the human gastric compared with other infection-relat-
■■ Human papillomavirus and
mucosa, where it causes chronic in- ed cancers. Because of global pop-
hepatitis B virus infections are
flammation, which may slowly lead to ulation growth and ageing, the total
amenable to primary prevention
fibrosis, atrophy, and ultimately can- number of H. pylori-related gastric through vaccination. Infections
cer in a small proportion of infected cancer cases is not expected to de- with hepatitis C virus, H. pylori,
individuals, usually after several decrease for decades. and the macroparasites are
cades. Infection often occurs during The treatment for H. pylori incurable. For HIV and hepatitis
childhood, and in the absence of fection comprises a combination of B virus, infections can be
treatment by an effective combina- antimicrobial drugs and a proton- controlled by antiviral treatment
tion of three or four antimicrobial pump inhibitor and is used widely to reduce the risk of cancer and
drugs, the infection is lifelong. in symptomatic individuals. Mass of transmission to others.
H. pylori transmission occurs via treatment provides a means of ■■ If existing strategies for prevention
oral–oral and faecal–oral routes cancer prevention, although stud- were more widely applied and
within the family and is considerably ies are bedevilled by the need for new infection control strategies
more frequent among people with large numbers and lengthy follow- developed, the global cancer
low socioeconomic status. In high- up; there may also be deleterious burden could be greatly reduced.
income countries, the prevalence of consequences in terms of drug
62
resistance and the unknown im- sexes in low- and lower-middle- disease, has resulted in substantial
pact of changes to the microbi- income countries, where screening declines in cervical cancer mortal-
ome. However, the evidence from for early cervical disease is limited ity in high-income countries but is
seven published studies (reviewed and where the prevalence of HPV often unavailable in low- and lower-
in [4] and [6], with an additional infection and of risk factors such as middle-income countries.
study published more recently [7]) early age at first sexual intercourse Over the past 10–15 years, safe
indicates that H. pylori eradication and co-infection with HIV is high. and effective HPV vaccination, in-
programmes can be effective. The The risk of cancer associated cluding bivalent, quadrivalent, and
adoption of further screen-and-treat with HPV can be reduced with a nonavalent vaccines, has been in-
strategies has been recommended, combination of factors that limit ei- troduced in more than 80 countries.
CHAPTER 2.2
SECTION 2
together with trials of screening for ther risk of infection or risk of dis- However, most of these are high-
early disease using non-invasive ease: using safe sexual practices and upper-middle-income countries
pepsinogen testing. The recently (including delayed start of sexual rather than low- and lower-middle-
initiated GISTAR study aims to test activity), male circumcision, and income countries, which have the
the impact of the combination of H. reduction in tobacco use, which is highest burden of HPV-associated
pylori eradication and screening for an important co-factor for cervical disease [10]. About 20 of these coun-
early disease on the gastric cancer cancer and oropharyngeal can- tries either already vaccinate boys
burden, and has a 15-year follow- cers in particular. Cervical cancer in addition to girls or plan to do so.
up period [8]. screening, for detection of early National vaccination programmes
An effective prophylactic or ther-
apeutic vaccine against H. pylori
would provide a cheaper and more Fig. 2.2.1. A woman and her baby daughter in Cochabamba, Bolivia. Like many other
effective way to reduce disease low- and middle-income countries, Bolivia has a high mortality rate for cervical cancer,
which is caused by human papillomavirus (HPV) infection. Cervical cancer screening,
risk, particularly in low- and lower-
which has resulted in substantial declines in mortality in high-income countries, is often
middle-income countries, which unavailable in low- and lower-middle-income countries.
have limited health infrastructure.
Vaccine-related activities are sum-
marized in [9]; all of the vaccines
currently under development are at
an early stage, and there appears
to be little, if any, investment from
large pharmaceutical companies,
without which progress is likely to
be limited.
Human papillomaviruses
Thirteen sexually transmitted mu-
cosal HPV subtypes have been clas-
sified as carcinogenic to humans.
Together, they are responsible for
all cervical cancer cases globally
(570 000 cases) and a variable pro-
portion of cases of other anogenital
and oropharyngeal cancers (totalling
120 000 cases) [2]. The most affect-
ed region of the world is sub-Saha-
ran Africa, where about 60% of all in-
fection-associated cancer cases are
caused by HPV (see Chapter 5.10).
In every world region, two sub-
types, HPV16 and HPV18, are re-
sponsible for about 70% of cervical
cancer cases. HPVs are responsi-
ble for more than half of all infec-
tion-associated cancers in women
worldwide and for about half of all
infection-associated cancers in both

with more than 50% coverage of Fig. 2.2.2. A girl in Zambia receives a human papillomavirus (HPV) vaccination. Safe
two- or three-dose schedules have and effective HPV vaccination has been introduced in more than 80 countries.
been shown to have a big impact
in decreasing HPV prevalence and
persistence and rates of cervical in-
traepithelial neoplasia (a precursor
of cervical cancer) [11]. The quad-
rivalent and nonavalent vaccines are
also highly effective at preventing
anogenital warts, caused by HPV6
and HPV11.
Although there has been con-
siderable progress in the deploy-
ment of HPV vaccination, many
years will need to go by before the
impact on cancer will be fully evi-
dent (see Chapter 6.3). Therefore,
cervical screening programmes, in
particular using HPV-based point-
of-care testing where available, will
need to be maintained for the fore-
seeable future, to protect cohorts of
unvaccinated women.
The barriers to HPV vaccination
are greatest in those countries with
the weakest health systems and the occur in Asia and in sub-Saharan of liver damage, a daily dose of anti-
highest burden of HPV-associated Africa, reflecting the prevalence of viral therapy, using widely available
disease. To maintain HPV vacci- the virus and the age at which infec- drugs, is effective in most people at
nation as a key element of cancer tion commonly occurs. reducing complications and trans-
control programmes globally, and The predominant modes of transmission to others, although treat-
to introduce it in other settings, will mission of HBV infection are peri- ment needs to be maintained for
require major international commit- natal, parenteral, and sexual. The life. Prevention of mother-to-child
ment and funding. If current efforts risk of chronic carriage, and hence transmission can be improved via
to establish the efficacy of single- of cancer, is related to the age at a combination of routine antenatal
dose vaccination prove viable, this screening, antiviral drugs during
infection. The risk is highest among
would remove some of the barriers pregnancy, and HBV vaccination
people infected as infants, of whom
to wider deployment in low- and of the baby at birth; administration
about 90% become chronic carri-
lower-middle-income countries. of HBV immunoglobulin can further
ers; this is the predominant mode
of transmission in Asia. The risk is reduce the risk of vertical transmis-
Hepatitis B virus intermediate among those infected sion. Coverage of the birth dose of
during childhood, of whom 30–50% HBV vaccine is thought to be about
Globally, more than 260 million peo-
become chronic carriers; this is the 39% globally. However, with a laten-
ple are estimated to be chronic carri-
predominant mode of transmission cy period from infection to cancer
ers of HBV, of whom 1–2% per year
in sub-Saharan Africa. The risk is of 30–40 years, it will be decades
will progress to liver disease; more
lowest among those infected as before the impacts of prevention ef-
than 90% are unaware of their status
forts are felt, highlighting the need
(see Chapter 5.6) [4]. Chronic HBV adults, of whom less than 5% be-
for screen-and-treat strategies in
infection resulted in about 360 000 come chronic carriers; this mode of
high-risk populations in the interim.
cases of hepatocellular carcinoma in transmission occurs mainly in high-
2018, amounting to about 55% of all income countries.
cases of hepatocellular carcinoma. Safe and effective preventive Hepatitis C virus
In addition, there is substantial mor- vaccines against HBV have been Approximately 200 million people
tality from non-malignant manifesta- available since 1982. Global cover- worldwide are estimated to be in-
tions of infection, with about 890 000 age is thought to be about 84%, al- fected with HCV. Chronic HCV infec-
HBV-related deaths, including those though there is evidence from rural tion resulted in about 160 000 new
from cancer, per year [12]. The larg- sub-Saharan Africa that this may cancer cases in 2018, predominant-
est proportion of HBV-associated be an overestimate [13]. For adults ly cases of hepatocellular carcino-
cases of hepatocellular carcinoma with chronic infection and evidence ma but also about 16 000 cases of
64
Fig. 2.2.3. A patient with liver cancer at the National Cancer Center of Mongolia, in middle-income countries, the peak
Ulaanbaatar. Chronic infection with hepatitis B virus resulted in about 360 000 cases prevalence of infection is within the
of hepatocellular carcinoma in 2018. first years of life, but in high-income
countries, only about 45–50% of
people are infected as infants.
Transmission is mainly via saliva,
although it can also occur via blood
[1,2,4].
In 2007, a vaccine against the
EBV gp350 antigen was shown
CHAPTER 2.2
SECTION 2
in a phase 2 trial to prevent infec-
tious mononucleosis, although it
did not prevent infection with EBV
(reviewed in [15]). However, since
then, further work both on that vac-
cine candidate and on others has
stalled, and currently no trials are
under way. A vaccine to prevent
EBV-related post-transplant lym-
phoma would provide an important
proof of principle for the prevention
of EBV-associated cancer. Trials to
reduce the incidence of other EBV-
associated cancers would be chal-
lenging, but feasible.
non-Hodgkin lymphoma [2]. In low- antiviral agents in 2014 has resulted Kaposi sarcoma-
and middle-income countries, the in cure rates of greater than 90% in associated herpesvirus
predominant cause of hepatocel- treated individuals, with minimal side-
KSHV is a necessary but not suf-
lular carcinoma is HBV, but in high- effects. However, the complexity of
ficient cause of Kaposi sarcoma,
income countries, 40% of cases are testing for HCV and the high cost of
primary effusion lymphoma, and
caused by HCV; in Japan, the pro- treatment mean that treatment is cur-
probably also multicentric Castleman
portion is up to 60% [1,2]. About 75– rently unavailable to most of the peo-
disease. KSHV caused about 42 000
85% of infections become chronic, ple who would benefit, even in high-
cancer cases in 2018, predominantly
and in the absence of treatment ap- income countries [14].
in HIV-infected people, in whom the
proximately half of the chronic carri- resulting immunosuppression facili-
ers will die of liver disease. Epstein–Barr virus tates the development of cancer [1,2].
The prevalence of HCV infec- In 2018, EBV was estimated to have KSHV is unique among the her-
tion varies widely; it is highest in caused 160 000 new cancer cases pesviruses in that it is not ubiquitous
Egypt, Pakistan, and Mongolia (up [2], including cases of African en- in human populations, but rather
to 20%), intermediate in parts of demic Burkitt lymphoma, which is shows marked geographical varia-
Italy and China (10%), and relatively also associated with exposure to tion in prevalence; the prevalence
lower elsewhere, except in high-risk malaria, as well as nasopharyn- of KSHV is highest in sub-Saharan
groups, such as intravenous drug geal cancer, Hodgkin lymphoma, Africa (50–95%), intermediate in
users and people who received a some non-Hodgkin lymphomas, Mediterranean countries (10%), and
transfusion before widespread HCV especially in immunocompromised generally low in other parts of the
testing of blood donors was imple- people, and a still ill-defined frac- world [1,16]. This distribution broad-
mented. Transmission is mainly tion of gastric cancer cases. EBV ly reflects that of Kaposi sarcoma,
parenteral, although it can occur is also the primary cause of infec- even before the HIV epidemic.
via sex and from mother to child, tious mononucleosis, which affects Transmission of KSHV is via sa-
although rarely; many infected peo- about half of people in whom EBV liva in both high-risk and low-risk
ple have no clear risk factors. infection occurs in adult life and has populations; in areas where the prev-
HCV is highly variable, with many been implicated as a cause of mul- alence is high, such as sub-Saharan
different genotypes. This significantly tiple sclerosis. Africa, infection occurs throughout
complicates vaccine development, EBV infection is extremely com- childhood and into adult life [4,16]. No
and currently no vaccines are avail- mon worldwide and affects about vaccines or treatments for KSHV are
able. The introduction of direct-acting 90% of the population. In low- and available, but management of HIV

greatly reduces the risk of develop- suppression, thereby facilitating the The liver flukes Opisthorchis
ing Kaposi sarcoma. Identification development of cancers caused by viverrini and Clonorchis sinensis af-
of the factors that sustain the high other infections. The cancers asso- fect up to 45 million people, primar-
transmission in sub-Saharan Africa ciated with HIV have been attributed ily in South-East Asia. In endemic
may provide opportunities for reduc- to those underlying infections men- areas, they are an important cause
ing the burden of associated cancer. tioned above. These include Kaposi of cholangiocarcinoma (bile duct
sarcoma, non-Hodgkin lymphoma, cancer), causing about 3600 cas-
Human T-cell lympho Hodgkin lymphoma, cervical cancer, es in 2018, although this number,
tropic virus type 1 anal cancer, and conjunctival cancer which is based on imperfect sta-
[1]. Perhaps of more relevance is the tistics, is probably a gross under-
HTLV-1 caused about 3600 cases
total morbidity and mortality associ- estimation [2]. Infection occurs via
of adult T-cell leukaemia/lymphoma
ated with HIV: in 2017, 36.9 million consumption of raw or undercooked
in 2018 [2]. It also causes progres-
people globally were living with HIV, contaminated fish, providing a key
sive myelopathy and other inflam-
21.7 million people were accessing target for prevention.
matory conditions [1]. Globally, an
antiretroviral therapy, and 940 000
estimated 10–20 million people
people died from AIDS-related ill- Conclusions
are infected, and 3–8 million of
nesses, despite the success of
them are in sub-Saharan Africa. Infections are an important cause
Although data from many parts of antiretroviral therapy in treating the
disease [17]. No vaccine is available, of cancer, especially in Asia and
the world are sparse, the preva- sub-Saharan Africa. In 2018, more
lence of infection appears to be but several are under development.
than one third of infection-related
highest in parts of Japan, Africa, cancer cases occurred in China,
the Caribbean, Central and South Macroparasites where 42% of all H. pylori-related
America, and northern Australasia. An estimated 200 million people cancers and 69% of all HBV-related
More than 90% of infections will re- worldwide are infected with one of cancers occurred. Adequate infec-
main asymptomatic. six species of Schistosoma, which tion control strategies, encompass-
The predominant route of trans- are prevalent to varying extents ing cheap and reliable point-of-care
mission of HTLV-1 is via breast- in tropical regions. All cause sig- diagnostic assays for particular in-
feeding, and interventions that limit nificant pathology and have been fectious agents for use in screen-
the duration of breastfeeding have
linked to several cancer types, but ing, effective treatments, and ther-
prevented up to 90% of mother-to-
only for Schistosoma haemato- apeutic and preventive vaccines,
child transmissions, in parts of the
bium in relation to bladder cancer should all play a more widespread
world where alternative feeding op-
is the evidence sufficiently robust; role in cancer control programmes.
tions are available. Surveillance of
S. haematobium infection caused Substantial international investment
the blood supply has also reduced
about 6000 cancer cases in 2018 is required to realize these aspira-
transfusion-related infections [4]. No
[2]. Infections occur after exposure tions. Further work is also justified
vaccines or treatments are available.
to contaminated freshwater and are to identify additional cancers with
treatable. However, evidence that an underlying infectious cause.
HIV large-scale pharmacological inter-
Although HIV is not directly carcino- ventions reduce the burden of can-
genic, HIV infection causes immuno- cer remains limited [1].
66
References
1. IARC (2012). Biological agents. IARC 7. Pan KF, Zhang L, Gerhard M, Ma JL, Liu 13. Asiki G, Newton R, Marions L, Seeley J,
Monogr Eval Carcinog Risks Hum. 100B:1– WD, Ulm K, et al. (2016). A large randomised Kamali A, Smedman L (2016). The impact of
441. Available from: http://publications.iarc. controlled intervention trial to prevent gas- maternal factors on mortality rates among
fr/119 PMID:23189750 tric cancer by eradication of Helicobacter children under the age of five years in a ru-
pylori in Linqu County, China: baseline re- ral Ugandan population between 2002 and
2. de Martel C, Georges D, Bray F, Ferlay J,
sults and factors affecting the eradication. 2012. Acta Paediatr. 105(2):191–9. https://
Clifford G (2020). Global burden of can-
Gut. 65(1):9–18. https://doi.org/10.1136/ doi.org/10.1111/apa.13252 PMID:26503711
cers attributable to infections in 2018: a
gutjnl-2015-309197 PMID:25986943
worldwide incidence analysis. Lancet
CHAPTER 2.2
SECTION 2
14. Kanwal F, El-Serag HB (2014). Hepatitis
Glob Health. 8(2):e180–90. https://doi. 8. Leja M, Park JY, Murillo R, Liepniece- C virus treatment: the unyielding chasm
org /10.1016/S2 214 -10 9X(19) 3 0 48 8 -7 Karele I, Isajevs S, Kikuste I, et al. between efficacy and effectiveness. Clin
PMID:31862245 (2017). Multicentric randomised study of Gastroenterol Hepatol. 12(8):1381–3.
3. Bray F, Jemal A, Grey N, Ferlay J, Forman Helicobacter pylori eradication and pepsino- https://doi.org/10.1016/j.cgh.2014.02.031
D (2012). Global cancer transitions ac- gen testing for prevention of gastric cancer PMID:24607698
cording to the Human Development Index mortality: the GISTAR study. BMJ Open.
7(8):e016999. https://doi.org/10.1136/bmj 15. Cohen JI (2018). Vaccine development
(2008-2030): a population-based study.
Lancet Oncol. 13(8):790–801. https:// open-2017-016999 PMID:28801429 for Epstein-Barr virus. Adv Exp Med Biol.
doi.org/10.1016/S1470-2045(12)70211-5 1045:477–93. https://doi.org/10.1007/978-
9. Sutton P, Boag JM (2019). Status of vaccine 981-10-7230-7_22 PMID:29896681
PMID:22658655
research and development for Helicobacter
4. Franceschi S, El-Serag HB, Forman D, pylori. Vaccine. 37(50):7295–9. https:// 16. Newton R, Labo N, Wakeham K, Miley
Newton R, Plummer M (2018). Infectious doi.org/10.1016/j.vaccine.2018.01.001 W, Asiki G, Johnston WT, et al. (2018).
agents. In: Thun MJ, Linet MS, Cerhan PMID:29627231 Kaposi’s sarcoma associated herpesvirus
JR, Haiman C, Schottenfeld D, editors. in a rural Ugandan cohort: 1992–2008.
10. Gallagher KE, LaMontagne DS, Watson- J Infect Dis. 217(2):263–9. https://doi.
Schottenfeld and Fraumeni cancer epi-
Jones D (2018). Status of HPV vaccine org/10.1093/infdis/jix569 PMID:29099933
demiology and prevention. 4th ed. New
York (NY), USA: Oxford University Press; introduction and barriers to country up-
pp. 433–60. take. Vaccine. 36(32 Pt A):4761–7. https:// 17. UNAIDS (2018). Global HIV & AIDS statis-
doi.org/10.1016/j.vaccine.2018.02.003 tics – 2018 fact sheet. Geneva, Switzerland:
5. GBD 2013 Mortality and Causes of Death PMID:29580641 Joint United Nations Programme on HIV
Collaborators (2015). Global, regional, and and AIDS. Available from: https://www.
national age-sex specific all-cause and 11. Drolet M, Bénard É, Boily MC, Ali H,
unaids.org/en/resources/fact-sheet.
cause-specific mortality for 240 causes of Baandrup L, Bauer H, et al. (2015).
death, 1990-2013: a systematic analysis Population-level impact and herd effects
for the Global Burden of Disease Study following human papillomavirus vaccina-
2013. Lancet. 385(9963):117–71. https:// tion programmes: a systematic review
doi.org/10.1016/S0140-6736(14)61682-2 and meta-analysis. Lancet Infect Dis.
PMID:25530442 15(5):565 –80. https://doi.org/10.1016/
S1473-3099(14)71073-4 PMID:25744474
6. IARC Helicobacter pylori Working Group
(2014). Helicobacter pylori eradication 12. O’Hara GA, McNaughton AL, Maponga T,
as a strategy for preventing gastric can- Jooste P, Ocama P, Chilengi R, et al. (2017).
cer. Lyon, France: International Agency Hepatitis B virus infection as a neglected
for Research on Cancer (IARC Working tropical disease. PLoS Negl Trop Dis.
Group Reports, No. 8). Available from: 11(10):e0005842. https://doi.org/10.1371/
http://publications.iarc.fr/391. journal.pntd.0005842 PMID:28981505

2.3
Alcohol consumption
A leading risk factor for cancer
Jürgen Rehm Veronika Fedirko (reviewer)

Kevin D. Shield Pietro Ferrari (reviewer)
Elisabete Weiderpass
World Cancer Research Fund of prostaglandins, by altering the

SUMMARY (WCRF)/American Institute for Can insulin-like growth factor 1 pathway,
cer Research (AICR) [2] have at- and by acting as a solvent for cel-
●● In 2016, alcohol consumption tributed the highest level of causal lular penetration of environmental
was one of the leading risk fac- evidence to the association between carcinogens (e.g. tobacco) [2]. The
tors for cancer development consumption of alcoholic beverag- biological pathways involved, and
and cancer death globally, es and the development of cancer. the relative contributions of these
causing an estimated 376 200 IARC classified alcohol consump- pathways to carcinogenesis, differ
cancer deaths, representing tion as carcinogenic to humans by cancer site.
4.2% of all cancer deaths, and (Group 1), and the WCRF/AICR On the basis of the evidence
10.3 million cancer disability- Continuous Update Project conclud- from epidemiological studies in
adjusted life years lost, repre- ed that there is convincing evidence humans, studies in experimental
senting 4.2% of all cancer dis- that consumption of alcoholic bever- animals, and mechanistic data, the
ability-adjusted life years lost. ages increases cancer risk. IARC Monographs and the WCRF/
●● The impact of alcohol consump- Alcoholic beverages contain nu- AICR Continuous Update Project
tion on cancer in 2016 varied merous carcinogenic compounds, have reported that alcohol con-
by age group; the proportion of but the majority of the risk relation- sumption causes cancers of the
cancer deaths attributable to al- ship between alcohol consumption oral cavity, oropharynx, hypophar-
cohol consumption ranged from and the development of cancer is ynx, oesophagus (squamous cell
13.9% of cancer deaths among due to ethanol [3]. Although car- carcinoma), colon, rectum, liver and
people aged 30–34 years to cinogenesis due to alcohol is far
2.7% of cancer deaths among from being fully understood, the Fig. 2.3.1. A farmer in Amani, Tanzania,
people aged 80–84 years. main pathophysiological carcino- drinks a cup of bamboo wine.
genic mechanisms of ethanol that
●● The burden of cancers caused have been postulated include its
by alcohol consumption might metabolism into the carcinogenic
be decreased through (i) individ-
metabolite acetaldehyde, its inhi-
ual-level and societal-level in-
bition of the one-carbon metabo-
terventions that reduce alcohol
lism pathway and DNA methyla-
consumption, and (ii) measures
tion (especially among people with
that target those risk factors that
a low dietary intake of folate), and
interact with alcohol consump-
its effect on increasing serum lev-
tion to increase the risk of can-
els of endogenous estrogens (see
cer or that directly affect the risk
Chapter 3.11) [2]. Ethanol has also
of alcohol-related cancers.
been hypothesized to increase the
risk of cancer through the produc-
tion of reactive oxygen species and
Alcohol consumption as polar metabolites, through the con-
a risk factor for cancer version of pro-carcinogens in the
The IARC Monographs [1] and the metabolic pathway of ethanol, by
Continuous Update Project of the lipid peroxidation, by the production
68
intrahepatic bile duct, larynx, and ship between alcohol consumption

female breast (both premenopausal and the development of cancer [9,10].
and postmenopausal as evaluated The WCRF/AICR Continuous Up
FUNDAMENTALS
by IARC [1]; postmenopausal only date Project concluded that there ■■ Alcohol (ethanol), an addictive
as evaluated by the WCRF/AICR is probable evidence that alcohol substance with carcinogenic
Continuous Update Project [2]). For consumption is associated with the properties, was consumed
all of these sites, there are dose– risk of non-cardia stomach cancer, by 42.9% of adults globally in
response relationships, with almost and limited–suggestive evidence
2016 (yearly prevalence).
linear gradients of relative risks and that alcohol consumption is asso-
no apparent lower risk threshold ciated with the risk of cancers of ■■ A relationship between alco-
CHAPTER 2.3
SECTION 2
[4,5]. The risk relationships depend the lung, pancreas, and skin (bas- hol consumption and the
mainly on the level of lifetime expo- al cell carcinoma and malignant development of cancer was
sure to alcohol [5,6]. However, for melanoma) [2]. However, alcohol first suggested by Lamy in
female breast cancer, in addition consumption is associated with 1910, when he noted that a
to the dose–response relationship other risk factors, including diet and high proportion of patients
between level of exposure and can- smoking, and therefore confound- with either cancer of the
cer incidence, patterns of alcohol ing may explain these associations. oesophagus or cancer of the
consumption, especially episodic Furthermore, there are inconsistent cardiac region of the stomach
heavy drinking, may play an impor- epidemiological findings for a rela- were alcohol misusers.
tant role [7]. tionship between alcohol consump-
The risk relationships have been tion and the development of cancers ■■ The IARC Monographs
shown to differ by population. For of the gall bladder and prostate [4]. and the Continuous Update
example, Mendelian randomization In addition, there is no evidence that Project have identified the
studies have found genetic variations alcohol consumption affects breast contribution of alcohol to
that affect the metabolism of acetal- cancer survival or recurrence [2]. carcinogenesis at numerous
dehyde in humans (see Chapter 3.3). The WCRF/AICR Continuous Up cancer sites. Alcohol con-
In particular, people with at least one date Project concluded that there sumption has been found to
copy of the aldehyde dehydrogenase is probable evidence that alcohol be causally associated with
ALDH2*2 allele (with the Glu487Lys consumption is associated with a the development of cancers
polymorphism), a variant that is prev- decreased risk of kidney cancer; of the oral cavity, oropharynx,
alent in eastern Asian populations, this may be due to improved insu- hypopharynx, oesophagus
have a higher risk of cancers of the lin sensitivity, improved blood lip (squamous cell carcinoma),
upper aerodigestive tract and of colo- id profiles, and higher adiponectin colon, rectum, liver and intra-
rectal cancer [8]. Variations in the levels among people with light and hepatic bile duct, larynx, and
alcohol dehydrogenase 1B (ADH1B) moderate alcohol consumption [2]. female breast.
and 1C (ADH1C), cytochrome P450 Resveratrol (a substance found in
2E1, and methylenetetrahydrofolate red wine) has received attention
reductase (MTHFR) genes are also for its hypothesized anticarcino- on the empirical evidence, for every
hypothesized to modify the relation- genic properties; however, based cancer case that the resveratrol in
wine might prevent, 100 000 cancer
cases are caused by ethanol [5].
Fig. 2.3.2. Young women in Japan drinking beer at a barbecue. Inconsistent inverse associations
between alcohol consumption and
the development of thyroid can-
cer, Hodgkin lymphoma, and non-
Hodgkin lymphoma also have been
found in epidemiological studies,
but there is currently not sufficient
evidence to determine the causality
of these relationships [1,2].
As a result of its effects on the
propensity to engage in unpro-
tected sex and its weakening of
the immune system, alcohol also
may indirectly increase the risk of
infection with sexually transmit-
ted viruses that potentially cause
Chapter 2.3 • Alcohol consumption 69

cancer (including Kaposi sarcoma- 19.3%, respectively, of all alcohol- ents of the dose–response curves –
associated herpesvirus and hu- attributable cancer deaths. between levels of alcohol consump-
man papillomavirus) and of HIV-1; Among all cancers types, al- tion and cancers of the upper aero
the immunosuppression caused by cohol consumption had the larg- digestive tract compared with can-
HIV-1 is thought to increase the car- est impact on cancers of the upper cers of the colorectum, liver, and
cinogenic effect of other infectious aerodigestive tract. Alcohol was re- breast [11].
agents [5]. However, more research sponsible for 26.4% of all cancers Like with cancer deaths, in 2016
is needed to further establish and of the lip and oral cavity, 30.5% of the largest contributors to the alco-
quantify any indirect effect of alco- all other pharyngeal cancers (ex- hol-attributable cancer DALYs lost
hol on an increased risk of cancers cluding nasopharyngeal cancers), were cancers of the liver, colorec-
caused by infectious diseases. 21.6% of all laryngeal cancers, and tum, and oesophagus, responsible
16.9% of all oesophageal cancers. for 22.5%, 20.6%, and 18.5%, re-
The global cancer burden These findings reflect the stronger spectively, of all alcohol-attributa-
due to alcohol associations – i.e. the higher gradi- ble cancer DALYs lost.
In 2016, alcohol consumption caused

an estimated 3.0 million deaths from
Fig. 2.3.3. Alcohol-attributable cancer deaths (top) and alcohol-attributable cancer
all causes worldwide, representing disability-adjusted life years (DALYs) lost (bottom) in 2016, by age group.
5.3% of all deaths [11]. A large pro-
portion of the health burden caused
by alcohol consumption stems from
cancer. In 2016, alcohol caused an
estimated 376 200 (95% uncertainty
interval, 324 900–439 700) cancer
deaths, representing 4.2% (95% un-
certainty interval, 3.6–4.9%) of all
cancer deaths, and an age-stand-
ardized rate (ASR) of 4.8 deaths
(95% confidence interval, 4.2–5.7)
per 100 000 people (Table 2.3.1).
Here, the term “alcohol-attributable
cancers” is used to refer to cancers
caused by alcohol. The proportion of
alcohol-attributable cancers is thus
defined by the proportion of cancers
that would not have occurred if there
had been no alcohol exposure (for
definitions of causality, see [12]; for al-
cohol-attributable fractions, see [13]).
Of the 245 million disability-
adjusted life years (DALYs) lost in
2016 due to cancer, 10.3 million
(95% uncertainty interval, 8.7 mil-
lion–12.0 million) were due to alco-
hol consumption, representing 4.2%
(95% uncertainty interval, 3.6–4.9%)
of all cancer DALYs lost (Table 2.3.2).
The majority (97.7%) of these alco-
hol-attributable cancer DALYs lost
were due to years of life lost because
of premature death resulting from
high cancer fatality rates.
In 2016, cancers of the colorec-
tum, liver, and oesophagus were
the largest contributors to the al-
cohol-attributable cancer burden,
responsible for 23.9%, 22.3%, and
70
Table 2.3.1. Alcohol-attributable cancer deaths in 2016, by sex and cancer site
Outcome and ICD-10 Number of alcohol-attributable Percentage of deaths attributable to Percentage

cancer site code deaths/1000 alcohol consumption of the total
(95% uncertainty interval) (95% uncertainty interval) alcohol-
attributable
Men Women Both sexes Men Women Both sexes cancer
deaths
Cancer C00–97 297.6 78.6 376.2 5.8 2.0 4.2 100.0

(246.9–346.1) (66–115.4) (324.9–439.7) (4.8–6.8) (1.7–3.0) (3.6–4.9)
Lip and oral C00–08 38.9 5.2 44.0 34.7 9.4 26.4 11.7
CHAPTER 2.3
SECTION 2
cavity (30.4–46.0) (3.8–7.3) (35.3–52.3) (27.1–41.0) (7.0–13.3) (21.2–31.4)
Other pharynx C09–10, 31.7 2.1 33.8 35.3 9.9 30.5 9.0
C12–14 (24.9–37.7) (1.5–3.0) (27.0–39.9) (27.8–42.1) (7.3–14.2) (24.4–36.1)
Oesophagus C15 66.9 5.8 72.7 21.7 4.8 16.9 19.3

(51.6–79.7) (3.9–8.9) (56.8–87.2) (16.7–25.8) (3.2–7.4) (13.2–20.3)
Colorectum C18–21 75.9 13.8 89.8 17.6 3.8 11.3 23.9

(61.5–89.6) (6.6–25.2) (73.1–107.4) (14.3–20.7) (1.8–6.9) (9.2–13.5)
Liver C22 65.1 18.9 84.0 11.1 7.8 10.1 22.3

(31.5–102.5) (9.5–34.4) (49.8–125.3) (5.4–17.5) (3.9–14.1) (6.0–15.1)
Larynx C32 19.1 0.8 19.9 23.7 6.7 21.6 8.5

(14.8–23.1) (0.6–1.0) (15.6–24.0) (18.4–28.6) (5.2–9.2) (16.9–26.1)
Breast C50 – 32.0 32.0 – 5.5 5.5 5.3

(26.8–51.1) (26.8–51.1) (4.6–8.8) (4.6–8.7)
All causes A00–Z99 2307.3 681.0 2988.3 7.7 2.6 5.3 –

(1929.7– (536.4– (2596.8– (6.4–9.0) (2.0–3.8) (4.6–6.2)
2720.1) 990.7) 3523.9)
ICD-10, International Statistical Classification of Diseases and Related Health Problems, 10th revision.
Table 2.3.2. Alcohol-attributable cancer disability-adjusted life-years lost in 2016, by sex and cancer site
Outcome and ICD-10 Number of alcohol-attributable DALYs Percentage of DALYs lost attributable to Percentage
cancer site code lost/100 000 alcohol consumption of the total
(95% uncertainty interval) (95% uncertainty interval) alcohol-
attributable
Men Women Both sexes Men Women Both sexes cancer
DALYs lost
Cancer C00–97 81.6 21.1 102.6 5.9 2.0 4.2 100.0

(67.0–95.9) (18.0–31.4) (87.3–120.0) (4.9–7.0) (1.7–2.9) (3.6–4.9)
Lip and oral C00–08 12.2 1.4 13.6 33.2 8.6 25.7 13.3
cavity (9.2–14.7) (1.0–2.0) (10.6–16.5) (25.0–40.0) (6.3–12.2) (19.9–31.0)
Other pharynx C09–10, 9.7 0.6 10.3 35.5 9.3 30.6 10.1
C12–14 (7.6–11.6) (0.4–0.9) (8.2–12.3) (27.6–42.5) (6.8–13.4) (24.1–36.5)
Oesophagus C15 17.7 1.4 19.0 22.1 4.7 17.5 18.5

(13.8–20.9) (0.9–2.1) (15.0–22.6) (17.2–26.1) (3.2–7.2) (13.8–20.7)
Colorectum C18–21 18.0 3.2 21.2 16.7 3.8 11.1 20.6

(14.4–21.4) (1.6–5.8) (17.2–25.3) (13.4–19.9) (1.9–6.9) (9.0–13.2)
Liver C22 18.6 4.5 23.1 10.7 7.2 9.7 22.5

(8.9–29.7) (2.3–8.2) (13.2–35.0) (5.1–17.0) (3.7–13.1) (5.6–14.8)
Larynx C32 5.4 0.2 5.6 23.8 6.7 21.8 9.6

(4.2–6.5) (0.2–0.3) (4.4–6.7) (18.5–28.8) (5.2–9.0) (17.0–26.3)
Breast C50 – 9.9 9.9 – 5.2 5.2 5.4

(8.2–16.4) (8.2–16.4) (4.4–8.7) (4.3–8.6)
All causes A00–Z99 1065.4 261.0 1326.4 7.6 2.2 5.1 –

(903.2– (234.4– (1164.1– (6.5–8.9) (1.9–2.7) (4.5–5.9)
1240.8) 331.5) 1539.8)
DALYs, disability-adjusted life years; ICD-10, International Statistical Classification of Diseases and Related Health Problems, 10th revision.

Similarly, alcohol consumption 80–84 years. At younger ages, can- complex; however, the proportion of
had the largest contributory impact cers of the liver, breast, and colorec- alcohol-attributable cancers among
on DALYs lost due to cancers of the tum were the leading contributors to this age group is hypothesized to be
upper aerodigestive tract. Alcohol the alcohol-attributable cancer bur- relatively small [15].
was responsible for 25.7% of all den, responsible for 32.2%, 19.4%, In 2016, there were large vari-
lip and oral cavity cancer DALYs and 18.4%, respectively, of all al- ations between countries and geo-
lost, 30.6% of all other pharyngeal cohol-attributable cancer deaths graphical regions in the ASRs of
cancer DALYs lost (excluding naso- among people aged 30–34 years. alcohol-attributable cancer deaths
pharyngeal cancers), 21.8% of all At older ages, cancers of the colo- (Fig. 2.3.4) and cancer DALYs lost
laryngeal cancer DALYs lost, and rectum, liver, and oesophagus were (Fig. 2.3.5). Based on the regions
17.5% of all oesophageal cancer the leading contributors to the al- as defined by the Institute for Health
DALYs lost. cohol-attributable cancer burden, Metrics and Evaluation’s Global
Based on different consumption responsible for 39.1%, 20.1%, and Burden of Disease study, the bur-
levels by age [14], the impact of alco- 14.9%, respectively, of all alcohol- den of alcohol-attributable cancers
hol consumption on cancer in 2016 attributable cancer deaths among was lowest in North Africa and the
varied by age group (Fig. 2.3.3); the people aged 80 years and older. Middle East (ASRs of 0.8 cancer
proportion of cancer deaths attribu The impact of alcohol on cancer deaths and 24.2 cancer DALYs lost
table to alcohol consumption ranged deaths and DALYs lost among peo- per 100 000 people) and highest in
from 13.9% of cancer deaths among ple aged 29 years and younger is eastern Europe (ASRs of 12.0 can-
people aged 30–34 years to 2.7% of unknown, because data are lacking cer deaths and 360.4 cancer DALYs
cancer deaths among people aged and the etiology of these cancers is lost per 100 000 people).
Fig. 2.3.4. Global burden of cancer deaths caused by alcohol consumption in 2016: (top) age-standardized cancer deaths
attributable to alcohol consumption per 100 000 people; (bottom) percentage of cancer deaths attributable to alcohol.
72
Fig. 2.3.5. Global burden of cancer disability-adjusted life years (DALYs) lost caused by alcohol consumption in 2016: (top) age-
standardized cancer DALYs lost attributable to alcohol consumption per 100 000 people; (bottom) percentage of cancer DALYs lost
attributable to alcohol.
CHAPTER 2.3
SECTION 2
Similarly, the proportion of alco- with high or very high levels of the HDI and countries with high HDI, and
hol-attributable cancer deaths and Human Development Index (HDI). cancers of the lip and oral cavity in
cancer DALYs lost also varied be- Both the consumption of alcohol countries with medium HDI.
tween countries and regions. The and the burden of cancer increase The alcohol-attributable cancer
proportions were lowest in North as countries develop [11,16]. In 2016, deaths and cancer DALYs lost dis-
Africa and the Middle East (0.8% of the ASRs of the alcohol-attributable cussed above include only cancer
cancer deaths and 0.8% of cancer cancer burden were highest for coun- sites for which sufficient causal evi-
DALYs lost) and highest in eastern tries with very high HDI (7.3 cancer dence exists, as determined by the
Europe (8.1% of cancer deaths and deaths and 203.8 cancer DALYs lost IARC Monographs, and do not in-
8.6% of cancer DALYs lost). per 100 000 people) and lowest for clude cancer sites for which there
The burden of cancer by site also countries with medium HDI (2.5 can- was insufficient evidence of carcino-
varied across geographical regions cer deaths and 78.8 cancer DALYs genicity in humans [1]. However, an
(Fig. 2.3.6). In particular, alcohol-at- lost per 100 000 people) (Fig. 2.3.7). analysis conducted for France in 2015
tributable cancers of the colorectum The site-specific alcohol-attributable found that the proportion of cancer
(see Chapter 5.5) were prominent cancer burden also varied by HDI. incidence due to alcohol increased
in southern Latin America, high-in- The largest contributors to the ASRs from 7.9% when limited to cancers for
come North America, high-income of alcohol-attributable cancer deaths which sufficient causal evidence ex-
Asia Pacific, Australasia, and cen- were colorectal cancer in countries ists to 8.4% when including cancers
tral, eastern, and western Europe; with very high HDI, liver cancer (see for which at least limited evidence of a
all of these regions have countries Chapter 5.6) in countries with low causal association exists [17].

Country- and region-specific ana leading risk factor for cancer devel- study [22] and in an analysis of 13
lyses of the relative contributions opment and cancer death. In some risk factors for France in 2015 [15].
of risk factors to the cancer burden analyses and countries, alcohol is
in the USA [18], France [15], the the second most important risk fac- Trends in the cancer
United Kingdom [19], Australia [20], tor for cancer development and can-
and the Nordic countries (Denmark, cer death after tobacco, for example
burden due to alcohol
Finland, Iceland, Norway, Sweden, in an analysis of nine behavioural
from 2010 to 2016
the Faroe Islands, and Greenland) and environmental risk factors for Trends in the alcohol-attributable
[21] have shown that alcohol is a the Global Burden of Disease 2000 cancer burden depend on changes
in alcohol consumption as well as
in cancer incidence, treatment, and
Fig. 2.3.6. Age-standardized alcohol-attributable cancer deaths per 100 000 people mortality. As a result of population
(top) and alcohol-attributable cancer disability-adjusted life years (DALYs) lost per growth and ageing and the eco-
100 000 people (bottom) in 2016, by geographical region. nomic development of countries,
the total number of cancer deaths
worldwide increased from 8.1 mil-
lion in 2010 to 9.0 million in 2016
[11]. However, the ASR of cancer
mortality decreased by 6.0% (from
122.4 per 100 000 in 2010 to 115.0
per 100 000 in 2016), less than the
9.0% decrease in the ASR of overall
mortality (from 791.3 per 100 000 in
2010 to 720.1 per 100 000 in 2016).
The ASRs of alcohol-attributable
mortality decreased less than over-
all cancer mortality rates in gen-
eral (by 4.8%, from 5.1 deaths per
100 000 in 2010 to 4.8 deaths per
100 000 in 2016), resulting in an in-
crease of 1.5% in the proportion of
cancer deaths attributable to alcohol
consumption (from 4.1% in 2010 to
4.2% in 2016). Thus, the relative im-
pact of alcohol on cancer mortality
increased slightly from 2010 to 2016.
Trends in the ASRs of alcohol-
attributable cancer mortality and in
the proportion of cancers attributa-
ble to alcohol consumption showed
heterogeneous patterns by cancer
site. In particular, the ASR of mortal-
ity due to cancers of the lip and oral
cavity was the only ASR to increase
(from 2.1 deaths per 100 000 in 2010
to 2.2 deaths per 100 000 in 2016),
and the ASR of mortality due to oe-
sophageal cancer (see Chapter 5.3)
decreased the most (from 6.2 deaths
per 100 000 in 2010 to 5.5 deaths per
100 000 in 2016).
In the long term, increases in the
economic wealth of countries are
likely to lead to further increases in
life expectancies, resulting in higher
incidence of and mortality from can-
cer and a concomitant higher rela-
tive importance of cancer as a cause
74
Fig. 2.3.7. Age-standardized alcohol-attributable cancer deaths per 100 000 people large, and this burden is expected
(top) and alcohol-attributable disability-adjusted life years (DALYs) lost per 100 000 to increase in the future. Therefore,
people (bottom) in 2016, by level of Human Development Index (HDI). programmes designed to reduce
alcohol consumption in the general
population are an effective and cost-
effective means of targeting and im-
proving cancer control (see Chapter
6.1). The observed differences be-
tween countries and regions in alco-
hol-attributable fractions of cancer
CHAPTER 2.3
SECTION 2
deaths and cancer DALYs lost pro-
vide an evidence base for how to re-
duce this burden through individual-
level and societal-level programmes
that reduce alcohol consumption,
such as the WHO intervention strat-
egies known as alcohol policy “best
buys”, which include increasing
excise taxation of alcoholic bever-
ages, restricting access to retailed
alcoholic beverages, and limiting
advertising and promotion of alco-
holic products [24].
Furthermore, the burden of alco-
hol-attributable cancers could be re-
duced through measures that target
those risk factors that interact with
alcohol consumption to increase the
risk of cancer or that directly affect the
risk of alcohol-related cancers, such
as tobacco smoking (see “Tobacco
cessation: the WHO perspective”).
In addition, early recognition of the
signs and symptoms of cancer, as
well as prompt diagnosis of precan-
cerous lesions and tumours, are in
many cases vital to patient survival,
and therefore screening for colorec-
tal cancer and breast cancer may
also reduce the burden of alcohol-
attributable cancers [25].
Finally, despite the evidence
of death (http://www.healthdata.org/ data.org/results/country-profiles). of the causal relationship between
results/country-profiles), as well as Accordingly, whereas in high-in- alcohol consumption and the devel-
to higher per capita alcohol con- come countries alcohol consump- opment of cancer, the majority of
sumption [11,14]. Furthermore, be- tion, cancer mortality rates, and the general population is unaware
cause the median latency between alcohol-attributable cancer mortality of this causal link [26]. Warning
mean alcohol consumption and the rates have declined, and may con- labels can be used to raise aware-
diagnosis of cancer is 10 years [23], tinue to decline, the overall global ness of the link between alcohol
it is expected that alcohol-attributa- burden of alcohol-attributable can- and cancer; however, the effective-
ble cancer mortality will continue to cers is not expected to decrease, ness of these labels to reduce al-
increase in the countries that have and may increase in the long term. cohol consumption is currently un-
had the most pronounced increases known [11]. In addition, explaining
in alcohol consumption over the past the causal link between alcohol and
few years. Examples of such coun-
The cancer burden due cancer could be part of brief inter-
tries are China and India, countries to alcohol is preventable ventions by medical professionals
in which life expectancies have also The current burden of cancers in primary care, to reduce alcohol
increased ([11]; http://www.health caused by alcohol consumption is consumption [27].

References
1. IARC (2012). Personal habits and in- 11. WHO (2018). Global status report on alco- 21. Andersson TM, Engholm G, Pukkala E,
door combustions. IARC Monogr Eval hol and health 2018. Geneva, Switzerland: Stenbeck M, Tryggvadottir L, Storm H,
Carcinog Risks Hum. 100E:1–575. World Health Organization. Available from: et al. (2018). Avoidable cancers in the
PMID:23193840. Available from: http:// https://www.who.int /substance_abuse/ Nordic countries – the impact of alcohol
publications.iarc.fr/122. publications/global_alcohol_report/en/. consumption. Eur J Cancer. 103:299–307.
https://doi.org/10.1016/j.ejca.2018.03.027
2. WCRF/AICR (2018). Diet, nutrition, physi- 12. Rothman KJ, Greenland S, Lash TL (2008).
PMID:29739641
cal activity and cancer: a global perspec- Modern epidemiology. 3rd ed. Philadelphia
tive. Continuous Update Project Expert (PA), USA: Lippincott Williams & Wilkins. 22. Danaei G, Vander Hoorn S, Lopez AD,
Report 2018. World Cancer Research Murray CJ, Ezzati M; Comparative Risk
13. Shield KD, Parkin DM, Whiteman DC,
Fund/American Institute for Cancer Assessment collaborating group (Cancers)
Rehm J, Viallon V, Micallef CM, et al.
Research. Available from: https://www. (2005). Causes of cancer in the world:
(2016). Population attributable and pre-
wcrf.org/dietandcancer. ventable fractions: cancer risk factor sur- comparative risk assessment of nine be-
veillance, and cancer policy projection. havioural and environmental risk factors.
3. Pflaum T, Hausler T, Baumung C, Lancet. 366(9499):1784–93. https://doi.
Curr Epidemiol Rep. 3(3):201–11. https://
Ackermann S, Kuballa T, Rehm J, et al. org/10.1016/S0140 - 673 6(05) 67725 -2
d o i.o r g /10.10 07/s 4 0 471- 016 - 0 0 8 5 - 5
(2016). Carcinogenic compounds in alco- PMID:16298215
PMID:27547696
holic beverages: an update. Arch Toxicol.
90(10):2349–67. https://doi.org/10.1007/ 14. Manthey J, Shield KD, Rylett M, Hasan 23. Holmes J, Meier PS, Booth A, Guo Y,
s00204-016-1770-3 PMID:27353523 OSM, Probst C, Rehm J (2019). Global Brennan A (2012). The temporal relation-
alcohol exposure between 1990 and 2017 ship between per capita alcohol consump-
4. Bagnardi V, Rota M, Botteri E, Tramacere tion and harm: a systematic review of
and forecasts until 2030: a modelling study.
I, Islami F, Fedirko V, et al. (2015). Alcohol Lancet. 393(10190):2493–2502. https:// time lag specifications in aggregate time
consumption and site-specific cancer doi.org/10.1016/S0140-6736(18)32744-2 series analyses. Drug Alcohol Depend.
risk: a comprehensive dose-response PMID:31076174 123(1–3):7–14. https://doi.org/10.1016/j.
meta-analysis. Br J Cancer. 112(3):580– drugalcdep.2011.12.005 PMID:22197480
93. https://doi.org/10.1038/bjc.2014.579 15. Soerjomataram I, Shield K, Marant-
PMID:25422909 Micallef C, Vignat J, Hill C, Rogel A, et 24. Chisholm D, Moro D, Bertram M, Pretorius
al. (2018). Cancers related to lifestyle C, Gmel G, Shield K, et al. (2018). Are the
5. Rehm J, Gmel GE Sr, Gmel G, Hasan and environmental factors in France in “best buys” for alcohol control still valid?
OSM, Imtiaz S, Popova S, et al. (2017). 2015. Eur J Cancer. 105:103–13. https:// An update on the comparative cost-ef-
The relationship between different dimen- d o i . o r g /10 .1016 / j . e j c a . 2 018 . 0 9 . 0 0 9 fectiveness of alcohol control strategies
sions of alcohol use and the burden of dis- PMID:30445359 at the global level. J Stud Alcohol Drugs.
ease – an update. Addiction. 112(6):968– 79(4):514–22. https://doi.org/10.15288/
1001. https://doi.org/10.1111/add.13757 16. Bray F, Jemal A, Grey N, Ferlay J, Forman jsad.2018.79.514 PMID:30079865
PMID:28220587 D (2012). Global cancer transitions ac-
cording to the Human Development Index 25. Smith RA, Andrews KS, Brooks D,
6. Bergmann MM, Rehm J, Klipstein- (2008-2030): a population-based study. Fedewa SA, Manassaram-Baptiste D,
Grobusch K, Boeing H, Schütze M, Drogan Lancet Oncol. 13(8):790–801. https:// Saslow D, et al. (2018). Cancer screen-
D, et al. (2013). The association of pattern doi.org/10.1016/S1470-2045(12)70211-5 ing in the United States, 2018: a review
of lifetime alcohol use and cause of death PMID:22658655 of current American Cancer Society
in the European Prospective Investigation guidelines and current issues in cancer
17. Shield KD, Marant Micallef C, Hill C,
into Cancer and Nutrition (EPIC) study. Int screening. CA Cancer J Clin. 68(4):297–
Touvier M, Arwidson P, Bonaldi C, et
J Epidemiol. 42(6):1772–90. https://doi. 316. https://doi.org/10.3322/caac.21446
al. (2018). New cancer cases in France
org/10.1093/ije/dyt154 PMID:24415611 PMID:29846940
in 2015 attributable to different lev-
7. Liu Y, Nguyen N, Colditz GA (2015). Links els of alcohol consumption. Addiction. 26. Scheideler JK, Klein WMP (2018).
between alcohol consumption and breast 113(2):247–56. https://doi.org/10.1111/add. Awareness of the link between alco-
cancer: a look at the evidence. Womens 14009 PMID:28833736 hol consumption and cancer across
Health (Lond). 11(1):65–77. https://doi. 18. Islami F, Goding Sauer A, Miller KD, the world: a review. Cancer Epidemiol
org/10.2217/WHE.14.62 PMID:25581056 Siegel RL, Fedewa SA, Jacobs EJ, et al. Biomarkers Prev. 27(4):429–37. https://
(2018). Proportion and number of cancer doi.org/10.1158/1055-9965.EPI-17-0645
8. Pierce BL, Kraft P, Zhang C (2018). PMID:29615419
cases and deaths attributable to poten-
Mendelian randomization studies of can-
tially modifiable risk factors in the United 27. Babor TF, Higgins-Biddle JC (2000). Alcohol
cer risk: a literature review. Curr Epidemiol
States. CA Cancer J Clin. 68(1):31–54.
Rep. 5(2):184–96. https://doi.org/10.1007/ screening and brief intervention: dissemi-
h t t p s: //d o i . o r g /10 . 3 3 2 2 /c a a c . 214 4 0
s40471-018-0144-1 PMID:30034993 nation strategies for medical practice and
PMID:29160902
public health. Addiction. 95(5):677–86.
9. Druesne-Pecollo N, Tehard B, Mallet Y, 19. Parkin DM, Boyd L, Walker LC (2011). 16. https://doi.org/10.1046/j.1360-0443.2000.
Gerber M, Norat T, Hercberg S, et al. (2009). The fraction of cancer attributable to life- 9556773.x PMID:10885042
Alcohol and genetic polymorphisms: ef- style and environmental factors in the UK
fect on risk of alcohol-related cancer. in 2010. Br J Cancer. 105(Suppl 2):S77–
Lancet Oncol. 10(2):173–80. https://doi. 81. https://doi.org/10.1038/bjc.2011.489
org /10.1016 /S1470 -20 4 5(0 9)70 019 -1 PMID:22158327
PMID:19185835
20. Whiteman DC, Webb PM, Green AC,
10. LoConte NK, Brewster AM, Kaur JS, Neale RE, Fritschi L, Bain CJ, et al. (2015).
Merrill JK, Alberg AJ (2018). Alcohol and Cancers in Australia in 2010 attributa-
cancer: a Statement of the American ble to modifiable factors: summary and
Society of Clinical Oncology. J Clin Oncol. conclusions. Aust N Z J Public Health.
36(1):83–93. https://doi.org/10.1200/JCO. 39(5):477–84. https://doi.org/10.1111/1753-
2017.76.1155 PMID:29112463 6405.12471 PMID:26437735
76
2.4
Sunlight and ultraviolet radiation

Affecting skin cancer incidence
in many countries
Chikako Nishigori Steffen Emmert (reviewer)
CHAPTER 2.4
SECTION 2
Nagarajan Rajendra Prasad (reviewer)
as the promotion phase in pho- Solar radiation encompasses a broad

SUMMARY tocarcinogenesis. Dipyrimidine range of wavelengths of photon ener-
photoproducts trigger ultraviolet- gy in the electromagnetic spectrum,
●● Ultraviolet radiation directly and induced immunosuppression, including ionizing radiation, ultravio-
indirectly induces DNA lesions, which leads to the failure of im- let (UV) radiation, visible light, and
which cause mutations and munosurveillance and enables infrared radiation (Fig. 2.4.1). UV
trigger inflammation and immu- the cancer cells to grow and radiation is conventionally classified
nosuppression, which mediate progress. into three types: UVA (wavelengths
tumour growth. Both ultraviolet of 315–400 nm), UVB (280–315 nm),
radiation itself and ultraviolet- ●● People who are taking immuno- and UVC (100–280 nm). Solar UV
induced inflammation lead to suppressants or some other kinds radiation has beneficial biological
the generation of reactive oxygen of medication, including voricon- effects, including enabling vitamin
species. These reactive oxygen azole and hydrochlorothiazide, D synthesis, but its adverse effects
species also cause DNA lesions should be careful to protect them- include the induction of skin cancers
and increase the frequency of selves from exposure to sunlight. (see Chapter 5.8).
mutations. Furthermore, lipid
peroxidation caused by ultravio-
let radiation and reactive oxy- Fig. 2.4.1. Schematic diagram of bands of solar radiation, classified by wavelength.
UVR, ultraviolet radiation.
gen species also contributes to
immunosuppression.
●● The incidence of skin cancers is
increasing worldwide, and espe-
cially in older people.
●● The most effective way to reduce
skin cancer incidence is to avoid
unnecessary sun exposure, use
protective measures when in the
sun, and avoid tanning devices.
●● Photocarcinogenesis is a com-
plicated, multistep pathway,
which is initiated by the forma-
tion of dipyrimidine photoprod-
ucts, which lead to the forma-
tion of mutations (the initiation
phase). Sunburn and inflamma-
tion caused by the presence of
persistent DNA lesions, including
dipyrimidine photoproducts and
oxidative DNA lesions, function
Chapter 2.4 • Sunlight and ultraviolet radiation 77

Fig. 2.4.2. Schematic summary of photocarcinogenesis as detailed in the text. COX-2,

cyclooxygenase 2; IL-1β, interleukin-1β; iNOS, inducible nitric oxide synthase; MAPK,
mitogen-activated protein kinase; 8-OHdG, 8-hydroxydeoxyguanosine; ROS, reactive FUNDAMENTALS
oxygen species; TNF-α, tumour necrosis factor α.
■■ Solar radiation encompasses
a broad range of wavelengths
of photon energy, including
ionizing radiation, ultraviolet
radiation, visible light, and
infrared radiation. Ultraviolet
radiation is conventionally
classified into three types: UVA
(wavelengths of 315–400 nm),
UVB (280–315 nm), and UVC
(100–280 nm).
■■ Solar ultraviolet radiation has
beneficial biological effects,
including enabling vitamin
D synthesis, but its adverse
effects include sunburn and the
development of solar lentigines,
immunosuppression, and
skin cancers.
■■ Irradiating mice with ultraviolet
radiation induces skin cancer,
and the action spectrum in the
mouse model of ultraviolet-
induced photocarcinogenesis
A simple perspective is that UVB- dition, pyrimidine dimers play a role falls into the UVB range.
induced DNA photolesions cause mu- in UV-induced immunosuppression,
■■ Ultraviolet radiation generates
tations, which may be equated with which also plays an important role
DNA photolesions, and such
initiation, a term originally used to de- in photocarcinogenesis [3], partly by
dipyrimidine photoproducts
scribe the first phase of chemically upregulation of interleukin 10 (IL-10),
include cyclobutane pyrimidine
induced carcinogenesis in rodents; an immunosuppressive cytokine [4].
dimers and (6–4) photo-
on the same basis, UVB-induced in- In skin cells, UV radiation also pro-
products, which are muta-
flammation, and specifically sunburn, duces oxidative stress and oxidative
genic and contribute to cancer
equates to the promotion phase of DNA damage, which cause alteration
development.
carcinogenesis. However, recent of the genes involved in apoptosis
findings have revealed that the pho- and modification of cell signalling by
tocarcinogenesis pathway is more redox regulation, resulting in inflam-
complex; each of these processes mation (Fig. 2.4.2). in the course of hardening resin and
is mediated by various cellular, bio- In this chapter, knowledge about
coating. Modern factories have pro-
chemical, and molecular changes, photocarcinogenesis is summarized.
duction processes designed so that
which are closely interrelated (see employees are well protected, and
Chapter 3.11). Sources of ultraviolet therefore such lamps are rarely as-
The accumulation of DNA pho- radiation sociated with harmful impacts on hu-
tolesions caused by UV radiation in The main source of human exposure man health. Germicidal UV lamps are
several cancer-related genes, which to UV radiation is solar radiation. In commonly used to disinfect rooms,
may still be regarded as the initiation addition, many people have been the floors of laboratories, and some-
phase, plays a crucial role in car- exposed through the use of tanning times public spaces, including hos-
cinogenesis. These DNA photole- devices (sunlamps and sunbeds), pitals, gymnasiums, and swimming
sions contribute to the development which are artificial sources of UV pools.
of skin cancers through specific mu- radiation; this warrants concern for Special UV lamps are used
tations that lead to the upregulation human health (as discussed later). therapeutically to treat certain skin
or downregulation of signal trans- In some occupational circum- diseases, including vitiligo vulgaris,
duction pathways of cell growth and stances, UV lamps are used for the psoriasis, and atopic dermatitis.
cell-cycle dysregulation [1,2]. In ad- purpose of polymerization, typically Currently, for therapeutic purposes,
78
Fig. 2.4.3. Incidence of skin cancers in Japan in 1975–2010 in different age groups. Epidemiology of skin
cancers
The incidence of both melanoma
and non-melanoma skin cancers
is increasing worldwide, not only
in White populations [5] but also in
Asian populations. In addition, there
is marked variation in incidence by
geographical location between and
within countries. Epidemiological
CHAPTER 2.4
SECTION 2
studies have demonstrated a nega-
tive correlation between the latitude
of residence and the incidence and
mortality rates of melanoma and non-
melanoma skin cancers in homoge-
neous populations.
According to statistics from
the Ministry of Health, Labour and
Welfare of Japan, the incidence of
skin cancers in Japan has increased
dramatically over the past decades,
especially in people older than
65 years (Fig. 2.4.3). A longer life ex-
pectancy contributes to this increase
in risk, because non-melanoma skin
cancer is more common in older
narrow-band UVB sources that emit in the air and is reflected by buildings people. Furthermore, the incidence
specifically radiation of wavelength and land surfaces. The reflection of of non-melanoma skin cancer in men
311 nm are widely used, to reduce solar UV radiation varies depending is strikingly higher than that in women
exposure to wavelengths shorter than on the condition of the land surface. in Japan as well as in Australasia,
305 nm, which are most harmful in Snow, sand, and other surfaces reflect Europe, and North America, probably
relation to developing skin cancer. UV radiation to varying degrees: new because the effects of lifestyle fac-
During the welding process, UV snow reflects 80%, a sandy beach tors are similar in different countries.
radiation is emitted, and therefore reflects 10–25%, concrete or asphalt The IARC Monographs classi-
welders should use personal protec- reflects 10%, the surface of water re- fied UV-emitting tanning devices
tive equipment in the course of their flects 10–20%, and a lawn or grassy (sunlamps and sunbeds) as carcino-
work (see Chapter 2.10). plain reflects 10%. The intensity of so- genic to humans (Group 1). Although
The ozone layer in the strato- lar UV radiation depends on the height commercial use of tanning devices
sphere absorbs solar UV radiation of the sun in the sky; it is strongest is prohibited in some states of the
of wavelengths shorter than 300 nm. at solar noon and during the summer USA, in almost all states and territo-
Therefore, only UVA radiation and months. ries of Australia, and in some other
UVB with wavelengths longer than Some weather services provide countries, many people continue to
300 nm reach the Earth’s surface. daily forecasts of the intensity of solar use them. The association of sunbed
The radiation reaching the Earth’s UV radiation. Such information may exposure with predicted increased
surface is largely composed of UVA be helpful as a rough indication, but risk of induction of squamous cell
(95%), with a small UVB component caution should be exercised, be- carcinoma has been confirmed [6],
(5%). cause the intensity of solar UV radia- and people should be aware of the
The level of solar UV exposure at tion differs greatly between locations risk associated with use of tanning
the Earth’s surface varies with latitude, where relevant measurements are devices.
altitude, time of day and time of year, conducted. Although several UV do-
cloud cover, other atmospheric fac- simetry instruments are commercially
tors (specifically including pollution), available, not all of the equipment is Ultraviolet-induced DNA
and reflection from nearby surfaces. accurate and reliable. The best way photolesions
UV radiation is stronger at high alti- to protect oneself from the sun is to The photon energy of UV radiation is
tudes than at ground level, because adopt multiple personal measures, not capable of causing ionization but
the thinner atmosphere blocks less such as wearing protective clothing, results only in excitation at the atomic
UV radiation. About 80% of solar UVB wearing a hat, applying sunscreen, level. Therefore, all the biological
penetrates thin cloud. UVB scatters and using shade. consequences of UV radiation are

attributable to excited chemical reac- mental models is maximal within nucleus, and that the response may
tions in the molecules of the skin. DNA the UVB range, with the peak at be elicited by oxidative stress caused
directly absorbs more energy from 293 nm [11]. Formation of dipyrimi- by UV radiation [14]. There is plenty of
UVB photons than from UVA photons. dine photoproducts can lead to UV evidence that various antioxidants at-
UVB specifically acts on DNA by di- signature mutations in DNA. UV tenuate erythema or oedema induced
rectly exciting the nucleobases, result- signature mutations are associ- by UVB radiation [15]. Low levels of
ing in the instant formation of dimeric ated with transition-type mutations oxidants can modify cell signalling
photoproducts at dipyrimidine sites. In such as C:G → T:A at dipyrimidine via redox regulation, and these signal
contrast, UVA and visible light primar- sequences, where a transition is de- modifications have functional conse-
ily exert a biological impact directly by fined as a change from one pyrimi- quences [16].
participating in the formation of reac- dine (cytosine or thymine) or purine UV radiation triggers sequential
tive oxygen species in the presence (guanine or adenine) to the other. molecular responses, thereby acti-
of photosensitizers, and indirectly The molecular changes observed vating cell-surface growth factors and
produce oxidative DNA lesions. UVB in skin cancers have been analysed pro-inflammatory cytokine receptors.
produces dipyrimidine photoproducts in many studies. In White people, Mice deficient in tumour necrosis
by direct excitation, and also gener- TP53 mutations are present at much factor α (TNF-α), a pro-inflammatory
ates oxidative DNA lesions. higher frequencies (~50–90%) in cytokine, are resistant to skin carcino-
Studies have suggested that dipy- non-melanoma skin cancers than genesis, although both deficient and
rimidine photoproducts are the most they are in internal malignancies [1]. wild-type mice exhibited the same
important UV-induced DNA photole- These mutations are predominantly c-Ha-ras mutations after treatment
sions, because they are involved in C:G → T:A at dipyrimidine sites, the with 7,12-dimethylbenz[a]anthracene
cytotoxicity and mutagenesis [7]. UV signature mutations. [17]. In animal photocarcinogenesis
Reactive oxygen species cause vari- In Asian people, the UV signa- studies, some antioxidant nutrition that
ous biological effects via the redox ture mutations are significantly more suppresses UV-induced inflammation
signalling pathway and produce oxi- frequent in skin cancers at sun-ex- has been shown to suppress cancer
dative DNA lesions, which also play posed body sites than in those at development.
a role in carcinogenesis [8]. Among non-sun-exposed sites [12], sug- Earlier, it was reported that in this
oxidative DNA lesions, 8-hydroxyde- gesting that UV radiation is also mouse photocarcinogenesis model,
oxyguanosine (8-OHdG) has been closely involved in the development the accumulation of 8-oxoG, an oxida-
established as a sensitive marker of of non-melanoma skin cancer in tive DNA photolesion, increases the
oxidative DNA damage. The guanine Asian people. Several other reports development of skin cancers; this re-
base in genomic DNA is highly sus- have demonstrated that the types of sult is attributable to the upregulation
ceptible to oxidative stress, because mutations that are not considered of genes related to the inflammatory
guanine has the lowest oxidation po- to be caused by dipyrimidine pho- response pathway, such as Cxcl1 and
tential of all the bases. toproducts are frequently observed Il-6, but not to the mutations caused
Recent work has shown that in human skin cancers at sun-ex- by oxidative DNA lesions [8]. Rodier
cyclobutane pyrimidine dimers are posed body sites [13], thereby sug- et al. reported that large doses of UV
produced at higher yields than 8-hy- gesting that oxidative DNA lesions radiation, which cause irreparable
droxyguanine (8-oxoG) after expo- may also play a role to some extent damage to cells, induce DNA double-
sure to UVA in human skin cells and in photocarcinogenesis. strand breaks and increase secretion
human skin in vivo [9]. The diuretic of IL-6 [18].
medication hydrochlorothiazide sig- Inflammation caused
nificantly increases the production by sunburn promotes Melanoma and ultraviolet-
of thymine dimers by UVA, indepen- carcinogenesis, and
dent of the presence of oxygen [10].
induced inflammation
This indicates that excited hydro-
particular DNA lesions Recently, much attention has been
chlorothiazide molecules function
are implicated paid to melanoma formation and
as UVA-absorbing chromophores, The sunburn process is dependent UV-induced inflammation. It is gen-
which transfer energy to adjacent on several factors, including UV dose, erally accepted that chronic inflam-
pyrimidines, thereby resulting in the UV wavelength, and photoskin type. mation increases the risk of cancer;
formation of thymine dimers. After cellular molecules absorb UV this is consistent with the finding that
radiation, photochemical reactions excessive intense, intermittent sun
occur, and these processes are re- exposure is one of the most important
Ultraviolet-induced DNA sponsible for biological changes that risk factors for melanoma.
lesions and mutations in culminate in sunburn. The findings In hepatocyte growth factor/scat-
skin cancers of Devary et al. suggested that the ter factor transgenic mice, a single
The action spectrum for UV-induced UV response is initiated at or near dose of burning UV radiation to neo-
carcinogenesis in animal experi- the cell membrane rather than in the nates, but not to adults, is necessary
80
Fig. 2.4.4. Crowds at Bondi Beach, Sydney, Australia. A relatively high incidence of tion spectrum for photocarcinogenesis
melanoma and other skin cancers is attributable to exposure of fair-skinned populations in a mouse model revealed that UVA
to intense ultraviolet radiation in countries such as Australia. The incidence of both is partly responsible for photocarcino-
melanoma and non-melanoma skin cancers is also increasing in Asian populations,
specifically including those in Japan.
genesis [11].
UVA seems to cause cancer-
promoting biological changes apart
from DNA lesions that result in ge-
nomic mutations. Many of the carci-
nogenic functions of UVA have been
attributed to the production of reac-
CHAPTER 2.4
SECTION 2
tive oxygen species and the subse-
quent induction of the inflammatory
signalling pathway. Reactive oxygen
species generated by UV radiation
upregulate the expression of many
signalling molecules, including in-
ducible nitric oxide synthase (iNOS),
nuclear factor kappa-light-chain-
enhancer of activated B cells (NF-
κB), activator protein 1 (AP-1), signal
transducer and activator of transcrip-
tion (STAT), and cyclooxygenase 2
(COX-2), resulting in inflammation,
which is followed, in turn, by gen-
eration of reactive oxygen species,
and sufficient to induce melanoma Role of UVA in depending on the strength of the
inflammation.
with a high incidence [19]. This pro- photocarcinogenesis
vides an experimental basis for the Until recently, studies on carcino-
epidemiological evidence that child- genesis induced by UV radiation Ultraviolet-induced
hood sunburn is a major risk factor for have focused on UVB-induced DNA immunosuppression
the development of melanoma [20]. mutations. However, the role of UVA The immune system plays an impor-
Whether UVA or UVB radiation is in photocarcinogenesis is now re- tant role in UV-induced carcinogen-
more dangerous for the development ceiving much more attention. One esis by contributing to host resis-
of melanoma is still controversial. reason for this is increasing aware- tance to skin cancer development.
Both non-melanocytic skin cancers ness of the involvement of UVA- However, UV radiation may circum-
and melanomas are induced by so- induced reactive oxygen species vent immunosurveillance against skin
lar UV radiation, but there are some in the development of melanoma. cancers by modulating the immune
differences. Melanocytes show re- Another reason is that many studies response in a way that favours tu-
sistance to UVB-induced apoptosis. have revealed that UVA generates mour development.
Consequently, melanocytes survive not only reactive oxygen species but Skin cancers induced by UV ra-
after acute sunburn, while harbouring also cyclobutane pyrimidine dimers diation are highly antigenic, and can
high levels of DNA lesions, whereas in vivo. therefore be recognized by the im-
keratinocytes tend to undergo apopto- Cyclobutane pyrimidine dimers mune system. This is evident from UV-
are now known to be produced at induced murine skin cancers, many
sis after large doses of UV radiation.
higher yields than 8-oxoG after UVA of which are immunologically rejected
The most frequent body sites for the
irradiation in rodent and human skin upon transplantation into normal syn-
development of superficial spreading
cells [9], prompting a paradigm shift in geneic mice [3]. The exceptionally
melanoma, which is the most com-
the theory of photocarcinogenesis. A high incidence of skin cancers, par-
mon type of malignant melanoma in recent series of studies demonstrating ticularly squamous cell carcinoma, in
the White population, are the trunk that UVA induces thymine dimers at the sun-exposed skin of immunosup-
and thigh; these anatomical regions much higher levels than other types pressed renal transplant recipients or
are often particularly exposed to the of pyrimidine dimers, and that UVA patients who received phototherapy
sun when sunbathing. Eumelanin does not induce (6–4) photoproducts together with an immunosuppressant
protects the skin against UV-induced [9], explains the mutation spectrum of [21] suggests that UV-induced human
damage, whereas pheomelanin acts the relevant genes in cancers at sun- skin cancers are also highly antigenic.
as a photosensitizer and causes oxi- exposed areas of the skin in humans However, despite the potential for
dative DNA damage in melanocytes. [2]. An in vivo study analysing the ac- immunological control, skin cancers

occur with a high frequency in sus- Fig. 2.4.5. Sun protection during play. Avoiding unnecessary sun exposure is critical to
ceptible, sun-exposed populations. avoiding sunburn and the associated risk of skin cancer.
Earlier studies, mainly those using
mouse models, have provided an
explanation for this paradox by dem-
onstrating that UV radiation not only
transforms cells by inducing muta-
tions but also interferes with host
immunity against the developing
skin tumours. These studies dem-
onstrated that UV irradiation of the
skin produces both local immunosup-
pression, which inhibits immune func-
tions within the irradiated skin, and
systemic immunosuppression against
antigens introduced at a critical time
after exposure to UV radiation.
Modulation of immune responses
initiated at non-irradiated sites is now
known to involve soluble mediators.
Among such soluble mediators, IL-10
is crucial in the photocarcinogenesis
pathway [22]. IL-10 polymorphisms wearing protective clothing, wearing exposed areas from early childhood,
and susceptibility to squamous cell a hat, applying sunscreen, and us- because of deficiency in the repair of
carcinoma have been reported in ing shade. Minimizing the time spent DNA photolesions [2].
several studies in humans. outdoors between the hours of 9:00 Recently, accelerated photoage-
Failure of immunosurveillance is a.m. and 3:00 p.m. – the period when ing and development of skin cancer
closely related in photocarcinogen- the intensity of sunlight is the strong- have been reported in patients who
esis, and in this context, the use of est – markedly reduces the risk of sun developed severe photosensitivity
a Toll-like receptor agonist recently damage. disorders after being treated with
emerged as a new strategy for cancer Members of the public should be voriconazole, an antifungal agent
treatment. Imiquimod, an agonist for advised that the strength of UV radia- [23]. Use of the diuretic antihyperten-
Toll-like receptor 7, is now clinically tion does not correlate with the tem- sive medication hydrochlorothiazide
used worldwide for the therapy of perature. For example, in March in the was associated with increased risk
actinic keratosis, a precancerous le- Northern Hemisphere, the intensity of of non-melanoma skin cancer in a
sion caused by sun damage that has UV radiation is strong even though nationwide case–control study in
the potential to progress to squamous temperatures may be low. Even on Denmark [24]. This epidemiological
cell carcinoma. cloudy days, about 80% of the solar result is consistent with the finding
UV radiation reaches ground level. that hydrochlorothiazide significantly
About 10% of solar UVB radiation increased the production of thymine
Prevention of damage passes through glass windows. dimers after exposure to radiation in
from solar ultraviolet In relation to photocarcinogene- the UVA range [10]. Taking account of
radiation sis, the heritable disease xeroderma these data and results from studies in
The most effective way to reduce pigmentosum should be kept in mind. animals and in humans, increased at-
skin cancer incidence is to avoid un- Xeroderma pigmentosum is charac- tention should be paid to any severe
necessary sun exposure and adopt terized by an extreme sensitivity to inflammatory lesions that are subject
personal preventive measures for sunlight and a greatly increased risk to UV radiation.
protection from sunlight, such as of developing skin cancers at sun-
82
References
1. Ziegler A, Jonason AS, Leffell DJ, Simon 9. Mouret S, Baudouin C, Charveron M, Favier 18. Rodier F, Coppé JP, Patil CK, Hoeijmakers
JA, Sharma HW, Kimmelman J, et al. A, Cadet J, Douki T (2006). Cyclobutane WA, Muñoz DP, Raza SR, et al. (2009).
(1994). Sunburn and p53 in the onset of skin pyrimidine dimers are predominant DNA Persistent DNA damage signalling trig-
cancer. Nature. 372(6508):773–6. https:// lesions in whole human skin exposed to gers senescence-associated inflamma-
doi.org/10.1038/372773a0 PMID:7997263 UVA radiation. Proc Natl Acad Sci U S A. tory cytokine secretion. Nat Cell Biol.
103(37):13765–70. https://doi.org/10.1073/ 11(8):973–9. https://doi.org/10.1038/ncb
2. Nishigori C, Hattori Y, Toyokuni S pnas.0604213103 PMID:16954188 1909 PMID:19597488
(2004). Role of reactive oxygen spe-
cies in skin carcinogenesis. Antioxid 10. Kunisada M, Masaki T, Ono R, Morinaga 19. Noonan FP, Recio JA, Takayama H, Duray
CHAPTER 2.4
SECTION 2
Redox Signal. 6(3):561–70. https://doi. H, Nakano E, Yogianti F, et al. (2013). P, Anver MR, Rush WL, et al. (2001).
o r g /10 .10 8 9 /15 2 3 0 8 6 0 47 7 3 9 3 4 314 Hydrochlorothiazide enhances UVA- Neonatal sunburn and melanoma in mice.
PMID:15130282 induced DNA damage. Photochem Nature. 413(6853):271–2. https://doi.
Photobiol. 89(3):649–54. https://doi.org/
org/10.1038/35095108 PMID:11565020
3. Kripke ML (1974). Antigenicity of murine 10.1111/php.12048 PMID:23331297
skin tumors induced by ultraviolet light. J 20. Whiteman DC, Whiteman CA, Green
11. de Gruijl FR (1995). Action spectrum for
Natl Cancer Inst. 53(5):1333–6. https://doi. AC (2001). Childhood sun exposure as
photocarcinogenesis. Recent Results
org/10.1093/jnci/53.5.1333 PMID:4139281 a risk factor for melanoma: a system-
Cancer Res. 139:21–30. https://doi.
o r g /10 .10 0 7/ 9 7 8 - 3 - 6 42 -7 8 7 71- 3 _ 2 atic review of epidemiologic studies.
4. Nishigori C, Yarosh DB, Ullrich SE, Vink AA, Cancer Causes Control. 12(1):69–82.
Bucana CD, Roza L, et al. (1996). Evidence PMID:7597292
https://doi.org/10.1023/A:1008980919928
that DNA damage triggers interleukin 10 12. Nishigori C (2006). Cellular aspects of pho- PMID:11227927
cytokine production in UV-irradiated mu- tocarcinogenesis. Photochem Photobiol
rine keratinocytes. Proc Natl Acad Sci U S Sci. 5(2):208–14. https://doi.org/10.1039/ 21. Naldi L (2010). Malignancy concerns with
A. 93(19):10354–9. https://doi.org/10.1073/ B507471A PMID:16465307 psoriasis treatments using phototherapy,
pnas.93.19.10354 PMID:8816804 methotrexate, cyclosporin, and biologics:
13. Pierceall WE, Goldberg LH, Tainsky facts and controversies. Clin Dermatol.
5. Eisemann N, Waldmann A, Geller AC, MA, Mukhopadhyay T, Ananthaswamy 28(1):88–92. https://doi.org/10.1016/j.
Weinstock MA, Volkmer B, Greinert R, HN (1991). Ras gene mutation and am- clindermatol.2009.03.003 PMID:20082957
et al. (2014). Non-melanoma skin can- plification in human nonmelanoma skin
cer incidence and impact of skin cancer cancers. Mol Carcinog. 4(3):196–202. 22. Loser K, Apelt J, Voskort M, Mohaupt M,
screening on incidence. J Invest Dermatol. https://doi.org/10.1002/mc.2940040306 Balkow S, Schwarz T, et al. (2007). IL-10
134(1):43–50. https://doi.org/10.1038/jid. PMID:2064725 controls ultraviolet-induced carcinogen-
2013.304 PMID:23877569 esis in mice. J Immunol. 179(1):365–71.
14. Devary Y, Gottlieb RA, Smeal T, Karin M https://doi.org/10.4049/jimmunol.179.1.365
6. Tierney P, de Gruijl FR, Ibbotson S, (1992). The mammalian ultraviolet response
PMID:17579057
Moseley H (2015). Predicted increased is triggered by activation of Src tyrosine
risk of squamous cell carcinoma induction kinases. Cell. 71(7):1081–91. https://doi. 23. Sheu J, Hawryluk EB, Guo D, London
associated with sunbed exposure habits. org/10.1016/S0092- 8674(05)80058 -3 WB, Huang JT (2015). Voriconazole pho-
Br J Dermatol. 173(1):201–8. https://doi. PMID:1473146 totoxicity in children: a retrospective re-
org/10.1111/bjd.13714 PMID:25645571 15. Köpcke W, Krutmann J (2008). Protection view. J Am Acad Dermatol. 72(2):314–20.
from sunburn with beta-carotene – a https://doi.org/10.1016/j.jaad.2014.10.023
7. Ellison MJ, Childs JD (1981). Pyrimidine PMID:25481710
dimers induced in Escherichia coli DNA meta-analysis. Photochem Photobiol.
by ultraviolet radiation present in sunlight. 84(2):284–8. https://doi.org/10.1111/j.1751- 24. Pedersen SA, Gaist D, Schmidt SAJ,
1097.2007.00253.x PMID:18086246
Photochem Photobiol. 34(4):465–9. https:// Hölmich LR, Friis S, Pottegård A (2018).
doi.org/10.1111/j.1751-1097.1981.tb09387.x 16. Sun Y, Oberley LW (1996). Redox regu- Hydrochlorothiazide use and risk of non-
PMID:7031709 lation of transcriptional activators. Free melanoma skin cancer: a nationwide
Radic Biol Med. 21(3):335–48. https:// case-control study from Denmark. J
8. Kunisada M, Sakumi K, Tominaga Y, Am Acad Dermatol. 78(4):673–681.e9.
doi.org/10.1016/0891-5849(96)00109-8
Budiyanto A, Ueda M, Ichihashi M, et al. https://doi.org/10.1016/j.jaad.2017.11.042
PMID:8855444
(2005). 8-Oxoguanine formation induced PMID:29217346
by chronic UVB exposure makes Ogg1 17. Moore RJ, Owens DM, Stamp G, Arnott
knockout mice susceptible to skin carcino- C, Burke F, East N, et al. (1999). Mice de-
genesis. Cancer Res. 65(14):6006–10. ficient in tumor necrosis factor-alpha are
https://doi.org/10.1158/0008-5472.CAN- resistant to skin carcinogenesis. Nat Med.
05-0724 PMID:16024598 5(7):828–31. https://doi.org/10.1038/10552
PMID:10395330

2.5
Ionizing radiation and radiofrequency

electromagnetic fields
Further clarification of particular risks
Dominique Laurier Maria Blettner (reviewer)
Martin Röösli Ausrele Kesminiene (reviewer)
Colin R. Muirhead (reviewer)
γ-radiation) and energetic subatom- and construction, and this can lead
SUMMARY ic particles (neutrons, β-particles, to environmental, occupational, or
and α-particles). This type of radia- medical exposures. Environmental
●● Epidemiological studies involv- tion carries enough energy to liber- exposures include fallout from weap-
ing people exposed to low lev- ate electrons from atoms and thus ons testing, nuclear power plant ac-
els of ionizing radiation from is able to break chemical bonds. cidents (such as those at Chernobyl
the environment (natural and Biological effects of ionizing radi- and Fukushima), and routine re-
artificial sources), occupations, ation are determined by the amount leases from nuclear installations.
or medical diagnostic proce- of energy absorbed by the exposed Exposures to medical radiation pro-
dures demonstrate that the risk organ or tissue. Low doses are gen- vide a direct benefit to the exposed
of leukaemia and other cancers erally defined as effective doses be- individuals. These exposures arise
increases with radiation dose. low 100 millisieverts (mSv). from some diagnostic procedures,
●● The latency between exposure such as radiography, nuclear medi-
to ionizing radiation and occur-
Sources and exposures cine, and computed tomography
rence of an excess risk of cancer Humans have always been exposed (CT), or as a consequence of treat-
varies from several years to sev- to ionizing radiation from natural ment, most commonly radiotherapy
eral decades. In addition, host sources. Natural radiation exposure for cancer. Medical uses of radia-
factors such as age at exposure, comes from four main sources: cos- tion have increased rapidly as tech-
attained age, and sex modify the mic radiation, terrestrial radiation, niques have been developed and
dose–risk relationship. ingestion of radionuclides present widely disseminated.
in the soil and ground, and inhala- The contributions of the main
●● Most of the epidemiological tion of radon. Exposure to cosmic components of average population
research does not support an radiation is higher at high altitudes. exposure are detailed in Fig. 2.5.1.
association between mobile Exposure to natural radionuclides The worldwide average annual ef-
phone use and tumours occur- varies considerably from place to fective dose is about 3 mSv, and
ring in the head, which is the place according to geology. Radon individual doses vary from tenths of
body part with the highest ex- is a gas that is formed during the millisieverts to several tens of mil-
posure to radiofrequency elec- decay of natural uranium in the soil. lisieverts, according to place of resi-
tromagnetic fields. In studies Exposure to indoor radon varies dence and behaviour [1].
reporting positive associations, depending on the geology, build-
it is difficult to exclude various ing construction, and household Cancer causation
forms of bias, such as recall lifestyle. Worldwide, inhalation of Ionizing radiation is one of the most
bias in retrospective exposure radon accounts for about half of the intensely studied carcinogens [2].
assessment. average exposure to natural radia- The mechanisms by which radiation
tion sources [1]. may produce carcinogenic changes
In addition, artificial sources of include mutations, alterations in the
Ionizing radiation exposure have developed over the structure of genes or chromosomes
Ionizing radiation is made up of past century. Today, ionizing radia- (see Chapter 3.11), and changes in
electromagnetic waves on the tion is encountered in a wide variety gene expression. Radiobiological
high-energy end of the electromag- of fields, such as medicine, nuclear research in recent decades has
netic spectrum (X-radiation and power, research, manufacturing, shown the biological complexity of
84
Fig. 2.5.1. Average annual doses of ionizing radiation by source. The worldwide average
annual effective dose is about 3 millisieverts (mSv). Natural sources (2.4 mSv; 80%) are
shown in green, and artificial sources (0.6 mSv; 20%) are shown in pink. Environmental FUNDAMENTALS
artificial sources include atmospheric nuclear testing (0.2%), releases from the Chernobyl
accident (0.1%), and routine releases from the nuclear fuel cycle (0.01%). ■■ The electromagnetic spectrum
is divided into non-ionizing and
Occupational Environmental ionizing radiation.
exposure artificial sources
(0.01%) (0.3%) ■■ On average, natural sources
contribute 80% to the
average total dose of ionizing
CHAPTER 2.5
SECTION 2
Medical diagnosis Inhalation radiation in the population.
(not therapy) of radon gas
(20%) (41%)
The remaining 20% originates
from artificial sources, such as
medical diagnostic procedures,
Ingestion External terrestrial atmospheric nuclear testing,
(10%) (16%)
and nuclear power plant
accidents. Inhalation of radon
Cosmic radiation is the single highest source
(13%) of exposure.
■■ Ionizing radiation is able to
produce carcinogenic changes
the carcinogenic impact of radiation, cludes more than 86 000 people of by mutations, alterations in
and many uncertainties still remain, both sexes and all ages, with acute the structure of genes or
especially at low doses. external radiation exposure. The chromosomes, and changes
Evidence that ionizing radia- range of doses was 0–4 Sv, but in gene expression. Although
tion can cause cancer in humans about 80% of the survivors received considerable uncertainty still
comes from epidemiological stud- less than 100 mSv. Recent results exists about the form of the
ies, especially from studies of pa- confirmed the existence of a dose– dose–response relationship
tients irradiated for therapeutic risk relationship for a large variety in the low-dose range,
reasons and from the follow-up of of cancer types, such as leukaemia an increasing number of
Japanese atomic bomb survivors. and cancer of the bladder, breast, epidemiological studies
In recent decades, other studies colon, liver, lung, skin, stomach, indicate the carcinogenicity of
have provided complementary re- and thyroid, and improved the es- ionizing radiation at relatively
sults in populations exposed to low- timation of how the risk varies with low dose levels.
er doses, from environmental (e.g. age at exposure and attained age.
■■ Most of the exposure
natural exposure, consequences A statistically significant dose–re-
to radiofrequency
of nuclear accidents), occupational sponse relationship was observed
electromagnetic fields arises
(e.g. miners, nuclear workers), or for incidence of solid cancers in the
from people’s own mobile
medical (e.g. diagnostic proce- 0–100 mSv dose range [3].
phone calls, and thus the head
dures) situations. Patients is the most exposed body part.
The latency between exposure
CT is a highly informative medi- ■■ Despite considerable research
to ionizing radiation and occur-
cal imaging technique, but it leads efforts, no mechanism
rence of an excess risk of cancer
to much higher doses than con- relevant for carcinogenesis
varies from several years to several
ventional radiology. Therefore, of radiofrequency
decades. In addition, host factors
the increasing use of CT scans in electromagnetic fields has
such as age at exposure, attained
paediatric populations raised the been consistently identified
age, and sex modify the dose–risk
question of a possible health impact to date. Also, most of the
relationship.
of radiation exposure. epidemiological research does
Recent epidemiological Cohort studies in Australia and not indicate carcinogenicity of
results the United Kingdom showed a sta-
radiofrequency electromagnetic
tistically significant dose–response
Atomic bomb survivors fields. This implies that any
relationship between the dose to
potentially undetected risk is
The follow-up of cancer mortality the red bone marrow due to CT ex-
expected to be small from an
and incidence in the cohort of atom- aminations and the risk of leukae-
individual perspective.
ic bomb survivors exceeds 60 years mia, and between the dose to the
after exposure. This large cohort in- brain and the risk of brain tumours.
Chapter 2.5 • Ionizing radiation and radiofrequency electromagnetic fields 85

Fig. 2.5.2. A man undergoes a computed tomography (CT) scan, which involves ex- sation). More recent studies tried
posure to ionizing radiation. to address these issues, and they
suggest that these potential biases
should be small [4]. The European
EPI-CT project, which includes
more than 1 million children, will
provide new results on cancer
risks associated with paediatric CT
scans [5].
More information about thy-
roid cancer risks at low doses was
provided by a pooled analysis of
nine cohorts of more than 100 000
children (eight medical cohorts of
children treated for benign and ma-
lignant diseases and the cohort of
atomic bomb survivors). It showed
a statistically significant linear
dose–response relationship for thy-
roid doses of 0–100 mSv [6].
Workers
Nuclear industry workers are ex-
posed to protracted low-dose radia-
tion and are individually monitored
for their occupational exposure.
Several studies were published in re-
cent years in France, Japan, Taiwan
These results raised controversies linked to underlying medical con- (China), the United Kingdom, and
about the impact of uncertainties ditions (e.g. higher prevalence of the USA, including results from the
in dosimetry and potential bias predisposing factors, inverse cau- INWORKS project (see “INWORKS:
INWORKS: a pooled analysis of cancer risks associated with ionizing radiation

among nuclear workers
The International Nuclear Workers Analyses demonstrated a sig- for the protection of adults from
Study (INWORKS) is a multina- nificant association between the low-dose, low-dose-rate exposures
tional research project coordi- dose to the red bone marrow and to ionizing radiation.
nated by IARC. It evaluated the the risk of leukaemia (excluding
exposures of more than 300 000 chronic lymphoblastic leukae- References
workers in the nuclear industry in mia), and between the dose to 1. Leuraud K, Richardson DB, Cardis E,
Daniels RD, Gillies M, O’Hagan JA, et al.
France, the United Kingdom, and the colon and the risk of solid
(2015). Ionising radiation and risk of death
the USA, with detailed individual cancers [1,2]. These associations from leukaemia and lymphoma in radi-
monitoring data for external expo- were significant even when the ation-monitored workers (INWORKS):
an international cohort study. Lancet
sure to ionizing radiation. analyses were restricted to a low- Haematol. 2(7):e276–81. https://doi.
Over an average follow-up dose range of 0–300 mSv. The org/10.1016/S2352-3026(15)00094- 0
duration of 27 years, there were estimated dose–risk coefficients PMID:26436129
17 957 deaths due to solid can- were very consistent with those 2. Richardson DB, Cardis E, Daniels RD,
cers and 1791 deaths due to hae- derived from the cohort of atomic Gillies M, O’Hagan JA, Hamra GB, et
al. (2015). Risk of cancer from occu-
matological cancers. The average bomb survivors, which form the pational exposure to ionising radiation:
individual cumulative external main basis for the system of ra- retrospective cohort study of workers
dose over the period 1945–2005 diological protection. in France, the United Kingdom, and
the United States (INWORKS). BMJ.
was 21 mSv to the colon and INWORKS is contributing to 351:h5359. https://doi.org/10.1136/bmj.
16 mSv to the red bone marrow. strengthening the scientific basis h5359 PMID:26487649
86
a pooled analysis of cancer risks Fig. 2.5.3. Debris from the upper levels of unit 4 at the Fukushima Daiichi power plant
associated with ionizing radiation in December 2012, 21 months after the nuclear accident.
among nuclear workers”). These re-
sults strengthen the quantification of
risks associated with external expo-
sures to ionizing radiation at a low
dose rate.
Other studies quantified a dose
relationship for specific internal ex-
posures. A recent analysis of co-
CHAPTER 2.5
SECTION 2
horts of uranium miners confirmed
the association between radon
exposure and risk of lung cancer,
even among miners with low levels
of exposure [7]; the results were
consistent with those from studies
of indoor radon. Also, an analy-
sis of the cohort of workers from
the Mayak nuclear facility in the
Russian Federation confirmed the
existence of a relationship between
lung dose due to plutonium and
lung cancer risk, compatible with a
linear model without threshold [8].
Nuclear accidents
The largest nuclear accident in the in much lower thyroid doses to the children in Switzerland, including
world occurred on 26 April 1986 resident populations than after the 530 leukaemia cases and 1252
at the Chernobyl nuclear plant in Chernobyl accident. The estimated cases of other childhood cancers,
Ukraine. This accident resulted in doses are low and are limited to a suggested a positive relationship
a large release of radionuclides, small population, and no observ- between exposure to background
which were deposited over a very able radiation-induced excess risk radiation and both leukaemia risk
wide area; the greatest deposits of cancer is expected. and cancer risk, at the limit of sta-
were in Belarus, the western part A large project has been tistical significance [13].
of the Russian Federation, and launched, called the Fukushima
Ukraine. Recent results confirmed Health Management Survey, which Prevention
the excess risk of thyroid cancer includes systematic thyroid exami- A comprehensive system of pro-
associated with exposure to io- nations of children and adolescents. tection against ionizing radiation
dine-131 among people exposed A large number of thyroid cancer has been developed, based espe-
during childhood, and demonstrat- cases have been recorded [10], but cially on recommendations from
ed the persistence of this excess these are mostly attributable to the the International Commission on
risk among people who are now implementation of screening, which Radiological Protection. Recent
adults (see Chapter 5.18). About led to the detection of small, indolent studies have improved our knowl-
25% of thyroid cancer cases in the cancers and to overdiagnosis [11]. edge of radiation-induced risks at
contaminated area among people low doses, down to a few hundreds
who were children or adolescents Other environmental exposures of millisieverts for solid cancers
at the time of the accident have A case–control study in Great [14] and a few tens of millisie-
been attributed to this exposure [9]. Britain that included more than verts for childhood leukaemia [15];
The Fukushima Daiichi nuclear 9000 cases of leukaemia and these results have contributed to
accident occurred on 11 March 18 000 cases of other childhood the strengthening of the radiation
2011 in Japan. Compared with the cancers observed a statistically sig- protection system. In the medical
Chernobyl accident, this accident nificant dose–response relationship field, the benefits of radiation appli-
resulted in a much lower release of between leukaemia and the cumu- cations for medical diagnostic pro-
radionuclides, which were essen- lative dose to the red bone marrow cedures are undeniable, but recent
tially deposited over some parts of due to background radiation expo- results from epidemiology and the
Fukushima Prefecture. Furthermore, sure, but found no clear evidence increasing use of CT scans high-
preventive measures, such as evac- of a relationship for other childhood light the need to enhance aware-
uation and food restrictions, resulted cancers [12]. A cohort of 2 million ness among medical practitioners

and to reinforce prevention, through Fig. 2.5.4. Contribution of various sources that emit radiofrequency electromagnetic
the use of dose optimization and fields (RF-EMF) to the average daily dose to grey matter in the brain in a cohort of
procedure justification. adolescents in Switzerland. The total average RF-EMF dose was 900 millijoules per
kilogram (mJ/kg) per day. Mobile phone calls and data traffic contributed 85%. Far-
field sources (6%) included mobile phone base stations (3.4%), other people’s mobile
Radiofrequency phones (2.0%), access points in wireless local area networks (0.2%), broadcasting
electromagnetic fields (0.2%), and cordless phone base stations (< 0.1%).
Sources and exposures Mobile phone Standby mobile

Radiofrequency electromagnetic data traffic phone on body
(5%) (0.3%)
fields (RF-EMF) are emitted from
various sources. For the public, the Mobile phone Cordless phone use
most relevant sources in daily life calls (9%)
are wireless communication devic- (80%)
es and transmitters. Computer, tablet
(wireless local
Wireless phones and other de- area networks)
vices that are used close to the (0.01%)
body produce a near-field exposure,
which is characterized by the spe- Far-field sources
(6%)
cific absorption rate (expressed in
watts per kilogram of tissue weight)
[16]. Transmitters that are further
away, such as access points in
wireless local area networks, base
stations for mobile and cordless System for Mobile Communications evant for carcinogenesis has been
phones, broadcast transmitters, and (GSM) network (2G) was shown consistently identified to date [21].
other people’s mobile phones, are to be 100–500 times that for
far-field sources, and the most com- calls on the Universal Mobile Recent epidemiological
mon exposure metric is the incident Telecommunications System (UMTS) results
electric field (in volts per metre). network (3G) [18,19]. This implies In the past 5 years, epidemiologi-
Combining the two exposure mea- that in new epidemiological research cal research on mobile phone use
sures into a single dose measure re- since the introduction of UMTS, and tumours occurring in the head
quires dosimetric calculations. one would expect a lower cumula- has slowed down compared with
In a recent cohort study of ado- tive dose to the brain for the same the previous decade. Most new
lescents in Switzerland, contribu- amount of mobile phone use. The and previous case–control studies
tions of various RF-EMF sources to increased variability in the output do not indicate an association be-
the dose to grey matter in the brain power of mobile phones implies that tween mobile phone use and risk
were estimated [17]. In this cohort of in new studies, duration of mobile of glioma, meningioma, acoustic
moderate users of mobile phones phone use has become a less valid neuroma, pituitary tumours, or sali-
and cordless phones (with calls last- surrogate of the RF-EMF exposure vary gland tumours [22]. Sporadic
ing on average 11 minutes and 6 min- of the brain than in older studies. It associations observed in a few
utes per day, respectively), mobile is not yet known what the situation case–control studies are inconsis-
and cordless phone calls contrib- will be for the Long-Term Evolution tent in terms of exposure–response
uted 80% and 8%, respectively, to (LTE) network (4G) or for 5G. associations. For example, in a new
the average grey matter dose from analysis of pooled case–control
RF-EMF (Fig. 2.5.4). The proportion Cancer causation studies in Sweden, with cases di-
from all far-field sources combined Because RF-EMF belong to the agnosed in 1997–2003 and 2007–
was 6%, including 3% from mobile non-ionizing part of the electro- 2009, glioma risk was higher for
phone base stations and 2% from magnetic spectrum, the photon en- people with at least 123 hours of
other people’s mobile phones. ergy is too weak to ionize molecules cumulative use [23], whereas in a
As technology and knowledge [20] and thereby cause direct DNA case–control study in France with
advance, these dose estimates may damage. Absorption of RF-EMF 253 glioma cases and 504 controls,
change. One of the main uncertain- is known to heat biological tissue, glioma risk was significantly higher
ties in such calculations is the adap- but a minimal temperature increase for people with at least 339 hours of
tive power control of mobile phones below the regulatory limits is not cumulative use [24]. In contrast, in
in response to the network quality. expected to increase the risk of the large international Interphone
For instance, the average output cancer [16]. Despite considerable study, which included 2708 glioma
power for calls made on the Global research efforts, no mechanism rel- cases, 2409 meningioma cases,
88
and 1105 acoustic neuroma cases, have used a mobile phone for a few topographic or morphological sub-
no indication of higher risk was hundred hours, and simple calcula- types of brain cancer. However, in
observed for cumulative use up to tions demonstrate that some of the the same studies, a decrease in the
1640 hours [25]. Thus, there is con- reported excess risks for brain tu- incidence of other subtypes of brain
cern that some studies are affected mours would have been noticed by cancer was seen, suggesting that
by recall bias, because cases may now. For instance, the populations these time trends may be explained
overestimate their previous mobile of the Nordic countries were among by changes in cancer coding prac-
phone use as a potential cause of the first to use mobile phones regu- tices over time.
their disease. larly, and in Europe a 50% pen- Research on exposures from
A recent study followed up 806 etration rate was achieved in 2000. transmitters has not progressed
CHAPTER 2.5
SECTION 2
glioma cases previously enrolled Thus, substantially more than 50% much in the past 5 years, and the
in a collaborative population-based of the population in European coun- evidence base has not expanded.
case–control study in Denmark, tries are now long-term mobile Several reported clusters of child-
Finland, and Sweden for survival and phone users, and reported excess hood cancer in the vicinity of in-
found no evidence of reduced sur- risks on the order of 60–70% for dividual transmitters could not
vival in relation to mobile phone use long-term users would produce an be confirmed in large population-
[26]. Strikingly, this study found some increase in the incidence of gli- based studies on childhood cancer
indications that prodromal symptoms oma of at least 30%, which is not in relation to RF-EMF emissions
of the tumour may prevent cases the case in people younger than from broadcast transmitters and
from starting to regularly use mobile 70 years [22]. mobile phone base stations [32].
phones, which may explain some of In addition, a very comprehen- For adults, even fewer studies have
the seemingly protective effects of sive analysis of global trends of tu- been conducted. However, RF-
mobile phone use observed previ- mours of the brain and central ner- EMF from transmitters will rarely
ously in the Interphone study [25]. vous system, which included data be a relevant exposure source for
In summary, such kinds of re- from 1993–2007 from 96 registries adults who at least occasionally use
verse causality, recall bias, and se- in 39 countries, did not find a pat- wireless communication devices.
lection bias are potential issues in a tern supporting the hypothesis of in-
case–control study. The continuing creasing incidence rates following, Prevention
prospective COSMOS study, which with some latency, the time period The large amount of research on
is using operator-recorded data for of mobile phone uptake in different RF-EMF suggests that any poten-
mobile phone use, is less vulner- populations [28]. This analysis is in tially undetected risk is expected
able to such kinds of recall bias and line with the results of several other to be small from an individual per-
exposure misclassification [27]. time trend studies [29], although spective. To address such small
Nowadays it is common for a a few studies [30,31] reported in- risks needs high-quality research
large proportion of the population to creases in the incidence of specific with accurate exposure assess-
ment, taking into account that the
duration of mobile phone calls
Fig. 2.5.5. Women using mobile phones in Kolkata, India. alone is not expected to adequately
reflect the RF-EMF exposure of the
brain. In the meantime, for tumours
of the head with few other risk fac-
tors, monitoring of incidence rates
is a suitable approach to detect rel-
evant changes in incidence rates
possibly related to the use of wire-
less phones.
Given the research uncertain-
ties, precautionary measures might
be taken. Because mobile phones
are the most relevant exposure
source and because the strength
of RF-EMF decreases rapidly with
distance from the source, the sim-
plest and most effective precaution-
ary measure is to hold the mobile
phone away from the body during
transmission; this will result in a
substantial reduction in exposure.

References
1. UNSCEAR (2010). Sources and effects of 10. Yamashita S, Suzuki S, Suzuki S, 19. Persson T, Törnevik C, Larsson LE,
ionizing radiation. UNSCEAR 2008 Report Shimura H, Saenko V (2018). Lessons Lovén J (2012). Output power distribu-
to the General Assembly, with scientific an- from Fukushima: latest findings of thy- tions of terminals in a 3G mobile commu-
nexes. Volume I: Sources. New York (NY), roid cancer after the Fukushima Nuclear nication network. Bioelectromagnetics.
USA: United Nations Scientific Committee Power Plant accident. Thyroid. 28(1):11– 33(4):320–5. https://doi.org/10.1002/bem.
on the Effects of Atomic Radiation. 22. https://doi.org/10.1089/thy.2017.0283 20710 PMID:22012866
Available from: http://www.unscear.org/ PMID:28954584
unscear/publications/2008_1.html. 20. Challis LJ (2005). Mechanisms for in-
11. IARC Expert Group on Thyroid Health teraction between RF fields and bio-
2. IARC (2012). Radiation. IARC Monogr Monitoring after Nuclear Accidents (2018). logical tissue. Bioelectromagnetics.
Eval Carcinog Risks Hum. 100D:1–437. Thyroid health monitoring after nuclear 26(Suppl 7):S98–106. https://doi.org/
PMID:23189752. Available from: http:// accidents. Lyon, France: International 10.1002/bem.20119 PMID:15931683
publications.iarc.fr/121. Agency for Research on Cancer (IARC
Technical Publications, No. 46). Available 21. Parham F, Portier CJ, Chang X, Mevissen
3. Grant EJ, Brenner A, Sugiyama H, Sakata from http://publications.iarc.fr/571. M (2016). The use of signal-transduction
R, Sadakane A, Utada M, et al. (2017). Solid and metabolic pathways to predict hu-
cancer incidence among the Life Span Study 12. Kendall GM, Little MP, Wakeford R, Bunch
man disease targets from electric and
of atomic bomb survivors: 1958–2009. KJ, Miles JC, Vincent TJ, et al. (2013).
magnetic fields using in vitro data in hu-
Radiat Res. 187(5):513–37. https://doi. A record-based case-control study of
man cell lines. Front Public Health. 4:193.
org/10.1667/RR14492.1 PMID:28319463 natural background radiation and the in-
https://doi.org/10.3389/fpubh.2016.00193
cidence of childhood leukaemia and other
4. Berrington de Gonzalez A, Salotti JA, cancers in Great Britain during 1980- PMID:27656641
McHugh K, Little MP, Harbron RW, Lee C, 2006. Leukemia. 27(1):3–9. https://doi. 22. Röösli M, Lagorio S, Schoemaker MJ,
et al. (2016). Relationship between pae- org/10.1038/leu.2012.151 PMID:22766784 Schüz J, Feychting M (2019). Brain and
diatric CT scans and subsequent risk of salivary gland tumours and mobile phone
leukaemia and brain tumours: assessment 13. Spycher BD, Lupatsch JE, Zwahlen M,
Röösli M, Niggli F, Grotzer MA, et al.; Swiss use: evaluating the evidence from various
of the impact of underlying conditions. epidemiological study designs. Annu Rev
Br J Cancer. 114(4):388–94. https://doi. Pediatric Oncology Group; Swiss National
Cohort Study Group (2015). Background Public Health. 40:221–38. https://doi.org/
org/10.1038/bjc.2015.415 PMID:26882064 10.1146/annurev- publhealth - 0 40218 -
ionizing radiation and the risk of child-
5. Bernier MO, Baysson H, Pearce MS, hood cancer: a census-based nationwide 044037 PMID:30633716
Moissonnier M, Cardis E, Hauptmann M, et cohort study. Environ Health Perspect.
23. Hardell L, Carlberg M (2015). Mobile
al. (2019). Cohort profile: the EPI-CT study: 123(6):622–8. https://doi.org/10.1289/
ehp.1408548 PMID:25707026 phone and cordless phone use and
a European pooled epidemiological study
the risk for glioma – analysis of pooled
to quantify the risk of radiation-induced
14. Shore R, Walsh L, Azizova T, Rühm W case-control studies in Sweden, 1997-
cancer from paediatric CT. Int J Epidemiol.
(2017). Risk of solid cancer in low dose- 2003 and 2007-2009. Pathophysiology.
48(2):379–81g. https://doi.org/10.1093/ije/
rate radiation epidemiological studies 22(1):1–13. https://doi.org/10.1016/j.patho
dyy231 PMID:30388267
and the dose-rate effectiveness factor. phys.2014.10.001 PMID:25466607
6. Lubin JH, Adams MJ, Shore R, Holmberg Int J Radiat Biol. 93(10):1064–78. https://
doi.org/10.1080/09553002.2017.1319090 24. Coureau G, Bouvier G, Lebailly P,
E, Schneider AB, Hawkins MM, et al.
PMID:28421857 Fabbro-Peray P, Gruber A, Leffondre
(2017). Thyroid cancer following child-
K, et al. (2014). Mobile phone use and
hood low-dose radiation exposure: a 15. Little MP, Wakeford R, Borrego D, French
pooled analysis of nine cohorts. J Clin brain tumours in the CERENAT case-
B, Zablotska LB, Adams MJ, et al. (2018). control study. Occup Environ Med.
Endocrinol Metab. 102(7):2575–83. Leukaemia and myeloid malignancy
ht t ps: //doi.org /10.1210 / jc. 2016 - 3 52 9 71(7):514–22. https://doi.org/10.1136/
among people exposed to low doses oemed-2013-101754 PMID:24816517
PMID:28323979 (<100 mSv) of ionising radiation during
7. Kreuzer M, Sobotzki C, Schnelzer M, childhood: a pooled analysis of nine his- 25. INTERPHONE Study Group (2010). Brain
Fenske N (2018). Factors modifying the torical cohort studies. Lancet Haematol. tumour risk in relation to mobile telephone
radon-related lung cancer risk at low 5(8):e346–58. https://doi.org/10.1016/ use: results of the INTERPHONE interna-
exposures and exposure rates among S2352-3026(18)30092-9 PMID:30026010 tional case-control study. Int J Epidemiol.
German uranium miners. Radiat Res. 39(3):675–94. https://doi.org/10.1093/ije/
16. International Commission on Non-
189(2):165–76. https://doi.org/10.1667/ dyq079 PMID:20483835
Ionizing Radiation Protection (1998).
RR14889.1 PMID:29215327 Guidelines for limiting exposure to time- 26. Olsson A, Bouaoun L, Auvinen A,
varying electric, magnetic, and electro- Feychting M, Johansen C, Mathiesen
8. Gillies M, Kuznetsova I, Sokolnikov M,
magnetic fields (up to 300 GHz). Health T, et al. (2019). Survival of glioma pa-
Haylock R, O’Hagan J, Tsareva Y, et al.
Phys. 74(4):494–522. PMID:9525427 tients in relation to mobile phone use
(2017). Lung cancer risk from plutoni-
um: a pooled analysis of the Mayak and 17. Foerster M, Thielens A, Joseph W, in Denmark, Finland and Sweden. J
Sellafield worker cohorts. Radiat Res. Eeftens M, Röösli M (2018). A prospec- Neurooncol. 141(1):139–49. https://
188(6):645–60. https://doi.org/10.1667/ tive cohort study of adolescents’ memory doi.org/10.10 07/s110 6 0 - 018 - 03 019 - 5
RR14719.1 PMID:28985139 performance and individual brain dose of PMID:30421160
microwave radiation from wireless com-
9. UNSCEAR (2018). Evaluation of data 27. Toledano MB, Auvinen A, Tettamanti
munication. Environ Health Perspect.
on thyroid cancer in regions affected by G, Cao Y, Feychting M, Ahlbom A, et
126(7):077007. https://doi.org/10.1289/
the Chernobyl accident: a white paper to al. (2018). An international prospective
EHP2427 PMID:30044230
guide the Scientific Committee’s future cohort study of mobile phone users and
programme of work. New York (NY), USA: 18. Gati A, Hadjem A, Wong M-F, Wiart J health (COSMOS): factors affecting va-
United Nations Scientific Committee on the (2009). Exposure induced by WCDMA mo- lidity of self-reported mobile phone use.
Effects of Atomic Radiation. Available from: biles phones in operating networks. IEEE Int J Hyg Environ Health. 221(1):1–8.
http://www.unscear.org/docs/publications/ Trans Wirel Commun. 8(12):5723–7. https:// https://doi.org/10.1016/j.ijheh.2017.09.008
2017/Chernobyl_WP_2017.pdf. doi.org/10.1109/TWC.2009.12.080758 PMID:29056311
90
28. Miranda-Filho A, Piñeros M, Soerjoma- 30. Philips A, Henshaw DL, Lamburn G, 32. Hauri DD, Spycher B, Huss A, Zimmermann
taram I, Deltour I, Bray F (2017). Cancers O’Carroll MJ (2018). Brain tumours: rise F, Grotzer M, von der Weid N, et al.;
of the brain and CNS: global patterns in glioblastoma multiforme incidence in Swiss National Cohort; Swiss Paediatric
and trends in incidence. Neuro Oncol. England 1995–2015 suggests an adverse Oncology Group (2014). Exposure to ra-
19(2):270–80. PMID:27571887 environmental or lifestyle factor. J Environ dio-frequency electromagnetic fields from
Public Health. 2018:7910754. https://doi. broadcast transmitters and risk of child-
29. Chapman S, Azizi L, Luo Q, Sitas F org/10.1155/2018/7910754 hood cancer: a census-based cohort study.
(2016). Has the incidence of brain can- Am J Epidemiol. 179(7):843–51. https://doi.
cer risen in Australia since the introduc- 31. Hardell L, Carlberg M (2015). Increasing org/10.1093/aje/kwt442 PMID:24651167
tion of mobile phones 29 years ago? rates of brain tumours in the Swedish nation-
Cancer Epidemiol. 42:199–205. https:// al inpatient register and the causes of death
doi.org /10.1016/ j.c anep. 2016.0 4.010 register. Int J Environ Res Public Health.
PMID:27156022 12(4):3793–813. https://doi.org/10.3390/
CHAPTER 2.5
SECTION 2
ijerph120403793 PMID:25854296

2.6
Diet and nutrition

Understanding which factors
are critical
Marjorie L. McCullough Giuseppe Grosso (reviewer)
Walter C. Willett Marc Gunter (reviewer)
Edward L. Giovannucci Sarah Lewis (reviewer)
For many decades, studies in ani- Dietary factors

SUMMARY mals and comparisons of cancer
rates across countries have raised Plant foods
●● Multiple aspects of diet influ- hypotheses that various aspects of Fruits, vegetables, nuts, legumes,
ence cancer risk, some ad-
diet might influence risk of cancer and whole grains are naturally rich
versely and some beneficially.
in humans. Recently, the results of in vitamins, phytochemicals, and
●● Probably most important are long-term epidemiological studies dietary fibre – constituents that are
the influences of diet on adipos- have provided a wealth of informa- thought to inhibit carcinogenesis
ity, a major risk factor for many tion about the relationships between [1]. During the late 20th century,
cancer types. Avoidance of sug- diet and risk of many cancer types. there was a great deal of research
ar-sweetened beverages and Some of the recent evidence has on the role that plant foods may
replacement of refined carbohy- not supported earlier beliefs, for ex- play in reducing the risk of can-
drates with whole-grain alterna- ample that high total fat intake and cer, with initially promising findings
tives is particularly important. low intake of fruits and vegetables originating primarily from case–
are key cancer risk factors. Other control studies. Although the evi-
●● Limiting consumption of red
factors related to nutrition, such dence that fruits and vegetables in-
meat and processed meat,
as overweight (see Chapter 2.7) dependently decrease cancer risk
especially of processed meat,
and alcohol consumption (see has weakened during recent de-
may decrease risk of colorec-
Chapter 2.3), have emerged as cades, the World Cancer Research
tal cancer.
clearly important, and evidence for Fund/American Institute for Cancer
●● Generous consumption of fruits Research (WCRF/AICR) conclud-
a role of overall healthy dietary pat-
and vegetables has less impact ed that greater consumption of
terns has strengthened.
on cancer risk than was thought non-starchy vegetables or fruits
Because dietary and other ex-
earlier, but some benefits exist. probably protects against several
posures many years before the
cancers of the aerodigestive tract
●● An overall healthy dietary pat- diagnosis of cancer, including dur-
[1]. Emerging studies of molecu-
tern that emphasizes low ining childhood, can influence can-
larly defined tumour subtypes can
take of red meat and processed cer risk, current evidence on diet identify different associations with
meat, generous intake of fruits and cancer remains incomplete, plant foods and/or their constitu-
and vegetables, whole grains and continued research is needed. ents. For example, higher concen-
rather than refined grains, and In addition, more research on diet trations of β-carotene, α-carotene,
plant sources of protein and fat and cancer is needed in countries and other carotenoids found in
will reduce risk of cancer as well undergoing the nutrition transition fruits and vegetables are associat-
as of cardiometabolic disease. towards a lifestyle typical of indus- ed with lower risk of more aggres-
●● Although data on the effects of trialized countries, where the inci- sive and deadly breast tumours [2],
diet after cancer diagnosis on dence of diet-related cancer types including estrogen receptor-nega-
overall and cancer-specific sur- (e.g. colorectal cancer) is rising. tive breast tumours [3].
vival are sparse, recent findings This chapter briefly describes The evidence that consumption
support adopting a similar di- the current state of knowledge, with of whole grains (i.e. grains in which
etary pattern as for prevention. an emphasis on findings during the 100% of the original kernel is re-
past 5 years. tained) decreases risk of colorectal
92
cancer was categorized by WCRF/ of N-nitroso compounds catalysed

AICR as probable [1]. Whole grains, by haem iron, and genotoxic com- FUNDAMENTALS
which are rich in dietary fibre and pounds formed during smoking or
phytochemicals, may decrease risk high-temperature cooking of meat ■■ Diets in childhood and
of colorectal cancer by diluting car- [5] (see Chapter 2.8). Additional throughout adult life can
cinogens in the colon, through pro- research is needed on the mecha- influence the carcinogenic
duction of short-chain fatty acids, nisms involved and on mediating process at various stages.
and also by limiting growth of pro- factors (e.g. cooking methods and
inflammatory bacterial species [4]. concomitant dietary components). ■■ Because dietary and other
WCRF/AICR also categorized as exposures many years before
CHAPTER 2.6
SECTION 2
probable the evidence that con- Dietary fat the diagnosis of cancer
sumption of dietary fibre, which is From the 1980s until recently, die- can influence cancer risk,
found in plant foods including whole tary fat intake was widely believed to various types of studies are
grains, fruits and vegetables, nuts, be the most important cause of can- needed, including long-term
and seeds, is associated with lower cers of the breast, colorectum, and epidemiological studies,
risk of colorectal cancer, weight prostate and some other common randomized trials, and shorter-
gain, overweight, and obesity [1]. cancer types in developed coun- term studies with cancer risk
tries. This belief was based largely factors as the outcome.
Red meat and processed meat on correlations between national per ■■ The available dietary assess-
In 2015, IARC classified consump- capita fat intake and rates of these ment methods complemented
tion of processed meat as carcino- cancer types, which were potentially by biomarkers of diet have
genic to humans (Group 1), based confounded by many aspects of diet proven value for the study of
on sufficient evidence in humans for and lifestyle. In subsequent large diet and cancer.
colorectal cancer, and consumption cohort studies with long follow-up,
of red meat as probably carcino- dietary fat has not been associated ■■ Overweight and obesity are
genic to humans (Group 2A), based with risk of these cancer types [6], major risk factors for many
on evidence for colorectal cancer, and in two large randomized trials, cancer types and account for
with strong mechanistic support [5]. women assigned to low-fat diets did much of the impact of diet.
Similarly, WCRF/AICR concluded not have lower risks of breast can-
■■ Studies of specific nutrients
that the evidence was convincing cer or other cancer types [7,8]. Also,
and foods provide important
that consumption of processed meat the type of fat, whether assessed
insights on diet and cancer,
increases risk of colorectal cancer by diet or biomarkers, has not been
but studies of overall dietary
[1], whereas the evidence for con- clearly associated with risk of breast
patterns may provide the most
sumption of unprocessed red meat cancer, but more research is need-
useful guidance for individuals
was classified as probable [1]. ed. Although excess body fatness,
and policies.
Processed meat is defined as most commonly assessed as body
meat that has been transformed mass index, increases risk of many ■■ A dietary pattern that
through salting, smoking, curing, cancer types (see Chapter 2.7), a emphasizes abundant intake
and/or fermentation to enhance higher percentage of energy intake of fruits and vegetables,
flavour or for preservation (exam- from dietary fat is not a major fac- whole grains rather than
ples are frankfurters, bacon, sa- tor in weight control; in randomized refined grains, and low intake
lami, deli meats, and similar prod- trials with balanced intensity of in- of red meat and processed
ucts), whereas red meat refers to tervention, weight loss is somewhat meat, sugar-sweetened
unprocessed mammalian muscle greater in diets with higher fat intake beverages, and salt will
meat (e.g. beef, veal, lamb, pork, and lower carbohydrate intake [9]. reduce risk of cancer, as well
and goat) [5]. For each 50 grams However, higher overall diet qual- as of cardiovascular disease,
of processed meat consumed per ity, including higher intakes of fruits, diabetes, and overall mortality.
day, the risk of colorectal cancer vegetables, nuts, and whole grains
increases by approximately 16%, and lower intakes of red meat and ■■ Although research on the
and for each 100 grams of red meat refined starch, is associated with relationship between diet after
consumed per day, it increases by less overall weight gain [10]. the diagnosis of cancer and
about 12% [1]. For colon cancer, survival is still limited, recent
these estimates are 23% and 22%, Dairy products and calcium evidence supports a benefit
respectively [1]. The effects of intake of dairy prod- on survival from the same
Potential biological mecha- ucts and calcium on cancer risk are dietary pattern recommended
nisms underlying these associa- complex. Intake of dairy products has to lower risk, for at least some
tions include oxidative damage re- been associated with increased risk cancer types.
sulting from endogenous formation of prostate cancer in many studies,
Chapter 2.6 • Diet and nutrition 93

Diet and cancer across the life-cycle
Most cancers, especially adult-on- the extent that dietary factors influ- use lowers risk of sporadic colo-
set cancers, represent a multistage ence cancer risk, similar tempo- rectal cancer, but only after about a
process that occurs over decades. ral associations exist. Importantly, decade from onset of use. Notably,
Several well-established non-die- emerging data suggest that early intake of micronutrients such as
tary risk factors demonstrate spe- dietary exposures, particularly dur- folate and calcium appears to be
cific temporal associations with ing adolescence, may influence related to lower risk of colorectal
cancer. For example, breast tissue risk of breast cancer (Fig. B2.6.1). cancer only after latency periods
may be particularly susceptible Some factors may act on early of more than a decade [1,2]. A ran-
to carcinogenic exposures during stages of carcinogenesis, so a domized trial of multivitamin use
childhood, adolescence, and early time lag (latency period) may be re- with up to 14 years of follow-up did
adult life, as observed in women quired to elicit an effect on cancer. not show a significant reduction in
exposed to ionizing radiation. It For example, from randomized tri- the incidence of colorectal cancer,
is reasonable to anticipate that to als and observational data, aspirin but intriguingly did suggest a pos-
sible decrease in risk after about a
decade of use [3], consistent with
Fig. B2.6.1. Dietary protein sources during adolescence in relation to risk of observational studies.
premenopausal breast cancer. The graph shows the hazard ratios (HR; circles)
and 95% confidence intervals (95% CI; bars) for breast cancer in premenopausal References
women associated with replacement of adolescent intake of one serving per day of
1. Lee JE, Willett WC, Fuchs CS, Smith-
total red meat with other sources of dietary protein.
Warner SA, Wu K, Ma J, et al. (2011).
Folate intake and risk of colorectal
cancer and adenoma: modification by
time. Am J Clin Nutr. 93(4):817–25.
Legumes for total https://doi.org/10.3945/ajcn.110.007781
red meat
PMID:21270374
Nuts for total
red meat 2. Zhang X, Keum N, Wu K, Smith-Warner
SA, Ogino S, Chan AT, et al. (2016).
Poultry for total Calcium intake and colorectal cancer risk:
red meat results from the Nurses’ Health Study and
Health Professionals Follow-up Study. Int
Fish for total
red meat J Cancer. 139(10):2232–42. https://doi.
org/10.1002/ijc.30293 PMID:27466215
Eggs for total
red meat 3. Gaziano JM, Sesso HD, Christen WG,
Bubes V, Smith JP, MacFadyen J, et
Total legumes, nuts,
poultry, and fish al. (2012). Multivitamins in the preven-
for total red meat tion of cancer in men: the Physicians’
Health Study II randomized controlled
trial. JAMA. 308(18):1871–80. https://
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6
d o i . o r g / 10 .10 0 1/ j a m a . 2 0 12 .14 6 41
HR (95% CI)
PMID:23162860
including in a recent meta-analysis, in the intestinal lumen. In addition, Vitamins and minerals
possibly through increases in levels intraluminal calcium binds to the Vitamin D
of insulin-like growth factor 1 [11]. calcium-sensing receptor, a cell
Whether this association is attribu- surface receptor that is expressed The potential role of vitamin D in
table to the calcium in dairy products on colonocytes and increases ex- lowering risk of cancer, particularly
is unclear on the basis of the existing pression of E-cadherin, p21, and of colorectal cancer, is of great in-
evidence. p27, which have anticancer effects. terest. An international consortium
WCRF/AICR categorized as The lower risk of colorectal cancer of 21 prospective cohorts (stud-
probable the evidence that higher appears to be related specifically ies of breast cancer and colorectal
intake of calcium and dairy prod- to calcium intake, because intakes cancer) reported that higher pre-di-
ucts decreases risk of colorectal of supplemental calcium and non- agnostic levels of circulating 25-hy-
cancer [1]. Calcium binds to po- dairy dietary sources of calcium are droxyvitamin D (25(OH)D), the ac-
tentially toxic secondary bile acids also related to lower risk [12]. cepted measure of vitamin D status,
94
were associated with lower risk of those with deficient 25(OH)D con- risk [13]. Mendelian randomization
colorectal cancer [13]. Compared centrations (< 30 nmol/L) had a 31% using four single-nucleotide poly-
with men and women with suffi- higher risk of colorectal cancer, morphisms associated with vitamin
cient 25(OH)D concentrations (50– whereas those with concentrations D to predict a 25 nmol/L increase in
< 62.5 nanomoles per litre [nmol/L]), of 75–100 nmol/L had a 22% lower 25(OH)D concentrations in relation
Cancer survivors
CHAPTER 2.6
SECTION 2
In recent decades, because of in- a lower risk of mortality from out- disease-free survival. Consuming
creases in the size of the popula- comes other than breast cancer, five or more servings per day of
tion, ageing of the population, and such as death from cardiovascular vegetables and fruits and choosing
enhanced use of screening tech- disease [1]. whole grains over refined grains
niques, the number of cancer sur- Emerging evidence suggests were associated with a 35–40%
vivors has skyrocketed. The role of potential benefits for cancer-spe- lower mortality [2].
dietary factors in the prognosis of cific mortality. A recent study of Among men with prostate can-
cancer is just beginning to be stud- stage III colon cancer examined cer, replacing animal fat or car-
ied. The specific role of diet is likely adherence to the American Cancer bohydrates with vegetable fat in
to differ by cancer type. Society guidelines on nutrition and the post-diagnostic period was
For cancer types with high physical activity, which include associated with a reduced risk of
long-term survival rates (e.g. ear- maintaining a healthy body weight, all-cause mortality and possibly
ly-stage colorectal, breast, and being physically active, and eating prostate cancer-specific mortality
prostate cancer), deaths from a diet that includes ample amounts [3]. Fig. B2.6.2 illustrates the as-
other chronic diseases, such as of vegetables, fruits, and whole sociation of higher intake of veg-
cardiovascular diseases, diabetes, grains, in relation to survival over a etable fat (replacing animal fat and
and second cancers, exceed those median follow-up period of 7 years, trans fat) with prostate cancer-spe-
from the cancer itself. Therefore, during which the majority of deaths cific and all-cause mortality among
general dietary guidelines for over- were cancer-related. Compared prostate cancer survivors [3].
all health (including cancer pre- with those who did not adhere to
vention) are likely to be most ben- the guidelines, those who adhered References
eficial for the patient. For example, to the combined guidelines had a 1. George SM, Ballard-Barbash R, Shikany
breast cancer survivors who fol- 42% lower risk of death during the JM, Caan BJ, Freudenheim JL, Kroenke
CH, et al. (2014). Better postdiagnosis
low healthy dietary patterns have study period and a 31% improved
diet quality is associated with reduced
risk of death among postmenopausal
women with invasive breast cancer in
Fig. B2.6.2. Relationship of higher intake of vegetable fat with prostate cancer- the Women’s Health Initiative. Cancer
specific and overall survival among 4577 men with prostate cancer. Epidemiol Biomarkers Prev. 23(4):575–
83. https://doi.org/10.1158/1055-9965.
EPI-13-1162 PMID:24493629
1.2
Prostate-cancer specific mortality
2. Van Blarigan EL, Fuchs CS, Niedzwiecki
P trend=.06
D, Zhang S, Saltz LB, Mayer RJ, et al.
1
(2018). Association of survival with ad-
herence to the American Cancer Society
Relative Risk
0.8 Nutrition and Physical Activity Guidelines

for cancer survivors after colon cancer
diagnosis: the CALGB 89803/Alliance
0.6 Overall mortality Trial. JAMA Oncol. 4(6):783–90. https://
P trend=<.001 doi.org/10.1001/jamaoncol.2018.0126
0.4 PMID:29710284
3. Richman EL, Kenfield SA, Chavarro

0.2 JE, Stampfer MJ, Giovannucci EL,
Willett WC, et al. (2013). Fat intake af-
ter diagnosis and risk of lethal prostate
0 cancer and all-cause mortality. JAMA
Q1 Q2 Q3 Q4 Q5 Intern Med. 173(14):1318–26. https://doi.
Quintiles of Vegetable Fat Intake org/10.1001/jamainternmed.2013.6536
PMID:23752662
1. Richman EL, Kenfield SA, Chavarro JE, Stampfer MJ, Giovannucci EL,
Willett WC, Chan M: Fat intake after diagnosis and risk of lethal
prostate cancer and all-cause mortality Chapter 2.6 • Diet and nutrition 95
JAMA Intern Med 2013, 173:1318-1326.
to colorectal cancer risk was non- enhance carcinogenesis in rapidly high doses have mostly shown no
significant (relative risk, 0.92; 95% growing tumours that are reliant on benefit, and some have shown the
confidence interval, 0.76–1.10) [14] DNA synthesis. potential for harm [1] (see Chapter
but overlapped with (and was con- Epidemiological data have tend- 6.4). In contrast, multivitamin trials
sistent with) estimates from the ed to support that higher folate in- of multiple nutrients at recommend-
consortium (relative risk, 0.87; 95% take is associated with a lower risk ed dietary amounts have not shown
confidence interval, 0.82–0.92 per of colorectal cancer [16]. An analy- harm, and some have shown ben-
25 nmol/L increase). In contrast, sis examining timing of folate intake efit in men [19]. Currently, cancer
preliminary findings from the same in relation to risk found a protective organizations recommend against
pooling project showed no asso- association, but only after a latency taking supplements for cancer pre-
ciation of pre-diagnostic 25(OH) period of at least 12–16 years [16]. vention, and recommend obtaining
D levels with risk of breast cancer Despite proven benefits of folic acid nutrients from food whenever pos-
across a wide range of concentra- supplementation on incidence of sible [1,20].
tions (< 20 nmol/L to > 125 nmol/L) neural tube defects and strokes, fol-
(unpublished data). Although ran- ic acid fortification efforts have been Processed foods
domized trials would be desirable to hindered in some countries because Processing modifies food from
confirm a protective effect of taking of concerns of higher cancer risk. its natural state for safety, conve-
vitamin D for prevention of colorectal Yet, reassuringly, no evidence of an nience, palatability, or taste [21].
cancer, supplementation trials usu- increased risk of colorectal cancer However, the term “processed
ally cannot achieve this wide range or other cancer types was observed foods” reflects a wide range of al-
of 25(OH)D levels, and tend to in- in an analysis of individual partici- terations, from washing, cutting,
clude smaller numbers of cases fol- pant data of 50 000 subjects from all and freezing fresh produce to form-
lowed up for limited time periods. placebo-controlled trials of folic acid ing new products that do not exist
Vitamin D can be obtained by for prevention of cardiovascular disin nature, such as sugar-sweetened
exposure to sunlight or consump- ease or colorectal adenoma, with a beverages, chicken nuggets, and
tion of fatty fish, fortified foods, mean follow-up of 5.5 years [17], or cheese puffs, items termed ultra-
and supplements [15]. However, in a study of time trends and colo- processed foods [22] or highly pro-
excessive exposure to ultraviolet rectal cancer incidence and death cessed foods [23]. Moreover, fast
radiation is a strong risk factor for rates in the USA [18]. foods are readily available conve-
skin cancer (see Chapter 5.8) and nience foods that tend to have a
should therefore be limited. Vitamin Vitamin supplementation high energy density and be seduc-
D intakes above 4000 international Cancer prevention trials of vitamin tively flavoured, affordable, easy
units per day are not recommend- and/or mineral supplementation at to access, aggressively marketed,
ed, because of potential adverse ef-
fects [15]. People at risk of vitamin
D inadequacy include those living Fig. 2.6.1. Women eating together in Chhattisgarh, India. In many countries, the
at high latitudes or in areas without proportion of highly processed foods consumed has risen markedly as large numbers
vitamin D fortification, the elderly, of people move from rural to urban areas, often with a transition from traditional diets
obese individuals, those with dark to global industrial diets.
skin, and those who cover most of
their skin for cultural, religious, or
other reasons.
Folate
Folate, which is found primarily
in plant foods and is added to the
food supply in certain countries as
folic acid, is essential as a carrier
of single-carbon units; as such, it
is critical for DNA methylation and
DNA biosynthesis and repair. It
has been proposed that folate has
a dual role in cancer, particularly
colorectal cancer. Folate deficien-
cy, particularly early in carcinogen-
esis, may increase risk, whereas at
a late stage, excess folate (particu-
larly in the form of folic acid) may
96
and consumed in large portions. Fig. 2.6.2. Traditional diets, such as those in India (left), tend to be rich in whole
Both fast foods and sugar-sweet- grains, fruits, vegetables, and nuts. In contrast, dietary patterns typical of industrialized
ened beverages are considered a countries, particularly in the context of fast foods (right), tend to be high in meat, refined
grains, fried potatoes, and sugar, and low in fruits and vegetables.
cause of weight gain, overweight,
and obesity [1]. Processed meats
and foods preserved by salting (e.g.
pickled vegetables and dried fish)
increase the risk of gastrointestinal
cancers [1,5].
In a large study in France, com-
CHAPTER 2.6
SECTION 2
pared with men and women with
less than 12% of energy intake
from ultra-processed foods, those
with more than 25% of energy in-
take from ultra-processed foods
had a 23% higher risk of any can-
cer, a 23% higher risk of colorec-
tal cancer, and a 38% higher risk
of postmenopausal breast cancer
[22]. Although it is unknown which
aspects may be related to cancer
risk, possible factors include ex-
cess sugar and energy, low dietary studies [25,26]. Such diets tend to was due to the Mediterranean diet,
fibre and micronutrients, added pre- be rich in whole grains, fruits, veg- the olive oil intervention, or chance,
servatives and other ingredients, etables, nuts, and unsaturated fats given the small number of breast
carcinogens formed during pro- (e.g. monounsaturated and/or poly- cancer cases (n = 35).
cessing, and/or lifestyle correlates unsaturated fat) and contain lower
of highly processed foods, such as amounts of processed meat, red
sedentary behaviours. Coffee
meat, sugar, and saturated and/
Over the past century, the global or trans fats [25,26]. In contrast, Studies conducted in the 1970s
food system has shifted dramatical- a dietary pattern typical of indus- concluded that coffee consumption
ly from that of consumption of local trialized countries, high in meat, may increase risk of cancer, par-
staple foods and home cooking to refined grains, fried potatoes, and ticularly of bladder cancer and pan-
increasing intake of ready-to-con- sugar and low in fruits and vegeta- creatic cancer. It is now thought that
sume, processed, and packaged bles, is associated with increased these early retrospective case–
foods, available globally. In 2012, risk of colorectal cancer [27]. The control studies had been largely
highly processed foods comprised Alternate Healthy Eating Index rep- confounded by tobacco use among
about 60% of per capita daily ener- resents an overall healthy dietary coffee drinkers or other sources of
gy consumption in North America, pattern (see “Distribution of global bias. More recent research sug-
and this percentage has remained diet quality”). gests that coffee consumption may
stable since 2000 [23], whereas In the Prevención con Dieta lower the risk of liver cancer and
the proportion of food intake made Mediterránea (PREDIMED) trial in endometrial cancer [1], and possi-
up of highly processed foods has Spain [28], women were assigned bly other cancer types [29,30].
risen markedly since 2000 in sev- to follow a Mediterranean diet sup- In a pooling project of nine co-
eral countries that are undergoing plemented with either extra virgin horts in the USA including more
a transition from traditional diets to olive oil or nuts, or were advised than 1 million people, compared
global industrial diets [23,24]. to follow a low-fat diet (control with not drinking coffee, drinking
group). Compared with controls, a 3 cups of coffee per day was as-
Dietary patterns 68% lower risk of invasive breast sociated with a 27% lower risk
The study of overall dietary pat- cancer was seen in women on the of hepatocellular carcinoma [31].
terns and cancer risk has grown Mediterranean diet supplemented Biologically active compounds in
markedly in recent decades. Diet with olive oil, and a non-significant coffee, including chlorogenic acid,
scores reflecting greater concor- 41% lower risk was seen in the kahweol, and N-methylpyridinium,
dance with hypothesized healthy group on the Mediterranean diet have been found to induce apopto-
eating patterns, and with traditional supplemented with nuts [28]. It is sis, improve insulin sensitivity, and
and regional dietary patterns, are unclear whether the lower risk of inhibit inflammation and angiogen-
associated with lower cancer risk breast cancer among women in the esis, among other potential antican-
and mortality in many prospective arm with olive oil supplementation cer mechanisms [32].

Distribution of global diet quality
Diet has many components that plant-sourced proteins, fruits, and and some parts of Africa and
ultimately need to be combined vegetables. Europe in part reflect low intakes
in an overall eating pattern. The high scores in some of fruits and vegetables and high in-
Fig. B2.6.3 shows the global dis- Mediterranean countries are con- takes of red meat, processed meat,
tribution of scores in 2017 for the sistent with the well-documented sugars, and refined grains; in some
Alternate Healthy Eating Index health benefits of the traditional of these countries, intake of indus-
[1], a measure of diet quality that diets of this region, although the trial trans fat remains high.
has predicted lower risks of weight region has generally experienced Although scores vary widely
gain and major chronic disease in declines in dietary quality over across the globe, even those
many populations. Higher scores time. The relatively high scores of countries with the highest scores
are given to lower amounts of red countries in some parts of Africa re-
(60–65) have considerable room
meat, sugar-sweetened bever- flect the positive aspects of many
for improvement, because the ideal
ages, salt, and trans fat, and traditional diets and are consistent
diet would score 100. Many coun-
higher amounts of fruits, vegeta- with low rates of chronic disease.
tries lacked current representative
bles, whole grains, nuts and le- However, in many of these same
dietary surveys, requiring imputa-
gumes, omega-3 fatty acids, and areas childhood mortality remains
tion of national food intakes and
omega-6 polyunsaturated fatty ac- high, in part because of inadequate
ids (alcohol is not included). food availability and unmet nutrition emphasizing the need for improved
Countries in the Mediterranean needs of growing children. These dietary surveillance.
region, South-East Asia (e.g. Viet countries are undergoing rapid eco-
Reference
Nam), the Caribbean, and some nomic and nutrition transitions, and
parts of Africa tend to have rela- it will be important to retain healthful 1. Wang DD, Li Y, Afshin A, Springmann M,
Mozaffarian D, Stampfer MJ, et al. (2019).
tively high scores, as do Brazil, aspects of traditional diets, because Global improvement in dietary quality
the Islamic Republic of Iran, and these are often lost with growing af- could lead to substantial reduction in
Japan. These scores reflect rela- fluence and the industrialization of premature death. J Nutr. 149(6):1065–74.
tively low consumption of red meat, food systems. The low scores for h t t p s : / / d o i . o r g / 10 .10 9 3 / j n / n x z 0 10
PMID:31049577
sugar-sweetened beverages, and Afghanistan, Argentina, Finland,
trans fat, and higher intakes of Mongolia, Pakistan, Turkmenistan,
Fig. B2.6.3. Geographical distribution of scores for the Alternate Healthy Eating Index (AHEI) in men and women aged 25
years and older in 190 countries or territories in 2017. The AHEI scores range from 0 (worst) to 100 (best). White areas
indicate that dietary data were not available.
98
Mechanisms the literature for each risk factor and cluded that 9.2% of cancer cases
cancer, and estimating a population are attributable to diet [34]. The
Many pathways are thought to un-
prevalence of each risk factor from higher estimate is mostly a result
derlie a role of diet in carcinogen-
the available data. Then, the percent- of a greater weight given to intake
esis, including those involved in
age of cases of the cancer that are of fruits and vegetables, for which
cell-cycle regulation, growth factors
accounted for by that factor can be the estimates have trended down-
(e.g. insulin and insulin-like growth
estimated. As science evolves and wards in recent years.
factors), inflammation, immunity,
dietary exposures change, these fig- The range of estimates for
and angiogenesis. Potential, but
as yet unproven, effects of the mi- ures will be updated. population attributable fraction is
crobiome are currently a topic of For example, a recent analysis approximately 5–10%. These es-
CHAPTER 2.6
SECTION 2
great interest [33]. Contemporary from the American Cancer Society timates do not account for syner-
research on diet and cancer, using relied on findings from WCRF/ gies among dietary factors, or for
tumour molecular pathology and AICR to estimate the total numbers the important indirect effect of diet
–omics research, including genet- of cancer cases and deaths attribu on obesity. Also, these estimates
ics (see Chapter 3.2), metabo- table to diet (independent of obesi- do not account for errors in mea-
lomics (see Chapter 3.7), and the ty) in the USA [34]. The risk factors suring diet or the potential effect of
microbiome (see Chapter 3.10), will identified included consumption diet during childhood or early adult
continue to elucidate the role of diet of red meat, consumption of pro- life. Continued research on dietary
in cancer etiology. cessed meat, and low intake of assessment measures, uniform as-
fruits and vegetables, dietary fibre, sessment of dietary patterns, and
and dietary calcium. These factors contemporary dietary exposures,
Population attributable were estimated to account for ap- as well as large harmonized pooled
fractions proximately 5.1% of cancer deaths analyses, randomized trials (where
Estimating the population attributable in the USA. The largest proportion feasible), and research across the
fraction for diet and cancer involves of these cancer deaths was from lifespan, will continue to contribute
identifying relevant dietary factors, colorectal cancer. A previous anal- information on the impact of diet on
deriving a relative risk estimate from ysis for the United Kingdom con- cancer risk.
References
1. WCRF/AICR (2018). Diet, nutrition, physi- 4. Singh RK, Chang HW, Yan D, Lee KM, 7. Martin LJ, Li Q, Melnichouk O, Greenberg
cal activity and cancer: a global perspec- Ucmak D, Wong K, et al. (2017). Influence C, Minkin S, Hislop G, et al. (2011). A ran-
tive. Continuous Update Project Expert of diet on the gut microbiome and impli- domized trial of dietary intervention for
Report 2018. World Cancer Research cations for human health. J Transl Med. breast cancer prevention. Cancer Res.
Fund/American Institute for Cancer 15(1):73. https://doi.org/10.1186/s12967- 71(1):123–33. https://doi.org/10.1158/0008-
Research. Available from: https://www. 017-1175-y PMID:28388917 5472.CAN-10-1436 PMID:21199800
wcrf.org/dietandcancer.
5. Bouvard V, Loomis D, Guyton KZ, Grosse 8. Prentice RL, Caan B, Chlebowski RT,
2. Eliassen AH, Liao X, Rosner B, Tamimi Y, Ghissassi FE, Benbrahim-Tallaa L, et Patterson R, Kuller LH, Ockene JK, et al.
RM, Tworoger SS, Hankinson SE (2015). al.; International Agency for Research (2006). Low-fat dietary pattern and risk of
Plasma carotenoids and risk of breast can- on Cancer Monograph Working Group invasive breast cancer: the Women’s Health
cer over 20 y of follow-up. Am J Clin Nutr. (2015). Carcinogenicity of consumption of Initiative Randomized Controlled Dietary
101(6):1197–205. https://doi.org/10.3945/ red and processed meat. Lancet Oncol. Modification Trial. JAMA. 295(6):629–42.
ajcn.114.105080 PMID:25877493 16(16):1599–600. https://doi.org/10.1016/ https://doi.org/10.1001/jama.295.6.629
S1470-2045(15)00444-1 PMID:26514947 PMID:16467232
3. Bakker MF, Peeters PH, Klaasen VM,
Bueno-de-Mesquita HB, Jansen EH, Ros 6. Smith-Warner SA, Spiegelman D, Adami 9. Tobias DK, Chen M, Manson JE, Ludwig
MM, et al. (2016). Plasma carotenoids, HO, Beeson WL, van den Brandt PA, DS, Willett W, Hu FB (2015). Effect of
vitamin C, tocopherols, and retinol and Folsom AR, et al. (2001). Types of dietary low-fat diet interventions versus other
the risk of breast cancer in the European fat and breast cancer: a pooled analysis diet interventions on long-term weight
Prospective Investigation into Cancer of cohort studies. Int J Cancer. 92(5): change in adults: a systematic review
and Nutrition cohort. Am J Clin Nutr. 767–74. https://doi.org/10.1002/1097-0215 and meta-analysis. Lancet Diabetes
103(2):454–64. https://doi.org/10.3945/ (20010601)92:5<767::AID-IJC1247>3.0. Endocrinol. 3(12):968–79. https://doi.
ajcn.114.101659 PMID:26791185 CO;2-0 PMID:11340585 org /10.1016/S2 213 - 8 5 87(15) 0 03 67- 8
PMID:26527511

10. Mozaffarian D, Hao T, Rimm EB, Willett WC, 19. Gaziano JM, Sesso HD, Christen WG, 27. Grosso G, Bella F, Godos J, Sciacca S, Del
Hu FB (2011). Changes in diet and lifestyle Bubes V, Smith JP, MacFadyen J, et Rio D, Ray S, et al. (2017). Possible role of
and long-term weight gain in women and al. (2012). Multivitamins in the preven- diet in cancer: systematic review and mul-
men. N Engl J Med. 364(25):2392–404. tion of cancer in men: the Physicians’ tiple meta-analyses of dietary patterns,
https://doi.org/10.1056/NEJMoa1014296 Health Study II randomized controlled lifestyle factors, and cancer risk. Nutr Rev.
PMID:21696306 trial. JAMA. 308(18):1871–80. https:// 75(6):405–19. https://doi.org/10.1093/nutrit/
d o i . o r g / 1 0 .1 0 0 1 / j a m a . 2 0 1 2 .1 4 6 41 nux012 PMID:28969358
11. Aune D, Navarro Rosenblatt DA, Chan PMID:23162860
DS, Vieira AR, Vieira R, Greenwood 28. Toledo E, Salas-Salvadó J, Donat-Vargas
DC, et al. (2015). Dairy products, calci- 20. Kushi LH, Doyle C, McCullough M, Rock C, Buil-Cosiales P, Estruch R, Ros E, et
um, and prostate cancer risk: a system- CL, Demark-Wahnefried W, Bandera al. (2015). Mediterranean diet and inva-
atic review and meta-analysis of cohort EV, et al.; American Cancer Society sive breast cancer risk among women at
studies. Am J Clin Nutr. 101(1):87–117. 2010 Nutrition and Physical Activity high cardiovascular risk in the PREDIMED
https://doi.org/10.3945/ajcn.113.067157 Guidelines Advisory Committee (2012). trial: a randomized clinical trial. JAMA
PMID:25527754 American Cancer Society guidelines on Intern Med. 175(11):1752–60. https://doi.
nutrition and physical activity for cancer org/10.1001/jamainternmed.2015.4838
12. Keum N, Aune D, Greenwood DC, Ju prevention: reducing the risk of cancer PMID:26365989
W, Giovannucci EL (2014). Calcium in- with healthy food choices and physical
take and colorectal cancer risk: dose- activity. CA Cancer J Clin. 62(1):30–67. 29. Grosso G, Godos J, Galvano F,
response meta-analysis of prospective h t t p s: //d o i . o r g /10 . 3 3 2 2 /c a a c . 2 014 0 Giovannucci EL (2017). Coffee, caffeine,
observational studies. Int J Cancer. PMID:22237782 and health outcomes: an umbrella review.
135(8):1940–8. https://doi.org/10.1002/ Annu Rev Nutr. 37(1):131–56. https://doi.
ijc.28840 PMID:24623471 21. Adams J, White M (2015). Characterisation org/10.1146/annurev-nutr-071816-064941
of UK diets according to degree of food PMID:28826374
13. McCullough ML, Zoltick ES, Weinstein processing and associations with socio-
SJ, Fedirko V, Wang M, Cook NR, et al. demographics and obesity: cross-sectional 30. Poole R, Kennedy OJ, Roderick P,
(2019). Circulating vitamin D and colorec- analysis of UK National Diet and Nutrition Fallowfield JA, Hayes PC, Parkes J (2017).
tal cancer risk: an international pooling Survey (2008-12). Int J Behav Nutr Phys Coffee consumption and health: umbrella
project of 17 cohorts. J Natl Cancer Inst. Act. 12(1):160. https://doi.org/10.1186/ review of meta-analyses of multiple health
111(2):158–69. https://doi.org/10.1093/ s12966-015-0317-y PMID:26684833 outcomes. BMJ. 359:j5024. https://doi.
jnci/djy087 PMID:29912394 org/10.1136/bmj.j5024 PMID:29167102
22. Fiolet T, Srour B, Sellem L, Kesse-
14. Dimitrakopoulou VI, Tsilidis KK, Haycock Guyot E, Allès B, Méjean C, et al. 31. Petrick JL, Freedman ND, Graubard BI,
PC, Dimou NL, Al-Dabhani K, Martin RM, (2018). Consumption of ultra-processed Sahasrabuddhe VV, Lai GY, Alavanja
et al.; GECCO Consortium; PRACTICAL foods and cancer risk: results from MC, et al. (2015). Coffee consumption
Consortium; GAME-ON Network NutriNet-Santé prospective cohort. BMJ. and risk of hepatocellular carcinoma and
(CORECT, DRIVE, ELLIPSE, FOCI- 360:k322. https://doi.org/10.1136/bmj. intrahepatic cholangiocarcinoma by sex:
OCAC, TRICL-ILCCO) (2017). Circulating k322 PMID:29444771 the liver cancer pooling project. Cancer
vitamin D concentration and risk of seven Epidemiol Biomarkers Prev. 24(9):1398–
cancers: Mendelian randomisation study. 23. Popkin BM (2017). Relationship between 406. https://doi.org/10.1158/1055-9965.
BMJ. 359:j4761. https://doi.org/10.1136/ shifts in food system dynamics and acceler- EPI-15-0137 PMID:26126626
bmj.j4761 PMID:29089348 ation of the global nutrition transition. Nutr
Rev. 75(2):73–82. https://doi.org/10.1093/ 32. Bøhn SK, Blomhoff R, Paur I (2014). Coffee
15. Institute of Medicine (US) Committee to nutrit/nuw064 PMID:28395033 and cancer risk, epidemiological evidence,
Review Dietary Reference Intakes for and molecular mechanisms. Mol Nutr Food
Vitamin D and Calcium (2011). Dietary 24. PAHO (2016). Pan American Health Res. 58(5):915–30. https://doi.org/10.1002/
reference intakes for calcium and vita- Organization Nutrient Profile Model. mnfr.201300526 PMID:24668519
min D. Washington (DC), USA: National Washington (DC), USA: PAHO.
Academies Press. PMID:21796828 Available from: http://iris.paho.org/xmlui/ 33. Arkan MC (2017). The intricate connection
handle/123456789/18621. between diet, microbiota, and cancer: a
16. Lee JE, Willett WC, Fuchs CS, Smith- jigsaw puzzle. Semin Immunol. 32:35–42.
Warner SA, Wu K, Ma J, et al. (2011). Folate 25. Schwingshackl L, Schwedhelm C, Galbete https://doi.org/10.1016/j.smim.2017.08.009
intake and risk of colorectal cancer and C, Hoffmann G (2017). Adherence to PMID:28870704
adenoma: modification by time. Am J Clin Mediterranean diet and risk of cancer: an
Nutr. 93(4):817–25. https://doi.org/10.3945/ updated systematic review and meta-anal- 34. Islami F, Goding Sauer A, Miller KD,
ajcn.110.007781 PMID:21270374 ysis. Nutrients. 9(10):E1063. https://doi. Siegel RL, Fedewa SA, Jacobs EJ, et al.
org/10.3390/nu9101063 PMID:28954418 (2018). Proportion and number of cancer
17. Vollset SE, Clarke R, Lewington S, Ebbing cases and deaths attributable to poten-
M, Halsey J, Lonn E, et al.; B-Vitamin 26. Schwingshackl L, Bogensberger B, tially modifiable risk factors in the United
Treatment Trialists’ Collaboration (2013). Hoffmann G (2018). Diet quality as as- States. CA Cancer J Clin. 68(1):31–54.
Effects of folic acid supplementation on sessed by the Healthy Eating Index, h t t p s: //d o i . o r g /10 . 3 3 2 2 /c a a c . 214 4 0
overall and site-specific cancer incidence Alternate Healthy Eating Index, Dietary PMID:29160902
during the randomised trials: meta- Approaches to Stop Hypertension score,
analyses of data on 50,000 individuals. and health outcomes: an updated systemat-
Lancet. 381(9871):1029–36. https://doi. ic review and meta-analysis of cohort stud-
org /10.1016/S 014 0 - 673 6 (12) 620 01-7 ies. J Acad Nutr Diet. 118(1):74–100.e11.
PMID:23352552 https://doi.org/10.1016/j.jand.2017.08.024
PMID:29111090
18. Keum N, Giovannucci EL (2014). Folic
acid fortification and colorectal can-
cer risk. Am J Prev Med. 46(3 Suppl 1):
S65–72. https://doi.org/10.1016/j.amepre.
2013.10.025 PMID:24512932
100
2.7
Physical activity, sedentary behaviour,

and obesity
Established and emerging modifiable
risk factors
SECTION 2
CHAPTER 2.7
Christine M. Friedenreich Steven C. Moore (reviewer)
Michael Leitzmann Leandro Fórnias Machado de Rezende (reviewer)
40% for all cancers associated tematic reviews performed in 2008–

SUMMARY with these risk factors. 2017, to assess the strength of the
evidence for an etiological role for
●● Strong epidemiological evidence cancer risk [1]. The World Cancer
exists that being physically ac- Three main modifiable factors have Research Fund/American Institute
tive reduces the risk of cancers emerged in the past 30–40 years of Cancer Research (WCRF/AICR)
of the bladder, breast, colon, en- that are associated with an in- 2018 Expert Report also provided
dometrium, kidney, oesophagus, creased risk of cancer at several an expert synthesis of the evidence
and stomach. sites: physical inactivity, sedentary [2]. These recent reports have con-
●● Emerging evidence suggests behaviour, and overweight or obesi- cluded that there is strong evidence
that sedentary behaviour is as- ty. This chapter reviews the observa- for an etiological role of physical ac-
tional epidemiological evidence that tivity associated with the incidence
sociated with an increased risk
has been synthesized in systematic of several cancer types. In addition,
of cancers of the breast, colon,
reviews and meta-analyses, and a pooling project coordinated by
endometrium, and lung.
highlights the strength of the associ- the United States National Cancer
●● Strong evidence exists for an as- ations, evidence for dose–response Institute examined these associa-
sociation between obesity and relationships, and the biological tions for 26 cancer sites with data
increased risk of cancers of the plausibility of these associations. In from more than 1 million study par-
postmenopausal breast, colo- addition, the prevalence of these ex- ticipants [3].
rectum, endometrium, kidney, posures worldwide is discussed, as From these two main reviews
liver, oesophagus, and pancre- well as the population attributable and this large pooling project, the
as, and moderate evidence ex- fractions that have been estimated current state of the evidence is that
ists for an association with can- for these exposures. The efficacy physical activity is associated with
cers of the gall bladder, mouth, of programmes to improve physical a reduced risk of 13 cancer types.
pharynx, larynx, ovary, prostate activity, decrease sedentary time, The PAGAC report provided the
(advanced), and stomach. and control obesity that have been most recent and comprehensive
evaluated are highlighted. review of the evidence on the as-
●● Several common biological mech-
anisms are likely to be involved sociation between physical activ-
in the association between
Physical activity ity and cancer as well as a stan-
More than 450 studies have been dardized evidence grading system.
physical activity, sedentary be-
conducted that have examined some Based on the PAGAC review, there
haviour, and obesity and can-
aspect of physical activity and its reis strong evidence that physical ac-
cer risk, including an effect on
lationship to cancer risk, and dozens tivity reduces the risk for cancers of
endogenous sex and metabolic
of meta-analyses and systematic the bladder, breast, colon, endome-
hormones, insulin resistance,
reviews have been published that trium, kidney, and gastric cardia and
and chronic inflammation.
have examined the associations for oesophageal adenocarcinoma.
●● The population attributable frac- for specific cancer sites. Most re- There is also moderate evidence for
tions associated with physical cently, the scientific report of the an association of physical activity
inactivity, sedentary behaviour, 2018 Physical Activity Guidelines with decreased risk of lung cancer,
and obesity are estimated to Advisory Committee (PAGAC) re- although confounding by smoking
range, collectively, from 20% to viewed 45 meta-analyses and sys- remains a concern for this cancer
Chapter 2.7 • Physical activity, sedentary behaviour, and obesity 101

site (see Chapter 5.1). The evidence women for cancers of the colon
is classified as limited for a protec- and kidney, and there is limited
tive effect of physical activity against evidence that effect modification FUNDAMENTALS
cancers of the ovary, pancreas, by sex may exist for other cancers,
■■ Research on the association
prostate, and mouth, pharynx, and such as those of the bladder, gas-
larynx. There is limited evidence of tric cardia, lung, oesophagus, and between physical activity and
no effect of physical activity on risk pancreas. There is insufficient evi- cancer risk began to emerge
of cancers of the thyroid and rectum. dence to determine whether the as- in the mid to late 1980s; early
The magnitude of the risk re- sociation between physical activity studies focused on athletes
duction is approximately 10–20% and cancer incidence varies by age and their risk of cancer over a
for most of these cancer sites, with or socioeconomic status, and some lifetime, as determined through
stronger reductions of about 25% limited information suggests that long-term follow-up.
for lung cancer, when the highest the benefits of physical activity ap-
■■ During the past 30–40 years,
versus the lowest levels of physi- pear to be equal for all racial and
cal activity are compared. There is ethnic groups. more than 450 observational
evidence for a dose–response rela- There are several hypothesized epidemiological studies have
tionship between increasing levels biological mechanisms involved in been published that have
of physical activity and decreasing the association between physical examined some aspect of
cancer risk. However, the methods activity and cancer risk, including an physical activity – however
used to measure and categorize effect on adiposity, endogenous sex that is defined – and the risk of
physical activity have been incon- and metabolic hormones, chronic developing cancer.
sistent across studies. Therefore, it inflammation, oxidative stress, and
is currently impossible to determine ■■ In the past 10–15 years,
genomic instability [4]. Randomized
the exact levels of physical activity controlled trials have been inves- there has been a focus on
that are needed to provide benefits tigating how these mechanisms how sedentary behaviour,
in reduced cancer incidence for any are changed with year-long exer- independent of physical activity,
particular cancer site. cise interventions and have dem- is associated with cancer risk,
Currently, limited information is onstrated direct effects on several and evidence is now emerging
available on how the association mechanisms (see “Randomized exon these associations for a few
between physical activity and can- ercise intervention trials of biologi- cancer sites.
cer varies by cancer subtype. There cal mechanisms between physical
is evidence that physical activity activity and cancer risk”). Not only ■■ Some randomized controlled
is equally beneficial for men and were these trials able to demonstrate trials of exercise interventions
have been conducted to
investigate how physical
Fig. 2.7.1. Potential biological mechanisms linking increased physical activity and activity influences several
decreased sedentary behaviour to reduced risk of cancer. (Inter-relationships between hypothesized biological
mechanisms are not shown.) CRP, C-reactive protein; IGF-1, insulin-like growth factor mechanisms involved in
1; IGFBP, insulin-like growth factor-binding protein; IL-6, interleukin 6; SHBG, sex
the association between
hormone-binding globulin; TNF-α, tumour necrosis factor α.
physical activity and cancer
risk, and these studies are
demonstrating an impact on
adiposity, endogenous sex
hormones, metabolic factors,
insulin resistance, and chronic
inflammation.
■■ Research on the association
between obesity and cancer
risk has accumulated over
the past 40 years, and there
is now strong evidence for an
association between obesity
and increased risk for several
cancer sites.
102
Randomized exercise interventions trials of biological mechanisms between

physical activity and cancer risk
Randomized controlled trials of represents the widely recommend- Taken together, these trials
exercise interventions [1–3] have ed level of physical activity for gen- have provided some evidence that
been conducted using healthy eral health that is often prescribed regular aerobic activity at a mod-
populations to address the ques- by public health agencies world- erate to vigorous intensity level for
tion of how aerobic exercise influ- wide. The high-volume arm was at least 150 minutes per week has
ences biomarkers hypothesized chosen because larger volumes of beneficial effects on biomarkers
SECTION 2
CHAPTER 2.7
to be associated with cancer risk, activity may provide more benefit associated with cancer risk.
with the main focus on breast can- for cancer prevention.
cer and colon cancer. These year- In BETA, participants in the References
long randomized controlled trials high-volume arm had statistically
1. Friedenreich CM, Woolcott CG,
have demonstrated that increased significantly greater decreases McTiernan A, Terry T, Brant R, Ballard-
levels of aerobic activity do de- in adiposity compared with the Barbash R, et al. (2011). Adiposity
crease the levels of endogenous moderate-volume arm for all mea- changes after a 1-year aerobic exercise
sex hormones, insulin, glucose, sures of body fat that were taken intervention among postmenopaus-
al women: a randomized controlled
insulin resistance as assessed by [4]. For the remaining biomark-
trial. Int J Obes (Lond). 35(3):427–35.
homeostatic model assessment ers, there were similar decreases ht tps://doi.org/10.1038/ijo. 2010.147
of insulin resistance (HOMA-IR), in both arms of the trial, but there PMID:20820172
inflammatory markers, and sev- was evidence for greater de-
2. Irwin ML, Yasui Y, Ulrich CM, Bowen
eral measures of body fat. The creases in insulin resistance and D, Rudolph RE, Schwartz RS, et al.
exercise interventions have used in inflammatory markers for those (2003). Effect of exercise on total and
varying volumes of aerobic activ- participants who had the highest intra-abdominal body fat in postmen-
ity, ranging from 150 minutes to exercise adherence and spent a opausal women: a randomized con-
300 minutes per week of a combi- greater amount of their prescribed trolled trial. JAMA. 289(3):323–30.
https://doi.org/10.1001/jama.289.3.323
nation of supervised and unsuper- exercise in their heart rate zone,
PMID:12525233
vised activity. i.e. exercising at a higher intensity.
The most recent of these tri- These studies have focused 3. Monninkhof EM, Velthuis MJ, Peeters
als was the Breast Cancer and PH, Twisk JW, Schuit AJ (2009). Effect
on aerobic exercise, and there re-
of exercise on postmenopausal sex hor-
Exercise Trial in Alberta (BETA) mains a need to understand how mone levels and role of body fat: a ran-
[4], which specifically examined resistance exercise influences domized controlled trial. J Clin Oncol.
the question of the optimal dose of these biomarkers. Additional po- 27(27):4492–9. https://doi.org/10.1200/
activity needed for the most ben- tential pathways have been ex- JCO.2008.19.7459 PMID:19687339
eficial effect on these biomarkers. amined, with a focus on chronic 4. Friedenreich CM, Neilson HK, O’Reilly
In BETA, 400 healthy postmeno- stress, oxidative stress, genomic R, Duha A, Yasui Y, Morielli AR, et al.
pausal women were randomized instability as assessed by DNA (2015). Effects of a high vs moderate
to a year-long intervention of either methylation, and leukocyte telo- volume of aerobic exercise on adiposity
outcomes in postmenopausal women: a
150 minutes per week (moderate mere length. The evidence for a
randomized clinical trial. JAMA Oncol.
volume) or 300 minutes per week direct effect of aerobic exercise 1(6):766–76. https://doi.org/10.1001/
(high volume). The moderate-vol- on these additional pathways has jamaoncol.2015.2239 PMID:26181634
ume arm was selected because it been inconsistent to date.
the efficacy of the exercise interven- monly expressed in metabolic equiva- mended that this target for weekly
tions in increasing physical activity lents (METs); 1 MET is the amount of physical activity should be the
levels in the study participants, but energy expended at rest. Moderate minimum that is done for cancer
they also clearly found that these activity is defined as any activity of prevention [2]. Physical inactivity
interventions resulted in beneficial 3–6 METs, and vigorous activity is is defined as an activity level that
changes in these biological mecha- 6 or more METs. The most recent is insufficient to meet the current
nisms. Hence, biological plausibility WHO public health recommenda- recommendations [5].
for the association between physical tions for physical activity are to Globally, the prevalence of
activity and cancer risk exists from achieve 150 minutes per week of physical inactivity is high, with an
efficacy trials. moderate activity or 75 minutes estimated median prevalence of
In epidemiological studies, the per week of vigorous activity. The 24% (range, 4.1–65.0%) world-
intensity of physical activity is com- WCRF/AICR committee recom- wide [6]. Estimates of the median

Fig. 2.7.2. Sedentary behaviour in the household includes sitting while watching television. population attributable fractions for
physical inactivity for cancers of the
breast, colon, and endometrium
range from 12% to 19% worldwide,
and the highest estimates are more
than 25% [7]. Hence, the global
burden of cancer that could be pre-
vented by regular physical activity
is considerable (see Chapter 6.2).
Sedentary behaviour
Sedentary behaviour is defined as
“any waking behaviour character-
ized by an energy expenditure less
than or equal to 1.5 METs while in
a sitting or reclining posture” [8].
Sedentary behaviour comprises sit-
ting in the workplace, during leisure
time, while commuting, and in the
Improvements in measurement of sedentary behaviour to improve understand-

ing of the association between sedentary behaviour and cancer
Prolonged sedentary behaviour Further progress in measurement behaviour in isolation, but activity
plays a potentially important role technology has recently enabled behaviours do not occur indepen-
in cancer etiology. Despite con- advanced activity monitoring that dently of one another. Rather, time
siderable research progress in this distinguishes between sitting, ly- spent in one behaviour ultimately
area over the past years, the epi- ing down, and standing. Integrating replaces time spent in another be-
demiology of sedentary behaviour this new generation of thigh-worn haviour. Therefore, sophisticated
and cancer is still in its infancy. sensors or combinations of sensor statistical approaches such as
Numerous uncertainties and limi- placements on the thigh and the isotemporal substitution modelling
tations persist [1], in part because hip or lower back into new and con- and compositional data analysis
of concerns that self-reported tinuing prospective epidemiological are required to appropriately handle
questionnaires and traditional ac- studies represents a major step for- the interdependent elements of dai-
celerometers measure sedentary ward in validly quantifying the vol- ly energy expenditure within the 24-
behaviour with too much error. ume and patterns of accumulation hour continuum to identify optimal
Early investigations used ques- of daily sedentary time. combinations of sitting, standing,
tionnaires, which proved useful for Ideally, such technology should light activity, moderate to vigorous
large-scale observational studies be combined with self-reports to activity, and sleep. The joint capac-
but may have introduced a cer- gather relevant information about ity of these approaches will help to
tain degree of measurement error the social and environmental con- further develop the epidemiological
due to reporting bias. Subsequent texts in which sedentary behaviour evidence base that is needed to ad-
investigations helped overcome takes place (e.g. location and pur- vance what is known about seden-
some of the limitations of self- pose). In addition, measurements tary behaviour and cancer.
report assessments by deploying should not be limited to a single time
objective measures of sedentary point – at study baseline – but should Reference
behaviour in epidemiological stud- be performed repeatedly during fol- 1. Schmid D, Jochem C, Leitzmann MF
ies. However, those studies mainly low-up to capture information about (2018). Limitations in sedentary behav-
used waist-worn accelerometers, changes in sedentary behaviour iour research and future research needs.
which are unable to determine pos- over time and to identify potential In: Leitzmann MF, Jochem C, Schmid D,
ture and thus produce output that time-sensitive effects of sedentary editors. Sedentary behaviour epidemi-
ology. Springer Series on Epidemiology
does not represent sedentary be- behaviour on cancer incidence. and Public Health. Springer International
haviour itself but rather lack of am- Also, most of the available Publishing; pp. 629–38. https://doi.org/
bulatory movement. studies have examined sedentary 10.1007/978-3-319-61552-3_28
104
household. Examples of sedentary analysis reported that sedentary duced partly inconsistent findings.
behaviour include computer use, behaviour is related to an increased Nevertheless, several experimen-
television viewing, reading, and sit- risk of breast cancer [13]. There is tal studies have demonstrated that
ting while commuting by car, bus, insufficient evidence to determine interrupting prolonged bouts of sit-
train, and airplane. whether the relationship of seden- ting by standing or stepping has
Data on sedentary behaviour tary behaviour to cancer risk varies a beneficial impact on circulating
in relation to risk of cancer are according to age, sex, race and eth- levels of insulin and glucose [15],
far less abundant than those on nicity, or other factors. supporting a link between seden-
physical activity and cancer risk. Very little is known about wheth- tary behaviour and type 2 diabe-
However, a growing body of evi- er prolonged sedentary behaviour tes, which is itself a risk factor for
SECTION 2
CHAPTER 2.7
dence demonstrates that prolonged affects biological pathways of can- numerous cancer types.
sedentary behaviour is associated cer risk. One possible etiological Quantifying the global burden
with increased cancer risk, inde- mechanism involves obesity, which of cancer due to sedentary behav-
pendent of physical activity level. may contribute to cancer risk direct- iour is challenging, because global
Specifically, a meta-analysis of 14 ly, or indirectly through enhanced surveillance programmes for sed-
observational studies showed that circulating concentrations of sex entary behaviour have not yet been
high versus low levels of time spent and metabolic hormones and of adi- established. However, a study that
sitting are related to a 24% higher pokines, and chronic inflammation estimated the population attributa-
risk of cancer incidence after ad- (see Chapter 3.5). Time spent in sed- ble fractions for sitting-related over-
justment for physical activity [9]. entary behaviour typically replaces all mortality from all causes (not
Another meta-analysis of prospec- time spent in light-intensity activ- cancer mortality specifically) for 54
tive studies reported a 2% increase ity, which is associated with greater countries found that time spent sit-
in risk of cancer mortality for each energy expenditure. However, data ting accounted for 4% of mortality
additional hour per day of television showing that sedentary behaviour from all causes [16].
viewing when adjusted for physical leads to weight gain are inconsis- The volume and patterns of ac-
activity [10]. A recent meta-analysis tent, and the relationship of seden- cumulation of daily sedentary be-
showed that physical activity modi- tary behaviour to weight gain is po- haviour related to the risk of cancer
fies the relationship of sedentary tentially bidirectional [14]. have not been determined. In ad-
behaviour to cancer mortality: in- Studies examining prolonged dition, it remains unclear whether
creased risk associated with longer sedentary behaviour in relation there are specific periods across
time spent sitting was noted only to putative molecular markers of the life-course during which an in-
among individuals with low levels of cancer risk have been restricted dividual may be particularly sus-
physical activity, and no increased
to cross-sectional study designs ceptible to the adverse effects of
risk of cancer mortality with pro-
or small-scale interventions in se- prolonged sedentary behaviour. To
longed sedentary behaviour was
lected populations and have pro- date, there is inadequate evidence
noted in individuals with higher lev-
els of physical activity [11].
Like for the above-mentioned
Fig. 2.7.3. Most office work is characterized by prolonged sedentary time.
data on total cancer risk, epidemi-
ological evidence is sparse about
the relationship of sedentary be-
haviour to risk of cancer at individu-
al sites. The strongest evidence has
been reported for cancers of the
breast, colon, and endometrium.
Weaker evidence has been found
for lung cancer, a site for which as-
sociations are particularly prone to
confounding by smoking. A meta-
analysis of observational studies
reported that each increment of
2 hours per day in time spent sitting
was associated with an increase of
8% in risk of colon cancer, an in-
crease of 10% in risk of endometrial
cancer, and a borderline statistical-
ly significant increase of 6% in risk
of lung cancer [12]. Another meta-

to formulate specific recommen- Fig. 2.7.4. Children playing football in Pakistan.

dations about restrictions on daily
sedentary time or sitting breaks.
Therefore, current guidelines from
government organizations and can-
cer control agencies are limited to
generic, non-quantitative reduc-
tions in sedentary behaviour [17].
Obesity
Overweight and obesity are general-
ly assessed through various anthro-
pometric measures. In population
studies of cancer, the predominant
measures used are body mass index
(BMI), which is obtained by dividing
the body weight (in kilograms) by
the square of the height (in metres),
and waist circumference. There is
now considerable epidemiological
evidence supporting an association The associations between obesi- are observed in the never-smoker
between overweight and obesity ty and cancer risk differ within sub- category. Among ever users of hor-
and cancer risk (Table 2.7.1). This groups of the population: stronger mone replacement therapy, there
evidence has been systematically effects are observed for some can- are no associations between BMI
reviewed in dozens of meta-analy- cers for women than men, and for and postmenopausal breast cancer
ses based on hundreds of studies older versus younger populations. and ovarian cancer, and there is an
conducted worldwide, including by There is also some evidence that attenuated association with endo-
WCRF/AICR [2]. the effect of obesity on cancer risk metrial cancer. However, for nev-
There is currently convincing differs by race and ethnicity. For er-users of hormone replacement
evidence that being overweight or example, a stronger adverse effect therapy, there are clearly increased
obese in adulthood is associated of obesity on breast cancer risk risks associated with elevated BMI
with increased risks of cancers of the was found for women of Asian eth- for these three cancer sites [20].
postmenopausal breast, colorectum, nicity than for women of Hispanic, There are several plausible
endometrium, kidney, liver, oesopha- African, or non-Hispanic White an- biological mechanisms that could
gus, and pancreas, and probable cestry [19]. The observed ethnici- explain the association between
evidence for an association with can- ty-associated variation in cancer obesity and cancer risk. The main
cers of the gall bladder, gastric car- risk at similar levels of adiposity ones are an increase in endog-
dia, mouth, pharynx, larynx, ovary, is thought to be, in part, related to enous sex hormones (see Chapter
and prostate (advanced), and limited differences in distribution of body 3.6), insulin and insulin-like growth
suggestive evidence for an asso- fat. Larger waist circumference, factors, circulating adipokines,
ciation with cervical cancer [2]. For as a measure of central adiposity, and systemic inflammation [21].
breast cancer, being overweight or is now a recognized risk factor for WCRF/AICR reported that world-
obese as an adult before menopause several cancer sites independent wide in 2016, 1.97 billion adults and
decreases the risk of premenopausal of body size [2,18]. more than 338 million children and
breast cancer risk, but greater weight Other cancer risk factors are also adolescents were classified as over-
gain in adulthood increases the risk being recognized as important effect weight or obese [2]. Furthermore,
of postmenopausal breast cancer. modifiers of the association between the increase in the prevalence of
The IARC Handbooks volume obesity and cancer; the most impor- obesity is being observed in both
that reviewed the evidence on obe- tant ones are smoking (see Chapter high-income countries and low- and
sity and cancer in 2016 concluded 2.1) and use of hormone replace- middle-income countries, given the
that there was sufficient evidence ment therapy (see Chapter 2.11). increased industrialization and the
for an association between obesity Meta-analyses have generally dem- decrease in active occupations and
and 13 cancer sites, and included onstrated an inverse association be- active transport (e.g. walking and
thyroid cancer, multiple myeloma, tween obesity and smoking-related cycling) that have occurred glob-
and meningioma in this category cancers (e.g. lung cancer and oe- ally. Over the next two decades,
along with the sites previously listed sophageal cancer), which can be the largest proportional increase in
by WCRF/AICR [18]. explained by null associations that overweight and obesity is projected
106
Table 2.7.1. Evidence on the relationships of physical activity, sedentary behaviour, and obesity to risk of cancer
Cancer site Physical activity Sedentary behaviour Obesity
Colorectum Strong evidence for decreased risk Limited evidence for increased risk Strong evidence for increased risk
(colon) (colon)
Endometrium Strong evidence for decreased risk Limited evidence for increased risk Strong evidence for increased risk
Breast (postmenopausal) Strong evidence for decreased risk Limited evidence for increased risk Strong evidence for increased risk
Breast (premenopausal) Strong evidence for decreased risk Limited evidence for increased risk
SECTION 2
CHAPTER 2.7
Oesophageal adenocarcinoma Strong evidence for decreased risk Strong evidence for increased risk
Kidney Strong evidence for decreased risk Strong evidence for increased risk
Bladder Strong evidence for decreased risk
Gastric cardia Strong evidence for decreased risk Strong evidence for increased risk
Liver Strong evidence for increased risk
Lung Limited evidence for decreased risk Limited evidence for increased risk
Prostate Limited evidence for decreased risk Strong evidence for increased risk
(advanced)
Ovary Limited evidence for decreased risk Strong evidence for increased risk
Pancreas Limited evidence for decreased risk Strong evidence for increased risk
Gall bladder Strong evidence for increased risk
Mouth, pharynx, and larynx Strong evidence for increased risk
Cervix Limited evidence for increased risk
Thyroid Limited evidence for increased risk
Multiple myeloma Limited evidence for increased risk
Meningioma Limited evidence for increased risk
to occur in low- and middle-income population attributable fractions are the food environment, food systems,
countries [22]. Countries that are generally similar for men and wom- and the built environment. WCRF/
undergoing an economic transition en, although variations by sex do AICR has provided some recommen-
are particularly relevant to investi- occur, depending on the prevalence dations on how these policy changes
gate, because the impact of rapid of obesity in those populations and can be made at a governmental and
weight gain on cancer risk can be the risks associated with obesity for societal level [2].
evaluated. These trends in the prev- specific cancer sites. At a global lev- The recommendations of the
alence rates of obesity are expect- el, obesity is ranked the third most WCRF/AICR report [2] for healthy
ed to result in a substantial increase important risk factor for cancer, with weight are to keep weight within
in cancer incidence worldwide. respect to attributable fractions, af- the healthy range of BMI for adults,
Globally, the median fraction of ter smoking and infections [20]. which is 18.5–24.9 kg/m2, and to
cancers that are attributable to over- The determinants of overweight avoid weight gain in adult life. To
weight and obesity, as measured by and obesity are complex and mul- achieve this overall recommenda-
BMI, has recently been estimated tifactorial, and it is now increasingly tion, three goals were provided:
to range from less than 1% to 9.5%, recognized that a multilevel approach (i) to ensure that body weight dur-
depending on the cancer site and is necessary to decrease the preva- ing childhood and adolescence pro-
the country [23]. The highest frac- lence of obesity globally. Several ini- jects towards the lower end of the
tions are found in North America, tiatives are needed that target behav- healthy adult BMI range; (ii) to keep
the Middle East, and Europe, and iour change not only at the individual weight as low as possible within the
lower fractions are observed in sub- level but also at the societal level. healthy range throughout life; and
Saharan Africa and Asia, which Policies are required that enable (iii) to avoid weight gain, measured
corresponds to the prevalence of populations to achieve and maintain as body weight or waist circumfer-
obesity in those regions. These a healthy weight and that consider ence, throughout adulthood.

References
1. 2018 Physical Activity Guidelines Advi- 9. Biswas A, Oh PI, Faulkner GE, Bajaj RR, 16. Rezende LFM, Sá TH, Mielke GI, Viscondi
sory Committee (2018). 2018 Physical Silver MA, Mitchell MS, et al. (2015). JYK, Rey-López JP, Garcia LMT (2016).
Activity Guidelines Advisory Committee Sedentary time and its association with All-cause mortality attributable to sitting
Scientific Report. Washington (DC), USA: risk for disease incidence, mortality, and time: analysis of 54 countries worldwide.
U.S. Department of Health and Human hospitalization in adults: a systematic re- Am J Prev Med. 51(2):253–63. https://
Services. Available from: https://health. view and meta-analysis. Ann Intern Med. doi.org/10.1016/j.amepre.2016.01.022
gov/paguidelines/second-edition/report/. 162(2):123–32. https://doi.org/10.7326/ PMID:27017420
M14-1651 PMID:25599350
2. WCRF/AICR (2018). Diet, nutrition, physi- 17. Stamatakis E, Ekelund U, Ding D, Hamer M,
cal activity and cancer: a global perspec- 10. Patterson R, McNamara E, Tainio M, Bauman AE, Lee IM (2019). Is the time right
tive. Continuous Update Project Expert de Sá TH, Smith AD, Sharp SJ, et al. for quantitative public health guidelines on
Report 2018. World Cancer Research (2018). Sedentary behaviour and risk sitting? A narrative review of sedentary be-
Fund/American Institute for Cancer of all-cause, cardiovascular and cancer haviour research paradigms and findings.
Research. Available from: https://www. mortality, and incident type 2 diabetes: a Br J Sports Med. 53(6):377–82. https://
wcrf.org/dietandcancer. systematic review and dose response me- doi.org/10.1136/bjspor ts-2018 - 099131
ta-analysis. Eur J Epidemiol. 33(9):811– PMID:29891615
3. Moore SC, Lee IM, Weiderpass E, 29. https://doi.org/10.1007/s10654-018-
Campbell PT, Sampson JN, Kitahara 18. Lauby-Secretan B, Scoccianti C, Loomis
0380-1 PMID:29589226
CM, et al. (2016). Association of leisure- D, Grosse Y, Bianchini F, Straif K;
time physical activity with risk of 26 types 11. Ekelund U, Brown WJ, Steene- International Agency for Research on
of cancer in 1.44 million adults. JAMA Johannessen J, Fagerland MW, Owen Cancer Handbook Working Group (2016).
Intern Med. 176(6):816–25. https://doi. N, Powell KE, et al. (2019). Do the asso- Body fatness and cancer – viewpoint of
org/10.1001/jamainternmed.2016.1548 ciations of sedentary behaviour with car- the IARC Working Group. N Engl J Med.
PMID:27183032 diovascular disease mortality and cancer 375(8):794–8. https://doi.org/10.1056/
mortality differ by physical activity level? A NEJMsr1606602 PMID:27557308
4. Neilson HK, Conroy SM, Friedenreich CM systematic review and harmonised meta-
(2013). The influence of energetic factors 19. Bandera EV, Maskarinec G, Romieu
analysis of data from 850 060 participants.
on biomarkers of postmenopausal breast I, John EM (2015). Racial and ethnic
Br J Sports Med. 53(14):886–94. https://
cancer risk. Curr Nutr Rep. 3(1):22–34. disparities in the impact of obesity on
doi.org/10.1136/bjspor ts-2017- 098963
ht t p s: //d o i.o r g /10.10 07/s13 6 6 8 - 013 - breast cancer risk and survival: a global
PMID:29991570
0069-8 PMID:24563822 perspective. Adv Nutr. 6(6):803–19.
12. Schmid D, Leitzmann MF (2014). https://doi.org/10.3945/an.115.009647
5. Lee IM, Shiroma EJ, Lobelo F, Puska P, Television viewing and time spent PMID:26567202
Blair SN, Katzmarzyk PT; Lancet Physical sedentary in relation to cancer risk:
Activity Series Working Group (2012). 20. Renehan AG, Soerjomataram I (2016).
a meta-analysis. J Natl Cancer Inst.
Effect of physical inactivity on major Obesity as an avoidable cause of can-
106(7):dju098. https://doi.org/10.1093/
non-communicable diseases worldwide: cer (attributable risks). Recent Results
jnci/dju098 PMID:24935969
an analysis of burden of disease and life Cancer Res. 208:243–56. https://doi.
expectancy. Lancet. 380(9838):219–29. 13. Zhou Y, Zhao H, Peng C (2015). o r g /10 .10 07/ 97 8 - 3 - 319 - 42 5 42- 9 _13
https://doi.org/10.1016/S0140- 6736(12) Association of sedentary behavior with PMID:27909911
61031-9 PMID:22818936 the risk of breast cancer in women: update
21. Renehan AG, Zwahlen M, Egger M
meta-analysis of observational studies.
6. Sallis JF, Bull F, Guthold R, Heath (2015). Adiposity and cancer risk: new
Ann Epidemiol. 25(9):687–97. https://doi.
GW, Inoue S, Kelly P, et al.; Lancet mechanistic insights from epidemiology.
org/10.1016/j.annepidem.2015.05.007
Physical Activity Series 2 Executive Nat Rev Cancer. 15(8):484–98. https://
PMID:26099193
Committee (2016). Progress in physical doi.org/10.1038/nrc3967 PMID:26205341
activity over the Olympic quadrennium. 14. Ekelund U, Brage S, Besson H, Sharp
22. Malik VS, Willett WC, Hu FB (2013). Global
Lancet. 388(10051):1325–36. https://doi. S, Wareham NJ (2008). Time spent
obesity: trends, risk factors and policy im-
org /10.1016/S0140 - 673 6(16) 3 0 5 81- 5 being sedentary and weight gain in
plications. Nat Rev Endocrinol. 9(1):13–27.
PMID:27475270 healthy adults: reverse or bidirectional
https://doi.org/10.1038/nrendo.2012.199
causality? Am J Clin Nutr. 88(3):612–7.
7. Whiteman DC, Wilson LF (2016). The PMID:23165161
ht tps://doi.org/10.10 93/ajcn/8 8.3.612
fractions of cancer attributable to
PMID:18779275 23. Arnold M, Leitzmann M, Freisling H, Bray
modifiable factors: a global review.
F, Romieu I, Renehan A, et al. (2016).
Cancer Epidemiol. 44:203–21. https:// 15. Henson J, Dunstan DW, Davies MJ, Yates Obesity and cancer: an update of the
doi.org /10.1016/ j.c anep. 2016.0 6.013 T (2016). Sedentary behaviour as a new global impact. Cancer Epidemiol. 41:8–15.
PMID:27460784 behavioural target in the prevention and https://doi.org/10.1016/j.canep.2016.01.003
treatment of type 2 diabetes. Diabetes PMID:26775081
8. Sedentary Behaviour Research Network
Metab Res Rev. 32(Suppl 1):213–20.
(2012). Letter to the editor: standardized
h t t p s : / / d o i . o r g / 1 0 .1 0 0 2 / d m r r. 2 7 5 9
use of the terms “sedentary” and “seden-
PMID:26813615
tary behaviours”. Appl Physiol Nutr Metab.
37(3):540–2. https://doi.org/10.1139/h2012-
024 PMID:22540258
108
2.8
Dietary carcinogens
A continuing concern in various contexts
John D. Groopman Peter P. Fu (reviewer)
CHAPTER 2.8
SECTION 2
Koraljka Gall Trošelj (reviewer)
J. David Miller (reviewer)
Historical context context of dietary carcinogens, it is

SUMMARY For nearly 50 years, the IARC
reasonable to assume that many
tumours diagnosed today had their
●● Dietary carcinogens include Monographs have summarized the
etiological roots in about 1975–
single specific agents, such as proportionate role that dietary car-
cinogens play in the development 1995. Therefore, it remains a sig-
aflatoxin and aristolochic acid, nificant issue whether those dietary
as well as complex mixtures, of the spectrum of cancer types in
humans. Most of Volume 1 of the carcinogen factors of 20–40 years
such as processed meat. ago will continue to be risk factors
Monographs, published in 1972,
●● Consumption of processed meat was devoted to N-nitroso com- for the cancers that will be diag-
was recently classified as car- pounds formed in foods and their nosed 20–40 years from now [2].
cinogenic to humans (Group 1), consumption, and natural products This is a critical question for future
joining the individual dietary that contaminate dietary staples, risk assessment analysis and for
contaminants aflatoxin and aris- such as aflatoxins, cycasin, safrole, the informed deployment of preven-
tolochic acid in that category. and sterigmatocystin [1]. These die- tion strategies.
tary contaminants had been identi- Some dietary carcinogens,
●● Many cohort, case–control, such as aflatoxin, that were pre-
fied using the tools of epidemiology
and other observational studies dominant in the past are still sig-
and experimental toxicology, as a
have associated and causally nificant risk factors for many cancer
result of their potency as initiating
linked exposures in the diet to types in different populations today.
agents of the carcinogenic pro-
a spectrum of human cancer cess. At that time, the mechanistic Furthermore, since 1972 some new
types, including cancers of the understanding of cancer biology, dietary carcinogens, such as aris-
breast, colorectum, liver, pan- including DNA adduct formation tolochic acid, have been identified
creas, and prostate. and resultant mutations, the role of and formally classified as carcino-
●● The acceleration of the obe- oncogenes and tumour suppressor genic to humans (Group 1). In recent
sity pandemic and the rising genes in cancer development, and years, the pace of the discovery of
incidence of type 2 diabetes in the multiple stages of cancer that new single potent agents in the diet
many populations are changing span decades before diagnosis, as carcinogens has slowed down.
the potential toxicological haz- was still in its infancy. Therefore, Now, greater attention is being fo-
ard from dietary carcinogens, during that period only the most po- cused on dietary exposures from
tent carcinogens, or those carcino- complex mixtures, such as red meat
which could, in turn, increase
gens with high exposure across the and processed meat as documented
the incidence of several human
lifespan, were identified. This pro- in Monographs Volume 114 [3].
cancer types.
vided a clear focus for pursuing ba- Projecting future risk from die-
●● New technologies that use sic and population studies. tary carcinogens will require knowl-
deep sequencing methods may Most solid tumours, irrespective edge of the dramatic change that is
reveal unique mutational signa- of their organ site, evolve through occurring in country after country
tures that can inform future risk a 15–25-year period of biological with respect to population health
analyses, providing evidence development. The current under- and overall chronic disease burden
for the role of dietary carcino- standing of these molecular pro- (see [4]). Simply put, experimental
gens in cancer development. cesses is extensively reviewed in toxicology models have explored the
Section 3 of this volume. Within the potency, biology, and mechanisms
Chapter 2.8 • Dietary carcinogens 109

Fig. 2.8.1. Harvested groundnuts lying on the ground are susceptible to fungal and
mould growth and hence, among other things, aflatoxin contamination.
FUNDAMENTALS
■■ For nearly 50 years, naturally
occurring, cooking-derived,
and complex constituents of
the diet have been determined
to be risk factors for a wide
variety of cancer types
occurring at different organ
sites in humans.
■■ Assessing dietary exposures
and their roles in the
development of cancer in
humans has been a daunting
task, because exposure
assessment has proven to be
challenging. The development
of chemically specific and
mechanistically justified
biomarkers has been shown
to be very important in the
of action based on single com- Naturally occurring identification of several potent
pounds and animal models that use dietary carcinogens human carcinogens that
balanced nutrition and growth man- contaminate dietary staples.
The potency of various naturally
agement. Over the past 30 years,
occurring dietary carcinogens has ■■ The translation of basic
the average energy intake has been
spurred many investigations, be- science and mechanistic
rising rapidly in many economically
cause these contaminants pose studies has proven to be
developing countries. This trend is
a hazard across the lifespan. essential for the development
dramatically changing the physiol-
Examples of this category of agents of risk models from dietary
ogy, across the lifespan, of people
are aflatoxin, aristolochic acid, and exposures.
who are chronically exposed to di-
etary carcinogenic agents. For ex- fumonisins. These chemicals have ■■ Knowledge gleaned from
ample, by 2020 400 million people in common the range of expo- epidemiological studies of
across all continents will have been sures from major staple grains and dietary exposures has been
diagnosed with type 2 diabetes [5]. foodstuffs consumed worldwide. successfully used in cancer
This disease will contribute to an Therefore, prevention strategies prevention, particularly with
increase in the incidence of liver will have to include source mitiga- respect to aflatoxin and liver
cancer (see Chapter 5.6) and is a tion, primary and secondary precancer development.
sentinel for chronic disease result- vention, and appropriate regulatory
levels in commerce and trade. ■■ The most challenging problem
ing from the obesity pandemic. in the analysis of dietary
From a regulatory and policy carcinogens as risk factors
perspective, the current experi-
Aflatoxin
remains the difficulty involved
mental models do not necessarily Since the early 1970s, aflatoxin
in exposure assessment,
provide the data to judge whether has been repeatedly examined
particularly across the lifespan.
carcinogenic risk from specific di- as a human carcinogen, eventu-
etary carcinogens will be potentially resulting in its classification as
ated or antagonized by chronic carcinogenic to humans (Group 1)
diseases such as type 2 diabetes. in Monographs Volume 56 [6].
Fortunately, when specific carcino- Recently, an IARC Working Group which is important in liver cancer
gens or risk factors are identified, Report summarized exposures and development in Africa and Asia [8].
prevention can be successfully im- health consequences from afla- More recently, as a result of the
plemented. The targeted prevention toxin in low- and middle-income availability of aflatoxin-specific bio-
programmes that have reduced the countries [7]. Classic investigations markers, new investigations have
burden of lung cancer by decreas- have documented the greater-than- been conducted to explore expo-
ing the use of tobacco are a model multiplicative interaction between sures in populations that consume
for the future. aflatoxin and hepatitis B virus, very high levels of maize and maize
110
products. Some populations, par- the universal vaccination of new- carcinogen are herbal medicines
ticularly those in Central America, born babies against hepatitis B vi- that have been demonstrated to
consume up to 500 grams of maize rus. Median levels of the aflatoxin be contaminated with this group
per day, and even low concentra- biomarker decreased by more than of compounds.
tions of aflatoxin in this food source 95% from 1989 to 2009. A popula- During the past 25 years, suf-
can lead to substantial exposures tion attributable benefit of 65% for ficient evidence has accrued for
on a daily basis. reduced liver cancer mortality was aristolochic acid to be classified as
A study in Guatemala that estimated from a government-facil- carcinogenic to humans (Group 1),
used the aflatoxin-specific serum itated switch of dietary staple from as summarized in Monographs
albumin biomarkers found levels maize to rice [12]. Thus, economic Volume 100A [15]. The specific
CHAPTER 2.8
SECTION 2
comparable to those detected dur- growth is leading to market basket mutational signature found in the
ing the 1980s in high-risk coun- diversity, which will help to reduce TP53 tumour suppressor gene that
tries in Africa and Asia [9,10]. exposure to aflatoxin from a single is a result of aristolochic acid–ad-
Remarkably, Guatemala has the source that is susceptible to high enine adducts has formed a basis
highest liver cancer incidence rate levels of contamination. for biomarkers to explore this car-
in the Western Hemisphere [11], cinogen as a risk factor in many
but preliminary studies have found Aristolochic acid populations [14].
low levels of hepatitis B virus and A coalescence of epidemiological A recent population-based
hepatitis C virus infection. Thus, the research – focused on the etiology case–control study involving near-
availability of sensitive and specific of Balkan endemic nephropathy, ly 6000 cases and about 23 000
biomarkers is expanding the under- an investigation of rare urothelial controls investigated the linkage
standing of at-risk populations and cancers in people who participated between history of prescription of
communities in previously underin- in certain weight-reduction inter- medicines containing Aristolochia,
vestigated regions of the world. ventions, and a unique mutational cumulative consumption of aris-
In eastern China, the availabili- signature in TP53 in tumours – led tolochic acid, and renal cell carcino-
ty of serum samples collected over to the discovery of the role of aris- ma in Taiwan, China. The presence
a 20-year period has enabled the tolochic acids in human cancer and level of mutagenic aristolochic
measurement of changing aflatoxin [13,14]. Aristolochic acid emerged acid-derived DNA adducts were
exposure patterns by using the bio- as a dietary carcinogen as a re- determined. Cumulative ingestion
sult of inadvertent contamination of more than 250 milligrams of aris-
marker strategy described above.
of staple grains as they grow in the tolochic acid increased the risk of
The population-based cancer regis-
field, because Aristolochia plants renal cell carcinoma, with an odds
try in Qidong, China, documented a
ratio of 1.25. Furthermore, the dis-
reduction of more than 50% in mor- encroach on the fields. During har-
tinctive mutational signature de-
tality rates from primary liver cancer vest, the Aristolochia plant is har-
scribed above was evident in 6 of
across birth cohorts from the 1960s vested together with the foodstuff
10 sequenced renal cell carcinoma
to the 1980s for people younger (such as wheat). Other widespread
exomes [16]. This study and others
than 35 years; all were born before sources of human exposure to this
provide strong evidence implicat-
ing aristolochic acid in a significant
fraction of renal cell carcinoma
Fig. 2.8.2. An Aristolochia plant. These plants encroach on fields where staple grains
in Taiwan, China, and thus aris-
are growing, and hence cause contamination of the grains during harvesting.
tolochic acid may contribute more
broadly to this cancer type in many
other settings.
Fumonisin
The initial reports implicating fu-
monisins in human cancer were in
association with high rates of oe-
sophageal cancer in residents of
Transkei, South Africa, in 1988 [8].
Mechanistically, fumonisin causes
toxic effects through inhibition of
ceramide synthase, an enzyme
needed for sphingolipid metabolism.
Elevated levels of fumonisin
in animal feed cause diseases
such as leukoencephalomalacia

Fig. 2.8.3. Maize contaminated with fu- This evaluation reviewed numerous multiple chronic diseases contribut-
monisin. Studies have supported a role cohort, case–control, and other ing to the development of cancers.
for fumonisins in the development of a observational studies across many It is clear that consumption of
variety of human cancer types.
different populations. Consumption red meat and processed meat plays
of red meat was classified as a role in the development of cancers
probably carcinogenic to humans of the colorectum, pancreas, and
(Group 2A), and consumption of prostate. These findings suggest
processed meat was classified as opportunities for prevention, par-
carcinogenic to humans (Group 1). ticularly for colorectal cancer and
Similarly to other Monographs prostate cancer, for which screen-
evaluations that reviewed complex ing methods exist and for which
mixtures and culminated in the iden- the incidence is rising as countries
tification of carcinogenic hazards to transition to higher levels of eco-
humans, such as the evaluation of nomic development. Furthermore,
outdoor air pollution (Monographs because pancreatic cancer is a
Volume 109) [23], this evaluation major contributor to overall cancer
of red meat and processed meat mortality, these findings provide
transcends traditional compound- further justification for the develop-
by-compound approaches to haz- ment of biomarkers in early detec-
ard assessments. From a policy tion strategies for this cancer, which
and regulatory perspective, this is nearly always fatal [24,25].
has enormous implications for the Although it has been revealed
translation of these findings in both in numerous epidemiological in-
individual and population public vestigations that consumption of
health prevention. red meat and processed meat
Collectively, the findings in contributes to the development of
Monographs Volume 114 point to- cancer in humans, the proportion-
wards major lifestyle factors that ate roles of individual agents or
clearly underlie the development classes of chemical carcinogens in
in horses and pulmonary oedema, of many cancers that will be diag- these products remain unresolved.
reduced weight gain, and liver dam- nosed throughout the rest of this Since the early 1970s, N-nitroso
age in swine. Fumonisin has also century. Diets high in meat con- compounds have been evaluated
been shown to cause liver cancer sumption also have impacts on for their carcinogenic hazard to hu-
and kidney cancer in rats and liver the development of other chronic mans. Controversy has surrounded
cancer in mice, as summarized in diseases, such as type 2 diabetes, the role that nitrates and nitrites play
World Cancer Report 2014 [17]. further illustrating the complexity of in a balance between preservation
Collectively, studies in China and
South Africa have supported a role
for fumonisins in the development Fig. 2.8.4. Frankfurters and other processed meats. Numerous epidemiological
of a variety of human cancer types, investigations have revealed that consumption of processed meat contributes to the
development of cancer in humans.
and because of its widespread con-
tamination of maize, this agent may
also interact with other mycotoxins,
amplifying their effect in the initia-
tion of cancer [18,19].
Red meat and processed

meat
Among the most significant recent
advances in the understanding of
the role of dietary carcinogens in
cancer at several organ sites in hu-
mans was the evaluation published
in Monographs Volume 114 [3] on
the contribution of red meat and
processed meat to cancer devel-
opment (see Chapter 2.6) [20–22].
112
of foods, bacteriological resis- to the formation of these chemical analytical strategies for eventual
tance, and general organolep- agents and the biological safety of translation to prevention and inter-
tic presentation, given that many the cooked product. This remains an ventions. The enormous variation
specific N-nitroso compounds are unresolved issue that needs to be on a day-to-day basis due to cook-
potent experimental carcinogens. addressed in future research. ing practices and sources of these
Biomarkers have been developed foods contributes to major uncer-
to attempt to evaluate internal and Future insights and tainty in exposure assessment for
biological effective dose from ex- strategies the compounds present or formed
posures to these agents. However, in different food components.
Over the past several decades,
many different chemical com- New technologies that use deep
CHAPTER 2.8
SECTION 2
there has been tremendous prog-
pounds form identical adducts, and sequencing methods may reveal
ress in the identification of single
this has confounded the ability to unique mutational signatures that
chemical carcinogens in the diet
obtain precise measurements of can be used as integrative metrics
that are associated with – and, in
exposure or dose. for cancer risk assessment be-
some cases, causally linked to –
Various compounds are chemi- the development of cancer in hu- fore a tumour diagnosis. Advances
cally formed during the cooking of mans. Recent findings have shown achieved with these new deep se-
red meat. These include acrylamide, that a reduction in exposure to afla- quencing technologies and their at-
many heterocyclic aromatic amines, toxin, as documented by biomarker tendant biostatistical approaches
and many different polycyclic aro- measurements, has produced a have shown mutational fingerprints
matic hydrocarbons. Each group reduction in the incidence of liver for specific carcinogens, such as
or class of these compounds has cancer in a high-risk population. aflatoxin and aristolochic acid, and
deleterious biological potency in ex- This reduction is similar in trajec- the patterns are also suggestive for
perimental models, and the hetero- tory over time to the decrease in oxidative damage [27,28]. Use of the
cyclic aromatic amines have been the risk of lung cancer seen in in- accumulated damage that survives
demonstrated to cause cancers of dividuals who quit tobacco smoking to a tumour diagnosis as a metric of
the breast, colon, and prostate in [26]. These data provide a roadmap the area under the curve for long-
experimental models. Collectively, for translation to other agents that term dosages of carcinogens is an
these agents represent intriguing have been identified as being po- exciting prospect for future work.
hypotheses for their contribution to tent human carcinogens. However, It will be a challenge to the cancer
the development of cancer in hu- recent analyses indicate that com- prevention community not only to
mans. Similarly to the issue with ni- plex dietary situations, such as that develop these analytical strategies
trates and nitrites, there is a balance found with red meat and processed but also to validate them in investi-
between the processes that lead meat, pose particularly challenging gations in human populations.

References
1. IARC (1972). Some inorganic substanc- 10. Smith JW, Kroker-Lobos MF, Lazo 20. Farinetti A, Zurlo V, Manenti A, Coppi
es, chlorinated hydrocarbons, aromatic M, Rivera-Andrade A, Egner PA, F, Mattioli AV (2017). Mediterranean
amines, N-nitroso compounds, and natu- Wedemeyer H, et al. (2017). Aflatoxin and diet and colorectal cancer: a system-
ral products. IARC Monogr Eval Carcinog viral hepatitis exposures in Guatemala: atic review. Nutrition. 43–44:83–8.
Risk Chem Man. 1:1–184. Available from: molecular biomarkers reveal a unique https://doi.org/10.1016/j.nut.2017.06.008
http://publications.iarc.fr/19. profile of risk factors in a region of high PMID:28935150
liver cancer incidence. PLoS One.
2. Fitzmaurice C, Akinyemiju TF, Al Lami 12(12):e0189255. https://doi.org/10.1371/ 21. Domingo JL, Nadal M (2017).
FH, Alam T, Alizadeh-Navaei R, Allen C, journal.pone.0189255 PMID:29236788 Carcinogenicity of consumption of red
et al.; Global Burden of Disease Cancer meat and processed meat: a review of
Collaboration (2018). Global, regional, 11. Ferlay J, Ervik M, Lam F, Colombet M, scientific news since the IARC deci-
and national cancer incidence, mortality, Mery L, Piñeros M, et al. (2018). Global sion. Food Chem Toxicol. 105:256–61.
years of life lost, years lived with disabil- Cancer Observatory: Cancer Today. https://doi.org/10.1016/j.fct.2017.04.028
ity, and disability-adjusted life-years for 29 Lyon, France: International Agency for PMID:28450127
cancer groups, 1990 to 2016: a systematic Research on Cancer. Available from:
analysis for the Global Burden of Disease https://gco.iarc.fr/today. 22. Alexander DD, Weed DL, Miller PE,
Study. JAMA Oncol. 4(11):1553–68. Mohamed MA (2015). Red meat and colo-
12. Chen JG, Egner PA, Ng D, Jacobson rectal cancer: a quantitative update on
https://doi.org/10.1001/jamaoncol.2018.2706
LP, Muñoz A, Zhu Y-R, et al. (2013). the state of the epidemiologic science.
PMID:29860482
Reduced aflatoxin exposure presages J Am Coll Nutr. 34(6):521–43. https://
3. IARC (2018). Red meat and processed decline in liver cancer mortality in an doi.org/10.1080/07315724.2014.992553
meat. IARC Monogr Eval Carcinog Risks endemic region of China. Cancer Prev PMID:25941850
Hum. 114:1–502. Available from: http:// Res (Phila). 6(10):1038–45. https://doi.
publications.iarc.fr/564. org/10.1158/1940 - 6207.CAPR-13- 0168 23. IARC (2015). Outdoor air pollution. IARC
PMID:23963804 Monogr Eval Carcinog Risks Hum. 109:1–
4. WHO (2017). Evaluation of certain con- 448. Available from: http://publications.
taminants in food: eighty-third report of 13. Aydin S, Ambroise J, Cosyns J-P, Gala iarc.fr/538.
the Joint FAO/WHO Expert Committee J-L (2017). TP53 mutations in p53-
on Food Additives. Geneva, Switzerland: negative dysplastic urothelial cells from 24. Abid Z, Cross AJ, Sinha R (2014).
World Health Organization (WHO Belgian AAN patients: new evidence Meat, dairy, and cancer. Am J Clin
Technical Report Series, No. 1002). for aristolochic acid-induced molecular Nutr. 100(Suppl 1):386S–93S. https://
Available from: https://apps.who.int/iris/ pathogenesis and carcinogenesis. Mutat d o i . o r g / 1 0 . 3 9 4 5 / a j c n .11 3 . 0 7 1 5 9 7
handle/10665/254893. Res. 818:17–26. https://doi.org/10.1016/j. PMID:24847855
mrgentox.2017.03.003 PMID:28477877
5. Afshin A, Forouzanfar MH, Reitsma MB, 25. Jankovic N, Geelen A, Winkels RM,
14. Hoang ML, Chen C-H, Sidorenko VS, He Mwungura B, Fedirko V, Jenab M, et
Sur P, Estep K, Lee A, et al.; GBD 2015
J, Dickman KG, Yun BH, et al. (2013). al.; Consortium on Health and Ageing:
Obesity Collaborators (2017). Health ef-
Mutational signature of aristolochic Network of Cohorts in Europe and the
fects of overweight and obesity in 195
acid exposure as revealed by whole- United States (CHANCES) (2017).
countries over 25 years. N Engl J Med.
exome sequencing. Sci Transl Med. Adherence to the WCRF/AICR dietary
377(1):13–27. https://doi.org/10.1056/
5(197):197ra102. https://doi.org/10.1126/ recommendations for cancer prevention
NEJMoa1614362 PMID:28604169
scitranslmed.3006200 PMID:23926200 and risk of cancer in elderly from Europe
6. IARC (1993). Some naturally occurring and the United States: a meta-analysis
substances: food items and constituents, 15. IARC (2012). Pharmaceuticals. IARC within the CHANCES project. Cancer
Monogr Eval Carcinog Risks Hum. Epidemiol Biomarkers Prev. 26(1):136–
heterocyclic aromatic amines and myco-
100A:1–437. Available from: http://publi 44. https://doi.org/10.1158/1055-9965.
toxins. IARC Monogr Eval Carcinog Risks
cations.iarc.fr/118 PMID:23189749 EPI-16-0428 PMID:27793797
Hum. 56:1–599. Available from: http://
publications.iarc.fr/74. 16. Hoang ML, Chen C-H, Chen P-C, Roberts 26. Jha P, Peto R (2014). Global effects
NJ, Dickman KG, Yun BH, et al. (2016). of smoking, of quitting, and of taxing
7. Wild CP, Miller JD, Groopman JD, edi-
Aristolochic acid in the etiology of re- tobacco. N Engl J Med. 370(1):60–8.
tors (2015). Mycotoxin control in low- and
nal cell carcinoma. Cancer Epidemiol https://doi.org/10.1056/NEJMra1308383
middle-income countries. Lyon, France:
Biomarkers Prev. 25(12):1600–8. https:// PMID:24382066
International Agency for Research on
doi.org/10.1158/1055-9965.EPI-16-0219
Cancer (IARC Working Group Reports,
PMID:27555084 27. Petljak M, Alexandrov LB (2016).
No. 9). Available from: http://publications.
Understanding mutagenesis through de-
iarc.fr/535. 17. Stewart BW, Wild CP, editors (2014).
lineation of mutational signatures in hu-
World cancer report 2014. Lyon, France:
8. Wu F, Groopman JD, Pestka JJ (2014). man cancer. Carcinogenesis. 37(6):531–
International Agency for Research on
Public health impacts of foodborne mycotox- 40. https://doi.org/10.1093/carcin/bgw055
Cancer. Available from: http://publica
ins. Annu Rev Food Sci Technol. 5(1):351– PMID:27207657
tions.iarc.fr/396.
72. https://doi.org/10.1146/annurev-food-
28. Rosenquist TA, Grollman AP (2016).
030713-092431 PMID:24422587 18. Pitt JI, Miller JD (2017). A concise history
Mutational signature of aristolochic acid:
of mycotoxin research. J Agric Food Chem.
9. Groopman JD, Egner PA, Schulze KJ, Wu clue to the recognition of a global disease.
65(33):7021–33. https://doi.org/10.1021/
LS-F, Merrill R, Mehra S, et al. (2014). DNA Repair (Amst). 44:205–11. https://
acs.jafc.6b04494 PMID:27960261
Aflatoxin exposure during the first 1000 doi.org/10.1016/j.dnarep.2016.05.027
days of life in rural South Asia assessed 19. De Ruyck K, De Boevre M, Huybrechts I, PMID:27237586
by aflatoxin B1-lysine albumin biomark- De Saeger S (2015). Dietary mycotoxins,
ers. Food Chem Toxicol. 74:184–9. co-exposure, and carcinogenesis in hu-
https://doi.org/10.1016/j.fct.2014.09.016 mans: short review. Mutat Res Rev Mutat
PMID:25308602 Res. 766:32–41. https://doi.org/10.1016/j.
mrrev.2015.07.003 PMID:26596546
114
2.9
Contamination of air, water, soil, and food

The challenge is to characterize
specific risks
Pietro Comba Aaron J. Cohen (reviewer) Michelle C. Turner (reviewer)
CHAPTER 2.9
SECTION 2
Ivano Iavarone Kathryn Z. Guyton (reviewer)
Manolis Kogevinas Josiah Ochieng (reviewer)
alkylated substances, and met- tics: (i) They are widespread (e.g.

SUMMARY als). Water pollution can be due air pollution, which affects billions of
to leaks from contaminated soils, people worldwide). (ii) They frequent-
●● Exposures to environmental car- and can result in contamination ly occur at low doses (e.g. endocrine
cinogens are widespread, and in- of the food chain. disrupters in numerous foods and
clude a large number of agents
●● The prevention of exposure to products). In specific populations, en-
emitted by different sources to
which human populations are carcinogenic environmental pol- vironmental exposures may be high
exposed through various routes. lutants requires both regulatory (e.g. air pollution in low- and middle-
Many people may be exposed to action and community commit- income countries or in the case of
relatively low levels of environment. At the global level, the accidents). (iii) They frequently oc-
mental carcinogens, thus poten- situation is currently improving in cur in mixtures (e.g. the hundreds
tially accounting for a substantial high-income countries and wors- of chemicals in drinking-water).
number of excess cancer cases. ening in low- and middle-income (iv) They occur throughout the lifetime
countries. (e.g. exposure may begin in utero and
●● Air pollution, both outdoor and
indoor, is the most widely in- ●● Exposome approaches to re- continue in childhood and adult life).
vestigated and most important search on environment and can- (v) They may concern single agents
contributor to the environmental cer have been applied recently, and routes (e.g. dioxins originating
cancer burden in human popu- based on extensive technologi- from incomplete combustion of waste
lations. Air pollution alone was cal advances that opened up and ingested through contaminated
responsible for an estimated new opportunities to collect and food), or they may concern mixtures
350 167 deaths from lung cancer analyse large data sets and proof chemicals from multiple sources
worldwide in 2017. mote effective preventive actions and routes (e.g. heavy metals, gas-
and policies. Exposome studies eous pollutants, particulate matter,
●● The most consistent predictor of
the carcinogenicity of air pollu- promote interdisciplinarity in re- and dioxins from complex industrial
tion is the concentration of air- search, encompass a wide spec- settings such as smelters, steel fac-
borne particulate matter with par- trum of environmental exposures tories, and chemical plants).
ticles of aerodynamic diameter experienced by humans from The high prevalence of such ex-
less than 2.5 µm. This complex conception onward, and integrate posures and the lifetime duration of
mixture of pollutants originates the external exposome with com- exposure result in high population
mainly from fuel combustion for plex mechanistic interactions and attributable risks, even though the
transportation, power genera- cross-omics responses. relative risks may be low. Recently,
tion, industrial activity, combus- technological developments have
tion of biomass, and domestic been applied to studies on envi-
heating and cooking. Throughout life, people are involun- ronmental carcinogens, and an
●● Drinking-water, or water used for tarily exposed to a wide range of pol- exposome approach has enabled
agricultural or recreational activi- lutants at home and in the general extensive assessments of multiple
ties, can be polluted by naturally environment, and many of these pol- exposures and linked them with bio-
occurring carcinogenic contami- lutants are established or suspected logical pathways [1–5].
nants (e.g. arsenic) or by anthro- carcinogens (Table 2.9.1). Exposure to specific environmen-
pogenic pollutants (e.g. chlo- Such environmental exposures tal carcinogens may differ widely
rinated agents, perfluorinated have several common characteris- across populations, and the mixture
Chapter 2.9 • Contamination of air, water, soil, and food 115

of environmental carcinogens to Outdoor air pollution

which populations are exposed var- Outdoor air pollution is a complex FUNDAMENTALS
ies in time and space. Multiple major mixture of pollutants originating
environmental pollutants have been mainly from fuel combustion for trans- ■■ Environmental carcinogenesis
evaluated by the IARC Monographs portation, power generation, indus- has been extensively studied
in terms of carcinogenic hazard to since the 1980s, but only recently
trial activity, combustion of biomass, has the evolution of study
humans (Table 2.9.1). and domestic heating and cooking protocols, integrating population-
The characteristics of environ- (https://www.who.int/airpollution/ based observational and
mental exposures have implications ambient/pollutants/en/). mechanistic experimental studies,
for risk assessment that are complex Outdoor air pollution comprises provided a comprehensive
and frequently depend on extrapola- a multitude of chemical and physi- evidence base for causal
tion from higher doses. The preven- cal constituents that vary globally as inference and supported reliable
estimation of the burden of
tion of exposure to environmental a result of differences in emission cancer attributable to pollution.
pollutants, which derives mainly sources, climate, and meteorology.
from uncontrolled urbanization and Among these constituents, several ■■ Exposure to outdoor air pollution
agents or mixtures have been estab- from multiple sources, including
industrialization, requires both regu-
diesel engine exhaust and
latory action and community com- lished to be carcinogenic to humans,
industrial processes, causes
mitment [6]. including benzene, 1,3-butadiene, lung cancer, and continuing
This chapter focuses on chemi- diesel engine exhaust, silica dust, household use of solid fuels
cal pollutants; for information on ra- benzo[a]pyrene, chromium, arsenic, causes lung cancer.
diation of various types, please see and asbestos (Table 2.9.1).
■■ Contamination of drinking-water
Chapters 2.4 and 2.5. In long-term longitudinal studies of by arsenic causes lung cancer,
exposure to outdoor air pollution, the bladder cancer, and skin cancer.
most consistent predictor of adverse
Air pollution health effects is the concentration of ■■ A variety of other potentially
carcinogenic pollutants occur in
The Global Burden of Disease Study PM2.5. On the basis of results from various communities worldwide,
2017 considered 84 behavioural, en- these studies and on strong experi- but their impact on cancer
vironmental, occupational, and meta- mental and mechanistic evidence, the causation is still not well known.
bolic risk factors with convincing or IARC Monographs classified overall
■■ Research on environment and
probable evidence of causation of outdoor air pollution as well as par-
cancer has focused largely on
human diseases [7]. (Estimates and ticulate matter in outdoor air pollution high-income countries, where
metadata from the Global Burden as carcinogenic to humans (Group 1), exposure to environmental
of Disease Study 2017 are avail- causing lung cancer [8]. The IARC carcinogens is in many instances
able from the Institute for Health Monographs also reviewed the evi- decreasing as a result of
dence for exposure to air pollution and regulatory action.
Metrics and Evaluation at http://
ghdx.healthdata.org/gbd-2017, http:// other cancer types, including bladder ■■ The impact of regulation can be
ghdx.healthdata.org/gbd-results- cancer, breast cancer, leukaemia and seen as resulting in the reloca-
lymphoma, childhood cancers, and all tion of certain industrial pro-
tool, and http://www.healthdata.org/
cancers combined, and concluded cesses to low-income countries,
data-visualization/gbd-compare.) exposing the local population to
that the evidence was positive but
Air pollution – which includes carcinogenic products or waste.
limited for bladder cancer only. More
airborne particulate matter with par- International cooperation is need-
recently, large studies, including the ed to redress this phenomenon.
ticles of aerodynamic diameter less
European Study of Cohorts for Air
than 2.5 µm (PM2.5), ambient ozone, ■■ The exposome approach aims to
Pollution Effects (ESCAPE) project,
and household PM2.5 due to the use have not identified an association be- assess and prevent health risks
of solid cooking fuel – was the fifth due to environmental exposures
tween air pollution and risk of incident
highest cause of death among the 84 by integrating information
bladder cancer; however, there was on the external environment
risk factors in the Global Burden of some additional evidence that long- (contaminants, lifestyle factors,
Disease Study 2017, with 4.9 million term exposure to outdoor air pollution diet, socioeconomic status, etc.)
attributable deaths and 147.4 million may be associated with risk of kidney and the internal environment
disability-adjusted life years (DALYs). cancer, breast cancer, brain cancer, (biological factors such as
For lung cancer, the overall burden and liver cancer [9–13]. genetic and metabolic factors).
attributable to indoor and outdoor WHO provides air quality guide- The exposome approach is
PM2.5 pollution was estimated to particularly relevant in assessing
lines and interim targets for the environmental exposures to
be 350 167 deaths and 7.8 million concentration of outdoor PM2.5 [14]. complex chemical mixtures,
DALYs, related mostly to outdoor In 2017, 92% of the world’s popula- which are possibly related to
PM2.5 pollution (265 267 deaths and tion lived in areas that exceeded the cancer and other health effects.
5.9 million DALYs) [7]. WHO air quality guideline of 10 µg/m3
116
Table 2.9.1. Environmental pollutants evaluated in terms of carcinogenic hazard to humans, the main associated cancer sites or
types, and the level of evidence (IARC Monographs classification)
Agent Cancer site or type IARC Monographs classification a
Outdoor air pollution
Outdoor air pollution, particulate matter in outdoor air Lung Group 1

pollution
Outdoor air pollutants, other b Lung, leukaemia Group 1

Diesel engine exhaust, silica dust, benzene
CHAPTER 2.9
SECTION 2
Indoor air pollution
Indoor emissions from household combustion of coal Lung Group 1
Indoor emissions from household combustion of biomass Lung Group 2A

fuel (primarily wood)
Second-hand tobacco smoke Lung Group 1
Indoor air pollutants, other b

Lung, leukaemias, lymphoma, Group 1
Benzene, 1,3-butadiene, diesel engine exhaust, ethylene nasopharynx, and others
oxide, formaldehyde, polychlorinated biphenyls
Asbestos and other fibres
Asbestos Lung, mesothelioma, larynx, ovary Group 1
Erionite, fluoro-edenite Mesothelioma Group 1
Drinking-water contaminants
Arsenic Lung, skin, bladder Group 1
Disinfection by-products Bladder Group 2B and Group 3
Nitrates Stomach Group 2A
Contaminants of soil and food, including pesticides
Dioxin (2,3,7,8-tetrachlorodibenzo-para-dioxin) All neoplasms Group 1
Polychlorinated biphenyls Skin, melanoma Group 1
Lindane Lymphomas Group 1
Several other pesticides Mostly leukaemia and lymphoma Group 2A
Metals in water and soil
Cadmium, lead, chromium(VI) Lung Group 1
Endocrine disrupters
Food, cosmetics, and other productsc Breast, testis Specific Group 1 carcinogens (e.g.
2,3,7,8-tetrachlorodibenzo-para-
dioxin) are endocrine disrupters
Ionizing and ultraviolet radiation
Radon-222 and its decay products (indoor air) Lung Group 1
Solar radiation Skin, malignant melanoma Group 1
Tanning devices that emit ultraviolet radiation Cutaneous malignant melanoma, Group 1
ocular melanoma
Non-ionizing radiation
Extremely low frequency magnetic fields Childhood leukaemia Group 2B
Radiofrequency electromagnetic fields Brain Group 2B

a
Group 1, carcinogenic to humans; Group 2A, probably carcinogenic to humans; Group 2B, possibly carcinogenic to humans; Group 3, not classifiable as
to its carcinogenicity to humans.
b
Identified primarily in the occupational environment but also present in the general environment.
c
Not evaluated by the IARC Monographs.

Fig. 2.9.1. Global map comparing concentrations of outdoor fine particulate matter (with particles of aerodynamic diameter less than
2.5 µm [PM2.5]) in 2017 with the WHO air quality guideline and Interim Target levels.
for outdoor PM2.5; 82% lived in ar- deaths worldwide, including those DALYs per 100 000 people in 2017
eas that exceeded Interim Target 3 from lung cancer [15]. (Fig. 2.9.2).
(15 µg/m3), 67% lived in areas that Ambient PM2.5 is the second It should be noted that the burden
exceeded Interim Target 2 (25 µg/m3), leading cause of lung cancer deaths of disease estimates for air pollution
and 54% lived in areas that exceeded (265 267 deaths and 5.9 million and lung cancer and other causes of
Interim Target 1 (35 µg/m3) (Fig. 2.9.1). DALYs globally), after smoking (see death have considerable uncertainty.
Among the world’s most populous This is because they are estimated
Chapter 2.1). The global burden of
countries, wide disparities exist in the by extrapolating the results of stud-
lung cancer deaths due to outdoor
changes in air quality from 1990 to ies in high-income countries with low
PM2.5 increased from 53 DALYs
2017. The largest improvements in PM2.5 concentrations to the high lev-
per 100 000 people in 1990 to 77 els of exposure measured in China
PM2.5 levels occurred in only a few
countries (Brazil, Japan, the Russian DALYs per 100 000 people in 2017; and other low- and middle-income
Federation, the USA, and countries in this increase was more rapid in Asia countries, using an integrated ex-
the European Union), whereas large and particularly in China, where the posure–response function for PM2.5.
percentages of the populations of burden increased from 75 DALYs This approach may underestimate
Bangladesh, China, India, Nigeria, per 100 000 people in 1990 to 220 the actual burden of lung cancer and
and Pakistan continue to live in ar-
eas with PM2.5 levels that still exceed
the less stringent WHO Interim Target Fig. 2.9.2. Disability-adjusted life years (DALYs, rate per 100 000 people) due to lung
cancer attributable to outdoor fine particulate matter (with particles of aerodynamic
1 (35 µg/m3).
diameter less than 2.5 µm [PM2.5]) in China, in Asia, and globally from 1990 to 2017,
Outdoor PM2.5 was the eighth for both sexes and all ages.
highest cause of death among the 84
risk factors in the Global Burden of
Disease Study 2017, responsible for
an overall burden of 2.9 million deaths
and 83.0 million DALYs. Large propor-
tions of the global burden of disease
due to outdoor PM2.5 occurred in
China (851 660 deaths and 19.8 mil-
lion DALYs) and India (673 129 deaths
and 21.3 million DALYs) [7].
More recent assessments of the
disease burden of outdoor PM2.5,
incorporating new evidence from
studies in countries with high levels
of pollution, produced much higher
estimates ranging up to 8.9 million
118
other causes of death in low- and is household air pollution caused by Although the global situation has
middle-income countries. Numerous the incomplete combustion of solid improved recently, in 2017 household
exposome studies have examined fuels for cooking and heating [19]. air pollution from the combustion of
personal measurements of air pol- Indoor emissions from the house- solid fuels still contributed to 1.6 mil-
lution using sensors or have used hold combustion of coal have been lion deaths (almost 3% of all deaths
other advanced models for ex- classified as carcinogenic to humans globally) and 59.5 million DALYs. Of
posure assessment in relation to (Group 1), and indoor emissions those deaths, almost one half (46%)
different –omics data, such as DNA from the household combustion of occurred in China and India, and
methylation, and provide new evi- biomass fuel are currently classified about one quarter (24%) occurred
dence on biological pathways that as- as probably carcinogenic to humans in sub-Saharan Africa – the parts of
CHAPTER 2.9
SECTION 2
sociate air pollution with disease [16]. (Group 2A) (Table 2.9.1). the world in which use of solid fuel is
New research on noncommuni- Trials that are currently under way most prevalent.
cable diseases has examined the have shown benefits from the use of Other important contributors
influence of urban environments in an advanced combustion cookstove to indoor air pollution, from non-
a wider perspective than examining that reduces indoor air pollutants and combustion sources, are radon and
only air pollution. Features of the thus the associated health effects, construction and building materials
built environment and green spaces including lung cancer [20]. (glues, formaldehyde, lead in paint or
have been associated with improve- Globally, the proportion of house- pipes, and asbestos). Second-hand
ments in various health outcomes, holds that rely on solid fuels for cook- tobacco smoke also contributes to
including psychological well-being, ing decreased from about 57% in indoor air pollution, and although
birth outcomes, cardiovascular dis- 2005 to 47% in 2017. Although this progress in combating tobacco smok-
eases, cancer, and overall mortality. proportion is decreasing in many ing has resulted in global declines,
Results from a large cohort study of countries, the number of people who the most recent estimates from the
women in the USA indicated that sur- are potentially exposed to household Global Burden of Disease Study
rounding greenness (vegetation) at air pollution may remain the same or 2017 showed that the burden of lung
the place of residence is associated even increase as populations con- cancer attributable to second-hand
with reduced cancer mortality [17]; tinue to grow. In 2017, the numbers tobacco smoke was still increasing:
this effect was mediated only to a and proportions of people exposed from 77 635 deaths and 1.8 million
small extent by physical activity. In a to household air pollution from the DALYs in 2007 to 99 579 deaths and
study in Spain, residential proximity combustion of solid fuels for cooking 2.2 million DALYs in 2017 [7]. Most of
to green spaces was found to be re- were as follows: in India, 846 million the above-mentioned contributors to
lated to a reduced risk of breast can- people (60% of the population); in indoor air pollution have been clas-
cer; physical activity did not seem to China, 452 million people (32% of the sified by the IARC Monographs as
mediate these results [18]. population); in Bangladesh, 124 mil- carcinogenic to humans (Table 2.9.1).
lion people (79% of the population),
Indoor air pollution and in the Democratic Republic of the Asbestos and other fibres
At a global level, by far the most im- Congo, 78 million people (96% of the The majority of mesothelioma cases
portant contributor to indoor pollution population) (Fig. 2.9.3). worldwide are due to occupational
Fig. 2.9.3. Global map comparing the proportion of the population exposed to household air pollution from the combustion of solid
fuels for cooking in 2017.

exposure to asbestos. In addition, taminants and risk of cancer is for Exposure to low levels of arsenic
an etiological role of environmental arsenic in drinking-water. Numerous is widespread. Evidence on risk of
exposure is well assessed with re- studies have associated exposure bladder cancer at low to moderate
spect to occurrence of asbestos in to water disinfection by-products levels of exposure to arsenic comes
the home or the presence of asbestos with risk of bladder cancer. The epi- mostly from studies in Europe and
industrial facilities in the vicinity [21]. demiological evidence is limited or the USA, and the findings are less
Although a few cases of mesothelio- inconsistent for other water contami- consistent. The excess incidence of
ma have been reported in individu- nants, including nitrates, perfluori- bladder cancer in the New England
als who had indoor asbestos expo- nated alkylated substances, metals, region of the USA has been attribut-
sure, the available evidence on risk and radionuclides. The United States ed, in part, to the high arsenic content
for inhabitants of asbestos-roofed Environmental Protection Agency proof well water [28].
houses is inadequate to assess risk vides a list of drinking-water contami-
of cancer. nants, which identifies various car- Water disinfection
The available estimates of the cinogens (https://www.epa.gov/sites/ by-products
proportion of mesothelioma cases production/files/2016-06/documents/ Chlorination by-products in drinking-
caused by environmental asbestos npwdr_complete_table.pdf). water have been consistently as-
exposure range from 4% to 20% [22]. Use of water is also associated sociated with risk of bladder cancer
Naturally occurring asbestos or as- with risk of cancer through the trans- [29,30]. Chlorination of drinking-water
bestiform fibres in soils have been mission of infectious agents, for ex- is used for disinfection. During chlo-
reported in different geographical ample squamous cell carcinoma of rination, chlorine reacts with organic
areas. Erionite has been shown to the bladder in relation to infection matter in water to produce a mixture
cause mesothelioma in studies in by Schistosoma haematobium (see of by-products, including trihalometh-
Turkey, and these findings have re- Chapter 2.2). anes, haloacetic acids, and hundreds
cently been confirmed in a study in of other compounds. Several of these
Mexico [23]. The most recent find- Arsenic in drinking-water compounds are mutagenic to bacteria,
ings concern fluoro-edenite, an am- Evidence linking arsenic in drinking- and some are carcinogenic to animals.
phibolic fibre. Fluoro-edenite is found water with risk of lung cancer, skin A pooled analysis of case–con-
in Sicily, Italy, in a volcanic area near cancer, and bladder cancer comes trol studies identified a 50% higher
Mount Etna. It was classified by the mainly from populations in areas risk of bladder cancer among indi-
IARC Monographs as carcinogenic with naturally occurring very high ar- viduals with long-term exposure to
to humans (Group 1) [24]. senic content, including Argentina, trihalomethanes in tap water at con-
The evidence that asbestos is Bangladesh, northern Chile, West centrations of about 50 mg/L [31];
carcinogenic to humans is over- Bengal in India, and Taiwan, China such levels are currently observed
whelming, and bans on the produc- [26]. The average exposure to arse- in many high- and middle-income
tion and use of asbestos have been nic varies, and in areas of high ar- countries. Exposure to chlorina-
adopted by many countries, including senic content the concentrations are tion by-products in water through
former asbestos producers such as typically above 100 µg/L. inhalation and dermal absorption
Brazil and Canada. However, the ma- Blackfoot disease is a severe form contributes to the total exposure to
jority of the world’s population lives in of peripheral vascular disease that is trihalomethanes more than expo-
countries where the use of asbestos linked to arsenic exposure from drink- sure through ingestion does, and
is still legal [25]. An asbestos ban ing-water and is endemic in areas of one study identified increased risks
alone, in the absence of thorough Taiwan, China, where well water with of bladder cancer for exposure in
environmental remediation, does not a high concentration of arsenic has showers and baths and for swim-
ensure the prevention of asbestos- been used for many years. Ecological, ming in pools [29]. Recent studies
related disease. Therefore, the long- case–control, and cohort studies of the water exposome examined
lasting legacy of the carcinogenicity have been conducted in those areas, metabolomics, transcriptomics, and
of asbestos is likely to affect countries and excess risks of bladder cancer, proteomics in subjects exposed to
where environmental health preven- lung cancer, skin cancer, and other disinfection by-products and identi-
tive interventions are less stringent. cancer types have been consistent- fied novel biological pathways and
ly found in both sexes, with an expo- genomic responses indicative of in-
sure–response relationship by years creased risk of cancer [32,33].
Water contaminants of consumption and by concentration
Drinking-water, or water used for of arsenic in well water. In an area of Nitrates, perfluorinated
agricultural or recreational activities, high arsenic exposure in southwest- alkylated substances, and
can be polluted by naturally occurring ern Taiwan, China, a progressive other water contaminants
carcinogenic contaminants or by an- decrease in bladder cancer mortality Nitrate is a widespread contaminant
thropogenic pollutants. The strongest was observed after the installation of in drinking-water. Nitrate levels above
evidence on exposure to water con- a tap-water supply system [27]. the WHO guideline concentration of
120
50 mg/L as nitrate are observed in (cadmium, nickel, and lead), radio- contamination, routes of exposure,
several countries, mainly in ground- nuclides, and tetrachloroethylene, and mortality and morbidity data.
water sources from agricultural areas in relation to risk of bladder cancer. The Superfund Research Program,
where use of nitrogen-containing The evaluation of new contaminants, coordinated by the United States
fertilizers is common. The evaluation such as pharmaceuticals and micro- National Institute for Environmental
of ingested nitrate and nitrite is com- plastics, and of mixtures of agents Health Sciences [39], has provided
plex, because there is an active en- is limited. clues to understanding the health
dogenous nitrogen cycle in humans impact of hazardous waste dumping
that under certain conditions gener- sites, including mechanisms through
Soil
ates N-nitroso compounds, a class of which environmental chemicals may
CHAPTER 2.9
SECTION 2
genotoxic compounds of which many Contamination of the soil may be a contribute to cancer.
are carcinogenic to animals. risk factor for cancer, because car- Estimates of cancer risk for popu-
Exposure to nitrates in drinking- cinogenic agents present in the soil, lations living near contaminated sites
water has been examined in case– either naturally or as a result of hu- are available in a few countries. An
control and cohort studies in relation man activities, may be inhaled (as in example is in Italy, where an epide-
to several cancer types, includ- the case of asbestos or other min- miological surveillance project of 44
ing stomach cancer, oesophageal eral fibres, as previously discussed), sites designated as national priority
cancer, brain cancer, lymphomas, accidentally ingested (especially by contaminated sites has specifically
bladder cancer, colorectal cancer, children playing in direct contact with considered 23 sites served by can-
and breast cancer. Several studies the ground), or absorbed through the cer registries (Fig. 2.9.4). For each
have identified positive associations food chain, as a consequence of their contaminated site, the incidence of
with estimates of nitrate uptake from release from soil into both groundwa- all cancers combined and of 35 can-
water, particularly for stomach can- ter and surface water. cer sites was analysed for the period
cer, but the evidence, overall, is not According to a report by the 1995–2005. In both sexes, an excess
consistent. The IARC Monographs European Joint Research Centre was observed for overall cancer inci-
concluded that there is inadequate [38], there are estimated to be dence (9% in men and 7% in women)
evidence in humans for the carcino- 342 000 sites in European Union as well as for specific cancer sites
genicity of nitrate in drinking-water countries with soil contamination, and [40]. An excess of mesothelioma has
but that ingested nitrate or nitrite only 15% of those sites have been been subsequently demonstrated,
under conditions that result in en- subject to remediation interventions. with an ascertained role of environ-
dogenous nitrosation is probably Industrial activities, including indus- mental, non-occupational exposure
carcinogenic to humans (Group 2A) trial waste disposal and treatment, to asbestos at three sites and to
[34]. A subsequent study suggested are responsible for about two thirds of fluoro-edenite at one site [41,24].
a positive association between wa- the overall contamination. The main Both in the USA and in Europe,
terborne ingested nitrates and risk of contaminants are heavy metals, min- a large proportion of contaminated
colorectal cancer [35]. eral oils, and aromatic hydrocarbons. sites, including those designated as
Perfluorinated alkylated sub- The United States Environmental national priority contaminated sites,
stances are chemicals that are widely Protection Agency has developed are characterized by the presence
used as surfactants and are classi- tools for risk assessment in indus- of hazardous waste, which has
fied as persistent organic pollutants. trially contaminated sites (https:// been dumped, burned, or otherwise
Evidence on perfluorinated alkylated www.epa.gov/risk/superfund-risk- improperly managed. Hazardous
substances in water and risk of can- assessment). waste may be defined, in general
cer is available for perfluorooctanoic A comprehensive public health terms, as non-household waste that
acid, after widespread exposure of assessment encompassing health includes hazardous chemicals (see
residents in the Mid-Ohio Valley, outcome data, including cancer oc- “Hazardous waste and cancer”).
USA, through drinking-water con- currence in affected communities,
taminated by chemical plant emis- is provided by the United States
sions. In this population, increased Agency for Toxic Substances and Food
risks were found for kidney cancer Disease Registry in the Public Health Contaminants can enter the food
and testicular cancer [36]. The IARC Assessment Guidance Manual (www. chain at various stages: during pri-
Monographs classified perfluoroocta- atsdr.cdc.gov/hac/PHAManual/ mary production, transformation, and
noic acid as possibly carcinogenic to toc.html). The Agency for Toxic distribution. Therefore, control is re-
humans (Group 2B) after evaluating Substances and Disease Registry quired at each of these stages. In this
the carcinogenicity of perfluoroocta- investigates the occurrence of a wide context, a priority is prevention of the
noic acid in animals and humans [37]. range of chemical agents in a large occurrence of endocrine disrupters
Few ecological or case–control number of affected communities, and in food.
studies have examined other wa- conducts health assessments con- Endocrine disrupters interfere with
ter contaminants, such as metals sidering the available information on the production, release, metabolic

Fig. 2.9.4. Air pollution at the industrial area of Priolo, in eastern Sicily, Italy, which has the only ascertained environmental
been designated a contaminated site of national priority for remediation. carcinogen (see Chapter 2.5).
For many agents, such as ben-
zene, arsenic, and dioxins, the evi-
dence of carcinogenicity is well es-
tablished in adults but only limited in
children. Nevertheless, many can-
cers in children, like in adults, are
thought to be activated by somatic
mutations. In adults, this is associ-
ated with ageing and long-term expo-
sure to carcinogens; in children, the
rarity of cancers and the difficulties in
evaluating what children might have
been exposed to early in life make it
difficult to establish a causal role of
the environment (https://www.cancer.
gov/types/childhood-cancers).
Compared with adults, children
are more vulnerable to environmental
agents, because of their unique activ-
ity patterns, behaviour, and physiol-
action, and elimination of hormones northern Italy living near a plant that ogy, as well as the immaturity of their
and may act at low doses, with no de- produced polychlorinated biphenyls, organs; in addition, many children –
tectable threshold [42,43]. Endocrine which had contaminated the soil, the especially those living in low-income
disrupters that are present in the en- surface water, and the food chain; af- regions of the world – are involved in
vironment and are involved in cancer ter public health measures were im- hazardous work, such as that involv-
causation include dioxins, furans, plemented, serum concentrations of ing contact with pesticides, and are
polychlorinated biphenyls, various polychlorinated biphenyls decreased exposed to emerging threats such as
solvents, heavy metals, pesticides, significantly [45]. toxic components of electronic waste
cosmetics, plastics, and numerous In the absence of preventive in- (e-waste) [48,49].
chemicals in consumer products. terventions and appropriate commu- Cancer types in children are dif-
Human exposure to persistent nication strategies, vulnerable popu- ferent from those in adults; in children,
organic pollutants and heavy metals lations may experience hazardous the most common cancer types are
occurs mainly from foods of animal exposures (see Chapter 6.8). For leukaemia, lymphoma, and tumours
origin, because of bioaccumulation example, Arctic Indigenous popula- of the central nervous system. This
and biomagnification. Despite the tions, whose traditional diet is based pattern should be further explored,
numerous positive effects of breast- on consumption of the meat of marine with investigation of specific mutation
feeding, which should be promoted, mammals, are thus exposed to po- profiles that are possibly related to
maternal milk can be a carrier of a lybrominated diphenyl ethers, which environmental carcinogens. Several
wide range of toxic chemicals, in- may disrupt thyroid homeostasis [46]. large-scale studies, for example
cluding polychlorinated biphenyls, The main cancer sites for which the International Childhood Cancer
4,4′-dichlorodiphenyltrichloroethane an etiological role of environmental Cohort Consortium, are currently ad-
(DDT) and its metabolites, dioxins, endocrine disrupters has been sug- dressing the issues of carcinogenic
and dibenzofurans. gested are the thyroid, together with risk in children associated with ex-
Plants can also absorb and ac- the breast, testis, and prostate. posure to chemical contaminants and
cumulate carcinogenic chemicals, electromagnetic fields.
such as arsenic, from contaminated
soils (for more details, see [44]). The
Cancer and environment
contribution of pesticides to cancer in children Conclusions
risk deserves special attention (see Cancer is a major cause of death in Environmental exposure to car-
“Pesticides and cancer”). children, and the incidence of child- cinogens is a well-defined and pre-
Public health interventions enforc- hood cancers is increasing worldwide ventable contributor to the global
ing prohibition of consumption of food in both high- and low-income regions cancer burden. The most important
produced at contaminated sites have [47]. However, the causes of childhood environmental cancer risk is from
been shown to be effective in reduc- neoplasms are largely unknown; only breathing polluted air that con-
ing absorption of toxic chemicals. An about 5% of tumours are of heredi- tains known human carcinogens.
example is a study of a community in tary origin, and ionizing radiation is Contamination of water and of
122
Hazardous waste and cancer
The potential adverse health ef- the IARC Monographs approach), to cause adverse health effects in
fects associated with waste man- essentially indicating a decreasing populations living in areas where
agement practices have been ex- gradient of confidence in a causal the waste was dumped, burned,
tensively investigated [1], although link (for more details, see [2]). or not suitably processed. Despite
firm conclusions have not been Limited evidence of an associa- a growing awareness of these is-
reached with respect to cancer risk tion was detected for cancer of the sues, illegal trafficking of hazard-
in terms of causal link or burden of liver, breast, testis, and bladder, and ous waste still occurs, especially
CHAPTER 2.9
SECTION 2
disease. However, the specific is- for non-Hodgkin lymphoma. Among towards low- and middle-income
sue of hazardous waste has been the chemical agents reported in the countries where environmental
the subject of a large body of stud- studies that showed excesses of regulation is still absent or is poorly
ies, and the findings of those stud- bladder cancer were heavy metals, enforced [1].
ies are summarized here. β-hexachlorocyclohexane, benzyl
A systematic review of the chloride, organic sulfur compounds, References
scientific literature on the health chlorobenzenes, sodium sulfide/
1. WHO (2016). Waste and human health:
impact of exposure to hazardous sulfhydrates, and dioxins. The evidence and needs. WHO Meeting
waste for populations living near studies that showed excesses of Report, 5–6 November 2015, Bonn,
dumping sites was conducted for non-Hodgkin lymphoma reported, Germany. Copenhagen, Denmark:
World Health Organization Regional
studies published in 1999–2015 among others, the presence of vinyl
Office for Europe. Available from: http://
[2]. The reliability of the studies was chloride, β-hexachlorocyclohexane, www.euro.who.int/_ _data/assets/pdf_
assessed by evaluating exposure heavy metals, and benzene. file/0003/317226/Waste-human-health-
and outcome assessment in terms Both for breast cancer and Evidence-needs-mtg-report.pdf?ua=1.
of possible bias, random error, for testicular cancer, the hypoth- 2. Fazzo L, Minichilli F, Santoro M,
and confounding. The evaluation esis of an etiological role of en- Ceccarini A, Della Seta M, Bianchi F, et
of the evidence of an association docrine disrupters was discussed. al. (2017). Hazardous waste and health
impact: a systematic review of the scien-
between exposure to hazardous In this context, it should be noted tific literature. Environ Health. 16(1):107.
waste and each health outcome that an excess of one or more ht tps://doi.org/10.118 6/s129 40 - 017-
was assessed on the basis of the hormone-sensitive cancer types 0311-8 PMID:29020961
reliability of the studies, the magni- was recently reported in a study of 3. Benedetti M, Zona A, Beccaloni E,
tude and accuracy of the estimat- contaminated sites in Italy charac- Carere M, Comba P (2017). Incidence
ed association, and concordance terized by the presence of endo- of breast, prostate, testicular, and thy-
between the findings of studies. crine disrupters [3]. roid cancer in Italian contaminated
sites with presence of substances with
The evidence of an association Hazardous waste includes elec- endocrine disrupting properties. Int J
between exposure to hazardous tronic waste (e-waste), the occur- Environ Res Public Health. 14(4):E355.
waste and each health outcome rence of which is increasing rapidly. https://doi.org/10.3390/ijerph14040355
was rated as sufficient, limited, or If hazardous waste is inappropri- PMID:28353667
inadequate (partly derived from ately managed, it has the potential
the food chain as a result of both more vulnerable to environmental Acknowledgement

naturally occurring carcinogens agents. The situation with respect to The authors wish to express their
and anthropogenic pollutants has exposure to environmental carcino- gratitude to Dr Anna Bastone of
been less extensively investigated, gens is currently improving in high- Istituto Superiore di Sanità, Rome,
but such contamination appears to income countries and worsening in Italy, for her most valuable contribu-
significantly affect high-risk popu- low- and middle-income countries, tion to the process of information re-
lations, such as those living near because of different standards of trieval and the editing of this chapter.
industrially contaminated sites. environmental protection and mech-
Compared with adults, children are anisms of economic globalization.

Pesticides and cancer
Laura E. Beane Freeman and Manolis Kogevinas

Pesticides encompass a large application of non-arsenical insecti- Work practices – including the
and diverse number of chemicals cides [3]. Of those, only glyphosate amount and types of pesticides
designed to kill pests, including and malathion are extensively used used – and application methods
weeds, insects, rodents, algae, and today. Several pesticides are clas- vary around the world. Therefore,
moulds, for agricultural, residential, sified as possibly carcinogenic to there is a need for additional large,
and public health purposes. These humans (Group 2B), and even more diversified epidemiological cohort
chemicals make important contribu- are categorized as not classifiable as studies applying modern research
tions to the production and protec- to their carcinogenicity to humans approaches. It is important for fu-
tion of agricultural commodities and (Group 3), largely due to inadequate ture research to also assess the ef-
the control of insect disease vec- evidence in humans, although there fects of environmental exposures,
tors. They also present potential are indicators from animal bioassays because of the widespread use of
hazards to human health. or mechanistic studies that require these chemicals. Future studies
Unlike many other chemical further investigation. should evaluate specific chemicals
agents, pesticides are designed Exposure assessment is a major and mixtures, and consider potential
for release into the environment, challenge in epidemiological studies mechanisms of action to support the
and exposure can occur occupa- of pesticides. Some issues include biological plausibility of the epide-
tionally, through environmental the seasonal nature of many expo- miological observations. Exposome
bystander exposure, and through sures, which may be either indoor or approaches may open up new pos-
ingestion of foods containing pes- outdoor, and the large number and sibilities for research and advanced
ticides or pesticides residues. In types of agents, as well as variabil- risk assessment, bridging toxicology
2012, 2.6 million tonnes (5.8 bility in exposure intensity, duration, and epidemiology.
lion pounds) of pesticide active and frequency, depending on the
ingredients were applied world- application and the purpose. There References
wide (https://www.epa.gov/sites/ are multiple routes of exposure, and 1. Loomis D, Guyton K, Grosse Y, El
production/files/2017- 01/docu pesticide products can include both Ghissasi F, Bouvard V, Benbrahim-Tallaa
L, et al.; International Agency for Research
ments/pesticides-industry-sales- active ingredients and inert ingredi- on Cancer Monograph Working Group
usage-2016_0.pdf). ents such as adjuvants. In addition, (2015). Carcinogenicity of lindane, DDT,
Despite widespread potential ex- most pesticides in use today have and 2,4-dichlorophenoxyacetic acid.
Lancet Oncol. 16(8):891–2. https://doi.
posure, cancer risks associated with short half-lives, which are mea-
org/10.1016/S1470-2045(15)00081-9
long-term exposure to specific pes- sured in days or even hours. Finally, PMID:26111929
ticides are generally not well char- the general population may also be
2. Guyton KZ, Loomis D, Grosse Y, El
acterized. Only one group of pesti- exposed, but exposure assessment Ghissassi F, Benbrahim-Tallaa L, Guha N,
cides (inorganic arsenic compounds, in the general population poses its et al.; International Agency for Research
which are not currently used), one own set of challenges. on Cancer Monograph Working Group
(2015). Carcinogenicity of tetrachlorvin-
pesticide contaminant (the dioxin Because of these and other phos, parathion, malathion, diazinon, and
2,3,7,8-tetrachlorodibenzo-para- challenges, few studies are cur- glyphosate. Lancet Oncol. 16(5):490–1.
dioxin), and two insecticides with rently available that can evaluate https://doi.org/10.1016/S1470 -2045
(15)70134-8 PMID:25801782
limited current usage (lindane and associations between exposure
pentachlorophenol, which is also to specific pesticides and risk of 3. IARC (1991). Occupational exposures in
insecticide application, and some pesti-
used as a biocide) are classified cancer. One study that has ac- cides. IARC Monogr Eval Carcinog Risks
by the IARC Monographs as carci- complished this is the Agricultural Hum. 53:5–586. Available from: http://
nogenic to humans (Group 1). The Health Study in the USA (https:// publications.iarc.fr/71 PMID:1688189
fungicide captafol, the insecticides aghealth.nih.gov/). Another study 4. Levêque-Morlais N, Tual S, Clin B,
4,4′-dichlorodiphenyltrichloroethane that has more recently been evalu- Adjemian A, Baldi I, Lebailly P (2015).
The AGRIculture and CANcer (AGRICAN)
(DDT), malathion, diazinon, and ating pesticides and cancer risk is
cohort study: enrollment and causes of
dieldrin (and aldrin metabolized to the AGRICAN study in France [4]. death for the 2005-2009 period. Int Arch
dieldrin) [1,2], the fumigant ethylene These unique studies provide de- Occup Environ Health. 88(1):61–73.
https://doi.org/10.1007/s00420 - 014-
dibromide, and the herbicide glypho- tailed exposure and outcome infor-
0933-x PMID:24599726
sate are classified as probably car- mation, but they examine specific
cinogenic to humans (Group 2A), work environments in only two ag-
as is occupational exposure in the ricultural regions.
124
References
1. Wild CP (2005). Complementing the genome 10. Andersen ZJ, Stafoggia M, Weinmayr 18. O’Callaghan-Gordo C, Kogevinas M,
with an “exposome”: the outstanding chal- G, Pedersen M, Galassi C, Jørgensen Cirach M, Castaño-Vinyals G, Aragonés N,
lenge of environmental exposure measure- JT, et al. (2017). Long-term exposure Delfrade J, et al. (2018). Residential prox-
ment in molecular epidemiology. Cancer to ambient air pollution and incidence imity to green spaces and breast cancer
Epidemiol Biomarkers Prev. 14(8):1847–50. of postmenopausal breast cancer in 15 risk: the multicase-control study in Spain
https://doi.org/10.1158/1055-9965.EPI-05- European cohorts within the ESCAPE (MCC-Spain). Int J Hyg Environ Health.
0456 PMID:16103423 project. Environ Health Perspect. 221(8):1097–106. https://doi.org/10.1016/j.
2. Vineis P, Chadeau-Hyam M, Gmuender 125(10):107005. https://doi.org/10.1289/ ijheh.2018.07.014 PMID:30076044
CHAPTER 2.9
SECTION 2
H, Gulliver J, Herceg Z, Kleinjans J, et EHP1742 PMID:29033383
19. WHO (2016). Burning opportunity: clean
al.; EXPOsOMICS Consortium (2017). 11. Andersen ZJ, Pedersen M, Weinmayr G, household energy for health, sustainable
The exposome in practice: design of the
Stafoggia M, Galassi C, Jørgensen JT, et development, and wellbeing of women
EXPOsOMICS project. Int J Hyg Environ
al. (2018). Long-term exposure to ambient and children. Geneva, Switzerland: World
Health. 220(2 Pt A):142–51. https://
air pollution and incidence of brain tumor: Health Organization. Available from: https://
doi.org /10.1016 / j.ijheh. 2016.0 8 .0 01
the European Study of Cohorts for Air www.who.int/airpollution/publications/
PMID:27576363
Pollution Effects (ESCAPE). Neuro Oncol. burning-opportunities/en/.
3. Herceg Z, Ghantous A, Wild CP, Sklias A, 20(3):420–32. https://doi.org/10.1093/
Casati L, Duthie SJ, et al. (2018). Roadmap neuonc/nox163 PMID:29016987 20. Quansah R, Semple S, Ochieng CA,
for investigating epigenome deregula- Juvekar S, Armah FA, Luginaah I, et al.
tion and environmental origins of cancer. 12. Pedersen M, Andersen ZJ, Stafoggia M, (2017). Effectiveness of interventions to
Int J Cancer. 142(5):874–82. https://doi. Weinmayr G, Galassi C, Sørensen M, et al. reduce household air pollution and/or im-
org/10.1002/ijc.31014 PMID:28836271 (2017). Ambient air pollution and primary prove health in homes using solid fuel in
liver cancer incidence in four European co- low-and-middle income countries: a sys-
4. Sarigiannis D, Karakitsios SP (2018). horts within the ESCAPE project. Environ tematic review and meta-analysis. Environ
Addressing complexity of health impact Res. 154:226–33. https://doi.org/10.1016/j. Int. 103:73–90. https://doi.org/10.1016/j.
assessment in industrially contami- envres.2017.01.006 PMID:28107740 envint.2017.03.010 PMID:28341576
nated sites via the exposome paradigm.
Epidemiol Prev. 42(5–6S1):37–48. https:// 13. Turner MC, Krewski D, Diver WR, Pope 21. Ferrante D, Mirabelli D, Tunesi S, Terracini
doi.org/10.19191/EP18.5-6.S1.P037.086 CA 3rd, Burnett RT, Jerrett M, et al. B, Magnani C (2016). Authors’s response:
PMID:30322234 (2017). Ambient air pollution and can- Pleural mesothelioma and occupational
cer mortality in the Cancer Prevention and non-occupational asbestos exposure:
5. Vineis P, Fecht D (2018). Environment,
Study II. Environ Health Perspect. a case-control study with quantitative risk
cancer and inequalities – the urgent need
125(8):087013. https://doi.org/10.1289/ assessment. Occup Environ Med. 73(10):
for prevention. Eur J Cancer. 103:317–26.
EHP1249 PMID:28886601 713–4. https://doi.org/10.1136/oemed-
PMID:29903684 14. WHO (2006). Air quality guidelines. 2016-103851 PMID:27298458
Global update 2005. Particulate matter, 22. Fazzo L, Minelli G, De Santis M, Bruno
6. WHO (2017). Declaration of the Sixth
Ministerial Conference on Environment ozone, nitrogen dioxide and sulfur diox- C, Zona A, Conti S, et al. (2018).
and Health, Ostrava, Czech Republic, ide. Copenhagen, Denmark: World Health Epidemiological surveillance of mesothe-
13–15 June 2017. Available from: http:// Organization Regional Office for Europe. lioma mortality in Italy. Cancer Epidemiol.
www.euro.who.int/en/media-centre/events/ Available from: http://www.euro.who.int/__ 55:184–91. https://doi.org/10.1016/j.canep.
events/2017/06/sixth-ministerial-confer data/assets/pdf_file/0005/78638/E90038. 2018.06.010 PMID:29990795
ence-on-environment-and-health/docu pdf.
mentation/declaration-of-the-sixth-ministeri 23. Ortega-Guerrero MA, Carrasco-Núñez G,
15. Burnett R, Chen H, Szyszkowicz M, Fann
al-conference-on-environment-and-health. Barragán-Campos H, Ortega MR (2015).
N, Hubbell B, Pope III CA, et al. (2018).
High incidence of lung cancer and malig-
7. GBD 2017 Risk Factor Collaborators (2018). Global estimates of mortality associated
nant mesothelioma linked to erionite fibre
Global, regional, and national comparative with long-term exposure to outdoor fine
exposure in a rural community in Central
risk assessment of 84 behavioural, envi- particulate matter. Proc Natl Acad Sci U S
Mexico. Occup Environ Med. 72(3):216–8.
ronmental and occupational, and metabolic A. 115(38):9592–97. https://doi.org/10.1073/
https://doi.org/10.1136/oemed-2013-101957
risks or clusters of risks for 195 countries and pnas.1803222115 PMID:30181279
PMID:25231672
territories, 1990–2017: a systematic analy-
sis for the Global Burden of Disease Study 16. Mostafavi N, Vermeulen R, Ghantous A,
24. IARC (2017). Fluoro-edenite. Some nano-
2017. Lancet. 392(10159):1923–94. https:// Hoek G, Probst-Hensch N, Herceg Z, et al.
materials and some fibres. IARC Monogr
doi.org/10.1016/S0140-6736(18)32225-6 (2018). Acute changes in DNA methylation
Eval Carcinog Risks Hum. 111:215–42.
PMID:30496105 in relation to 24 h personal air pollution ex-
Available from: http://publications.iarc.fr/552.
posure measurements: a panel study in four
8. IARC (2016). Outdoor air pollution. IARC European countries. Environ Int. 120:11–21. 25. Marsili D, Terracini B, Santana VS,
Monogr Eval Carcinog Risks Hum. 109:1– https://doi.org/10.1016/j.envint.2018.07.026 Ramos-Bonilla JP, Pasetto R, Mazzeo
448. Available from: http://publications.iarc. PMID:30055357 A, et al. (2016). Prevention of asbestos-
fr/538.
related disease in countries currently
17. James P, Hart JE, Banay RF, Laden F
9. Raaschou-Nielsen O, Andersen ZJ, Beelen using asbestos. Int J Environ Res Public
(2016). Exposure to greenness and mor-
R, Samoli E, Stafoggia M, Weinmayr G, et Health. 13(5):E494. https://doi.org/10.3390/
tality in a nationwide prospective cohort
al. (2013). Air pollution and lung cancer in- ijerph13050494 PMID:27187433
study of women. Environ Health Perspect.
cidence in 17 European cohorts: prospec-
124(9):1344–52. https://doi.org/10.1289/ 26. IARC (2004). Some drinking-water disin-
tive analyses from the European Study of
Cohorts for Air Pollution Effects (ESCAPE). ehp.1510363 PMID:27074702 fectants and contaminants, including ar-
Lancet Oncol. 14(9):813–22. https://doi. senic. IARC Monogr Eval Carcinog Risks
org/10.1016/S1470 -20 45(13)70279 -1 Hum. 84:1–477. Available from: http://
PMID:23849838 publications.iarc.fr/102 PMID:15645577

27. Su CC, Lu JL, Tsai KY, Lian IeB (2011). 35. Espejo-Herrera N, Gràcia-Lavedan E, Boldo 44. Mancini FR, Busani L, Tait S, La Rocca C
Reduction in arsenic intake from water has E, Aragonés N, Pérez-Gómez B, Pollán M, (2016). The relevance of the food produc-
different impacts on lung cancer and blad- et al. (2016). Colorectal cancer risk and ni- tion chain with regard to the population ex-
der cancer in an arseniasis endemic area trate exposure through drinking water and posure to chemical substances and its role
in Taiwan. Cancer Causes Control. 22(1): diet. Int J Cancer. 139(2):334–46. https:// in contaminated sites. Ann Ist Super Sanita.
101–8. https://doi.org/10.1007/s10552-010- doi.org/10.1002/ijc.30083 PMID:26954527 52(4):505–10. https://doi.org/10.4415/ANN_
9679-2 PMID:21052815 16_04_08 PMID:27999220
36. Barry V, Winquist A, Steenland K (2013).
28. Baris D, Waddell R, Beane Freeman Perfluorooctanoic acid (PFOA) exposures 45. Raffetti E, Speziani F, Donato F, Leonardi
LE, Schwenn M, Colt JS, Ayotte JD, et and incident cancers among adults liv- L, Orizio G, Scarcella C, et al. (2017).
al. (2016). Elevated bladder cancer in ing near a chemical plant. Environ Health Temporal trends of polychlorinated biphe-
northern New England: the role of drink- Perspect. 121(11–12):1313–8. https://doi.
nyls serum levels in subjects living in a
ing water and arsenic. J Natl Cancer Inst. org/10.1289/ehp.1306615 PMID:24007715
highly polluted area from 2003 to 2015: a
108(9):djw099. https://doi.org/10.1093/jnci/ 37. IARC (2017). Perfluorooctanoic acid. Some follow-up study. Int J Hyg Environ Health.
djw099 PMID:27140955 chemicals used as solvents and in polymer 220(2 Pt B):461–7. https://doi.org/10.1016/
manufacture. IARC Monogr Eval Carcinog j.ijheh.2017.01.002 PMID:28108193
29. Villanueva CM, Cantor KP, Grimalt JO,
Risks Hum. 110:37–110. Available from:
Malats N, Silverman D, Tardon A, et al. 46. Byrne SC, Miller P, Seguinot-Medina S,
http://publications.iarc.fr/547.
(2007). Bladder cancer and exposure to wa- Waghiyi V, Buck CL, von Hippel FA, et al.
ter disinfection by-products through ingestion, 38. van Liedekerke M, Prokop G, Rabl- (2018). Associations between serum po-
bathing, showering, and swimming in pools. Berger S, Kibblewhite M, Louwagie G lybrominated diphenyl ethers and thyroid
Am J Epidemiol. 165(2):148–56. https://doi. (2014). Progress in the management of hormones in a cross sectional study of a
org/10.1093/aje/kwj364 PMID:17079692 contaminated sites in Europe. Report remote Alaska Native population. Sci Rep.
EUR 26376. Luxembourg: Publications 8(1):2198. https://doi.org/10.1038/s41598-
30. Beane Freeman LE, Cantor KP, Baris D, Office of the European Union. https://doi.
Nuckols JR, Johnson A, Colt JS, et al. 018-20443-9 PMID:29396447
org/10.2788/4658
(2017). Bladder cancer and water dis- 47. Steliarova-Foucher E, Colombet M, Ries
infection by-product exposures through 39. Landrigan PJ, Wright RO, Cordero JF, LAG, Moreno F, Dolya A, Bray F, et al.;
multiple routes: a population-based case- Eaton DL, Goldstein BD, Hennig B, IICC-3 contributors (2017). International
control study (New England, USA). Environ et al. (2015). The NIEHS Superfund
incidence of childhood cancer, 2001-
Health Perspect. 125(6):067010. https://doi. Research Program: 25 years of trans-
10: a population-based registry study.
org/10.1289/EHP89 PMID:28636529 lational research for public health.
Lancet Oncol. 18(6):719–31. https://doi.
Environ Health Perspect. 123(10):909–
31. Costet N, Villanueva CM, Jaakkola JJ, org/10.1016/S1470 -2045(17)30186 - 9
18. https://doi.org/10.1289/ehp.1409247
Kogevinas M, Cantor KP, King WD, et al. PMID:25978799 PMID:28410997
(2011). Water disinfection by-products
40. Comba P, Ricci P, Iavarone I, Pirastu R, 48. WHO (2017). Inheriting a sustainable
and bladder cancer: is there a European
Buzzoni C, Fusco M, et al.; ISS-AIRTUM world? Atlas on children’s health and the
specificity? A pooled and meta-anal-
Working Group for the study of cancer environment. Geneva, Switzerland: World
ysis of European case-control stud-
incidence in contaminated sites (2014). Health Organization. Available from:
ies. Occup Environ Med. 68(5):379–85.
Cancer incidence in Italian contaminated https://www.who.int /ceh/publications/
https://doi.org/10.1136/oem.2010.062703
sites. Ann Ist Super Sanita. 50(2):186–91. inheriting-a-sustainable-world/en/.
PMID:21389011
https://doi.org/10.4415/ANN_14_02_13
49. Laborde A, Tomasina F, Bianchi F, Bruné
32. van Veldhoven K, Keski-Rahkonen P, PMID:24968919
MN, Buka I, Comba P, et al. (2015).
Barupal DK, Villanueva CM, Font-Ribera
41. Binazzi A, Marinaccio A, Corfiati M, Children’s health in Latin America: the influ-
L, Scalbert A, et al. (2018). Effects of ex-
Bruno C, Fazzo L, Pasetto R, et al. (2017). ence of environmental exposures. Environ
posure to water disinfection by-products
Mesothelioma incidence and asbestos ex- Health Perspect. 123(3):201–9. https://doi.
in a swimming pool: a metabolome-wide
posure in Italian national priority contami- org/10.1289/ehp.1408292 PMID:25499717
association study. Environ Int. 111:60–70.
nated sites. Scand J Work Environ Health.
https://doi.org/10.1016/j.envint.2017.11.017 43(6):550–9. https://doi.org/10.5271/
PMID:29179034 sjweh.3676 PMID:28985440
33. Espín-Pérez A, Font-Ribera L, van 42. Bergman A, Heindel JJ, Jobling S, Kidd
Veldhoven K, Krauskopf J, Portengen L, KA, Zoeller RT, editors (2013). State of the
Chadeau-Hyam M, et al. (2018). Blood science of endocrine disrupting chemicals
transcriptional and microRNA responses 2012. Geneva, Switzerland: World Health
to short-term exposure to disinfection by- Organization. Available from: https://www.
products in a swimming pool. Environ Int. who.int/ceh/publications/endocrine/.
110:42–50. https://doi.org/10.1016/j.envint.
2017.10.003 PMID:29122314 43. Vandenberg LN, Colborn T, Hayes TB,
Heindel JJ, Jacobs DR Jr, Lee DH, et al.
34. IARC (2010). Ingested nitrate and nitrite, (2012). Hormones and endocrine-disrupt-
and cyanobacterial peptide toxins. IARC ing chemicals: low-dose effects and non-
Monogr Eval Carcinog Risks Hum. 94:1– monotonic dose responses. Endocr Rev.
448. Available from: http://publications.iarc. 33(3):378–455. https://doi.org/10.1210/er.
fr/112 PMID:21141240 2011-1050 PMID:22419778
126
2.10
Occupation
The need for continuing vigilance
Jack Siemiatycki Robert D. Daniels (reviewer) Elizabeth Ward (reviewer)
CHAPTER 2.10
SECTION 2
Lesley Rushton Mary K. Schubauer-Berigan (reviewer) Elizabeth A. Whelan (reviewer)
known causes of human cancer carcinogens also occur in the gen-

SUMMARY were occupational circumstances eral environment (see Chapter 2.9)
[1]. In most such instances of in- and/or in the course of using con-
●● To date, 38 occupational agents creased risk, the relevant informa- sumer products, and, reciprocally,
and 12 occupational expo- tion concerned a particular occu- most environmental exposures and
sure circumstances have been pation or industry, with little or no those associated with using certain
classified as carcinogenic to information that enabled risk to be consumer products, including medi-
humans, and 41 occupational attributed to particular chemicals. cations, foods, and others, also oc-
agents and 6 occupational ex- Since then, many more causes cur in some occupational context.
posure circumstances have of cancer have been identified, both For instance, whereas exposures to
been classified as probably car- occupational and non-occupation- tobacco smoke, solar radiation, and
cinogenic to humans. al. However, even today occupa- immunosuppressive medications are
tional carcinogens make up a large generally not identified as occupa-
●● Workplace exposure to several
fraction of all known human carcin- tional exposures, there are people
well-recognized carcinogens,
ogens. Although the discovery of whose occupation results in them
such as asbestos, polycyclic
occupational carcinogens provides being in contact with these agents
aromatic hydrocarbons, heavy a means for preventing occupation-
metals, diesel engine exhaust, to a degree that would not otherwise
al cancer, the potential benefit of occur. Also, whereas asbestos, ben-
and silica, is still widespread. such discoveries goes beyond the zene, diesel engine exhaust, and
●● The proportion of cancer cases factory walls, because most occu- radon gas are considered to be oc-
attributable to occupational car- pational carcinogens are also found cupational carcinogens, exposure to
cinogens may be substantial. in the general environment and in these agents is also experienced by
consumer products, sometimes at
●● Prevention of occupational can- the general population, and indeed
concentrations as high as those en-
cer is feasible, and during re- many more people are probably ex-
countered in the workplace.
cent decades there have been posed to these substances in the
many successful regulations course of day-to-day life than are ex-
and programmes to eliminate
Specifying occupational posed at work.
or reduce exposure to carcino-
carcinogens Given the definitional ambiguity,
gens in the workplace, particu- This chapter includes tables list- the following operational conven-
larly in high-income countries. ing established and probable oc- tion is adopted here: a carcinogen
cupational carcinogens, as well as is considered to be “occupational”
●● Little information is available on the occupations and industries in if there is significant human expo-
occupational cancer risk in low- which exposure to them occurs and sure to the agent in the workplace,
income countries, but it can be their target organs. Although it may in terms of either prevalence or
reasonably expected to become seem simple, drawing up an unam- level of exposure, and/or if the main
a large problem in the future. biguous list of occupational carcin- epidemiological studies that led
ogens is challenging [2,3]. to the identification of an elevated
The first source of ambiguity risk of cancer were undertaken
Until the recognition in the 1950s is the definition of an occupational among workers. This operational
of the cancer-causing effects of carcinogen. As mentioned above, definition requires judgement in its
cigarette smoking, almost the only exposures to most occupational implementation.
Chapter 2.10 • Occupation 127

Fig. 2.10.1. If there is a requirement to continue using and distributing a chemical

known to be carcinogenic to humans, specifically in an occupational context, a range
of preventive measures may be implemented, in this case illustrated by warning signs
FUNDAMENTALS
concerning benzene in the USA.
■■ It has long been recognized
that a large fraction of known
human carcinogens are agents
that are found in the workplace.
■■ Recognized carcinogens
include chemical, physical, and
biological agents of various
families of agents. Important
occupational carcinogens
are polycyclic aromatic
hydrocarbons, aromatic
amines, certain metals
involved in smelting and
related work, and dusts that
involve exposure to asbestos
and crystalline silica.
■■ Some of the most frequent
cancer types for which excess
risk has been observed from
one or more occupational
carcinogens are lung cancer,
bladder cancer, and skin cancer.
Another source of ambiguity de- who were studied and other work- ■■ Among the challenges in
rives from the nature of those oc- ers in the same occupation but in discovering occupational
cupations, circumstances, and in- different times or places. carcinogens is the fact that
dustries that have been determined there is typically a long time
to involve increased risk of cancer, Occupational agents period between exposure
although the responsible agent has
or exposure circum- to carcinogens and onset
not been identified. Examples are
work as a painter, as a hairdresser,
stances evaluated as of cancer, and therefore
carcinogenic or probably information is required about
or in aluminium production. Such workers’ exposures many years
determinations have somewhat dif- carcinogenic
before the onset of cancer.
ferent implications from the deter- The IARC Monographs provide au-
minations that a particular chemi- thoritative information for compiling ■■ Prevention of occupational
cal, or related chemicals, confers a list of occupational carcinogens cancer can be achieved
an excess risk, as is the case for [4]. The objective of the Monographs through the use of less-
benzene and nickel compounds. programme, which has been oper- hazardous materials,
A determination of carcino- ating since 1971, is to publish criti- engineering controls, optimal
genicity of a specified chemical is cal reviews of epidemiological, ex- procedures and training, and
a statement about the properties of perimental, and mechanistic data the use of personal protective
that chemical that are invariant in on carcinogenicity for chemicals, equipment, together with
time and place; conditional on the groups of chemicals, industrial pro- the monitoring of exposure
level of exposure to the agent, the cesses, other complex mixtures, levels. Such measures may be
chemical or chemicals should al- physical agents, and biological supported by regulation.
ways be considered to be capable agents to which humans are known
of causing cancer. A determination to be exposed, and to evaluate data
that a given occupation involves indicative of carcinogenicity. logical data primarily involve occupa-
a carcinogenic risk does not have Expert Working Groups are con- tional exposure. IARC Monographs
such a universal quality. Cancer vened to evaluate all relevant data. As evaluations are respected worldwide
risks associated with an occupation of 2018, 123 Monographs meetings and are widely used.
or industry may well change if there have been held and more than 1000 A review was performed of all
are differences in technologies or agents have been evaluated, includ- Monographs that were based on the
processes between the workers ing many for which relevant epidemio- 125 meetings held up to November
128
Fig. 2.10.2. Workers in Kolkata, India, tend a furnace in the course of producing at-risk occupational circumstances
fertilizer and fish feed. “Dirty” workplaces are still the norm in many countries. (food frying and shift work). Most
of the agents listed in Table 2.10.2
are carcinogenic in experimental
animals, with little or no epidemio-
logical evidence to confirm or con-
tradict the evidence in animals. For
a few of the agents, including night
shift work, lead compounds, and
creosotes, there is a reasonable
CHAPTER 2.10
SECTION 2
body of epidemiological evidence.
However, the studies in humans and
in experimental animals, taken to-
gether, provide limited evidence of
carcinogenicity to humans by IARC
Monographs criteria. The relevant
epidemiological evidence is not suf-
ficient, because bias, confounding,
or chance cannot be excluded as
contributing to the association that is
evident, or because different studies
provide conflicting results.
The family of polycyclic aro-
matic hydrocarbons poses a par-
ticular challenge. This class of
chemicals includes several potent
experimental carcinogens, such as
benzo[a]pyrene. However, humans
2019. Table 2.10.1 lists 50 occupa- clude ionizing radiation and asbes- are always exposed to mixtures of
tional agents, occupations, and in- tos, which are each associated with polycyclic aromatic hydrocarbons;
dustries that have been classified as multiple target organs. several sources of such mixtures
carcinogenic to humans (Group 1). are indicated in Tables 2.10.1 and
Among the high-risk occupa-
2.10.2, including coal tars, soot, and
The table explicitly distinguishes tions and industries shown in the
creosotes. Because of the difficulty
between 38 chemical or physical second part of Table 2.10.1, most
of isolating the impact of specific
agents and 12 occupations and in- are industries in which the number
polycyclic aromatic hydrocarbons
dustries that involve an increased of workers is quite small, at least in
in exposure assessment, it is diffi-
risk of cancer but for which the re- high-income countries. However,
cult to evaluate human cancer risks
sponsible agent has not been speci- one occupational group – painters –
associated with individual members
fied. The table also indicates which stands out as an occupation that
of this family. Only for benzo[a]py
agents have been added to the list of is very prevalent. The excess risk rene has the evidence warranted
Group 1 agents since 2014. of bladder cancer among painters an evaluation of carcinogenic to hu-
Some of the carcinogens listed may be due to aromatic amines in mans (Group 1), based on mecha-
occur naturally (e.g. wood dust, so- paints, and the excess risk of lung nistic data taken together with other
lar radiation), whereas some are cancer may be due to exposures to available evidence, but there are
anthropogenic (e.g. 1,3-butadiene, asbestos or silica in the construc- probably more individual polycyclic
vinyl chloride). Some are single tion industry. aromatic hydrocarbons that are car-
chemical compounds (e.g. ben- Table 2.10.2 lists occupational cinogenic to humans.
zene, trichloroethylene). Others are agents, occupations, and indus- Loomis et al. [3] recently un-
families of compounds that include tries that have been classified as dertook a similar effort to list oc-
some carcinogens, and still oth- probably carcinogenic to humans cupational carcinogens. They used
ers are mixtures of varying chemi- (Group 2A). The table explicitly slightly different criteria for defining
cal composition (e.g. diesel engine distinguishes between 41 chemi- an agent as occupational, and their
exhaust, mineral oils). Most known cal or physical agents and 4 oc- resulting list is slightly different.
human carcinogens have been es- cupations and industries that have Even when the criteria are identical,
tablished to induce only one type been found to present a probable implementing them requires judge-
of cancer or a few different types risk but for which a causative agent ment, and this can legitimately vary
of cancer; notable exceptions in- has not been identified, and 2 other between experts.

Table 2.10.1. Occupational exposures, occupations, industries, and occupational circumstances classified as carcinogenic to humans
(Group 1) by the IARC Monographs, Volumes 1–125
Agent, occupation, or industry Cancer site or type Where exposure occurs (industry, occupation, or use)
Chemical or physical agent
Acid mists, strong inorganic Larynx, lung Pickling operations, steel and petrochemical industries,
manufacturing of phosphate fertilizer
4-Aminobiphenyl Bladder Rubber
Arsenic and inorganic arsenic Lung, skin, bladder Glass, metals, pesticides
compounds
Asbestos (all forms) Larynx, lung, mesothelioma, Insulation, construction, renovation

ovary
Benzene Leukaemia (acute non- Starter and intermediate in chemical production, solvent
lymphocytic leukaemia, acute
myeloid leukaemia)
Benzidine Bladder Pigments
Benzo[a]pyrene Uncertain Coal liquefaction and gasification, coke production, coke ovens,
coal-tar distillation, roofing, paving, aluminium production, and
others
Beryllium and beryllium Lung Aerospace, metals, nuclear industry

compounds
Bis(chloromethyl)ether; Lung Production of bis(chloromethyl)ether; manufacturing of plastics,

chloromethyl methyl ether resins, and polymers
1,3-Butadiene Leukaemia and/or lymphoma Plastics, rubber
Cadmium and cadmium Lung Pigments, batteries

compounds
Chromium(VI) compounds Lung Metal plating, pigments
Coal-tar pitch Lung, skin Construction, electrodes
1,2-Dichloropropane a
Biliary tract Production of chlorinated chemicals
Diesel engine exhaust Lung Transportation, mining
Ethylene oxide Uncertain Many, including chemical, sterilizing agent
Formaldehyde Nasopharynx, leukaemia Formaldehyde production; plastics, textiles
Ionizing radiation (including Thyroid, leukaemia, salivary Radiology, nuclear industry, underground mining
radon-222 progeny) gland, lung, bone, oesophagus,
stomach, colon, rectum, skin,
breast, kidney, bladder, brain
Leather dust Nasal cavity Shoe manufacture and repair
Lindanea Non-Hodgkin lymphoma Pesticide
4,4′-Methylenebis(2-chloro Uncertain Rubber

aniline) (MOCA)
Mineral oils, untreated or mildly Skin Lubricant

treated
2-Naphthylamine Bladder Pigments
Nickel compounds Nasal cavity, lung, paranasal Metal alloy

sinus
Outdoor air pollutiona Lung Outdoor workers
Pentachlorophenola Non-Hodgkin lymphoma Pesticide
Polychlorinated biphenyls Melanoma of skin Transformer manufacturing, electric power workers

(PCBs) a
Shale oils Skin Lubricant, fuel
Silica dust, crystalline, in the Lung Construction, mining

form of quartz or cristobalite
Solar radiation Skin, melanoma Outdoor work
Soot Lung, skin Chimney sweeps, masons, firefighters
130
Table 2.10.1. Occupational exposures, occupations, industries, and occupational circumstances classified as carcinogenic to humans
(Group 1) by the IARC Monographs, Volumes 1–125 (continued)
Tobacco smoke, second-hand Lung Bars, restaurants, offices
ortho-Toluidine Bladder Pigments
Trichloroethylene Kidney Solvent, dry cleaning
Ultraviolet radiation from Melanoma of eye Welding

weldinga
CHAPTER 2.10
SECTION 2
Vinyl chloride Liver Plastics
Welding fumesa Lung Welders, construction workers
Wood dust Nasal cavity, nasopharynx Wood sawing, construction, furniture
Occupation or industry, without specification of the responsible agent
Acheson processa Lung Production of silicon carbide fibres
Aluminium production Lung, bladder –
Auramine production Bladder –
Coal gasification Lung –
Coal-tar distillation Skin –
Coke production Lung –
Haematite mining (underground) Lung –
Iron and steel founding Lung –
Isopropyl alcohol manufacture Nasal cavity –

using strong acids
Magenta production Bladder –
Painter Bladder, lung, mesothelioma –
Rubber manufacture Stomach, bladder, leukaemia –

a
Added to the list of Group 1 agents since 2014.
Challenges and Because of the long latency be- tion on potential confounding vari-
trends in establishing tween exposure to carcinogens and ables, such as smoking. It would
and understanding onset of cancer, it is necessary to help if physicians or government
be able to ascertain occupational agencies such as cancer registries
lists of occupational circumstances many years before routinely recorded the occupations
carcinogens the onset of cancer. The documen- of patients, but this does not often
Although the lists of occupational tation to enable this to be done is occur. Although epidemiological
carcinogens and associated expo- often fragmentary, unreliable, or and toxicological studies are best
sures shown in Tables 2.10.1 and non-existent. Although large com- suited to the investigation of single
2.10.2 are long, they are not companies may have industrial hygiene agents, the occupational environ-
plete. There are likely to be many data for their workforce, these data ment is complex and shifting and
more occupational carcinogens are often of dubious representa- comprises many agents; this poses
that have not yet been discovered tiveness. Small companies rarely significant difficulties in assessing
or properly documented. For most have any such data. Companies in risks. The statistical power of epi-
occupational circumstances, there low- and middle-income countries demiological studies is often limited
is no relevant epidemiological evi- are even less likely to have and by the size of various workforces;
dence about carcinogenic risk. One maintain such data over long pe- this limitation could sometimes be
of the foremost challenges in occu- riods. Even if long-term exposure overcome by collaborative pooling
pational epidemiology is to reveal data can be obtained, there are of data among investigators.
as-yet-unrecognized carcinogens significant challenges in the statis- In the past, epidemiological re-
and carcinogenic risks. tical modelling of such time-related search on occupational risk factors
There are many obstacles to information. In many occupational has focused largely on occupation-
the discovery and characteriza- cancer studies, it is difficult or im- al exposures associated with “dirty”
tion of occupational carcinogens. possible to obtain reliable informa- industrial environments.

Table 2.10.2. Occupational exposures, occupations, industries, and occupational circumstances classified as probably carcinogenic
to humans (Group 2A) by the IARC Monographs, Volumes 1–125
Chemical or physical agent
Acrylamide – Plastics
Bitumens (combustion products) Lung Roofing
Captafol – Fungicide
α-Chlorinated toluenes – Pigments, chemicals

combined with benzoyl chloride
4-Chloro-ortho-toluidine Bladder Pigments, textiles
Cobalt metal with tungsten Lung Hard-metal production

carbide
Creosotes Skin Wood preserving, brick making
Diazinona – Insecticide
4,4′-Dichlorodiphenyltrichloro – Biocide
ethane (DDT) a
Dichloromethane (methylene – Organic solvent

chloride) a
Dieldrin, and aldrin metabolized Breast Biocide

to dieldrin
Diethyl sulfate – Production of dyes, pigments, textiles
Dimethylcarbamoyl chloride – Production; manufacture of pharmaceuticals; pesticides and dyes
Dimethylformamide a
– Solvent in production of acrylic fibres, plastics, pharmaceuticals,
pesticides, adhesives, synthetic leathers, and surface coatings
1,2-Dimethylhydrazine – Laboratory use only; DNA methylation
Dimethyl sulfate – Used in methylation of phenols, amines, and thiols; plastics,

pharmaceuticals, herbicides
Epichlorohydrin – Plastics
Ethylene dibromide – Fumigant
Glycidol – Pharmaceutical industry
Glyphosate a
Non-Hodgkin lymphoma Herbicide, agriculture
Hydrazinea Lung Production of gases, propellants, pharmaceuticals, pesticides,

solvent
Indium phosphide – Semiconductors
Lead compounds, inorganic Lung, stomach Metals, pigments
Malathiona – Organophosphate insecticide
2-Mercaptobenzothiazole a
– Sulfur vulcanization of rubber
Methyl methanesulfonate – Methylating agent
6-Nitrochrysenea – Transportation, vehicle mechanic
1-Nitropyrene a
– Transportation, vehicle mechanic
2-Nitrotoluene – Production of dyes
Non-arsenical insecticides – Agriculture
Polycyclic aromatic – Combustion of organic matter, coal liquefaction and gasification,

hydrocarbons coke production, coke ovens, coal-tar distillation, roofing, paving,
Cyclopenta[cd]pyrene aluminium production, foundries, steel mills, firefighters, vehicle
Dibenz[a,h]anthracene mechanics
Dibenz[a,j]acridine
Dibenzo[a,l]pyrene
1,3-Propane sultonea – Laboratory use, photographic chemicals, pharmaceuticals,

insecticides, dyes, chemical industry
132
Table 2.10.2. Occupational exposures, occupations, industries, and occupational circumstances classified as probably carcinogenic
to humans (Group 2A) by the IARC Monographs, Volumes 1–125 (continued)
Silicon carbide whiskers a

– Mineral, abrasives
Styrene and styrene-7,8-oxide – Plastics
Tetrabromobisphenol A a
– Fire retardant
Tetrachloroethylene – Solvent
(perchloroethylene)
CHAPTER 2.10
SECTION 2
Tetrafluoroethylenea – Alkylating agent used in production of polymers, non-stick
coatings, resistant tubing
1,2,3-Trichloropropane – General-purpose solvent
Tris(2,3-dibromopropyl) – Plastics, textiles

phosphate
Vinyl bromide – Plastics, textiles
Vinyl fluoride – Production of various polymers, solar panels
Occupation or industry, without specification of the responsible agent
Art glass, glass containers, and Lung, stomach –

pressed ware (manufacture of)
Carbon electrode manufacture Lung –
Hairdressers or barbers Bladder, lung –
Petroleum refining – –
Occupational circumstance, without specification of the responsible agent
Food frying at high temperature – –
Night shift work Breast, prostate, colon, rectum Health care, transportation, services
a
Added to the list of Group 2A agents since 2014.
However, in recent decades oc-

cupational hygiene in many indus- Fig. 2.10.3. This factory worker in Thailand has a degree of protection from occupational
exposures, including gloves to reduce dermal exposure.
tries has improved or different tech-
nology has been adopted such that
the historical risks no longer apply,
at least in high-income countries.
Increasing attention is now be-
ing paid to non-chemical agents
in the work environment. Physical
agents such as solar radiation and
electromagnetic fields have been
investigated, as have behavioural
and ergonomic characteristics of
particular occupations, such as
physical activity and shift work. For
almost all of these risk factors, the
distinction between occupational
and non-occupational exposure is
becoming more blurred.
Industries and occupations are
constantly evolving. Even if we knew
all there was to know about the can-
cer risks in today’s occupational

environments (which we do not), Fig. 2.10.4. Hazmat suits (hazardous materials suits) are an example of personal pro-
continuing to monitor cancer risks in tective equipment, an option to be used to control workplace exposures to occupa-
occupational settings would remain tional carcinogens.
an important activity, because oc-
cupational exposure circumstances
change over time and novel exposure
circumstances may be introduced;
recent examples include video dis-
play terminals and nanoparticles.
Estimates of the
fraction of cancer
that is attributable to
occupational exposures
Estimates have been made in
various countries, using various
methodologies, of the fraction of
cancer that may be attributable to
occupational exposures, and that
could potentially be prevented if
those hazards were eliminated.
In general, it has been estimated
that the fraction of cancer attri
butable to occupational exposures
crease over time). The main cancer 660 000 deaths, and asbestos is
is between 2% and 8% in high-in-
types attributable to occupational the exposure that contributes the
come countries [5]. The estimates
carcinogens were mesothelioma, largest proportion.
vary considerably among different
lung cancer, bladder cancer, breast The WHO Global Burden of
types of cancer.
cancer, non-melanoma skin cancer, Disease Study 2017 estimated
The estimates of occupation-
and sinonasal cancer. Among the that in 2017, about 334 000 cancer
al burden of cancer vary among
main occupational exposures con- deaths were due to occupational
countries, depending on the in-
tributing to this burden were asbes- exposures, and the major contribu-
dustrial profiles of the countries,
tos, shift work (night work), mineral tors were asbestos, silica, and die-
and will change over time as new
oils, solar radiation, silica, diesel sel engine exhaust [9].
occupational carcinogens are dis-
engine exhaust, and the follow- Studies on occupational cancer
covered or the impact of old ones
ing industries: construction, metal burden are influencing the prioriti-
diminishes. The estimates also
working, service industries, mining, zation and development of strate-
vary with the methodology used,
and several manufacturing sectors. gies for risk reduction, galvanizing
including whether the estimates
The total annual economic cost of campaigns to raise awareness of
are based only on established car-
new cases of work-related cancer issues related to occupational can-
cinogens or on both established
in Great Britain in 2010 was esti- cer [10], and encouraging the intro-
and probable carcinogens.
mated to be £12.3 billion, of which duction or reduction of occupational
The most detailed and inten-
98% was due to “human” costs – a limit values. In Europe, a socioeco-
sive effort to date to estimate oc-
monetary value on the effects of nomic health and environmental
cupational burden of cancer was
cancer on quality of life, or loss of impact assessment has already led
conducted in Great Britain [6]. The
to binding occupational exposure
study, which took into account can- life for fatal cancers [7].
limits being set for all 28 European
cer latency, workforce turnover, and The International Labour Orga
Union Member States. Such stud-
changing employment trends and nization and WHO have estimated
ies have also drawn attention to the
life expectancy over time, estimat- that 5–7% of global deaths are at-
inequalities of occupational cancer
ed that 5.3% of all cancers (8.2% tributable to work-related illnesses
burden between different sectors of
in men, 2.3% in women) were attri and occupational injuries, corre-
society [11].
butable to past exposure to occu- sponding to 2.3 million occupation-
pational carcinogens, correspond- related deaths per year, of which the
ing to about 13 600 new cancers majority, 2.0 million, are due to oc- Prevention
per year and about 8000 deaths cupational diseases [8,9]. Overall, The designation of an agent as carci-
per year in Great Britain in 2004 cancer makes up the largest com- nogenic is an important public health
(the numbers are expected to in- ponent (~32%), corresponding to statement, as well as a scientific one.
134
Such a designation, together with Fig. 2.10.5. In this scrapyard associated with “ship breaking”, located north of Chittagong
findings from occupational research, in Bangladesh, workers have no protection from toxic agents they may encounter.
has implications for engineering
and/or industrial hygiene measures
to reduce or eliminate occupational
exposure to the agent.
Approaches to preventing work-
place exposures to occupational
carcinogens and reduction of oc-
cupational cancer include eliminat-
CHAPTER 2.10
SECTION 2
ing the production or use of car-
cinogens and controlling exposure
to below a minimal risk exposure
level, for example an occupational
exposure limit (Table 2.10.3).
Even though older, “dirty” indus-
tries are declining in importance as
a source of employment in high-
income countries, it remains true –
and will for the foreseeable future –
that small companies in all coun-
tries may continue to operate with
older and dirtier technologies and come countries. Furthermore, the Concurrent exposure to mul-
processes without appropriate pre- rapid growth of industry in low- and tiple carcinogens is of concern,
ventive measures. For high-income middle-income countries is often and in some situations a concerted
countries and rapidly industrializing unregulated and has inadequate industry-focused strategy may be
countries, risk reduction strategies, occupational hygiene. needed. Protection measures for a
such as improvement of compliance Effective regulation and control single carcinogen may also simul-
with current occupational exposure measures need to be appropriately taneously reduce exposure from
limits (e.g. for silica exposure) and adapted to different circumstances. others (e.g. measures to reduce
targeting small- and medium-sized For some agents, reduction of ex- general dust); measures to protect
industries, have been demonstrat- posure levels is feasible and ap- against carcinogens will also poten-
ed to be effective (see Chapter propriate; for others, more extreme tially reduce the incidence of non-
6.8) [12]. The problem is more measures, such as banning use, malignant occupational disease,
acute in low- and middle-income may be appropriate. Large numbers such as respiratory ill health.
countries. Some particularly dirty of workers continue to be exposed Monitoring of the workplace can
and dangerous industrial work, to low levels of occupational carcin- rely on various types of approach-
like removing asbestos from ships ogens; some of these workers may es, from industrial hygiene to bio-
that have been decommissioned, well develop cancers as a result of monitoring. Technical advances in
is now being performed in low-in- these exposures. these areas should be encouraged.
Table 2.10.3. Measures to control workplace exposures to occupational carcinogens
Control method Examples of good practice
Elimination Remove the hazard from the workplace, for example change a process so that the chemicals,
materials, or equipment are no longer required.
Substitution Replace a hazardous material or piece of equipment with a less-hazardous one.
Engineering controls Redesign the equipment or process so that the hazard is controlled at its source, for example through
a physical barrier.
Worker education Provide information and training on all workplace carcinogens and the use of appropriate control
methods.
Use information media (e.g. posters, leaflets, data sheets) imaginatively and strategically.
Administrative controls Design and operate effective and reliable processes and activities to minimize exposure.
Provide safe storage, handling, and transportation, and disposal facilities.
Personal protective equipment Use suitable personal protective equipment, for example gloves, coveralls, respirators, hard hats,
safety glasses, high-visibility clothing, and safety footwear.

All stakeholders, including regu- able, or are not available in a form Conclusions
lators, employers, and employees, that facilitates intervention. In addi- Prevention of cancer depends on
should be encouraged to work to- tion, reliable reporting systems for the identification and management
gether on prevention and to develop occupational disease are scarce, of cancer-causing circumstances.
effective policies and procedures. particularly for cancers with long la- The workplace remains an important
Unfortunately, precise and reliable tency. Increased efforts are needed locus for research to identify carcin-
data on the magnitude of risks as- to push for more education on oc- ogens and for mitigating or eliminat-
sociated with different agents, and cupationally related ill health, for ing the impact of carcinogens.
on the nature of dose–response example in medical training and
relationships, are not always avail- more generally.
References
1. Siemiatycki J, Richardson L, Boffetta P 6. Rushton L, Bagga S, Bevan R, Brown 11. Cherrie JW, Hutchings S, Gorman Ng M,
(2006). Occupation. In: Schottenfeld D, TP, Cherrie JW, Holmes P, et al. Mistry R, Corden C, Lamb J, et al. (2017).
Fraumeni JF Jr, editors. Cancer epidemiol- (2010). Occupation and cancer in Prioritising action on occupational car-
ogy and prevention. 3rd ed. New York (NY), Britain. Br J Cancer. 102(9):1428–37. cinogens in Europe: a socioeconomic and
USA: Oxford University Press; pp. 322–54. https://doi.org/10.1038/sj.bjc.6605637 health impact assessment. Br J Cancer.
h t t p s : / / d o i . o r g / 10 .10 9 3 / a c p r o f : o s o / PMID:20424618 117(2):274–81. https://doi.org/10.1038/
9780195149616.003.0018 bjc.2017.161 PMID:28609433
7. Zand M, Rushbrook C, Spencer I, Donald
2. Siemiatycki J, Richardson L, Straif K, K, Barnes A (2016). Costs to Britain of 12. Hutchings S, Cherrie JW, Van Tongeren
Latreille B, Lakhani R, Campbell S, et work-related cancer. Health and Safety M, Rushton L (2012). Intervening to re-
al. (2004). Listing occupational carcino- Executive, Research Report RR1074. duce the future burden of occupational
gens. Environ Health Perspect. 112(15): Available from: http://www.hse.gov.uk/ cancer in Britain: what could work?
1447–59. https://doi.org/10.1289/ehp.7047 research/rrhtm/rr1074.htm. Cancer Prev Res (Phila). 5(10):1213–22.
PMID:15531427
8. Takala J (2015). Eliminating occupa- https://doi.org/10.1158/1940-6207.CAPR-
3. Loomis D, Guha N, Hall AL, Straif K (2018). tional cancer. Ind Health. 53(4):307–9. 12-0070 PMID:22961776
Identifying occupational carcinogens: an https://doi.org/10.2486/indhealth.53-307
update from the IARC Monographs. Occup PMID:26377441
Environ Med. 75(8):593–603. https://
doi.org /10.113 6 /oemed -2017-10 49 4 4 9. GBD 2017 Risk Factor Collaborators
PMID:29769352 (2018). Global, regional, and national
comparative risk assessment of 84 be-
4. IARC (2019). IARC Monographs on the havioural, environmental and occupa-
Evaluation of Carcinogenic Risks to tional, and metabolic risks or clusters of
Humans, Volumes 1–123. Lyon, France: risks for 195 countries and territories,
International Agency for Research on 1990–2017: a systematic analysis for the
Cancer. Available from: http://publications. Global Burden of Disease Study 2017.
iarc.fr/. Lancet. 392(10159):1923–94. https://doi.
org /10.1016/S0140 - 673 6(18) 32225 - 6
5. Purdue MP, Hutchings SJ, Rushton L, PMID:30496105
Silverman DT (2015). The proportion of can-
cer attributable to occupational exposures. 10. IOSH (2016). Institution of Occupational
Ann Epidemiol. 25(3):188–92. https://doi. Safety and Health “No Time to Lose” cam-
org/10.1016/j.annepidem.2014.11.0 09 paign: working together to beat occupation-
PMID:25487971 al cancer. https://www.notimetolose.org.uk/
136
2.11
Pharmaceutical drugs
A current focus on hormones
Lisa Iversen Anssi Auvinen (reviewer)
CHAPTER 2.11
SECTION 2
Michael E. Jones (reviewer)
Both the health benefits (often im- ceptives were classified as possibly
SUMMARY mediate) and the risks of adverse carcinogenic to humans (Group 2B)
outcomes (often associated with (Table 2.11.1).
●● Evaluating any possible cancer dose and duration of treatment, and Most of the evidence about
effects of pharmaceutical drugs experienced at a later time) of using hormonal contraceptives relates
is problematic, even if a drug is pharmaceutical drugs need to be to combined estrogen–progesto-
used by many people, given fully considered by health profes- gen products, and in particular oral
the long surveillance period re- sionals and patients [1]. Evaluating contraceptives (Fig. 2.11.1). Current
quired for any cancer risks or or recent users of combined oral
any possible cancer effects of phar-
benefits to emerge. contraceptives have an increased
maceutical drugs is problematic, risk of breast cancer and cervical
●● Hormonal contraceptives are even if a drug is used by many peo- cancer and, in regions at low risk
often used for prolonged peri- ple, given the long surveillance pe- of hepatitis B virus infection, an in-
ods; the very long-term cancer riod required for any cancer risks or creased risk of liver cancer. Users
effects of combined oral con- benefits to emerge. of combined oral contraceptives
traceptives can now be inves- Over decades, causation of have a reduced risk of ovarian can-
tigated, because the women cancer by pharmaceutical drugs cer; this protective effect increases
who were the first users of has been discovered in a variety with duration of use and persists
these products, in the 1960s, of circumstances. This chapter fo- for many years after stopping use.
are now entering later life. cuses on research during the past Combined oral contraceptives may
also be associated with a reduced
●● Evidence is starting to emerge 5 years, and the central issue has
risk of colorectal cancer, although
about the cancer risks associ- been hormonal agents. no consistent relationship has been
ated with contemporary hor- demonstrated with duration or re-
monal contraceptives, includ- Hormonal contraceptives cency of use.
ing new routes of delivery, new In 2015, an individual participant
Hormonal contraceptives are used,
progestogens, and progesto- meta-analysis of 27 276 women with
often for prolonged periods, to endometrial cancer (see Chapter
gen-only contraceptives.
prevent pregnancy, not as a treat- 5.11) found that use of oral contra-
●● The relationship between meno- ment for a disease. Hormonal con- ceptives for 10–15 years halves
pausal hormone therapy and the traceptives are commonly used – the risk of endometrial cancer, and
risk of cancer of the ovary and every day, at least 100 million that a significant protective effect
colorectum has been examined. women worldwide are using hor- remains more than 30 years after
monal contraception [2]. The IARC stopping use [5]. These effects var-
●● Fertility drugs are being used by
Monographs programme has evalu- ied by histological type: ever use of
increasing numbers of women; oral contraceptives was strongly as-
studies examining the risk of ated the carcinogenic hazards as-
sociated with combined estrogen– sociated with a reduced risk of type
cancer of the breast, ovary, and I and type II endometrial cancer but
endometrium have many meth- progestogen contraceptives [3] and
was not associated with a reduced
odological challenges, particu- progestogen-only contraceptives [4] risk of uterine sarcoma, which is
larly because subfertile women and concluded that there was suffi- a much rarer type. During the 50-
have an inherently increased cient evidence for combined hormo- year period from 1965 to 2014, an
cancer risk independent of any nal contraceptives to be classified as estimated 400 000 cases of en-
fertility treatments. carcinogenic to humans (Group 1), dometrial cancer in women youn-
whereas progestogen-only contra- ger than 75 years were avoided in
Chapter 2.11 • Pharmaceutical drugs 137

Fig. 2.11.1. Most of the evidence about the cancer risks and benefits of hormonal
contraception relates to combined estrogen–progestogen products, and mainly oral
products, such as the contraceptive pills shown here. FUNDAMENTALS
■■ Over decades, a range of
pharmaceutical drugs has been
recognized as causing particu-
lar cancers among the people
using them. Cytotoxic drugs,
either alone or in combination,
may cause second cancers, and
their use must take into account
these and other adverse effects.
■■ Some drugs, for example
diethylstilbestrol and phenace-
tin, have been withdrawn from
widespread use as a result of
cancer causation.
■■ The IARC Monographs pro-
gramme concluded that there
high-income countries as a result of Study in the United Kingdom fol- was sufficient evidence for com-
use of oral contraceptives. bined hormonal contraceptives
lowed up an initial cohort of 46 022
to be classified as carcinogenic
women for up to 44 years and in-
Long-term cancer effects to humans (Group 1), whereas
cluded more than 1.2 million per- progestogen-only contraceptives
The very long-term cancer risks or son-years of observation. It found were classified as possibly carci-
benefits of combined oral contra- that an increased risk of incident nogenic to humans (Group 2B).
ceptives can now be investigated, breast cancer and cervical can-
■■ Most of the evidence about hor-
because the women who were the cer seen in current and recent us- monal contraceptives relates to
first users of these products, in the ers of oral contraception was lost combined estrogen–progesto-
1960s, are now entering the later within approximately 5 years of gen products, and in particular
stages of their lives. stopping use, with no evidence of oral contraceptives.
The most recent findings from an increased risk of either cancer ■■ Current or recent users of
the Nurses’ Health Study in the type in ever users later in life [7]. combined oral contraceptives
USA, after 36 years and 3.6 mil- When risks were stratified by time have an increased risk of breast
lion person-years of follow-up, since last use, ever users had a cancer, cervical cancer, and (in
were that overall ever use of oral reduced risk of endometrial can- regions at low risk of hepatitis B
contraceptives was not associated cer 25–35 years after stopping use virus infection) liver cancer.
with risk of death from cancer of the (incidence rate ratio, 0.58; 99% ■■ Users of combined oral contra-
breast, cervix, uterus/endometrium, CI, 0.38–0.88). The risk of ovarian ceptives have a reduced risk of
or large bowel and rectum [6]. A re- cancer (incidence rate ratio, 0.50; ovarian cancer; this protective
duced risk of death from ovarian 99% CI, 0.29–0.84) and colorectal effect increases with duration
cancer was of borderline statisti- cancer (incidence rate ratio, 0.67; of use and persists for many
cal significance (hazard ratio [HR], 99% CI, 0.49–0.91) was reduced years after stopping use. The
0.86; 95% confidence interval [CI], 35 years or more since last use. risk of colorectal cancer may be
0.74–1.00). However, use of oral If it is assumed that the incidence reduced, although no consis-
contraceptives for 5 years or more rate ratios represent a causal rela- tent relationship has been found
was associated with an increased tionship, approximately one third of with duration or recency of use.
risk of death from breast cancer endometrial cancers and ovarian ■■ The IARC Monographs pro-
(P trend < 0.0001) and a decreased cancers and one fifth of colorectal gramme concluded that estro-
risk of death from ovarian cancer cancers among ever users in this gen-only menopausal hormone
(P trend = 0.002). The increased risk study might have been prevented therapy is associated with cancer
of death from breast cancer dimin- by the use of oral contraceptives. of the endometrium, ovary,
and breast, and that combined
ished with time since last use, with Importantly, the study found no evi-
estrogen–progestogen hormone
no increased risk 10 years or more dence of new cancer risks appear-
therapy is associated with cancer
after stopping use. For risk of death ing later in life among ever users, of the breast and endometrium
from ovarian cancer, no trends were providing strong evidence that most and is unlikely to increase the risk
found by time since last use. women do not expose themselves of colorectal cancer or alter the
The Royal College of General to long-term cancer harm if they risk of ovarian cancer.
Practitioners’ Oral Contraception use oral contraceptives.
138
Table 2.11.1. Summary of hormonal contraceptives, hormone therapy, and fertility drugs and cancer risks
Drug IARC Monographs Cancer site Increased or decreased risk?

evaluation
Combined estrogen–progestogen Carcinogenic to Breast Increased in current or recent users;

oral contraceptives humans (Group 1) evidence emerging of similar risk patterns
associated with contemporary productsa
Cervix Increased in current or recent users
CHAPTER 2.11
SECTION 2
Liver b Increased in current or recent users
Ovary Decreased in current or recent users;

decreased in ever users; persistent reduced
risk many years after stopping use; evidence
emerging of similar risk patterns associated
with contemporary productsa
Endometrium Decreased in ever users; persistent reduced

risk many years after stopping use
Colorectum May be decreased in ever users; no

consistent relationship shown for duration or
recency of use
Progestogen-only contraceptives Possibly carcinogenic Breast Evidence emerging of increased risk

to humans (Group 2B) associated with current or recent use
of contemporary oral productsa and the
levonorgestrel-releasing intrauterine system
Ovary Mixed evidence, with one study finding no

reduced risk associated with contemporary
productsa; others found a reduced risk
associated with the levonorgestrel-releasing
intrauterine system but did not examine risk
in exclusive users
Estrogen-only hormone therapy Carcinogenic to Endometrium Increased

humans (Group 1)
Ovary Increased
Breast Increased
Combined estrogen–progestogen Carcinogenic to Breast Increased

hormone therapy humans (Group 1)
Endometrium Increased (risk of endometrial cancer
reduced proportionally by number of days
per month that progestogens are added to
regimen)
Ovary Increased (based on prospective studies)

and associated with recency of use
Colorectum Possible reduced risk, but current evidence

insufficient
Fertility drugs (can include Not assessed Breast No association, but possible concerns raised
clomiphene citratec, gonadotropins, about clomiphene citrate. Lack of good-
gonadotropin-releasing hormone quality evidence
agonists and antagonists, and
human chorionic gonadotropin) Ovary No evidence of an association; possible
increased risk of borderline tumours. Lack of
good-quality evidence
Endometrium Lack of good-quality evidence
a
Hormonal contraceptives available on the market during 1995–2014.
b
In regions at low risk of hepatitis B virus infection.
c
IARC Monographs evaluation: not classifiable as to its carcinogenicity to humans (Group 3).

The National Institutes of Health- The relative risk of breast can- Fig. 2.11.2. Evidence is starting to emerge
AARP Diet and Health Study of cer among current or recent users about the cancer risks associated with
196  536 mostly postmenopaus- of combined oral contraceptives contemporary hormonal contraceptives,
including new routes of delivery, new pro-
al women at recruitment reported was 1.19 (95% CI, 1.13–1.26). The gestogens, and progestogen-only contra-
reductions in the risk of incident strength of the association increased ceptives such as the levonorgestrel-re
ovarian cancer (HR, 0.74; 95% with duration of use. The relative risk leasing intrauterine system, shown here.
CI, 0.65–0.84), endometrial can- estimate was similar to that previ-
cer (HR, 0.78; 95% CI, 0.70–0.86), ously reported [11] but, importantly,
and any cancer (HR, 0.97; 95% CI, was based on contraceptive prod-
0.95–0.99) among users of oral ucts available since 1995, whereas
contraceptives [8]. For longer dura- the earlier estimate was based on
tions of use, the risk reductions were products prescribed in the 1980s or
stronger for both ovarian cancer and earlier. There were no major differ-
endometrial cancer. The effects of ences between the risk associated
time since last use (recency) were with combined oral contraceptives
not examined. An increased risk of containing different progestogens.
breast cancer was of borderline sta- The same study also examined
tistical significance (HR, 1.04; 95% progestogen-only contraceptives
CI, 1.00–1.09) and was not associ- and found that both the levonor-
ated with duration of use. gestrel-only pill and the levonor-
A study that combined data gestrel-releasing intrauterine sys-
from 310  290 women who were tem (LNG-IUS) (Fig. 2.11.2) were
participants in three large cohorts associated with an increased risk
in the USA (the National Institutes of breast cancer. The absolute in-
of Health-AARP Diet and Health crease in the risk of breast cancer in
Study, the California Teachers current and recent users was small:
Study, and the Women’s Health 13 (95% CI, 10–16) per 100 000
Initiative) found that the reduction in person-years, or 1 extra breast
the risk of epithelial ovarian cancer cancer for every 7690 women using in 1995–2014 investigated use of
per 5 years of oral contraceptive use hormonal contraception for 1 year. contemporary combined hormonal
did not wane with age (50–64 years: The results of the study in contraceptives and risk of ovarian
HR, 0.88; 95% CI, 0.80–0.98; 65– Denmark concurred with those of a cancer [15]. Both current or recent
74 years: HR, 0.82; 95% CI, 0.74– study of women with menorrhagia use (relative risk [RR], 0.58; 95%
0.91; ≥ 75 years: HR, 0.85; 95% CI, aged 30–49 years, which investigat- CI, 0.49–0.68) and former use (RR,
0.71–1.02; Pinteraction = 0.79) [9]. ed the cancer risks of the LNG-IUS 0.77; 95% CI, 0.66–0.91) of hor-
In all of these studies [6–9], the using national registries in Finland monal contraceptives was associ-
combined oral contraceptives as- [12]. The study in Finland found a ated with a reduced risk of ovarian
sessed usually contained a higher higher-than-expected incidence of cancer; this effect was directly as-
dose of estrogen combined with an breast cancer (standardized inci- sociated with duration of use and
older progestogen compared with dence ratio, 1.19; 95% CI, 1.13–1.25) persisted for several years after
the products that are currently avail- among users of the LNG-IUS. The stopping use. There was little evi-
able. Evidence is starting to emerge users had an increased risk of both dence of major differences in risk
about the cancer risks associated ductal and lobular breast cancer, estimates by the progestogen con-
with contemporary hormonal con- and the risk estimates were high- tent of combined oral contracep-
traceptives, including new routes est in women who had purchased tives or by tumour type. There was
of delivery, new progestogens, and the contraceptive at least twice [13]. no evidence of a protective effect
progestogen-only contraceptives. These results contradict those of for ovarian cancer associated with
the Norwegian Women and Cancer use of progestogen-only contracep-
Contemporary hormonal Study, which did not find an in- tives, although the evidence was
contraceptives creased risk of breast cancer in ever limited because few women were
A study of 1 797 932 women living or current users of the LNG-IUS, al- exclusive users of progestogen-
in Denmark and aged 15–49 years though few participants in that study only products.
in 1995–2012 examined the risk of were younger than 46 years and the Both the Finnish study [12,13]
breast cancer associated with cur- mean time since stopping use was and the Norwegian study [14]
rently available hormonal contra- 7.5 years [14]. found a decreased risk of ovar-
ceptives [10]. During 19.6 million Another recent study of more ian cancer and endometrial cancer
person-years of follow-up, 11 517 than 1.8 million women living in among ever users of the LNG-IUS.
incident breast cancers occurred. Denmark and aged 15–49 years Although the studies adjusted for
140
some possible confounding factors, ovarian cancer (RR, 1.37; 95% CI, occurred [18]. Current users of any
neither was able to calculate risks 1.27–1.48). systemic hormones (all types of
among exclusive users of this pro- The risk was highest among hormone therapy) had a decreased
gestogen-only product. Therefore, women last recorded as current us- risk of colon cancer (RR, 0.84; 95%
it is possible that the findings were ers (RR, 1.41; 95% CI, 1.32–1.50). CI, 0.78–0.90) and of rectal cancer
due to a persisting protective ef- Even relatively short duration of use (RR, 0.87; 95% CI, 0.79–0.95) com-
fect from previous use of combined (< 5 years of current use) was asso- pared with never users. A stronger
oral contraceptives. Such limitations ciated with an increased risk (RR, reduction in the risk of colon can-
highlight the need for more studies 1.43; 95% CI, 1.31–1.56). The risk cer was found in long-term current
of the possible cancer effects of appeared to decline with time since users with 10 years or more of use
CHAPTER 2.11
SECTION 2
progestogen-only contraceptives. stopping use of hormone therapy, (RR, 0.72; 95% CI, 0.61–0.85). Use
although there was the suggestion of tibolone, vaginal estrogen, and
of a small increased risk remaining transdermal combined preparations
Menopausal hormone
in past users who had used hor- was not associated with colorectal
therapy mone therapy for at least 5 years cancer. There was little evidence
Hormone therapy to manage meno- and who had stopped use 5 years for differences in risk for different
pausal symptoms such as vasomo- or more ago. progestogen doses or progestogen
tor hot flushes, night sweats, and The risk of ovarian cancer was types. Risk estimates were gener-
vaginal atrophy includes estrogen- increased in both users of estrogen- ally lower among current users of
only therapy (which is prescribed only therapy and users of combined transdermal estrogen-only therapy
mainly to women who have had a estrogen–progestogen therapy. Risk compared with oral estrogen. The
hysterectomy) and combined estro- estimates were similar regardless benefits of hormone therapy ap-
gen–progestogen preparations. of the age when hormone therapy peared to be stronger for advanced
The IARC Monographs pro- started. There were differences in re- stage 4 colorectal cancer.
gramme has evaluated these drugs sults by tumour type, with increased Over the 4-year period from
[3] and concluded that estrogen- risks found only for serous or en- 2004 to 2008, 3799 colorectal
only hormone therapy is associ- dometrioid tumours (see Chapter cancers occurred in a cohort of
ated with cancer of the endome- 5.12). The Collaborative Group esti- 466 822 women aged 55–79 years
trium, ovary, and breast, and that mated that use of hormone therapy who were born in Norway and
combined estrogen–progestogen for 5 years from about age 50 years were living in Norway in 2004 [19].
hormone therapy is associated results in 1 additional ovarian cancer Current, but not past, use of hor-
with cancer of the breast and en- per 1000 users and 1 additional ovar- mone therapy was associated with
dometrium (the risk of endometrial ian cancer death per 1700 users. a reduced risk of colorectal cancer
cancer is reduced proportionally by Critics of the findings of the (RR, 0.88; 95% CI, 0.80–0.98). The
the number of days per month that Collaborative Group have high- short follow-up period of the study
progestogens are added to the regi- lighted the absence of a relationship meant that the influence of duration
men). The IARC Monographs also with duration of use, the potential for of use could not be examined.
concluded that combined hormone diagnostic bias, the smaller risk es- Risk estimates were similar for
therapy is unlikely to increase the timates from retrospective studies, estrogen-only therapy and com-
risk of colorectal cancer or alter the and inadequate adjustment for some bined estrogen–progestogen thera-
risk of ovarian cancer. important confounders; therefore, py and for colon cancer and rectal
Since the IARC Monographs causality could not be established cancer. Similarly to the findings of
evaluation, the Collaborative Group [17]. Nevertheless, the work of the the Danish study, use of hormone
on Epidemiological Studies of Ovar- Collaborative Group is the most therapy was associated with a re-
ian Cancer [16] analysed data from comprehensive so far and forms the duction in the risk of regionally
52 observational studies involving basis for many current clinical guide- advanced tumours (by 19%) and
21 488 women with ovarian can- lines for the prescribing of menopau- of metastatic colorectal cancer (by
cer; more than half of the cancers sal hormone therapy. 21%) but not of localized tumours.
(12 110) occurred in prospective Two recent large observational Although the association was not
studies. In the prospective studies, studies have both linked national statistically significant, in current
ever users of hormone therapy had registries to investigate use of hor- users the risk of colorectal cancer
an increased risk of ovarian cancer mone therapy and risk of colorectal tended to decrease with higher
(RR, 1.20; 95% CI, 1.15–1.26) com- cancer [18,19]. doses of oral estrogen, but not
pared with never users, and the risk A cohort of 1 006 219 wom- of progestogen.
was strongly associated with re- en living in Denmark and aged Both of the recent studies ac-
cency of use. Current use or recent 50–79 years was followed up from counted for several confounders but
use (within the past 5 years) was 1995 to 2009; 8377 incident colon were unable to adjust for previous
associated with an increased risk of cancers and 4742 rectal cancers use of oral contraceptives and for

some colorectal cancer risk factors, initiated worldwide, in addition to risk of endometrial cancer when
including body mass index, physical an unknown number of ovulation used at high doses or when used
activity, and smoking (see Chapter induction cycles. Concerns have for more than seven cycles, but the
5.5). Therefore, the evidence about been raised about the long-term effect of clomiphene citrate could
menopausal hormone therapy and effects of fertility drugs on the risk not be separated from the underly-
a possible reduced risk of colorec- of cancers of the breast, ovary, and ing clinical reasons for such usage
tal cancer remains inconclusive. endometrium [23]. patterns. Accordingly, the review
The Collaborative Group on A systematic review and meta- reported that because of very low-
Hormonal Factors in Breast Cancer analysis of 20 cohort studies in- quality evidence, robust conclu-
recently published an individual par- cluding 207 914 women who had sions could not be reached.
ticipant meta-analysis of the world- hormonal treatments for infertil- These systematic reviews [21,
wide epidemiological evidence on ity concluded overall that there 24,25] highlight several methodo-
the type and timing of menopausal was no association with risk of logical limitations of research to
hormone therapy and risk of breast breast cancer (summary RR, 1.05; date. Many studies have a relatively
cancer in 143 887 women with breast 95% CI, 0.96–1.14) [24]. However, short follow-up period, are limited
cancer and 424 972 controls [20]. All there was significant heterogene- by risk estimates based on small
types of menopausal hormone ther- ity among the studies (I 2 = 58.5%; event numbers, lack adjustment
apy, except vaginal estrogens, were P = 0.001). Subgroup analysis for confounders, could be prone to
associated with an increased risk of found an increased risk of breast detection or surveillance bias, and
breast cancer, which increased with cancer in three studies of women do not provide details of the fertil-
duration of use. Risks were larger who were treated before 1980 and ity drugs used (including regimens,
for combined estrogen–progesto- therefore did not have in vitro fer- doses, and number of cycles). The
gen therapy than for estrogen-only tilization (summary RR, 1.26; 95%
choice of comparator varies be-
therapy, especially with daily rather CI, 1.06–1.50). This finding raised
tween studies; the comparator can
than intermittent progestogen. After concerns about the association of
be the general population, subfer-
cessation of use, an increased risk clomiphene citrate treatment with
tile women, or both groups. It is also
of breast cancer remained for more breast cancer risk, although it was
important to note that women who
than 10 years, which was depen- noted that during the time period
take fertility drugs are a heteroge-
dent on duration of prior use. Risks before in vitro fertilization, use of
neous group, and for many of them
were similar regardless of whether this agent was not limited to anovu-
the underlying reasons for subfer-
women were aged 40–44, 45–49, latory women.
tility are risk factors for cancers of
50–54, or 55–59 years when start- A systematic review of 14 co-
the breast, ovary, or endometrium
ing menopausal hormone therapy. hort studies and 11 case–control
studies (including a total of 182 972 independent of any fertility treat-
It was estimated that approximately ments. Such limitations mean that
1 million of the 20 million breast women) was conducted to evaluate
the risk of ovarian cancer in women it is challenging to interpret the find-
cancers diagnosed in high-income ings of studies of the association
countries since 1990 would have treated with ovary-stimulating drugs
[25]. Because of the heterogeneity between use of fertility drugs and
been caused by use of menopausal
among the studies, meta-analysis risk of cancer.
hormone therapy [20].
was not performed. The review Based on the evidence to date,
concluded that there was no con- the Practice Committee of the
Fertility drugs vincing evidence of an increased American Society for Reproductive
Treatment for subfertility typically risk of invasive ovarian cancer and Medicine has concluded that there
involves the use of ovary-stimu- that there may be an increased risk does not appear to be a meaning-
lating agents, including selective of borderline ovarian tumours with ful increase in the risk of breast
estrogen-receptor modulators such use of fertility drugs. cancer, invasive ovarian cancer,
as clomiphene citrate, gonadotro- The association between use of or endometrial cancer associated
pins, gonadotropin-releasing hor- ovary-stimulating drugs and risk of with the use of fertility drugs, and
mone agonists and antagonists, endometrial cancer has been ex- that although there may be an in-
and human chorionic gonadotropin amined in a systematic review of creased risk of borderline ovarian
[21]. Use of these drugs is becom- 19 studies (16 retrospective cohort tumours, any absolute risk is small
ing increasingly common. During studies and 3 case–control stud- [23]. Given the growing numbers of
2011, more than 1.5 million assist- ies) including 1  937 880 women women using fertility drugs, good-
ed reproductive technology cycles [21]. Clomiphene citrate appeared quality evidence about their possi-
[22] were estimated to have been to be associated with an increased ble cancer effects is required.
142
References
1. Friis S, Kesminiene A, Espina C, Auvinen 10. Mørch LS, Skovlund CW, Hannaford 17. Gompel A, Burger H (2015). A
A, Straif K, Schüz J (2015). European Code PC, Iversen L, Fielding S, Lidegaard Ø Commentary on a recent update of the
Against Cancer 4th edition: medical expo- (2017). Contemporary hormonal con- ovarian cancer risk attributable to meno-
sures, including hormone therapy, and can- traception and the risk of breast can- pausal hormone therapy. Climacteric.
cer. Cancer Epidemiol. 39(Suppl 1):S107– 18(3):376–8. https://doi.org/10.3109/1369
cer. N Engl J Med. 377(23):2228–39.
19. https://doi.org/10.1016/j.canep.2015. 7137.2015.1023615 PMID:25812672
https://doi.org/10.1056/NEJMoa1700732
08.003 PMID:26390952 PMID:29211679 18. Mørch LS, Lidegaard Ø, Keiding N,
Løkkegaard E, Kjær SK (2016). The influ-
CHAPTER 2.11
SECTION 2
2. United Nations, Department of Economic 11. Collaborative Group on Hormonal Factors ence of hormone therapies on colon and
and Social Affairs, Population Division in Breast Cancer (1996). Breast can- rectal cancer. Eur J Epidemiol. 31(5):481–9.
(2015). Trends in contraceptive use world- cer and hormonal contraceptives: col- https://doi.org/10.1007/s10654-016-0116-z
wide 2015 (ST/ESA/SER.A/349). New
laborative reanalysis of individual data PMID:26758900
York (NY), USA: United Nations. Available
on 53 297 women with breast cancer
from: http://www.un.org/en/development/ 19. Botteri E, Støer NC, Sakshaug S, Graff-
and 100 239 women without breast can-
desa/population/publications/pdf/family/ Iversen S, Vangen S, Hofvind S, et al.
cer from 54 epidemiological studies. (2017). Menopausal hormone therapy and
trendsContraceptiveUse2015Report.pdf.
Lancet. 347(9017):1713–27. https://doi. colorectal cancer: a linkage between na-
3. IARC (2012). Pharmaceuticals. IARC org/10.1016/S0140 - 673 6(9 6) 9 0 8 0 6 - 5 tionwide registries in Norway. BMJ Open.
Monogr Eval Carcinog Risks Hum. 100A: PMID:8656904 7(11):e017639. https://doi.org/10.1136/
1–437. Available from: http://publications. bmjopen-2017-017639 PMID:29146641
iarc.fr/118 PMID:23189749 12. Soini T, Hurskainen R, Grénman S,
Mäenpää J, Paavonen J, Pukkala 20. Collaborative Group on Hormonal Factors
4. IARC (1999). Hormonal contraception and E (2014). Cancer risk in women us- in Breast Cancer (2019). Type and timing of
post-menopausal hormonal therapy. IARC ing the levonorgestrel-releasing in- menopausal hormone therapy and breast
Monogr Eval Carcinog Risks Hum. 72: trauterine system in Finland. Obstet cancer risk: individual participant meta-
1–660. Available from: http://publications. Gynecol. 124(2 Pt 1):292–9. https://doi. analysis of the worldwide epidemiological
iarc.fr/90. evidence. Lancet. 394:1159–68. https://doi.
org/10.1097/AOG.0000000000000356
org/10.1016/S0140-6736(19)31709-X
5. Collaborative Group on Epidemiological PMID:25004338
Studies on Endometrial Cancer (2015). 21. Skalkidou A, Sergentanis TN, Gialamas
13. Soini T, Hurskainen R, Grénman S, SP, Georgakis MK, Psaltopoulou T,
Endometrial cancer and oral contracep- Mäenpää J, Paavonen J, Joensuu H, et Trivella M, et al. (2017). Risk of endome-
tives: an individual participant meta-anal-
al. (2016). Levonorgestrel-releasing in- trial cancer in women treated with ovary-
ysis of 27 276 women with endometrial
trauterine system and the risk of breast stimulating drugs for subfertility. Cochrane
cancer from 36 epidemiological studies.
cancer: a nationwide cohort study. Database Syst Rev. 3:CD010931. https://
Lancet Oncol. 16(9):1061–70. https://
Acta Oncol. 55(2):188–92. https://doi. doi.org/10.10 02 /14651858.CD010931.
doi.org/10.1016/S1470-2045(15)00212-0 pub2 PMID:28349511
org/10.3109/028 418 6X. 2015.10 62538
PMID:26254030
PMID:26243443 22. Adamson GD, de Mouzon J, Chambers GM,
6. Charlton BM, Rich-Edwards JW, Colditz Zegers-Hochschild F, Mansour R, Ishihara
GA, Missmer SA, Rosner BA, Hankinson 14. Jareid M, Thalabard J-C, Aarflot M,
O, et al. (2018). International Committee
SE, et al. (2014). Oral contraceptive use Bøvelstad HM, Lund E, Braaten T (2018).
for Monitoring Assisted Reproductive
and mortality after 36 years of follow-up Levonorgestrel-releasing intrauterine sys-
Technology: world report on assisted re-
in the Nurses’ Health Study: prospective tem use is associated with a decreased productive technology, 2011. Fertil Steril.
cohort study. BMJ. 349:g6356. https://doi. risk of ovarian and endometrial cancer, 110(6):1067–80. https://doi.org/10.1016/
org/10.1136/bmj.g6356 PMID:25361731 without increased risk of breast can- j.fertnstert.2018.06.039 PMID:30396551
cer. Results from the NOWAC Study.
7. Iversen L, Sivasubramaniam S, Lee AJ, Gynecol Oncol. 149(1):127–32. https:// 23. Pfeifer S, Butts S, Dumesic D, Fossum G,
Fielding S, Hannaford PC (2017). Lifetime Gracia C, La Barbera A, et al.; Practice
doi.o rg /10.1016 / j.ygy no. 2018 .02.0 0 6
cancer risk and combined oral contra- Committee of the American Society for
PMID:29482839
ceptives: the Royal College of General Reproductive Medicine (2016). Fertility
Practitioners’ Oral Contraception Study. 15. Iversen L, Fielding S, Lidegaard Ø, Mørch drugs and cancer: a guideline. Fertil Steril.
Am J Obstet Gynecol. 216(6):580.e1–9. LS, Skovlund CW, Hannaford PC (2018). 106(7):1617–26. https://doi.org/10.1016/j.
https://doi.org/10.1016/j.ajog.2017.02.002 fertnstert.2016.08.035 PMID:27573989
Association between contemporary hor-
PMID:28188769 monal contraception and ovarian cancer 24. Gennari A, Costa M, Puntoni M, Paleari
in women of reproductive age in Denmark: L, De Censi A, Sormani MP, et al. (2015).
8. Michels KA, Brinton LA, Pfeiffer RM,
prospective, nationwide cohort study. Breast cancer incidence after hormonal
Trabert B (2018). Oral contraceptive use
BMJ. 362:k3609. https://doi.org/10.1136/ treatments for infertility: systematic re-
and risks of cancer in the NIH-AARP
bmj.k3609 PMID:30257920 view and meta-analysis of population-
Diet and Health Study. Am J Epidemiol. based studies. Breast Cancer Res Treat.
187(8):1630–41. https://doi.org/10.1093/ 16. Beral V, Gaitskell K, Hermon C, Moser K, 150(2):405–13. https://doi.org/10.1007/
aje/kwx388 PMID:29394309 Reeves G, Peto R; Collaborative Group on s10549-015-3328-0 PMID:25744295
9. McGuire V, Hartge P, Liao LM, Sinha R, Epidemiological Studies of Ovarian Cancer
25. Rizzuto I, Behrens RF, Smith LA (2013).
Bernstein L, Canchola AJ, et al. (2016). (2015). Menopausal hormone use and Risk of ovarian cancer in women
Parity and oral contraceptive use in rela- ovarian cancer risk: individual participant treated with ovarian stimulating drugs
tion to ovarian cancer risk in older wom- meta-analysis of 52 epidemiological stud- for infertility. Cochrane Database
en. Cancer Epidemiol Biomarkers Prev. ies. Lancet. 385(9980):1835–42. https:// Syst Rev. 8(8):CD008215. https://doi.
25(7):1059–63. https://doi.org/10.1158/ doi.org/10.1016/S0140-6736(14)61687-1 org/10.1002/14651858.CD008215.pub2
1055-9965.EPI-16-0011 PMID:27197274 PMID:25684585 PMID:23943232

World Cancer Research Fund International/

American Institute for Cancer Research
Martin J. Wiseman
The global network of World and cancer, and in translating the The WCRF/AICR recommenda-
Cancer Research Fund (WCRF) findings into recommendations for tions developed by the independent
International comprises registered cancer prevention for use by health expert panel based on the system-
charities in the United Kingdom professionals, individuals, and gov- atic evidence reviews now consti-
and the Netherlands as well as ernments worldwide. tute the most authoritative state-
the American Institute for Cancer The first WCRF/AICR Expert ment of the opportunity to prevent
Research (AICR) in the USA. AICR Report, Food, Nutrition, Physical cancer through food, nutrition, and
was established in 1982 after the re- Activity, and the Prevention of physical activity, highlighting the
view by the United States National Cancer: a Global Perspective, was importance of maintaining a healthy
Academy of Sciences in that year, published in 1997. This synthesis of body weight through appropriate
which drew attention to the increas- mostly epidemiological research on levels of physical activity and a bal-
ing epidemiological evidence of nutrition and cancer laid the foun- anced diet, predominantly based on
links between food and nutrition and dations for the following decades of plant foods, with no more than mod-
several cancer types, as well as the scientific interest in this area. est amounts of meat and dairy, and
growing understanding of the influ- An initially sceptical scientific limiting the amounts of processed
ence of nutritional factors on the pro- community has been persuaded meat, salt, and alcohol, as well as
cess of carcinogenesis. However, not only by the now large number of high-energy foods with high lev-
even then, scientific research into the of studies of increasingly high qual- els of fat, sugar, and salt (so-called
link between diet and cancer was in ity but also by a series of state-of- fast foods).
its infancy. The WCRF International the-art reviews conducted by WCRF The same rigorous approach to
network was the first organization and AICR. In particular, the second the evidence underpinned the next
to focus exclusively on the links WCRF/AICR Expert Report, pub- phase of development, the WCRF/
between cancer and nutrition, and lished in 2007, explored the epide- AICR Continuous Update Project,
more recently physical activity. The miology of the links between food, in which the database of information
WCRF International network has nutrition (in particular adiposity), and extracted for the articles identified
a vision to live in a world where no physical activity and cancers as well by systematic review is maintained
one develops a preventable cancer, as the potential mechanistic under- on a continuous basis. The past de-
and over the past decades WCRF pinning of those links; that created cade of research was summarized
International has funded millions a step change in the perception of in 2018 in the third WCRF/AICR
of dollars in cancer prevention re- the importance of these exposures Expert Report. The revised rec-
search and awareness-raising for the global distribution, and bur- ommendations were not strikingly
programmes. Through its Expert den, of cancer, second only to that different from those in the previous
Reports and now the Continuous of smoking. The importance of the reports, but there was a shift in em-
Update Project, WCRF International 2007 Expert Report lay in the rig- phasis away from individual foods
has set the standard for the syn- orous systematic methods used to and nutrients and towards an over-
thesis and analysis of published review the evidence, as well as the all package, with healthy patterns
research on the links between diet, care taken in developing criteria to of food and beverage consumption
body weight, and physical activity evaluate the evidence. and physical activity, and with an
144
additional emphasis on the impor- identified that, although cancer immune surveillance, is likely to pro-
tance of body weight. appears clinically mostly after the vide important insights in the future.
The 2018 Expert Report also age of 50 years, events that occur The WCRF/AICR series of
identified some areas where more early in life (marked by, for example, Expert Reports and the Continuous
work would help to derive bet- birth weight or adult attained height) Update Project are recognized as
ter recommendations. First, there seem to be important in determin- the most authoritative summary
remains a dearth of high-quality ing cancer susceptibility in later life. statement of the links between diet,
studies to inform nutritional guid- Finally, new research on nutritional nutrition, and physical activity and
ance to people living with and be- influences in developing areas such the risk and progression of cancer.
yond cancer. Second, the report as the colonic microbiome, and in
World Cancer Research Fund International/American Institute for Cancer Research 145
3 Biological
processes in cancer
development
Knowledge of how normal cells become can- that is categorized as sporadic, for which no
cerous – the process of malignant transfor- such exposure is evident. Cancer development
mation – may underpin cancer prevention. after exposure includes the induction of carcin-
Changes evident in premalignant tissues or at ogen-related mutations; critical mutations may
the earliest stage of tumour development are also occur spontaneously. DNA repair may be
key to improve screening and to monitor peo- protective, epigenetic events may be as impor-
ple with an increased risk of cancer because of tant as mutations, and chronic inflammation
their genetic makeup, and also have implica- plays a key role. Malignant transformation is
tions for cancer treatment. Two scenarios are marked by metabolic, immunological, and hor-
covered: cancer that develops after exposure to monal changes. Knowledge of such biological
carcinogens, including hazardous chemicals, processes has contributed to reducing cancer
radiation, or infectious organisms, and cancer incidence and mortality.
3.1
Sporadic cancer
Tumorigenesis in the absence of an
established or avoidable cause
David Schottenfeld Paul Brennan (reviewer)
George Davey Smith (reviewer)
of stem/progenitor cells medi- marks of the neoplastic phenotype

SUMMARY ated by aberrant regulation of include sustaining proliferative sig-
tumour progenitor genes. nalling, evading growth suppres-
●● Multiple factors are recognized sion, avoiding immune destruction,
as contributing to the develop-
enabling replicative immortality,
ment of sporadic cancers.
Sporadic cancers occur ostensibly resisting apoptosis, deregulating
●● Telomeric DNA shortens pro- in the absence of a demonstrable cellular energetics, inducing an-
gressively as cell lineages cause or history of familial suscep- giogenesis, and activating invasion
pass through repeated division tibility. At the germline or somatic and metastasis (Fig. 3.1.1).
cycles and ultimately senes- cellular level, the biology of the More recently, additional em-
cence. The immortalization of cancer cell is viewed as a complex phasis has been placed on the in-
cancer cells may occur through genetic disorder. teraction of tumour cells and the
activating expression of the tel- The publications of Hanahan mesenchymal cells forming the tu-
omerase polymerase. and Weinberg have provided a logi- mour-associated stroma or tumour
●● Stem cell quiescence may cal framework for comprehending microenvironment. The above-
be viewed as an evolutionar- the multistep process of human mentioned essential functional ca-
ily conserved mechanism that tumour pathogenesis [1]. The hall- pabilities of cancer cells to survive,
modulates stochastic events of
cell replication and the acquisi-
tion of tumorigenic mutations. Fig. 3.1.1. The hallmarks of cancer.
●● Cancer stem cells are a selec- Sustaining

tive clonal subset of tumour proliferative
signaling
cells that have avoided various
Deregulating Evading
cell regulatory mechanisms, cellular growth
including terminal differentia- energetics suppressors
tion, and yet have retained the
self-renewal properties and
proliferative potential of adult
stem cells.
Resisting Avoiding
●● Epigenetic events are intimately cell immune
death destruction
associated with fetal organ de-
velopment, pathological events
associated with ageing, bio-
chemical effects of micronutri-
ents, and the tumorigenic ef- Enabling
Inducing
fects of cytokine mediators of angiogenesis
replicative
chronic inflammation. The pro- immortality
Activating
posed tumorigenic event is a invasion &
polyclonal epigenetic disruption metastasis
148
proliferate, and disseminate are In the context of tumour sup-

enabled by genomic instability and pression, factors secreted by se- FUNDAMENTALS
inflammatory responses mediated nescent cells attract components
by the immune cells recruited by the of the innate and adaptive immune ■■ At the germline or somatic
stroma of malignant cells. system that serve to remove dam- cellular level, the biology of the
aged and stressed senescent cells. cancer cell and its nurturing
In addition to arrested growth and microenvironment is viewed as
Ageing, telomeres, and failure to re-enter the cell cycle, a complex genetic disorder.
cancer susceptibility senescent cells show widespread
changes in chromatin organiza- ■■ The convergence of biological
Ageing mechanisms in ageing and
tion. Senescent cells may also se-
Ageing is a complex biological phe- crete pro-inflammatory cytokines, neoplasia reflects the effects
nomenon that is exhibited by all liv- chemokines, and growth factors of telomere dysfunction on
ing organisms and is accompanied cellular senescence and
that are demonstrated to enhance
by a gradual decline in physiologi- genomic instability.
cell proliferation and transformation
cal functions. The convergence of [6]. Pro-angiogenic factors secreted ■■ Adult stem cells are observed
biological mechanisms in ageing from senescent cells promote tissue in close association with
and neoplasia is explored by relat- vascularization and increase inva- differentiated cells of various
ing the effects of telomere dysfunc- siveness of premalignant cells by organs and tissues, and exhibit
tion on cellular senescence and driving epithelial-to-mesenchymal properties of self-renewal and
genomic instability. transitions (Fig. 3.1.2). DNA double- asymmetric division.
SECTION 3
CHAPTER 3.1
Increasing age is a major predic- strand breaks or telomere dysfunc-
■■ Epigenetic events are
tor of adult-onset cancer incidence. tion caused by oxidative stress may
stochastic, discrete, and
A logarithmic pattern of overall can- induce a senescent response.
heritable, may confer the
cer incidence and age (i.e. the inci-
Telomeres propensity for aberrant
dence of cancer increases approxi- growth, and are influenced
mately exponentially as a function Human telomeres, which are special-
by environmental
of age) has suggested a multistep ized structures at the ends of chromo-
oncogenic agents.
biological mechanism in human somes, consist of tandem repetitive
carcinogenesis [2,3]. In industrial- arrays of the hexameric sequence ■■ The terminology “sporadic
ized countries, the overall cancer TTAGGG. Functional telomeres are cancer” reflects a currently
incidence rates more than doubled required to protect chromosome dynamic but incomplete
with each increase of 10 years in at- ends, provide chromosome stability, knowledge of the etiology and
tained age. In an analysis of adults and ensure, upon cell division, the pathogenesis of a biologically
in the USA in 2012–2014 [4], the fidelity of segregation of genetic ma- and morphologically
terial into daughter cells. Telomeric heterogeneous class
probability (as a percentage) of
dysfunction has consequences for of diseases.
developing invasive cancer at at-
ageing and carcinogenesis [7,8].
tained ages 50–59 years was 6%,
The mechanisms that govern ex-
as contrasted with 26% in women
posure of cells to metabolic stress
and 32% in men at ages 70 years
or crisis involve the cell genome, ing expression of the telomerase
and older.
and more specifically the telo- polymerase, a ribonucleoprotein
Cellular senescence meres. The ends of the telomeric enzyme, which restores and main-
DNA are not copied completely dur- tains telomeric DNA length [10].
Cellular senescence refers to ir- ing each cycle of DNA replication, The enzyme telomerase consists
reversible arrest of cell prolifer- because of an intrinsic limitation in of a subunit that has reverse tran-
ation. Although senescent cells are the DNA polymerases responsible scriptase activity, an RNA element
not dividing, they remain metaboli- for DNA replication. In addition, the that is the template on which DNA
cally active, secreting factors that ends of telomeric DNA are suscep- is synthesized, and the protein
may stimulate or inhibit the growth tible to the action of exonucleases, dyskerin, which has the ability to
of tumours. In vitro, senescent which contribute to erosion of telo- bind to and stabilize the RNA ele-
cells display an enlarged and flat- meric DNA length [9]. As a con- ment [11]. Upregulated telomerase
tened morphology, have elevated sequence, the telomeres shorten expression is a characteristic of
β-galactosidase activity, and ex- progressively as cell lineages pass pluripotent stem cells. Telomerase
press markers consistent with ac- through repeated division cycles activity is detectable in most human
tivation of tumour suppressor path- and ultimately senescence. tumours as a result of induction of
ways, cell-cycle arrest, and DNA The immortalization of cancer expression by a complex array of
damage response signalling [5]. cells may occur through activat- trans-activating oncoproteins.
Chapter 3.1 • Sporadic cancer 149

Fig. 3.1.2. Senescent cells secrete multiple factors that can have effects on the tis- Cancer stem cells and
sue microenvironment. progenitor cells
TUMOR-SUPPRESSIVE ROLE Cancer stem cells are a selective
clonal subset of tumour cells that
Immune system (NK cells,,
pro-inflammatory macrophages, T cells) have avoided various cell regula-
cytokines
tory mechanisms, including ter-
Reinforcement of minal differentiation, and yet have
senescence surveillance, killing retained the self-renewal properties
arrest (autocrine) and removal of
senescent cells and proliferative potential of adult
Senescent cell stem cells. Most tumours are main-
tained by a subpopulation of clonal
stem cells.
protease As defined by the American
secretion Association for Cancer Research
Bystander [14], a cancer stem cell is “a cell
senescence Extracellular within a tumor that possesses the
(paracrine) matrix remodeling
capacity to self-renew and to cause
the heterogeneous lineages of can-
cer cells that comprise the tumor”.
By maintaining at least some of the
properties of their tissue of origin,
TUMOR-PROMOTING ROLE
cancer stem cells give rise to tu-
mours that phenotypically share in
Chronic their morphological features and pat-
pro-inflammatory inflammation
terns of expression of tissue-specific
growth factor
cytokines
genes. Progenitor cells are progeny
Tumor cell secretion
proliferation of tissue-specific stem cells with lim-
ited potential for self-renewal.
Senescent cell
Two models of carcinogenesis
growth factor have been proposed. A stochas-
secretion tic model proposes that neoplasia
evolves potentially in any somatic
cell through a sequence of muta-
Epithelial tional and epigenetic events that
mesenchymal Angiogenesis
transition are amplified by selective clonal
increased growth. In contrast to the stochastic
invasiveness model, the cancer stem cell model
hypothesizes that the cellular origin
of cancer resides in tissue-specific
Somatic stem cells and sion results when a stem cell divides stem cells or progenitor cells that
human carcinogenesis into one daughter cell that replicates possess or acquire the property of
a stem cell, while the other daughter self-renewal [15]. The development
Stem cells cell proceeds along some differen- of biomarkers to identify cancer
Adult stem cells are observed in tiating pathway (Fig. 3.1.3). The ho- stem cells has facilitated the iso-
close association with differentiated meostatic balance between self-relation and characterization of cells
cells of a given tissue. They are usu- newal and differentiation is essential from human tumours. The neoplas-
ally located within specialized tis- for physiological maintenance of the tic evolution from normal tissue cells
sue microenvironments or stem cell architecture and functioning of adult is signalled by the loss of homeo-
“niches” composed of stromal cells organs and tissues [13]. static mechanisms that regulate mi-
and paracrine signalling factors [12]. Although adult somatic stem totic activity and differentiation.
Stem cells exhibit properties of cells have the potential to proliferate A contemporary view would
self-renewal and asymmetric divi- actively, they are relatively dormant tend to combine biological features
sion. Self-renewal signifies that in in their microenvironment. Stem cell advanced by both experimental
mitotic activity of stem cells there quiescence may be viewed as an models. Cancer stem cells are
is resistance to genetic and epige- evolutionarily conserved mechanism regulated by and interact with the
netic mechanisms that trigger se- that modulates stochastic events of tumour microenvironment. Cells
nescence or a permanent state of cell replication and the acquisition of recruited to the microenvironment
cell-cycle arrest. Asymmetric divi- tumorigenic mutations. include growth factors, cytokine
150
Fig. 3.1.3. Properties of stem cells.
1 A stem cell can self-renew and give

Stem cells are maintained in
rise to either cells of its own type or cells
microenvironmental niches
entering a terminal differentiation pathway.
consisting of stromal cells
Depending on tissue requirements,
stemness-associated genes can dictate Stromal cell
whether stem cell can remain transiently
dormant or undergo steady-state cycling. Stem cells have three characteristics: self-
Stem cell renewal, proliferation, and differentiation
replenishment into mature cells. Stem cells are housed in
2 Proliferation (self-renewal) niches consisting of stromal cells that
3 A precursor cell can provide factors for their maintenance.
undergo several rounds of Stem cells of the embryo can give rise to
cell divisions. As a cell precursors that generate all the tissues
pecursor cell differentiates, of the body. This property defines stem cells
it acquires distinctive as multipotent.
features characteristic of Stem cells are difficult to identify
each lineage. morphologically. Their identification is based
on specific cell surface markers (cell
surface antigens recognized by specific
SECTION 3
CHAPTER 3.1
monoclonal antibodies) and on the lineage
they generate following transplantation.
4 Differentiated
Three typical examples are the stem
cells are nonmitotic
cells of bone marrow, intestine, and testes.
with a finite life span.
5 Differentiating cells of a lineage
networks, and immunomodulatory thyl group to the 5-carbon position confer the propensity for aber-
T cells and macrophages. The no- of cytosine in a CpG dinucleotide. rant growth, and are influenced
tion of interaction between a stem Methylated cytosine residues have by environmental factors, namely
cell (the “seed”) and the tumour a tendency to deaminate spon- physical and chemical carcinogens
microenvironment (the “soil”) has taneously, causing C → T transi- and oncogenic infectious agents.
relevance to understanding tumour tions. Histone proteins are sub- Abnormal epigenetic programmes
metastasis and resistance to anti- ject to diverse post-translational may silence large groups of
cancer therapy. modifications, such as acetylation, genes, causing genomic instability.
methylation, phosphorylation, and Epigenetic post-translational modi-
ubiquitination [17]. fications of core histone patterns
Epigenetic mechanisms Epigenetic mechanisms are and DNA methylation may influence
in tumour development essential for normal functioning or accompany the ageing process.
Epigenetic events are composed of and development of human cells Feinberg et al. [18] proposed an
potentially heritable alterations in and tissues, as well as for mainte- epigenetic progenitor cell or epige-
gene expression that do not entail a nance of gene expression patterns. netic mediator model as a strategic
structural change in DNA sequenc- Epigenetic events are intimately as- step in human carcinogenesis. The
ing. Epigenetic events are associ- sociated with fetal organ develop- proposed tumorigenic event is a
ated with patterns of DNA methyla- ment, pathological events associat- polyclonal epigenetic disruption of
tion and histone modification that ed with ageing, biochemical effects stem/progenitor cells mediated by
serve to modulate the expression of of micronutrients, and the tumori- aberrant regulation of tumour pro-
proto-oncogenes and tumour sup- genic effects of cytokine mediators genitor genes. The authors’ pro-
pressor genes [16]. of chronic inflammation. posed terminology of “epigenetic
The methylation of DNA refers Epigenetic events are stochas- mediators” underscores functions
to the covalent addition of a me- tic, discrete, and heritable, may that affect the emergence and

maintenance of cancer stem cells, fraction) is generally interpreted as cer deaths in the 1970s could have
and the facilitation of cancer initiation the proportion of cases, or excess been avoided [22]. A review by Parkin
and progression (see Chapter 3.8). number of cases, that – based on et al. estimated that for the United
current knowledge – could be elimi- Kingdom in 2010, 14 lifestyle and
nated if the exposed people were to environmental risk factors (tobacco
Population attributable
experience the same risks as the smoke, ethanol consumption, obesi-
risks of sporadic cancers unexposed people [21]. The popu- ty and overweight, physical inactivity,
The terminology “sporadic cancers” lation attributable fraction reflects dietary factors including consump-
reflects a currently dynamic but the magnitude of the relative risk tion of red meat and processed meat,
incomplete knowledge of the etiol- of the association of the exposure cancer-causing infectious agents,
ogy and pathogenesis of a biologi- and the disease outcome, and the occupational exposures, ionizing and
cally and morphologically heteroge- prevalence of the exposure in the solar radiation, and exogenous hor-
neous class of diseases. The subpopulation. This assumes that the mones) were associated with 45%
text of the terminology, namely the estimation of population attributa- of cancer cases in men and 40% in
absence of a demonstrable cause, ble fraction is unbiased, that the women [23]. Colditz and Wei, in their
underscores the view of assigning exposure is causal, and that elimi- review of biological agents, lifestyle
“bad luck” in the affected populace. nation of the risk factor has no ef- behavioural patterns, and physical
Tomasetti and Vogelstein have hy- fect on the distribution of other risk environmental factors, concluded
pothesized that the patterns of can- factors. It is important to establish that 50–60% of cancer deaths and
cer incidence in various cells and that the measure of the prevalence more than 60% of cancer cases in
tissues are highly correlated with of the exposure in the population the USA were potentially avoidable
the estimated lifetime number of matches as closely as possible the [24]. The World Cancer Research
stem cell divisions [19,20]. Each so- population source for deriving the Fund/American Institute for Cancer
matic stem cell division entails a risk measure of relative risk. Research report in 2015 estimated
of random mutations. The variable Is there a consensus on the pop- that 20–22% of all incident cancers
number of divisions appears to be ulation cancer burden that may be at- in the United Kingdom and the USA
a major determinant of differences tributable to lifestyle behavioural and were due to the combined risk fac-
in cancer risks in different organs. environmental risk factors that would tors of diet, physical inactivity, and
The authors reviewed the risks of be interactive with stem cell replica- overweight or obesity [25]. Specific
17 types of cancers in 69 countries. tion activity? In the 1981 publication aspects of dietary factors included
The median correlation coefficient by Doll and Peto on the avoidable high consumption of red meat and
between the lifetime risk of cancer risks of cancer in the USA, the au- processed meat and low consump-
in each tissue and the reported life- thors concluded that 75–80% of can- tion of folate (see Chapter 2.6).
time number of stem cell divisions
within that tissue was r = 0.80 (95%
confidence interval, 0.67–0.84). Fig. 3.1.4. In the absence of a demonstrable cause, the view of assigning “bad luck”
The linearity of the positive corre- to cancer development arose from the proposal that the patterns of cancer incidence
lations was observed consistently in various cells and tissues are highly correlated with the estimated lifetime number of
stem cell divisions within those cells or tissues. Each somatic stem cell division entails
among the countries studied.
a risk of random mutations.
The estimated proportion of to-
tal variation in cancer incidence ex-
plained by the number of stem cell
divisions may be estimated by r 2
or 0.64 (95% confidence interval,
0.45–0.71). The authors concluded
that approximately two thirds of
global cancer incidence may be at-
tributed to random replication errors,
with a confidence boundary as low
as 45% and as high as 71%. Would
this be a measure of the global bur-
den of “sporadic cancers”?
A counterpoint epidemiologi-
cal perspective on the stem cell
hypothesis in human carcinogen-
esis will now be summarized. The
attributable fraction in the popula-
tion at risk (population attributable
152
Tomasetti and Vogelstein have hypothesis must be viewed in the dress the complexity of interactions
described a biological mechanism context of diverse and contrasting already established in epidemiologi-
of tissue-specific stem cell replica- global trends and patterns of types cal and experimental studies that
tion patterns that are positively cor- and “causes” of cancers that are describe the burden of cancers that
related with, and universally applica- closely linked with economic devel- may be attributable to avoidable or
ble in comprehending the diversity opment and cultural lifestyle prac- remediable risk factors.
of, organ-specific cancer incidence tices. The terminology “sporadic
patterns. The unifying nature of their cancer” does not adequately ad-
References
1. Hanahan D, Weinberg RA (2011). Hallmarks 11. Donate LE, Blasco MA (2011). Telomeres 19. Tomasetti C, Vogelstein B (2015). Cancer
of cancer: the next generation. Cell. in cancer and ageing. Philos Trans R etiology. Variation in cancer risk among tis-
144(5):646–74. https://doi.org/10.1016/j. Soc Lond B Biol Sci. 366(1561):76–84. sues can be explained by the number of
cell.2011.02.013 PMID:21376230 https://doi.org/10.1098/rstb.2010.0291 stem cell divisions. Science. 347(6217):78–
PMID:21115533 81. https://doi.org/10.1126/science.1260825
2. Armitage P, Doll R (1954). The age distri- PMID:25554788
bution of cancer and a multi-stage theory 12. Rossi DJ, Jamieson CHM, Weissman IL
of carcinogenesis. Br J Cancer. 8(1):1– (2008). Stems cells and the pathways to 20. Tomasetti C, Li L, Vogelstein B (2017). Stem
12. https://doi.org/10.1038/bjc.1954.1 aging and cancer. Cell. 132(4):681–96. cell divisions, somatic mutations, cancer
PMID:13172380 https://doi.org/10.1016/j.cell.2008.01.036 etiology, and cancer prevention. Science.
SECTION 3
CHAPTER 3.1
3. Hiller J, Vallejo C, Betthauser L, Keesling science.aaf9011 PMID:28336671
J (2017). Characteristic patterns of can- 13. Beachy PA, Karhadkar SS, Berman DM
cer incidence: epidemiological data, bio- (2004). Tissue repair and stem cell renewal 21. Poole C (2015). A history of the population
logical theories, and multistage models. in carcinogenesis. Nature. 432(7015):324– attributable fraction and related measures.
Prog Biophys Mol Biol. 124:41–8. https:// 31. https://doi.org/10.1038/nature03100 Ann Epidemiol. 25(3):147–54. https://doi.
doi.org/10.1016/j.pbiomolbio.2016.11.002 PMID:15549094 org/10.1016/j.annepidem. 2014.11.015
PMID:27836510 PMID:25721747
14. Clarke MF, Dick JE, Dirks PB, Eaves
4. Siegel RL, Miller KD, Jemal A (2018). CJ, Jamieson CHM, Jones DL, et al. 22. Doll R, Peto R (1981). The causes of can-
Cancer statistics, 2018. CA Cancer J (2006). Cancer stem cells – perspectives cer: quantitative estimates of avoidable
Clin. 68(1):7–30. https://doi.org/10.3322/ on current status and future directions: risks of cancer in the United States to-
caac.21442 PMID:29313949 AACR Workshop on cancer stem cells. day. J Natl Cancer Inst. 66(6):1191–308.
Cancer Res. 66(19):9339–44. https://doi. ht tps://doi.org/10.1093/jnci/6 6.6.1192
5. Ershler WB, Longo DL (1997). The biology of org/10.1158/0008 - 5 472.CAN - 06 -3126 PMID:7017215
aging: the current research agenda. Cancer. PMID:16990346
80(7):1284–93. https://doi.org/10.1002/(SICI) 23. Parkin DM, Boyd L, Walker LC (2011). The
1097- 0142(19971001)80:7<1284::AID - 15. Sugihara E, Saya H (2013). Complexity fraction of cancer attributable to lifestyle
CNCR14>3.0.CO;2-3 PMID:9317181 of cancer stem cells. Int J Cancer. and environmental factors in the UK in
132(6):1249–59. https://doi.org/10.1002/ 2010. Br J Cancer. 105(Suppl 2):S77–81.
6. Bolden JE, Lowe SW (2015). Cellular senes- ijc.27961 PMID:23180591 ht t ps: //d o i.o r g /10.10 3 8 / bjc . 2 011. 4 8 9
cence. In: Mendelsohn J, Gray JW, Howley PMID:22158327
PM, Israel MA, Thompson CB, editors. The 16. Herceg Z, Ghantous A, Wild CP, Sklias A,
molecular basis of cancer. Philadelphia Casati L, Duthie SJ, et al. (2018). Roadmap 24. Colditz GA, Wei EK (2012). Preventability of
(PA), USA: Elsevier; pp. 229–38. for investigating epigenome deregula- cancer: the relative contributions of biologic
tion and environmental origins of cancer. and social and physical environmental deter-
7. Aubert G, Lansdorp PM (2008). Telomeres Int J Cancer. 142(5):874–82. https://doi. minants of cancer mortality. Annu Rev Public
and aging. Physiol Rev. 88(2):557–79. org/10.1002/ijc.31014 PMID:28836271 Health. 33(1):137–56. https://doi.org/10.1146/
https://doi.org/10.1152/physrev.00026.2007 a n n u r ev - p u b l h e a l t h - 0 31811-12 4 6 2 7
PMID:18391173 17. Sharma S, Kelly TK, Jones PA (2010). PMID:22224878
Epigenetics in cancer. Carcinogenesis.
8. Finkel T, Serrano M, Blasco MA (2007). 31(1):27–36. https://doi.org/10.1093/carcin/ 25. WCRF/AICR (2015). Cancer preventabil-
The common biology of cancer and ageing. bgp220 PMID:19752007 ity estimates for diet, nutrition, body fat-
Nature. 448(7155):767–74. https://doi.org/ ness, and physical activity. Available from:
10.1038/nature05985 PMID:17700693 18. Feinberg AP, Koldobskiy MA, Göndör A ht tp:// w w w.wcr f.org / int /c anc er-facts -
(2016). Epigenetic modulators, modifiers figures/preventability-estimates/cancer-
9. Blasco MA (2005). Telomeres and hu- and mediators in cancer aetiology and
man disease: ageing, cancer and beyond. preventability-estimates-diet-nutrition.
progression. Nat Rev Genet. 17(5):284–
Nat Rev Genet. 6(8):611–22. https://doi. 99. https://doi.org/10.1038/nrg.2016.13
org/10.1038/nrg1656 PMID:16136653 PMID:26972587
10. Haycock PC, Burgess S, Nounu A, Zheng
J, Okoli GN, Bowden J, et al.; Telomeres
Mendelian Randomization Collaboration
(2017). Association between telomere
length and risk of cancer and non-neoplas-
tic diseases: a Mendelian randomization
study. JAMA Oncol. 3(5):636–51. https://
doi.org/10.10 01/jamaoncol. 2017. 2316
PMID:28241208

3.2
Genomics
Susceptibility and somatic patterns
Stephen J. Chanock Marc Ladanyi (reviewer)

James McKay (reviewer)
genes, and even more mutations Principles of germline

SUMMARY that appear to be passengers. genetic susceptibility
●● Next-generation sequencing has ●● The density of single-nucleotide to cancer
accelerated the pace of discov- mutations across a genome dif- The concept of familial cancer was
ery of new genes in which one fers by nearly 4 orders of mag- appreciated before the discovery of
or more mutations can confer an nitude (> 10 000-fold) between genes. In 1866, the astute French
increased risk of cancer. More cancer types with strong envi- physician Broca described a cluster
than 120 such genes have been ronmental factors and tumours of breast cancers in his wife’s family,
identified, of which more than with little such evidence, such heralding the idea of familial risk for
50% are also somatically altered as paediatric cancers. breast cancer. Although the herita
in cancers. ble contribution of cancer has been
investigated for a century and a half
●● Genome-wide association stud- The advent of the age of geno- through family and twin studies, it is
ies have accelerated the pace mic analyses has dramatically ac- the advent of genetic technologies,
of discovery of common genetic celerated the pace of discovery including the rise of next-generation
susceptibility variants. More than and characterization of susceptibil- sequencing in the past 15 years, that
85% of the loci identified in can- ity to cancer and of the hallmarks has accelerated the pace of discov-
cer genome-wide association of the genomic changes that can- ery of mutations in cancer predis-
studies have been discovered in cer cells undergo, both as conse- position genes and, more recently,
individuals of European ances- quential events and as a result of cancer susceptibility alleles.
try, with approximately 10% in the genomic changes in the cancer The annotation of the human
Asian ancestry and less than 5% cells (see Chapter 3.1). The devel- genome revealed a wide spectrum
in African ancestry; this reflects opment of a cancer represents a of genetic variation, from the most
the scope of studies undertaken new, distinct cell population charac- frequent variant – the single base
to date. terized by a range of genetic events, change – to large structural changes
●● Although the pace of discover- some of which drive the cancer. in copy number. Early studies in fam-
ies from genome-wide asso- The germline genome (i.e. the ge- ilies identified damaging mutations
ciation studies has accelerated nome that a person has at birth) in BRCA1, the first hereditary breast
with large collaborative net- confers susceptibility to or protec- and ovarian cancer gene discovered
works, the investigation of each tion against contributions to the [1]. The search for familial cancer
individual susceptibility locus cancer and its clinical course. The genes has identified more than 120
has not advanced at a compa- next generation of studies will inte- genes in which rare mutations can
rable speed. grate these two genomes, providing confer an increased risk of cancer
more precise insights into how the [2]; most of these mutations are also
●● Landscape analyses of events environment, including lifestyle fac- seen in tumours, serving as somatic
across entire cancer genomes tors, contributes to cancer etiology drivers of the cancer. From a public
have revealed a wide range of and the outcomes associated with health perspective, these account
types of somatic genetic events a cancer. This chapter discusses for less than 10% of cancers.
(from single base mutations to major trends in elucidating how the More recently, the focus has been
the shattering of entire chromo- germline genome informs the un- on the identification of many common
somes), many involving driver derstanding of the cancer genome. variants, each of which provides a
154
small contribution to cancer risk [3,4]. discover highly penetrant mutations,

The search has been to scan across such as those in BRCA1 or TP53, FUNDAMENTALS
the most common variant, the single- which are rare but have a strong pre-
nucleotide polymorphism (SNP), de- dictive value for cancer over time. ■■ Investigations of the
fined as a substitution of one base, Common susceptibility alleles, which contribution of the germline
with at most minimal impact on the bi- confer a smaller cancer risk, are genome have successfully
ology of the gene or the genomic re- discovered by association studies, identified many new
gion. The frequency of the alternative which compare the frequency of sets susceptibility variants, most
base pair, known as the minor allele of alleles between affected and unaf- of which are unique to a
frequency, varies greatly by popula- fected individuals [5]. The estimated cancer type; these variants
tion genetics history. Often, the ef- effect sizes are smaller for common
vary substantially in both
fect of common SNP variants is on variants and are neither necessary
effect size and distribution in
the regulation of a gene and not the nor sufficient for cancer susceptibility.
distinct populations.
gene or protein function itself. The For each cancer type and subtype, it
combined effects of selection and has emerged that there is a distinct ■■ The discovery of germline
background drift in allele frequencies underlying genetic architecture, com- variants that contribute to
are etched in the patterns of genetic prising common variants with small cancer susceptibility has
variation; this includes both the cor- effects, rare variants with strong ef- provided new mechanistic
relation between nearby variants, fects, and the still-to-be-defined less clues to cancer etiology,
known as linkage disequilibrium, and common variants with moderate ef- including changes in the
the actual frequencies of common fects (Fig. 3.2.1) [6,7]. Moreover, the regulation of key genes and
variants, measured by the minor al-
CHAPTER 3.2
set of common variants can be com-
SECTION 3
pathways. The relationship
lele frequency. In turn, these differ- bined to generate a polygenic model between germline
ences have become attractive for for cancer susceptibility [8]. susceptibility alleles and
investigating differences in incidence
somatic alterations may
for distinct cancer types, by either
population or exposure.
The search for regions of uncover new pathways and
Cancer susceptibility alleles can the genome that confer targets for therapeutic and
be discovered by different approach- susceptibility to cancer preventive measures.
es, including linkage, association, ■■ Understanding the underlying
and now next-generation sequencing Cancer predisposition genes genetic architecture of
analyses. Not all alleles have com- For decades, cancer geneticists common and rare cancer
parable estimated effects. Linkage have investigated families or spe- types provides a foundation
analyses in family studies are used to cial populations in which multiple for developing effective
approaches towards precision
prevention in oncology.
Fig. 3.2.1. Distribution of susceptibility alleles by frequency and strength of genetic effect,
illustrating the distribution of susceptibility alleles as well as the feasibility of identifying ■■ One of the hallmarks of
variants through genome-wide association studies (GWAS) and sequence analysis. cancer is an altered genome,
which features mutations that
Effect size
drive abnormal growth and
50.0
can lead to cancer-related
High
deaths. The disruption of
Rare alleles Few examples of
causing high-effect normal functions by cancer
Mendelian common variants mutations can also generate
3.0 disease influencing many passenger mutations.
common disease
Intermediate Low-frequency
variants with ■■ Globally, major differences
intermediate in the patterns of mutations
effect
1.5 for distinct types and subtypes
Common of cancer correlate with
Rare variants of
Modest variants distinct exposure and
small effect very
implicated in
1.1
hard to identify
common population diversity, providing
by genetic means
disease by etiological clues that could
Low GWAS
be used to develop new
0.001 0.005 0.05 prevention, detection, and
Very rare Rare Low-frequency Common
treatment strategies.
Allele frequency
Chapter 3.2 • Genomics 155

members developed the same type genetic variants in a cancer predis- Common susceptibility alleles
or types of cancer. Most of the early position gene confer risk; this un- in cancer
studies were based on collections derscores the importance of careful The advent of genome-wide as-
of families with similar cancers, annotation of variants in particular sociation studies (GWAS) has
and these, in turn, provided an op- genes, with the data ideally shared substantially accelerated the pace
portunity to identify rare mutations publicly. Large consortium efforts of discovery of common genetic
that confer a high risk of cancer in are under way to publicly annotate susceptibility variants for a wide
other family members. The concept and classify iconic and rare can- range of human diseases and traits
of penetrance – i.e. the likelihood cer predisposition genes, such as (Box 3.2.1). The previous decades
that other family members carrying BRCA1 and BRCA2, based on the of candidate gene studies yielded
the same variant will develop can- accumulation of data from many re- very few results that have withstood
cer – has been intensely studied sources [11]. the rigours of multiple testing. After
in families, yielding estimates in a Until recently, the field was domi- a draft human genome sequence
small subset of genes that genetic nated by reports of families with high and its annotation were available,
counsellors and health-care provid- cancer burdens, not always due to advances in microarray technolo-
ers use to guide patients and family a particular cancer type. In 1969, gies, together with new analytical
members to consider early detec- Li and Fraumeni reported multiple tools and standards, enabled re-
tion or prevention strategies [9]. cancers in families who were later searchers to interrogate hundreds
Many of these genes are now tested determined to harbour loss-of-func- of thousands of SNPs in parallel.
in clinical settings, but the number tion mutations in TP53 [12]. Somatic The resultant success of GWAS
of variants identified has exceeded mutations in TP53 are common in has been based on an agnostic
the threshold for adequate inter- many adult cancers and constitute approach to the discovery of mark-
pretation [10]. Consequently, many the most common set of drivers [13]. ers, based primarily on statistical
variants are known as variants of For the set of more than 120 known grounds [4]. Rarely does a GWAS
unknown significance, and further cancer predisposition genes, it is es- initially find the causal or functional
work is required before classifica- timated that more than 50% are also variant [17]. This is because SNP
tion can be determined – as either somatically altered in cancers, serv- microarrays have been designed
a pathogenic mutation or a benign ing as key drivers of carcinogenesis to provide varying degrees of cov-
mutation [10]. These two categories [2]. Population and clinic-based se- erage of the blocks of haplotypes
are key for clinicians to recommend quencing (targeted to cancer genes, across the genome with optimal
next steps when encountering exomes, and whole genomes) has genetic surrogate markers, which
these variants in families or genetic shown that the prevalence of cancer usually do not include the functional
testing venues (see Chapter 6.5). gene mutations could be higher than variant or variants.
The advent of next-generation anticipated, suggesting that not all In GWAS, many statistical tests
sequencing has accelerated the mutations alone confer cancer risk are conducted, raising the spectre of
pace of discovery of new genes in [14,15]. Even highly penetrant muta- false positives. The community has
which one or more mutations can tions are complex and are modified embraced a threshold of genome-
confer an increased risk of can- by environmental factors and other wide significance for reporting
cer. More than 120 such genes genetic factors, which are not yet GWAS results, defined as a trend
have already been identified, and well explained. In some settings, the association test with P ≤ 5 × 10 −8
the expectation is that more will presence of pathogenic mutations is after adjustment as per the GWAS
be discovered [2]. However, not all much higher than expected [16]. study design [18]. Follow-up studies
Box 3.2.1. Current status of genome-wide association studies.
1. Discovery of new regions in 2. Clues for mechanistic insights 3. Challenge of genetic markers
the genome associated with into the contribution of com- for risk prediction for individ-
diseases or traits mon genetic variation to can- ual or public health decisions
• New candidate genes and cer biology • Common variants represent
regions • Etiology a fraction of the genetic con-
• Gene–environment/lifestyle tribution to risk
interactions • Polygenic risk models
• Outcomes and
pharmacogenomics
156
or large meta-analyses are required or newly generated next-generation odds ratios of 1.1–1.3; in paediatric
to establish a conclusive finding. sequencing of populations [20]. cancers, estimates of 1.6–1.8 are
Independent replication guards not unusual – this may be related
against the pursuit of false posi- Discoveries from cancer GWAS to their rapid development but could
tives; this is particularly important Cancer GWAS are scalable with also be due to the homogeneity of
because mapping and laboratory respect to discovery. Large inter- the tumours studied [23]. In tes-
investigation are expensive with national collaborative efforts have ticular cancer, a disease that has a
respect to time and resources. The yielded the discovery of more than very high heritability but is relatively
actual functional marker does not 1000 independent loci (specific rare, the per-allele effect estimate
have to be tested; instead, a sur- regions harbouring one or more is greater than 2.5 for KITLG on
rogate in linkage disequilibrium can functional variants) in at least 30 chromosome 12 [24].
be replicated in subsequent studies different cancer types. The larger More than 85% of the loci
(Fig. 3.2.2). Occasionally, a com- consortia for breast cancer and identified in cancer GWAS have
mon genetic marker may point to- prostate cancer, two of the most been discovered in individuals of
wards a less common variant with a common cancer types, have es- European ancestry, with approxi-
stronger effect, known as a synthet- tablished more than 180 distinct mately 10% in Asian ancestry and
ic association [19]. Because GWAS regions in each of these cancer less than 5% in African ancestry
genotyping has been performed types, and each region harbours an [4]. This is not surprising, because
with different commercial and cus- allele with a small effect [21,22]. most studies to date have been
tom SNP microarrays, techniques Cancer GWAS have discovered conducted in cases and controls
for imputation of data have been common susceptibility alleles. To of European ancestry. Because
CHAPTER 3.2
SECTION 3
developed to combine data sets. date, nearly all markers discov- the population genetics of differ-
Imputation programs successfully ered by cancer GWAS have a mi- ent continental ancestry can yield
infer untested and highly correlated nor allele frequency greater than different allele frequencies, which
SNPs based on reference data sets, 10%, with a handful in the 5–10% are key for discovery, a small frac-
such as the International HapMap range. The per-allele estimated ef- tion appear to be specific to distinct
Project, the 1000 Genomes Project, fect sizes are small, with estimated populations. However, with further
fine-mapping, it is likely that most
signals from GWAS will yield one or
more SNPs in distinct populations.
Fig. 3.2.2. Genetic analysis of a genome-wide association study. Multiple steps
are conducted, including the choice of single-nucleotide polymorphisms (SNPs) With rare exceptions, the etiolog-
across the genome (usually included on a commercial SNP microarray) based on ical markers are not associated with
linkage disequilibrium in a region, enabling the selection of a surrogate to test for clinical outcomes, including meta-
the region. Association analysis is conducted in a case–control setting, examining static disease or survival. Several
all SNPs in a Manhattan plot, followed by replication analyses that pinpoint markers of the markers for neuroblastoma
on chromosomes, which are fine-mapped and investigated in the laboratory. MAF, appear to discriminate between ag-
minor allele frequency.
gressive and milder disease [25].
Of the more than 180 independent
loci identified for prostate cancer,
not one accurately discriminates be-
tween aggressive and non-aggres-
sive prostate cancer [22].
Investigation of cancer GWAS

susceptibility alleles
Although the pace of discoveries
from GWAS has accelerated with
large collaborative networks, the
investigation of each individual sus-
ceptibility locus has not advanced
at a comparable speed; therefore,
the ability to gain new mechanistic
insights has lagged behind [17]. This
is because it is necessary to con-
duct a series of studies to determine
the variants that are actually respon-
sible for the functional effect identi-
fied in the large population-based

GWAS, based on correlated mark- Consortium (ICGC) and the Cancer py, such as human leukocyte anti-
ers. Moreover, most variants map to Genome Atlas (TCGA), have laid gen (HLA) alleles or neo-antigens.
non-coding sequences, and in the the foundation for understanding Landscape analyses of events
more than 40 susceptibility alleles the scope and complexity of cancer across entire cancer genomes have
that have been well investigated, genomes and have already identi- revealed several key points: a wide
the vast majority confer effect by fied many driver mutations (defined range of types of genetic events
altering the regulation of expression as mutations that initiate or perpetu- (from single base mutations to the
or function of one or more genes ate carcinogenesis). Building on the shattering of entire chromosomes),
nearby [26]. Only a handful of vari- many involving driver genes, and
success of these consortia, investi-
ants appear to map to actual coding even more mutations that appear
gators worldwide are continuing to
changes, resulting in non-synony- to be passengers, arising as a con-
search for distinct characteristics
mous base changes, which lead to sequence of the sloppy proliferative
an alteration of an amino acid. In this in rare and common cancer types
process of cancer genomes [13,28].
regard, one of the major themes of that can shed light on the etiology
Moreover, characterization of can-
cancer GWAS is the appreciation of of cancer, lead to the discovery of
cer genomes has revealed that
the accumulation of many small reg- new targets, and provide a deeper
the origins of cancer are complex.
ulatory changes in cancer etiology, understanding of clinical successes
Although the hallmark processes
unlike the strong effects of highly and failures with known anticancer
of driver genes frequently become
penetrant mutations, which often agents based on genetic mutations dysregulated through somatic alter-
co-occur in known oncogenes or tu- [13]. Accordingly, substantial efforts ations in the genome, many differ-
mour suppressor genes. have been focused on the princi- ent events can occur. Accordingly,
Risk stratification based on ple of precision oncology, i.e. the the list of recurrently mutated can-
many GWAS susceptibility alleles matching of drugs tailored to indi- cer genes is relatively short, but
holds great promise for improving viduals based on specific tumour there are many rarely mutated
screening and prevention strate- mutations [27]. The use of geno- genes (Fig. 3.2.3) [28].
gies, especially for common can- mics to guide therapy has emerged There is substantial heterogene-
cer types with substantial absolute as a major effort in oncology, wheth- ity of cancer mutations across the
risks, such as breast cancer and
er it is defining specific targets for globe, reflecting distinct geographi-
prostate cancer. Recent studies
new drugs or identifying the predic- cal exposures and differences in
have demonstrated the value of
tors of success with immunothera- underlying population ancestry. The
combining data sets of the common
GWAS variants in a polygenic risk
score [8]. The proof of principle has
been established with goodness- Fig. 3.2.3. Mutation rates across cancer types. The frequency of point mutations can
vary by 4 orders of magnitude; the lowest frequencies are found in haematological
of-fit tests in breast cancer, showing and paediatric tumours and the highest in tumours induced by carcinogens such as
that the polygenic risk score can be tobacco smoke and ultraviolet radiation. The patterns and signatures of the distribution
calibrated and predicts risk accu- of base changes are highly variable and can reflect distinct environmental or genetic
rately in the tails of the highest and mechanisms (e.g. homologous recombination deficiency of the APOBEC family of
lowest risk distribution. It is likely cytidine deaminases).
that the polygenic risk score, com-
bined with classic epidemiological
risk factors, will drive major advanc-
es in early detection and prevention
strategies during the next decade.
The landscape of
mutational changes in
cancer genomes
The application of next-genera-
tion sequencing technology to the
analysis of somatic mutations in
cancer genomes has transformed
the understanding of cancer, be-
ginning with the identification of
key drivers of tumorigenesis. Large
international consortia, such as
the International Cancer Genome
158
identification of a more comprehen- genome) (Table 3.2.1). There is an ways or that, in some cases, sets of
sive set of cancer genes has set extensive “dark matter” space out- genes can contribute to cancer [29].
in motion the process of mapping side the protein-coding regions that Because cancer is a disease
them against different cancer types has emerged from landscape analy- that alters the genome, mutational
and subtypes. Distinct environmen- ses but cannot be easily interpreted. events can range in size from a sin-
tal exposures (e.g. chemical expo- Widely available data sharing within gle nucleotide to an entire chromo-
sures, dietary and lifestyle factors, the research community, albeit with- some [30]. Although gains and loss-
and infections) as well as different in controlled circumstances, is criti- es of entire chromosomes occur in
population genetic backgrounds cal to better understand what has many cancers, it is daunting to sep-
can partially explain the geographi- already been generated, because arate the driver gene events from
cal and biological differences. new algorithms and perspectives those that result from alterations
Mapping genomic features against regularly uncover novel biological in genome structure. Previously, a
different environmental exposures processes underlying carcinogen- handful of driver fusion genes had
should lead to new discoveries and esis, especially with respect to can- been identified in elegant molecu-
eventually generate new approach- cers across the globe [29]. lar genetics studies. An example
es to early detection or prevention. By definition, somatic alterations is the Philadelphia translocation in
A multitude of international ar- arise as a postzygotic event. When chronic myeloid leukaemia cells, in
ticles on landscape analyses have cancer develops, it is because of a which the ABL1 gene on chromo-
detailed the mutational events disruption of one or more key cellu- some 9 is juxtaposed onto the BCR
across a wide range of cancer types lar functions that confer a selective gene on chromosome 22 to yield a
and have begun to reveal important advantage for tumour growth [30] tyrosine kinase signal that is per-
CHAPTER 3.2
SECTION 3
patterns that overlap between dif- (Fig. 3.2.4). Some mutations inacti- petually “on”. Fusion genes have
ferent types of cancer (but not all vate genes that protect the cell from been identified in a wide range of
cancer types); these are known as abnormal growth, known as tumour cancer types. For instance, a sub-
pan-cancer analyses [13]. Major ef- suppressor genes, whereas other stantial fraction of papillary thyroid
forts are under way to catalogue and mutations activate genes that ac- cancer is driven by fusion genes
understand the underlying biology celerate abnormal growth, known as involving the RAS pathway [31].
for the hundreds of cancer genes oncogenes. More recently, studies Concatenation of somatically al-
that have been identified, but to have shown that mutations can also tered regions (either within a chro-
date, most studies have reported on disrupt pathways of expression or mosome or between chromosomes)
protein-coding regions (~2% of the epigenetic regulators of gene path- can occur in most cancer types
Table 3.2.1. Large resources for cancer genomics data
Resource Website Description
International Cancer Genome https://dcc.icgc.org/ The ICGC Data Portal provides access to cancer
Consortium (ICGC) genome data and project data from ICGC members.
The Cancer Genome Atlas (TCGA) https://www.cancer.gov/about- The TCGA Data Portal provides a platform for
nci/organization/ccg/research/ researchers to search, download, and analyse cancer
structural-genomics/tcga genome data sets generated by institutions in the USA
contributing to TCGA.
Genomic Data Commons (GDC), https://portal.gdc.cancer.gov The GDC Data Portal includes data from TCGA and
National Cancer Institute, USA other cancer genome sequencing projects supported
by the National Cancer Institute, as well as analytical
pipelines.
Catalogue of Somatic Mutations in https://cancer.sanger.ac.uk/cosmic COSMIC stores and displays curated somatic mutation
Cancer (COSMIC) data and other information related to human cancer.
Pan-Cancer Analytical Framework https://www.cell.com/pb-assets/ The Pan-Cancer Atlas resource includes many linked
consortium/pancanceratlas/pancani3/ articles detailing analyses across cancer types using
index.html TCGA and ICGC, known as pan-cancer analyses.
Broad Institute Integrative Genomics https://broadinstitute.org/igv/ IGV is a visualization tool for interactive exploration of
Viewer (IGV) large, integrated genomic data sets.
University of California Santa Cruz https://genome-cancer.ucsc.edu The UCSC Cancer Genomics Browser is a suite of
(UCSC) Cancer Genomics Browser web-based tools to visualize, integrate, and analyse
cancer genomics and associated clinical data.

Fig. 3.2.4. Figurative depiction of somatic mutations present in a cancer cell in the adenocarcinoma) and the Sherlock-
small cell lung cancer cell line NCI-H2171. Individual chromosomes are depicted on lung study (of lung adenocarcino-
the outer circle. Concentric circles show point mutations, copy number events, and ma in never-smokers) are conduct-
rearrangements. Arrows indicate distinct types of somatic mutational events.
ing landscape genomic analyses in
common cancer types with distinct
geographical patterns and linking
the epidemiological factors with
signatures of carcinogenesis to un-
derstand geographical differences
in cancer incidence and subtypes.
It is plausible to identify signatures
that point towards an environmental
or lifestyle risk factor that could be
avoided or controlled. Similarly, sig-
natures could be used to determine
driver events that could be targeted
with specific therapeutic strategies,
including immunotherapy.
The pattern of distribution of mu-
tations in the DNA binding region of
TP53, an iconic tumour suppressor
gene, has led to new insights into its
role in responding to cellular stress
and approximating genomic stabil-
ity. The distribution of TP53 muta-
tions varies widely by cancer types,
including by age and by geographi-
cal distribution, which suggests key
opportunities to investigate the role
of environmental triggers in car-
but to varying degrees of density. ultraviolet radiation and melanoma) cinogenesis [35].
Distinct types of structural variants and paediatric cancers (e.g. Ewing
can occur. Chromosomal shat- sarcoma and retinoblastoma) [28]. Future use of genomics
tering, known as chromothripsis, The patterns of specific mutations
can result in thousands of rear- can leave mutational signatures –
in cancer research
rangements that occur in a sin- or footprints – based on the specific In the process of characterizing
gle crisis due to imperfect DNA types of mutations and their adja- cancer genomes as well as cancer
repair mechanisms (see Chapter cent base pair context [32]. Some of susceptibility alleles, it has become
3.4). Similarly, hypermutation of a the signatures have been correlated apparent that as cells divide, they ac-
region can result in kataegis, of- with tobacco use (see Chapter 2.1), cumulate somatic mutations. Recent
ten due to the APOBEC family of exposure to potent mutagens such analyses of normal tissues have
genes. Major shifts in the balance as aflatoxins or aristolochic acid (see shown that mutations can accumu-
between regulators of genes – Chapter 2.8), or host defence sys- late in healthy individuals with age,
i.e. epigenetic mechanisms – have tems (e.g. APOBEC3 genes, which particularly in response to strong en-
emerged as an important driver protect against small pathogens) vironmental mutagens (e.g. ultravio-
in some cancer types, either with [33,34]. New efforts are under way let radiation and the skin, nutritional
overactive methylation (which usu- to search for mutational signatures elements and the oesophagus, and
ally silences a genetic fragment) that could point to novel risk factors inhalants like tobacco smoke and
or with low levels of methylation for specific cancer types by looking the lung) [36,37]. Surprisingly, even
(known as hypomethylation). for epidemiological factors associ- if the mutations are known cancer
Mutational rates vary greatly by ated with specific signatures. The drivers (e.g. in TP53 and NOTCH1),
type of cancer. So far, the density of Mutographs project (which results cancer may not have developed yet;
single-nucleotide mutations across a from a Cancer Research UK Grand this clearly signals that additional
genome differs by nearly 4 orders of Challenge grant and is analysing local tumour microenvironmental
magnitude (> 10 000-fold) between five cancer types: colorectal cancer, and immune interactions contrib-
cancer types with strong environ- kidney cancer, pancreatic cancer, ute to malignant transformation
mental factors (e.g. tobacco use and and both oesophageal squamous [38]. The assessment of genomic
lung adenocarcinoma or exposure to cell carcinoma and oesophageal changes in precancerous states has
160
Fig. 3.2.5. Mosaicism and ageing. Distribution and types of postzygotic somatic events leading to clonal mosaicism, from single
point events to large chromosomal events, always in a subset of cells. Mosaic events can vary by tissue of origin and can be driven
by new mutations that achieve selective advantage balanced by effective or ineffective surveillance repair mechanisms.
CHAPTER 3.2
SECTION 3
tremendous potential for early de- cancer and other complex adult dis- tumour DNA or tumour cells. Early
tection and prevention. eases (e.g. cardiovascular disease, studies suggest that it is possible to
Genetic mosaicism (defined diabetic diseases, or neurodegen- detect circulating DNA in advanced
as the presence of a subpopula- erative diseases) [40,41]. For hae- cases, but major questions remain
tion of cells with an alternative ge- matological cancers, it is possible about the sensitivity and timing of
notype) has been well established to detect a subset of mutations well such diagnostic tools, especially be-
across the spectrum of mutational before the diagnosis of cancer. This cause genetic mosaicism could be
events, generally accumulating is known as clonal haematopoiesis, more common than previously ap-
with age (Fig. 3.2.5) [39]. Whether and it has been shown to be an im- preciated. Large studies will be re-
large structural events increase portant risk factor for subsequent quired to define the utility of a liquid
with age or single base pair muta- leukaemia [42]. biopsy in cancer diagnosis and care.
tions emerge, current research is The technology of next-genera-
focused on how detection of these tion sequencing holds the promise
events could be a biomarker for of detecting either free circulating

References
1. Zuppan P, Hall JM, Lee MK, 11. Cline MS, Liao RG, Parsons MT, Paten 20. Auton A, Brooks LD, Durbin RM, Garrison
Ponglikitmongkol M, King MC (1991). B, Alquaddoomi F, Antoniou A, et al.; EP, Kang HM, Korbel JO, et al.; 1000
Possible linkage of the estrogen recep- BRCA Challenge Authors (2018). BRCA Genomes Project Consortium (2015). A
tor gene to breast cancer in a family with Challenge: BRCA Exchange as a global global reference for human genetic varia-
late-onset disease. Am J Hum Genet. resource for variants in BRCA1 and tion. Nature. 526(7571):68–74. https://doi.
48(6):1065–8. PMID:2035527 BRCA2. PLoS Genet. 14(12):e1007752. org/10.1038/nature15393 PMID:26432245
h t t p s : / / d o i . o r g / 1 0 .1 3 7 1 / j o u r n a l .
2. Rahman N (2014). Realizing the promise 21. Michailidou K, Lindström S, Dennis J,
pgen.1007752 PMID:30586411 Beesley J, Hui S, Kar S, et al.; NBCS
of cancer predisposition genes. Nature.
505(7483):302–8. https://doi.org/10.1038/ 12. Li FP, Fraumeni JF Jr (1969). Soft- Collaborators; ABCTB Investigators;
tissue sarcomas, breast cancer, and ConFab/AOCS Investigators (2017).
nature12981 PMID:24429628
other neoplasms. A familial syndrome? Association analysis identifies 65
3. Chung CC, Chanock SJ (2011). Current new breast cancer risk loci. Nature.
Ann Intern Med. 71(4):747–52. https://
status of genome-wide association stud- 551(7678):92–4. https://doi.org/10.1038/
doi.org /10.7326/0 0 03 - 4819 -71- 4 -747
ies in cancer. Hum Genet. 130(1):59–78. nature24284 PMID:29059683
PMID:5360287
ht t p s: //d o i.o r g /10.10 07/s 0 0 4 3 9 - 011- 22. Schumacher FR, Al Olama AA, Berndt
1030-9 PMID:21678065 13. Bailey MH, Tokheim C, Porta-Pardo E,
SI, Benlloch S, Ahmed M, Saunders EJ,
Sengupta S, Bertrand D, Weerasinghe
4. Park SL, Cheng I, Haiman CA (2018). et al.; Profile Study; Australian Prostate
A, et al.; MC3 Working Group; Cancer Cancer BioResource (APCB); IMPACT
Genome-wide association studies of Genome Atlas Research Network (2018). Study; Canary PASS Investigators; Breast
cancer in diverse populations. Cancer Comprehensive characterization of can- and Prostate Cancer Cohort Consortium
Epidemiol Biomarkers Prev. 27(4):405–17. cer driver genes and mutations. Cell. (BPC3); PRACTICAL (Prostate Cancer
https://doi.org/10.1158/1055-9965.EPI-17- 174(4):1034–5. https://doi.org/10.1016/j. Association Group to Investigate Cancer-
0169 PMID:28637795 cell.2018.07.034 PMID:30096302 Associated Alterations in the Genome)
5. Risch NJ (2000). Searching for genetic Consortium; Cancer of the Prostate in Swe
14. Carlo MI, Mukherjee S, Mandelker D,
determinants in the new millennium. den (CAPS); Prostate Cancer Genome-
Vijai J, Kemel Y, Zhang L, et al. (2018).
Nature. 405(6788):847–56. https://doi. wide Association Study of Uncommon
Prevalence of germline mutations in
org/10.1038/35015718 PMID:10866211 Susceptibility Loci (PEGASUS); Genetic
cancer susceptibility genes in patients Associations and Mechanisms in Oncology
with advanced renal cell carcinoma. (GAME-ON)/Elucidating Loci Involved in
6. Park JH, Gail MH, Weinberg CR, Carroll
JAMA Oncol. 4(9):1228–35. https:// Prostate Cancer Susceptibility (ELLIPSE)
RJ, Chung CC, Wang Z, et al. (2011).
doi.org/10.10 01/jamaoncol.2018.1986 Consortium (2018). Association anal-
Distribution of allele frequencies and effect
PMID:29978187 yses of more than 140,000 men iden-
sizes and their interrelationships for com-
mon genetic susceptibility variants. Proc tify 63 new prostate cancer suscepti-
15. Mandelker D, Zhang L, Kemel Y, Stadler
Natl Acad Sci U S A. 108(44):18026–31. bility loci. Nat Genet. 50(7):928–36.
ZK, Joseph V, Zehir A, et al. (2017).
https://doi.org/10.1073/pnas.1114759108 https://doi.org/10.1038/s41588-018-0142-8
Mutation detection in patients with ad-
PMID:29892016
PMID:22003128 vanced cancer by universal sequencing
of cancer-related genes in tumor and 23. Machiela MJ, Grünewald TGP, Surdez
7. Zhang Y, Qi G, Park JH, Chatterjee N
normal DNA vs guideline-based germ D, Reynaud S, Mirabeau O, Karlins E,
(2018). Estimation of complex effect-size
line testing. JAMA. 318(9):825–35. et al. (2018). Genome-wide association
distributions using summary-level statis-
https://doi.org/10.1001/jama.2017.11137 study identifies multiple new loci as-
tics from genome-wide association stud- sociated with Ewing sarcoma suscep-
PMID:28873162
ies across 32 complex traits. Nat Genet. tibility. Nat Commun. 9(1):3184. https://
50(9):1318–26. https://doi.org/10.1038/ 16. de Andrade KC, Frone MN, Wegman- doi.org /10.103 8/s 41467- 018 - 0 5 5 37-2
s41588-018-0193-x PMID:30104760 Ostrosky T, Khincha PP, Kim J, Amadou PMID:30093639
A, et al. (2019). Variable population
8. Mavaddat N, Michailidou K, Dennis J, 24. Kanetsky PA, Mitra N, Vardhanabhuti S,
prevalence estimates of germline TP53
Lush M, Fachal L, Lee A, et al.; ABCTB Li M, Vaughn DJ, Letrero R, et al. (2009).
variants: a gnomAD-based analysis.
Investigators; kConFab/AOCS Inves- Common variation in KITLG and at 5q31.3
Hum Mutat. 40(1):97–105. https://doi.
tigators; NBCS Collaborators (2019). predisposes to testicular germ cell can-
org/10.1002/humu.23673 PMID:30352134
Polygenic risk scores for prediction of cer. Nat Genet. 41(7):811–5. https://doi.
breast cancer and breast cancer sub- 17. Freedman ML, Monteiro AN, Gayther org/10.1038/ng.393 PMID:19483682
types. Am J Hum Genet. 104(1):21–34. SA, Coetzee GA, Risch A, Plass C, et
https://doi.org/10.1016/j.ajhg.2018.11.002 25. Capasso M, Devoto M, Hou C,
al. (2011). Principles for the post-GWAS
PMID:30554720 Asgharzadeh S, Glessner JT, Attiyeh EF,
functional characterization of cancer risk
et al. (2009). Common variations in BARD1
loci. Nat Genet. 43(6):513–8. https://doi.
9. Rehm HL, Berg JS, Brooks LD, influence susceptibility to high-risk neuro-
org/10.1038/ng.840 PMID:21614091
Bustamante CD, Evans JP, Landrum blastoma. Nat Genet. 41(6):718–23. https://
MJ, et al.; ClinGen (2015). ClinGen – the 18. Chanock SJ, Manolio T, Boehnke M, doi.org/10.1038/ng.374 PMID:19412175
Clinical Genome Resource. N Engl J Med. Boerwinkle E, Hunter DJ, Thomas G, 26. Bigot P, Colli LM, Machiela MJ, Jessop
372(23):2235–42. https://doi.org/10.1056/ et al.; NCI-NHGRI Working Group on L, Myers TA, Carrouget J, et al. (2016).
NEJMsr1406261 PMID:26014595 Replication in Association Studies (2007). Functional characterization of the 12p12.1
Replicating genotype-phenotype associa- renal cancer-susceptibility locus impli-
10. Tavtigian SV, Greenblatt MS, Goldgar DE, tions. Nature. 447(7145):655–60. https:// cates BHLHE41. Nat Commun. 7(1):12098.
Boffetta P; IARC Unclassified Genetic doi.org/10.1038/447655a PMID:17554299 https://doi.org/10.1038/ncomms12098
Variants Working Group (2008). Assessing
PMID:27384883
pathogenicity: overview of results from 19. Goldstein DB (2011). The importance
the IARC Unclassified Genetic Variants of synthetic associations will only 27. Collins FS, Varmus H (2015). A new initia-
Working Group. Hum Mutat. 29(11):1261–4. be resolved empirically. PLoS Biol. tive on precision medicine. N Engl J Med.
ht t p s: //d o i.o r g /10.10 0 2 / h u m u. 2 0 9 0 3 9(1):e1001008. https://doi.org/10.1371/ 372(9):793–5. https://doi.org/10.1056/
PMID:18951436 journal.pbio.1001008 PMID:21267066 NEJMp1500523 PMID:25635347
162
28. Lawrence MS, Stojanov P, Polak P, Kryukov 33. Alexandrov LB, Ju YS, Haase K, Van 38. Chanock SJ (2018). The paradox of muta-
GV, Cibulskis K, Sivachenko A, et al. (2013). Loo P, Martincorena I, Nik-Zainal S, et tions and cancer. Science. 362(6417):893–
Mutational heterogeneity in cancer and the al. (2016). Mutational signatures asso- 4. https://doi.org/10.1126/science.aav5697
search for new cancer-associated genes. ciated with tobacco smoking in human PMID:30467157
Nature. 499(7457):214–8. https://doi.org/ cancer. Science. 354(6312):618–22.
10.1038/nature12213 PMID:23770567 39. Machiela MJ, Chanock SJ (2017). The age-
https://doi.org/10.1126/science.aag0299
ing genome, clonal mosaicism and chronic
29. Hoadley KA, Yau C, Hinoue T, Wolf DM, PMID:27811275
disease. Curr Opin Genet Dev. 42:8–13.
Lazar AJ, Drill E, et al. (2018). Cell-of- https://doi.org/10.1016/j.gde.2016.12.002
34. Turesky RJ, Yun BH, Brennan P, Mates
origin patterns dominate the molecular PMID:28068559
D, Jinga V, Harnden P, et al. (2016).
classification of 10,000 tumors from 33
types of cancer. Cell. 173(2):291–304.e6. Aristolochic acid exposure in Romania
40. Jacobs KB, Yeager M, Zhou W, Wacholder
https://doi.org/10.1016/j.cell.2018.03.022 and implications for renal cell carcinoma.
S, Wang Z, Rodriguez-Santiago B, et
PMID:29625048 Br J Cancer. 114(1):76–80. https://doi.
al. (2012). Detectable clonal mosaicism
org/10.1038/bjc.2015.402 PMID:26657656 and its relationship to aging and can-
30. Stratton MR, Campbell PJ, Futreal PA
(2009). The cancer genome. Nature. 35. Ding L, Bailey MH, Porta-Pardo E, cer. Nat Genet. 44(6):651–8. https://doi.
458(7239):719–24. https://doi.org/10.1038/ Thorsson V, Colaprico A, Bertrand D, et org/10.1038/ng.2270 PMID:22561519
nature07943 PMID:19360079 al. (2018). Perspective on oncogenic pro- 41. Forsberg LA, Gisselsson D, Dumanski JP
cesses at the end of the beginning of can- (2017). Mosaicism in health and disease –
31. Agrawal N, Akbani R, Aksoy BA, Ally
cer genomics. Cell. 173(2):305–320.e10. clones picking up speed. Nat Rev Genet.
A, Arachchi H, Asa SL, et al.; Cancer
https://doi.org/10.1016/j.cell.2018.03.033 18(2):128–42. https://doi.org/10.1038/
Genome Atlas Research Network (2014).
Integrated genomic characterization PMID:29625049 nrg.2016.145 PMID:27941868
of papillary thyroid carcinoma. Cell. 36. Martincorena I, Campbell PJ (2015). 42. Abelson S, Collord G, Ng SWK, Weissbrod
159(3):676–90. https://doi.org/10.1016/j. Somatic mutation in cancer and nor-
cell.2014.09.050 PMID:25417114 O, Mendelson Cohen N, Niemeyer E, et
mal cells. Science. 349(6255):1483–9. al. (2018). Prediction of acute myeloid leu-
32. Alexandrov LB, Nik-Zainal S, Wedge DC, https://doi.org/10.1126/science.aab4082 kaemia risk in healthy individuals. Nature.
Aparicio SA, Behjati S, Biankin AV, et al.; PMID:26404825
CHAPTER 3.2
SECTION 3
559(7714):400–4. https://doi.org/10.1038/
Australian Pancreatic Cancer Genome s41586-018-0317-6 PMID:29988082
Initiative; ICGC Breast Cancer Consortium; 37. Martincorena I, Fowler JC, Wabik A,
ICGC MMML-Seq Consortium; ICGC Lawson ARJ, Abascal F, Hall MWJ, et al.
PedBrain (2013). Signatures of mutation- (2018). Somatic mutant clones colonize
al processes in human cancer. Nature. the human esophagus with age. Science.
500(7463):415–21. https://doi.org/10.1038/ 362(6417):911–7. https://doi.org/10.1126/
nature12477 PMID:23945592 science.aau3879 PMID:30337457

3.3
Gene–environment interactions
The preventive implications are
still not clear
Paolo Vineis Anja Rudolph (reviewer)
Ghislaine Scelo (reviewer)
●● Approaches to achieving optimal highly penetrant inherited mutations.

SUMMARY prevention are still debated and The remaining cancers are due to
include a stratified or precision environmental agents, exposure to
●● Genetic susceptibility is related prevention approach that focus- endogenous carcinogens, or the
to changes in gene structure or es on high-risk populations. interaction between weak genetic
function that predispose to dis- susceptibility and external or endog-
ease, including cancer. enous agents (Fig. 3.3.1).
●● Generally, about 5–10% of all What is genetic Structural changes, in the form of
base substitutions in the sequence of
cancers are estimated to be susceptibility?
due to highly penetrant inher- DNA, can have higher or lower pen-
ited mutations. The remaining Genetic susceptibility is related to etrance, and hence have a higher
cancers are due to environ- changes in gene structure or func- or lower strength of association with
mental agents, exposure to en- tion that predispose to disease, in- disease. Rare variants, indicated by
dogenous carcinogens, or the cluding cancer. Here, only structural minor allele frequency lower than 1%,
interaction between weak ge- changes are considered; suscepti- are called mutations and tend to have
netic susceptibility and external bility due to epigenetic modifications high penetrance. Common variants,
or endogenous agents. is not addressed (see Chapter 3.8). as described by single-nucleotide
Gene–environment interactions oc- polymorphisms (SNPs), have low
●● Some gene–environment interac- cur when different genotypes, as in- penetrance (i.e. their association
tions are due to low-penetrance
dicated by gene variants, respond to with cancer is weaker). Examples of
gene variants as indicated by sin-
environmental variation in different rare variants are inherited mutations
gle-nucleotide polymorphisms.
ways. Generally, about 5–10% of all in the BRCA1 gene predisposing to
●● Phenotypes described in relation cancers are estimated to be due to breast cancer or in the RB1 gene
to the key characteristics of car-
cinogens can be modulated by
single-nucleotide polymorphisms. Fig. 3.3.1. The risk of cancer and degenerative diseases is determined by a complex
●● An example of gene–environ- interplay of genetic and environmental factors. The contribution of genetic factors to
the risk varies, but several lines of evidence show that non-genetic (“environmental”)
ment interactions is the carcino- factors have high attributable risks, often in the range of 80–90%.
genicity of alcohol, specifically
in relation to ADH and ALDH
gene variants. The strength of ‘Environmental’ Risk
association between ALDH var-
iants and aerodigestive cancers
is such that ALDH has been
successfully used in Mendelian
randomization studies.
10% Cancer and
degenerative diseases 90%
●● The assessment of causality
is not straightforward, and few
gene–environment interactions
in cancer have been replicated
in a convincing way. Genetic Risk
164
predisposing to retinoblastoma (see types were also discovered that

Chapter 3.2). In this chapter, low- predisposed individuals to the ac- FUNDAMENTALS
penetrance variants as indicated by tion of carcinogens. Examples are
SNPs are considered. N-acetyltransferase 2 (NAT2) and ■■ Many gene variants that
SNPs can occur in all genes in- its modulation of the risk of blad- interact with environmental
volved in the modulation of the ef- der cancer in subjects exposed to
agents have been identified.
fects of environmental agents. For aromatic amines, and the modula-
historical reasons, the most stud- tion of outcomes from exposure to However, the assessment
ied SNPs are in genes involved in polycyclic aromatic hydrocarbons of causal evidence is often
carcinogen metabolism and in DNA due to CYP1A1 variants [2]. uncertain, because of the very
repair. However, expression of all The literature grew exponen- large sample sizes required to
key characteristics of carcinogens tially when gene variants related to investigate interactions.
[1] (see Chapter 3.11) can be mod- metabolic phenotypes were discov-
ulated by SNPs. For each of the ered and polymerase chain reaction ■■ For each of the key character-
key characteristics of carcinogens, (PCR) techniques enabled system- istics of carcinogens, genes
which in some instances loosely atic searches to be done for such with inherited variants can be
correspond to the hallmarks of can- candidate variants in populations. found, but the real impact of
cer, there are examples of genes Given the large amount of evidence,
these variants in modulating
whose SNPs may modulate the this chapter refers to reviews and
mechanism of action (Table 3.3.1). presents some examples of gene– the effect of environmental
In the early phases of gene– environment interactions. An early exposures is largely unknown.
CHAPTER 3.3
SECTION 3
environment interaction studies, review was published by IARC in
■■ The gene–environment
genotyping was not available, and 1999 [2], but much more evidence is
interactions investigated most
evidence came from a phenotypic currently available. To synthesize the
characterization of susceptibility. evidence, a set of criteria – known as frequently have included
People were known to react differ- the Venice criteria because they were environmental factors
ently to drugs, including with re- proposed by the Human Genome categorized as energy
spect to adverse effects, because Epidemiology (HuGE) Network at a balance (e.g. indicated by
of more or less rapid metabolism, meeting in Venice – is used [3]. The body mass index, diet),
usually related to enzymes of criteria assess the quality of the evi-
exogenous hormonal factors
the cytochrome P450 (CYP) sys- dence based on three general cate-
(e.g. oral contraceptives),
tem, often identified as members gories: amount of evidence, degree of
of the CYP family. Some pheno- replication, and protection from bias. endogenous hormonal factors
(e.g. indicated by menopausal
status), particular chemical
Table 3.3.1. Key characteristics of carcinogens and examples of genes with low-pene- exposures (e.g. consumption
trance variants (single-nucleotide polymorphisms) that may modulate the mechanism of grilled meats), and lifestyle
of action
factors (e.g. smoking, alcohol
Key characteristic Examples of genes consumption).
1. Is electrophilic or can be Phase I (CYP) or phase II (NAT2, GSTM1) ■■ The magnitudes of the
metabolically activated to metabolizing genes
electrophiles
interactions reported were
usually modest, with risks
2. Is genotoxic DNA repair genes (e.g. XRCC1)
increased or decreased by
3. Alters DNA repair or causes DNA repair genes
genomic instability 20–50%.
4. Induces epigenetic alterations Genes involved in DNA methylation or histone ■■ There are very few examples
acetylation
of actionable gene–environ-
5. Induces oxidative stress OGG1 ment interactions prompting
6. Induces chronic inflammation Interleukin-1 gene family specific prevention strategies,
7. Is immunosuppressive Several genes involved in immunosuppression partly because a large number
8. Modulates receptor-mediated effects AHRR of people at a small risk
9. Causes immortalization Genes involved in senescence (e.g. pRB and
may give rise to more cases
p53 cell-cycle control pathways) of disease than the small
10. Alters cell proliferation, cell death, NOTCH1 number who are at a high risk.
or nutrient supply
Chapter 3.3 • Gene–environment interactions 165

ADH and ALDH, aero Fig. 3.3.2. Smoke from tobacco cigarettes is a major source of human exposure to
polycyclic aromatic hydrocarbons. Through gene–environment interaction studies,
digestive cancers, and some phenotypes were discovered that predisposed individuals to the action of
Mendelian randomization carcinogens, including the modulation of outcomes from exposure to polycyclic
One example that has been stud- aromatic hydrocarbons due to CYP1A1 variants.
ied extensively and belongs to the
highest categories according to the
Venice criteria is the different ability
that individuals have to metabolize
ethanol to acetaldehyde. Alcohol
consumption has been associated
with the risk of cancer at different
organ sites (see Chapter 2.3), and
acetaldehyde is believed to be the
active agent.
Individuals have different sus-
ceptibilities to the acute effects of
ethanol (notably, some people of
Asian descent are particularly sus-
ceptible), and this has been related
to common variants as indicated by
SNPs of the alcohol dehydrogenase
(ADH) and aldehyde dehydroge-
nase (ALDH) genes. Such variants
are also associated with greater sus- ining, as an example, the association Fig. 3.3.3. People in Turbo, a small town
ceptibility to the carcinogenic effects between the ALDH2 polymorphisms in Kenya, brew a traditional alcoholic
of ethanol, for example for laryngeal and oesophageal cancer. beverage called changaa. Individuals
cancer and oesophageal cancer. have different susceptibilities to the acute
The ALDH2*2 allele produces an
In one study, six ADH gene vari- effects of ethanol, and this has been
inactive protein, which is unable to related to common variants as indicated
ants were investigated in more than metabolize acetaldehyde. An indi- by SNPs of the ADH and ALDH genes.
3800 people with aerodigestive vidual’s genotype at this locus may
cancer and 5200 controls [4]. The influence their risk of developing
gene variants rs1229984 (ADH1B) oesophageal cancer via two mecha-
and rs1573496 (ADH7) were signifi- nisms: by influencing alcohol intake,
cantly protective against aerodiges- and by influencing acetaldehyde
tive cancers. These effects became levels. In a meta-analysis of studies
more apparent with increasing alco- investigating the ALDH2 genotype
hol consumption. The gene effects and oesophageal cancer, the risk
were independent of each other, im- was reduced among *2*2 homozy-
plying that multiple ADH genes may gotes (odds ratio, 0.36; 95% confi-
be involved in the etiology of upper dence interval, 0.16–0.80) and in-
aerodigestive cancers. creased among heterozygotes (odds
ADH gene variants have been ratio, 3.19; 95% confidence interval,
included in studies on alcoholism 1.86–5.47) relative to *1*1 homozy-
based on a Mendelian randomiza- gotes. This provides evidence that
tion design (e.g. [5]). In turn, ALDH acetaldehyde plays a carcinogenic
variants have been successfully in- role in oesophageal cancer [6].
vestigated in relation to aerodigestive Mendelian randomization can
cancers with Mendelian randomiza- also be used to clarify dose–re-
tion. In brief, gene variants are trans- sponse relationships. For example,
mitted randomly from parents to their the relationship between alcohol
offspring, because of random assort- consumption (using a variant in the
ment in meiosis. Therefore, they are ADH1B gene as an instrumental vari-
expected not to be affected by con- able) and risk of cardiovascular dis-
founding in epidemiological studies ease was investigated. Alcohol con-
and are used as instrumental vari- sumption was found to increase risk
ables to assess causality between of cardiovascular disease, with no
environmental exposures and can- evidence of a cardioprotective effect
cer. Here, this is illustrated by exam- at moderate consumption levels [7].
166
A review of the literature of Colorectal Cancer Consortium sociated with dietary habits. Larger
(GECCO). An example is a study on sample sizes are probably needed
Many gene variants that interact
the gene–environment interaction to elucidate modest or weak inter-
with environmental agents have
for use of aspirin and non-steroidal action in GWAS of gene–diet inter-
been identified. However, the as-
anti-inflammatory drugs and risk of action [11]. Potential chemopreven-
sessment of causal evidence is of-
colorectal cancer [10]. tion of colorectal cancer mediated
ten uncertain, because of the very
For colon cancer, several stud- by aspirin and related drugs is not
large sample sizes required to in-
ies have evaluated the role of necessarily an exception, because
vestigate interactions. For each of
gene–diet interactions. Results in this case (in spite of very low P
the key characteristics of carcino- values), the relative risks are about
gens, genes with inherited variants from candidate gene studies were
inconsistent, with little replication 0.66–0.69 for gene variants [10,12].
can be found (Table 3.3.1), but the
real impact of these variants in mod- across studies. In recent years,
ulating the effect of environmental GWAS have identified several colo- Functional interpretation
exposures is largely unknown. rectal cancer susceptibility loci, but The underlying biological mecha-
Simonds et al. [8] performed a limited evidence was provided that nism contributing to disease risk is
systematic review of published litera- these loci may modify the risk as- known for only a small proportion of
ture from two databases of genetic
association studies: the HuGE litera-
ture finder and the Cancer Genome- Fig. 3.3.4. Distribution of the number of gene–environment interactions examined
Wide Association and Meta Analy- by environmental exposure category in (A) primary and (B) supplemental literature
searches. A total of 3526 interactions were examined in the primary search, and 1370
ses Database (Cancer GAMAdb). Of
interactions were examined in the supplemental search from relevant publications.
CHAPTER 3.3
SECTION 3
3019 articles identified in the search-
es, only 272 articles met the inclusion A 1600
criteria. In both searches, the majority 1457 (41%)
Number of Interactions Examined
1400
of the publications examined gene–
environment interactions in cancers 1200
of the colon, rectum, colorectum, 1000

breast, or lung. The interactions ex- 800
760 (22%)
amined most frequently included
600
environmental factors categorized 456 (13%)
393 (11%)
as energy balance (e.g. indicated by 400 283 (8%)
body mass index, diet), exogenous 200 114 (3%)
39 (1%)
hormonal factors (e.g. oral contracep- 0
24 (<1%)
tives), endogenous hormonal factors ce e es t es t

ion
t
tyl en en en
(e.g. indicated by menopausal status), alan ifes m
on nm m
on
atm m
at nm
B L r iro r re iro
particular chemical exposures (e.g. gy Ho nv Ho /T lam nv
ner ous lE ous ugs inf lE
E a r d a
consumption of grilled meats), and en ic
ge
n D an ic
og em n ys
Ex Ch do tio Ph
lifestyle factors (e.g. smoking, alcohol E n
fe
c
consumption) (Fig. 3.3.4). In
Interestingly, the majority of the Environmental Exposure Category
interactions examined used loci B 500
from candidate gene studies, and 456 (33%)
Number of Interactions Examined
450
none of the studies were genome-
400 379 (28%)
wide interaction studies (i.e. studies
350
based on genome-wide association 291 (21%)
300
studies [GWAS]). The magnitudes
250 221 (16%)
of the interactions reported were
200
modest, as is usually the case in
150
the literature on gene–environment
100
interactions in cancer: the risks in-
50 22 (2%)
creased or decreased by 20–50% 1 (<1%)
0
in carriers of the minor allele com- t
es le es ce en ion
pared with wild-type individuals for on esty on lan tm at
rm Lif rm Ba ea m
o r m
the same exposure [9] (some exam- Ho sH
y
s/T la
s
ne
rg
ug inf
ples are given below). More recent- ou nou E r nd
gen ge D
n
a
ly, GWAS gene–environment inter- Ex
o do tio
En fec
action studies have been published In
by the Genetics and Epidemiology Environmental Exposure Category

the loci identified through GWAS (see particular in relation to screening or therefore the motivation to do so is
Chapter 3.2). More research is need- surveillance for chronic diseases. enhanced. Also, the “high-risk” ap-
ed to functionally characterize risk Also for primary prevention, fo- proach generally offers a more cost-
loci. This includes: using functional cusing on individuals who are at effective use of limited resources,
annotations for discovery and valida- higher risk (e.g. because of their and it has a more favourable ratio of
tion; studying molecular phenotypes, genetic background) has been re- the benefits to the risks. (If an inter-
including epigenetics or gene expres- peatedly proposed. A typical exam- vention has some adverse effects,
sion, to improve gene–environment ple is screening for phenylketonuria then the ratio of the benefits to the
interaction discovery; and leveraging at birth, where the detection of a risks will be more favourable if the
in vitro and in vivo models for these particular set of mutations enables benefits are greater.) However, the
studies [13]. Large public databases, the identification of individuals who “high-risk” strategy has drawbacks.
such as the Encyclopedia of DNA will benefit enormously from simple The first disadvantage is related to
Elements (ENCODE), the Epigeno- preventive actions, such as avoiding the difficulties and costs of screen-
mics Roadmap, and Genotype-Tissue phenylalanine in the diet. In this ex- ing individuals to identify those who
Expression (GTEx), enable functional ample, the screening test has high are most susceptible, even with the
annotation and interpretation of many sensitivity and specificity and the more refined measures of suscepti-
genomic regions; this can be used to preventive action is highly effective; bility that result from the improved
prioritize candidate gene–environ- hence, precision prevention is high- molecular understanding of cancer.
ment interaction markers [14]. ly attractive for phenylketonuria. The second disadvantage is that
it is a temporary solution and not
Can genetic susceptibility Sick individuals and sick a definitive – or what Rose called
be used to select high- populations “radical” – solution: with a popula-
tion-based approach the risk fac-
risk populations? The strategic problems of the popu-
lation science of primary prevention tor can in principle be eradicated,
The concept of precision medi- whereas with the “high-risk” strat-
cine has recently attracted signifi- were already addressed in 2001 by
Rose in an article titled “Sick indi- egy it is not. The main problem that
cant attention [15]. As is stated on Rose identified with this approach,
the website of the United States viduals and sick populations” [18].
Rose compared the advantages which is also the case for the con-
National Institutes of Health [16], cept of precision prevention, is that
“Precision medicine is an emerg- and disadvantages of an approach
to prevention that is focused on “a large number of people at a small
ing approach for disease treatment risk may give rise to more cases of
and prevention that takes into ac- high-risk individuals or subgroups –
which today would be termed strati- disease than the small number who
count individual variability in genes, are at a high risk” [18].
environment, and lifestyle for each fied or personalized or precision
prevention – and of the population- Hence, the preference is for pop-
person. While significant advances
based approach. ulation-based approaches, which
in precision medicine have been
The first advantage of the “high- have multiple advantages. They are
made for selected cancers, the
risk” strategy is that it produces definitive, because they attempt to
practice is not currently applied to
interventions that are appropriate remove the underlying causes of
most diseases. Many efforts are
for the particular individuals who disease, and they may lead to large
under way to help make precision
are advised to follow them, and dividends, because they target the
medicine the norm rather than the
exception.” Prevention is mentioned
side by side with treatment, and the
potential impact of environment and Fig. 3.3.5. Blood from the heel of a newborn baby is applied to a card for a phenylke-
tonuria test. If a particular set of mutations is detected, precision prevention can be
lifestyle is also cited.
implemented by avoiding phenylalanine in the diet.
According to Collins and Varmus
[17], “The concept of precision
medicine – prevention and treat-
ment strategies that take individual
variability into account – is not new;
blood typing, for instance, has been
used to guide blood transfusions for
more than a century.” The concept
of taking inter-individual variation
into account – which seems key to
the definition of precision preven-
tion – is indeed an old one: focus-
ing on more susceptible subgroups
has been discussed for decades, in
168
whole population instead of a rela- backs. In particular, it offers only a less susceptible to prostate cancer
tively small fraction of it. Rose used small benefit to each individual, be- or breast cancer, but the risk still re-
data from the Framingham Heart cause most of the treated individuals mains in the residual portion of the
Study to calculate that a lowering of will not develop the disease anyway. population. Second, an intervention
the blood pressure distribution of the This leads to the so-called preven- may be potentially targetable to a
population as a whole by 10 millime- tion paradox: “a preventive measure subgroup in a population but may
tres of mercury would correspond to which brings much benefit to the not be easily applicable in such a se-
a reduction of about 30% in the to- population offers little to each par- lective manner. Therefore, for prag-
tal attributable mortality [18]. Today, ticipating individual” [18]. matic reasons of service delivery, to
the evidence indicates that elimina- achieve effectiveness in a national
tion of certain risk factors such as programme one may have to trade
smoking, and hence a reduction of Conclusions off the precision against the practi-
exposure to the main carcinogenic In general, the literature on gene– calities of the intervention and aim
agents in tobacco smoke, might pre- environment interactions in cancer at everyone. The practicalities of
vent 40–50% of cancers, a goal that contains few convincing and repli- implementation are where the theo-
is not achievable by selecting only cated examples that can be trans- retical strategies of prevention often
high-risk populations [19]. However, ferred into practice. First, risks are fail, even among susceptible sub-
the population-based approach to not all or nothing. One can identify groups, as exemplified by strategies
prevention does have some draw- people who are more susceptible or to encourage smokers to quit [15].
References
CHAPTER 3.3
SECTION 3
1. Smith MT, Guyton KZ, Gibbons CF, Fritz JM, 7. Holmes MV, Dale CE, Zuccolo L, 13. Ritchie MD, Davis JR, Aschard H, Battle A,
Portier CJ, Rusyn I, et al. (2016). Key charac- Silverwood RJ, Guo Y, Ye Z, et al.; Conti D, Du M, et al. (2017). Incorporation
teristics of carcinogens as a basis for orga- InterAct Consortium (2014). Association of biological knowledge into the study
nizing data on mechanisms of carcinogene- between alcohol and cardiovascular dis- of gene-environment interactions. Am
sis. Environ Health Perspect. 124(6):713–21. ease: Mendelian randomisation analysis J Epidemiol. 186(7):771–7. https://doi.
ht t p s: //d o i.o r g /10.128 9 /e h p.15 0 9 912 based on individual participant data. BMJ. org/10.1093/aje/kwx229 PMID:28978191
PMID:26600562 349:g4164. https://doi.org/10.1136/bmj.
g4164 PMID:25011450 14. McAllister K, Mechanic LE, Amos C,
2. Vineis P, Malats N, Lang M, d’Errico A, Aschard H, Blair IA, Chatterjee N, et al.
Caporaso N, Cuzick J, et al., editors (1999). 8. Simonds NI, Ghazarian AA, Pimentel CB, (2017). Current challenges and new op-
Metabolic polymorphisms and susceptibil- Schully SD, Ellison GL, Gillanders EM, et portunities for gene-environment inter-
ity to cancer. Lyon, France: International al. (2016). Review of the gene-environ- action studies of complex diseases. Am
Agency for Research on Cancer (IARC ment interaction literature in cancer: what J Epidemiol. 186(7):753–61. https://doi.
Scientific Publications, No. 148). do we know? Genet Epidemiol. 40(5): org/10.1093/aje/kwx227 PMID:28978193
356–65. https://doi.org/10.1002/gepi.21967
3. Ioannidis JP, Gwinn M, Little J, Higgins 15. Vineis P, Wild CP (2017). The science of
PMID:27061572
JP, Bernstein JL, Boffetta P, et al.; Human precision prevention of cancer. Lancet
Genome Epidemiology Network and 9. Hutter CM, Mechanic LE, Chatterjee Oncol. 18(8):997–8. https://doi.org/10.1016/
the Network of Investigator Networks N, Kraft P, Gillanders EM; NCI Gene- S1470-2045(17)30331-5 PMID:28759370
(2006). A road map for efficient and reli- Environment Think Tank (2013). Gene-
able human genome epidemiology. Nat 16. United States National Institutes of Health
environment interactions in cancer ep-
Genet. 38(1):3–5. https://doi.org/10.1038/ (2017). The future of health begins with
idemiology: a National Cancer Institute
ng0106-3 PMID:16468121 All of Us. https://allofus.nih.gov/about/
Think Tank report. Genet Epidemiol. 37(7):
about-all-us-research-program
643–57. https://doi.org/10.1002/gepi.21756
4. Hashibe M, McKay JD, Curado MP,
PMID:24123198 17. Collins FS, Varmus H (2015). A new initia-
Oliveira JC, Koifman S, Koifman R, et
al. (2008). Multiple ADH genes are as- tive on precision medicine. N Engl J Med.
10. Nan H, Hutter CM, Lin Y, Jacobs EJ, Ulrich 372(9):793–5. https://doi.org/10.1056/
sociated with upper aerodigestive can- CM, White E, et al.; CCFR; GECCO (2015). NEJMp1500523 PMID:25635347
cers. Nat Genet. 40(6):707–9. https://doi. Association of aspirin and NSAID use with
org/10.1038/ng.151 PMID:18500343 risk of colorectal cancer according to ge- 18. Rose G (2001). Sick individuals and sick
netic variants. JAMA. 313(11):1133–42. populations. Int J Epidemiol. 30(3):427–32,
5. Wium-Andersen MK, Ørsted DD, Tolstrup
https://doi.org/10.1001/jama.2015.1815 discussion 433–4. https://doi.org/10.1093/
JS, Nordestgaard BG (2015). Increased al-
PMID:25781442 ije/30.3.427 PMID:11416056
cohol consumption as a cause of alcohol-
ism, without similar evidence for depres- 11. Kantor ED, Giovannucci EL (2015). 19. Vineis P, Wild CP (2014). Global cancer
sion: a Mendelian randomization study. Gene-diet interactions and their impact patterns: causes and prevention. Lancet.
Int J Epidemiol. 44(2):526–39. https://doi. on colorectal cancer risk. Curr Nutr 383(9916):549–57. https://doi.org/10.1016/
org/10.1093/ije/dyu220 PMID:25433705 Rep. 4(1):13–21. https://doi.org/10.1007/ S0140-6736(13)62224-2 PMID:24351322
s13668-014-0114-2 PMID:25844273
6. Lewis SJ, Smith GD (2005). Alcohol,
ALDH2, and esophageal cancer: a meta- 12. Wender RC (2015). Aspirin and NSAID che-
analysis which illustrates the potentials moprevention, gene-environment interac-
and limitations of a Mendelian ran- tions, and risk of colorectal cancer. JAMA.
domization approach. Cancer Epidemiol 313(11):1111–2. https://doi.org/10.1001/jama.
Biomarkers Prev. 14(8):1967–71. https:// 2015.1032 PMID:25781439
doi.org/10.1158/1055-9965.EPI-05-0196
PMID:16103445

3.4
DNA repair and genetic instability

Endogenous and exogenous sources
of damage and hereditary syndromes
Eugenia Dogliotti Janet Hall (reviewer)
Margherita Bignami Jiri Zavadil (reviewer)
suppression. Multiple, redundant Exogenous agents such as

SUMMARY DNA damage repair and signalling ionizing radiation and chemicals
pathways combine to avoid errors damage DNA in a variety of ways.
●● Environmental exposures and during DNA replication and to re- Ionizing radiation and endogenous
reactive species generated dur- move DNA lesions from endogenous oxidizing metabolites induce simi-
ing normal cellular processes or exogenous sources. This chapter lar DNA lesions, although to differ-
can damage DNA, which can highlights the role of DNA repair in ent extents. Ultraviolet radiation,
lead to genetic instability. DNA preventing mutation and cancer de- which is non-ionizing, causes di-
damage repair and signalling velopment and suggests how this merization of adjacent DNA pyrim
pathways operate to maintain knowledge can be exploited for can- idines. Simple alkylating agents
genome integrity. cer prevention and therapy. and polycyclic aromatic hydrocar-
●● Some highly cancer-prone inher- bons generate addition products
ited human diseases are associ- DNA damage and repair (adducts) with DNA bases. In some
ated with DNA repair deficien- pathways cases, second reactions generate
cies. This indicates that cancer DNA interstrand and intrastrand
In the 1920s, well before the struc-
can be a disease of mutation cross-links.
ture of DNA was elucidated, work
resulting from DNA damage. The relative contributions of in-
in Drosophila melanogaster re-
trinsic and extrinsic factors to hu-
●● Mutational analysis of individual vealed that exposure to exogenous
agents, such as ionizing radiation man mutagenesis remain unclear.
cancer genes and sequencing Exogenous carcinogens were long
and chemicals, may cause muta-
of cancer genomes provides di- considered to be the main source of
tions. Only much later was it recog-
rect evidence that DNA insults mutation, but large-scale sequenc-
nized that spontaneous hydrolysis
leave mutational fingerprints on ing of cancer genomes suggests a
and reactive species generated en-
tumour DNA. significant contribution from endog-
dogenously during normal metabo-
●● Environmental factors, heredi- lism are also potentially mutagenic enous DNA damage factors [2].
ty, and random DNA damage all and that this reflects their ability to Several DNA repair pathways
contribute to the burden of can- damage DNA. The human genome provide protection against both
cer mutations. The relative con- sustains approximately 70 000 le- endogenous and exogenous DNA
tributions of these factors are sions per day [1]. The majority are damage. These operate either
currently under investigation. single-strand DNA breaks, which through direct reversal of DNA dam-
arise from oxidation or base loss age or by excision of DNA lesions.
●● Knowing how DNA lesions are via glycosyl bond hydrolysis. Single- Fig. 3.4.2 is a schematic repre-
generated, processed, and re- strand breaks may be converted into sentation of the main DNA repair
paired will continue to provide in- double-strand breaks, a particularly pathways. Nucleotide excision re-
sights and opportunities for can- hazardous form of damage that can pair removes bulky DNA lesions by
cer prevention and treatment. cause cell death or chromosomal two distinct subpathways: global
rearrangements. Furthermore, the genome nucleotide excision repair,
addition, deletion, and incorporation which operates throughout the ge-
Genetic information must be pre- of erroneous bases during DNA rep- nome, and transcription-coupled
served for cellular homeostasis, or- lication contribute to spontaneous nucleotide excision repair, which tar-
ganismal development, and cancer mutation (Fig. 3.4.1). gets transcribed DNA regions [3,4].
170
Fig. 3.4.1. Types of endogenous DNA damage and estimated frequency (per cell per
day) in human cells. The frequencies shown for an abasic site refer to depurination and
depyrimidination events, respectively. 8-oxoG, 8-hydroxyguanine; DSB, double-strand
FUNDAMENTALS
break, SSB, single-strand break.
■■ Many chemical carcinogens
Damage SSB Abasic site 8-oxoG cause tumours as a result of
being metabolized to reactive
intermediates, which may
become covalently bound to
Estimated DNA and give rise to mutation.
frequency 55 000 12 000/600 2800 Carcinogen adducts may be
(/cell/day)
eliminated from DNA in vivo via
a range of enzyme-mediated
Damage Deamination DSB Mismatch DNA repair processes.
■■ Human skin cancers that
are attributable to exposure
to ultraviolet radiation occur
Estimated
frequency 192 25 n.d. at a markedly increased
(/cell/day) rate in individuals with the
autosomal recessive disease
xeroderma pigmentosum, a
CHAPTER 3.4
SECTION 3
Base excision repair removes more very extensive, activation of death
condition arising from defects
subtly damaged DNA bases by ei- pathways occurs. In the context of
in a particular DNA repair
ther short-patch or long-patch base cancer avoidance, both the DNA
damage response and DNA repair pathway. This was the first
excision repair [5,6]. Homologous
recombination and non-homologous play major roles in the maintenance example to indicate the role
end joining repair double-strand of genome stability and in cancer of DNA repair in preventing
breaks [7]. Mismatch repair corrects avoidance [10]. cancer development.
DNA replication errors [8]. ■■ The enzymes that mediate
Homologous recombination, non- DNA repair disorders and DNA repair, and their genes,
homologous end joining, and mis- cancer have been characterized and
match repair contribute to replica- are specific for particular
The formal proof of the underlying
tion fidelity and to the recovery from
role of DNA damage repair in can- categories of DNA damage.
replication fork stalling or collapse.
cer development is the presence of
In the case of lesions that are com- ■■ DNA damage may also occur
germline mutations in specific DNA
plete blocks for DNA replication, spontaneously as a result of
repair or DNA damage response
such as interstrand and intrastrand various biological processes,
genes in cancer-prone hereditary
cross-links, repair is achieved by syndromes (Table 3.4.1). including the production of
subpathways that contain compo- The autosomal recessive dis- reactive oxygen species.
nents of both homologous recom- ease xeroderma pigmentosum was
bination and nucleotide excision re- ■■ Failure of effective DNA repair,
the first example that linked defec-
pair [9]. Direct reversal of damage as exemplified by a range of
tive DNA repair to cancer develop-
is provided by O 6 -methylguanine- heritable syndromes, may
ment. Defects in the global genome
DNA methyltransferase, which trans- contribute to increased mutation
nucleotide excision repair subpath-
fers a methyl group from a pro- way in individuals with xeroderma rates and related chromosomal
mutagenic DNA base to itself, pigmentosum increase sun sensi- structural changes, leading to
and by AlkB human homologues, tivity and skin cancer risk more than tumour development.
which perform dealkylation re- 1000-fold [11]. Defects in transcrip- ■■ Malignant cells have a high
pair of N1-methyladenine and N3- tion-coupled nucleotide excision mutation rate and manifest
methylcytosine [5]. repair are associated with several
chromosomal instability, which
DNA repair is part of a wider pathologies, including ultraviolet-
DNA damage response in which facilitates the development of
sensitive syndrome and severe
DNA damage triggers signalling to drug-resistant cell populations
premature ageing conditions such
a checkpoint response that arrests as Cockayne syndrome and tri- and leads to the failure, in the
cell-cycle progression, inhibits tran- chothiodystrophy. However, these longer term, of some cancer
scription and translation, and initi- syndromes do not exhibit increased therapies.
ates DNA repair. If DNA damage is cancer predisposition.
Chapter 3.4 • DNA repair and genetic instability 171

Fig. 3.4.2. The main DNA repair pathways. (A) Nucleotide excision repair with its two subpathways, global genome nucleotide
excision repair (GG-NER) and transcription-coupled nucleotide excision repair (TC-NER). (B) Base excision repair takes place
by short-patch or long-patch repair. (C) Pathways of double-strand break (DSB) repair: homologous recombination (HR) and non-
homologous end joining (NHEJ). (D) Mismatch repair.
A Nucleotide excision repair B Base excision repair
C Double-strand break repair D Mismatch repair
172
Defects in mismatch repair are pair destabilizes repetitive DNA se- from mechanistic studies in bacteria
associated with both familial and quences that are prone to replication to human cancer”). MUTYH, a base
sporadic colon cancer (see Chapter errors. Frameshift mutations and excision repair DNA glycosylase,
5.5). Colorectal cancer in autoso- microsatellite instability are the hall- participates in the removal of DNA
mal dominant Lynch syndrome (also marks of HNPCC. A milder type of 8-hydroxyguanine, a pre-mutagenic
called hereditary non-polyposis colo- colon cancer predisposition in some lesion. Homozygosity for mutations
rectal cancer [HNPCC]) is caused by cases of familial adenomatous pol- in NTHL1, which encodes a DNA
a germline mutation in a mismatch yposis is associated with mutations glycosylase involved in the base ex-
repair gene (MLH1, MSH2, MSH6, or in the MUTYH gene (see “The 8-hy- cision repair of oxidized pyrimidines,
PMS2) [12]. Defective mismatch re- droxyguanine mutational signature: also causes adenomatous polyposis
Table 3.4.1. Inherited mutations in DNA repair and DNA damage response genes and cancer risk
Syndrome Genes Pathway Tumours Neurological Immunological
abnormalities defects
Xeroderma pigmentosum 7 genes (XPA NER Skin cancer No/Yes No/Yes

to XPG)
MUTYH-associated MUTYH, BER Colorectal cancer and gastric No No

polyposis (MAP) NTHL1 cancer
Lynch syndrome (hereditary MSH2, MMR Colorectal cancer; carcinoma No No
CHAPTER 3.4
SECTION 3
non-polyposis colorectal MSH6, of the stomach, endometrium,
cancer [HNPCC]) MLH1, biliary and pancreatic
PMS2 system, urinary tract
Werner syndrome WRN HR, RFR Soft tissue sarcomas, No No

osteosarcomas,
meningiomas, malignant
melanomas, thyroid
carcinomas
Bloom syndrome BLM HR, RFR Lymphoma, leukaemia, No Yes

carcinoma
Rothmund–Thomson RECQL4 HR, RFR Osteosarcoma, skin cancer No No/Yes

syndrome
Ataxia–telangiectasia ATM DDR Leukaemia, lymphomas, Yes Yes

breast cancer
Ataxia–telangiectasia-like MRE11 DDR Leukaemia, lymphomas, Yes Yes

disorder 1 breast cancer
Nijmegen breakage NBS1 DDR Lymphoid malignancies and Yes Yes

syndrome cancer at multiple sites
Nijmegen breakage RAD50 DDR Lymphoid malignancies and Yes Yes

syndrome-like cancer at multiple sites
Li–Fraumeni syndrome TP53 DDR Multiple primary sites (brain, No No

breast, ovary, prostate,
osteosarcoma)
Seckel syndrome type 1 ATR, ATRIP DDR, RFR Acute myeloid leukaemia Yes Yes
Fanconi anaemia 19 genes ICLR, RFR Acute myeloid leukaemia, Yes/No Yes
(FANCA to squamous cell carcinoma
FANCT)
Hereditary breast and BRCA2, ICLR, RFR Breast cancer and ovarian No No
ovarian cancer BRCA1 cancer
Severe combined Artemis NHEJ Lymphoma No Yes

immunodeficiency with
radiosensitivity (RS-SCID)
DNA ligase IV syndrome LIG4 NHEJ Lymphoma Yes Yes
BER, base excision repair; DDR, DNA damage response; HR, homologous recombination; ICLR, interstrand cross-link repair; MMR, mismatch repair;
NER, nucleotide excision repair; NHEJ, non-homologous end joining; RFR, replication fork repair.

Fig. 3.4.3. Extreme measures taken to protect French children diagnosed with xero The link between
derma pigmentosum from sunlight. This autosomal recessive disease provided the
first evidence that linked defective DNA repair to cancer development.
DNA damage repair,
mutagenesis, and
carcinogenesis
In vitro and in vivo models
Work in Ames’s laboratory con-
firmed the functional link between
carcinogenicity and mutagenicity
[20] and led to the incorporation
of mutagenicity tests into regula-
tory and industrial decision-making.
Knowledge of the importance of
DNA repair in counteracting mu-
tagenesis informed the design of
DNA repair-defective Salmonella
tester strains with increased sen-
sitivity to chemical mutagenesis.
Assays based on cultured mammal
ian cells were developed in parallel.
The bacterial reversion (Ames) as-
say together with the mammalian
and colorectal cancer [13]. Germline erozygous for dominant missense chromosomal aberration, gene mu-
mutations in DNA polymerase δ or ε ATM mutations have a higher risk tation, and micronucleus tests com-
have also been shown to be respon- of breast cancer, colorectal can- prise the standard battery of assays
sible for some types of early-onset cer, and stomach cancer. Somatic of in vitro genotoxicity. These are
colon cancer and endometrial can- ATM mutations or deletions are currently an essential component of
cer characterized by a massive mu- also commonly found in lymphoid the safety assessment of chemicals.
tational burden [14]. Defective repair malignancies and a variety of solid In vitro bacterial or mammalian
of interstrand and intrastrand cross- tumours. Inherited mutations af- cell systems have also been used
links and of double-strand breaks fecting the MRE11–NBS1–RAD50 to determine the relative biologi-
characterizes Fanconi anaemia. complex cause disorders that pres- cal importance of DNA lesions by
Patients with mutations in genes of ent similar clinical and cellular fea- transfecting into host cells plasmid
the Fanconi anaemia pathway have tures to those seen in patients with or viral vectors either globally modi-
growth retardation, infertility, bone ataxia–telangiectasia, although the fied by a DNA-damaging agent or
marrow failure, and a susceptibility to features do not fully overlap. These engineered to contain a single DNA
leukaemia and various solid tumours lesion [21]. Mutational analysis of
disorders include Nijmegen break-
[9]. Inherited mutations significantly chromosomal reporter genes (lacI,
age syndrome and ataxia–telangi-
influence risk of breast cancer and HPRT) also enabled the identifica-
ectasia-like disorder. Patients with
ovarian cancer. Most familial breast tion of specific mutational spectra
Nijmegen breakage syndrome are
and ovarian cancers can be ascribed generated by exposures to DNA-
highly cancer-prone; in ataxia–
to highly penetrant germline muta- damaging agents. The use of cells
telangiectasia-like disorder, the
tions in the BRCA1 or BRCA2 ho- defective in a specific DNA repair
mologous recombination genes [15]. cancer predisposition is somewhat enzyme or expressing a special-
The ATM protein is a major regu- milder [17]. ized DNA polymerase has defined
lator of the DNA damage response. In addition to cancer, defective the roles of specific enzymes as
The importance of the DNA dam- DNA repair is often associated with protectors from damage or inducers
age response in cancer prevention pleiotropic phenotypes including of damage. These basic studies of
is emphasized by the clinical profile immunodeficiency, neurodegenera- mutagenesis have been instrumen-
of individuals with ataxia–telangiec- tion, and developmental abnormali- tal for the decoding of cancer mu-
tasia who carry homozygous ATM ties. This is not surprising, because tational signatures and associated
mutations. In addition to hypersen- several DNA repair proteins con- clinical developments (see below).
sitivity to ionizing radiation, patients tribute to immune development and Animal models provide an alter-
with ataxia–telangiectasia exhibit a tight control of genome stability is native means to explore the con-
chromosomal instability and cancer required for the function of the ner tribution of DNA repair to genome
predisposition, particularly to lym- vous system and the development stability and tumour suppression.
phoid tumours [16]. Individuals het- of the whole organism [18,19]. Nucleotide excision repair-defective
174
animal models have been largely sponsible for the increased risk of cutaneous tumours (see Chapter
used to understand the molecular colorectal cancer. 2.4). The C → A transversion mu-
mechanisms underlying the asso- tations that are characteristic of
ciation between DNA repair defect Mutational signatures in DNA adducts formed by benzo[a]
and cancer risk. However, remark- human cancer pyrene, the major carcinogen in
able differences in these animal Sequencing of human cancer ge- tobacco smoke, comprise the main
models in clinical phenotype and/ nomes revealed a great variation in signature in smoking-associated
or DNA repair abilities weaken their the mutational load among cancer cancers of the lung and larynx. This
use as models of human disease types: the number of mutations per signature is absent in tumours from
[3]. Cancer in patients with HNPCC tumour ranged from 500 in acute never-smokers [29].
is due to heterozygous germline mu- myeloid leukaemia to 100 000 in Examples of mutational signa-
tations, predominantly in MSH2 or melanoma [22,23]. More than 40 tures associated with exposure to
MLH1, and the subsequent somatic years ago, it was hypothesized that genotoxic natural products include
inactivation of the remaining wild- human cancers express a mutator those of aflatoxin B1 and aristolo-
type allele in the colonic epithelium. phenotype, because of the antici- chic acid (see Chapter 2.8). Various
HNPCC mouse models in which in- pated impact of mutations in DNA experimental systems indicate that
activation of mismatch repair genes polymerases and/or repair genes, aflatoxin B1 induces a mutational
is targeted to the intestinal epithe- and as a result of the progressive spectrum dominated by G  → T
lium exhibit a high frequency of in- accumulation of large numbers of transversions. This signature has
testinal adenocarcinomas within mutations during tumour progres- been found in hepatocellular carci-
the first year of life. It is currently sion [24]. This hypothesis has been nomas from regions with possible
CHAPTER 3.4
SECTION 3
unclear why HNPCC mouse models controversial for many years, and exposure to this mycotoxin. Some
develop tumours in the small intes- recently an argument was advanced hepatocellular carcinomas harbour
tine rather than the colorectal can- that the number of stem cell divisions the TP53 R249S G → T transver-
cers that are associated with Lynch alone is sufficient to generate the sion, which occurs in about half of
syndrome in humans. large number of mutations found in the hepatocellular carcinomas of
The effects of mutational in- human tumours, and that increased aflatoxin B1-exposed people with
activation of enzymes in the base mutation rates are not required [25]. hepatitis B virus infection. The vari-
excision repair pathway are more The relative contributions of environ- able prevalence of this mutation
complex. Mice with targeted disrup- mental factors, heredity, and chance is probably due to different levels
tions of DNA glycosylases often ex- (random mutations during DNA rep- of aflatoxin B1 exposure [30]. The
hibit moderately increased mutation lication) are currently a matter of de- mutational signature of aristolochic
frequencies without overt disease. bate (see Chapter 3.1). acid, characterized by A → T trans-
The limited effect of inactivation of Although the origin of mutations versions, was initially associated
single DNA glycosylases is prob- in tumours remains to be firmly es- with upper tract urothelial carcino-
ably due to redundancy in repair tablished, the spectra of mutations mas [31] and, more recently, was
pathways. As a consequence, the in many tumours provide some widely implicated in liver cancer
phenotypes are enhanced in dou- clues. Mutational analysis of indi- (see “The aristolochic acid muta-
ble-knockout mice, affecting back- vidual cancer genes, in particular tional signature in many tumours:
up functions. Therefore, a cancer- TP53, provided the first evidence a warning”). The DNA lesions re-
prone phenotype is observed only that carcinogenic insults leave mu- sponsible for these mutations are
in double-knockout mice deficient tational fingerprints on tumour DNA all substrates for nucleotide exci-
in NTHL1 and NEIL1, two enzymes [26]. A compilation of mutant DNA sion repair, and mutational strand
that repair oxidized pyrimidines and sequences from specific tumour bias is consistent with incomplete
ring-opened purines, with some types has identified mutational sig- repair by this pathway.
overlapping substrate specificities. natures. These define both the type A similar example of overloading
Similarly, only double inactivation and the sequence context of muta- of DNA repair is provided by analy-
of two DNA glycosylases involved tions [23] and provide a record of sis of the genomic landscape of
in the removal of 8-hydroxyguanine the multiple mutagenic processes recurrent glioma in patients treated
from the genome, i.e. OGG1 and that have been operative over the with the chemotherapeutic alkylat-
MUTYH, leads to a cancer-prone lifetime of an individual. ing temozolomide. In this case, loss
phenotype and a shortened life Some mutational signatures re- of expression of the repair enzyme
span (see “The 8-hydroxyguanine flect environmental exposures [27,28]. O 6 -methylguanine-DNA methyl-
mutational signature: from mecha- For example, the distinctive dipy- transferase, which reverses poten-
nistic studies in bacteria to human rimidine mutations known to be as- tially mutagenic DNA methylation
cancer”). However, in humans, sociated with ultraviolet radiation- damage induced by temozolomide,
single germline mutations in the induced DNA lesions comprise the is associated with a characteristic
MUTYH or NTHL1 genes are re- predominant mutational class in G → A mutational signature [32].

The 8-hydroxyguanine mutational signature: from mechanistic studies in bac-

teria to human cancer
Endogenously or exogenously gen- frequent insertion of dAMP op- these mutations by the combined
erated reactive oxygen species in- posite 8-oxoG by replicative DNA action of the base excision repair
duce pre-mutagenic DNA lesions. polymerases causes G:C  → T:A glycosylases OGG1 and MUTYH.
8-Hydroxyguanine (8-oxoG) – one transversions. A three-tier error- Removal of 8-oxoG from 8-oxoG:C
of many oxidized DNA bases – has avoidance repair system, discov- pairs by OGG1 and subsequent
been extensively studied because ered in Escherichia coli [1] and base excision repair restores the
of its miscoding properties. The conserved in eukaryotes, prevents normal G:C base pairing. When
Fig. B3.4.1. Top panel: The three-tier system for removal of 8-hydroxyguanine (8-oxoG). Oxidative stress can introduce
oxidized lesions in DNA. 8-oxoG can be removed by OGG1, and subsequent base excision repair restores the normal G:C
base pairing. In the case of unrepaired 8-oxoG, adenine (A) is misincorporated opposite the 8-oxoG (G*) as a consequence
of inaccurate replication. A removal by MUTYH is followed by resynthesis via long-patch base excision repair by a much less
error-prone DNA polymerase (Polλ). This results in a C:8-oxoG pair, again a substrate for OGG1. Inset: Oxidative damage can
also produce an oxidized pool of dNTPs. MTH1 hydrolyses 8-oxo-dGTP to 8-oxo-dGMP, effectively preventing its incorpora-
tion into DNA. Bottom panel: The mutational signature in MUTYH-associated polyposis tumours identifies the location at which
mutations arise because of unrepaired 8-oxoG:A mispairs by a defective MutY DNA glycosylase. In the bar graphs, the triplets
where the mutation is located (including the 5′ and 3′ bases) are shown on the horizontal axes and the mutation type probability
is shown on the vertical axes. CRC, colorectal cancer.
176
adenine is incorporated opposite of colorectal cancer from patients References

8-oxoG during replication, its re- with MAP offered the unique op- 1. Michaels ML, Miller JH (1992). The GO
moval by MUTYH is followed by portunity to identify a mutational system protects organisms from the mu-
resynthesis by a less error-prone fingerprint of persistent 8-oxoG:A tagenic effect of the spontaneous lesion
8-hydroxyguanine (7,8-dihydro-8-oxo-
polymerase (Polλ). This generates mismatches. A distinct mutational guanine). J Bacteriol. 174(20):6321–5.
OGG1 substrate C:8-oxoG base signature of G:C → T:A transver- https://doi.org/10.1128/jb.174.20.6321-
pairs. A third level of protection is sions (signature 36) was identified 6325.1992 PMID:1328155
provided by the MTH1 hydrolase, in MAP colorectal cancer. This 2. Sieber OM, Lipton L, Crabtree M,
which degrades 8-oxodGTP to the mutational signature is reflected in Heinimann K, Fidalgo P, Phillips RKS,
monophosphate to prevent incor- et al. (2003). Multiple colorectal ad-
the specific pattern of oncogenes enomas, classic adenomatous pol
poration of pro-mutagenic 8-oxo- and tumour suppressor genes in- yp sis, and germ-line mutations in
o
dGMP into DNA. Inactivation of any MYH. N Engl J Med. 348(9):791–9.
volved in colorectal carcinogen-
of these genes confers a mutator https://doi.org/10.1056/NEJMoa025283
esis and associated with inactive PMID:12606733
phenotype.
MUTYH. It is remarkable that the
Although no human disease 3. Viel A, Bruselles A, Meccia E, Fornasarig
MAP-specific signature 36 has M, Quaia M, Canzonieri V, et al. (2017). A
has been associated with defective
never been identified in sporadic specific mutational signature associated
OGG1 or MTH1 activities, germ with DNA 8-oxoguanine persistence
line biallelic MUTYH mutations un- colorectal cancer. However, it was in MUTYH-defective colorectal can-
derlie MUTYH-associated polyp noted that signature 36 [3] closely cer. EBioMedicine. 20:39–49. https://
resembles signature 18 [4], which doi.org/10.1016/j.ebiom.2017.04.022
osis (MAP), a recessively heritable PMID:28551381
colorectal polyposis syndrome is particularly prevalent in neuro-
CHAPTER 3.4
SECTION 3
blastoma and at lower levels in 4. Alexandrov LB, Kim J, Haradhvala
with a predisposition to colorec- NJ, Huang MN, Ng AWT, Wu Y, et al.;
tal cancer. Colorectal cancer in pancreatic cancer, breast cancer, PCAWG Mutational Signatures Working
patients with MAP bears distinc- and gastric cancer. Therefore, it is Group and the ICGC/TCGA Pan-Cancer
possible that oxidative DNA dam- Analysis of Whole Genomes Network
tive somatic G:C → T:A transver-
(2020). The repertoire of mutational sig-
sions in the APC gene [2]. Thus, age also contributes to cancer eti- natures in human cancer. Nature. https://
whole-exome DNA sequencing ology in these organ sites. doi.org/10.1101/322859 (forthcoming)
Disruption of DNA repair path- tumour evolution in a single catas from mechanistic studies in bacte-
ways acting on endogenous lesions trophic event known as chro mo ria to human cancer”).
also leaves a molecular mark on thripsis [34]. These signatures might
the genome and results in specific originate from sporadic bursts of Therapeutic approaches
mutational signatures. The muta- massive endogenous or oncogenic
tional signature associated with
that target DNA repair
stress [2], leading to a temporary
inactive mismatch repair in both Current cancer therapy is based
saturation of the DNA repair capac-
HNPCC and sporadic gastrointes- largely on DNA damage and sat-
ity or to activation of an error-prone
tinal cancers involves the expect- uration of DNA repair in highly
DNA repair pathway or pathways.
ed increase in base substitution proliferative tumour cells. These
Several different mutational sig-
mutations as well as insertions or treatments frequently result in side-
natures can be linked to modifica-
deletions at repetitive sequences. effects, such as secondary tumours
tion of DNA bases occurring spon- and drug resistance. Precision ther-
Similarly, the homologous recombi-
taneously. Specific signatures have apies targeted to cancer-specific
nation pathway was altered in near-
been ascribed to deamination of a DNA repair defects, either by syn-
ly 40% of cancers, for example in
canonical cytosine or 5-methylcyt thetic lethality or by immunothera-
BRCA1/2-mutated ovarian cancers
osine in DNA. In the deamination py, aim to reduce collateral damage
and triple-negative breast cancers
[33]. Because of the central role of of a canonical cytosine, overactiv- to normal cells.
DNA repair and DNA damage re- ity of members of the AID/APOBEC
sponse genes in cell survival after family of cytidine deaminases has Synthetic lethal interaction
DNA damage, mutations in these been implicated [23]. As an exam- In 2005, a description was published
genes, which have been observed ple of endogenous DNA oxidation, of the synthetic lethal interaction
in several tumour types, provide a tumours in which repair by the between mutations in the homolo-
predictive marker of likely therapeu- MUTYH DNA glycosylase was ge- gous recombination genes BRCA1
tic response and clinical outcome. netically impaired bear a signature and BRCA2 and inhibitors of poly-
In some tumours, the majority associated with unrepaired DNA (ADP-ribose) polymerase 1 (PARP1).
of genomic rearrangements appear 8-hydroxyguanine (see “The 8-hy- PARP1 acts as a sensor of DNA
to be acquired at an early stage of droxyguanine mutational signature: single-strand breaks and prevents

The aristolochic acid mutational signature in many tumours: a warning

Aristolochic acids (AAs) are a fam- aristolactam I (dA–AAI), which endemic nephropathy. More re-
ily of carcinogenic, mutagenic, leads to A:T → T:A transversions cently, two groups [1,2] indepen-
and nephrotoxic compounds that in vitro. AAI-induced tumours in ro- dently determined the mutational
are present in plants of the gen- dents show this same transversion signature of AA-exposed upper
era Aristolochia and Asarum (wild mutation in codon 61 of the H-ras tract urothelial carcinomas from
ginger), which are commonly used oncogene, suggesting that dA– Taiwan, China. Both groups ob-
in Chinese herbal medicines (see AAI may be the critical lesion in the served a very high mutation rate
Chapter 2.8). The main compo- carcinogenic process in rodents. in exposed tumours and identified
nents of the plant extract, AAI and These mechanistic key steps, the typical AA mutational signature
AAII, have been shown to form i.e. typical DNA adducts and mu- (A:T → T:A transversions) occur-
DNA adducts after metabolic ac- tation type, have been identified ring preferentially at splice sites.
tivation, preferentially targeting and are consistent with events This signature was then found in a
purines. In vivo the most persis- occurring in patients with upper variety of tumour types, such as re-
tent of these adducts in target tis- tract urothelial carcinomas associ- nal cell carcinoma, intrahepatic bile
sue is 7-(deoxyadenosin-N6 -yl)- ated with AA poisoning and Balkan duct carcinoma, and hepatocellular
Fig. B3.4.2. Left panel: Mechanisms of mutagenesis of aristolochic acid (AA). AA is derived from plants of the genus Aristo
lochia. AAI is shown. The metabolic activation to aristolactam nitrenium ions is followed by DNA binding preferentially to
adenosine and production of specific A:T → T:A transversion mutations. Right panel: Proportion of hepatocellular carcinomas
with the AA signature in various geographical regions. The percentage for South-East Asia comprises data from several
countries, including Viet Nam.
178
carcinoma. In particular, the anal- compounds. Moreover, the hepa- 2. Poon SL, Pang ST, McPherson JR, Yu
tocellular carcinomas from Taiwan, W, Huang KK, Guan P, et al. (2013).
ysis of the role of AA in hepato-
Genome-wide mutational signatures
cellular carcinomas [3] revealed China, that present heavy burdens
of aristolochic acid and its application
that countries in Asia, especially of AA signature mutations may as a screening tool. Sci Transl Med.
Taiwan, China, are highly affect- be good candidates for immune 5(197):197ra101. https://doi.org/10.1126/
ed, and almost half of the hepa- checkpoint inhibitors. scitranslmed.3006086 PMID:23926199
tocellular carcinomas from China 3. Ng AWT, Poon SL, Huang MN, Lim
References
showed the AA signature, consis- JQ, Boot A, Yu W, et al. (2017).
1. Hoang ML, Chen CH, Sidorenko VS, He
tent with exposure through herbal J, Dickman KG, Yun BH, et al. (2013).
Aristolochic acids and their derivatives
are widely implicated in liver cancers
medicines. Because exposure to Mutational signature of aristolochic
acid exposure as revealed by whole- in Taiwan and throughout Asia. Sci
AA seems to be widespread, ad- Transl Med. 9(412):eaan6446. https://
exome sequencing. Sci Transl Med.
ditional measures should be taken 5(197):197ra102. https://doi.org/10.1126/ doi.org/10.1126/scitranslmed.aan6446
to avoid exposure to these harmful scitranslmed.3006200 PMID:23926200 PMID:29046434
their conversion into double-strand not expressed in normal cells (neo- [38]. Indeed, the clinical benefit of
breaks, which are selectively lethal antigens). This renders the tumour anti-PD-1 immune checkpoint in-
in homologous recombination-defec- cells more immunogenic and prone hibitors is correlated with tumour
tive cells [35]. Clinical trials clearly to recognition by cytotoxic T cells. somatic mutation frequency. The
showed that PARP inhibitors, such The burden of neo-antigens is par- efficacy of this approach is not con-
CHAPTER 3.4
SECTION 3
as olaparib, are effective therapy for ticularly high in mismatch repair-defined to mismatch repair-defective
BRCA1/2-mutated cancers. Although ficient tumours with a tendency to tumours. Any tumour with a high
tumour resistance developed in the frameshift mutation. Consistent with somatic mutation burden (these
overwhelming majority of patients, this phenotype, mismatch repair- include mutagen-induced cancers
PARP inhibitor combination regi- defective colorectal cancers re- such as cutaneous cancers and
mens provide promising alternative spond well to the anti-programmed smoking-related non-small cell lung
therapeutic approaches [36]. cell death 1 (PD-1) immune check- tumours) is likely to respond to im-
point inhibitor pembrolizumab [37]. munotherapy, and this approach
Immunotherapy response and Responsiveness is independent of offers considerable promise in the
DNA repair deficiencies the tumour histology and is driven treatment of a significant subgroup
Hypermutated tumours express only by the mutator phenotype as of human cancers.
numerous mutant peptides that are defined by microsatellite instability
References
1. Lindahl T, Barnes DE (2000). Repair of 5. Lindahl T (2013). My journey to DNA repair. 9. Ceccaldi R, Sarangi P, D’Andrea AD
endogenous DNA damage. Cold Spring Genomics Proteomics Bioinformatics. (2016). The Fanconi anaemia path-
Harb Symp Quant Biol. 65(0):127–33. 11(1):2–7. https://doi.org/10.1016/j.gpb. way: new players and new functions.
https://doi.org/10.1101/sqb.2000.65.127 2012.12.001 PMID:23453014 Nat Rev Mol Cell Biol. 17(6):337–49.
PMID:12760027 ht t p s: //d o i.o r g /10.10 3 8 /n r m. 2 016 . 4 8
6. Dogliotti E, Fortini P, Pascucci B, Parlanti
2. Tubbs A, Nussenzweig A (2017). PMID:27145721
E (2001). The mechanism of switching
Endogenous DNA damage as a source among multiple BER pathways. Prog 10. Ciccia A, Elledge SJ (2010). The DNA dam-
of genomic instability in cancer. Cell.
Nucleic Acid Res Mol Biol. 68:3–27. https:// age response: making it safe to play with
168(4):644–56. https://doi.org/10.1016/j.
doi.org/10.1016/S0079-6603(01)68086-3 knives. Mol Cell. 40(2):179–204. https://
cell.2017.01.002 PMID:28187286
PMID:11554307 doi.org/10.1016/j.molcel. 2010.0 9.019
3. Marteijn JA, Lans H, Vermeulen W, PMID:20965415
7. Kakarougkas A, Jeggo PA (2014). DNA
Hoeijmakers JHJ (2014). Understanding
DSB repair pathway choice: an or- 11. Cleaver JE (1968). Defective repair repli-
nucleotide excision repair and its roles in
cancer and ageing. Nat Rev Mol Cell Biol. chestrated handover mechanism. Br J cation of DNA in xeroderma pigmentosum.
15(7):465–81. https://doi.org/10.1038/ Radiol. 87(1035):20130685. https://doi. Nature. 218(5142):652–6. https://doi.
nrm3822 PMID:24954209 org/10.1259/bjr.20130685 PMID:24363387 org/10.1038/218652a0 PMID:5655953
4. Hanawalt PC, Spivak G (2008). 8. Modrich P (2016). Mechanisms in E. 12. Lynch HT, Snyder CL, Shaw TG, Heinen
Transcription-coupled DNA repair: two de- coli and human mismatch repair (Nobel CD, Hitchins MP (2015). Milestones
cades of progress and surprises. Nat Rev lecture). Angew Chem Int Ed Engl. of Lynch syndrome: 1895-2015. Nat
Mol Cell Biol. 9(12):958–70. https://doi. 55(30):8490–501. https://doi.org/10.1002/ Rev Cancer. 15(3):181–94. https://doi.
org/10.1038/nrm2549 PMID:19023283 anie.201601412 PMID:27198632 org/10.1038/nrc3878 PMID:25673086

13. Weren RD, Ligtenberg MJ, Kets CM, de 22. Cancer Genome Atlas Network (2012). 31. Grollman AP (2013). Aristolochic acid
Voer RM, Verwiel ET, Spruijt L, et al. (2015). Comprehensive molecular characteriza- nephropathy: harbinger of a global iat-
A germline homozygous mutation in the tion of human colon and rectal cancer. rogenic disease. Environ Mol Mutagen.
base-excision repair gene NTHL1 causes Nature. 487(7407):330–7. https://doi. 54(1):1–7. https://doi.org/10.1002/em.21756
adenomatous polyposis and colorectal org/10.1038/nature11252 PMID:22810696 PMID:23238808
cancer. Nat Genet. 47(6):668–71. https://
doi.org/10.1038/ng.3287 PMID:25938944 23. Alexandrov LB, Nik-Zainal S, Wedge DC, 32. Johnson BE, Mazor T, Hong C, Barnes
Aparicio SA, Behjati S, Biankin AV, et al.; M, Aihara K, McLean CY, et al. (2014).
14. Palles C, Cazier JB, Howarth KM, Domingo Australian Pancreatic Cancer Genome Mutational analysis reveals the origin
E, Jones AM, Broderick P, et al.; CORGI Initiative; ICGC Breast Cancer Consortium; and therapy-driven evolution of recur-
Consortium; WGS500 Consortium (2013). ICGC MMML-Seq Consortium; ICGC rent glioma. Science. 343(6167):189–93.
Germline mutations affecting the proof- PedBrain (2013). Signatures of mutation- https://doi.org/10.1126/science.1239947
reading domains of POLE and POLD1 pre- al processes in human cancer. Nature. PMID:24336570
dispose to colorectal adenomas and carci- 500(7463):415–21. https://doi.org/10.1038/
nomas. Nat Genet. 45(2):136–44. https:// nature12477 PMID:23945592 33. Knijnenburg TA, Wang L, Zimmermann
doi.org/10.1038/ng.2503 PMID:23263490 MT, Chambwe N, Gao GF, Cherniack AD,
24. Loeb LA, Springgate CF, Battula N (1974). et al.; Cancer Genome Atlas Research
15. Nielsen FC, van Overeem Hansen T, Errors in DNA replication as a basis of ma- Network (2018). Genomic and molecu-
Sørensen CS (2016). Hereditary breast lignant changes. Cancer Res. 34(9):2311– lar landscape of DNA damage repair
and ovarian cancer: new genes in confined 21. PMID:4136142 deficiency across The Cancer Genome
pathways. Nat Rev Cancer. 16(9):599– Atlas. Cell Rep. 23(1):239–254.e6. https://
612. https://doi.org/10.1038/nrc.2016.72 25. Tomasetti C, Vogelstein B (2015). Cancer doi.org /10.1016/ j.c elrep. 2018.03.076
PMID:27515922 etiology. Variation in cancer risk among PMID:29617664
tissues can be explained by the num-
16. Shiloh Y, Lederman HM (2017). Ataxia- ber of stem cell divisions. Science. 34. Zhang CZ, Leibowitz ML, Pellman D
telangiectasia (A-T): an emerging dimen- 347(6217):78–81. https://doi.org/10.1126/ (2013). Chromothripsis and beyond: rapid
sion of premature ageing. Ageing Res science.1260825 PMID:25554788 genome evolution from complex chro-
Rev. 33:76–88. https://doi.org/10.1016/j. mosomal rearrangements. Genes Dev.
arr.2016.05.002 PMID:27181190 26. Olivier M, Hollstein M, Hainaut P (2010). 27(23):2513–30. https://doi.org/10.1101/
TP53 mutations in human cancers: ori- gad.229559.113 PMID:24298051
17. Stracker TH, Roig I, Knobel PA, Marjanović gins, consequences, and clinical use. Cold
M (2013). The ATM signaling network in de- Spring Harb Perspect Biol. 2(1):a001008. 35. Bryant HE, Schultz N, Thomas HD, Parker
velopment and disease. Front Genet. 4:37. ht t p s: //d o i.o r g /10.1101/c s h p e r s p e c t . KM, Flower D, Lopez E, et al. (2005).
https://doi.org/10.3389/fgene.2013.00037 a001008 PMID:20182602 Specific killing of BRCA2-deficient tu-
PMID:23532176 mours with inhibitors of poly(ADP-ribose)
27. Helleday T, Eshtad S, Nik-Zainal S (2014). polymerase. Nature. 434(7035):913–7.
18. Gennery AR, Cant AJ, Jeggo PA (2000). Mechanisms underlying mutational signa- ht t p s: //d o i.o r g /10.10 3 8 /nat ur e 0 3 4 4 3
Immunodeficiency associated with DNA tures in human cancers. Nat Rev Genet. PMID:15829966
repair defects. Clin Exp Immunol. 121(1): 15(9):585–98. https://doi.org/10.1038/
1–7. https://doi.org/10.1046/j.1365-2249. nrg3729 PMID:24981601 36. Brown JS, O’Carrigan B, Jackson SP, Yap
2000.01257.x PMID:10886231 TA (2017). Targeting DNA repair in cancer:
28. Hollstein M, Alexandrov LB, Wild CP, Ardin beyond PARP inhibitors. Cancer Discov.
19. McKinnon PJ (2013). Maintaining ge- M, Zavadil J (2017). Base changes in tu- 7(1):20–37. https://doi.org/10.1158/2159-
nome stability in the nervous system. mour DNA have the power to reveal the 8290.CD-16-0860 PMID:28003236
Nat Neurosci. 16(11):1523–9. https://doi. causes and evolution of cancer. Oncogene.
org/10.1038/nn.3537 PMID:24165679 36(2):158–67. https://doi.org/10.1038/onc. 37. Le DT, Uram JN, Wang H, Bartlett BR,
2016.192 PMID:27270430 Kemberling H, Eyring AD, et al. (2015).
20. McCann J, Choi E, Yamasaki E, Ames PD-1 blockade in tumors with mis-
BN (1975). Detection of carcinogens 29. Alexandrov LB, Ju YS, Haase K, Van Loo P, match-repair deficiency. N Engl J Med.
as mutagens in the Salmonella/micro- Martincorena I, Nik-Zainal S, et al. (2016). 372(26):2509–20. https://doi.org/10.1056/
some test: assay of 300 chemicals. Proc Mutational signatures associated with to- NEJMoa1500596 PMID:26028255
Natl Acad Sci U S A. 72(12):5135–9. bacco smoking in human cancer. Science.
https://doi.org/10.1073/pnas.72.12.5135 354(6312):618–22. https://doi.org/10.1126/ 38. Le DT, Durham JN, Smith KN, Wang H,
PMID:1061098 science.aag0299 PMID:27811275 Bartlett BR, Aulakh LK, et al. (2017).
Mismatch repair deficiency predicts re-
21. Delaney JC, Essigmann JM (2008). 30. Huang MN, Yu W, Teoh WW, Ardin M, sponse of solid tumors to PD-1 block-
Biological properties of single chemical- Jusakul A, Ng AWT, et al. (2017). Genome- ade. Science. 357(6349):409–13. https://
DNA adducts: a twenty year perspec- scale mutational signatures of aflatoxin in d o i . o r g / 1 0 .11 2 6 / s c i e n c e . a a n 6 7 3 3
tive. Chem Res Toxicol. 21(1):232–52. cells, mice, and human tumors. Genome PMID:28596308
h t t p s : / / d o i . o r g / 1 0 .1 0 2 1 / t x 7 0 0 2 9 2 a Res. 27(9):1475–86. https://doi.org/10.1101/
PMID:18072751 gr.220038.116 PMID:28739859
180
3.5
Inflammation
Playing a pivotal role in cancer
pathogenesis
Muthu K. Shanmugam Massimo Tommasino (reviewer)
Alan Prem Kumar Ioannis P. Trougakos (reviewer)
Gautam Sethi
cases are estimated to be linked to scription factors plays a critical role

SUMMARY inflammation [5]. Inflammation by in the sustained cell proliferation
itself may not lead to cancer; ad- observed in cancers [5]. The major-
●● Factors linking chronic inflam- ditional mutations and epigenetic ity of cancers are a consequence of
mation and cancer are of great events that occur in the genome of chronic inflammation, infection, dys-
CHAPTER 3.5
SECTION 3
interest, and increasing evi- cells as a result of environmental ex- functional cell death mechanisms,
dence suggests that constitu- posures or changes in immunity are and dysregulation of cell-cycle
tive activation of pro-inflamma-
also important contributors to onco- molecules. Chronic inflammation
tory transcription factors can
genesis [6]. is associated with the production
mediate carcinogenesis.
Through the immune response to of pro-inflammatory cytokines and
●● An inflammatory condition often acute inflammation, activated cells, chemokines, which constitutively ac-
precedes the development of including macrophages, monocytes, tivate pro-survival transcription fac-
cancer, and pro-inflammatory lymphocytes, neutrophils, and leu- tors that may act as key regulators
transcription factors such as NF- kocytes, are attracted to the injured of carcinogenesis [6].
κB and STAT3 are constitutively site and reduce the inflammation There are some exceptions; for
active in various cancer types. (see Chapter 3.9). However, in cases example, chronic inflammation of the
●● Chemotherapeutic agents and of severe inflammation, these cells joint or muscle may not lead to the
gamma irradiation can activate contribute to excessive production of development of cancer. Nonetheless,
NF-κB and/or STAT3, which pro-inflammatory molecules, such as tumour-associated persistent infec-
can lead to chemoresistance the cytokine tumour necrosis factor tion and inflammation are associ-
and radioresistance. α (TNF-α), interleukin-1β (IL-1β) and ated with 15–20% of cancer deaths
IL-6, the chemokine receptor CXCR4 worldwide (see Chapter 2.2), and
●● Suppression of NF-κB and and its ligand CXCL12, cyclooxy- obesity-associated inflammation is
STAT3 may inhibit the prolif- genase 2 (COX-2), prostaglandins, likely to contribute further to cancer-
eration and invasion of cancer nitric oxide, and leukotrienes, which related deaths (see Chapter 2.7) [8].
cells, and most chemopreven- dysregulate signal transduction path- Tumour-caused inflammation, such
tive agents mediate their effects ways, thereby contributing to the de- as necrotic death of cancer cells,
through inhibition of the NF-κB velopment of cancer [6]. insufficient blood supply, and viral
and STAT3 activation pathways.
Inflammation is a tightly regu- infections in the tumour bed, con-
●● Modulation of these pro-inflam- lated process that can be very ef- tributes to malignant progression of
matory pathways may provide fectively turned on or off under organ-specific cancers such as liver
opportunities for both prevention normal physiological conditions [7]. cancer (see Chapter 5.6) and colon
and treatment of chronic diseas- Acute inflammation is mainly a self- cancer (see Chapter 5.5) [9]. In addi-
es, including cancer. limiting process and can be treated tion, in patients who are undergoing
therapeutically; however, prolonged chemotherapy or radiotherapy, in-
chronic inflammation is mostly det- duced tumour necrosis is often asso-
Virchow (in the 19th century) and rimental [6]. Factors linking chronic ciated with an increase in tumour-as-
others (in the early 20th century) inflammation and cancer are of sociated inflammation, leading to the
proposed an association between great interest, and several lines of development of resistance to therapy
inflammation and cancer [1–4]. evidence suggest that constitutive and/or the induction of anti-tumour
Worldwide, about 15% of all cancer activation of pro-inflammatory tran- immunity. Therefore, inflammation is
Chapter 3.5 • Inflammation 181

an important factor driving tumour ogy domain of about 300 amino ac-
growth in most solid and haemato- ids that plays a critical role in their FUNDAMENTALS
poietic malignancies [10]. functions, such as dimerization and
The molecular mechanisms that DNA binding via the N-terminal part ■■ Historically, inflammation was
connect chronic inflammation to of the Rel homology domain, and described in Latin by four
cancer development have become a heterodimerization interaction with major signs: rubor (redness),
major area of research. This chapter inhibitory κBs (IκBs) involving the tumor (swelling), calor (heat),
focuses on the role of the transcrip- C-terminal part of the Rel homology
and dolor (pain).
tion factor nuclear factor kappa-light- domain, both of which are intracel-
chain-enhancer of activated B cells lular inhibitors of NF-κB [12]. NF-κB ■■ In 1858, the German physician
(NF-κB). Other notable transcription family members can also form di- Virchow postulated that
factors that are implicated in inflam- verse homodimers or heterodimers, micro-inflammation that results
mation and tumorigenesis are also and the subunits RelA, c-Rel, and from irritation may lead to the
discussed, i.e. the signal transducer RelB contain a C-terminal transcrip-
development of cancer.
and activator of transcription (STAT) tional activation domain (absent in
family as well as the mitogen-acti- p50 and p52), which enables them ■■ Inflammation is typically
vated protein kinase (MAPK) family. to dimerize and physically bind designated by adding the
Finally, opportunities for the preven- via promoter/enhancer molecules suffix “–itis”. Such conditions,
tion and treatment of inflammation- to specific DNA sequences in κB for example colitis and
driven cancers are described. sites: 5′-GGGRNYYYCC-3′, where
pancreatitis, often predispose
R is a purine, Y is a pyrimidine, and
to cancer.
N is any nucleotide [15].
NF-κB signalling in In resting cells, most NF-κB ■■ Alcohol consumption, smoking,
inflammation and cancer subunit complexes are primarily cy- chronic infections, obesity,
The first evidence for the link be- toplasmic and exist as homodimers exposure to environmental
tween chronic inflammation and or heterodimers bound to IκBs and
pollutants, radiation exposure,
cancer involved a proposed relation- present in an inactive form. This is
ship between NF-κB and cancer de- a high energy intake, and other
because their binding to IκB pro-
velopment. This hypothesis gained teins prevents DNA binding and, as factors have been recognized
prominence from the similarities in a consequence, prevents nuclear ac- as risk factors for most chronic
structure between the v-Rel protein cumulation [6]. The IκB family of pro- diseases, including cancer.
and the NF-κB c-Rel protein [11]. teins is composed of the typical IκBs All of these risk factors may
Cancer development in the presence (IκBα, IκBβ, and IκBε), the atypical be linked to cancer through
of chronic inflammation involves the IκBs (Bcl-3 and IκBζ), and the pre- the process of chronic
constant presence of activated on- cursor IκBs (p100 and p105). They inflammation.
cogenes and major transcription fac- have been characterized, and con-
tors, such as NF-κB and STAT3. tain in their C terminus up to seven ■■ Inflammation may be caused
The NF-κB family, which was 33-amino acid consensus ankyrin re- by a range of diseases due to
discovered in 1986 by Baltimore peats, which regulate protein–protein infectious organisms that are
and Sen [12], plays a pivotal role in interaction and bind to Rel proteins, recognized to cause cancer.
wide-ranging processes, including thereby masking their nuclear locali- These include hepatitis
immunity, inflammation, apoptosis, zation signal. The IκB kinase (IKK) from hepatitis B virus and
learning, and memory [13]. These complex is composed of two catalytic hepatitis C virus, gastritis
proteins have a key role in innate and kinases (IKKα and IKKβ) and one
from Helicobacter pylori,
adaptive immune functions that can non-catalytic subunit, called IKKγ or
and cervicitis from human
regulate proliferation and survival NF-κB essential modulator (NEMO).
and stimulate angiogenesis, inva- Upon activation, IKK can phosphor- papillomavirus.
sion, and migration, thereby leading ylate IκB and abrogate the suppres- ■■ The major defence response
to metastasis [14]. sive effect of IκBs on NF-κB dimers initiated by the human body
[6]. This effectively releases NF-κB upon injury or infection is
Structural components and for subsequent phosphorylation and
organization of the NF-κB the activation of the immune
acetylation, and promotes nuclear
pathway system via active recruitment
translocation (Fig. 3.5.1).
The mammalian NF-κB family of of diverse cells such as
transcription factors is composed NF-κB signalling pathways macrophages, monocytes,
of RelA (p65), c-Rel, RelB, NF-κB1 Activation of NF-κB is fairly rapid, lymphocytes, neutrophils,
(p50), and NF-κB2 (p52). They all and it can be activated by expo- and leukocytes.
contain a conserved Rel homol- sure to diverse stimuli. There are
182
The STAT signalling pathway
Several published studies have cytokines and chemokines, and STAT3 dimerization, increase DNA
indicated the pivotal role of the the oncogenic signalling pathway. binding and transcriptional activa-
signal transducer and activator of Binding of a ligand, for example IL- tion, and mediate cancer progres-
transcription (STAT) family as pro- 6, to its specific receptor subunit sion [3].
inflammatory transcription factors can induce dimerization of glyco-
that are found to be constitutively protein 130 and activation of non- References
activated in several cancer types. receptor tyrosine kinases called 1. Wegenka UM, Buschmann J, Lütticken
STAT3 was first discovered as an Janus kinases (JAKs). This, in C, Heinrich PC, Horn F (1993). Acute-
phase response factor, a nuclear fac-
acute-phase response protein, turn, can phosphorylate STAT3 at tor binding to acute-phase response
thereby indicating its causal link to tyrosine 705, and activated STAT elements, is rapidly activated by in-
inflammation [1]. The STAT family dimers can translocate to the nu- terleukin-6 at the posttranslational
level. Mol Cell Biol. 13(1):276–88.
of transcription factors was discov- cleus, bind to specific elements, https://doi.org/10.1128/MCB.13.1.276
ered in 1994 during the evaluation and regulate gene transcription. PMID:7678052
of the molecular pathways involved In addition, it has been report-
2. Darnell JE Jr, Kerr IM, Stark GR (1994).
in interferon-triggered gene regu- ed that STAT3 may directly inter- Jak-STAT pathways and transcriptional
lation [2]. A total of seven STAT act with the NF-κB family member activation in response to IFNs and
proteins (STAT1, STAT2, STAT3, RelA, thereby increasing the pro- other extracellular signaling proteins.
Science. 264(5164):1415–21. https://
STAT4, STAT5a, STAT5b, and duction of pro-inflammatory mole- d o i . o r g / 1 0 .112 6 / s c i e n c e . 819 74 5 5
CHAPTER 3.5
SECTION 3
STAT6) have been identified to cules such as IL-6, TNF, and growth PMID:8197455
date in mammalian cells [3]. factors, which in turn act in and can 3. Zhong Z, Wen Z, Darnell JE Jr (1994).
Among the STAT family of pro- sustain a chronic inflammatory mi- Stat3: a STAT family member activated
teins, STAT3 is the most active. croenvironment in tumours. STAT3 by tyrosine phosphorylation in re-
can also be acetylated at lysine sponse to epidermal growth factor and
STAT3 plays a critical role in the
interleukin-6. Science. 264(5155):95–8.
regulation of intracellular signalling, K685 by lysine acetyltransferase https://doi.org/10.1126/science.8140422
the synthesis of pro-inflammatory p300/CBP, which may upregulate PMID:8140422
Fig. 3.5.1. The structure of the mammalian NF-κB family of transcription factors. ANK, ankyrin repeats; ATD, amino terminal
domain; DD, death domain; DMD, dimerization domain; G, glycine-rich region; L, 10-amino acid polypeptide linker to nuclear
localization signal; LZ, leucine zipper; TAD, transcriptional activation domain.
Key serine (S)

Rel homology domain phosphorylation sites
1 551
RelA COOH S276, 311, 529, 536
NH2 ATD DMD L TAD
(NF- Bp65)
1 619
c-Rel NH2 ATD DMD L TAD COOH S267, 454, 460, 471
1 579
RelB NH2 LZ ATD DMD L TAD COOH S368, 552
1 447/899
p52/p100 S99, 108, 115, 123
NH2 ATD DMD L G ANK DD COOH
(NF- B2)
1 433/969
p50/p105 S337, 903, 907, 927, 932
NH2 ATD DMD L G ANK DD COOH
(NF- B1)
two
ATDmajor types
– amino of NF-κB
terminal signalling be activated by upstream trans-
domain known as the non-canonical path-
DMD – dimerization domain path- forming growth factor β-activated
pathways [6]. The classical way) leads to the formation of the
way (also
L – 10 known polypeptide
aminoacid as the canonical kinase localization
linker to nuclear (TAK) upon induction by
signal RelB–NF-κB2 (p52/p100) complex.
TAD - transcription
pathway) leads to activation domain pro-inflammatory cytokines such as
the generation Activation of this pathway is medi-
LZ the
of - leucine-Zipper
active RelA–NF-κB1 (p50/ TNF-α, IL-1β, and lipopolysaccha- ated through the catalytic activity of
G – glycine
p105) rich region
complex. This pathway can ride. The alternative pathway (also NF-κB-inducing kinase (NIK), and
ANK – ankyrin repeats
DD - death domain
can be initiated by lymphotoxin, miR-301a) that target messenger thereby modulate inflammation, tu-
receptor activator of NF-κB ligand RNAs regulating NF-κB subunits, mour progression, and metastasis
(RANKL), CD40 ligand, and B cell- IκBs, and IKKs; in turn, NF-κB can [17] (Fig. 3.5.3).
activating factor of the TNF family regulate microRNA expression [6].
(BAFF) [6,16] (Fig. 3.5.2). Therefore, NF-κB may have a key Epigenetic modifications in
Upon activation of the classi- role in the inflammatory responses NF-κB
cal pathway, NF-κB can transcribe in normal cells coordinating both Chronic inflammation, which is
various genes encoding the pro-in- acute inflammation and chronic in- often driven by inflammatory re-
flammatory enzyme COX-2, induc- flammation, and any dysregulation sponse mediated through NF-κB
ible nitric oxide synthase, cytokines of this signalling pathway can lead activation, is associated with epi-
such as TNF-α, IL-1, and IL-6, to diverse malignancies. genetic modifications such as ly-
chemokines, growth factors, matrix sine acetylation and methylation
metalloproteinases, cell-cycle pro- Role of NF-κB in the tumour and arginine methylation [18]. The
teins, anti-apoptotic proteins such microenvironment major modification is lysine acety-
as Bcl-2, Bcl-xL, and FLIP, vascu- Tumorigenesis is often associ- lation, which has been reported to
lar endothelial growth factor, adhe- ated with the presence of tumour- be an important regulator of expres-
sion molecules such as ICAM-1 and associated macrophages, mast sion of pro-inflammatory genes.
VCAM-1, and inhibitors of NF-κB cells, neutrophils, dendritic cells, Acetylation of distinct lysine resi-
signalling, including IκBs and A20. myeloid-derived suppressor cells, dues of RelA at K218, K221, and
Recent studies have also in- T cells, B cells, natural killer cells, K310 by lysine acetyltransferase
dicated that NF-κB can be posi- natural killer T cells, endothelial p300/CBP can regulate NF-κB tran-
tively or negatively regulated by cells, and cancer-associated fibro- scriptional activation, DNA binding
microRNAs (such as miR-21, blasts. NF-κB signalling can regu- affinity, IκBα assembly, and sub-
miR-146, miR-155, miR-181b, and late recruitment of these cells and cellular localization [19]. However,
Fig. 3.5.2. The canonical (or classical) and non-canonical (or alternative) NF-κB signalling pathways. BAFF, B cell-activating
factor of the TNF family; CD40L, CD40 ligand; IκB, inhibitor of NF-κB; IKK, inhibitory κB (IκB) kinase; IL-1β, interleukin-1β; LPS,
lipopolysaccharide; LT-β, lymphotoxin β; NEMO, NF-κB essential modulator; NIK, NF-κB-inducing kinase; RANKL, receptor
activator of NF-κB ligand; TAK, transforming growth factor β-activated kinase; TNF-α, tumour necrosis factor α.
Canonical Pathway Non-Canonical Pathway

(classical pathway) (alternative pathway)
TNF, IL1, LPS BAFF, CD40L, LT, RANKL
P
IKK IKK NIK
P
IKK /IKK
P
IKK /IKK
P
I B- P I B-
degradation P
RelA, (NF B1) p100 p100 P
(p50/p105) degradation
CYTOPLASM RelB (NF B2)
(p52/p100)
RelA p50 RelB p52
NUCLEUS
P P
Gene transcription Gene transcription
184
Fig. 3.5.3. The role of NF-κB in the tumour microenvironment. Different types of cells in the tumour microenvironment, including
tumour-associated macrophages, mast cells, neutrophils, dendritic cells, myeloid-derived suppressor cells, T cells, B cells, natural killer
cells, natural killer T cells, endothelial cells, and cancer-associated fibroblasts, can augment NF-κB activation, modulate inflammation,
and lead to sustained tumorigenesis and metastasis. IFN-γ, interferon γ; IL-1, interleukin-1; TNF-α, tumour necrosis factor α.
Cytokines
Tumour cells
TNF-
IFN-
IL-1
Extrinsic
(tobacco smoke, Chemokines
viral infections)
or intrinsic Activated transcription factors
(genetic mutations
and epigenetic
Matrix metalloproteases
alterations) Angiogenic factors
factors of
inflammation
NF-κB activation
CHAPTER 3.5
Cells of the tumour microenvironment
SECTION 3
Metastatic tumour cells
acetylation of RelA at K122 and to prevent the development of can- Infection with the Gram-negative
K123 by p300/CBP was found to re- cer can work, because of the long bacterium Helicobacter pylori is a
duce DNA binding and increase IκB time frame required for the cancer major risk factor for gastritis, gastric
binding to RelA, thereby indicating to progress from a benign state to a ulcers, and stomach cancer (see
negative regulation of inflammation. malignant phenotype. Chapter 5.4). A significant decline
Another NF-κB family member, in the incidence of stomach cancer
p50 (NF-κB1), can be acetylated at Preventable risk factors has been observed as a result of im-
K431, K440, and K441, which may for cancer initiation and proved sanitation, refrigeration, and
also upregulate transcriptional ac- progression food preservation as well as the use
tivation, thereby indicating positive Primary prevention is aimed at pre- of antibiotics to effectively eradicate
regulation of inflammation [18,19]. venting the development of cancer H. pylori infection [20].
Acetylation of histone H3 is often in the first place by reducing the Lifestyle factors such as obe-
found in cytokine-mediated inflam- exposures of individuals to risk fac- sity, unhealthy diet, and physical
mation and NF-κB activation, and tors, through strategies such as inactivity have also been identified
thus histone-modifying enzymes smoking cessation; abstaining from as potential risk factors for cancer
can have critical functions in tu- chronic alcohol consumption; vac- (see Chapter 2.6). All of these risk
mour progression. cination against oncogenic viruses; factors are linked to cancer through
reducing or eliminating environ- the process of chronic inflammation.
mental, occupational, or behaviour- In addition, consumption of fruits,
Opportunities for al exposures to carcinogens; the legumes, and green leafy vegeta-
prevention and treatment use of novel screening methods; bles has been found to considerably
Early detection or screening for and the possibility of delaying age- reduce the risk of cancer develop-
pre-symptomatic cancers or cancer ing, thereby preventing or delaying ment, potentially through an anti-
precursors as a potential strategy the development of cancer. oxidant activity. Skin cancer can be

The MAPK signalling pathway
Mitogen-activated protein kinases In the MAPK signalling pathway, phosphorylated and active in the in-
(MAPKs) are a family of serine/ MAPK kinase kinase (MAPKKK) flamed intestinal mucosa of patients
threonine-specific protein kinas- phosphorylates and activates MAPK with inflammatory bowel disease [2].
es. MAPKs regulate cellular pro- kinase (MAPKK), which in turn can
phosphorylate and activate vari- References
cesses such as cell proliferation,
differentiation, cell survival, and ous MAPKs during the inflamma- 1. Dhillon AS, Hagan S, Rath O, Kolch W
(2007). MAP kinase signalling pathways
apoptosis in response to a variety tory response. Dysregulated p38 in cancer. Oncogene. 26(22):3279–90.
MAPK signalling is highly active in https://doi.org/10.1038/sj.onc.1210421
of external stimuli, including mito-
different cancer types, favouring PMID:17496922
gens, heat shock, osmotic stress,
tumour growth. p38 MAPKs are 2. Docena G, Rovedatti L, Kruidenier L,
and inflammatory cytokines, and central to inflammatory processes Fanning A, Leakey NA, Knowles CH, et al.
MAPKs are often found to be and to the production of pro-inflam- (2010). Down-regulation of p38 mitogen-
activated protein kinase activation and
dysregulated in cancer cells. The matory molecules that contribute to proinflammatory cytokine production by
mammalian MAPKs comprise ex- colitis-associated colorectal cancer mitogen-activated protein kinase inhibi-
tracellular signal-regulated kinase pathogenesis. p38α can also me- tors in inflammatory bowel disease. Clin
Exp Immunol. 162(1):108–15. https://doi.
1/2 (ERK1/2), c-Jun N-terminal ki- diate inflammation in inflammatory org/10.1111/j.1365-2249.2010.04203.x
nases (JNKs), and p38 MAPK [1]. bowel disease and is substantially PMID:20731675
prevented by reducing exposure to processed meat has been associ- Avoiding chronic alcohol con-
ultraviolet radiation from sunlight or ated with a decreased risk of colo- sumption has been found to lower the
artificial sources (see Chapter 2.4). rectal cancer [22]. risk of liver cancer by reducing inflam-
The role of inflammation as a Microbial pathogens can also mation and cirrhosis of the liver (see
crucial mediator of colorectal cancer drive tumorigenesis in 15–20% of Chapter 2.3). The success of can-
is also well established, and the use cancer cases. The gut microbiota cer prevention strategies will require
of non-steroidal anti-inflammatory has been shown to alter cancer comprehensive planning and the in-
drugs such as aspirin and ibupro- susceptibility and progression by corporation of diverse approaches,
fen has been found to significantly modulating inflammation and by including public policy, education,
reduce the risk of colorectal cancer producing metabolites that may and research, to identify acceptable
in some patient populations (see be involved in either oncogen- and effective ways to modify people’s
Chapter 6.4) [21]. In addition, re- esis or tumour suppression (see behaviour over long periods of time.
duced consumption of red meat and Chapter 3.10). For example, in the Ageing is also closely asso-
colon Clostridium scindens bac- ciated with the development of
teria can produce toxic secondary chronic inflammation, which forms
Fig. 3.5.4. Potential cancer risk factors in- bile acids in response to dietary the basis for the development of
clude obesity, unhealthy diet, and physi- fat. Furthermore, diets high in fats various age-related disorders (see
cal inactivity. Such factors may mediate induce blooms of Bilophila wads Chapter 3.1). Epidemiological data
cancer risk by provoking inflammatory worthia, a sulfite-reducing bacterium clearly indicate that elevated levels
change in relevant tissues.
that has been found to be associat- of IL-6 and C-reactive protein in the
ed with increased risk of inflamma- blood may lead to multiple cellular
tory bowel disease and malignan- changes. Compared with younger
cies. However, there are examples people, those aged 64–102 years
of whole foods and dietary compo- were found to have higher levels of
nents, such as soy-based products, inflammatory biomarkers, including
cruciferous vegetables containing IL-6, TNF-α, IL-8, and C-reactive
sulforaphane and isothiocyanates, protein [24], which may contribute
and berries containing ellagic acid, to tumour development by forming
that can inhibit COX-2 production a pro-tumorigenic inflammatory en-
and subsequent development of vironment and by recruiting various
cancer. The diet may also dictate immune cells that can promote tu-
whether the gut microbiota can pro- mour progression by both autocrine
duce active metabolites that may and paracrine mechanisms.
aggravate or ameliorate tumour de- Chronic inflammation is a low-
velopment and progression [23]. grade sustained process driven by
186
Fig. 3.5.5. Histological section from a cirrhotic liver. Avoiding chronic alcohol consump tory cascade may aid in the develop-
tion has been found to lower the risk of liver cancer by reducing inflammation and ment of novel anti-cancer treatment
cirrhosis of the liver. strategies [25].
Compounds from natural

products as inhibitors of
NF-κB- and STAT3-mediated
inflammation-driven cancers
Targeting NF-κB and STAT3 has
become an attractive strategy, and
various pharmacological inhibitors
can modulate NF-κB and STAT3
activation in tumour models. Some
important natural compounds have
been shown to inhibit inflammatory
mediators involved in cancer pro-
gression; examples are curcumin,
ursolic acid, oleanolic acid, gar-
cinol, zerumbone, resveratrol, thy-
moquinone, diosgenin, celastrol,
butein, sulforaphane, and epigallo-
CHAPTER 3.5
SECTION 3
catechin gallate [26].
The link between inflammation
and cancer is well established, and
strategies to prevent chronic cancer
continuous activation of various tran- such as diabetes, obesity, and can- inflammation include (i) reducing
scription factors, such as NF-κB and cer. Long-term administration of non- the recruitment of inflammatory re-
STAT3, leading to oncogenesis. The steroidal anti-inflammatory drugs has sponse elements to the tumour site
bacterial population in the gut micro- been shown to reduce the risk of de- and (ii) blocking pro-tumorigenic in-
biota has been found to have an im- velopment of various inflammation- flammatory elements or redirecting
portant function in the development driven ailments. Therefore, a better inflammation with properties that
of inflammatory bowel disease and understanding of the diverse molecu- are anti-tumour, immunostimula-
in increased risk of chronic diseases lar players involved in the inflammatory, or both.

References
1. Virchow R (1858). Reizung und Reiz 11. Gilmore TD, Starczynowski DT, Kalaitzidis 20. Crew KD, Neugut AI (2006). Epidemiology
barkeit. Arch Pathol Anat Physiol Klin D (2004). RELevant gene amplification in of gastric cancer. World J Gastroenterol.
Med. 14(1–2):1–63. https://doi.org/10.1007/ B-cell lymphomas? Blood. 103(8):3243–4, 12(3):354–62. https://doi.org/10.3748/
BF01877355 author reply 3244–5. https://doi.org/10.1182/ wjg.v12.i3.354 PMID:16489633
blood-2003-11-4019 PMID:15070712
2. Yamagiwa K, Ichikawa K (1916). Ex 21. Todoric J, Antonucci L, Karin M (2016).
perimental study on the pathogenesis 12. Sen R, Baltimore D (2006). Multiple nu- Targeting inflammation in cancer preven-
of epithelial tumors (the first report) [in clear factors interact with the immuno- tion and therapy. Cancer Prev Res (Phila).
Japanese]. Tokyo Igakukai Zasshi. 30:1–43. globulin enhancer sequences. Cell 1986. 9(12):895–905. https://doi.org/10.1158/1940-
46: 705–716. J Immunol. 177(11):7485– 6207.CAPR-16-0209 PMID:27913448
3. Yamagiwa K, Ichikawa K (1918). Ex 96. PMID:17114415
perimental study of the pathogenesis of 22. O’Keefe SJ (2016). Diet, microorgan-
carc inoma. J Cancer Res. 3(1):1–29. 13. Perkins ND (2007). Integrating cell-signal- isms and their metabolites, and colon
ling pathways with NF-κB and IKK function. cancer. Nat Rev Gastroenterol Hepatol.
4. Fujiki H (2014). Gist of Dr. Katsusaburo Nat Rev Mol Cell Biol. 8(1):49–62. https:// 13(12):691–706. https://doi.org/10.1038/
Yamagiwa’s papers entitled “Experimen doi.org/10.1038/nrm2083 PMID:17183360 nrgastro.2016.165 PMID:27848961
tal study on the pathogenesis of epithe-
lial tumors” (I to VI reports). Cancer Sci. 14. Puar YR, Shanmugam MK, Fan L, 23. Bhatt AP, Redinbo MR, Bultman SJ
105(2):143–9. https://doi.org/10.1111/cas. Arfuso F, Sethi G, Tergaonkar V (2018). (2017). The role of the microbiome in
12333 PMID:24313817 Evidence for the involvement of the mas- cancer development and therapy. CA
ter transcription factor NF-κB in cancer Cancer J Clin. 67(4):326–44. https://doi.
5. Colotta F, Allavena P, Sica A, Garlanda C, initiation and progression. Biomedicines. org/10.3322/caac.21398 PMID:28481406
Mantovani A (2009). Cancer-related inflam- 6(3):E82. https://doi.org/10.3390/biomed
mation, the seventh hallmark of cancer: icines6030082 PMID:30060453 24. Claesson MJ, Jeffery IB, Conde S, Power
links to genetic instability. Carcinogenesis. SE, O’Connor EM, Cusack S, et al.
30(7):1073–81. https://doi.org/10.1093/ 15. Oeckinghaus A, Ghosh S (2009). The (2012). Gut microbiota composition cor-
carcin/bgp127 PMID:19468060 NF-κB family of transcription factors and relates with diet and health in the elderly.
its regulation. Cold Spring Harb Perspect Nature. 488(7410):178–84. https://doi.
6. Taniguchi K, Karin M (2018). NF-κB, in- Biol. 1(4):a000034. https://doi.org/10.1101/ org/10.1038/nature11319 PMID:22797518
flammation, immunity and cancer: coming cshperspect.a000034 PMID:20066092
of age. Nat Rev Immunol. 18(5):309–24. 25. Zinger A, Cho WC, Ben-Yehuda A
h t t p s: //d o i .o r g /10 .10 3 8 /n r i . 2 017.142 16. Sethi G, Tergaonkar V (2009). Potential (2017). Cancer and aging – the inflamma-
PMID:29379212 pharmacological control of the NF-κB path- tory connection. Aging Dis. 8(5):611–27.
way. Trends Pharmacol Sci. 30(6):313–21. https://doi.org/10.14336/AD.2016.1230
7. Medzhitov R (2010). Inflammation 2010: https://doi.org/10.1016/j.tips.2009.03.004 PMID:28966805
new adventures of an old flame. Cell. PMID:19446347
140(6):771–6. https://doi.org/10.1016/j. 26. Shanmugam MK, Lee JH, Chai EZ, Kanchi
cell.2010.03.006 PMID:20303867 17. Sethi G, Shanmugam MK, Ramachandran MM, Kar S, Arfuso F, et al. (2016). Cancer
L, Kumar AP, Tergaonkar V (2012). prevention and therapy through the modu-
8. Plummer M, de Martel C, Vignat J, Ferlay Multifaceted link between cancer and lation of transcription factors by bioactive
J, Bray F, Franceschi S (2016). Global inflammation. Biosci Rep. 32(1):1–15. natural compounds. Semin Cancer Biol.
burden of cancers attributable to infec- https://doi.org/10.1042 /BSR 2010 0136 40–41:35–47. https://doi.org/10.1016/j.
tions in 2012: a synthetic analysis. Lancet PMID:21981137 semcancer.2016.03.005 PMID:27038646
Glob Health. 4(9):e609–16. https://doi.
org /10.1016/S2 214 -10 9X(16) 3 014 3 -7 18. Kaypee S, Sudarshan D, Shanmugam
PMID:27470177 MK, Mukherjee D, Sethi G, Kundu TK
(2016). Aberrant lysine acetylation in
9. Grivennikov SI, Greten FR, Karin M (2010). tumorigenesis: implications in the de-
Immunity, inflammation, and cancer. Cell. velopment of therapeutics. Pharmacol
140(6):883–99. https://doi.org/10.1016/j. Ther. 162:98–119. https://doi.org/10.1016/j.
cell.2010.01.025 PMID:20303878 pharmthera.2016.01.011 PMID:26808162
10. Shalapour S, Karin M (2015). Immunity, 19. Shanmugam MK, Sethi G (2012). Role of
inflammation, and cancer: an eternal epigenetics in inflammation-associated dis-
fight between good and evil. J Clin Invest. eases. In: Kundu TK, editor. Epigenetics:
125(9):3347–55. https://doi.org/10.1172/ development and disease. London, UK:
JCI80007 PMID:26325032 Springer Dordrecht; pp. 627–58.
188
3.6
Reproductive and hormonal factors

Important contributors to several
cancer sites
Louise A. Brinton Jennifer D. Brooks (reviewer)
Silvia Franceschi (reviewer)
Esther Roura Fornells (reviewer)
that estrogens are important in if it is a full-term pregnancy; there is

SUMMARY the etiology of female breast little evidence for relationships with
cancer, endometrial cancer, and short-term pregnancies, including
●● Reproductive and hormonal fac- male breast cancer, and possi- miscarriages and abortions.
tors appear to have particular as- bly advanced prostate cancer. The reduced risk associated with
CHAPTER 3.6
SECTION 3
sociations for different subtypes parity may be further enhanced if a
of cancers in women, including woman breastfeeds. However, the
those defined by either histology It is now well recognized that repro- protection appears to be dependent
or hormone receptor status. ductive and hormonal factors play on longer periods of breastfeeding;
●● Use of oral contraceptives is a major role in the etiology of many therefore, in most high-income coun-
related to substantial reduc- cancer types in women. This is par- tries, in which numbers of births are
ticularly true for breast cancer, endo- limited and each child is breastfed for
tions in the risk of endometrial
metrial cancer, and ovarian cancer, a relatively short period, there is little
cancer and ovarian cancer, and
in which such factors are likely to ex- evidence of a relationship of risk with
the reduction in risk persists for
plain large proportions of disease oc- breastfeeding. The most conclusive
extended durations after dis-
currence. A few cancer types in men findings on the protective effects of
continuation of use. Use of oral
may also be influenced by hormonal breastfeeding derive from studies of
contraceptives appears to be
factors, although the relationships women who have given birth to multi-
related to an increased risk of
are less well defined. ple children and have breastfed them
cervical cancer, consistent with
growing evidence for a possible for long periods (e.g. 2 years or more
role of hormonal factors in cer- Female breast cancer per child), leading to long durations of
vical carcinogenesis. The role of parity in the etiology of cumulative breastfeeding.
breast cancer is well established. In contrast to the other estab-
●● Obese women are at increased lished reproductive risk factors,
Parous women have approximately
risk of postmenopausal breast half the risk of nulliparous women, use of oral contraceptives is not
cancer and endometrial cancer, and multiparous women have even generally associated with risk of
presumably through hormonal lower risk. Women with early age at breast cancer, although there may
mechanisms; further support for first birth also have a reduced risk, be some increased risk in younger
this derives from findings that and risk rises steadily with later ages women as well as in those who
obesity can affect risks associat first birth. Women with a first birth have either used oral contracep-
ated with use of menopausal at age 30 years or older are generally tives recently or used them before a
hormone therapy. at higher risk than nulliparous women, first birth (see Chapter 2.11).
presumably because of promotional Menstrual factors are also pre-
●● Studies are beginning to empha-
effects of pregnancy on previously dictive of risk. Early age at menarche
size the role of reproductive and
initiated cells in older mothers. These and late age at natural menopause
hormonal factors in the etiology
relationships are generally strong- are associated with the highest
of some cancer types in men, al-
est for hormone receptor-positive risks, presumably reflecting in part
though further studies are need-
tumours, and less conclusive effects an influence of ovulatory activity
ed to clarify risk relationships.
have been found for other breast (Fig. 3.6.1) [2]. These relationships
●● Recent advances in measuring cancer subtypes [1]. Pregnancy has appear to be consistent across risk
endogenous hormones support an effect on breast cancer risk only subgroups, including those defined
Chapter 3.6 • Reproductive and hormonal factors 189

by use of exogenous hormones. who undergo this operation before

Women who have an early surgi- age 40 years have approximately FUNDAMENTALS
cal menopause involving removal of half the risk of those who have a nat-
both ovaries have a lower risk; those ural menopause after age 55 years. ■■ Parity is strongly and nega
tively related to the risk of
breast cancer, endometrial
Fig. 3.6.1. Relative risk of breast cancer by (A) age at menarche and (B) age at
menopause, based on multiple studies. Calculated stratifying by study, age, year of cancer, ovarian cancer, and
birth, parity, age at first birth, smoking, alcohol consumption, height, and current body cervical cancer, supporting the
mass index. CI, confidence interval; gs, group-specific; RR, relative risk. notion that hormonal factors
are important contributors for
these cancer sites. Breast
cancer risk is further affected
by the woman’s age when her
first child is born.
■■ Use of oral contraceptives is
related to long-term reduced
risks of endometrial cancer
and ovarian cancer, but
does not have a generalized
effect on breast cancer risk.
Although use of menopausal
hormone therapy has been
recognized for some time as
being related to increased
risks of breast cancer and
endometrial cancer, it has
been more difficult to resolve
how changing prescribing
patterns (including the addition
of progestins to estrogen
therapy) affect risk.
■■ A variety of menstrual factors,
including age at menarche,
age at menopause, and type
of menopause, appear to be
related to risk of breast cancer,
endometrial cancer, and
ovarian cancer.
■■ Additional support for the
importance of hormonal factors
derives from findings that
obese women are at increased
risk of postmenopausal breast
cancer and endometrial
cancer, and that obesity
can affect the influence of
exogenous hormones.
■■ Until recently, investigations
that have attempted to assess
the influence of endogenous
hormones on various cancer
sites have been hindered by
the limitations of assays for
measuring hormones.
190
Reproductive and menstrual fac- risk, and conversion of androgens to being increasing recognized. Pooling
tors are major risk factors and can be estrogens in adipose tissue appears efforts have provided evidence that
used to estimate individual risks via to influence the increased risk. estrogens and androgens are direct-
the Breast Cancer Risk Assessment Menopausal hormone use has ly related to both hormone receptor-
Tool (http://www.cancer.gov/bcrisk been associated with increased positive and hormone receptor-nega-
tool/) and other risk prediction mod- breast cancer risk in post men o tive breast cancers [4], and additional
els. Despite the well-recognized role paus al women, and the highest analyses that use more precise hor-
of reproductive and menstrual fac- risks have been observed in thin mone measurement techniques may
tors in breast cancer etiology, stud- women. The type of hormones provide further clarity about relation-
ies have been unable to relate these used is also a major predictor of ships. Mass spectrometry–liquid
factors to specific underlying biologi- risk; higher risks are observed for chromatography assays that enable
cal mechanisms. It is generally as- use of estrogen plus progestin than measurements of 15 individual es-
sumed that changes in endogenous for use of unopposed estrogen ther- trogen metabolites have shown an
hormonal profiles are involved, but apy. This has been hypothesized as important etiological role for parent
additional research is needed to being due to mitotic influences of estrogens and individual estrogens,
clarify the effects. It is also unclear progestins on breast tissues. as well as for certain hydroxylation
how hormonally induced changes in Endogenous hormones are im- pathways (Fig. 3.6.3) [5]. Additional
breast tissue are involved. Recent portant predictors of breast cancer research is needed to assess the
attention has focused on the effects risk, although it has been difficult for influence of other endogenous hor-
of parity on involution of lobules, the studies to fully define relationships mones, such as androgens and pro-
structures from which the majority of with either breast cancer risk or pat- gestogens, on risk, both overall and
CHAPTER 3.6
SECTION 3
breast cancers are thought to arise terns of risk factors (see Chapter according to the hormone receptor
(Fig. 3.6.2) [3]. 5.9). This probably reflects difficul- status of the tumours.
The relationship of obesity with ties in measuring hormones or the
breast cancer risk is complex (see complexity of patterns of many in- Endometrial cancer
Chapter 2.7). Obesity is inversely re- terrelated markers, including not Endometrial tissue is extremely
lated to risk of premenopausal-onset only estrogens but also androgens, hormonally responsive, and endo-
breast cancer and is directly asso- progesterone, prolactin, and insulin- metrial cancer is believed to arise
ciated with risk of postmenopausal like growth factors. In addition, the as a result of estrogen stimulation
breast cancer. Obesity-associated importance of large inter-individual that is unopposed by progestins.
anovulation has been hypothesized differences in metabolism, which One of the strongest risk factors for
as responsible for the decreased may have etiological implications, is postmenopausal-onset endometrial
cancer is obesity (see Chapter 5.11),
presumably reflecting the conver-
Fig. 3.6.2. Assessment of terminal ductal lobular unit (TDLU) involution in the Susan
G. Komen Tissue Bank. Three quantitative measures (TDLU count, TDLU span, and sion of androstene dione to estrone
number of acini per TDLU) associated with reduced levels of TDLU involution were in adipose tissue. Particularly high
assessed. (A) Digital haematoxylin–eosin section with multiple TDLUs (TDLU count). risks have also been noted for use of
For up to 10 TDLUs per section, the longest TDLU span was measured and the counts unopposed estrogen therapy, which
of acini per TDLU were categorized. (B) Representative TDLUs for which the longest has been associated with 2–10-fold
TDLU span was measured. A representative acinus is circled in red and indicated with increases in risk, depending on the
an arrow.
duration of use and the woman’s
body size (higher relative risks are
observed in thin women). Use of
tamoxifen has also been strongly
related to an increased risk of endo-
metrial cancer.
In contrast to breast cancer, for
which especially elevated risks are
associated with use of estrogen
plus progestin menopausal hor-
mone therapy (combination ther-
apy), endometrial cancer shows a
favourable risk profile for such us-
ers. Data from the Women’s Health
Initiative clinical trial support that
relative risks are substantially lower
for users of combination therapy

Fig. 3.6.3. Odds ratios (ORs) and 95% confidence intervals comparing the risk of breast cancer in individuals with a higher analyte
or pathway concentration (90th percentile) with that in individuals with a lower concentration (10th percentile).
than for non-users of hormones obese non-hormone users (who are to elevated risks of endometrial can-
(Fig. 3.6.4) [6]. at higher risk than obese users of cer in premenopausal women, for the
These risks also appear to be continuous estrogen plus progestin more commonly used combined oral
modified by body mass, although in therapy), although the confidence in- contraceptives (a combination of es-
contrast to the situation for use of un- tervals on these risks are often broad trogen and progestin), use has been
opposed estrogen therapy, the great- and overlapping (Fig. 3.6.5) [7]. The related to substantial reductions in
est reductions in relative risks are effects of combination therapy may risk. Long-term users have the lowest
seen in heavier women. Because of also be influenced by how it is pre- risk, and the reduction in risk persists
these complexities, more meaning- scribed (estrogens given sequentially for some time after discontinuation of
ful insights can be derived by a focus vs continuously), but studies are only use [8]. Although the progesterone
on absolute risks. The lowest risks beginning to investigate this issue. content of the pills used may affect
are seen in thin women (either non- Although use of sequential oral risk, studies have not been able to
hormone users or users of continu- contraceptives (estrogen-only pills confirm this hypothesis.
ous estrogen plus progestin therapy), followed by progestin pills for a limit- Nulliparous women have high
and the highest risks are observed in ed number of days) has been related risks of developing endometrial
192
Fig. 3.6.4. Kaplan–Meier estimates of cumulative hazards of endometrial cancer in tumours, including serous cancers).
the Women’s Health Initiative randomized trial of continuous combined estrogen plus The tumour subtypes have been
progestin with the intention-to-treat principle. CI, confidence interval; HR, hazard ratio;
shown to be etiologically heteroge-
y, years.
neous, and stronger relationships of
hormonal risk factors (such as obe-
sity and parity) are seen for type 1
tumours than for type 2 tumours.
Ovarian cancer
Nulliparity is a well-recognized risk
factor for ovarian cancer, as is in-
fertility. Although there has been
extensive controversy about the
potential effects of fertility drugs,
the latest studies suggest that the
indications for use are more im-
portant than the drugs themselves
(see Chapter 2.11). Endometriosis
is a well-established predictor of
certain types of ovarian cancer, in-
CHAPTER 3.6
SECTION 3
cluding clear cell and endometrioid
cancers (Table 3.6.1) [10]. Unlike for
breast cancer and endometrial can-
cer, body size is not strongly related
to risk of ovarian cancer, although
it may have some modest effect for
certain tumour subtypes.
Some studies have suggested
elevated risks with early age at men
arche and late age at menopause,
but the results are not entirely con-
sistent. Substantially reduced risks
have been observed in women who
have had a simple hysterectomy or
tubal ligation. Although this finding
cancer, and multiparous women Although it is recognized that may reflect detection of abnormali-
have the lowest risks, but no effect hormonal factors have a strong role ties and removal of ovaries during
on risk has been demonstrated ac- in the etiology of endometrial cancer, either of these procedures, more
cording to age at first birth. Instead, relatively few studies have assessed recent attention has focused on the
age at last birth or interval since last the role of endogenous hormones effects of partial devascularization
birth may be important contributors in the etiology of endometrial can- or partial removal of tubes, given
cer, and it has often been difficult to increasing evidence of the tubal ori-
to risk, although studies are still at-
disentangle effects of endogenous gin of many serous cancers.
tempting to understand these rela-
hormones from those associated Use of oral contraceptives is
tionships. Early age at menarche
with obesity. A recent study showed related to substantial reductions
and late age at menopause are even
that parent estrogens and individual in the risk of ovarian cancer, par-
stronger risk factors for endometrial
estrogen metabolites all appear to ticularly when long-term use is
cancer than for breast cancer, pre- involved. However, use of meno-
exert uterotropic activity [9], but fur-
sumably because these parameters ther studies are needed to clarify the pausal hormones has been linked
indicate an enhanced opportunity effects on endometrial cancer risk with increases in risk (Fig. 3.6.6)
for circulating estrogens to influ- of additional hormones, including [11]. This has been most clearly
ence risk. Like for breast cancer, androgens. In such studies, it will demonstrated for unopposed es-
recent efforts have been made to be important to distinguish patterns trogen therapy, but there is growing
develop individualized risk predic- of risk according to specific tumour evidence that combined estrogen
tion models based on identified risk subtypes (e.g. type 1 or endometri- plus progestin therapy may also be
factors for endometrial cancer. oid vs the rarer type 2 endometrial linked with elevated risk [12].

Fig. 3.6.5. Age-standardized incidence of endometrial cancer by use of menopausal which there is growing evidence of
hormone therapy and body mass index, from the United States National Institutes of etiological heterogeneity [14].
Health-AARP (NIH-AARP) Diet and Health Study. Error bars indicate 95% confidence
interval on the age-standardized incidence rate. EPT, estrogen plus progestin therapy;
ET, unopposed estrogen therapy; MHT, menopausal hormone therapy. Cervical cancer
Infection with human papillomavirus
(HPV) is recognized as a neces-
sary cause of cervical cancer, but
other co-factors are important (see
Chapter 5.10). Although the rela-
tionship of reproductive factors with
cervical cancer risk is controversial,
one project that involved combining
data from 25 epidemiological stud-
ies demonstrated that risk of inva-
sive cervical cancer increased with
the number of full-term pregnancies
within each stratum of age at first
full-term pregnancy, and vice versa
(Fig. 3.6.7) [15].
The same investigation found
an increased risk of cervical cancer
related to current and long-term use
of oral contraceptives. The relation-
ship of risk with use of menopausal
hormone therapy remains less clear.
There is some evidence that endog-
Although many of the identified Conflicting results have emerged enous sex steroids, particularly tes-
risk factors for ovarian cancer are about the respective roles of estro- tosterone and estradiol, may play an
consistent with a protective effect gens, androgens, follicle-stimulat- etiological role [16], but it remains
of reduced ovulation, this does not ing hormone, sex hormone-binding unclear how hormonal factors might
appear to entirely explain all of the globulin, and insulin-like growth interact with HPV. Studies are also
identified risk factors (see Chapter factor [13]. Further investigation needed to separately examine rela-
5.12). Recent attention has focused appears to be warranted, particu- tionships for squamous cell cancers
on the possible role of hormonal and larly with respect to specific ovarian versus adenocarcinomas, given sug-
immunological factors (including in- cancer subtypes, especially serous gestions that adenocarcinomas may
flammation) and their interplay. versus non-serous tumours, for be more affected by hormonal risk
Table 3.6.1. Associations between history of endometriosis and the histological subtypes of ovarian cancer
Histological subtype Stratified and adjusted ORa (95% CI) P value
Invasive 1.46 (1.31–1.63) < 0.0001
Clear cell 3.05 (2.43–3.84) < 0.0001
Endometrioid 2.04 (1.67–2.48) < 0.0001
Mucinous 1.02 (0.69–1.50) 0.93
High-grade serous 1.13 (0.97–1.32) 0.13
Low-grade serous 2.11 (1.39–3.20) < 0.0001
Borderline 1.12 (0.93–1.35) 0.24
Mucinous 1.12 (0.84–1.48) 0.45
Serous 1.20 (0.95–1.52) 0.12
CI, confidence interval; OR, odds ratio

a
Stratified by age (5-year categories) and ethnic origin (non-Hispanic White, Hispanic White, Black, Asian, other), and adjusted for duration of oral contra
ceptive use (never, < 2 years, 2–4.99 years, 5–9.99 years, ≥ 10 years) and parity (0, 1, 2, 3, ≥ 4). Pooled analysis of 13 ovarian cancer case–control studies:
1 in Australia, 3 in Europe, and 9 in the USA.
194
Fig. 3.6.6. Relative risk of ovarian cancer by duration of use in current and past users of hormone therapy. * Risk relative to
never-users of hormone therapy, stratified by age at diagnosis, study, and body mass index, and adjusted for age at menopause,
hysterectomy, oral contraceptive use, and parity. CI, confidence interval.
CHAPTER 3.6
SECTION 3
Fig. 3.6.7. Relative risks (RRs) of invasive cervical carcinoma and corresponding 95% floating confidence intervals (FCIs) by number
of full-term pregnancies (FTPs) stratified by age at first FTP. 1 Conditioned on age and study or study centre. 2 As in 1, and conditioned
on age at first sexual intercourse and lifetime number of sexual partners.

factors such as obesity and use of tors, with suggestions of increased served was an inverse relationship
exogenous hormones. risks related to obesity, Klinefel with sex hormone-binding globulin.
ter syndrome, and gynaecomastia Use of finasteride reduces risk
Testicular cancer (Table 3.6.2) [18]. Data are also of prostate cancer by blocking the
beginning to emerge that implicate conversion of testosterone to di-
Hormonal factors play a role in the
the importance of endogenous hor- hydrotestosterone; use has also
etiology of testicular cancer, as evi-
mones (particularly estrogens) in the been associated with increases
denced by the rise in incidence start-
etiology of male breast cancer [19]. in estradiol levels. The Prostate
ing at adolescence and a variety of
Cancer Prevention Trial has shown
risk factors, including height, subfer-
substantial reductions in prostate
tility, and possibly exposure to endo- Prostate cancer cancer incidence associated with
crine disrupters (see Chapter 5.14).
Prostate cancers respond well to exposure to finasteride (https://
Several risk factors also support an
anti-androgen therapies, and both w w w.cancer.gov/types/prostate/
influence of exposures received in
surgical and medical castration research/prostate-cancer-preven
utero, including cryptorchidism, hy-
results in substantial reductions tion-trial-qa). This has raised ques-
pospadias, inguinal hernia, low birth
in the risk of metastatic disease. tions about whether estrogen levels
weight, short gestational age, and
Although it has been assumed that may play a role in prostate cancer
being a twin, some of which may
androgens play a role in the etiol- etiology (see Chapter 5.13). The
reflect the influence of endogenous
ogy of prostate cancer, studies to fact that trial participants who de-
hormones [17]. Recent studies have
veloped prostate cancer while tak-
attempted to assess the role of en- date have provided conflicting evi-
ing finasteride experienced high-
dogenous hormones in the etiology dence of a role for any hormones
er-grade tumours has prompted
of testicular cancer, but further stud- as risk factors. One large pooling
interest in examining subgroup re-
ies are needed to fully understand project showed no association be-
lationships. The most recent study
the relationships. tween risk of prostate cancer and that assessed such relationships
circulating concentrations of tes- observed a strong inverse asso-
Male breast cancer tosterone, calculated free testoster- ciation between the ratio of estra-
The incidence of breast cancer in one, and conversion products; the diol to testosterone and aggressive
men is only about 1% that in wom- major conversion product is dihy- prostate cancer (Table 3.6.3) [21].
en, complicating the evaluation of drotestosterone, to which testoster- However, given the conflicting data
etiological factors. However, the few one is converted in the prostate by from other studies on the role of
available studies appear to implicate 5α-reductase (Fig. 3.6.8) [20]. The both estrogens and androgens in
several hormonally related risk fac- only evidence of association ob- the etiology of prostate cancer [22],
Table 3.6.2. Relationship of anthropometric and hormonal risk factors with risk of male breast cancer: results from the Male
Breast Cancer Pooling Project
Factors Odds ratioa (95% confidence interval)
Meta-analysis Case–control studies Cohort studies
Adult body mass index (kg/m2)
Lowest tertile, ≤ 24.6 1.00 (referent) 1.00 (referent) 1.00 (referent)
Middle tertile, 24.7–27.4 1.15 (1.00–1.33) 1.22 (1.02–1.46) 1.04 (0.81–1.32)
Highest tertile, > 27.4 1.30 (1.12–1.51) 1.39 (1.16–1.67) 1.16 (0.91–1.49)
Klinefelter syndrome
No 1.00 (referent) 1.00 (referent) 1.00 (referent)
Yes 24.73 (8.94–68.38) 22.30 (1.98–251.70) 25.28 (8.24–77.54)
Gynaecomastia
No 1.00 (referent) 1.00 (referent) 1.00 (referent)
Yes 9.78 (7.52–12.71) 14.57 (10.13–20.96) 6.34 (4.34–9.27)

a
Estimated via unconditional logistic regression, with adjustment for study and age.
196
additional studies are needed to ous reproductive and hormonal fac- attempted to assess whether repro-
clarify the relationship of hormones tors on other cancer types, there are ductive and hormonal factors are as-
to prostate cancer risk, both overall many inconsistent findings. Findings sociated with the risk of cancers of
and according to tumour subtypes. with respect to some of the better the stomach, thyroid, and central ner
studied cancer types, including can- vous system as well as melanomas,
Other cancer types cers of the colorectum [23], liver [24], again without conclusive results.
Although some studies have sug- and lung [25], are particularly dif-
gested possible influences of vari- ficult to decipher. Studies have also
Table 3.6.3. Associations between circulating sex steroid hormone concentrations and aggressive prostate cancer
Estrogen and estrogen metabolism measures Odds ratioa (95% confidence interval) P trend
Quartile 1 Quartile 2 Quartile 3 Quartile 4
All estrogens and estrogen metabolites 1.00 1.27 (0.72–2.23) 1.28 (0.72–2.27) 0.84 (0.46–1.54) 0.65
2-Hydroxylation pathway 1.00 1.53 (0.87–2.70) 1.27 (0.71–2.28) 0.94 (0.51–1.72) 0.69
2-Hydroxylation pathway catechols 1.00 1.35 (0.75–2.41) 1.63 (0.92–2.87) 0.87 (0.47–1.62) 0.95
2-Hydroxyestrone 1.00 1.48 (0.82–2.65) 1.49 (0.84–2.64) 0.91 (0.49–1.68) 0.89
2-Hydroxyestradiol 1.00 1.23 (0.70–2.16) 0.94 (0.52–1.68) 0.92 (0.52–1.65) 0.79
2-Hydroxylation pathway methylated catechols 1.00 0.54 (0.30–0.98) 0.71 (0.41–1.24) 0.59 (0.33–1.06) 0.14
CHAPTER 3.6
SECTION 3
2-Methoxyestrone 1.00 0.47 (0.26–0.85) 0.65 (0.37–1.13) 0.53 (0.29–0.94) 0.06
2-Methoxyestradiol 1.00 1.07 (0.61–1.88) 0.96 (0.54–1.70) 0.80 (0.44–1.45) 0.38
2-Hydroxyestrone-3-methyl ether 1.00 0.75 (0.42–1.33) 0.65 (0.36–1.15) 0.82 (0.47–1.44) 0.34
4-Hydroxyestrone 1.00 1.85 (1.05–3.28) 1.20 (0.66–2.17) 1.07 (0.58–1.97) 0.89
4-Hydroxylation pathway methylated catechols 1.00 0.63 (0.32–1.27) 0.60 (0.30–1.20) 0.54 (0.27–1.10) 0.09
4-Methoxyestrone 1.00 0.48 (0.24–0.97) 0.67 (0.34–1.30) 0.45 (0.22–0.92) 0.08
4-Methoxyestradiol 1.00 0.58 (0.29–1.17) 0.70 (0.36–1.37) 0.52 (0.25–1.06) 0.12
16α-Hydroxyestrone 1.00 1.54 (0.86–1.77) 1.73 (0.97–3.07) 0.84 (0.44–1.58) 0.87
Estriol 1.00 1.00 (0.57–1.75) 0.77 (0.43–1.38) 0.82 (0.46–1.49) 0.33
17-Epiestriol 1.00 0.73 (0.41–1.30) 0.85 (0.48–1.50) 0.83 (0.47–1.47) 0.72
16-Ketoestradiol 1.00 1.23 (0.69–2.19) 1.41 (0.80–2.48) 0.87 (0.47–1.60) 0.80
16-Epiestriol 1.00 0.75 (0.42–1.34) 0.75 (0.42–1.34) 0.72 (0.41–1.28) 0.30
Estrogen metabolic pathway ratios
2-Hydroxylation pathway:parent estrogens 1.00 1.54 (0.86–2.76) 0.85 (0.45–1.60) 1.69 (0.95–3.02) 0.24
2-Hydroxylation pathway:16-hydroxylation pathway 1.00 1.31 (0.71–2.42) 1.73 (0.96–3.12) 1.53 (0.84–2.79) 0.10
2-Hydroxyestrone:16-hydroxyestrone 1.00 1.24 (0.65–2.37) 1.87 (1.01–3.44) 2.44 (1.34–4.45) 0.001
2-Hydroxylation pathway methylated
1.00 1.12 (0.65–1.94) 0.56 (0.31–1.02) 0.73 (0.41–1.30) 0.08
catechols:catechols
4-Hydroxylation pathway methylated
1.00 0.60 (0.30–1.22) 1.01 (0.52–1.94) 0.37 (0.17–0.80) 0.08
catechols:catechols
a
Adjusted for age at blood draw, body mass index, and sex hormone-binding globulin. Boldface indicates findings that are statistically significant.

Fig. 3.6.8. Association between risk of prostate cancer and increasing fifths of hormone concentrations, from a collaborative analysis
of 18 prospective studies. The position of each square indicates the magnitude of the relative risk (RR), and the area of the square
is proportional to the amount of statistical information available. The length of the horizontal line through the square indicates the
95% confidence interval (CI). The chi-square 1 degree of freedom statistic for linear trend (χ21 for trend) is calculated by replacing the
categorical variables with a continuous variable scored as 0, 0.25, 0.5, 0.75, and 1. The P value was two-sided for statistical signif
icance of χ21 for trend. DHEA-S, dehydroepiandrosterone sulfate; DHT, dihydrotestosterone; SHBG, sex hormone-binding globulin.
198
References
1. Anderson KN, Schwab RB, Martinez ME 10. Pearce CL, Templeman C, Rossing MA, 19. Brinton LA, Key TJ, Kolonel LN, Michels
(2014). Reproductive risk factors and breast Lee A, Near AM, Webb PM, et al.; Ovarian KB, Sesso HD, Ursin G, et al. (2015).
cancer subtypes: a review of the literature. Cancer Association Consortium (2012). Prediagnostic sex steroid hormones
Breast Cancer Res Treat. 144(1):1–10. Association between endometriosis and in relation to male breast cancer risk.
https://doi.org/10.1007/s10549-014-2852-7 risk of histological subtypes of ovarian can- J Clin Oncol. 33(18):2041–50. https://
PMID:24477977 cer: a pooled analysis of case-control stud- d o i . o r g / 10 .12 0 0 /J C O . 2 0 14 . 5 9 .16 0 2
ies. Lancet Oncol. 13(4):385–94. https:// PMID:25964249
2. Collaborative Group on Hormonal Factors
doi.org/10.1016/S1470-2045(11)70404-1
in Breast Cancer (2012). Menarche, men- 20. Roddam AW, Allen NE, Appleby P, Key
PMID:22361336
opause, and breast cancer risk: individ- TJ; Endogenous Hormones and Prostate
ual participant meta-analysis, including 11. Beral V, Gaitskell K, Hermon C, Moser K, Cancer Collaborative Group (2008).
118 964 women with breast cancer from Reeves G, Peto R; Collaborative Group on Endogenous sex hormones and prostate
117 epidemiological studies. Lancet Oncol. Epidemiological Studies of Ovarian Cancer cancer: a collaborative analysis of 18
13(11):1141–51. https://doi.org/10.1016/ (2015). Menopausal hormone use and prospective studies. J Natl Cancer Inst.
S1470-2045(12)70425-4 PMID:23084519 ovarian cancer risk: individual participant 100(3):170–83. https://doi.org/10.1093/jnci/
meta-analysis of 52 epidemiological stud- djm323 PMID:18230794
3. Figueroa JD, Pfeiffer RM, Patel DA,
Linville L, Brinton LA, Gierach GL, et al. ies. Lancet. 385(9980):1835–42. https:// 21. Black A, Pinsky PF, Grubb RL 3rd, Falk
(2014). Terminal duct lobular unit involu- doi.org/10.1016/S0140-6736(14)61687-1 RT, Hsing AW, Chu L, et al. (2014). Sex
tion of the normal breast: implications for PMID:25684585 steroid hormone metabolism in relation to
breast cancer etiology. J Natl Cancer Inst. 12. Trabert B, Wentzensen N, Yang HP, risk of aggressive prostate cancer. Cancer
106(10):dju286. https://doi.org/10.1093/jnci/ Sherman ME, Hollenbeck A, Danforth Epidemiol Biomarkers Prev. 23(11):2374–
dju286 PMID:25274491 KN, et al. (2012). Ovarian cancer and 82. https://doi.org/10.1158/1055-9965.EPI-
menopausal hormone therapy in the NIH- 14-0700 PMID:25178985
4. James RE, Lukanova A, Dossus L, Becker
AARP Diet and Health Study. Br J Cancer.
CHAPTER 3.6
SECTION 3
S, Rinaldi S, Tjønneland A, et al. (2011). 22. Boyle P, Koechlin A, Bota M, d’Onofrio
Postmenopausal serum sex steroids and 107(7):1181–7. https://doi.org/10.1038/bjc. A, Zaridze DG, Perrin P, et al. (2016).
risk of hormone receptor-positive and -neg- 2012.397 PMID:22929888 Endogenous and exogenous testoster-
ative breast cancer: a nested case-control 13. Brown SB, Hankinson SE (2015). one and the risk of prostate cancer and in-
study. Cancer Prev Res (Phila). 4(10): Endogenous estrogens and the risk of creased prostate-specific antigen (PSA)
1626–35. https://doi.org/10.1158/1940-6207. breast, endometrial, and ovarian can- level: a meta-analysis. BJU Int. 118(5):
CAPR-11-0090 PMID:21813404 cers. Steroids. 99(Pt A):8–10. https:// 731–41. https://doi.org/10.1111/bju.13417
doi.org/10.1016/j.steroids.2014.12.013 PMID:26779889
5. Sampson JN, Falk RT, Schairer C, Moore
SC, Fuhrman BJ, Dallal CM, et al. (2017). PMID:25555473 23. Rennert G (2017). Reproductive factors,
Association of estrogen metabolism with 14. Wentzensen N, Poole EM, Trabert B, hormones and colorectal cancer-still unre-
breast cancer risk in different cohorts of White E, Arslan AA, Patel AV, et al. (2016). solved. Br J Cancer. 116(1):1–3. https://doi.
postmenopausal women. Cancer Res. org/10.1038/bjc.2016.388 PMID:27898659
Ovarian cancer risk factors by histologic
77(4):918–25. https://doi.org/10.1158/0008-
subtype: an analysis from the Ovarian 24. Zhong GC, Liu Y, Chen N, Hao FB, Wang
5472.CAN-16-1717 PMID:28011624
Cancer Cohort Consortium. J Clin Oncol. K, Cheng JH, et al. (2016). Reproductive
6. Chlebowski RT, Anderson GL, Sarto GE, 34(24):2888–98. https://doi.org/10.1200/ factors, menopausal hormone thera-
Haque R, Runowicz CD, Aragaki AK, et JCO.2016.66.8178 PMID:27325851 pies and primary liver cancer risk: a
al. (2015). Continuous combined estrogen systematic review and dose-response
15. International Collaboration of Epidemio
plus progestin and endometrial cancer: meta-analysis of observational stud-
logical Studies of Cervical Cancer (2006).
the Women’s Health Initiative randomized ies. Hum Reprod Update. 23(1):126–38.
Cervical carcinoma and reproductive fac-
trial. J Natl Cancer Inst. 108(3):djv350. https://doi.org/10.1093/humupd/dmw037
tors: collaborative reanalysis of individual
h t t p s : / / d o i . o r g / 10 .10 9 3 / j n c i / d j v 3 5 0 PMID:27655589
data on 16,563 women with cervical carci-
PMID:26668177
noma and 33,542 women without cervical 25. Siegfried JM, Stabile LP (2014). Estrogenic
7. Trabert B, Wentzensen N, Yang HP, carcinoma from 25 epidemiological studies. steroid hormones in lung cancer. Semin
Sherman ME, Hollenbeck AR, Park Y, Int J Cancer. 119(5):1108–24. https://doi. Oncol. 41(1):5–16. https://doi.org/10.1053/j.
et al. (2013). Is estrogen plus proges- org/10.1002/ijc.21953 PMID:16570271 seminoncol.2013.12.009 PMID:24565577
tin menopausal hormone therapy safe
with respect to endometrial cancer risk? 16. Rinaldi S, Plummer M, Biessy C, Cas
Int J Cancer. 132(2):417–26. https://doi. tellsagué X, Overvad K, Krüger Kjær S,
org/10.1002/ijc.27623 PMID:22553145 et al. (2011). Endogenous sex steroids
and risk of cervical carcinoma: results
8. Collaborative Group on Epidemiological from the EPIC study. Cancer Epidemiol
Studies on Endometrial Cancer (2015). Biomarkers Prev. 20(12):2532–40. https://
Endometrial cancer and oral contracep- doi.org/10.1158/1055-9965.EPI-11-0753
tives: an individual participant meta-anal- PMID:21994406
ysis of 27 276 women with endometrial
cancer from 36 epidemiological studies. 17. McGlynn KA, Trabert B (2012). Adolescent
Lancet Oncol. 16(9):1061–70. https://doi. and adult risk factors for testicular can-
org /10.1016 /S1470 -20 4 5(15) 0 0212- 0 cer. Nat Rev Urol. 9(6):339–49. https://doi.
PMID:26254030 org/10.1038/nrurol.2012.61 PMID:22508459
9. Brinton LA, Trabert B, Anderson GL, Falk 18. Brinton LA, Cook MB, McCormack V,
RT, Felix AS, Fuhrman BJ, et al. (2016). Johnson KC, Olsson H, Casagrande JT,
Serum estrogens and estrogen metabolites et al.; European Rare Cancer Study Group
and endometrial cancer risk among post- (2014). Anthropometric and hormonal
menopausal women. Cancer Epidemiol risk factors for male breast cancer: Male
Biomarkers Prev. 25(7):1081–9. https:// Breast Cancer Pooling Project results. J
doi.org/10.1158/1055-9965.EPI-16-0225 Natl Cancer Inst. 106(3):djt465. https://doi.
PMID:27197275 org/10.1093/jnci/djt465 PMID:24552677

3.7
Metabolic change and metabolomics

Emerging approaches and new insights
Augustin Scalbert Demetrius Albanes (reviewer)

Marc Gunter A. Heather Eliassen (reviewer)
James R. Krycer (reviewer)
Metabolism refers to the sum of a the multivariate metabolic respons-

SUMMARY large number of chemical reactions es of a cell, tissue, or organism to
that occur in biological tissues. These pathophysiological stimuli or genet-
●● Metabolomics has been applied reactions involve several thousands ic modification [2]. Metabolomics
to blood, tissue, and other bio- of metabolites, organized in hundreds was initially proposed as an ap-
specimens in cancer research. of metabolic pathways. More than proach to compare metabolic pro-
Comparison of metabolic profiles 18 000 metabolites have been detect- files in various biological samples –
in tumour samples and in normal for example, in samples from indi-
ed in human tissues or biofluids, and
tissues leads to the identification
about 80 000 have been predicted viduals with specific diseases com-
of metabolic pathways that are
but still await precise characterization pared with those from healthy sub-
more specific for tumours.
[1]. The totality of these metabolites jects (Fig. 3.7.1).
●● The metabolites that are most constitutes the metabolome. Typically, samples are analysed
commonly reported as cancer Compared with the genome, the using nuclear magnetic resonance
discriminants in case–control epigenome, the proteome, and the (NMR) spectroscopy or mass spec-
studies include various amino transcriptome, the metabolome is trometry (MS), two universal ana-
acids, nucleotides, polyamines, the most downstream expression or lytical techniques that are able to
sugars, organic acids from the “readout” of the phenotype. A single measure a multiplicity of organic
tricarboxylic acid cycle, and determination of the metabolome compounds in complex matrices
bile acids. defines metabolic phenotypes that such as blood, urine, or tissues.
●● In the past 5 years, 15 pro- characterize an individual at a given NMR spectroscopy is a robust meth-
spective metabolomics studies time in their life. Metabolic pheno- od that is well adapted to the analysis
on cancers of the colorectum, types vary between individuals and of large series of samples. However,
liver, pancreas, prostate, and within individuals (e.g. repeated sam- MS is a much more sensitive tech-
breast have been published, ples over time) according to diverse nique that enables the measure-
with the number of case–con- factors, such as genotype, age, body ment of hundreds to thousands of
trol pairs varying from 100 to mass index, disease status, physical metabolites in a single sample. For
more than 1000. exercise, diet, and other environmen- this reason, it is now widely used in
tal factors. Metabolic phenotypes metabolomics studies.
●● Prospective studies that show can be measured in biofluids or tis- These techniques can be ap-
associations of blood metabo- sues, including, for example, tumour plied to various biospecimens, such
lites several years before diag- as blood, urine, tissue, saliva, faecal
tissues, and they provide valuable
nosis with risk of cancer suggest samples, or hair. Although many me-
information on the mechanisms that
new pathophysiological mecha- tabolites are shared between these
link metabolism and diseases.
nisms that lead to cancer. matrices, they also differ in some
●● Metabolomics is emerging an aspects, such as ease of collection,
Metabolomics as
essential tool, complementary to chemical composition, stability dur-
genomics, transcriptomics, and
a powerful tool to ing storage, and intra-individual re-
proteomics, to identify novel bio-
characterize metabolic producibility in a particular individual
markers for cancer and to better phenotypes over time. The selection of a matrix
understand cancer etiology. Metabolomics has been defined as or matrices will depend on the par-
the quantitative measurement of ticular study and its objectives.
200
Fig. 3.7.1. Metabolomics workflow. MS, mass spectrometry; NMR, nuclear mag-

netic resonance. FUNDAMENTALS
Case–control studies ■■ Metabolic profiles are
Prospective studies defined by the nature and
concentrations of low-
molecular-weight compounds,
which are naturally present in
human biospecimens such as
Controls Cases blood, urine, or tissues. These
Sample collection
N samples (blood, urine, tissue, etc.) N′ samples compounds are products of the
metabolism and are described
Sample analysis
as metabolites. Metabolic
(NMR spectroscopy, MS) profiles characterize the
human phenotypes.
Data pre-processing ■■ Metabolomics compares
metabolic profiles in various
Statistics individuals. When applied to
290.155@4.5
people at risk of developing
cancer and to healthy
0.08 424.898@0.6 272.948@0.6
SECTION 3
CHAPTER 3.7
Characteristic signals
183.049@4.5 300.213@10.6
286.126@1.9 205.098@5.6 246.169@5.1
110.008@0.5 247.13@2.7 181.069@5.5
0.06 297.11@4.6 272.186@8.1
384.115@6.2
241.155@1.7
252.074@6.2
454.208@11.1
189.075@5.6 139.001@1.3
147.042@9
266.139@6.3
330.227@10.1 231.145@1.3
209.089@8.6
138.052@0.7 262.163@1.4
367.148@3.4
201.124@4.3
257.114@3 170.039@1.7
262.039@2.9
302.226@11.2
244.155@4.8
201.124@5.1
216.034@2.9
181.068@6.3 167.051@3.8
177.054@9.3 235.17@9.9
182.081@1.7
120.078@2.5 230.158@1.5
241.156@1
197.066@6.2
278.176@11.3203.083@1.7
151.058@4.9217.131@0.9 164.924@0.6 202.049@6.8
281.099@7.9
165.053@1.7
413.142@7.9 531.43@22.5 288.923@0.6 147.041@6.5 167.054@4.4
397.072@6.8
226.954@0.6 144.063@1
464.186@11.5 282.121@1.1
145.047@2.4196.061@6.1
344.229@9.5
413.124@5.2
303.214@11.2
269.125@0.9
153.06@2.8
276.145@1.5 429.174@10.7 218.984@0.5
486.198@6.4 223.991@0.5
247.129@4.3276.161@11.1 261.144@5.5
241.094@3.9
310.182@7.3 195.139@10.7
137.042@1.5
326.196@11.1 153.048@4.9139.986@22 576.131@6.8
181.071@6.9
0.04 251.129@13.3
120.042@3.8
231.142@11.4257.15@9.3 290.16@3.4
215.017@1.3
414.146@7.9 330.061@4.4
239.14@10.1
196.084@1.9314.232@11.5 389.182@10.7
166.071@1.2
188.066@5.6
342.227@9.5 283.106@4.9
455.214@11.2 247.138@5.1 411.162@11.6
329.179@11.4
281.15@10.7 111.007@1.3 219.171@0.7 303.086@11.4 145.103@
290.136@2.1
175.078@5.9 452.192@11
203.154@0.7 217.119@2.5
209.121@6.4
241.182@11.6
116.069@1.3 267.061@1.1
195.139@11.6
213.108@5.3
182.067@4.6 291.165@4.5
223.169@11.6 226.085@7
347.226@11.6
223.022@1.3 130.048@2.3
343.228@9.6180.084@5.7
312.13@5.7
135.052@5.8 125.022@3.9
166.083@2.5
154.989@22.9 118.083@0.7
243.002@0.5 241.999@0.5
142.084@0.9
198.942@0.5
200.974@0.5
235.109@5.8 182.068@5.4
328.244@12.3
259.157@11.7 211.145@8.7
281.114@6.2
304.129@9285.186@10.1 195.088@7.6
279.112@8.3 123.039@0.5
166.071@4.9
286.221@16.2
265.018@0.5
208.049@1.4 287.199@0.7
369.154@10.6
331.228@11.3
310.202@10.4 185.123@5.1
229.152@1.5
166.068@1.8 326.094@4.4
114.055@1.2
223.098@8.8 209.083@9.6 340.176@7.6
339.095@5.6
129.015@1.4 206.084@10.4 247.143@6.5
220.119@4.9
186.113@8.6 146.078@1.3
273.08@11.4
(lipids, sugars, amino

300.108@2
240.107@1.9 663.462@21.8 312.212@10.5
subjects, it provides new data

0.02 145.047@1.8 286.197@9.9
310.2@10.9 319.165@8.6
245.151@8.5 209.057@1.1 154.058@5.2
195.114@5.4
365.194@10.7 105.023@6.9
567.178@7.5
130.047@1.2
106.037@6.8
475.405@22 517.424@22.6180.9@0.6 83.088@8.7
267.174@16.4
438.981@1.4 135.039@8.8
329.208@6.2 274.094@0.9
371.171@10.7 346.125@1.4
256.13@6
231.142@10.7
155.043@1.1
149.13@10.7
202.119@1.4
100.075@4.5
297.144@9.6 357.105@5.6
346.223@9.8
243.135@7.6 297.128@5.2
365.233@11.8 296.149@5.6
329.17@9.8
181.072@1.7
301.222@10.6 181.07@1
583.151@7.5
267.135@10.1
487.412@22.6 170.058@11
249.11@0.7
279.176@11.3 491.403@22.3 547.44@22.5
244@0.5 121.024@6.1
175.107@2.5449.112@11.1
465.197@11.5 400.27@10.5 385.12@6.2
225.186@12.6 207.138@8.6
279.133@9 419.017@4.5
229.151@0.8 253.151@10.6
175.123@5.3
237.125@5.3
130.048@8.5
365.096@4.4
129.052@5.2
184.054@4.5 247.136@11.2
286.221@17.6
259.094@0.8
209.119@13.3 337.066@1.7
167.052@4.8
180.062@6.8 242.928@0.6
245.093@10.9 328.268@16.3
311.276@19.2
247.115@10.6 160.098@0.7
86.098@1.2 95.013@4
271.165@10.5469.106@1.5
201.055@12.3
282.105@7.9
203.111@9.8
365.105@0.8
337.176@8.4
329.253@12.3
287.205@9.9
164.038@9.9 128.95@22.9 346.223@6.8
254.162@2.4
258.118@3
318.191@7.5 266.122@7.5
233.158@2.3
201.088@1.5
171.088@2.6 197.067@5.4
304.174@5.4 287.225@18.3
326.38@15.9
128.017@0.5
247.109@7.5 354.155@7.1
84.046@1.2368.259@18.8
286.218@19.6 179.071@8.9
395.096@8.9
153.052@7.1
285.183@11.5
217.099@6.9 287.133@1.9
330.225@11.3
243.135@7.2 136.072@1.7
176.07@9.8 220.946@0.6
243.16@11.2
385.311@12.8 288.178@5.1
366.237@11.8
268.108@2.3 303.073@4.5
190.049@8 136.073@7.5
347.282@19.6
328.266@18.8 153.038@2
304.231@15.3
147.074@7.5
129.064@7.5
245.082@1.1
239.969@0.5 151.033@0.5
391.287@22.1
304.897@0.6
260.187@7.1
171.087@1.5
328.266@18.2
198.073@5.3
171.087@4.3 182.986@0.4
287.225@18.8 144.097@0.7
402.285@10.7
252.133@13.3
147.041@1.3
146.057@5.6 136.062@2.3
450.176@11.3
226.145@11.1
211.06@6
204.123@0.8
121.027@8.2 179.046@1.1
354.13@5
204.088@12.7 421.157@8.7
86.062@2.1 279.162@22.1
196.092@7.6
341.172@8
304.23@18.3
235.12@1.1
304.212@8.8
215.104@2.9
169.093@3.7 169.032@1.7
136.046@0.7
413.043@6.8
127.037@4.5 311.176@10.7 211.165@10.6
229.139@11
358.256@11.7
445.034@6.1
304.23@17.8
305.3@14.5 395.389@22.2
346.279@15.8
274.202@9.5 347.166@10.6
-0.00 313.135@5.7 206.102@5.6
256.119@2 130.061@9.8
459.138@9 227.104@4.1
332.208@11.1 273.191@8.1
199.109@3.9
182.067@2.3
176.12@5.3
457.206@11.6
369.239@9.5 140.036@1.4
195.079@1.9
171.066@5.1
219.108@5.4 231.071@5
458.204@11.2
368.154@3.4
167.012@0.5
445.368@21.4
182.075@6.3
315.233@11.6 304.298@14.5
227.079@1.6 332.206@8.7
287.225@19.6
151.035@7130.047@6.2 512.827@22.9 153.123@10.6
304.093@11.4
263.141@7.4 169.978@22.9 124.039@3.9
374.255@10.1
306.861@0.6
250.04@1.5
302.197@9 354.101@14.9
372.233@9.4
258.063@2.1
237.149@12.9 298.116@5
304.068@2.1 288.217@10.2
232.153@2.3
241.146@10.6 149.022@22.1
175.086@1.7 205.093@0.8
214.92@0.5
198.071@6.2
329.271@16.8
327.08@16.8
305.235@16.8 117.071@18.8
368.259@18
197.068@4.8
292.19@11.3
157.015@1.3
216.953@0.5
278.126@0.9
183.051@3.7 131.045@10.5
413.265@22.1
414.273@22.1
331.231@10.2 183.081@15.3
109.064@8.5 313.223@10.5
279.135@4.8
341.048@1.3
302.197@7.1
214.084@1.4 476.193@12.2
414.129@5.2
467.154@13.3
255.169@11.7 185.068@1.7
368.254@16.1
356.244@11.5
206.046@8.2
221.093@1.7
209.13@10.6 555.311@13.1 267.143@6.3
124.085@20.8
338.088@6.8 84.962@0.5
440.872@0.6
503.41@21.9
311.154@9.9
561.45@22.6 377.142@8.1
418.173@4.8
685.451@21.9 288.224@19.6
415.079@6.8
71.064@22.4 211.171@9.7
218.14@1.2
504.282@11.7
414.302@13.7
111.043@6 440.212@9.1
251.131@10.6
213.154@10.7 394.246@8.7
169.122@10 745.465@18.8
197.126@6.7
332.244@9.5
475.245@10.9
209.064@1.7
875.363@6 71.064@21.9
168.058@3.8 285.155@10.7
124.085@21.3 112.017@22
304.231@16
w*c[2]
205.055@1.8
462.235@9.4
124.038@5.8
167.071@11.2
222.083@1.8
132.1@1.2
182.083@2.9 456.242@10.2
114.127@11.9
334.187@5
199.104@2.6
158.964@22.9 319.197@7.5
303.888@22.9
466.2@11.5 298.1@6
278.144@9 76.078@0.7
402.285@11.7
230.159@0.9
338.108@8.4
393.153@8.6
271.169@11 288.219@10.8
291.07@9.2
213.099@4.6 171.087@2.1
458.859@0.5
199.083@1.5 237.125@9
162.111@0.7
247.109@3.4
232.929@0.5
273.956@0.5 320.871@0.6
346.278@16.3 249.114@10.6175.148@9.7
368.259@19.6
413.269@22.6
329.272@18.2
124.085@21.9 170.06@11.6
305.235@17.6
523.07@2.9405.265@12.9
261.119@9
356.245@11.2
191.045@1.1
480.179@9.9 368.257@17.3
154.049@2.3 219.16@12289.119@10.1
259.167@9.9
232.146@11.4
242.079@3.8
147.074@10.5 333.116@7.2
181.087@10.3
240.125@9.3
281.114@5.6
326.039@7.7 149.022@17.7
368.255@16.8
346.279@15.3
194.104@4.6
346.278@17.6
329.271@19.6
308.18@10.6
227.127@0.7
304.229@19.6 267.133@8.5 265.135@10.6
286.222@15.6 442.88@0.5
97.077@22.5
234.963@0.5 377.147@8.8
217.108@9.1
211.072@4.8
304.23@18.8
230.215@19.6
188.071@6.5 487.362@18.8
322.079@2.1
374.086@9.8
321.208@11.3 283.124@1.1
130.048@5.6
129.052@6 237.167@12.2
258.155@9.3 346.278@18.3
271.207@11.8
327.092@9.1
345.239@9.6 114.056@0.7
744.462@19.6
117.071@19.6
153.009@22.9
168.056@4.9
333.151@8.6
431.065@6
293.149@10.5 126.09@5.2 255.213@13.2
344.244@11.5 287.101@7.5
304.231@17.3
346.276@19.6
149.021@14.2
438.236@8.7 531.39@18.7
300.134@11.4
328.264@19.6
200.095@2.5
328.268@15.6
465.25@12.3
302.16@3.9 330.27@19.6 286.221@16.8
227.127@12.1
acids, xenobiotics, etc.)

-0.02 448.306@13.7
220.107@9.7
204.135@1.3
123.043@1.7 139@2.5
255.136@5.4
269.141@13.3 191.056@8
139.047@4.2 378.152@8.8
260.151@1.4
248.17@11.4 284.06@6
207.112@5.2 124.084@22.5
300.218@11
287.225@17.6
327.202@11.1 497.146@11.4
177.054@14.2
329.272@18.8 804.556@22.1
281.119@11.4
117.072@16.8
199.084@2.1
355.284@19.1
186.973@22.9 121.023@9.4 195.095@9.2 232.064@2.4
272.187@9.1
295.118@8.7
284.187@8.8
342.173@11.8 744.461@18.6
242.159@1.8 453.177@12.9
305.208@11.3 305.235@18.2
553.291@12.6 255.946@0.5
273.223@13.2
144.981@22.5
745.465@19.6
226.182@13.1
274.095@6
257.228@12.9
240.233@13.7 483.151@10.7
202.111@5
167.036@22.1
304.231@16.8 117.072@18111.02@22.9226.955@22.7
239.091@9.1
271.093@4.2
265.111@7.5
283.129@9.3 195.073@7.2
212.104@0.8150.073@1.1
276.115@2.1
254.14@10.5 117.072@17.3
373.272@12.7
357.28@13.2
358.223@10 497.241@11.3
152.021@0.7 171.067@10.9
295.131@6.8
369.265@19.6587.491@22.6
453.791@22.9
130.061@11
130.042@7.5 138.995@22.9
141.977@22.9
195.105@12.6 188.103@1.1
232.134@2.9
189.136@0.8
330.083@6.9
132.078@0.7
201.03@6.5 191.085@10.2
283.172@6.7
302.188@4.9
154.068@2.8
248.1@5 293.141@10.8
360.094@9.4
463.16@11.2
330.192@7.9
267.125@10.6
519.222@11.3 359.263@11.8
326.113@9.1
305.236@19.6 248.093@7.5
258.902@0.6
162.054@6.8
417.374@22.2
369.258@18.3
291.165@3.4
347.282@18.8
301.143@17.6
306.234@19.6 256.105@1.1
207.067@6.3
201.164@11.3 227.126@9.9 479.229@11.3
305.235@18.8
253.193@11.8
429.229@11.8
196.875@0.6 369.262@18.8
328.267@17.6 131.052@7.5
160.107@1.3 111.019@22
316.249@12
346.261@11.5
449.311@13.7 110.018@22.9
102.034@22.9
329.272@17.6
97.077@21.9
347.281@17.6 328.214@9.5 112.02@22.9
71.064@22.8
299.128@11.4 292.156@6.9
115.068@0.7
on metabolic pathways that

-0.04 201.165@12
330.226@9.2 197.082@10.8
286.121@9 137.048@5.3
304.209@11.3 113.059@5.8 257.979@0.5
184.989@22.1
529.232@7.5 484.292@13.1
375.325@20.8
245.159@4.9
328.267@16.8 443.332@18.9
125.985@0.5
803.551@22.1
173.08@9.1
312.947@22.9
289.222@10.2
276.182@3.8 125.106@5.2
257.158@12745.465@18.3
173.08@2.3 346.277@18.8
399.31@19 249.099@7.5 213.148@11.7
353.343@20.8
265.146@11.2 184.988@22.9
320.17@8.7 300.182@6.2
257.192@11.3263.104@8.1 176.07@11
98.513@22.9 167.015@22.9 287.894@22.9
313.206@11.1316.213@8.5 121.081@2.5
150.053@6.4
468.159@13.3 545.447@22.6
130.062@11.7
744.462@17.8
409.151@13.5 347.282@18.2
98.984@0.8
431.244@13.2 182.987@22.9
196.06@6.8 271.919@22.9 140.994@22.9
336.219@11.2 316.177@6.7 253.114@5.9 181.087@13.3
316.192@6.1 427.162@13.5 340.104@7.1
110.521@22.9
287.202@12.1
327.092@4.3 494.816@22.9
332.244@10.1338.234@11.7 286.219@18.8
230.013@6.5305.217@8.8
249.127@11.5
575.454@22.5
306.169@10.3
225.113@9 219.115@7.5
184.06@2.9
307.199@12.3
307.201@11.4
192.065@8229.05@10.5
-0.06 274.181@12.1286.22@18.3 148.11@5.8 220.101@7.5
209.118@15 346.278@16.8
130.048@10.5269.06@5.4
272.17@10.5 392.292@22.1348.072@6.4
223.099@13.3111.523@22.9 311.072@8.3
328.259@14.8
489.249@13.1 322.843@0.6 158.153@13.1
274.877@0.6
245.08@14.2
393.3@22
139.986@22.9
287.108@9.1 338.818@0.6
162.054@9.6
305.133@9.1132.042@6.5 130.057@9.1 329.185@12.3
312.219@11.4 289.196@12.4
84.047@7.5
251.129@15 340.103@9
338.346@22.6 146.056@8.1
545.138@7.8
595.349@11.1 244.03@8.2
-0.08 291.234@13.1 266.105@9.1 635.379@12.3
250.094@4.2 247.182@12.3
597.364@14.4 274.12@4.6
174.123@1.4
contribute to cancer etiology.

-0.10 340.103@5.7
328.104@8.7
182.118@12.3
308.207@12.3
267.128@7.5
357.051@1.4
-0.12
-0.09 -0.08 -0.07 -0.06 -0.05 -0.04 -0.03 -0.02 -0.01 -0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 0.12 0.13 0.14 0.15
w *c[1]
Identification of
unknown signals ■■ Recent applications of
metabolomics to cancer
epidemiology have shown that
various metabolic pathways
Biomarkers, biological understanding are influenced by cancers,
and some of them are causally
linked to cancer development.
After data acquisition, metabo- tabolites generally belong to specif- ■■ Characterization of these
lite levels are statistically compared ic chemical classes, such as amino
metabolic changes is applied
in various groups of individuals to acids, bile acids, fatty acids, and
to the identification of new
identify metabolites that vary in lipids. In untargeted metabolomics,
biomarkers for early detection
their concentrations in any given thousands of metabolites can be
of cancer and new risk factors
condition. Data are interpreted on detected by MS, and the only limit
the basis of current knowledge of to the number of metabolites mea- for cancer.
factors that can influence concen- sured is the sensitivity of the ana-
trations of these metabolites and lytical instrument. The large volume
the corresponding metabolic path- of information collected makes this compounds that are present at low
ways. Novel hypotheses on mecha- approach ideal for biomarker dis- concentrations. The large number
nisms that lead to diseases can be covery studies [3]. of compounds measured makes
generated, and new biomarkers for However, untargeted metabo- calibration with chemical standards
diagnosis, prognosis, or disease lomics also has some limitations. impossible, and therefore measure-
susceptibility can be discovered. The first is that despite the large ments for any given metabolite are
Two different MS-based me- number of metabolites that can be expressed in study-based relative,
tabolomics approaches are com- measured in a single analytical run, rather than ab so
lute, concentra-
monly used: the targeted and un- no single method is able to compre- tions. This means that specific pro-
targeted approaches. In targeted hensively measure the metabolome. cedures are required to monitor the
metabolomics, a limited number of Combinations of methods are often stability of the response of the mass
metabolites (typically 50–200), de- recommended to maximize analyti- spectrometer over the analysis of
fined a priori, are measured by MS cal coverage. large series of samples (typically a
against calibration curves for each In addition, targeted MS as- few hundred to a few thousand) and
metabolite measured. These me- says may be needed to measure to check the quality of the data.
Chapter 3.7 • Metabolic change and metabolomics 201

A second important limitation of to the identification of 4 steps in the and those with low-grade bladder
untargeted metabolomics is related pathway that enhance glycolysis cancer (ROC AUC, 0.96) [11].
to the identification of the metabo- in the tumour cell and underlie the Less-common biomarkers, which
lites detected. There are about 8000 Warburg effect [7]. are often present at low concentra-
known metabolites in blood, and Metabolomics is also used to tions in blood or urine, may be more
about 1000 of those can be iden- compare metabolic profiles of cells specific for a particular cancer type
tified in untargeted metabolomics treated with various enzyme in- and better predictors of cancer or of
experiments. Many more signals hibitors or drugs. Koningic acid was specific stages of the cancer. Several
are detected but are still unknown, identified as a highly specific inhibitor such markers were identified in un-
because of the lack of reference of glyceraldehyde 3-phosphate de- targeted metabolomics studies using
mass spectra in metabolite data- hydrogenase, a rate-controlling en- high-resolution MS.
bases and of commercial standards zyme in the glycolytic pathway, with A conjugated steroid, 27-nor-5β-
needed for their identification. limited perturbations of other meta- cholestane-3,7,12,24,25 pentol glucu-
bolic pathways [8]. ronide, was significantly upregulated
Initial applications of metabo- in the serum of women with epithelial
Applications of
lomics to cancer epidemiology ovarian cancer in both early-stage
metabolomics to were case–control studies of small and late-stage patients when com-
understanding cancer sample size aimed at the identifi- pared with healthy women or women
development cation of biomarkers for diagnosis, with benign ovarian tumours [12].
Currently, applications of metabo- prognosis, and response to therapy Compared with α-fetoprotein, phe-
lomics to cancer research are quite [9,10]. Results of 106 case–control nylalanyl-tryptophan and glycochol
diverse. Tumour samples have been studies were systematically ana- ate were better able to differentiate
compared with normal tissues to lysed, showing that the cancer dis- individuals with hepatocellular carci-
identify metabolites that vary in their criminants most commonly report- noma from those with cirrhosis, with
concentrations in the two types of ed in blood or urine samples were ROC AUC values greater than 0.89
tissues. Metabolic alterations in tu- various amino acids, nucleotides, [13]. These two metabolites also had
mour samples were investigated in polyamines, sugars, organic acids higher diagnostic performance than
11 studies for 7 cancer types, and a from the tricarboxylic acid cycle, α-fetoprotein for early-stage hepato-
meta-analysis was performed of the bile acids, and closely related me- cellular carcinoma.
results from each individual study [4]. tabolites in their respective meta- In a similar metabolomics study,
Some metabolites were differentially bolic pathways [10]. Many of these several hydroxylated long-chain fatty
abundant in tumour samples and metabolites are affected by differ- acids with anti-inflammatory proper-
normal tissues for multiple cancer ent types of cancer, whereas oth- ties were identified and found to be
types; these included taurine, acyl- ers appear to be more specific for a downregulated in serum samples
carnitine, kynurenine, and lactate, particular cancer type; for example, from colorectal cancer cases in
reflecting common alterations in bilirubin and bile acids are associ- three independent groups of patients
pathways notably related to sugar ated with hepatocellular carcinoma. in Japan and the USA [14]. Two of
metabolism, glutathione metabolism, Most of these metabolites are these fatty acids were good predict
and fatty acid biosynthesis. Similarly, common, universally occurring me- ors of colorectal cancer cases, with
the comparison of metabolic profiles tabolites, which can also be influ- ROC AUC values ranging from 0.85
of 60 primary cancer cell lines from enced by various confounding fac- to 0.93. A later study in the European
9 tumour types showed that several tors, such as other diseases, age, or Prospective Investigation into Cancer
pathways were commonly affected body mass index (see Chapter 2.7). and Nutrition (EPIC) cohort confirmed
in the different cell lines, and that Therefore, there is little likelihood that the low concentrations of these two
glycine was highly correlated with any of these metabolites can be used fatty acids in pre-diagnostic samples
rate of proliferation [5], leading to the on their own as a biomarker for diag- of subjects who developed colorectal
recognition of the oncogenic role of nosis, but they may have applications cancer [15]. The differences in the
glycine decarboxylase. in the context of panels made up of levels of the fatty acids between cas-
Metabolomics and fluxomics several metabolites, proteins, and/or es and controls were seen 3–7 years
have been applied to tumour cell clinical biomarkers. A combination before diagnosis, suggesting possi-
cultures to identify metabolic alter of three serum metabolites differen- ble clinical applications as early bio-
ations and adaptations, which are tiated with high accuracy between markers of disease.
now recognized as a hallmark of individuals with low-grade bladder Applications of metabolomics to
cancer [6]. As an example, the sys- cancer and healthy controls, with prospective epidemiological studies
tematic overexpression of individual a receiver operating characteristic are relatively recent. The earliest
enzymes in the 12 steps linking ex- (ROC) area under the curve (AUC) application of metabolomics within
tracellular glucose to excreted lac- value of 0.99, and between individu- a prospective study involved 189
tate combined with flux analysis led als with high-grade bladder cancer individuals who developed type 2
202
diabetes and 189 matched controls sis with risk of cancer, suggest new betes. This finding confirms results
from the Framingham Offspring pathophysiological mechanisms that from earlier rat feeding studies that
Study [16]. Among 61 metabolites lead to cancer. Such studies face two showed a contribution of branched-
measured at baseline by MS, 5 main challenges. First, few of these chain amino acids to the development
metabolites (leucine, isoleucine, results have yet been replicated in of insulin resistance [25].
valine, tyrosine, and phenylalanine) independent cohorts [3,12]. They will
were associated with risk of type 2 need to be confirmed in future studies,
diabetes [16]. Subsequently, more
Metabolomics and
as has been done for type 2 diabetes.
studies were performed on risk of Second, complementary approaches
biomarkers of exposure
type 2 diabetes, and a recent meta- will be needed to interpret these new to cancer risk factors
analysis of results from eight origi- data. The combination of metabo- Anthropometric, lifestyle, and en-
nal publications showed consistent lomics with other –omics will help vironmental factors all influence
associations of levels of these five to establish the causal implications blood metabolic profiles (Fig. 3.7.2).
amino acids with the risk of devel- of specific metabolites or metabolic Their effects have been described
oping type 2 diabetes [17]. pathways in carcinogenesis, as illus- in an increasing number of inter-
In the past 5 years, 15 prospec- trated by a Mendelian randomization vention studies and observational
tive metabolomics studies on can- analysis on branched-chain amino studies, which aid in the interpre-
cers of the colorectum, liver, pan- acids and type 2 diabetes [24]. This tation of results from prospective
creas, prostate, and breast have analysis showed that genetic variants studies on cancer. Metabolites that
been published, with the number of associated with levels of branched- are simultaneously associated with
case–control pairs varying from 100 chain amino acids were significantly a specific risk factor for cancer
SECTION 3
CHAPTER 3.7
to more than 1000. All of the studies associated with risk of type 2 dia- and with cancer risk are possible
used blood samples that were ana-
lysed by NMR spectroscopy or MS.
In a case–control study on
Fig. 3.7.2. Metabolites associated with body mass index in two independent subcohorts
colorectal cancer nested in the of healthy subjects from the European Prospective Investigation into Cancer and
Prostate, Lung, Colorectal, and Nutrition (EPIC) study. A total of 22 metabolites replicate in both subcohorts (EPIC-Oxford
Ovarian Cancer Screening Trial and EPIC-Hepatobiliary controls). Arrows indicate the direction of the associations:
cohort, 676 metabolites, including positive association (), negative association (), and inconsistent direction (?). PCs,
447 metabolites of known identity, phosphatidylcholines; SM, sphingomyelin.
were measured [18]. The bile acid
glycochenodeoxycholate was asso- acylcarnitines biogenic amines
ciated with risk of colorectal cancer amino acids lysophosphatidylcholines
in women. In a case–control study phosphatidylcholines sphingomyelins
nested in the EPIC cohort, several
metabolites related to amino acid,
lipid, and carbohydrate metabo-
lism were associated with risk of
hepatocellular carcinoma [19,20].
45 22 40
In a prospective study involving EPIC-Hepatobiliary EPIC-Oxford
subjects from four cohorts in the Controls N = 392
USA (453 cases and 898 matched N = 327
controls), 83 metabolites were mea
sured; three branched-chain amino
acids – leucine, isoleucine, and
valine – were associated with risk
of pancreatic cancer, and these
associations were independent of
diabetes development [21]. In the
Alpha-Tocopherol, Beta-Carotene
Cancer Prevention Study and the 1 biogenic amine: Kynurenine();
EPIC study, several metabolites 1 amino acid: Glutamate();
related to energy and lipid metab- 1 sphingomyelin: SM C18:0();
olism were associated with risk of 3 lysoPCs: C18:1(), C18:2(), C28:0(?);
prostate cancer [22,23]. 8 diacyl PCs: C32:2(), C34:4(), C38:3(), C38:4(),
These prospective studies, which C40:6(), C42:0(), C42:1(), C42:2();
show associations of blood metabo- 8 acyl-alkyl PCs: C38:2(), C40:5(), C42:3(), C42:4(),
lites several years before diagno- C42:5(), C44:4(), C44:5(), C44:6()

mediators of the risk. In a nested of hundreds of these compounds, estrogen receptor-positive breast
case–control study that included which together constitute the inter- cancer, enabling the generation of
621 postmenopausal breast can- nal exposome [27,28]. Many of these novel hypotheses on the role of diet
cer cases and 621 matched con- compounds are direct indicators of in breast cancer risk [33].
trols, 4 metabolites (16α-hydroxy- exposure to environmental factors. Metabolomics data, once col-
dehydroepiandrosterone-3-sulfate, They have been measured in pop- lected in a prospective study, can be
3-methylglutarylcarnitine, allo-iso- ulation studies as proxy biomark- further mined to selectively examine
leucine, and 2-methylbutyrylcar- ers of exposures to environmental associations of specific markers of
nitine) were associated with both factors. Detailed information on dietary exposures with cancer risk
body mass index and risk of inva- these biomarkers is curated in the [28]. Among 657 metabolites mea
sive breast cancer. These four me- Exposome-Explorer database [29]. sured in serum samples, trigonelline,
tabolites may point towards meta- Metabolomics approaches have a biomarker of coffee intake, was
bolic pathways that contribute to the potential to play an important found to be inversely associated with
breast carcinogenesis and explain role in the discovery of new bio- risk of colorectal cancer, suggest-
the positive association of body markers of exposure in intervention ing a protective role of coffee intake
mass index with risk of postmeno- studies or cross-sectional studies against colorectal cancer [34].
pausal breast cancer [26]. [30,31]. A classic early example is
This example and those given in the identification of proline betaine Conclusions
the previous section illustrate how as a biomarker for intake of citrus The potential of metabolomics for
metabolomics aids in understand- fruit [32]. Since then, many other di- elucidating mechanisms of carcino-
ing mechanisms that link exposures etary biomarkers have been identi- genesis and for identifying novel risk
to risk factors for cancer. In all of fied. These biomarkers of exposure factors for cancer is established.
these examples, the focus was put can be simultaneously measured in Techniques for metabolomics have
on endogenous metabolites, as exposome-wide association stud- improved considerably over the past
indicators of changes in host me- ies (EWAS) using untargeted or few years, and there is now little
tabolism. Beyond endogenous me- targeted metabolomics approaches doubt that metabolomics is emerging
tabolites, a large array of exogenous to study their association with can- as an essential tool, complementary
compounds, directly derived from cer risk. In a case–control study to the well-established genomics,
the diet, pollutants, and drugs and on breast cancer nested in the transcriptomics, and proteomics ap-
mainly absorbed through the gut Prostate, Lung, Colorectal, and proaches, to identify novel biomark-
mucosa, are found in blood or urine, Ovarian Cancer Screening Trial ers for cancer and to better under-
often at low concentrations. The use cohort, of 113 metabolites related stand cancer etiology.
of sensitive MS techniques ena- to dietary exposures, 19 metabo-
bles the detection in blood or urine lites were associated with risk of
References
1. Wishart DS, Feunang YD, Marcu A, Guo 5. Jain M, Nilsson R, Sharma S, 8. Liberti MV, Dai Z, Wardell SE, Baccile
AC, Liang K, Vázquez-Fresno R, et al. Madhusudhan N, Kitami T, Souza AL, et JA, Liu X, Gao X, et al. (2017). A predic-
(2018). HMDB 4.0: the Human Metabolome al. (2012). Metabolite profiling identifies a tive model for selective targeting of the
Database for 2018. Nucleic Acids Res. key role for glycine in rapid cancer cell pro- Warburg effect through GAPDH inhibi-
46(D1):D608–17. https://doi.org/10.1093/ liferation. Science. 336(6084):1040–4. tion with a natural product. Cell Metab.
nar/gkx1089 PMID:29140435
https://doi.org/10.1126/science.1218595 26(4):648–659.e8. https://doi.org/10.1016/j.
2. Nicholson JK, Wilson ID (2003). Opinion: un- PMID:22628656 cmet.2017.08.017 PMID:28918937
derstanding ‘global’ systems biology: metab
onomics and the continuum of metabolism. 6. Hanahan D, Weinberg RA (2011). Hallmarks 9. Spratlin JL, Serkova NJ, Eckhardt SG
Nat Rev Drug Discov. 2(8):668–76. https:// of cancer: the next generation. Cell. (2009). Clinical applications of me-
doi.org/10.1038/nrd1157 PMID:12904817 144(5):646–74. https://doi.org/10.1016/j. tabolomics in oncology: a review. Clin
cell.2011.02.013 PMID:21376230 Cancer Res. 15(2):431–40. https://doi.
3. Scalbert A, Ferrari P (2020). Biomarker org/10.1158/1078 - 0432.CCR- 08 -1059
discovery. In: Adamski J, editor. 7. Tanner LB, Goglia AG, Wei MH, Sehgal T, PMID:19147747
Metabolomics for biomedical research. Parsons LR, Park JO, et al. (2018). Four
Elsevier. (forthcoming) key steps control glycolytic flux in mam- 10. Liesenfeld DB, Habermann N, Owen RW,
4. Reznik E, Luna A, Aksoy BA, Liu EM, malian cells. Cell Syst. 7(1):49–62.e8. Scalbert A, Ulrich CM (2013). Review of
La K, Ostrovnaya I, et al. (2018). A land- https://doi.org/10.1016/j.cels.2018.06.003 mass spectrometry-based metabolomics
scape of metabolic variation across tu- PMID:29960885 in cancer research. Cancer Epidemiol
mor types. Cell Syst. 6(3):301–313.e3. Biomarkers Prev. 22(12):2182–201. https://
https://doi.org/10.1016/j.cels.2017.12.014 doi.org/10.1158/1055-9965.EPI-13-0584
PMID:29396322 PMID:24096148
204
11. Tan G, Wang H, Yuan J, Qin W, Dong 20. Stepien M, Duarte-Salles T, Fedirko V, 28. Scalbert A, Huybrechts I, Gunter MJ
X, Wu H, et al. (2017). Three serum me- Floegel A, Barupal DK, Rinaldi S, et al. (2019). The food exposome. In: Dagnino
tabolite signatures for diagnosing low- (2016). Alteration of amino acid and bio- S, Macherone A, editors. Unravelling
grade and high-grade bladder cancer. Sci genic amine metabolism in hepatobiliary the exposome: a practical view. Cham,
Rep. 7(1):46176. https://doi.org/10.1038/ cancers: findings from a prospective Switzerland: Springer; pp. 217–45.
srep46176 PMID:28382976 cohort study. Int J Cancer. 138(2):348–
29. Neveu V, Moussy A, Rouaix H, Wedekind
60. https://doi.org/10.1002/ijc.29718
12. Chen J, Zhang X, Cao R, Lu X, Zhao S, R, Pon A, Knox C, et al. (2017). Exposome-
PMID:26238458
Fekete A, et al. (2011). Serum 27-nor-5β- Explorer: a manually-curated database
cholestane-3,7,12,24,25 pentol glucuro- 21. Mayers JR, Wu C, Clish CB, Kraft P, on biomarkers of exposure to dietary and
nide discovered by metabolomics as po- Torrence ME, Fiske BP, et al. (2014). environmental factors. Nucleic Acids Res.
tential diagnostic biomarker for epithelium Elevation of circulating branched-chain 45(D1):D979–84. https://doi.org/10.1093/
ovarian cancer. J Proteome Res. 10(5): amino acids is an early event in human nar/gkw980 PMID:27924041
2625–32. https://doi.org/10.1021/pr200173q pancreatic adenocarcinoma develop-
30. Scalbert A, Rothwell JA, Keski-Rahkonen
PMID:21456628 ment. Nat Med. 20(10):1193–8. https://doi.
P, Neveu V (2017). The food metabolome
org/10.1038/nm.3686 PMID:25261994
13. Luo P, Yin P, Hua R, Tan Y, Li Z, Qiu G, and dietary biomarkers. In: Schoeller
et al. (2018). A large-scale, multicenter 22. Mondul AM, Moore SC, Weinstein SJ, DA, Westerterp-Plantenga M, editors.
serum metabolite biomarker identifica- Karoly ED, Sampson JN, Albanes D Advances in the assessment of di-
tion study for the early detection of hepa- (2015). Metabolomic analysis of prostate etary intake. Boca Raton (FL), USA:
tocellular carcinoma. Hepatology. 67(2): cancer risk in a prospective cohort: the CRC Press; pp. 259–82. https://doi.
662–75. https://doi.org/10.1002/hep.29561 Alpha-Tocopherol, Beta-Carotene Cancer org/10.1201/9781315152288-16
PMID:28960374 Prevention (ATBC) Study. Int J Cancer. 31. Vlaanderen JJ, Janssen NA, Hoek G,
137(9):2124–32. https://doi.org/10.1002/ Keski-Rahkonen P, Barupal DK, Cassee
14. Ritchie SA, Ahiahonu PWK, Jayasinghe D,
ijc.29576 PMID:25904191 FR, et al. (2017). The impact of ambient
Heath D, Liu J, Lu Y, et al. (2010). Reduced
levels of hydroxylated, polyunsaturated ul- 23. Schmidt JA, Fensom GK, Rinaldi S, air pollution on the human blood metabo-
tra long-chain fatty acids in the serum of Scalbert A, Appleby PN, Achaintre D, et lome. Environ Res. 156:341–8. https://
doi.org/10.1016/j.envres. 2017.03.0 42
SECTION 3
CHAPTER 3.7
colorectal cancer patients: implications for al. (2017). Pre-diagnostic metabolite con-
early screening and detection. BMC Med. PMID:28391173
centrations and prostate cancer risk in
8(1):13. https://doi.org/10.1186/1741-7015- 1077 cases and 1077 matched controls 32. Heinzmann SS, Brown IJ, Chan Q,
8-13 PMID:20156336 in the European Prospective Investigation Bictash M, Dumas M-E, Kochhar S, et
15. Perttula K, Edmands WMB, Grigoryan H, into Cancer and Nutrition. BMC Med. al. (2010). Metabolic profiling strategy
Cai X, Iavarone AT, Gunter MJ, et al. (2016). 15(1):122. https://doi.org/10.1186/s12916- for discovery of nutritional biomarkers:
Evaluating ultra-long-chain fatty acids as 017-0885-6 PMID:28676103 proline betaine as a marker of citrus con-
biomarkers of colorectal cancer risk. Cancer sumption. Am J Clin Nutr. 92(2):436–43.
24. Lotta LA, Scott RA, Sharp SJ, Burgess
Epidemiol Biomarkers Prev. 25(8):1216–23. https://doi.org/10.3945/ajcn.2010.29672
S, Luan J, Tillin T, et al. (2016). Genetic
https://doi.org/10.1158/1055-9965.EPI-16- PMID:20573794
predisposition to an impaired metabo-
0204 PMID:27257090 lism of the branched-chain amino acids 33. Playdon MC, Ziegler RG, Sampson JN,
16. Wang TJ, Larson MG, Vasan RS, Cheng and risk of type 2 diabetes: a Mendelian Stolzenberg-Solomon R, Thompson
S, Rhee EP, McCabe E, et al. (2011). randomisation analysis. PLoS Med. HJ, Irwin ML, et al. (2017). Nutritional
Metabolite profiles and the risk of developing 13(11):e1002179. https://doi.org/10.1371/ metabolomics and breast cancer risk
diabetes. Nat Med. 17(4):448–53. https://doi. journal.pmed.1002179 PMID:27898682 in a prospective study. Am J Clin Nutr.
org/10.1038/nm.2307 PMID:21423183 106(2):637–49. https://doi.org/10.3945/
25. Newgard CB, An J, Bain JR, Muehlbauer ajcn.116.150912 PMID:28659298
17. Guasch-Ferré M, Hruby A, Toledo E, MJ, Stevens RD, Lien LF, et al. (2009). A
Clish CB, Martínez-González MA, Salas- branched-chain amino acid-related meta- 34. Guertin KA, Loftfield E, Boca SM,
Salvadó J, et al. (2016). Metabolomics in bolic signature that differentiates obese Sampson JN, Moore SC, Xiao Q, et al.
prediabetes and diabetes: a systematic and lean humans and contributes to in- (2015). Serum biomarkers of habitual
review and meta-analysis. Diabetes Care. sulin resistance. Cell Metab. 9(4):311–26. coffee consumption may provide insight
39(5):833– 46. https://doi.org/10.2337/ https://doi.org/10.1016/j.cmet.2009.02.002 into the mechanism underlying the as-
dc15-2251 PMID:27208380 PMID:19356713 sociation between coffee consumption
and colorectal cancer. Am J Clin Nutr.
18. Cross AJ, Moore SC, Boca S, Huang 26. Moore SC, Playdon MC, Sampson JN, 101(5):1000–11. https://doi.org/10.3945/
W-Y, Xiong X, Stolzenberg-Solomon R, et Hoover RN, Trabert B, Matthews CE, et ajcn.114.096099 PMID:25762808
al. (2014). A prospective study of serum al. (2018). A metabolomics analysis of
metabolites and colorectal cancer risk. body mass index and postmenopausal
Cancer. 120(19):3049–57. https://doi. breast cancer risk. J Natl Cancer Inst.
org/10.1002/cncr.28799 PMID:24894841 110(6):588–97. https://doi.org/10.1093/jnci/
djx244 PMID:29325144
19. Fages A, Duarte-Salles T, Stepien M,
Ferrari P, Fedirko V, Pontoizeau C, et al. 27. Rappaport SM, Barupal DK, Wishart D,
(2015). Metabolomic profiles of hepatocel- Vineis P, Scalbert A (2014). The blood ex-
lular carcinoma in a European prospec- posome and its role in discovering caus-
tive cohort. BMC Med. 13(1):242. https:// es of disease. Environ Health Perspect.
d o i .o r g /10 .118 6 /s12 916 - 015 - 0 4 6 2- 9 122(8):769–74. https://doi.org/10.1289/ehp.
PMID:26399231 1308015 PMID:24659601

3.8
Epigenetics
Potential in diagnostics, therapy,
and prevention
Toshikazu Ushijima James Flanagan (reviewer)
Zdenko Herceg Igor Pogribny (reviewer)
●● Suppressing the induction of epi- RNAs. All of these mechanisms are

SUMMARY genetic changes and reversing known to be critical for high-fidelity
induced epigenetic changes have propagation of gene activity states
●● DNA methylation, histone modi- potential for cancer prevention. in a cell type-specific manner. In ad-
fications, and non-coding RNAs, dition, some investigators include
the three main epigenetic mech- nucleosome positioning and forma-
anisms, are all known to be criti- In recent years, epigenetics has tion of higher-order chromatin struc-
cal for high-fidelity propagation been consolidated as a mainstream ture as epigenetic mechanisms.
of gene activity states in a cell field of cancer research, funda- Consistent with the importance
type-specific manner. mental to the understanding of of epigenetic mechanisms in critical
the etiology and biology of cancer. cellular processes, dysregulation of
●● Many cancer risk factors, includ-
The importance of epigenetic dys- epigenetic mechanisms has been
ing ageing, inflammation, tobacco
regulation in cancer initiation and linked to various noncommunicable
smoking, alcohol consumption,
progression has been highlighted diseases in humans, most notably
fungal toxins, biological agents, cancer [1,2]. Almost all critical pro-
at multiple levels, and many con-
and diet as well as air and water cesses in cancer cells – such as
ceptual breakthroughs in the field
pollution and certain endocrine self-sufficiency in growth signals,
have revolutionized the traditional
disrupters, are associated with concepts of cancer development. insensitivity to anti-growth signals,
epigenome dysregulation. In addition, the emergence of pow- tissue invasion and metastasis, lim-
●● Epigenetic changes, especially erful technologies that enable the itless replicative potential, sustained
DNA methylation, are use- detection of epigenetic changes in angiogenesis, and evasion of apo-
ful as biomarkers for cancer. high-throughput and genome-wide ptosis – can be caused not only by
settings has dramatically acceler- genetic changes but also by epige-
Methylation changes can be
ated cancer research and opened netic alterations (Fig. 3.8.1) [3].
detected in a large number of
up new perspectives. This has re- It has been proposed that the
cells in normal-appearing tis-
sulted in a broader appreciation of epigenome may function as an inter-
sues, and such change has
the importance of epigenetics in the face between environmental factors
been correlated with risk of
etiology of human cancer. and the genome; however, the epi-
cancer development for major genetic mechanisms by which risk
In the past, the term “epige-
cancer types in humans. factors induce dysregulation of the
netics” was used to describe all
●● Epigenetic change can be re- biological phenomena that do not epigenome and the functional im-
versed by drugs, and the rel- follow normal genetic principles. pact of this dysregulation in specific
evant agents have expanded Nowadays, epigenetics refers to the human cancers remain poorly un-
from those affecting DNA meth- study of all changes in gene expres- derstood [4]. The challenges posed
ylation and histone acetylation sion that are transmitted across cell by numerous efforts to sequence
to now include histone methyla- generations and that do not involve human cancers are to identify the
tion modifications. changes in the DNA sequence (i.e. epigenome changes and conse-
mutations). In this chapter, three quently dysregulated genes and
●● Epigenetic changes in normal main epigenetic mechanisms are pathways that precede and promote
cells and cancer cells can be described: DNA methylation, his- tumour development, and to distin-
used as diagnostic targets. tone modifications, and non-coding guish functionally important events
206
(“drivers”) from events that are ubiquitination at specific residues

merely “passengers”. Accordingly, of histone proteins, mostly in the FUNDAMENTALS
epigenetics may have potential in N-terminal “tails” of histone proteins.
the prevention, early detection, and Histone modifications regulate mul- ■■ Epigenetics refers to all
treatment of cancer. tiple cellular processes, including mitotically heritable changes
gene transcription, DNA repair, and in gene expression and
DNA replication [7]. associated phenotypic traits
Epigenetic mechanisms
Histone acetylation is regulated that are not coded in the
The three main epigenetic mecha- by histone acetyltransferases and DNA sequence itself. These
nisms described here – DNA meth- histone deacetylases (HDACs), and changes are mediated by
ylation, histone modifications, and HDACs consist of 11 different mol- DNA methylation, histone
non-coding RNAs – have been stud- ecules in classes I, IIa, IIb, and IV modifications, and non-coding
ied primarily in the context of regula- and sirtuins in class III. Histone RNAs.
tion of gene expression. In addition methylation at specific amino acid
to this well-established context, they ■■ Epigenetic changes can be
residues is regulated by histone
are now recognized as important for induced by environmental and
methyltransferases, such as EZH2,
other chromatin-based processes, nutritional factors, and they are
MLL, SETD2, and DOT1L, and
such as DNA repair, DNA replica- involved in a variety of human
by histone demethylases, such
tion, and formation of higher-order cancer types and in other
as KDM1A (LSD1) and KDM4A
chromatin structure [5]. chronic disorders.
(JMJD2A). Multiple histone modi-
fication enzymes are mutated or ■■ Growing evidence suggests
DNA methylation that epigenetic changes
CHAPTER 3.8
SECTION 3
dysregulated in human neoplasms
Of the three main epigenetic mech- [7]. Therefore, the importance of may be risk factor-specific
anisms, the best studied is DNA histone modifications in cancer and (“signatures”), which may
methylation. The methylation of other diseases is now recognized. prove instrumental in the
DNA refers to the covalent addition discovery of novel biomarkers
of a methyl group to the 5-carbon Non-coding RNAs of cancer.
position of cytosine in a CpG dinu- Non-coding RNAs consist of small ■■ Recent advances in epi
cleotide. DNA methylation, via the RNAs – microRNAs, Piwi-interacting genetics and epigenomics
function of maintenance DNA meth- RNAs (piRNAs), and small nucleolar present an exciting opportunity
yltransferase (mainly DNMT1), has RNAs (snoRNAs) – and long non- to incorporate epigenetic data
long been considered a highly stable coding RNAs, and many investi- into carcinogen identification
epigenetic modification. However, gators consider them as the third and safety assessment.
recent studies showed that the ten– class of epigenetic mechanisms [8].
eleven translocation (TET) family of MicroRNAs can regulate expression ■■ Epigenetic changes in cancer
proteins are involved in active DNA levels of messenger RNA, and piR- cells are now targets for
demethylation, and that DNA meth- NAs are important to suppress the cancer therapy, and correction
ylation can be dynamically regulated transcription of retrotransposons. of epigenetic changes can
at specific stages of life, such as Long non-coding RNAs, defined as form the basis for a cancer
during early embryogenesis. The endogenous cellular RNAs longer prevention strategy.
TET proteins hydrolyse methyl cy- than 200 base pairs, tend to be ex-
tosines, either fully methylated or pressed at lower levels compared
hemi-methylated, and produce 5-hy- with the majority of protein-coding tween these three epigenetic mech-
droxymethylcytosine and its further genes. Interest in long non-coding anisms in setting up and maintain-
metabolites, which will eventually be RNAs has been stimulated by the ing the genome-wide expression
removed by base excision repair [6]. recent finding that almost the entire programme in a tissue-specific and
mammalian genome is transcribed, lineage-specific manner.
Histone modifications although only a small fraction (~2%)
The second main epigenetic mecha- of the genome is established to en-
nism encompasses various modifi- code proteins [9]. A variety of human
Epigenomic changes in
cations of histone proteins. Typically, malignancies were found to exhibit
cancer
two copies each of the histones aberrant expression of long non- Consistent with the critical role of
H2A, H2B, H3, and H4 compose an coding RNAs, some of which were epigenetic mechanisms in the con-
octamer, which is wrapped by an ap- demonstrated to be involved in can- trol of cellular processes, a plethora
proximately 147-base-pair stretch of cer onset and progression [10]. of studies have revealed that the
DNA to form a nucleosome. Histone Experimental evidence sug- epigenome is markedly dysregulat-
modifications include acetylation, gests that there is intimate and ed in almost all malignancies [1,2]
methylation, phosphorylation, and mutually reinforcing cross-talk be- (Fig. 3.8.1).
Chapter 3.8 • Epigenetics 207

Fig. 3.8.1. Interplay between genetics and epigenetics in cancer development. Epigenetic mechanisms regulate key cellular processes
(such as gene transcription, DNA repair, and differentiation) and play critical roles in cellular responses to environmental exposures
and endogenous stimuli. Dysregulation of epigenetic mechanisms may promote the development of abnormal phenotypes and cancer.
There is cross-talk between epigenetic and genetic changes in the process of cancer development and progression. Given that
epigenetic and genetic changes coexist in all cancers, it is important to identify the functionally important changes (“drivers”) that are
pertinent to carcinogenesis, and to distinguish them from events that are not functionally important (“passengers”).
DNA methylation come unable to bind the factors that cellular processes and cell identity,
Traditionally, two forms of aberrant are responsible for gene expression dysregulation of histone modifica-
DNA methylation have been de- [12], and the gene is not transcribed. tion patterns has a global impact
scribed in human cancer: the overall A large number of studies have in- on regulation of gene expression
loss of 5-methylcytosine (global hy- dicated that the silencing of tumour across the genome. This notion is
pomethylation) and gene promoter- suppressor genes and other cancer- supported by recent studies show-
associated (CpG island-specific) related genes may occur through ing that recurrent mutations in the
hypermethylation [11]. Genome-wide hypermethylation of their promoters. genes encoding histone modifiers
hypomethylation can induce chro- and remodellers were associated
mosomal instability and hypometh-
Histone modifications with widespread transcriptome and
ylation of cancer/testis antigen Recent genetic and molecular stud- epigenome changes in many can-
genes. The impact of genome-wide ies have directly implicated histone cer types [13,14]. It has also been
or gene-specific hypomethylation modifications and histone-modi- observed that cancer cells exhibit
on the activation of cellular proto- fying and histone-remodelling en- dysregulated occupancy of the his-
oncogenes is still debated, but hy- zymes in human cancer. Consistent tone modifications H3K27ac (at en-
permethylation of gene promoters with the critical role of histone hancers) and H3K27me (at promot-
is well established to be associated modifications in the establishment ers), revealing distinct mechanisms
with gene inactivation. When hy- and maintenance of gene expres- underlying transcriptional dysregu-
permethylated, gene promoters be- sion programmes that underpin key lation in cancer [15].
208
Current and future studies aimed malignancies [14,18]. These findings toxin B1 (see Chapter 2.8) [21], and
at characterizing the functional im- should prove pivotal in facilitating inorganic arsenic and heavy metals
pact of dysregulation of chromatin functional studies, aimed at a better (see Chapter 2.9) [22], may leave
modifiers should provide valuable mechanistic understanding of tumour epigenetic signatures in the fetus
mechanistic insights into tumorigen- development (see Chapter 3.2) and that may be detected in neonatal
esis and reveal potential molecular of the plasticity of cancer cells that samples. These observations not
targets for biomarker discovery and underlies tumour resilience and ther- only suggest potential mechanisms
therapeutic intervention. A growing apy failure. of cancer development involving
emphasis in drug discovery on small epigenome dysregulation but also
molecules targeting HDACs, histone underscore that early life may rep-
Environmental influences
acetyltransferases, or histone meth- resent a critical period for interven-
yltransferases (“epigenetic drugs”)
on epigenomes
tion and cancer prevention.
may result in novel strategies for ef- A profound dysregulation of the Although the importance of the
ficient treatment and overcoming re- epigenome is a universal feature environment in the development of
sistance to therapies. across almost all cancer types, a wide variety of cancer types is
and increasing evidence points well supported by both epidemio-
Non-coding RNAs to an important role of epigenetic logical and laboratory-based stud-
Many recent studies also provided mechanisms in mediating gene– ies, the mechanisms by which en-
evidence that the dysregulation of environment interactions and their vironmental exposures dysregulate
non-coding RNAs is involved in the effect throughout the tumorigen- the epigenome remain poorly char-
development of human neoplasia esis process (see Chapter 3.3) [4]. acterized [4,23]. The recent estab-
Remarkable progress in the field of
CHAPTER 3.8
SECTION 3
[1,2]. Although epigenetic changes lishment of reference epigenomes
have been implicated in differ- epigenetics, in conjunction with the for normal cell types and cancer-
ent stages of tumour development emergence of powerful epigenomic specific epigenomes provided by
and progression, the challenge is technologies and computational several major international projects
to identify functionally important tools, has led to the establishment should facilitate the identification
epigenetic changes, which may be of the impact of different endog- of environmental factors that are
referred to as “epigenetic drivers” enous and external risk factors on associated with epigenomic al-
(“epidrivers”) in the same way that the epigenome. A wide range of es- terations. Ultimately, intervention
this term is used for mutations, and tablished and suspected cancer risk studies in animals or humans are
hence differentiate them from “pas- factors (including ageing, inflamma- important to establish causal asso-
senger” events, which are evident tion, tobacco smoking, alcohol con- ciations between environmental ex-
but not functionally important. sumption, fungal toxins, biological posures and epigenetic alterations.
One of the most remarkable and agents, and diet) as well as some
groundbreaking findings of the in- less widely studied exposures and
ternational high-resolution cancer lifestyle factors (such as air and wa- Epigenetic changes as
genome sequencing efforts, spear- ter pollution and certain endocrine biomarkers
headed by the Cancer Genome Atlas disrupters) have been shown to be Epigenetic changes, especially
(TCGA) and the International Cancer associated with epigenome dysreg- DNA methylation, are useful as
Genome Consortium (ICGC), is the ulation (Fig. 3.8.2). cancer biomarkers in multiple ways
high frequency of mutational and In addition to the type of environ- (Fig. 3.8.3). The accumulation lev-
non-mutational (expression) chang- mental exposure, the timing of ex- els of aberrant DNA methylation in
es in the genes encoding proteins posure may also play a critical role normal tissues can be correlated
that directly regulate the epigenome in influencing cancer risk. In utero with future cancer risk, and can be
in malignancies [14,16–18]. About and early life may represent particu- used for cancer risk diagnosis (see
half of all newly identified genes that larly vulnerable periods in humans, Chapter 6.7) [24,25]. Initially, the
are found to be recurrently mutated in because of the profound reconfigu- accumulation of DNA methylation
cancer encode proteins that are part ration of the epigenome during em- changes in normal-appearing tis-
of epigenetic machineries involved bryonic development. Epigenetic sues of cancer patients was shown
in DNA methylation and chromatin changes can be stably propagated for multiple cancer types. Unlike mu-
modifications [17,19]. Furthermore, it over many cell generations, and tations, methylation changes can be
is now evident that frequent dysregu- therefore epigenome dysregula- detected in a large number of cells in
lation of these epigenetic players may tion brought about by early-life ex- normal-appearing tissues, and can
be mediated not only through muta- posures may have lifelong health be readily measured [26]. The accu-
tional events but also through epige- outcomes. Accumulating evidence mulation can be associated with past
netic events; this suggests a potential suggests that in utero exposure to exposure to carcinogenic stimuli,
mechanism for epigenetic changes different agents, including tobacco and the genes that are methylated
that are rampant in almost all human smoke (see Chapter 2.1) [20], afla- can be specific to the exposure [27].

Fig. 3.8.2. Epigenetic mechanisms and cancer: an interface between the environment and the genome. Exposures arising from
external sources (such as environmental chemicals, air pollution, infectious agents, diet, tobacco smoking, alcohol consumption,
and endocrine disrupters) and internal processes (such as metabolism, hormones, inflammation, gut microflora, and ageing) may
induce stable and potentially reversible changes in the epigenome. The patterns (“signatures”) and persistence of these alterations
depend on multiple factors, including the type of epigenetic changes (some genomic regions remain methylated for longer periods
than others), the duration and dosage of the exposure (longer and more intense exposures could minimize the reversibility of DNA
methylation), the tissue type, and the developmental stage (in utero life or puberty may be particularly sensitive periods for some
exposures). Thus, epigenetic mechanisms may represent “sensors” of exposure and “mediators” of the outcomes, including cancer
development. Epigenomic alterations should prove instrumental in the discovery of new biomarkers for risk stratification and early
detection, and attractive targets for novel therapies and preventive strategies.
Exposures Windows of Cell responses Consequences

vulnerability
External
- Smoking
Reverse causality
In utero life
- Air pollution
Reversibility
- Nutrition
- Heavy metals Epigenome
- Biological adaptation
agents
- Endocrine
disrupters
etc.
Non-coding
RNAs
Birth
Dysregulated
cellular
methylation
Epigenome processes
DNA
dysregulation CANCER
Transcription
Childhood
modifications
DNA repair
Histone
Differentiation
Recombination
Endogenous “Normal”
- Hormones epigenome
- Microbiome
- Metabolome
- Inflammation Causality
Adulthood
- Ageing - Longitudinal studies

etc. Cell death - In vitro and animal models
- Twin cohort studies
- Optimized statistical approaches
Epigenetic - Transcription-dependent vs -independent
inheritance - Epidrivers
- Carcinogen evaluation
The accumulation levels of aberrant methylation can be sensitively de- mixed results [31,32]. In addition,
DNA methylation can be correlated tected by technologies based on distinct DNA methylation patterns
with risk of cancer development for polymerase chain reaction (PCR) according to cancer types have
gastric cancer, liver cancer, cervi- amplification of methylated DNA been established, and the specific
cal cancer, and other cancer types molecules, the detection of can- patterns were used to predict the
[24,25]. The usefulness in cancer cer cell-derived DNA has been at- origin of cancers, with a very prom-
risk diagnosis has been shown by a tempted for decades. As a result, ising result [33].
prospective clinical study for gastric there are many cancer detection Even in a specific cancer type,
cancer and cervical cancer [28,29]. systems using materials that are methylation of specific genes or
Similar approaches appear to be likely to contain cancer cells or methylation profiles can be asso-
promising for multiple cancer types cancer cell-derived DNA, such as ciated with the pathophysiology of
in which aberrant DNA methylation stool, urine, sputum, and cervical cancers, and may be useful to de-
is deeply involved. smear, and some of them are al- termine patient prognosis and re-
Cancer cell-specific DNA meth- ready commercially available [30]. sponsiveness to a particular therapy
ylation can be used as a biomarker In contrast, the attempts at using [32]. In sharp distinction to patterns
to detect cancer. Because DNA serum or plasma DNA have had of gene expression, DNA methylation
210
Fig. 3.8.3. DNA methylation as a biomarker. Normal tissues without exposure have minimal levels of aberrant methylation. Exposure
to environmental factors, such as Helicobacter pylori infection and tobacco smoke, is known to induce aberrant DNA methylation
of specific genes in normal-appearing tissue. Further accumulation of aberrant DNA methylation and genetic alterations will lead to
the development of cancer, and each cancer has individual epigenetic profiles that can be associated with pathophysiology, such
as aggressiveness, response to therapy, and prognosis. Measurement of methylation in normal-appearing tissues can be used as
a biomarker of cancer risk. Cancer cell-specific DNA methylation can be used as a biomarker for cancer detection. Methylation of
specific genes that are associated with pathophysiology can be used as a biomarker of pathophysiology.
Normal tissue
Gene A
B
C
D Unmethylated gene
Methylated gene
Exposures
(external and endogenous)
Normal-appearing tissue
Cancer risk biomarker
CHAPTER 3.8
SECTION 3
Very early cancer
Cancer #2
Cancer detection biomarker
Cancer #1
Pathophysiology biomarker
can indicate that a particular gene patient prognosis in several can- tinguishes such change from mu-
cannot be expressed even if its ex- cer types, including colorectal and tation, is the fact that it can be re-
pression is induced in the future. gastric malignancies as well as versed by drugs [1,19,36]. During
For example, if the promoter region neuroblastomas. Specifically, the the past decade, this field has rap-
of O 6 -methylguanine-DNA methyl- CpG island methylator phenotype idly expanded from agents affecting
transferase (MGMT) is determined in neuroblastoma provides prog- DNA methylation and histone acet-
to be methylated at biopsy of a nostic information that is more pre- ylation to now include histone meth-
brain tumour, this gene will never cise than that from the amplifica- ylation modifications (Fig. 3.8.4).
be expressed even after future che tion of the MYCN oncogene, one of DNA methylation can be reversed
mo therapy involving an alkylating the clearest prognostic indicators by DNA demethylating agents.
agent. In the absence of MGMT ex- in clinical oncology [35]. Two such drugs, azacitidine and
pression, such chemotherapy has decitabine, have been approved by
been shown to be effective [34]. the United States Food and Drug
DNA methylation of multiple Epigenetic therapy Administration and other regulatory
genes – the CpG island methyla- One of the most important aspects agencies for treating myelodysplas-
tor phenotype – is associated with of epigenetic change, which dis- tic syndrome and acute myeloid

Fig. 3.8.4. Epigenetic targets for drugs. DNA methylation and histone modifications are written, erased, and read by specific
proteins. Many of the writers, erasers, and readers are now used as drug targets. Among these, DNA methyltransferase (DNMT)
inhibitors and histone deacetylase (HDAC) inhibitors (targets shaded in blue) are already approved by the United States Food and
Drug Administration. Only major drug targets are shown here; novel targets are still being identified. BRDs, bromodomains; HATs,
histone acetyltransferases; TETs, ten–eleven translocation proteins; MBDs, methyl-CpG-binding domains.
M M
CG CG
GC GC
M M
DNA methylation Histone modifications
Acetylation H3K4me H3K9me H3K27me H3K79me
Writer DNMT1 HATs G9a EZH2 DOT1L
Eraser TETs HDACs LSD1
Reader MBDs BRDs
leukaemia, and are now being ex- have pleiotropic effects on cancer ered as drug targets, and their spe-
plored for treating solid tumours. In cell phenotypes. In addition, some cific inhibitors have been developed.
addition to these two drugs, multi- HDAC inhibitors induce acetylation Some histone demethylases are
ple new DNA demethylating agents, of non-histone proteins, including also targets for therapy. Currently,
such as SGI-110 and CC-486, are p53, signal transducer and activator the most successful target is LSD1,
being developed. All these drugs of transcription 1/3 (STAT1/3), and which demethylates di- and mono-
are incorporated into DNA and co- heat shock protein 90 (Hsp90). methylated H3K4. Inhibition of LSD1
valently bind to DNMT1, which ulti- In contrast to HDAC inhibitors, induces differentiation of leukaemia
mately leads to its degradation. As overactivity of oncogenes and other cells and apoptosis of brain tumour
a result, cell replication in the ab- genes due to the formation of ex- cells by activating enhancers and
sence of maintenance methylation tensively histone-acetylated en- promoters of related genes.
leads to DNA demethylation. DNA hancers (super-enhancers) can be
demethylation leads to the activa- targeted by inhibitors of proteins
tion of aberrantly silenced tumour that bind to acetylated histones,
Epigenetic cancer
suppressor genes and an increased namely bromodomain and extrater-
prevention
immune response. To achieve this minal domain (BET) proteins [37]. Suppressing the induction of epi-
mode of action, low-dose and long- Multiple BET inhibitors are being genetic changes and reversing
term administration are seen to be developed against haematological induced epigenetic changes are
important [19]. malignancies and brain tumours. also useful for cancer prevention
Histone deacetylation can be Mutations of histone methyl- [38]. As a proof of concept, in ex-
reversed by HDAC inhibitors [1,19, transferases and histone demethyl- perimental animals, tumours such
36]. Three such drugs have been ases have also provided novel ther- as those of the colon, prostate, and
approved for treating cutaneous apeutic targets [1,19,36]. Especially stomach have been suppressed
lymphoma, and one for treating the H3K27 methyltransferase EZH2 by repression of DNA methyltrans-
multiple myeloma, and many new is mutationally activated in some ferases by gene engineering and
HDAC inhibitors are being devel- tumour types, such as lymphomas, DNA demethylating agents [39–41].
oped. Individual HDAC inhibitors and is overexpressed in many tu- However, it must be recognized that
have different specificities to the mour types. Multiple EZH2 inhibitors DNA methylation is physiologically
individual molecules of HDAC1– are being developed. In addition, essential to repress transposons
HDAC11 in classes I, IIa, IIb, and IV. inhibitors of the H3K79 methyl- and some genes, and nonspe-
All the HDAC inhibitors induce ex- transferase DOT1L and the H3K9 cific demethylation is expected to
pression of many genes, and thus methyltransferase G9a are consid- lead to long-term adverse effects.
212
Therefore, to enable epigenetic appears to be more practical. Also, to benefit from effective chemopre-
cancer prevention by reversing it is now possible to identify individ- vention by balancing the benefit and
epigenetic changes in the human uals at extremely high risk of some the potential adverse effects (see
population, the specificity of precancers by assessing accumulated Chapter 6.4). Because epigenetic
ventive agents for genes with aber- levels of aberrant DNA methylation cancer prevention has great poten-
rant epigenetic modifications must in normal-appearing tissues, as pre- tial, multiple relevant studies are re-
be improved. Instead, suppressing viously discussed. These individuals quired in a timely manner.
the induction of epigenetic changes represent a population that is likely
References
1. Dawson MA (2017). The cancer epi ge
- 10. Lin C, Yang L (2018). Long noncod- 18. Plass C, Pfister SM, Lindroth AM,
nome: concepts, challenges, and thera- ing RNA in cancer: wiring signaling cir- Bogatyrova O, Claus R, Lichter P (2013).
peutic opportunities. Science. 355(6330): cuitry. Trends Cell Biol. 28(4):287–301. Mutations in regulators of the epigenome
1147–52. https://doi.org/10.1126/science. https://doi.org/10.1016/j.tcb.2017.11.008 and their connections to global chroma-
aam7304 PMID:28302822 PMID:29274663 tin patterns in cancer. Nat Rev Genet.
14(11):765–80. https://doi.org/10.1038/nrg
CHAPTER 3.8
SECTION 3
2. Feinberg AP, Koldobskiy MA, Göndör A
11. Sinčić N, Herceg Z (2011). DNA meth- 3554 PMID:24105274
(2016). Epigenetic modulators, modifiers
ylation and cancer: ghosts and angels
and mediators in cancer aetiology and 19. Jones PA, Issa JP, Baylin S (2016).
progression. Nat Rev Genet. 17(5):284– above the genes. Curr Opin Oncol.
23(1):69–76. https://doi.org/10.1097/CCO. Targeting the cancer epigenome for
99. https://doi.org/10.1038/nrg.2016.13
0b013e3283412eb4 PMID:21119515 therapy. Nat Rev Genet. 17(10):630–41.
PMID:26972587
h t t p s: //d o i . o r g /10 .10 3 8 / n r g . 2 016 . 9 3
3. Herceg Z, Lambert MP, van Veldhoven 12. Vaissière T, Sawan C, Herceg Z (2008). PMID:27629931
K, Demetriou C, Vineis P, Smith MT, et Epigenetic interplay between histone mod-
al. (2013). Towards incorporating epige- ifications and DNA methylation in gene 20. Joubert BR, Felix JF, Yousefi P, Bakulski
netic mechanisms into carcinogen identi- KM, Just AC, Breton C, et al. (2016). DNA
silencing. Mutat Res. 659(1–2):40–8.
fication and evaluation. Carcinogenesis. methylation in newborns and maternal
https://doi.org/10.1016/j.mrrev.2008.02.
34(9):1955–67. https://doi.org/10.1093/ smoking in pregnancy: genome-wide con-
004 PMID:18407786
carcin/bgt212 PMID:23749751 sortium meta-analysis. Am J Hum Genet.
13. Plass C, Pfister SM, Lindroth AM, 98(4):680–96. https://doi.org/10.1016/j.
4. Herceg Z, Ghantous A, Wild CP, Sklias A, ajhg.2016.02.019 PMID:27040690
Bogatyrova O, Claus R, Lichter P (2013).
Casati L, Duthie SJ, et al. (2018). Roadmap
Mutations in regulators of the epigenome
for investigating epigenome deregula- 21. Hernandez-Vargas H, Castelino J, Silver
tion and environmental origins of cancer. and their connections to global chromatin MJ, Dominguez-Salas P, Cros M-P,
Int J Cancer. 142(5):874–82. https://doi. patterns in cancer. Nat Rev Genet. 14(11): Durand G, et al. (2015). Exposure to
org/10.1002/ijc.31014 PMID:28836271 765–80. https://doi.org/10.1038/nrg3554 aflatoxin B1 in utero is associated with
PMID:24105274 DNA methylation in white blood cells of
5. Dabin J, Fortuny A, Polo SE (2016).
14. Timp W, Feinberg AP (2013). Cancer as infants in The Gambia. Int J Epidemiol.
Epigenome maintenance in response to
DNA damage. Mol Cell. 62(5):712–27. a dysregulated epigenome allowing cel- 44(4):1238–48. https://doi.org/10.1093/ije/
https://doi.org/10.1016/j.molcel.2016.04. dyv027 PMID:25855716
lular growth advantage at the expense
006 PMID:27259203 of the host. Nat Rev Cancer. 13(7): 22. Green BB, Karagas MR, Punshon T,
497–510. https://doi.org/10.1038/nrc3486 Jackson BP, Robbins DJ, Houseman EA,
6. Schübeler D (2015). Function and informa-
tion content of DNA methylation. Nature. PMID:23760024 et al. (2016). Epigenome-wide assessment
517(7534):321–6. https://doi.org/10.1038/ of DNA methylation in the placenta and
15. Hnisz D, Abraham BJ, Lee TI, Lau A,
nature14192 PMID:25592537 arsenic exposure in the New Hampshire
Saint-André V, Sigova AA, et al. (2013).
Birth Cohort Study (USA). Environ Health
7. Greer EL, Shi Y (2012). Histone methyla- Super-enhancers in the control of cell
Perspect. 124(8):1253–60. https://doi.
tion: a dynamic mark in health, disease identity and disease. Cell. 155(4):934–47. org/10.1289/ehp.1510437 PMID:26771251
and inheritance. Nat Rev Genet. 13(5): https://doi.org/10.1016/j.cell.2013.09.053
343–57. https://doi.org/10.1038/nrg3173 PMID:24119843 23. Hattori N, Ushijima T (2016). Epigenetic
PMID:22473383 impact of infection on carcinogenesis:
16. Gonzalez-Perez A, Jene-Sanz A, Lopez- mechanisms and applications. Genome
8. Kopp F, Mendell JT (2018). Functional Bigas N (2013). The mutational landscape Med. 8(1):10. https://doi.org/10.1186/
classification and experimental dis- of chromatin regulatory factors across
section of long noncoding RNAs. Cell. s13073-016-0267-2 PMID:26823082
4,623 tumor samples. Genome Biol.
172(3):393–407. https://doi.org/10.1016/j. 24. Ushijima T, Hattori N (2012). Molecular
14(9):r106. https://doi.org/10.1186/gb-2013-
cell.2018.01.011 PMID:29373828 pathways: involvement of Helicobacter
14-9-r106 PMID:24063517
9. Iyer MK, Niknafs YS, Malik R, Singhal pylori-triggered inflammation in the forma-
U, Sahu A, Hosono Y, et al. (2015). The 17. Vogelstein B, Papadopoulos N, Velculescu tion of an epigenetic field defect, and its
landscape of long noncoding RNAs in the VE, Zhou S, Diaz LA Jr, Kinzler KW (2013). usefulness as cancer risk and exposure
human transcriptome. Nat Genet. 47(3): Cancer genome landscapes. Science. markers. Clin Cancer Res. 18(4):923–9.
199–208. https://doi.org/10.1038/ng.3192 339(6127):1546–58. https://doi.org/10.1126/ https://doi.org/10.1158/1078-0432.CCR-
PMID:25599403 science.1235122 PMID:23539594 11-2011 PMID:22205689

25. Ushijima T (2007). Epigenetic field for 31. Worm Ørntoft MB (2018). Review of blood- 37. Stathis A, Bertoni F (2018). BET pro-
cancerization. J Biochem Mol Biol. 40(2): based colorectal cancer screening: how teins as targets for anticancer treatment.
142–50. PMID:17394762 far are circulating cell-free DNA methyla- Cancer Discov. 8(1):24–36. https://doi.
tion markers from clinical implementation? o r g /10.115 8 / 215 9 - 8 2 9 0.C D -17- 0 6 0 5
26. Ushijima T, Asada K (2010). Aberrant DNA Clin Colorectal Cancer. 17(2):e415–33. PMID:29263030
methylation in contrast with mutations. https://doi.org/10.1016/j.clcc.2018.02.012
Cancer Sci. 101(2):300–5. https://doi. PMID:29678513 38. Feinberg AP (2018). The key role of
org/10.1111/j.1349 -7006.2009.01434.x epigenetics in human disease pre-
32. Koch A, Joosten SC, Feng Z, de Ruijter vention and mitigation. N Engl J Med.
PMID:19958364
TC, Draht MX, Melotte V, et al. (2018). 378(14):1323–34. https://doi.org/10.1056/
27. Takeshima H, Ushijima T (2010). Analysis of DNA methylation in cancer: NEJMra1402513 PMID:29617578
Methylation destiny: Moira takes ac- location revisited. Nat Rev Clin Oncol.
15(7):459–66. https://doi.org/10.1038/ 39. McCabe MT, Low JA, Daignault S,
count of histones and RNA polymerase
s41571-018-0004-4 PMID:29666440 Imperiale MJ, Wojno KJ, Day ML
II. Epigenetics. 5(2):89–95. https://doi.
(2006). Inhibition of DNA methyltrans-
org/10.4161/epi.5.2.10774 PMID:20160507 33. Moran S, Martínez-Cardús A, Sayols S, ferase activity prevents tumorigenesis
Musulén E, Balañá C, Estival-Gonzalez A, in a mouse model of prostate cancer.
28. Asada K, Nakajima T, Shimazu T,
et al. (2016). Epigenetic profiling to classify Cancer Res. 66(1):385–92. https://doi.
Yamamichi N, Maekita T, Yokoi C, et al.
cancer of unknown primary: a multicen- org/10.1158/0008 - 5 472.CAN - 05 -2020
(2015). Demonstration of the usefulness
tre, retrospective analysis. Lancet Oncol. PMID:16397253
of epigenetic cancer risk prediction by a 17(10):1386–95. https://doi.org/10.1016/
multicentre prospective cohort study. Gut. S1470-2045(16)30297-2 PMID:27575023 40. Yoo CB, Chuang JC, Byun HM, Egger G,
64(3):388–96. https://doi.org/10.1136/ Yang AS, Dubeau L, et al. (2008). Long-
gutjnl-2014-307094 PMID:25379950 34. Jacinto FV, Esteller M (2007). Mutator path- term epigenetic therapy with oral zebu-
ways unleashed by epigenetic silencing in larine has minimal side effects and pre-
29. Maeda M, Moro H, Ushijima T (2017). human cancer. Mutagenesis. 22(4):247–53. vents intestinal tumors in mice. Cancer
Mechanisms for the induction of gas- https://doi.org/10.1093/mutage/gem009 Prev Res (Phila). 1(4):233–40. https://doi.
tric cancer by Helicobacter pylori infec- PMID:17412712 org/10.1158/1940- 6207.CAPR- 07- 0008
tion: aberrant DNA methylation pathway. PMID:19138966
35. Abe M, Ohira M, Kaneda A, Yagi Y,
Gastric Cancer. 20(Suppl 1):8–15. https://
Yamamoto S, Kitano Y, et al. (2005). CpG
d o i.o r g /10.10 07/s1012 0 - 016 - 0 6 5 0 - 0 41. Niwa T, Toyoda T, Tsukamoto T, Mori
island methylator phenotype is a strong
PMID:27718135 A, Tatematsu M, Ushijima T (2013).
determinant of poor prognosis in neuro-
Prevention of Helicobacter pylori-induced
30. Leygo C, Williams M, Jin HC, Chan blastomas. Cancer Res. 65(3):828–34.
gastric cancers in gerbils by a DNA de-
MWY, Chu WK, Grusch M, et al. (2017). PMID:15705880
methylating agent. Cancer Prev Res (Phila).
DNA methylation as a noninvasive epi- 36. Pfister SX, Ashworth A (2017). Marked 6(4):263–70. https://doi.org/10.1158/1940-
genetic biomarker for the detection of for death: targeting epigenetic changes in 6207.CAPR-12-0369 PMID:23559452
cancer. Dis Markers. 2017:3726595. cancer. Nat Rev Drug Discov. 16(4):241–
ht tps://doi.org/10.1155/2017/3726595 63. https://doi.org/10.1038/nrd.2016.256
PMID:29038612 PMID:28280262
214
3.9
Immune function
From the tumour microenvironment to
therapeutic targeting
Alberto Mantovani Terry Lichtor (reviewer)
Graham Pawelec (reviewer)
The ecological niche in which cell cluding macrophages, neutrophils,

SUMMARY transformation and tumour progres- mast cells, and eosinophils are pres
sion occur is an essential component ent in the TME. Tumour-associated
●● Immune cells and mediators of of malignancy [1,2]. Innate and adap- macrophages (TAMs) are prototypic
innate and adaptive immunity tive immunity play key roles in the inflammatory cells, playing a key
CHAPTER 3.9
SECTION 3
are essential components of the tumour microenvironment (TME) by role in the orchestration of the TME.
tumour microenvironment. interacting with cancer cells as well Mononuclear phagocytes are
as with stroma and the vascular bed. extremely plastic. In the context of
●● Innate and adaptive immunity Immunity in all its diversity and interferon-driven type 1 immune
in the tumour microenvironment plasticity acts as a double-edged responses, macrophages acquire
are double-edged swords. sword during carcinogenesis, inva- tumoricidal activity. Type 1 immunity
●● Appropriately activated adap- sion, and metastasis. Appropriately signatures are generally associated
activated T cells and innate immune with better prognosis in human tu-
tive immune responses medi-
effectors (natural killer [NK] cells) mours [7]. Moreover, type 1 immu-
ate resistance to carcinogen-
mediate early elimination of trans- nity resulting in M1 polarization of
esis and progression.
formed cells and limit progression macrophages mediates the initial
●● In contrast, cancer-related in- [3]. In contrast, inflammatory cells (elimination) phase in the natural
flammation orchestrated by in- and myeloid cells – in particular, history of carcinogenesis [8].
nate immunity, such as mac- macrophages – act as “corrupted During neoplastic progression,
rophages and the complement policemen”, promoting carcinogen- macrophage function is skewed in a
esis and tumour progression at pro-tumour direction (M2 or M2-like)
system, facilitates tumour pro-
different levels, including suppres- [7]. Signals responsible for the pro-
gression via several mecha-
sion of effective adaptive immune tumour function of TAMs are known
nisms, including suppression of
responses [4,5].
adaptive immune responses. to originate from tumour cells (e.g.
This chapter concisely summa-
interleukin-10 [IL-10], transforming
●● Progress has been made in de- rizes key aspects of the yin–yang
growth factor β [TGF-β]); T helper
fining the beneficial anti-cancer relationship between immunity and
type 2 (Th2) cells, eosinophils, or
immunity cycle, its cellular and cancer, emphasizing clinical impli-
basophils (IL-4 or IL-13, resulting in
cations. Inflammation and innate im-
molecular brakes (checkpoints), M2 activation); B cells (antibodies,
munity are discussed first, in a sche-
and its relevance to prognosis immune complexes); and stromal
matic way, followed by a description
and treatment of human cancers. cells (IL-1).
of lymphoid cell-mediated immune
There is evidence suggesting
●● A revised view of the role of responses that have impacts on pre-
vention, diagnosis, and treatment. that the relative importance of dif-
the tumour microenvironment ferent pathways for regulating the
in cancer progression, and function of TAMs varies in different
the dissection of molecular Inflammation, innate tissues [9]. Single-cell analysis has
mechanisms, has opened up a immunity, and cancer added a new dimension to the dis-
new frontier in oncology, repre- A connection between inflamma- section of myeloid cell diversity in
sented by tumour immunology tion and cancer (Fig. 3.9.1) has cancer [10]. Clusters of more than 10
and immunotherapy. long been perceived [1,4,6] (see differentiation/activation states have
Chapter 3.5). Inflammatory cells in- been identified. The microanatomical
Chapter 3.9 • Immune function 215

Fig. 3.9.1. Pathways connecting inflammation and cancer at the tissue level and at the
systemic level.
FUNDAMENTALS
Tissue level Systemic level ■■ Immune cells are a key
component of the tumour
Carcinogen
microenvironment.
Oncogene
activation Intrinsic ■■ Components of innate
pathway immunity drive tumour-
Obesity
promoting inflammation.
Inflammatory cells
and mediators ■■ Macrophages promote
Inflammaging tumour progression and
(e.g. macrophages,
complement) immunosuppression.
Lifestyle ■■ T cells eliminate and edit
Extrinsic (diet, physical exercise) cancer cells.
pathway
Chronic
non-resolving ■■ Checkpoints and other
inflammation pathways of suppression
restrain the anti-tumour activity
Tumour promotion
(genetic instability, angiogenesis, of T cells, natural killer cells,
suppression of immunity) and macrophages.
■■ Immune components have
strong prognostic significance.
signals responsible for the diversity plement system can also act as a ■■ Immunology and immunother-
of cancer-associated myeloid cells double-edged sword by mediating apy represent a new frontier in
remain to be defined. complement-dependent cytotoxic- the fight against cancer.
Phagocytosis is the eponymous ity in the presence of antibodies or,
function of mononuclear phagocytes. alternatively, by recruiting tumour-
CD47 on normal and tumour cells promoting myeloid cells. The long
delivers a “don’t eat me” signal via pentraxin PTX3 was shown to act as and mortality from lung cancer [17].
signal regulatory protein 1α (SIRP1α) an extrinsic oncosuppressor, which These and other results provide a
on macrophages [11]. CD47 is am- is epigenetically silenced in selected strong proof-of-principle rationale
plified downstream of the oncogene human tumours [15]. PTX3 silenc- for targeting tumour-promoting in-
MYC [5]. CD47, which is one of the ing unleashes complement-driven flammation in human tumours.
negative regulators (checkpoints) of recruitment and functional orienta-
myeloid cells, can serve as a thera- tion (M2-like) of TAMs, which is re- Anti-tumour immunity
peutic target [12]. Recent evidence sponsible for tumour promotion and and immunosuppression
suggests that blocking CD47 may increased genetic instability.
unleash antibody-dependent cel- Cytokines are a key component T-cell-orchestrated type 1 immune
lular cytotoxicity and phagocytosis of tumour-promoting inflammation. responses mediate host resistance
mediated by TAMs [13]. TAMs and In particular, IL-1 has been shown to during the early phases of carcino-
other myeloid cells – for example, op- drive myeloid cell infiltration, gener- genesis (Fig. 3.9.2). Moreover, in hu-
erationally defined myeloid-derived ation of MDSCs, and angiogenesis man tumours, the presence of T cells
suppressor cells (MDSCs) [14] – [8]. Recent evidence is consistent and type 1 immunity or interferon
have now been shown to have im- with IL-1 being an important driver signatures is associated with bet-
pacts on diverse aspects of cancer of progression in human tumours ter prognosis [7,19]. Genomics has
progression, including tumour cell [17,18]. The Canakinumab Anti- provided a more in-depth view of im-
proliferation and invasion, construc- inflammatory Thrombosis Outcomes mune cell recognition of tumour-spe-
tion of a metastatic niche, angio- Study (CANTOS) was originally de- cific antigens, arising from mutations,
genesis, and immunosuppression. signed to assess the impact of an or tumour-associated antigens, re-
Immunosuppression, a key function anti-IL-1β antibody (canakinumab) sulting from overexpression of normal
of myeloid cells, is discussed below. on atherosclerosis-related cardio- cell genes. Evidence in mouse and
Components of the humoral arm vascular pathology. In more than human tumours has indicated that
of innate immunity have recently 10 000 patients, blocking of IL-1β mutations and genetic instability rep-
been recognized as important ele- was associated with reductions of resent the fundamental molecular ba-
ments in the TME [15,16]. The com- more than 50% in the incidence of sis for T-cell-dependent anti-tumour
216
immunity [3,7,9,20]. The intersection sis has provided new vistas on the suggests that this pathway plays a
of genomics and the dissection of im- T-cell receptor repertoire and func- dominant role in carcinoma of the
munity is paving the way to personal- tional properties of tumour-infiltrating breast and in pancreatic ductal ad-
ized immunotherapy approaches. lymphocytes. Regulatory T cells enocarcinoma. Th17 cells activate
Failure of effective immunity (Treg cells) have long been associ- a neutrophil-dependent pathway of
is associated with progression ated with immunosuppression in immunity to extracellular pathogens,
and the appearance of clinical cancer. Single-cell analysis has led and neutrophils can contribute to
cancer. In the Darwinian TME, tu- to the identification of molecules ex- myeloid cell-mediated tumour pro-
mour cell-centred and host cell- pressed by infiltrating Treg cells [22]. motion [23].
centred mechanisms of immune For instance, the IL-1 decoy receptor Whereas a skewed, inappro-
evasion drive progression, inva- IL-1R2 was found to be expressed at priate response and exhaustion
sion, and metastasis (Fig. 3.9.2). very high levels in infiltrating Treg cells. are important determinants of the
Mechanisms of physical exclusion Whereas a Th1-orchestrated cy- failure of immunity to restrain can-
(e.g. extracellular matrix deposition totoxic T-cell-mediated response has cer, active immunosuppression has
[21]), and selection of less immuno- a protective function, Th2-polarized emerged as a dominant mechanism
genic variants, can hamper effec- T cells and Th17 cells trigger tu- of progression. Checkpoints are
tive recognition. mour-promoting cascades. IL-4 and physiological mechanisms to re-
T-cell exhaustion is an effector IL-13 produced by Th2 cells or by strain uncontrolled T-cell activation
T-cell-intrinsic mechanism for failure eosinophils elicit alternative M2 po- and tissue damage. Targeting of the
to mount an effective immune re- larization of macrophages, which reprogrammed cell death 1 (PD-1)/
sponse. Single-cell genomic analy- sults in tumour promotion. Evidence programmed death ligand 1 (PD-L1)
CHAPTER 3.9
SECTION 3
Fig. 3.9.2. The immunity–immunosuppression circle. ADCC, antibody-dependent cellular cytotoxicity; ADCP, antibody-dependent
cellular phagocytosis; CDC, complement-dependent cytotoxicity; IL-10, interleukin-10; NK, natural killer; PG, prostaglandin; TAA,
tumour-associated antigens; TGF-β, transforming growth factor β; Th2, T helper type 2; TSA, tumour-specific antigens.
Elimination Escape
Effective immunity Immunosuppression
Metastasis
Progression
T-cell exhaustion
Treg checkpoints
TSA/TAA recognition (T, NK, M0)
T cells Microbiome
Skewed T cells (Th2, Th17)
Type 1 immune responses M2-like macrophages
NK cells Lifestyle
(diet, exercise) Myeloid cell-mediated
M1 macrophages suppression (checkpoints;
Neutrophils IL-10, TGF-β, PG, aminoacid
B cells Organ
contexture metabolism)
ADCC Neutrophils
ADCP Mast cells
CDC B cells
Complement
Effective
immunotherapy

axis and cytotoxic T-lymphocyte- sive cytokines (IL-10) and/or produc- In many human tumours, in par-
associated protein 4 (CTLA-4) has tion of antibodies and formation of ticular colorectal cancer, T-cell infil-
had an unprecedented impact on immune complexes that skew TAMs tration is a positive prognostic indica-
cancer treatment. A host of mo- in an M2-like direction [6]. tor, independent of other parameters.
lecular brakes acting on T cells as Adaptive T-cell-orchestrated im- The so-called Immunoscore to as-
well as other cell types have been munity and its subversion are cen- sess T-cell infiltration was validated
identified [3], and these represent tral in the control of carcinogenesis in a large cooperative study. A recent
candidate therapeutic targets. and progression. Recent results study involving more than 3500 pa-
Immunosuppression in the TME have shed new light on the long- tients worldwide confirmed the value
is orchestrated by tumour cells overlooked role of innate lymphoid of the Immunoscore in colorectal
and/or by stromal cells, in particu- cells. NK cells are a population of cancer as an independent prognostic
lar myelomonocytic cells. Tumour innate lymphoid cells that has not factor [19]. That study proposed mov-
cells produce immunosuppressive been credited with playing a ma- ing from a tumour–node–metastasis
cytokines (IL-10, TGF-β) and ex- jor role in resistance against solid (TNM) classification to a TNM-I clas-
press triggers of checkpoint block- tumour carcinogenesis. Evidence sification of colorectal cancer, where
ade, such as PD-L1. PD-L1 gene suggests that NK cells mediate re- “I” stands for immunity.
amplification was found to occur in sistance against haematopoietic The results obtained in the past
Hodgkin lymphoma, in which PD- neoplasms and restrain haematog- few years prove that assessment
L1 is also prominently expressed of the quantity and diversity of im-
enous dissemination of cancer cells.
by TAMs. In general, the relative mune cell infiltration has prognostic
The differentiation and activity of NK
contribution of tumour cells versus significance. Genomic analysis of
cells are also controlled by nega-
myeloid cells to PD-L1 expression the TME has confirmed these ob-
tive regulators. Recently, novel NK
in the TME varies considerably in servations and has provided tools
cell checkpoints (e.g. IL-1R8) were
different human cancer types [15]. for TME-based classification of
identified, and unleashed NK cells
Myelomonocytic cells at dif- cancer, as illustrated by colorectal
were found to mediate resistance
ferent stages of differentiation or cancer [19]. Conventional immuno-
to carcinogenesis and metastasis
activation have the capacity to histology as well as gene expres-
strongly suppress T-cell-mediated at NK-cell-rich anatomical sites,
sion profiling are faced with the
responses. MDSCs are operation- such as the liver and the lung [25].
challenge of moving from prognosis
ally defined as a mixed population Elucidation of the molecular mecha-
to prediction, particularly in the con-
of relatively immature myeloid cells nisms that regulate the function of
text of immunotherapy.
with potent suppressive activity NK cells and innate lymphoid cells
[24]. Depending on the system ex- may pave the way to therapeutic
amined among MDSCs, suppres- strategies that are complementary Implications for
sion was mediated by neutrophils to the current checkpoint blockade. immunotherapy
or, more frequently, monocytes. Immunotherapy in the form of PD-
Monocytic MDSCs differentiate into Prognosis versus 1/PD-L1 and CTLA-4 checkpoint
TAMs in the TME [24]. prediction blockade inhibitors and chimeric
TAMs were found to exert immu- antigen receptor T cells is now part
As expected given the complexity of the anticancer armamentarium. A
nosuppressive activity via diverse
mechanisms. These include immuno- and diversity of the roles of innate recent review discussed the mech-
suppressive cytokines (IL-10, TGF-β), and adaptive immunity, infiltration anisms, resistance to, and stum-
triggers of checkpoint blockade (e.g. of different components of the im- bling blocks of this approach [20].
PD-L1), amino acid metabolism (ar- mune system has different, at times In spite of the unprecedented broad
ginase, tryptophan metabolites), and divergent, prognostic significance. impact of checkpoint blockade in-
prostaglandins. Prostaglandins are Infiltration of TAMs is generally as- hibitors, only approximately 20% of
particularly significant in view of the sociated with worse prognosis [5], treated patients benefit from current
protective effect of aspirin on several which is a reflection of their pro-tu- checkpoint blockade. As discussed
human tumour types. mour function. However, infiltration above, new vistas on fundamen-
B cells and antibodies are part of of TAMs is associated with better tal mechanisms, including novel
the anti-tumour response. However, prognosis in colorectal cancer. The checkpoints, targeting of tumour-
evidence suggests that B cells can positive prognostic significance of promoting myeloid cells [12], and
contribute to tumour progression in TAMs in colorectal cancer reflects harnessing NK cell potential, hold
certain epithelial tumours, such as the association with response to promise to help predict which pa-
prostate cancer. B-cell-mediated tu- chemotherapy. If these results are tients will be responsive, thus spar-
mour promotion has been shown to confirmed and extended, they raise ing toxicity and contributing to the
involve different mechanisms, such the possibility of using TAMs to financial sustainability and to the
as production of immunosuppres- guide eligibility to chemotherapy. improvement of therapeutic results.
218
Conclusions checkpoint blockade, and induction nessed to cope with advanced dis-
and recruitment of Treg cells. seminated neoplastic diseases.
Immunity is an essential component
Quantification of the immune Full exploitation of the diagnostic
of the TME and a key determinant
of metastasis [1,2,7]. Inflammatory and inflammatory landscape of the and therapeutic potential of innate
cells, in particular TAMs, pave the TME has provided novel prognos- and adaptive immunity will require:
way to tissue invasion and intra- tic indicators of cancer progres- an integrated in-depth analysis of its
vasation and provide a nurturing sion, as shown by quantification components in primary tumours ver-
microenvironment for metastasis, of tumour-infiltrating T cells and sus spreading, metastatic tumours;
serving as a component of the can- TAMs. Genomic technologies have the dissection of the diversity of me
cer cell niche at distant sites. NK added a new dimension to the char- tastatic niches; and the identification
cells are innate lymphoid cells that acterization of the TME and to the and development of new molecular
have long been considered to play classification of cancers. Finally, and cellular tools. Moreover, the in-
a role in resistance against haema- the elucidation of the mode of ac- tegration of –omics approaches with
togenous dissemination of cancer tion of conventional cytoreductive the elucidation of immunological
cells, in particular to the lungs. strategies, the impact of checkpoint complexity holds promise for the de-
Tumour progression and escape blockade inhibitors, the introduc- velopment of personalized immuno-
are associated with immunosup- tion of therapeutic antibodies, and, therapy, and for addressing the fun-
pressive pathways in innate and very recently, adoptive cell therapy damental issue of the sustainability
adaptive anti-tumour responses, for haematological malignancies of these innovative approaches for
which include, among others, sup- [8,26,27] have proven the principle health-care systems.
pressive myeloid cells, activation of that the immune system can be har-
CHAPTER 3.9
SECTION 3

References
1. Balkwill F, Mantovani A (2001). Inflammation 12. Mantovani A, Longo DL (2018). Macrophage 20. Wei SC, Duffy CR, Allison JP (2018). Fun
and cancer: back to Virchow? Lancet. checkpoint blockade in cancer – back to damental mechanisms of immune check-
357(9255):539–45. https://doi.org/10.1016/ the Future. N Engl J Med. 379(18):1777–9. point blockade therapy. Cancer Discov.
S0140-6736(00)04046-0 PMID:11229684 https://doi.org/10.1056/NEJMe1811699 8(9):1069–86. https://doi.org/10.1158/2159-
PMID:30380398 8290.CD-18-0367 PMID:30115704
2. Hanahan D, Weinberg RA (2011). Hallmarks
of cancer: the next generation. Cell. 144(5): 13. Advani R, Flinn I, Popplewell L, Forero A, 21. Pearce OMT, Delaine-Smith RM, Maniati E,
646–74. https://doi.org/10.1016/j.cell.2011. Bartlett NL, Ghosh N, et al. (2018). CD47 Nichols S, Wang J, Böhm S, et al. (2018).
02.013 PMID:21376230 Deconstruction of a metastatic tumor micro-
blockade by Hu5F9-G4 and rituximab in
3. Chen DS, Mellman I (2017). Elements of non-Hodgkin’s lymphoma. N Engl J Med. environment reveals a common matrix re-
cancer immunity and the cancer-immune set 379(18):1711–21. https://doi.org/10.1056/ sponse in human cancers. Cancer Discov.
point. Nature. 541(7637):321–30. https://doi. NEJMoa1807315 PMID:30380386 8(3):304–19. https://doi.org/10.1158/2159-
org/10.1038/nature21349 PMID:28102259 8290.CD-17-0284 PMID:29196464
14. Bronte V, Brandau S, Chen SH, Colombo
4. Mantovani A, Allavena P, Sica A, Balkwill MP, Frey AB, Greten TF, et al. (2016). 22. De Simone M, Arrigoni A, Rossetti G,
F (2008). Cancer-related inflammation. Recommendations for myeloid-derived sup- Gruarin P, Ranzani V, Politano C, et al.
Nature. 454(7203):436–44. https://doi.org/ pressor cell nomenclature and characteri- (2016). Transcriptional landscape of hu-
10.1038/nature07205 PMID:18650914 zation standards. Nat Commun. 7(1):12150. man tissue lymphocytes unveils unique-
https://doi.org/10.1038/ncomms12150 ness of tumor-infiltrating T regulatory
5. Mantovani A, Marchesi F, Malesci A, Laghi
L, Allavena P (2017). Tumour-associated PMID:27381735 cells. Immunity. 45(5):1135–47. https://
macrophages as treatment targets in on- doi.org/10.1016/j.immuni. 2016.10.021
15. Bonavita E, Gentile S, Rubino M, Maina PMID:27851914
cology. Nat Rev Clin Oncol. 14(7):399–416. V, Papait R, Kunderfranco P, et al. (2015).
https://doi.org/10.1038/nrclinonc.2016.217
PTX3 is an extrinsic oncosuppressor 23. Coffelt SB, Kersten K, Doornebal CW,
PMID:28117416
regulating complement-dependent inflam- Weiden J, Vrijland K, Hau CS, et al. (2015).
6. Coussens LM, Zitvogel L, Palucka AK mation in cancer. Cell. 160(4):700–14. IL-17-producing γδ T cells and neutrophils
(2013). Neutralizing tumor-promoting chron- https://doi.org/10.1016/j.cell.2015.01.004 conspire to promote breast cancer metas-
ic inflammation: a magic bullet? Science. PMID:25679762 tasis. Nature. 522(7556):345–8. https://doi.
339(6117):286–91. https://doi.org/10.1126/ org/10.1038/nature14282 PMID:25822788
science.1232227 PMID:23329041 16. Reis ES, Mastellos DC, Ricklin D, Mantovani
A, Lambris JD (2018). Complement in can- 24. Bronte V (2018). The mesenchymal and my-
7. Fridman WH, Zitvogel L, Sautès-Fridman cer: untangling an intricate relationship. eloid regulation of immunity: power is noth-
C, Kroemer G (2017). The immune con- Nat Rev Immunol. 18(1):5–18. https://doi. ing without control. Semin Immunol. 35:1–2.
texture in cancer prognosis and treat- org/10.1038/nri.2017.97 PMID:28920587 https://doi.org/10.1016/j.smim.2018.03.001
ment. Nat Rev Clin Oncol. 14(12):717–34. PMID:29566937
https://doi.org/10.1038/nrclinonc.2017.101 17. Ridker PM, MacFadyen JG, Thuren T,
PMID:28741618 Everett BM, Libby P, Glynn RJ; CANTOS 25. Molgora M, Bonavita E, Ponzetta A, Riva
Trial Group (2017). Effect of interleukin-1β F, Barbagallo M, Jaillon S, et al. (2017). IL-
8. Gubin MM, Artyomov MN, Mardis ER,
inhibition with canakinumab on incident 1R8 is a checkpoint in NK cells regulating
Schreiber RD (2015). Tumor neoanti-
lung cancer in patients with atherosclero- anti-tumour and anti-viral activity. Nature.
gens: building a framework for personal-
sis: exploratory results from a randomised, 551(7678):110–4. https://doi.org/10.1038/
ized cancer immunotherapy. J Clin Invest.
125(9):3413–21. https://doi.org/10.1172/ double-blind, placebo-controlled trial. nature24293 PMID:29072292
JCI80008 PMID:26258412 Lancet. 390(10105):1833–42. https://
doi.org/10.1016/S0140-6736(17)32247-X 26. Sadelain M, Rivière I, Riddell S (2017).
9. Schumacher TN, Schreiber RD (2015). PMID:28855077 Therapeutic T cell engineering. Nature.
Neoantigens in cancer immunotherapy. 545(7655):423–31. https://doi.org/10.1038/
Science. 348(6230):69–74. https://doi.org/ 18. Wu TC, Xu K, Martinek J, Young RR, nature22395 PMID:28541315
10.1126/science.aaa4971 PMID:25838375 Banchereau R, George J, et al. (2018).
IL1 receptor antagonist controls tran- 27. Anderson KG, Stromnes IM, Greenberg
10. Azizi E, Carr AJ, Plitas G, Cornish AE, scriptional signature of inflammation in PD (2017). Obstacles posed by the tu-
Konopacki C, Prabhakaran S, et al. (2018). patients with metastatic breast cancer. mor microenvironment to T cell activity:
Single-cell map of diverse immune pheno- a case for synergistic therapies. Cancer
Cancer Res. 78(18):5243–58. https://doi.
types in the breast tumor microenvironment. Cell. 31(3):311–25. https://doi.org/10.1016/j.
org/10.1158/0 0 08 - 5 472.CAN -18 - 0 413
Cell. 174(5):1293–1308.e36. https://doi.org/ ccell.2017.02.008 PMID:28292435
PMID:30012670
10.1016/j.cell.2018.05.060 PMID:29961579
19. Pagès F, Mlecnik B, Marliot F, Bindea
11. Tseng D, Volkmer JP, Willingham SB,
Contreras-Trujillo H, Fathman JW, Fernhoff G, Ou FS, Bifulco C, et al. (2018).
NB, et al. (2013). Anti-CD47 antibody- International validation of the consensus
mediated phagocytosis of cancer by mac- Immunoscore for the classification of colon
rophages primes an effective antitumor cancer: a prognostic and accuracy study.
T-cell response. Proc Natl Acad Sci U S A. Lancet. 391(10135):2128–39. https://doi.
110(27):11103–8. https://doi.org/10.1073/ org/10.1016/S0140 - 673 6(18) 3 0789 -X
pnas.1305569110 PMID:23690610 PMID:29754777
220
3.10
The microbiome
Its influence on tumorigenesis
and therapy
Georg Zeller Nele Brusselaers (reviewer)
Mazda Jenab (reviewer)
Herbert Tilg (reviewer)
●● Microbiota-targeted cancer pre- terns, which are conserved com-

SUMMARY vention strategies appear prom- ponents of bacterial cell walls [3].
ising, but they have yet to be Under homeostatic conditions, mu-
●● Changes in the human micro- evaluated in prospective studies. cus and epithelial cells shield host
biota – particularly in the large tissues from unrestricted exposure
CHAPTER 3.10
SECTION 3
intestine, but also in other lo- to microbe-associated molecular
cations – have been associ- The understanding of the complex patterns. However, many dysbi-
ated with multiple tumour types relationship between the human mi- otic microbiome states, both in the
in retrospective case–control crobiota and its host organism has intestine and in the oral cavity, are
studies. However, it often re- expanded rapidly in recent years, characterized by microbes degrad-
mains unclear whether these
fuelled by high-throughput meta ing and penetrating the mucus. This
alterations are consequential,
genomic sequencing technologies, compromised barrier eventually per-
or relevant to cancer etiology.
advanced bioinformatics analysis mits bacterial translocation and al-
Currently, evidence is strongest
methodology, and the development lows increased levels of microbe-as-
for an enrichment of pathogenic
of experimental model systems [1]. sociated molecular patterns to reach
species in the gut microbiota
Research focusing primarily on the the circulation. The inflammatory re-
associated with cancers of the
gut microbiome has led to a growing sponses that ensue both locally and
digestive tract.
appreciation of its key role in main- systemically are a central factor in
●● To date, bacterial mechanisms taining health, and of dysbiotic gut many pathologies and contribute to
that promote carcinogenesis microbiome states being associated neoplastic transformations in many
are still incompletely elucidated. with many common human disor- organs [3,8].
However, a few bacterial geno- ders, including cancer [1].
toxins and carcinogens are well The microbiota, in particular in
described, as well as mecha- the gut, is shaped by, and in turn
Cancers associated
nisms by which bacteria repro- modulates, many environmental and
with a single microbial
gramme host signalling towards host factors by chemical transfor- pathogen
neoplastic transformation, pro- mation of endogenous (host) and Helicobacter pylori is the best-un-
mote inflammation, or protect exogenous (diet, medication) me- derstood model bacterium with a
against immunosurveillance. tabolites as well as host–microbiota causal role in infection-related can-
●● Recent research has uncov- signalling. Recently, we have begun cer, and the only one that has been
ered profound effects of the to understand the contribution of classified as carcinogenic to humans
gut microbiota on cancer ther- these processes to individual-spe- (Group 1) by the IARC Monographs
apies. Strikingly, response to cific cancer risks and therapy out- (see Chapter 2.2). As a persistent
immunotherapy depends par- comes (Fig. 3.10.2) [1–7]. colonizer of gastric mucosa, H. py-
tially on an intact gut micro- Central to this host–microbiota lori can develop pathogenic traits,
biota with immunostimulatory cross-talk is the host immune sys- and its presence is a major risk fac-
function. Whereas antibiotics tem (see Chapter 3.9) [3,7]. Host tor for gastric cancer. Consistent
compromise immunotherapy re- cells sense commensal and patho- with its causal role, eradication of H.
sponse, microbiome reconstitu- genic bacteria through pattern rec- pylori was found to significantly re-
tion (e.g. by probiotics) improves ognition receptors. These bind to duce the incidence of gastric cancer,
outcomes in animal models. microbe-associated molecular pat- both in animals and in humans [9].
Chapter 3.10 • The microbiome 221

Fig. 3.10.1. A street in Busan, a large city in the Republic of Korea. The presence
of Helicobacter pylori bacteria, a major risk factor for gastric cancer, is particularly FUNDAMENTALS
relevant to parts of China, Japan, and the Republic of Korea.
■■ Epithelial and mucosal
surfaces of the human body
are colonized by complex
microbial communities
consisting of bacteria,
archaea, eukaryotes (mostly
unicellular in this context), and
viruses; collectively, they are
referred to as the microbiota.
■■ The microbiota is charac
terized by large taxonomic
diversity and inter-individual
heterogeneity, and also
possesses enormous
metabolic capabilities, which
far exceed the enzymatic
repertoire of the host.
■■ Collectively, the microbiota
and its genes and metab
olites, which shape the
environmental milieu, are
referred to as the microbiome.
■■ The microbiota has co-
Research on Helicobacter has Helicobacter largely evades the im- evolved with its host to fulfil
mune system to persist in the host many important physiological
unveiled many of the key molecu-
functions in co-metabolism
lar mechanisms by which bacteria (Fig. 3.10.3) [9,10]. with the rest of the organism;
persistently colonize host tissues Another well-studied example these include the digestion of
and create a pro-oncogenic milieu. of a single bacterial pathogen that dietary compounds and the
Many of these might be generaliz- may promote tumorigenesis during synthesis of micronutrients,
able to other cancer-associated chronic infection is Salmonella en- as well as the breakdown
pathogens (Fig. 3.10.3) [10]. Key terica serovar Typhi. Epidemiological of endogenous (host) and
xenobiotic compounds,
features of H. pylori virulence in- studies have associated persistent
including drugs.
clude bacterial surface proteins Salmonella colonization of the gall
facilitating attachment to epithelial bladder with strongly increased risk ■■ Culture-independent metage-
cells, enzymes capable of modify- of biliary cancer. This is further sup- nomic sequencing (and other
–omics technologies) has
ing the host environment to facilitate ported by research on mouse mod- enabled microbiome charac-
colonization (e.g. urease permitting els of long-term Salmonella infection terization in situ. Based on
survival in a low-pH environment), [10,11]. this technology, microbiome-
and manipulation of host signalling. In these etiologies, a single infec- wide association studies have
Reprogramming of cellular signal- tious agent is sufficient to promote linked many common human
ling can be achieved via diffusible neoplastic transformation. Based diseases, including cancers,
with changes in microbiota
toxins and/or export of effector pro- on culture-independent metagenomic
composition; disease-asso-
teins into host cells through a bac- sequencing of the more diverse mi- ciated microbiome states are
terial secretion system. This can crobial communities that inhabit the sometimes referred to as
locally alter mucus and acid secre- mouth and the gut, polymicrobial dysbiosis.
tion of the host, which further facili- signatures have been statistically as- ■■ Experimental studies based
tates colonization; it can also entail sociated with several other tumour on in vitro systems and
stimulation of host pathways that types, in particular with those that animal models complement
drive proliferation and cell survival are anatomically close to the gastro- microbiome-wide association
or compromise tumour suppres- intestinal tract. However, because of studies as they have started
sion and DNA damage response the complexity of cancer-associated to unravel causal relationships
and molecular mechanisms
(see Chapter 3.4). Manipulation alterations in these communities,
underlying microbe–host
of other host pathways can alter the timing of microbial tumour colo- interactions in health and
host cell morphology and polarity. nization, causalities, and molecular disease, including in the
Finally, despite its ability to sustain mechanisms remain to be elucidat- etiology of several cancers.
a chronic inflammatory response, ed in most cases. Although some
222
Fig. 3.10.2. Environment- and host-dependent effects of the microbiota on carcinogenesis, cancer prevention, and therapy. The
composition of the microbiota is shaped by many environmental factors, such as diet and xenobiotics (pharmaceuticals), as well as
host factors, which include lifestyle, metabolism, the immune system, and pathophysiological conditions (e.g. cancer) that alter mu-
cosal milieus. The microbiota itself modulates many of these effects, which contributes to individual-specific cancer risk and therapy
outcomes. Examples of such modulations are (i) gut microbial fermentation of dietary fibre into butyrate (and other short-chain fatty
acids), which promotes epithelial health and prevents neoplastic transformation; (ii) gut microbial metabolism of primary bile acids
into carcinogenic secondary bile acids; (iii) disruption of mucosal barriers by microbial mucus degradation and pro-inflammatory
metabolites, which promotes cancer development; (iv) gut microbial drug metabolism and reversal of host detoxification processes;
and (v) microbial immunostimulation. Both (iv) and (v) can affect the outcome of cancer therapy. Current knowledge of cancer-
preventing or cancer-promoting mechanisms is based on preclinical and observational studies. However, because large-scale
cohort studies in multiple countries are now collecting faecal samples, it will soon become possible to evaluate gut microbial risk
factors for several cancer types prospectively. Similarly, prospective follow-up studies of cancer patients will enable better definition
of prognostic microbial biomarkers for general survival or treatment success. (For more details, see [2–5,20].)
Environment Diet Colonisation

Xenobiotics resistance
Preventing
cancer
Microbiota Enhancing
Lifestyle epithelial health
Host Carcinogenic
CHAPTER 3.10
metabolites
SECTION 3
Promoting
Host metabolism
cancer
Disruption of
mucosal barrier
Immune system Immune

modulation
Modulating
therapy outcome
Xenobiotic
Pathologies metabolism
bacterial pathogens and their pro- genera Fusobacterium, Parvimonas, that promote epithelial health (e.g.
oncogenic mechanisms have been Porphyromonas, and Escherichia short-chain fatty acids, vitamins, and
characterized in animal models, the [4,12–15]. Preclinical studies have antioxidants) towards those that con-
evidence from clinical studies is still complemented these microbiome- tribute to carcinogenesis and inflam-
limited. To date, the role of the gut wide association studies by eluci- mation (including secondary bile
microbiota in gastrointestinal tumour dating the molecular mechanisms acids and protein degradation prod-
development has been most conclu- through which gut microbes may ucts) (see Chapter 5.5) [4,5,14–16].
sively defined. directly or indirectly promote colo- Because the liver is connected to
rectal carcinogenesis (Fig. 3.10.3). the intestine through the portal vein,
Cancers of the Mouse models have revealed sev- it is exposed to gut bacterial metabo-
gastrointestinal tract eral virulence factors and metabolites translocating through the epithe-
associated with lites from Fusobacterium nucleatum lium into the circulation. Especially
altered gut microbiota and strains of Bacteroides fragilis when the intestinal barrier is compro-
composition or Escherichia coli that can trigger mised, microbial metabolites and mi-
Many independent studies have pro-oncogenic signalling and cel- crobe-associated molecular patterns
linked colorectal cancer at the time of lular transformation programmes reach the liver in higher concentra-
diagnosis to alterations in gut (faecal (Fig. 3.10.3) [2,4,13]. In addition, colo- tions. There, upon binding to pattern
and mucosal) microbiota composi- rectal cancer appears to be linked to recognition receptors at multiple liver
tion. Metagenomic meta-analyses a shift in the metabolic products of cell types, they can elicit persistent
confirmed a broad agreement of bacterial digestion of dietary and host inflammatory programmes. This pro-
tumour-enriched bacterial taxa be- metabolites (contained in meat, fat, cess was found to be a hallmark of
tween studies. These include the fibre, or digestive juices) from those many chronic liver diseases that are

Fig. 3.10.3. Molecular mechanisms by which bacteria promote carcinogenesis. Although most mechanisms remain to be
characterized in detail, they can be broadly grouped into direct and indirect mechanisms. Bacteria can directly contribute to
genomic instability of host cells via diffusible genotoxins such as colibactin or cytolethal distending toxin (CDT). Another means
of directly promoting genomic instability is to interfere with host DNA damage repair (in some cases via manipulation of p53
activity). Bacterial pathogens have further evolved a range of mechanisms that divert host signalling processes to promote cell
survival and proliferation. As a consequence, various pro-oncogenic cellular programmes are triggered, for instance via activation
of oncogenes and deactivation of tumour suppressors, or via disruption of cell–cell junctions, cell polarity, and epithelial barrier
integrity (through interference with β-catenin/WNT signalling). Bacterial manipulation of the mitogen-activated protein kinase
(MAPK), the signal transducer and activator of transcription 3 (STAT3) protein, and the nuclear factor kappa-light-chain-enhancer of
activated B cells (NF-κB) pathways can result in pro-inflammatory signalling (see Chapter 3.5). Systemic inflammation can also be
promoted indirectly via signalling of host pattern recognition receptors (PRRs) binding to microbe-associated molecular patterns,
which include bacterial cell-wall antigens such as lipopolysaccharide. Bacteria are also capable of producing pro-inflammatory
metabolites, such as secondary bile acids, which can act systemically when reaching the circulation. Indirect ways of triggering local
inflammation include the production of reactive oxygen species (ROS), such as extracellular superoxide, or the induction of the host
spermine oxidase (SMO), an enzyme that is involved in generating hydrogen peroxide from polyamine breakdown. Finally, some
bacteria can also elicit immunosuppressive responses, thereby indirectly contributing to tumour survival through evasion of immune
surveillance. AvrA, avirulence A; BFT, Bacteroides fragilis toxin; CagA, cytotoxin-associated gene A; FadA, Fusobacterium adhesin
A; IpgD, inositol phosphate phosphatase; VacA, vacuolating cytotoxin A. (For a more detailed presentation of these mechanisms,
see [2,10].)
precursors to hepatocellular carci- findings support a causal role of an signalling [17]. However, larger clini-
noma (see Chapter 5.6) [8]. Another altered microbiome in liver inflamma- cal studies are needed to validate
process by which intestinal bacteria tion and malignancy [8]. individual microbial taxa enriched in
promote hepatocellular carcinoma There is also emerging evidence pancreatic tissue [18] or in the mouth
involves bile acids. Primary bile for a bacterial contribution to pan- and the gut of patients with pancre-
acids are secreted from the liver into creatic cancer development [13]. In atic ductal adenocarcinoma.
the gut, where they can be convert- mouse models, germ-free condi- Although microbiome-wide as-
ed into secondary bile acids, such tions or administration of antibiotics sociation studies of medium scale
as deoxycholic acid, by intestinal were shown to slow down progres- (with n ≈ 300 each) have investi-
Clostridium spp. After re-uptake, de- sion of pancreatic ductal adenocar- gated the oral microbiota in case–
oxycholic acid circulates back to the cinoma. Moreover, the microbiota control studies for oesophageal
liver, where it exerts its carcinogenic colonizing the pancreas was found cancer and head and neck cancers,
effects. In sum, several clinical stud- to play an important role in regulat- bacteria–tumour associations were
ies have revealed profound changes ing the inflammatory tone in the pan- relatively weak in these patient pop-
in the gut microbiota associated with creatic tumour microenvironment in ulations. In addition, it is currently
chronic liver diseases, and preclinical mice via pattern recognition receptor unclear whether microbial markers
224
Fig. 3.10.4. Low-temperature electron micrograph of a cluster of Escherichia coli Role of the gut
bacteria, ×10 000. E. coli strains can produce the carcinogenic colibactin toxin.
microbiome in cancer
therapy
The gut microbiota is increasingly
appreciated as a versatile “microbial
pharmacist within us” [22], because
evidence is accumulating that it can
also affect the pharmacokinetics,
efficacy, and toxicity of various
anticancer therapies (Fig. 3.10.5)
[6,22].
Chemotherapy
As one of the first examples, irinote-
can was reported to be metabolized
by intestinal bacteria. This chemo-
therapeutic drug, used to treat colo-
rectal cancer, is detoxified (glucu-
ronidated) in the liver to SN-38-G.
After SN-38-G is excreted into the
intestine, it can be reactivated by
CHAPTER 3.10
SECTION 3
bacterial β-glucuronidases, and
this causes intestinal toxicity, such
as severe diarrhoea [6].
Another example is the che
would have diagnostic or prognos- understanding of hormonal co-me- motherapeutic drug gemcitabine,
tic value for these tumour types tabolism between the host and its gut which can be rendered inactive by
[19,20]. microbiome has yet to be elucidated, bacterial enzymes, as has been
and its clinical significance remains demonstrated in mouse models.
Cancers in organs to be established [21]. Other stud- Bacteria capable of this biotrans-
outside the ies have examined microbiota resid- formation were found in tissue sam-
gastrointestinal tract ing in breast tissue of women with ples from patients with pancreatic
and without breast cancer. Whereas ductal adenocarcinoma, suggest-
Breast cancer structural alterations were not de- ing that this bacterial resistance
Among tumour types outside the tected in association with breast mechanism is clinically relevant
digestive tract, breast cancer has cancer, some studies found rare taxa [6,16,23,24].
been most extensively examined to differ in abundance in tumour tis- There is also recent evidence
for potential associations with mi- sue. However, among the published that the gut microbiota modulates the
crobiota at various body sites [7,21]. microbiome-wide association studies anti-tumour efficacy of platin-based
As in the liver, tumorigenesis in the there is little agreement on the pre- and cyclophosphamide chemother-
breast may potentially be influenced cise breast cancer-associated bacte- apies. The efficacy of cisplatin and
by the gut microbiota through pro- oxaliplatin is greatly decreased in
rial taxa [21].
inflammatory metabolites (microbe- mice under germ-free conditions or
associated molecular patterns). Lung cancer when their gut microbiome has been
Another potential connection occurs perturbed with broad-spectrum anti-
An involvement of the respiratory
via estrogen metabolism. Intestinal biotics. The immunogenic cell death
tract microbiota in lung cancer de-
bacteria may affect estrogen expo- that these drugs induce is depen
velopment is conceivable, based dent on inflammatory responses (par-
sure, a major risk factor for breast
on epidemiological studies showing tially mediated by signalling through
cancer (see Chapter 2.11), via acti-
vation (or reactivation) of estrogens bacterial lung infections (including pattern recognition receptors), which
(excreted in conjugated form from pneumonia) to be associated with in mouse models were enhanced by
the liver into the intestine) or dietary lung cancer risk [13]. However, only the administration of specific bacte-
xeno-estrogens [21]. few studies of relatively small scale rial species [6,7].
Clinical studies have found es- have directly investigated this ques-
trogen-dependent and estrogen-in- tion; hence, the evidence on the Immunotherapy
dependent microbiome associations role of the airway microbiota in lung Clinical and preclinical studies have
with breast cancer, but a mechanistic cancer is currently still inconclusive. indicated that the composition of

Fig. 3.10.5. Effects that the gut microbiota can have on cancer treatment. The gut improved efficacy of anti-PD-1
microbiota can influence cancer therapy either directly, via biotransformation of drugs, treatment. Similar effects were ob-
or indirectly, via immune modulation. The indirect effects have recently been found served in mouse tumour models
to play a critical role in response to cancer immunotherapy. Examples of intestinal
upon administration of defined bac-
bacterial genera with immunostimulatory effects are highlighted in dark magenta, and
cancer treatments that are known to be affected by bacteria are highlighted in cyan. terial marker species predictive of
(For more details, see the text and [6,7,20,21].) PD-1 response [6,24,26–28,30].
Allogeneic haematopoietic
stem cell transplantation
Allogeneic haematopoietic stem cell
transplantation can be seen as a
form of immunotherapy that is pri-
marily used to treat various hae-
matological malignancies (and also
immune disorders). Although poten-
tially curative, it is associated with
a range of serious, life-threatening
complications, which include graft-
versus-host disease and systemic
infections. Therefore, several pre-
clinical and clinical studies have ex-
amined whether the gut microbiome
influences relapse or mortality after
allogeneic haematopoietic stem
cell transplantation. They found
that general microbial diversity and
the abundance of specific micro-
the gut microbiota is an important Collectively, these studies es- bial taxa (from within the classes
cause of heterogeneous patient re- tablished that the gut microbiota of Clostridiales, Bacteroidia, and
sponse to cancer immunotherapy, has a systemic effect on the out- Actinobacteria) were prognostic
among several other factors that come of treatments targeting vari- markers of allograft maintenance
determine the cancer immune phe- ous cancers types, including some and survival [31,32].
notype [6,24,25]. that are distal to the gastrointestinal
tract (e.g. melanoma and non-small Probiotics/prebiotics and
These studies have shown that
cell lung cancer). The molecular dietary interventions for
the composition and diversity of a improved cancer therapies?
patient’s gut microbiota (assessed mechanisms through which the
before the start of treatment) are gut microbiota achieves immune The accumulating evidence that gut
activation are still poorly defined. microbes affect cancer therapy has
predictive of the response to im-
Consequently, elucidation of the reinforced interest in microbiome
munotherapy with checkpoint in-
cross-talk between the microbiota modulations that aim to improve
hibitors – primarily targeting the
and innate as well as adaptive im- response rates. Along these lines,
programmed cell death 1 (PD-1)/
munity has become a major re- preclinical studies have found ben-
programmed death ligand 1 (PD-
search focus [3,29]. eficial effects of probiotics (oral
L1) interaction, but also cytotoxic
Clinical studies published to date administration of defined live bac-
T-lymphocyte-associated protein 4 terial strains) on progression-free
have been limited in size (n < 100
(CTLA-4) [6,24,26–28]. In patients in most cases) and only partially survival in mice when administered
with melanoma, renal cell carcino- agree on the gut commensal mark- alone or in combination with immu-
ma, or non-small cell lung cancer, ers for response to immunotherapy. notherapy [7,24,26,30]. However,
the diversity of the gut microbiota However, by examining how the recurrent regulations impede the rap-
was predictive of a favourable prog- sponse phenotype from human paid clinical translation of these find-
nosis and response to immuno- tients can be transferred to animals, ings; strict regulation of probiotics
therapy [26–28]. These data are these studies have provided strong as combination therapies with im-
consistent with clinical observa- data supporting a causal role of gut munotherapeutic treatment modali-
tions that treatment with antibiot- microbes. When the faecal microbi- ties necessitates extensive clinical
ics can compromise the efficacy of ome from patients who responded trials [24].
immunotherapy, presumably due to to immunotherapy was transplanted An attractive alternative may
a dramatic loss of microbiota diver- into mice, the recipients showed be to instead focus on prebiotics
sity [24,26]. slower tumour progression and (dietary compounds that stimulate
226
the growth of certain gut microbial of the gut microbiome on cancer im- damage to commensal microbes, or
clades) or diets that are rationally munosurveillance suggests that ear- directly target pathogenic or carci-
designed to modulate the gut mi- ly interventions aiming to rectify gut nogenic processes with small-mol-
crobiome. These could promote microbiota dysbiosis and to promote ecule inhibitors (e.g. Fusobacterial
microbiota diversity and the expan- microbiota diversity may also help to adhesion proteins, required for
sion of gut commensal taxa that are prevent cancer. These questions are their virulence, or the Clostridial
predictive of therapy response and anticipated to also be addressed in 7α-dehydroxylation pathway, which
progression-free survival relative to prospective cohort studies or direct- results in carcinogenic secondary
those that are associated with non- ly in prospective intervention studies bile acids; see Fig. 3.10.3) [5,7].
response or severe complications aiming to modulate the microbiome. Secondary cancer prevention
[7,24,27,31]. However, these intervention studies strategies based on the microbiome
will have to be sufficiently powered are closer to actual implementation.
to overcome the large inter-individu- Several studies have suggested that
Microbiome-based microbiota alterations in colorectal
al heterogeneity in microbiota com-
approaches to cancer position and response [7]. cancer are characteristic enough
prevention Eradication of H. pylori has to hold promise for non-invasive
The current understanding of micro- proven to be an effective strategy cancer screening (potentially also
bial processes with a causal effect for the prevention of gastric cancer in combination with existing non-
on carcinogenesis is still very incom- [9]. However, studies of more com- invasive tests) [4,13–16]. However,
plete, and this limits primary cancer plex microbial communities have no microbial biomarkers for accu-
prevention approaches targeting the had difficulties to precisely pinpoint rate detection of precancerous co-
CHAPTER 3.10
microbiota. Nevertheless, a few di- cancer-associated bacterial strains lonic lesions (advanced adenomas)
SECTION 3
rections are emerging, and in par- and metabolic processes and to have been discovered yet [13]. Early
ticular approaches that closely inte- establish their carcinogenic effects. microbiome-wide association stud-
grate with nutrition appear promising At least for some tumour types, ies for several other cancer types –
(see Chapter 2.6). Reconsidering for example colorectal cancer, re- although they are of small scale and
dietary recommendations in view of search towards this goal has none- lack independent confirmation – fuel
the emerging knowledge of their di- theless progressed rapidly in the the hope for microbiome-based ear-
rect effects on, and their modulation past 5 years. Growing appreciation ly detection of cancer. Continuing
by, gut microbial metabolism may be of diverse microbial processes with efforts for liver cancer (primary can-
warranted. For instance, increasing potential roles in cancer etiology cer and metastases) and pancreatic
dietary fibre content (beyond current (Fig. 3.10.3) also drives the continu- cancer are particularly promising
recommendations) may help to pre- ing search for specific microbiome [13,33], but all these microbiome-
vent malignancies in a microbiota- modulation strategies. These could based secondary prevention ap-
dependent manner via stimulation either aim to suppress pathogenic proaches will also have to be evalu-
of butyrate production (Fig. 3.10.2) species with narrow-spectrum an- ated in large prospective trials.
[5]. The recently discovered impact tibiotics that minimize collateral
References
1. Lynch SV, Pedersen O (2016). The hu- 4. Tilg H, Adolph TE, Gerner RR, Moschen 7. Zitvogel L, Galluzzi L, Viaud S, Vétizou M,
man intestinal microbiome in health and AR (2018). The intestinal microbio- Daillère R, Merad M, et al. (2015). Cancer
disease. N Engl J Med. 375(24):2369–79. ta in colorectal cancer. Cancer Cell. and the gut microbiota: an unexpected
https://doi.org/10.1056/NEJMra1600266 33(6):954–64. https://doi.org/10.1016/j. link. Sci Transl Med. 7(271):271ps1. https://
PMID:27974040 ccell.2018.03.004 PMID:29657127 doi.org/10.1126/scitranslmed.3010473
PMID:25609166
5. O’Keefe SJ (2016). Diet, microorgan-
2. Garrett WS (2015). Cancer and the mi-
isms and their metabolites, and colon 8. Yu LX, Schwabe RF (2017). The gut
crobiota. Science. 348(6230):80–6.
cancer. Nat Rev Gastroenterol Hepatol. microbiome and liver cancer: mecha-
https://doi.org/10.1126/science.aaa4972
13(12):691–706. https://doi.org/10.1038/ nisms and clinical translation. Nat Rev
PMID:25838377 nrgastro.2016.165 PMID:27848961 Gastroenterol Hepatol. 14(9):527–39.
3. Belkaid Y, Hand TW (2014). Role of the mi- 6. Roy S, Trinchieri G (2017). Microbiota: a https://doi.org/10.1038/nrgastro.2017.72
crobiota in immunity and inflammation. Cell. key orchestrator of cancer therapy. Nat PMID:28676707
157(1):121–41. https://doi.org/10.1016/j.cell. Rev Cancer. 17(5):271–85. https://doi.
2014.03.011 PMID:24679531 org/10.1038/nrc.2017.13 PMID:28303904

9. Wroblewski LE, Peek RM Jr, Wilson KT 18. Geller LT, Barzily-Rokni M, Danino T, Jonas 27. Matson V, Fessler J, Bao R, Chongsuwat
(2010). Helicobacter pylori and gastric OH, Shental N, Nejman D, et al. (2017). T, Zha Y, Alegre ML, et al. (2018). The
cancer: factors that modulate disease Potential role of intratumor bacteria in commensal microbiome is associ-
risk. Clin Microbiol Rev. 23(4):713–39. mediating tumor resistance to the chemo- ated with anti-PD-1 efficacy in meta-
https://doi.org/10.1128/CMR.00011-10 therapeutic drug gemcitabine. Science. static melanoma patients. Science.
PMID:20930071 357(6356):1156–60. https://doi.org/10.1126/ 359(6371):104–8. https://doi.org/10.1126/
science.aah5043 PMID:28912244 science.aao3290 PMID:29302014
10. Gagnaire A, Nadel B, Raoult D, Neefjes
J, Gorvel JP (2017). Collateral dam- 19. Hayes RB, Ahn J, Fan X, Peters BA, Ma 28. Gopalakrishnan V, Spencer CN, Nezi
age: insights into bacterial mecha- Y, Yang L, et al. (2018). Association of oral L, Reuben A, Andrews MC, Karpinets
nisms that predispose host cells to can- microbiome with risk for incident head and TV, et al. (2018). Gut microbiome modu-
cer. Nat Rev Microbiol. 15(2):109–28. neck squamous cell cancer. JAMA Oncol. lates response to anti-PD-1 immuno-
https://doi.org/10.1038/nrmicro.2016.171 4(3):358–65. https://doi.org/10.1001/ therapy in melanoma patients. Science.
jamaoncol.2017.4777 PMID:29327043
science.aan4236 PMID:29097493
11. Gunn JS, Marshall JM, Baker S,
20. Peters BA, Wu J, Pei Z, Yang L, Purdue
Dongol S, Charles RC, Ryan ET (2014). 29. Schirmer M, Smeekens SP, Vlamakis H,
MP, Freedman ND, et al. (2017). Oral
Salmonella chronic carriage: epidemiol- Jaeger M, Oosting M, Franzosa EA, et
microbiome composition reflects pro-
ogy, diagnosis, and gallbladder persis- al. (2016). Linking the human gut micro-
spective risk for esophageal cancers.
tence. Trends Microbiol. 22(11):648–55. biome to inflammatory cytokine produc-
https://doi.org/10.1016/j.tim.2014.06.007 Cancer Res. 77(23):6777–87. https://doi.
tion capacity. Cell. 167(4):1125–36.e8.
PMID:25065707 org/10.1158/0 0 08 - 5 472.CAN -17-129 6
https://doi.org/10.1016/j.cell.2016.10.020
PMID:29196415
PMID:27814509
12. Sze MA, Schloss PD (2018). Leveraging
existing 16S rRNA gene surveys to iden- 21. Fernández MF, Reina-Pérez I, Astorga 30. Sivan A, Corrales L, Hubert N, Williams
tify reproducible biomarkers in individuals JM, Rodríguez-Carrillo A, Plaza-Díaz JB, Aquino-Michaels K, Earley ZM, et
with colorectal tumors. MBio. 9(3):e00630- J, Fontana L (2018). Breast cancer and al. (2015). Commensal Bifidobacterium
18. https://doi.org/10.1128/mBio.00630-18 its relationship with the microbiota. Int J promotes antitumor immunity and fa-
PMID:29871916 Environ Res Public Health. 15(8):1747. cilitates anti-PD-L1 efficacy. Science.
https://doi.org/10.3390/ijerph15081747 350(6264):1084–9. https://doi.org/10.1126/
13. Vogtmann E, Goedert JJ (2016). PMID:30110974 science.aac4255 PMID:26541606
Epidemiologic studies of the human
microbiome and cancer. Br J Cancer. 22. Spanogiannopoulos P, Bess EN, Car 31. Shono Y, van den Brink MRM (2018). Gut
114(3):237–42. https://doi.org/10.1038/ mo dy RN, Turnbaugh PJ (2016). The microbiota injury in allogeneic haema-
bjc.2015.465 PMID:26730578 microbial pharmacists within us: a meta- topoietic stem cell transplantation. Nat
genomic view of xenobiotic metabo- Rev Cancer. 18(5):283–95. https://doi.
14. Thomas AM, Manghi P, Asnicar F, Pasolli
lism. Nat Rev Microbiol. 14(5):273–87. org/10.1038/nrc.2018.10 PMID:29449660
E, Armanini F, Zolfo M, et al. (2019).
https://doi.org/10.1038/nrmicro.2016.17
Metagenomic analysis of colorectal cancer 32. Peled JU, Devlin SM, Staffas A, Lumish
PMID:26972811
datasets identifies cross-cohort microbial M, Khanin R, Littmann ER, et al. (2017).
diagnostic signatures and a link with cho- 23. Lehouritis P, Cummins J, Stanton M, Intestinal microbiota and relapse af-
line degradation. Nat Med. 25(4):667–78. Murphy CT, McCarthy FO, Reid G, et ter hematopoietic-cell transplantation.
https://doi.org/10.1038/s41591-019-0405-7 al. (2015). Local bacteria affect the ef- J Clin Oncol. 35(15):1650–9. https://
PMID:30936548 ficacy of chemotherapeutic drugs. Sci d o i . o r g /10 .12 0 0 /J C O. 2 016 .7 0 . 3 3 4 8
Rep. 5(1):14554. https://doi.org/10.1038/ PMID:28296584
15. Wirbel J, Pyl PT, Kartal E, Zych K, Kashani
A, Milanese A, et al. (2019). Meta-analysis srep14554 PMID:26416623
33. Ren Z, Li A, Jiang J, Zhou L, Yu Z, Lu H,
of fecal metagenomes reveals global mi- 24. Zitvogel L, Ma Y, Raoult D, Kroemer G, et al. (2018). Gut microbiome analysis
crobial signatures that are specific for as a tool towards targeted non-invasive
Gajewski TF (2018). The microbiome in
colorectal cancer. Nat Med. 25(4):679– biomarkers for early hepatocellular car-
cancer immunotherapy: diagnostic tools
89. https://10.1038/s41591-019-0406-6 cinoma. Gut. 68(6):1014–23. https://
and therapeutic strategies. Science.
PMID:30936547 d o i . o r g /10 .113 6 /g u t j n l - 2 017- 315 0 8 4
359(6382):1366–70. https://doi.org/10.1126/
science.aar6918 PMID:29567708 PMID:30045880
16. Zeller G, Tap J, Voigt AY, Sunagawa
S, Kultima JR, Costea PI, et al. (2014).
Potential of fecal microbiota for early-stage 25. Chen DS, Mellman I (2017). Elements of
detection of colorectal cancer. Mol Syst cancer immunity and the cancer-immune
Biol. 10(11):766. https://doi.org/10.15252/ set point. Nature. 541(7637):321–30.
msb.20145645 PMID:25432777 ht t p s: //d o i.o r g /10.10 3 8 /n at u r e213 4 9
PMID:28102259
17. Pushalkar S, Hundeyin M, Daley D,
Zambirinis CP, Kurz E, Mishra A, et al. 26. Routy B, Le Chatelier E, Derosa L, Duong
(2018). The pancreatic cancer microbi- CPM, Alou MT, Daillère R, et al. (2018).
ome promotes oncogenesis by induction Gut microbiome influences efficacy of
of innate and adaptive immune suppres- PD-1-based immunotherapy against epi-
sion. Cancer Discov. 8(4):403–16. https:// thelial tumors. Science. 359(6371):91–7.
doi.org/10.1158/2159-8290.CD-17-1134 https://doi.org/10.1126/science.aan3706
PMID:29567829 PMID:29097494
228
3.11
Identifying carcinogens
from 10 key characteristics
A new approach based on mechanisms
Martyn T. Smith Gloria M. Calaf (reviewer)
Kathryn Z. Guyton John D. Groopman (reviewer)
●● The key characteristics ap- (EPA) have been formally evalu-

SUMMARY proach can inform the design of ated by the United States National
high-throughput testing systems Toxicology Program (NTP) [1] or
●● The key characteristics of hu- and human biomarker studies other national testing programmes
man carcinogens were recently with greater relevance to cancer (e.g. the Japan Bioassay Research
CHAPTER 3.11
SECTION 3
introduced as the basis for a hazard identification – the first Center of the Japan Organization of
uniform approach to evaluating step in cancer prevention. Occupational Health and Safety). In
mechanistic evidence to support contrast, data on carcinogen mech-
cancer hazard identification. anisms from human biomarker stud-
●● The key characteristics reflect The IARC Monographs programme ies, in vivo animal tests, and in vitro
the chemical and biological identifies the causes of human can- cell culture models are increasing in
properties of established hu- cer, based on the systematic as- both volume and diversity [2–5].
sembly, review, and integration of When the evidence from human
man carcinogens, including
evidence of cancer in humans, can- epidemiological studies is less than
“is genotoxic”, “is immunosup-
cer in experimental animals, and sufficient, strong mechanistic data
pressive”, and “modulates re-
carcinogen mechanisms. Of the ap- can play a pivotal role in the over-
ceptor-mediated effects”. The
proximately 120 agents classified all carcinogen hazard classifica-
key characteristics are distinct
by the IARC Monographs as carci- tion [6]. For instance, even though
from the hallmarks of cancer,
nogenic to humans (Group 1), most the evidence from rodent cancer
which relate to the properties
have sufficient evidence of carci- bioassays provided sufficient evi-
of cancer cells.
nogenicity in humans, based on dence of carcinogenicity in experi-
●● The key characteristics ap- epidemiological studies. However, mental animals, d -limonene was
proach avoids a narrow focus epidemiological studies of cancer in categorized as not classifiable as
on specific pathways and hy- exposed humans are often limited in to its carcinogenicity to humans
potheses and provides for a number, and may have deficiencies (Group 3) on the basis of mecha-
broad, holistic consideration in terms of sample size, confound- nistic and other relevant data, be-
of the mechanistic evidence. ing, and exposure characterization. cause the probable mechanism
Therefore, data on the key Furthermore, for chemicals that of carcinogenicity in experimental
characteristics can provide have recently been introduced on animals was unlikely to operate in
independent evidence of car- the market, epidemiological studies humans. Other agents have been
cinogenicity when data from may not exist or may not be relevant, classified as probably carcinogenic
studies in humans are lacking, because of the long latency period to humans (Group 2A) or even as
and can help in establishing for cancer development. The num- carcinogenic to humans (Group 1)
biological plausibility. ber of lifetime rodent cancer bio- based on strong evidence for rec-
assays being performed is declin- ognized carcinogen mechanisms,
●● The key characteristics ap- ing, and only a fraction of the ap- such as genotoxicity (for ethylene
proach is being increasingly proximately 75 000 chemicals that oxide), inhibiting DNA repair (for
applied by agencies throughout are listed in the Toxic Substances etoposide), or binding to the aryl hy-
the world, and key characteris- Control Act Chemical Substance drocarbon receptor and subsequent
tics for other toxicological haz- Inventory of the United States downstream effects (for 2,3,7,8-tet-
ards are being developed. Environmental Protection Agency rachlorodibenzo-para-dioxin).
Chapter 3.11 • Identifying carcinogens from 10 key characteristics 229

A recent review of all the agents To address these challenges,

classified as carcinogenic to hu- the key characteristics of human
mans (Group 1) in IARC Mono carcinogens were recently intro- FUNDAMENTALS
graphs Volumes 1–99 revealed duced as the basis for a uniform
■■ The biological mechanisms
several issues relevant to improving approach to searching for, orga-
the evaluation of mechanistic data nizing, and evaluating mechanistic by which certain chemicals,
for carcinogen hazard identification evidence to support cancer hazard some types of radiation, and
[7]. First, many human carcinogens identification [7]. The key charac- some infectious agents cause
show a number of characteristics teristics comprise the properties of cancer in humans have been
that are shared among carcinogen- known human carcinogens. These intensively investigated.
ic agents. Second, different human characteristics are distinct from the
carcinogens may exhibit a different hallmarks of cancer, which relate to ■■ For chemical carcinogens, no
spectrum of these key character- the properties of cancer cells (see single sequence of biological
istics and operate through distinct Chapter 3.1) [9,10]; instead, they events is evident for all such
mechanisms. Third, for many car- reflect the chemical and biological
agents.
cinogens evaluated before Volume properties of cancer-causing agents
100 of the IARC Monographs, few (see Table 3.11.1). Established hu- ■■ Studies in experimental
data were available on some mech- man carcinogens commonly exhibit animals have established
anisms of recognized importance in one or more of these characteris-
that some classes of organic
carcinogenesis, such as epigenetic tics. Therefore, data on these char-
acteristics can provide independent compounds include multiple
alterations (see Chapter 3.8) [8].
Fourth, the evaluation of mechanis- evidence of carcinogenicity when carcinogens, and such
tic and other relevant data has been data from studies in humans are agents are metabolized in
further challenged by the lack of a lacking. Data on key characteristics mammalian tissue, causing
systematic and transparent method can also help in interpreting the rel- mutations as a result of
of searching for and assembling evance and importance of findings binding of these agents to
mechanistic data for cancer haz- of cancer in experimental animals
DNA. These carcinogens are
ard identification. Specifically, there and in humans.
was no widely accepted method to This chapter describes the key described as genotoxic.
systematically search for relevant characteristics and discusses their ■■ The distribution of cancer
mechanisms, and this resulted in a application in IARC Monographs
in humans has implicated
lack of uniformity in the mechanistic evaluations that have taken advan-
topics addressed across assess- tage of the systematic considera- a variety of inorganic and/
ments. Finally, there was no pro- tion of mechanistic evidence. The or naturally occurring
cedure to efficiently organize, ana- strengths and the weaknesses of compounds, including
lyse, and interpret the voluminous this approach are discussed, as are asbestos, as well as
data from mechanistic studies. opportunities for further progress immunosuppressive drugs,
which are not characterized
as genotoxic.
Table 3.11.1. Key characteristics of carcinogens
■■ For decades, mechanisms
1. Is electrophilic or can be metabolically activated to electrophiles of carcinogenesis involved
a primary reference to
2. Is genotoxic
genotoxicity, with binding
3. Alters DNA repair or causes genomic instability to critical protein receptors
being common to many non-
4. Induces epigenetic alterations
genotoxic carcinogens.
5. Induces oxidative stress
■■ The recent description of
6. Induces chronic inflammation certain key characteristics,
one or more of which is
7. Is immunosuppressive
exhibited by all established
8. Modulates receptor-mediated effects human carcinogens, is an
9. Causes immortalization
innovative approach to
identifying carcinogens.
10. Alters cell proliferation, cell death, or nutrient supply
230
and refinement. The last section of metabolic activation [12]. The ability cesses that can be usurped during
the chapter further discusses how to form adducts with nucleic acids the carcinogenic process, with im-
the paradigm could be expanded and proteins is a common property pacts on gene expression and DNA
to other end-points and how fu- of these inherently electrophilic repair dynamics [8]. A wide range
ture toxicological and molecular and/or metabolically activated hu- of carcinogens have been shown to
epidemiological studies could be man carcinogens [13]. dysregulate the epigenome [18].
developed to generate more useful
information for the process of car- Characteristic 2: Is genotoxic Characteristic 5: Induces
cinogen evaluation. A genotoxic agent induces dam- oxidative stress
age to a cell’s genetic material (see Many carcinogens are capable of
Descriptions of the Chapter 3.2). Examples of DNA influencing redox balance within
damage include DNA strand breaks target cells. If an imbalance occurs,
key characteristics of (breaks in the phosphodiester favouring the formation of reactive
carcinogens bonds), protein–DNA cross-links, oxygen species at the expense of
The number of ways in which and oxidative damage to DNA. their detoxification, this is referred
agents contribute to carcinogenesis Genotoxic agents may also induce to as oxidative stress. This may be
can be extensive. However, these damage at the chromosomal level, accompanied by the production of
mechanisms can be grouped into a including chromosomal aberra- reactive nitrogen species, or nitra-
limited number of categories (geno- tions, micronuclei, sister chroma- tive stress. Oxidative stress can
toxicity, immunosuppression, etc.). tid exchanges, and aneuploidy. A lead to the generation of mutations
Guyton et al. described 15 types of mutation, which is a change in the in DNA, and more than 100 differ-
CHAPTER 3.11
“key events” associated with human
SECTION 3
DNA sequence, usually arises as ent types of oxidative damage to
carcinogens that collectively rep- the cell attempts to repair the DNA DNA have been identified [19]. The
resented many carcinogen mecha- damage [14]. A large proportion of induction of oxidative stress and
nisms [1]. As part of its review of the agents classified by IARC in subsequent injury is a character-
the agents classified in Group 1, Group 1 are genotoxic. istic of a diverse group of carcino-
IARC convened two meetings in gens, including radiation, asbestos,
2012 to review mechanisms of es- Characteristic 3: Alters DNA chemicals, and carcinogenic infec-
tablished human carcinogens. At repair or causes genomic tious agents.
the first of the meetings, 24 mecha- instability
nistic end-points were identified. Carcinogens may act not only by Characteristic 6: Induces
However, these were considered producing DNA damage directly but chronic inflammation
too impractical as a guide for cat- also by altering the processes that Chronic inflammation from persis-
egorizing the evidence on carcino- control normal DNA replication or tent infections, such as that caused
gen mechanisms. Therefore, at the repair of DNA damage (see Chapter by Helicobacter pylori, has been as-
second meeting, these end-points 3.4). Examples include the inhibi- sociated with several forms of cancer
were merged into 10 categories. tion of DNA repair by cadmium [15] (see Chapter 3.5) [20]. Various other
The 10 key characteristics listed in and formaldehyde [16]. In cells ex- carcinogens also induce chronic
Table 3.11.1 represent the majority posed to ionizing radiation, genetic inflammation, including fibres (e.g.
of the chemical and biological prop- instability is a relatively late-occur- silica, asbestos) and chemicals (e.g.
erties of human carcinogens, as ring event that appears several cell polychlorinated biphenyls) [7].
described below and in more detail generations after irradiation and re-
elsewhere [7]. sults in a reduced ability to replicate Characteristic 7: Is
the genotype faithfully [17]. immunosuppressive
Characteristic 1: Is electro- Immunosuppression is a reduction
philic or can be metabolically Characteristic 4: Induces in the capacity of the immune sys-
activated to electrophiles epigenetic alterations tem to respond effectively to foreign
Electrophiles are electron-seeking The term “epigenetic” refers to sta- antigens, including antigens on tu-
molecules that form addition prod- ble changes in gene expression mour cells. Persistent immunosup-
ucts, commonly referred to as ad- and chromatin organization that are pression presents a risk of cancer
ducts, with cellular macromolecules not caused by changes in the DNA (see Chapter 3.9), especially ex-
including DNA, RNA, lipids, and sequence itself and can be inherit- cess risk of lymphoma. Several
proteins (see Chapter 3.3). Some ed over cell divisions [8]. Epigenetic carcinogens act entirely or largely
chemical carcinogens (e.g. sulfur phenomena – including changes in by immunosuppression, often in
mustard) are direct-acting electro- the DNA methylome, in chromatin concert with oncogenic infectious
philes, whereas others (e.g. afla- compaction states, and in histone agents. The Group 1 agents that
toxins, benzene) require chemical modification – are important as- act by immunosuppression include
conversion within the body [11] or pects of normal developmental pro- HIV-1 and the immunosuppressive

drug ciclosporin (also known as cy- Characteristic 10: Alters cell result of an impaired nutrient sup-
closporine) [21]. proliferation, cell death, or ply. The number of neoplastic cells
nutrient supply can increase exponentially, quickly
Characteristic 8: Modulates outstripping the supply capabilities
A component common to many
receptor-mediated effects of the existing tissue vasculature.
types of cancer is the evasion of
All actions of hormonally active Neo-angiogenesis, in which new
programmed cell death, via apopto-
agents are mediated by their abil- blood vessels grow into a tumour,
sis, or of other terminal program-
ity to interact with a receptor, with is key to providing a supply of nu-
ming, including autophagy, in at
the hormone acting as an endog- trients. Thus, agents that promote
least a proportion of the cell popu-
enous ligand (see Chapter 2.11). or inhibit angiogenesis, such as ar-
lation [25]. In contrast to apoptosis
For a chemical to interfere with senic, will promote or delay tumour
and autophagy, necrotic cell death
hormone signalling and produce growth [28,29].
releases pro-inflammatory signals
adverse effects, it must ultimately
into the surrounding tissue, which
interfere with hormone receptor
can enhance cancer cell prolifera- Using the key
activation – either directly or indi- characteristics to
tion and promote cancer metastasis
rectly. Numerous carcinogens act
as ligands to receptor proteins, in-
[26,27]. Many agents affect necro- identify carcinogens
sis, apoptosis, and/or autophagy, Recently, Guyton et al. [30] reviewed
cluding hormone replacement ther-
apy and 2,3,7,8-tetrachlorodibenzo- and they can have profoundly diver- the feasibility and the limitations of
para-dioxin. Many exogenous agents gent effects on cancer induction in applying the 10 key characteristics
act directly as agonists or antago- different tissues. of carcinogens to comprehensively
nists by competing for binding In addition to cell death caused search for, screen, and evaluate
with the endogenous ligand (e.g. directly by the toxicity of an agent, mechanistic evidence in cancer haz-
a hormone, such as testosterone). cells within a tumour may die as a ard identification. The methods and
However, there are also receptors
for which few or no endogenous
Fig. 3.11.1. Typically, various classes of oncogenic viruses mediate immortalization of
ligands have been identified, such
target cells, such that cell proliferation continues indefinitely, rather than being constrained
as the aryl hydrocarbon receptor by shortening of telomeres [7]. Telomeres are unique DNA sequences located at the
[22,23]; in these cases, the carci- ends of chromosomes, as visualized by red fluorescence in this micrograph.
nogenic chemical is the activating
ligand. Carcinogens may also act
indirectly on receptor-mediated ef-
fects by altering the bioavailability
of endogenous ligands by affecting
the biosynthesis, bioactivation, and/
or degradation of the ligand. These
direct and indirect effects all modu-
late receptor-based regulation of
gene transcription, and ultimately
cell growth and proliferation.
Characteristic 9: Causes
immortalization
Several human DNA and RNA vi-
ruses are carcinogenic to humans.
Although oncogenic viruses belong
to different families, their strategies
in human cancer development show
many similarities and involve viral-
encoded oncoproteins targeting the
key cellular proteins that regulate
cell growth [24]. These targets may
include important tumour suppres-
sor genes and/or oncogenes. The
result of these viral effects is to im-
mortalize the cells of the target tis-
sue such that they divide continu-
ously (see Chapter 3.1).
232
results of mechanistic data evalua- ation to classify two agents in utility of the key characteristics
tions were compiled from eight re- Group 2A: tetrabromobisphenol A approach is underscored by the
cent IARC Monographs meetings and tetrachloroazobenzene, both of fact that it is being increasingly ap-
in which expert Working Groups which modulate receptor-mediated plied by agencies throughout the
classified 34 diverse chemicals and effects in combination with other world, including at the EPA and the
complex exposures into Group 1, key characteristics. Fewer studies NTP Report on Carcinogens in the
Group 2A, Group 2B (possibly carci- were available for the 17 agents USA. In parallel, key characteris-
nogenic to humans), or Group 3. For classified in Group 2B or Group 3, tics for other toxicological hazards
these evaluations, the key charac- and only one agent classified in are being developed, in line with
teristics served as the basis for tar- Group 2B (1-bromopropane) had the recommendations of the re-
geted literature searches to identify strong evidence of more than one port Using 21st Century Science to
published mechanistic studies, and key characteristic. Thus, this ob- Improve Risk-Related Evaluations
the Health Assessment Workplace jective approach to identify and [31], which recognized that the key
Collaborative (https://HAWCproject. evaluate mechanistic studies re- characteristics approach “avoids a
org) was used to record the litera- vealed strong evidence for multiple narrow focus on specific pathways
ture search terms, sources, articles key characteristics for most agents and hypotheses and provides for a
retrieved, exclusion criteria, and cat- classified in Group 1 or Group 2A, broad, holistic consideration of the
egorization of included articles. but it also identified opportunities mechanistic evidence”. Thus, the
As illustrated by the resulting for improvement. Specifically, fur- key characteristics approach can
literature flow diagram for penta- ther development and mapping of aid in preventing bias and misinter-
chlorophenol (Fig. 3.11.2), a broad toxicological and biomarker end- pretation, even when disproportion-
CHAPTER 3.11
SECTION 3
literature encompassing multiple points and pathways relevant to the ate resources have been focused
key characteristics was identified key characteristics could advance on investigating a favoured mecha-
for most of the 16 carcinogens the systematic search for and eval- nism [6]. In contrast, focusing on
classified in Group 1 or Group 2A uation of mechanistic data in car- hypothesized modes of action or
at those eight IARC Monographs cinogen hazard identification. adverse outcome pathways can
meetings. Mechanistic data were Notwithstanding the opportuni- result in exclusion of data, leading
used as part of the overall evalu- ties for further development, the to analyses that favour a particular
Fig. 3.11.2. Literature flow diagram for pentachlorophenol (classified in Group 1 by the IARC Monographs in Volume 117) illustrates
the results of the search, screening, and organization of the published scientific literature, according to the key characteristics and
other topics relevant to the evaluation of mechanistic data.

viewpoint. As a related challenge, in exposed humans were avail- limitations underscore the need for
hypotheses are inherently limited able. Especially when mechanistic a testing battery with greater rel-
by the current understanding of data are sparse, high-throughput evance to cancer hazard identifica-
the disease process and may be testing systems such as the EPA’s tion – perhaps a Carcinogenicity
shown to be incorrect or incomplete Toxicity Forecaster (ToxCast) and Forecaster (CarciCast). In parallel,
as biological knowledge develops the NTP’s Toxicology in the 21st the report Applications of Toxico-
[1]. This limitation was recognized Century (Tox21) can aid as an ad- genomic Technologies to Predictive
by Hill [32], who noted that “what ditional or supportive source of Toxicology and Risk Assessment
is biologically plausible depends mechanistic data [30]. However, the [2] has encouraged human bio-
upon the biological knowledge of experience of applying an approach marker studies to improve hazard
the day”. based on key characteristics to the prediction; end-points related to the
The experience of applying the mechanistic data stream, as further key characteristics could be applied
key characteristics approach for elaborated by Chiu et al. [33], dem- in such studies to better forecast
34 sequentially evaluated chemi- onstrated the usefulness of high- carcinogenic activity in humans [3].
cals and complex exposures in throughput testing systems for the In summary, the application of the
the IARC Monographs has clearly key characteristic “modulates re- key characteristics to cancer haz-
revealed the variable extent of the ceptor-mediated effects” while also ard identification is a robust new
mechanistic information available, revealing significant gaps in their approach that complements other
even for carcinogens with wide- coverage for most other key char- efforts to advance identification of
spread human exposures [30]. acteristics. These and other chal- the causes of human cancer – the
Moreover, for most agents, few lenges have hampered carcino- first step in cancer prevention.
studies of biomarker end-points genicity prediction, which remains
relevant to the key characteristics imprecise [1,34]. Together, these
References
1. Guyton KZ, Kyle AD, Aubrecht J, Cogliano 6. IARC (2019). Preamble to the IARC 12. Miller JA (1970). Carcinogenesis by
VJ, Eastmond DA, Jackson M, et al. (2009). Monographs, amended January 2019. chemicals: an overview – G. H. A.
Improving prediction of chemical carcino- Available from: https://monographs.iarc. Clowes Memorial Lecture. Cancer Res.
genicity by considering multiple mecha- fr/preamble-to-the-iarc-monographs/. 30(3):559–76. PMID:4915745
nisms and applying toxicogenomic ap-
7. Smith MT, Guyton KZ, Gibbons CF, Fritz 13. Ehrenberg L (1984). Covalent binding of ge-
proaches. Mutat Res. 681(2–3):230–40.
JM, Portier CJ, Rusyn I, et al. (2016). Key notoxic agents to proteins and nucleic acids.
https://doi.org/10.1016/j.mrrev.2008.10.001
characteristics of carcinogens as a basis In: Berlin A, Draper M, Hemminki K, Vainio
PMID:19010444
for organizing data on mechanisms of H, editors. Monitoring human exposure to
2. National Research Council (US) Commit- carcinogenesis. Environ Health Perspect. carcinogenic and mutagenic agents. Lyon,
tee on Applications of Toxicogenomic Tech- 124(6):713–21. https://doi.org/10.1289/ France: International Agency for Research
nologies to Predictive Toxicology (2007). ehp.1509912 PMID:26600562 on Cancer (IARC Scientific Publications, No.
Applications of toxicogenomic technologies 59); pp. 107–14. PMID:6545273
8. Herceg Z, Lambert M-P, van Veldhoven
to predictive toxicology and risk assess- K, Demetriou C, Vineis P, Smith MT, et
ment. Washington (DC), USA: National 14. Shaughnessy DT, DeMarini DM (2009).
al. (2013). Towards incorporating epige- Types and consequences of DNA
Academies Press. netic mechanisms into carcinogen iden- damage. In: Knasmüller S, DeMarini
3. Fielden MR, Ward LD, Minocherhomji S, tification and evaluation. Carcinogenesis. DM, Johnson I, Gerhäuser C, editors.
Nioi P, Lebrec H, Jacobson-Kram D (2018). 34(9):1955–67. https://doi.org/10.1093/ Chemoprevention of cancer and DNA
Modernizing human cancer risk assess- carcin/bgt212 PMID:23749751 damage by dietary factors. Weinheim,
ment of therapeutics. Trends Pharmacol 9. Hanahan D, Weinberg RA (2011). Hallmarks Germany: Wiley-VCH; pp. 21–33.
Sci. 39(3):232–47. https://doi.org/10.1016/j. of cancer: the next generation. Cell.
tips.2017.11.005 PMID:29242029 15. Candéias S, Pons B, Viau M, Caillat S,
144(5):646–74. https://doi.org/10.1016/j.
Sauvaigo S (2010). Direct inhibition of
cell.2011.02.013 PMID:21376230
4. Tice RR, Austin CP, Kavlock RJ, Bucher excision/synthesis DNA repair activities
JR (2013). Improving the human haz- 10. Hanahan D, Weinberg RA (2000). The hall- by cadmium: analysis on dedicated bio-
ard characterization of chemicals: a marks of cancer. Cell. 100(1):57–70. https:// chips. Mutat Res. 694(1–2):53–9. https://
Tox21 update. Environ Health Perspect. doi.org/10.1016/S0092-8674(00)81683-9 doi.org/10.1016/j.mrfmmm.2010.10.001
121(7):756–65. https://doi.org/10.1289/ PMID:10647931 PMID:20969882
ehp.1205784 PMID:23603828
11. Salnikow K, Zhitkovich A (2008). Genetic 16. Luch A, Frey FCC, Meier R, Fei J, Naegeli
5. Collins FS, Gray GM, Bucher JR and epigenetic mechanisms in metal H (2014). Low-dose formaldehyde delays
(2008). Toxicology. Transforming envi- carcinogenesis and cocarcinogenesis: DNA damage recognition and DNA exci-
ronmental health protection. Science. nickel, arsenic, and chromium. Chem sion repair in human cells. PLoS One.
319(5865):906–7. https://doi.org/10.1126/ Res Toxicol. 21(1):28–44. https://doi. 9(4):e94149. https://doi.org/10.1371/journal.
science.1154619 PMID:18276874 org/10.1021/tx700198a PMID:17970581 pone.0094149 PMID:24722772
234
17. Kadhim M, Salomaa S, Wright E, 23. Baek SH, Kim KI (2014). Emerging roles 29. Wang F, Liu S, Xi S, Yan L, Wang H, Song
Hildebrandt G, Belyakov OV, Prise KM, et of orphan nuclear receptors in can- Y, et al. (2013). Arsenic induces the ex-
al. (2013). Non-targeted effects of ionis- cer. Annu Rev Physiol. 76(1):177–95. pressions of angiogenesis-related factors
ing radiation – implications for low dose https://doi.org/10.1146/annurev-physiol- through PI3K and MAPK pathways in SV-
risk. Mutat Res. 752(2):84–98. https:// 030212-183758 PMID:24215441 HUC-1 human uroepithelial cells. Toxicol
d o i.o r g /10.1016 / j.m r r ev. 2 012 .12 . 0 01 Lett. 222(3):303–11. https://doi.org/10.1016/j.
PMID:23262375 24. Saha A, Kaul R, Murakami M, Robertson toxlet.2013.08.008 PMID:23968725
ES (2010). Tumor viruses and cancer bi-
18. Pogribny IP, Rusyn I (2013). Environmental ology: modulating signaling pathways for 30. Guyton KZ, Rusyn I, Chiu WA, Corpet DE,
toxicants, epigenetics, and cancer. Adv therapeutic intervention. Cancer Biol Ther. van den Berg M, Ross MK, et al. (2018).
Exp Med Biol. 754:215–32. https://doi. 10(10):961–78. https://doi.org/10.4161/ Application of the key characteristics of car-
o r g /10 .10 0 7/ 9 7 8 -1 - 4 419 - 9 9 6 7- 2 _11 cbt.10.10.13923 PMID:21084867 cinogens in cancer hazard identification.
PMID:22956504 Carcinogenesis. 39(4):614–22. https://doi.
25. Ryter SW, Mizumura K, Choi AMK (2014). org/10.1093/carcin/bgy031 PMID:29562322
19. Klaunig JE, Wang Z, Pu X, Zhou S The impact of autophagy on cell death
(2011). Oxidative stress and oxidative modalities. Int J Cell Biol. 2014:502676. 31. National Academies of Sciences,
damage in chemical carcinogenesis. ht t p s: //d o i.o r g /10.115 5 / 2 014 / 5 0 2 676 Engineering, and Medicine (2017). Using
Toxicol Appl Pharmacol. 254(2):86–99. PMID:24639873 21st century science to improve risk-relat-
https://doi.org/10.1016/j.taap.2009.11.028 ed evaluations. Washington (DC), USA:
PMID:21296097 26. Pollard JW (2008). Macrophages define National Academies Press. Available
the invasive microenvironment in breast from: https://doi.org/10.17226/24635.
20. Grivennikov SI, Greten FR, Karin M (2010). cancer. J Leukoc Biol. 84(3):623–30.
Immunity, inflammation, and cancer. Cell. h t t p s : / / d o i . o r g / 10 .118 9 / j l b .110 7 7 6 2 32. Hill AB (1965). The environment and dis-
140(6):883–99. https://doi.org/10.1016/j. PMID:18467655 ease: association or causation? Proc R
cell.2010.01.025 PMID:20303878 Soc Med. 58:295–300. PMID:14283879
27. Coussens LM, Zitvogel L, Palucka AK
21. Rafferty P, Egenolf D, Brosnan K, (2013). Neutralizing tumor-promoting chron- 33. Chiu WA, Guyton KZ, Martin MT, Reif DM,
Makropoulos D, Jordan J, Meshaw K, et ic inflammation: a magic bullet? Science. Rusyn I (2018). Use of high-throughput
al. (2012). Immunotoxicologic effects of 339(6117):286–91. https://doi.org/10.1126/ in vitro toxicity screening data in cancer
CHAPTER 3.11
SECTION 3
cyclosporine on tumor progression in mod- science.1232227 PMID:23329041 hazard evaluations by IARC Monograph
els of squamous cell carcinoma and B-cell Working Groups. ALTEX. 35(1):51–64.
lymphoma in C3H mice. J Immunotoxicol. 28. Yang MH, Zang Y-S, Huang H, Chen K, https://doi.org/10.14573/altex.1703231
9(1):43–55. https://doi.org/10.3109/15476 Li B, Sun G-Y, et al. (2014). Arsenic triox- PMID:28738424
91X.2011.614646 PMID:22299716 ide exerts anti-lung cancer activity by in-
hibiting angiogenesis. Curr Cancer Drug 34. Rusyn I, Sedykh A, Low Y, Guyton KZ,
22. Ma Q (2011). Influence of light on aryl hy- Targets. 14(6):557–66. https://doi.org/ Tropsha A (2012). Predictive modeling of
drocarbon receptor signaling and con- 10.2174/1568009614666140725090000 chemical hazard by integrating numerical
sequences in drug metabolism, physiol- PMID:25088040 descriptors of chemical structures and
ogy and disease. Expert Opin Drug Metab short-term toxicity assay data. Toxicol
Toxicol. 7(10):1267–93. https://doi.org/10.15 Sci. 127(1):1–9. https://doi.org/10.1093/
17/17425255.2011.614947 PMID:21883026 toxsci/kfs095 PMID:22387746

The IARC Handbooks of Cancer Prevention
Béatrice Lauby-Secretan
The IARC Handbooks of Cancer referred to as preventive therapy; Monographs Preamble in 2006. The
Prevention series was launched see Chapter 6.4); the scope was later Handbooks programme undertook
in 1995 to complement the IARC enlarged to cover evaluation of other a formal update by convening an
Monographs series. The purpose of types of preventive interventions, in- Advisory Group at IARC in February
the IARC Handbooks is to evaluate cluding primary prevention and can- 2019. The Working Procedures are
scientific evidence on agents and cer screening. So far, the Handbooks now referred to as the Preambles.
interventions that may reduce the in- have covered cancer-preventive Planned future Handbooks in-
cidence of or mortality from cancer. agents, including non-steroidal anti- clude evaluations of screening for
The Handbooks assist national inflammatory drugs (such as aspirin), cervical cancer (updating Volume
and international authorities in as- vitamin A, carotenoids, and retinoids, 10, published in 2005) and oral cav-
sessing the benefits and risks of a preventive actions (e.g. use of sun- ity cancer (first-time evaluation).
particular intervention and in devis- screens, absence of excess body The IARC Handbooks of Cancer
ing programmes of health promotion fatness, physical activity, and con- Prevention have had a broad impact
and cancer prevention. There is a sumption of fruit and vegetables), on guidelines, public recommenda-
major demand worldwide for such screening (for breast cancer, cervical tions, and implementation of health
evaluations in order to improve pub- cancer, and colorectal cancer), and strategies, including the following:
lic health. IARC is ideally placed to the efficacy of tobacco control mea- ●● Numerous national health agen-
respond to this demand, because of sures (reversal of risk after quitting cies (including those of Australia,
its expertise, experience, reputation, smoking, smoke-free policies, and Canada, New Zealand, the
and independence. tax and price policies). United Kingdom, and the USA),
The principles, procedures, and After a 5-year hiatus due to re- the European Committees,
scientific criteria that guide the IARC structuring and financial restric- and offices of the World Health
Handbooks evaluations closely mir- tions, the Handbooks series was Organization have used the IARC
ror those of the IARC Monographs: relaunched in 2014. The first in the Handbooks as a basis for devel-
interdisciplinary Working Groups of new series, Volume 15, was a re- oping their public health strate-
experts review the published stud- assessment of breast cancer screen- gies and guidelines.
ies and evaluate the weight of evi- ing (updating Volume 7, published in ●● Both Handbooks on breast cancer
dence on the effectiveness of pri- 2002). Volume 16 dealt with a pre- screening (Volume 7 and Volume
mary and secondary interventions ventive action, absence of excess 15) have triggered national mea-
to prevent cancer. The full evalua- body fatness (updating Volume 6, sures to implement programmes
tions are then published in a volume published in 2002), and Volume 17 or update guidelines.
of the Handbooks series, and a was a first-time evaluation of colorec- ●● After the publication of the Hand-
summary is published as a Special tal cancer screening. books on tobacco control (Vol-
Report in a leading scientific jour- At the time of the relaunch, the umes 11–14), IARC was invited
nal, currently The New England original Working Procedures were to report to the Conference
Journal of Medicine. revised in accordance with devel- of the Parties to the World
The Handbooks were origi- opments in the Monographs pro- Health Organization Framework
nally developed for the evaluation gramme, incorporating many of the Convention on Tobacco Control.
of chemopreventive agents (now elements from the update to the
236
4 Inequalities
that affect
cancer prevention
This is the first time that a section primar- burden of cancer. The relevant factors may
ily concerned with inequalities and cancer be specific to particular countries or regions.
is being included in a World Cancer Report. Recently, there have been improvements in
Inequalities that affect cancer prevention in- the methods for investigating associations
clude those determined by educational at- between inequalities and cancer as well as
tainment and by limitations on circumstances; the ways in which adverse outcomes may be
examples are nutrition and housing, which minimized. Typically, data are available on
are determined by financial income. Such in- variations within a particular country, and the
equalities may perturb the efficacy of almost chapters in this section describe such data for
all initiatives that are aimed at reducing the certain countries.
4.1
Inequalities between and within

countries
Impact on cancer prevention
Salvatore Vaccarella David I. Conway (reviewer)
Johan P. Mackenbach Diana Sarfati (reviewer)
Paolo Vineis (reviewer)
●● Preventive policies, such as Inequalities between social

SUMMARY elimination of occupational ex- groups are observed in every
posure to carcinogens, tobacco country, whether it is a high-, mid-
●● On average, the incidence rates control measures, vaccination dle-, or low-income country. Such
for all cancers combined, in against cancer-causing infec- social inequalities may arise from
both sexes, increase with in- tious agents, and screening for the various dimensions that make
creasing levels of national so- early stages of cancer, are po- up the structure of society, includ-
cioeconomic development: the tentially powerful ways to reduce ing socioeconomic position, race
highest-income countries have not only the average incidence and ethnicity, area of residence,
much higher rates than the low- of and mortality from cancer but sex, and sexual orientation, among
est-income countries. In con- also socioeconomic inequalities others. Despite these complexities,
trast, for the mortality rates for in cancer occurrence. cancer disproportionately affects
all cancers combined, no clear
●● The low budget allocated to the most disadvantaged individuals
gradient is observed with aver-
cancer prevention contrasts and groups.
age levels of national socioeco-
with the large investments made Of all the potentially relevant
nomic development.
in the development of advanced dimensions of social inequalities
●● Within countries, the socioeco- technological devices and pre- within countries, this chapter fo-
nomic gradient for cancer inci- cision medicine, which may, in cuses mainly on the socioeconomic
dence may vary in magnitude some cases, increase social in- dimension. Socioeconomic factors
and direction across different equalities in cancer. shape the environments in which
cancer sites, but cancer mortal- individuals live as well as the dis-
ity is often higher, and cancer tribution of resources and services,
survival lower, in groups with Inequalities in cancer are the sys- and could therefore be considered
low socioeconomic position and tematic differences in cancer oc- the “causes of the causes” of dis-
other disadvantaged groups currence (i.e. in cancer incidence, eases such as cancer [1].
(e.g. ethnic and racial minorities mortality, and survival) that ex- Social factors may have a very
and Indigenous populations), for ist between and within countries. different impact on different cancer
cancer overall and for the large Cancer inequalities are driven types and on different steps along
majority of cancer types. by the interplay of many factors, the cancer continuum, from the
●● Individuals with higher socio- which largely reflect the cultures time of an individual’s exposure to
economic position tend to ben- and environments in which people a carcinogenic agent to early diag-
efit more from cancer preven- are born, live, and work, as well as nosis, treatment, and survival [2–7].
tion interventions and to have the uneven distribution of resourc- Some cancer types are related to
earlier detection and diagnosis es and services between and with- social conditions during childhood,
and better treatment, because in countries. Inequalities in cancer whereas others are more closely
they have better access to between countries may be due to related to circumstances during
health-care services, greater a combination of contextual fac- adult life. Multiple pathways are
health literacy, and fewer fi- tors – such as culture, geography, involved, resulting in differential
nancial barriers to health care politics, policies, societal struc- exposures to proximal risk factors,
compared with individuals with ture, and economic structure – such as tobacco smoking, alcohol
lower socioeconomic position. and individual factors. consumption, unhealthy diet, and
238
occupational exposures, and in dif- sions, or even opposite trends, both

ferences in access to health-care types of measures should be moni- FUNDAMENTALS
services. Therefore, different pro- tored when describing trends in so-
files of cancer types are often ob- cioeconomic inequalities in cancer ■■ Major differences in cancer
served in groups of individuals and and when assessing interventions occurrence exist between
in countries with different socioeco- aimed at reducing socioeconomic countries and, within countries,
nomic conditions. inequalities in cancer [8]. between groups with different
The large observed variations in At the area or country level, socioeconomic position.
cancer occurrence, even between socioeconomic conditions can be
otherwise similar populations, to- measured with macroeconomic indi- ■■ Cancer inequalities are driven
gether with the fact that changes in cators, such as national income (e.g. by the interplay of many
temporal trends may sometimes oc- as indicated by gross domestic prod- factors, which largely reflect
cur relatively quickly, indicate that uct) and years of schooling, or with the cultures and environments
these cancer differences could, in composite measures that include in which people are born,
principle, be substantially reduced. different combinations of indicators, live, and work, as well as
This chapter provides an overview such as the Human Development the uneven distribution of
of inequalities in cancer between Index (HDI), which is a composite resources and services
and within countries and then dis- indicator of health (based on life ex- between and within countries.
cusses possible interventions to re- pectancy at birth), education (based
duce these inequalities as well as on years of schooling), and standard ■■ Exposures to certain
research priorities, with a particular of living (based on gross national in- cancer risk factors, such as
focus on prevention. come per capita), or by proxy mea- tobacco smoking, alcohol
sures, such as levels of urbanicity consumption, unhealthy diet,
Measuring inequalities or rurality (see Chapter 1.3). occupational exposures, and
Another option is to use indica- cancer-causing infections,
in cancer tors of the extent of socioeconomic are highest predominantly
At the individual level, socioeco- inequality within an area or country, among individuals with low
nomic position reflects a complex such as the Gini index of income socioeconomic position
set of social and economic factors, inequality or the prevalence of pov- and among the most
often imperfectly correlated with erty or multiple deprivation. Such disadvantaged groups.
one another. Socioeconomic po- aggregate measures are often
sition is usually measured by the used in descriptive studies when ■■ The availability of and access
level of educational attainment, the individual-level data are not avail- to high-quality health-care
SECTION 4
CHAPTER 4.1
household income, and the occupa- able. However, caution should be services are often lower
tional classification, and sometimes exercised when linking aggregate- in lower-income countries
by the socioeconomic circumstanc- level indicators to health outcomes and among groups with low
es of the area or the location of the and attempting to draw conclusions socioeconomic position and
home residence. The choice be- about individual-level relationships. other disadvantaged groups.
tween these indicators may depend For more details about how to mea-
on the availability of data or on the sure inequalities in cancer, see [9]. ■■ Coordinated efforts could
objective of the study, because lead to efficient interventions,
these indicators may suggest dif- particularly those focusing on
ferent aspects and mechanisms for
Evidence of cancer prevention, and ultimately to a
the role of social determinants. inequalities between reduction of social inequalities
Several measures of associa- countries in cancer.
tion can be used to estimate the Large variations in cancer occur-
strength of the relationship between rence are observed between coun-
socioeconomic conditions and dis- tries (see Chapter 1.2), although a
ease, including cancer, or the ex- distinction must be made between of new cancer cases worldwide was
tent of inequality. Examples of ab- cancer incidence and cancer mor- 18.1 million, of which 44% occurred
solute measures of socioeconomic tality. On average, the incidence in countries with very high HDI, and
inequalities in cancer are rate dif- rates for all cancers combined, in 36%, 15%, and 4% occurred in coun-
ferences and the slope index of in- both sexes, increase with increasing tries with high, medium, and low HDI,
equality. Examples of relative mea- levels of national socioeconomic de- respectively [10]. In contrast, for the
sures are rate ratios, odds ratios, velopment: the highest-income coun- mortality rates for all cancers com-
and the relative index of inequality. tries have much higher rates than the bined, no clear gradient is observed
Because absolute and relative mea- lowest-income countries (Fig. 4.1.1). with average levels of national socio-
sures may lead to different conclu- In 2018, the estimated total number economic development.
Chapter 4.1 • Inequalities between and within countries 239

Fig. 4.1.1. Age-standardized (world population) incidence and mortality rates of all cancer types, by average level of socioeconomic
development in 2012. Socioeconomic development is measured by the education and income index (EDI), which is similar to the
Human Development Index (HDI) but excludes life expectancy. (HDI was not appropriate for this analysis because life expectancy
could be directly affected by cancer mortality.) EDI is calculated by taking the geometric mean of normalized indices of gross
national income per capita and of national education level (average and expected years of schooling). EDI is a dimensionless
variable between 0 and 1 (the higher a country’s score, the higher the level of development). Four categories of socioeconomic
development are shown: low (EDI ≤ 0.55), medium (0.55 < EDI ≤ 0.7), high (0.7 < EDI ≤ 0.8), and very high (EDI > 0.8).
All cancers excl. non−melanoma skin cancer, Male All cancers excl. non−melanoma skin cancer, Female
Low Medium High Very High Low Medium High Very High
400
Age−Standardized Rate (per 100 000)
300
200
100
0.4 0.6 0.8 1.0 0.4 0.6 0.8 1.0

EdI
population size: 1e+07 1e+08 1e+09 cancer incidence cancer mortality
Furthermore, the profile of can- socioeconomic development have, higher rates are currently observed
cer types varies markedly between on average, higher standards of liv- in high-income countries.
high- and low-income countries: lowing, improved hygienic conditions, In populations in which cancer
income countries have a higher rate higher life expectancy, and lower screening is widely available, “screen-
of infection-related cancers [11,12], rates of infection-related cancers. ing pressure” and increased detec-
such as stomach cancer, liver can- However, these improvements are tion of clinically irrelevant cancers in
cer, and cervical cancer (see Chapter often accompanied by changing individuals with higher access to the
2.2), whereas high-income countries environments, which may result in health-care system may contribute,
have higher rates of cancer types at least partly, to overdiagnosis and
increased exposure to other cancer
such as breast cancer, prostate can- overtreatment of certain cancers,
risk factors, particularly among low-
cer, colorectal cancer, thyroid cancer, such as prostate cancer, breast can-
income groups, and which may lead
and melanoma. cer, and thyroid cancer (see Chapter
Although there is considerable to national increases in cancer inci- 6.6). Overdiagnosis may have con-
heterogeneity in cancer patterns dence. In several low- and middle- tributed to the rise in incidence rates
between countries and there are income countries, particularly those observed in several high- and middle-
several exceptions, depending on that are undergoing rapid socioeco- income countries without substan-
the country or area and the cancer nomic transitions, the decreases in tially affecting mortality rates [13].
type, some general considerations rates of infection-related cancers In high-income countries, access to
apply. Countries that are undergoing are counterbalanced by increases screening and early detection pro-
a transition towards higher levels of in rates of cancer types for which grammes and to effective treatments
240
has contributed to keeping mortality infection-related cancers, such as with lower educational attainment,
rates relatively low, even when inci- stomach cancer, liver cancer, and cancer mortality has declined at a
dence rates have increased to very cervical cancer [14,18–20]. slower rate, has remained stable, or
high levels. Data on trends in cancer mor- has even, in some cases, increased.
The discrepancy between in- tality are available mainly in high-in- These differential trends can prob-
cidence and mortality is generally come countries and generally show ably be explained by the fact that
less pronounced in low- and mid- more favourable trends among peo- individuals with higher socioeco-
dle-income countries than in high- ple with higher socioeconomic posi- nomic position tend to benefit more
income countries, probably because tion. Among men and women with from cancer prevention interven-
of lower survival rates in low- and higher educational attainment, can- tions and to have earlier detection
middle-income countries as a result cer mortality has generally declined, and diagnosis and better treatment,
of later diagnosis and poorer access whereas among men and women because they have better access to
to treatment. It is not clear whether
it will be possible to provide an ad-
equate response to the growing Fig. 4.1.2. Relative social inequalities in cancer mortality by education level in 17
European countries, by country, for the most recent data available for each country
cancer epidemic in low- and middle-
(from 2004 to 2013). The charts show rate ratios and corresponding 95% confidence
income countries, given the organi- intervals of mortality from all cancers combined for men (above) and women (below)
zational constraints and the limited with a low versus high education level, and a pooled rate ratio estimate obtained from
resources available. a random effects meta-analysis.
Evidence of cancer
inequalities within
countries
Within countries, the socioeconom-
ic gradient for cancer incidence may
vary in magnitude and direction
across different cancer sites, but
cancer mortality is often higher, and
cancer survival lower, in groups with
low socioeconomic position and oth-
er disadvantaged groups (e.g. ethnic
and racial minorities and Indigenous
SECTION 4
CHAPTER 4.1
populations), for cancer overall and
for the large majority of cancer types
[7,12,14–17] (Fig. 4.1.2). There is a
clear gradient of higher overall can-
cer mortality and lower cancer sur-
vival from high to low socioeconomic
position [7], which shows that can-
cer inequalities affect (almost) the
entire population and are not limited
to low-income sectors of society.
Therefore, policies and interventions
to reduce cancer inequalities can
be beneficial for entire populations,
although the potential benefits are
largest for disadvantaged groups.
Relatively large socioeconomic
inequalities, with much higher can-
cer incidence and mortality in groups
with lower socioeconomic position,
have been consistently reported,
most markedly for smoking-related
cancers (see Chapter 2.1), such as
lung cancer, oral cancer, pharyn-
geal cancer, laryngeal cancer, and
oesophageal cancer, and also for

health-care services, greater health to the observed inequalities in can- the health-care system is often dif-
literacy, and fewer financial barriers cer survival, particularly in disad- ficult for disadvantaged groups, and
to health care compared with indi- vantaged populations. the availability of health-care ser-
viduals with lower socioeconomic vices is often lower in lower-income
position. The higher severity of co- Factors underlying countries [27]. Universal health cov-
morbidities in individuals with low erage, a current priority of WHO, is
cancer inequalities, and
socioeconomic position is also an key to improve access to essential
interventions to reduce components of cancer control for all
important factor that could reduce
inequalities individuals, without exposing them
cancer survival.
When data are available in low- Several factors, usually related and to financial hardships.
and middle-income countries, they intertwined, underlie the complicat- Preventive policies are potentially
often show similarly strong socio- ed patterns and socioeconomic gra- powerful ways to reduce not only the
economic gradients in cancer as dients in different cancer outcomes average incidence of and mortality
observed between and within coun- from cancer but also socioeconom-
observed in high-income countries,
tries. Exposures to certain cancer ic inequalities in cancer occurrence.
but they also reveal much poorer
risk factors, such as tobacco smok- National and international laws may
cancer outcomes than in high-in-
ing, alcohol consumption, unhealthy also have a powerful role, particu-
come countries [21,22], with very
diet, occupational exposures, and larly when used in coordination
high cancer mortality and low can-
cancer-causing infections, are high- with other initiatives (see Chapter
cer survival even for preventable or
est predominantly among individu- 6.8). Examples of legislative mea-
curable cancers, including cervical
als with low socioeconomic position sures are the banning of asbestos
cancer and childhood cancers. This and among the most disadvantaged
is generally due to the absence, or in workplaces and comprehensive
groups [24–26]. The reasons for this international tobacco control poli-
at best the limited availability, of re- are complex and include cultural, cies, such as the WHO Framework
sources and infrastructures at all economic, and psychosocial factors, Convention on Tobacco Control, in
phases of cancer control, from pre- as well as the availability, affordabili- which countries make commitments
vention to effective and timely treat- ty, and marketing of the products that to regulate tobacco use. Taxation is
ment to palliative care. Recently, cause cancer (e.g. tobacco and alco- a particularly efficient tool to reduce
there have been improvements hol) or prevent cancer (e.g. healthy consumption of tobacco, alcohol, and
in survival for most cancer types foods and sun-protective clothing). unhealthy foods.
in many low- and middle-income High-quality health-care servic- However, any intervention or
countries, although with a large var- es are key to control the burden of legislation that aims to reduce the
iability between cancer types and disease. Such services may reduce overall burden of a disease in a
between countries [23]. The stage cancer incidence and mortality at all population may result in either an
at diagnosis, the quality of treat- phases of cancer control, from pre- increase or a decrease in social in-
ment, and the quality of health-care vention to early detection, diagnosis, equalities in cancer, depending on
services are important contributors and treatment. However, accessing how it is designed, on the specific
context, and on many other factors.
Fig. 4.1.3. In almost all countries, graphic evidence of disparity within particular com- Therefore, there is a need to en-
munities may be illustrated. This photograph shows the physical divide that separates hance the use of evidence for the
Bloubosrand, a middle-class suburb northwest of Johannesburg, South Africa, from development, implementation, and
Kya Sands, an informal settlement consisting of improvised housing made of plywood
regulation of interventions, to en-
and corrugated metal.
sure that these would reduce or, at
least, would not exacerbate social
inequalities in cancer.
Interventions and policies are
likely to be more effective when
they are based on approaches that
combine a population strategy with
a vulnerable-population strategy –
an approach called proportionate
universalism. In the case of cervical
cancer, there is enormous potential
to eliminate the disease, and thus
reduce inequalities, through a com-
bination of human papillomavirus
(HPV) vaccination and screening
with HPV testing.
242
The increasing use of technol- Fig. 4.1.4. Access to state-of-the-art medical technology, such as this scanner, is re-
ogy in medical practice may be very stricted to high-income countries and is often available in a disproportionate manner.
useful, but in some cases it may Individuals with greater access to health-care services are most at risk of overdiagno-
sis and overtreatment.
also increase social inequalities in
cancer. This is because access to
innovative technology, and the re-
sulting benefits – like for any other
expensive intervention – are likely to
be enjoyed predominantly by high-
income individuals and countries.
In this context, it is relevant to high-
light an important phenomenon:
there is increasing evidence that
individuals and populations with
high socioeconomic position may
receive unnecessary care and that
the harms related to the use of tech-
nological advances and expensive
interventions may outweigh the
benefits. An example is the case of
thyroid cancer (see Chapter 5.18);
the increased medical surveillance
of the thyroid gland and the use of
advanced diagnostic techniques
have led to massive overdiagnosis in cancer, by implementing and im- which may, in some cases, increase
and overtreatment, affecting mainly proving the quality of cancer regis- social inequalities in cancer.
high-income countries and individu- tries, by carrying out surveys to moni- There is a strong need to expand
als with greater access to health- tor risk factors and access to health both the research focus on and in-
care, and by collecting other data in vestments in prevention, particularly
care services [28].
the context of surveillance, whether because of the low interest in invest-
national, regional, or global. In addi- ment in this area by the private sector.
Research priorities tion, etiological studies within a life- Of particular importance would be to
Research priorities have recently course framework, exploring oppor- ensure that all interventions and can-
SECTION 4
CHAPTER 4.1
been identified to inform approach- tunities to prevent the disease at all cer control initiatives, from prevention
es to tackle cancer inequalities [29]. stages of life, should be implemented to treatment measures, are explicitly
As a first step, the importance has to provide a more detailed analysis of designed and evaluated not only for
been recognized of (i) improving the inequalities in cancer. their overall effects but also, ideally,
collection of high-quality monitoring Furthermore, although social to decrease or eliminate social in-
data on the magnitude of social in- determinants affect all steps of the equalities or, at least, not exacerbate
equalities in cancer, (ii) increasing cancer continuum, including preven- them. This would represent an attain-
the scientific evidence base on the tion, diagnosis, treatment, and end- able, desirable, and ethical objective.
multidimensional aspects related to of-life care, it is prevention that has
social inequalities, particularly in the greatest potential to reduce can- Conclusions
low- and middle-income countries, cer disparities in all settings. This is Inequalities in cancer are consis-
where data are currently limited, particularly true in low- and middle- tently observed between and within
and (iii) improving the understand- income countries, where health-care countries. Although social inequal-
ing of the impact of social factors on services are lacking or are available ities affect the entire population,
all steps of the cancer continuum. almost exclusively for the highest- it is often the most disadvantaged
In all countries where data are income individuals. However, despite individuals and groups who suf-
available, there are striking differ- this great potential, investments in fer the most. This has an impact
ences in cancer occurrence between cancer prevention are disproportion- across societies, causing human
socioeconomic groups. Nevertheless, ately lower compared with other ar- and economic costs in the health
information on social characteristics eas, such as basic science and treat- system, which are borne by society
is often not collected in population- ment. The low budget allocated to but which could be, in large part,
based studies, including those based cancer prevention also contrasts with avoided. Coordinated, multisectoral
on cancer registry data. Improved ef- the large investments made in the de- efforts and efficient interventions
forts are needed to generate knowl- velopment of advanced technologi- could ultimately lead to a reduction
edge and monitor social inequalities cal devices and precision medicine, of social inequalities in cancer.

References
1. CSDH (2008). Closing the gap in a gen- 9. Conway DI, McMahon AD, Brown D, 17. Bryere J, Dejardin O, Launay L, Colonna M,
eration: health equity through action on the Leyland AH (2019). Measuring socio- Grosclaude P, Launoy G; French Network
social determinants of health. Final Report economic status and inequalities. In: of Cancer Registries (FRANCIM) (2018).
of the Commission on Social Determinants Vaccarella S, Lortet-Tieulent J, Saracci Socioeconomic status and site-specific
R, Conway DI, Straif K, Wild CP, edi- cancer incidence, a Bayesian approach in
of Health. Geneva, Switzerland: World
tors. Reducing social inequalities in can- a French Cancer Registries Network study.
Health Organization. Available from: https://
cer: evidence and priorities for research Eur J Cancer Prev. 27(4):391–8. https://
w w w.who.int /social _determinants/the doi.org/10.1097/CEJ.0000000000000326
(IARC Scientific Publications, No. 168).
commission/finalreport/en/. Lyon, France: International Agency for PMID:27879493
Research on Cancer. Available from:
2. Gallo V, Mackenbach JP, Ezzati M, 18. Mouw T, Koster A, Wright ME, Blank MM,
http://publications.iarc.fr/580.
Menvielle G, Kunst AE, Rohrmann S, et al. Moore SC, Hollenbeck A, et al. (2008).
(2012). Social inequalities and mortality in 10. Ferlay J, Ervik M, Lam F, Colombet M, Education and risk of cancer in a large
Europe – results from a large multi-nation- Mery L, Piñeros M, et al. (2018). Global cohort of men and women in the United
al cohort. PLoS One. 7(7):e39013. https:// Cancer Observatory: Cancer Today. Lyon, States. PLoS One. 3(11):e3639. https://
doi.org/10.1371/journal.pone.0039013 France: International Agency for Research doi.org/10.1371/journal.pone.0003639
on Cancer. Available from: https://gco.iarc. PMID:18982064
PMID:22848347
fr/today.
19. Spadea T, Zengarini N, Kunst A, Zanetti
3. Menvielle G, Boshuizen H, Kunst AE,
11. Plummer M, de Martel C, Vignat J, Ferlay R, Rosso S, Costa G (2010). Cancer risk
Vineis P, Dalton SO, Bergmann MM, et in relationship to different indicators of
J, Bray F, Franceschi S (2016). Global
al. (2010). Occupational exposures con- adult socioeconomic position in Turin,
burden of cancers attributable to infec-
tribute to educational inequalities in lung Italy. Cancer Causes Control. 21(7):1117–
tions in 2012: a synthetic analysis. Lancet
cancer incidence among men: evidence Glob Health. 4(9):e609–16. https://doi. 30. https://doi.org/10.1007/s10552-010-
from the EPIC prospective cohort study. org /10.1016/S2 214 -10 9X(16) 3 014 3 -7 9539-0 PMID:20349125
Int J Cancer. 126(8):1928–35. https://doi. PMID:27470177 20. Sharpe KH, McMahon AD, Raab
org/10.1002/ijc.24924 PMID:19810107
12. Bray F, Jemal A, Grey N, Ferlay J, Forman GM, Brewster DH, Conway DI (2014).
4. Merletti F, Galassi C, Spadea T (2011). D (2012). Global cancer transitions ac- Association between socioeconomic fac-
The socioeconomic determinants of cording to the Human Development Index tors and cancer risk: a population cohort
(2008-2030): a population-based study. study in Scotland (1991-2006). PLoS
cancer. Environ Health. 10(Suppl 1):S7.
Lancet Oncol. 13(8):790–801. https:// One. 9(2):e89513. https://doi.org/10.1371/
https://doi.org/10.1186/1476-069X-10-S1- journal.pone.0089513 PMID:24586838
S7 PMID:21489217 doi.org/10.1016/S1470-2045(12)70211-5
PMID:22658655 21. de Vries E, Arroyave I, Pardo C, Wiesner C,
5. Franceschi S, Plummer M, Clifford G, Murillo R, Forman D, et al. (2015). Trends
13. Sullivan R, Aggarwal A (2019). Technology
de Sanjosé S, Bosch X, Herrero R, et in inequalities in premature cancer mortal-
and cancer systems: creating better policy
al.; International Agency for Research ity by educational level in Colombia, 1998-
to enhance equality. In: Vaccarella S, Lortet-
on Cancer Multicentric Cervical Cancer Tieulent J, Saracci R, Conway DI, Straif K, 2007. J Epidemiol Community Health.
Study Groups; International Agency 69(5):408–15. https://doi.org/10.1136/jech-
Wild CP, editors. Reducing social inequal-
for Research on Cancer Human 2014-204650 PMID:25492898
ities in cancer: evidence and priorities for
Papillomavirus Prevalence Surveys Study research (IARC Scientific Publications, No. 22. Dikshit R, Gupta PC, Ramasundarahettige
Group (2009). Differences in the risk of 168). Lyon, France: International Agency for C, Gajalakshmi V, Aleksandrowicz L,
cervical cancer and human papillomavirus Research on Cancer. Available from: http:// Badwe R, et al.; Million Death Study
infection by education level. Br J Cancer. publications.iarc.fr/580. Collaborators (2012). Cancer mortality
101(5):865–70. https://doi.org/10.1038/sj. in India: a nationally representative sur-
14. Dalton SO, Steding-Jessen M, Engholm
bjc.6605224 PMID:19654578 vey. Lancet. 379(9828):1807–16. https://
G, Schüz J, Olsen JH (2008). Social in-
doi.org/10.1016/S0140-6736(12)60358-4
6. Dahlgren G, Whitehead M (2006). equality and incidence of and survival from
PMID:22460346
lung cancer in a population-based study
European strategies for tackling social
in Denmark, 1994-2003. Eur J Cancer. 23. Swaminathan R (2019). Cancer survival
inequities in health: levelling up (part 2).
44(14):1989–95. https://doi.org/10.1016/j. in countries in transition, with a focus on
Copenhagen, Denmark: World Health ejca.2008.06.023 PMID:18693111 selected Asian countries. In: Vaccarella
Organization Regional Office for Europe. S, Lortet-Tieulent J, Saracci R, Conway
Available from: http://www.who.int/iris/ 15. AIHW (2013). Cancer in Aboriginal DI, Straif K, Wild CP, editors. Reducing
handle/10665/107791. and Torres Strait Islander peoples social inequalities in cancer: evidence
of Australia: an overview. Canberra, and priorities for research (IARC Scientific
7. Menvielle G, Kunst AE, Stirbu I, Strand Australia: Australian Institute of Health Publications, No. 168). Lyon, France:
BH, Borrell C, Regidor E, et al. (2008). and Welfare. Available from: https://www. International Agency for Research on
Educational differences in cancer mortality aihw.gov.au/repor ts/cancer/cancer-in- Cancer. Available from: http://publications.
among women and men: a gender pattern indigenous-australians-overview/contents/ iarc.fr/580.
that differs across Europe. Br J Cancer. table-of-contents.
98(5):1012–9. https://doi.org/10.1038/sj. 24. Allen L, Williams J, Townsend N,
16. Blakely T, Shaw C, Atkinson J, Cunningham Mikkelsen B, Roberts N, Foster C, et al.
bjc.6604274 PMID:18283307 R, Sarfati D (2011). Social inequalities or (2017). Socioeconomic status and non-
inequities in cancer incidence? Repeated communicable disease behavioural risk
8. Mackenbach JP (2015). Should we aim to
census-cancer cohort studies, New factors in low-income and lower-middle-
reduce relative or absolute inequalities in Zealand 1981-1986 to 2001-2004. Cancer income countries: a systematic review.
mortality? Eur J Public Health. 25(2):185. Causes Control. 22(9):1307–18. https:// Lancet Glob Health. 5(3):e277–89. https://
https://doi.org/10.1093/eurpub/cku217 d o i .o r g /10 .10 07/s10 5 5 2- 011- 9 8 0 4 -x doi.org/10.1016/S2214-109X(17)30058-X
PMID:25818489 PMID:21717195 PMID:28193397
244
25. Casetta B, Videla AJ, Bardach A, Morello 27. Atun R, Jaffray DA, Barton MB, Bray F, 29. Vaccarella S, Lortet-Tieulent J, Saracci
P, Soto N, Lee K, et al. (2017). Association Baumann M, Vikram B, et al. (2015). R, Fidler MM, Conway DI, Vilahur N, et
between cigarette smoking prevalence Expanding global access to radiotherapy. al. (2018). Reducing social inequalities in
and income level: a systematic review cancer: setting priorities for research. CA
and meta-analysis. Nicotine Tob Res. Cancer J Clin. 68(5):324–6. https://doi.
19(12):1401–7. https://doi.org/10.1093/ntr/ doi.org/10.1016/S1470-2045(15)00222-3
org/10.3322/caac.21463 PMID:30152865
ntw266 PMID:27679607 PMID:26419354
26. Grittner U, Kuntsche S, Gmel G, Bloomfield 28. Vaccarella S, Franceschi S, Bray F, Wild CP,
K (2013). Alcohol consumption and social Plummer M, Dal Maso L (2016). Worldwide
inequality at the individual and coun-
thyroid-cancer epidemic? The increasing
try levels – results from an international
study. Eur J Public Health. 23(2):332–9. impact of overdiagnosis. N Engl J Med.
https://doi.org/10.1093/eurpub/cks044 375(7):614–7. https://doi.org/10.1056/NEJM
PMID:22562712 p1604412 PMID:27532827
SECTION 4
CHAPTER 4.1

4.2
Socioeconomic factors and cancer

prevention in Africa
Cervical cancer as an example
Lynette Denny Clement A. Adebamowo (reviewer) Rengaswamy Sankaranarayanan
(reviewer)
Filip Meheus (reviewer)
Robert Newton (reviewer)
(31.5 million) occurred in low- and cancer care and outcomes in Africa,

SUMMARY middle-income countries, and can- with cervical cancer as an example.
cer accounted for 9.0 million deaths
●● In sub-Saharan Africa, cervi- (22% of all deaths due to noncommu-
cal cancer is the second most
Overall cancer burden in
nicable diseases) [1]. Approximately Africa and globally
common cancer in women, after one third of cancer cases in sub-
breast cancer, but more women Saharan Africa were estimated to be The overall cancer burden in Africa
die from cervical cancer than attributable to infections, presenting in 2012 was estimated at 847 000
from breast cancer. new cancer cases and 591  000 can-
unique opportunities for prevention
cer deaths [5]. In women, the most
and treatment [2].
●● Although cervical cancer is pre- common cancer type was breast
Inequity in health care exists be-
ventable, services for prevention, cancer (133 900 cases), followed by
tween countries, within countries,
early detection, and treatment cervical cancer (99 000 cases). In
and across continents. The lowest-
are rare in low-income countries. men, prostate cancer was the most
income countries provide the worst
common (59 500 cases), followed
●● It was found that for women in quality of care and spend the small-
by liver cancer (38 700 cases) and
developing countries the cervi- est amount of national resources on
Kaposi sarcoma (23 800 cases) [5].
cal cancer incidence rates were health care. Access to high-quality
CONCORD-3 updated the world-
2-fold higher and the cervical care is a key factor in predicting
wide surveillance of cancer survival
cancer mortality rates were good outcomes in all forms of health trends to include patients diagnosed
3-fold higher than those for care; it requires an “ecosystem” up to 2014 [6]. Data were analysed
women in developed countries. of interrelated support, which infor 322 population-based cancer reg-
cludes arable land, adequate nutri- istries in 71 countries; for Africa, this
●● The poverty rate (a deprivation tion, safe drinking-water, sanitation, included 8 registries in 6 countries.
level measuring the propor- and transportation infrastructure as The 322 registries covered a com-
tion of the population living in a few examples of necessary inter- bined population of almost 1 billion
extreme poverty) was a strong ventions [3]. In addition, expenditure people in about 2014. Overall, the
predictor of cross-national vari- on health care, health-care profes- proportion of the population covered
ations in cervical cancer inci- sionals, and health infrastructure is by cancer registries in Africa was
dence and mortality. key to functional and strong health- 3.5% (Table 4.2.1) [6].
care systems [4]. There are vast differences in
Cancer is a leading cause of pre- cervical cancer mortality rates be-
Of the 56.9 million deaths recorded mature death and morbidity globally tween women in Africa and women in
globally in 2016, 40.5 million (71%) and is rapidly becoming a significant high-income countries (Table 4.2.2)
were due to noncommunicable dis- health problem in low- and middle-in- [7]. Singh et al. [7] computed age-
eases. The four main causes of death come countries, particularly in Africa, adjusted cervical cancer incidence
due to noncommunicable diseases where there is an epidemiological and mortality rates for women in 184
were cardiovascular diseases, can- shift from communicable to noncom- countries using the GLOBOCAN
cer, diabetes, and chronic respiratory municable diseases (see Chapter 2008 database. The authors’ analy-
diseases (see Chapter 6.9). In 2016, 1.3) [5]. sis indicated that overall, for women
more than three quarters of deaths This chapter explores the range of in developing countries the incidence
due to noncommunicable diseases effects of socioeconomic factors on rates were 2-fold higher and the
246
Table 4.2.1. Population covered by cancer registries in Africa (number of people and
percentage of the national population) and number of patients diagnosed during 2000– FUNDAMENTALS
2014, by country
■■ Cancer is becoming a
Cancer registry Population Percentage Total number significant health problem in
covered of population of patients
covered many low- and middle-income
countries, where both incidence
Algeria 2 447 075 6.3%  
15 602 and mortality rates are higher
Mali (Bamako) 764 245 9.0% 60 than those in some high-income
countries, but where the health
Mauritius 1 268 567 100.0%  
3 959 agenda has been dominated
by maternal mortality,
Morocco (Casablanca) 2 178 083 12.7%  
4 683
communicable diseases, and
Nigeria (Ibadan) 2 797 220 1.6%  
8 274 nutritional diseases.
South Africa (Eastern Cape) 1 078 572 2.0%  
7 619 ■■ The Human Development
Total 10 533 762 3.5% 40 197
Index (HDI) and the poverty
rate explain more than 50% of
the global variance in cervical
Table 4.2.2. Age-adjusted cervical cancer mortality rates per 100 000 (world standard cancer mortality.
population), in 2008
■■ Cervical cancer is known to be
Country Number of deaths Age-adjusted a preventable disease.
mortality rate
■■ Modern technology has the
Countries with the highest mortality rates
potential to enable greater
Guinea 1217 41.7 precision and sensitivity in the
application of screening and
Zambia 1276 38.6
early detection for many cancer
Malawi 1621 38.3 types, but it is not accessible in
Uganda 2464 34.9
low-income countries.
■■ There are still differences in the
CHAPTER 4.2
SECTION 4
Zimbabwe 1286 33.4
occurrence of cancer across
Lesotho 178 22.7
different groups, resulting in
Angola 1008 21.9 deepening health inequalities.
Countries with the lowest mortality rates ■■ An inadequately trained

health-care workforce, inad-
Australia 241 1.4
equate expenditure on health
Iceland 4 0.8 systems and infrastructure,
out-of-pocket expenses, and
lack of preventive health care
mortality rates were 3-fold higher and labour market participation), and
are major obstacles to health-
than those for women in developed socioeconomic factors (poverty rate
care delivery and development
countries. Cervical cancer rates [a deprivation level measuring the
varied widely across countries; rates proportion of the population living in low-income countries.
in many countries in sub-Saharan in extreme poverty], health expen-
Africa were 10–20-fold higher than diture per capita, urbanization rate,
those in some countries in North and literacy rate). All were found to middle-income countries in other
Africa, the Middle East, and Europe. be significantly related to cervical world regions is found in the differ-
Furthermore, Singh et al. mod- cancer incidence and mortality. HDI ent incidence and mortality rates of
elled the impact of the Human and the poverty rate each explained various cancer types. An estimated
Development Index (HDI), the Gender more than 52% of the global variance 18.1 million new cancer cases and
Inequality Index (a composite index in cervical cancer mortality [7]. 9.6 million cancer deaths occurred
that reflects women’s relative social The evidence of the impact of worldwide in 2018 [8]. The average
disadvantage in three dimensions: socioeconomic factors and cancer risk of developing cancer before age
reproductive health, empowerment, prevention in Africa and in low- and 75 years was 20%, and the average
Chapter 4.2 • Socioeconomic factors and cancer prevention in Africa 247

risk of dying from cancer before age Fig. 4.2.1. Total expenditure on health as a percentage of gross domestic product
75 years was 10%. In men, prostate (GDP) in 2000 and 2012, by WHO region.
cancer was the most frequently di-
16
agnosed cancer in 12 regions of
14 13.6
the world. In both sexes, lung can-
cer was the most frequent cause of 12 11.0
death from cancer in 14 regions of 10
8.9
the world. In women, breast cancer 8
7.9
6.6
was the most frequently diagnosed 6 5.6 5.8
cancer in all regions of the world, 4.2
3.6 3.7 4.0 4.6
4
and cervical cancer ranked fourth 2
for both incidence and mortality [8].
0
Of the 18.1 million new can-
Africa Americas South-East Europe Eastern Western
cer cases in 2018, 5.8% occurred Asia Mediterranean Pacific
in Africa, 21.0% in the Americas,
23.4% in Europe, 1.4% in Oceania, 2000 2012
and 48.4% in Asia. Of the 9.6 mil-
lion cancer deaths, 7.3% occurred Fig. 4.2.2. Per capita total expenditure on health (purchasing power parity at interna-
in Africa, 14.4% in the Americas, tional dollar rate) in 2000 and 2012, by WHO region.
20.3% in Europe, 0.7% in Oceania,
and 57.3% in Asia [9]. Although the 4000 3768
proportion of the global cancer bur- 3500
den is lower for Africa than for other
3000
regions of the world, cancer is also
2500 2402
low on the health agenda in Africa 2055
because of multiple competing 2000
health priorities and other needs. 1500 1216
1000 857
Costs of cancer care 500 260
561
315
110 208 88 208
In 2009, the global cost of treat- 0
ing 12.9 million patients diagnosed Africa Americas South-East Europe Eastern Western
with cancer was estimated to be Asia Mediterranean Pacific
US$ 285.8 billion [10]. The indirect
2000 2012
costs associated with premature
death and lost productivity from the
growing cancer burden were es-
timated to be US$ 1.16 trillion per disease that is largely preventable, nancial contributions to the health
year [10]. such as cervical cancer. The high system exceeded 40% of income
World Health Statistics 2015 incidence of cervical cancer in remaining after subsistence needs
presented data on the total expen- Africa is also related to the high had been met. The analysis showed
diture on health as a percentage rates of HIV infection, particularly in that certain groups were particu-
of gross domestic product (GDP) eastern and southern Africa, where larly vulnerable, such as older
in the six WHO regions (Fig. 4.2.1) HIV infection is epidemic and cervi- people, people with disabilities, un-
[11]. In most regions, there was very cal cancer is classified as an AIDS- employed people, people with low
little change in the percentage ex- defining illness [12]. incomes, and people with reduced
penditure between 2000 and 2012. Out-of-pocket expenditure on or no access to health insurance.
The percentage expenditure was health care is a major barrier to Wyszewianski [14] made the point
highest in the Americas. For per accessing health care in low- and that catastrophic health expendi-
capita total expenditure on health middle-income countries, and a ture is common in many countries
(Fig. 4.2.2) [11], the values were significant illness in a family can be and can lead to impoverishment
lowest in Africa and South-East catastrophic. Xu et al. [13] used a that has long been ignored by the
Asia and highest in the Americas cross-country analysis design and health system. There is a significant
and Europe. data from household surveys in amount of data showing that low-
The lack of access to screen- 59 countries to explore variables income households have a limited
ing and early detection and the high related to catastrophic health ex- capacity to cope with health-care
costs of treatment are often cited as penditure. Expenditure was defined expenditure compared with higher-
the causes for a high incidence of a as catastrophic if a household’s fi- income households.
248
The American Public Health As Moreover, access to anti-can- to-incidence ratio for prostate can-
sociation reported that before the cer therapies is very limited in al- cer, breast cancer, and lung cancer.
introduction of the Patient Protection most all African countries. A WHO They concluded that the population
and Affordable Care Act of 2010 in study in 2001 found that only 22% density of radiotherapy machines
the USA, about 20% of the popula- of African countries had access to is related to cancer mortality in-
tion younger than 65 years was medi- anti-cancer drugs, compared with dependently of other public health
cally uninsured, and that after the in- 91% in Europe. An analysis by the parameters. They also found a lin-
troduction of the act, about 13% (or International Atomic Energy Agency ear relationship between GDP per
one eighth) of people younger than found that in 2010 only 23 of the 52 capita and the population density of
65 years remained uninsured [15]. African countries included in the radiotherapy machines, until a GDP
The USA spends more on health analysis had facilities for teletherapy per capita of US$ 60 000 [19].
care than any other high-income (external radiation therapy), which
country (18% of the GDP), but in were concentrated in the northern Health-care workforce
terms of life expectancy it ranks 26th and southern regions of the conti- The African continent has 168 med-
out of the 36 member countries of nent [17]. Brachytherapy resources ical schools, located in 41 coun-
the Organisation for Economic Co- were available in only 20 of the 52 tries. However, facilities for training
operation and Development. Further countries. A total of 160 radiation in cancer prevention, diagnosis,
more, in the USA only about 3% of facilities were recorded in the conti- and management are found mainly
spending on health is allocated to nent, housing 277 radiotherapy main North Africa (Algeria, Egypt, and
preventive health care. chines (88 cobalt-60 units and 189 Morocco) and South Africa, with lim-
linear accelerators) [17]. ited facilities in Libya, Nigeria, and
Barton et al. [18] performed a de- Zimbabwe [20]. Overall, sub-Saha-
Barriers to prevention tailed analysis of the gap between ran Africa has a very low physician-
and treatment of cancer existing radiation facilities in low- to-population ratio of about 18 per
in Africa and middle-income countries and 100 000, compared with the ratios
Almost all of the 54 countries the needs of the population. They of India (60 per 100 000), Brazil
in sub-Saharan Africa have low concluded that the African conti- (170 per 100 000), and France (370
HDI values and high values of the nent had only 18% of the radiation per 100 000) [20].
Human Poverty Index [16]. Of the equipment needed for full coverage Adding to the complexity of the
total population of sub-Saharan of the population. Medenwald et al. challenges facing sub-Saharan
Africa, which was estimated to be [19] extracted data from a wide vari- Africa (including environmental
more than 1 billion in 2018, only ety of sources and found an inverse disasters, competing health needs,
CHAPTER 4.2
SECTION 4
7.2% were covered by medically linear relationship between the endemic civil strife, war, and lack of
certified causes of death and 8.3% number of radiotherapy machines safe drinking-water and sanitation,
by population-based registries. in the population and the mortality- to name just a few) has been the
HIV/AIDS epidemic. Sub-Saharan
Africa accounts for about 70% of
Fig. 4.2.3. Women signing up for free breast cancer and cervical cancer screening in people living with HIV worldwide
Senegal. Breast cancer and cervical cancer are the two most common cancer types [21]. HIV infection increases the
among women in Africa.
risk of developing certain types of
cancer, and Kaposi sarcoma, non-
Hodgkin lymphoma, and cervical
cancer have been classified as
AIDS-defining diseases since 1993
[12]. Women living with HIV have an
increased risk of being infected with
human papillomavirus (HPV) and
are therefore considered to be at a
higher risk for anogenital cancers.
Socioeconomic
determinants of health
The political determinants of
health inequity and socioeconom-
ic factors deserve careful analy-
sis. The Lancet-University of Oslo
Commission on Global Governance

for Health noted that the lowest- Fig. 4.2.4. Two girls aged 10–14 years are vaccinated against human papillomavirus
income population groups have the (HPV) by an outreach nursing team from Binga District Hospital in Matabeleland North
heaviest burden of disease; this can Province in Zimbabwe.
be attributed not only to poverty but
also to socioeconomic inequality
[22]. The commission identified five
dysfunctions of the global gover-
nance system that allow adverse
effects of global political determi-
nants of health inequity to persist:
(i) insufficient participation in deci-
sion-making by civil society, health
experts, and marginalized groups;
(ii) weak accountability mecha-
nisms; (iii) lack of response to
changing societal needs, enabling
entrenchment of power dispari-
ties, with adverse effects on health
(called “institutional stickiness” by
the authors); (iv) inadequate policy
space for health; and (v) lack of
international institutions to protect
and promote health [22].
The Commission on Social De
terminants of Health, led by Michael
Marmot, stated in its report: “The
poor health of the poor, the social ach cancer, liver cancer, oesopha- for late presentation were: negative
gradient in health within countries, geal cancer, and cervical cancer. interpretation of symptoms; fear;
and the marked health inequities Together, these cancers accounted lack of belief, trust, or confidence
between countries are caused by for 62% of the cancer burden in re- in orthodox medicine; poor social
the unequal distribution of power, gions with low HDI. In both settings, relations and networks; and lack of
income, goods, and services, glob- lung cancer was the most common access to health care [24].
ally and nationally, the consequent cancer diagnosed.
unfairness in the immediate, visible
circumstances of people’s lives ...
Challenges associated
and their chances of leading a flour-
Priorities for prevention, with cancer care in
ishing life. This unequal distribution research, policy, and Africa
of health-damaging experiences is development Analyses of the causes of ill health
not in any sense a ‘natural’ phenom- Men and women with cancer in low- are essential to prioritize public
enon but is the result of a toxic com- and middle-income countries, par- policy and to determine the re-
bination of poor social policies and ticularly in Africa, face multiple chal- search agenda and the allocation
programmes, unfair economic ar- lenges because of poor health-care of resources, particularly based on
rangements, and bad politics...” [23]. infrastructure. Access to diagnosis, the population-level risk. Attaining
Bray et al. [16] used four tiers treatment, and timely intervention the highest standard of health care
of HDI (low, medium, high, and are lacking, resulting in high case requires access to safe drinking-
very high HDI) to evaluate cancer- mortality rates, lack of trust in the water, adequate sanitation, educa-
specific patterns in 2008 and trends health-care system, stigmatization, tion, health-care education, nutri-
over the period 1988–2002. They and high rates of premature death. tion, and good employment, among
found that in the regions with the In a systematic review of nine many other factors. Cancer care is
highest HDI in 2008, breast cancer, eligible studies of late presenta- relatively expensive, and without
lung cancer, colorectal cancer, and tion of women with breast cancer effective means of prevention and
prostate cancer accounted for more conducted in Egypt, Ghana, Kenya, early detection, aligned with appro-
than half of the cancer burden. In Libya, and Nigeria, more than 50% priate interventions, the incidence
regions with low HDI, other cancer of women presented with advanced of and mortality from cancer will
types were more common: stom- disease. The most important drivers continue to rise.
250
References
1. WHO (2018). NCD mortality and morbid- 9. Bray F, Ferlay J, Soerjomataram I, Siegel 17. Abdel-Wahab M, Bourque JM, Pynda Y,
ity. Global Health Observatory (GHO) RL, Torre LA, Jemal A (2018). Global Iżewska J, Van der Merwe D, Zubizarreta
data. Geneva, Switzerland: World Health cancer statistics 2018: GLOBOCAN es- E, et al. (2013). Status of radiotherapy re-
Organization. Available from: https://www. timates of incidence and mortality world- sources in Africa: an International Atomic
who.int/gho/ncd/mortality_morbidity/en/. wide for 36 cancers in 185 countries. CA Energy Agency analysis. Lancet Oncol.
Cancer J Clin. 68(6):394–424. https://doi. 14(4):e168–75. https://doi.org/10.1016/
2. Plummer M, de Martel C, Vignat J, Ferlay
org/10.3322/caac.21492 PMID:30207593 S1470-2045(12)70532-6 PMID:23561748
J, Bray F, Franceschi S (2016). Global
burden of cancers attributable to infec- 10. Fitzmaurice C, Dicker D, Pain A, Hamavid 18. Barton MB, Frommer M, Shafiq J (2006).
tions in 2012: a synthetic analysis. Lancet H, Moradi-Lakeh M, MacIntyre MF, et Role of radiotherapy in cancer control in
Glob Health. 4(9):e609–16. https://doi. al.; Global Burden of Disease Cancer low-income and middle-income coun-
org /10.1016/S2 214 -10 9X(16) 3 014 3 -7 Collaboration (2015). The global burden of tries. Lancet Oncol. 7(7):584–95. https://
PMID:27470177 cancer 2013. JAMA Oncol. 1(4):505–27. doi.org/10.1016/S1470-2045(06)70759-8
3. Wagstaff A (2002). Poverty and health https://doi.org/10.1001/jamaoncol.2015. PMID:16814210
sector inequalities. Bull World Health 0735 PMID:26181261
19. Medenwald D, Vordermark D, Dietzel CT
Organ. 80(2):97–105. PMID:11953787 11. WHO (2015). World health statistics (2018). Number of radiotherapy treatment
4. WHO (2018). Global health workforce sta- 2015. Geneva, Switzerland: World machines in the population and cancer mor-
tistics. Geneva, Switzerland: World Health Health Organization. Available from: tality: an ecological study. Clin Epidemiol.
Organization. Available from: https://www. https://w w w.who.int /gho/publications/ 10:1249–73. https://doi.org/10.2147/CLEP.
who.int/hrh/statistics/hwfstats/. world_health_statistics/2015/en/. S156764 PMID:30288122
5. Parkin DM, Bray F, Ferlay J, Jemal A (2014). 12. CDC (1992). 1993 revised classifica-
20. Mullan F, Frehywot S, Omaswa F, Buch E,
Cancer in Africa 2012. Cancer Epidemiol tion system for HIV infection and ex-
Chen C, Greysen SR, et al. (2011). Medical
Biomarkers Prev. 23(6):953–66. https:// panded surveillance case definition for
schools in sub-Saharan Africa. Lancet.
doi.org/10.1158/1055-9965.EPI-14-0281 AIDS among adolescents and adults.
377(9771):1113–21. https://doi.org/10.1016/
PMID:24700176 MMWR Recomm Rep. 41(RR-17):1–19.
S0140-6736(10)61961-7 PMID:21074256
PMID:1361652
6. Allemani C, Matsuda T, Di Carlo V, 21. UNAIDS (2018). Geneva, Switzerland:
Harewood R, Matz M, Nikšić M, et al.; 13. Xu K, Evans DB, Kawabata K, Zeramdini
Joint United Nations Programme on HIV
CONCORD Working Group (2018). Global R, Klavus J, Murray CJL (2003).
and AIDS. Available from: https://www.
surveillance of trends in cancer survival Household catastrophic health expen-
unaids.org.
2000–2014 (CONCORD-3): analysis of diture: a multicountry analysis. Lancet.
individual records for 37 513 025 patients 362(9378):111–7. https://doi.org/10.1016/ 22. Ottersen OP, Dasgupta J, Blouin C, Buss P,
diagnosed with one of 18 cancers from 322 S0140-6736(03)13861-5 PMID:12867110 Chongsuvivatwong V, Frenk J (2014). The
population-based registries in 71 coun- political origins of health inequity: pros-
tries. Lancet. 391(10125):1023–75. https:// 14. Wyszewianski L (1986). Families with ca-
tastrophic health care expenditures. Health pects for change. Lancet. 383(9917):630–
doi.org/10.1016/S0140-6736(17)33326-3 67. https://doi.org/10.1016/S0140-6736(13)
Serv Res. 21(5):617–34. PMID:3102403
CHAPTER 4.2
SECTION 4
PMID:29395269
62407-1 PMID:24524782
7. Singh GK, Azuine RE, Siahpush M (2012). 15. APHA (2018). Health reform. American
Public Health Association. Available from: 23. CSDH (2008). Closing the gap in a gen-
Global inequalities in cervical cancer in-
https://www.apha.org/topics-and-issues/ eration: health equity through action on the
cidence and mortality are linked to dep-
health-reform. social determinants of health. Final Report
rivation, low socioeconomic status, and
human development. Int J MCH AIDS. of the Commission on Social De ter
mi-
16. Bray F, Jemal A, Grey N, Ferlay J, Forman nants of Health. Geneva, Switzerland: World
1(1):17–30. https://doi.org/10.21106/ijma.12
D (2012). Global cancer transitions ac- Health Organization. Available from: https://
PMID:27621956
cording to the Human Development Index w w w.who.int /social _determinants/the
8. Ferlay J, Colombet M, Soerjomataram I, (2008-2030): a population-based study. commission/finalreport/en/.
Mathers C, Parkin DM, Piñeros M, et al. Lancet Oncol. 13(8):790–801. https://
(2019). Estimating the global cancer inci- doi.org/10.1016/S1470-2045(12)70211-5 24. Donkor A, Lathlean J, Wiafe S, Vander
dence and mortality in 2018: GLOBOCAN PMID:22658655 puye V, Fenlon D, Yarney J, et al. (2015).
sources and methods. Int J Cancer. Factors contributing to late presentation
144(8):1941–53. https://doi.org/10.1002/ of breast cancer in Africa: a systematic
ijc.31937 PMID:30350310 literature review. Arch Med. 8:2.

4.3
Cancer in urban and rural communities

in China
Patterns reflect social dynamics
Wanqing Chen Chunxue Bai (reviewer)
He Li Partha Basu (reviewer)
Zhixun Yang Zhengming Chen (reviewer)
new cases of liver cancer, oesoph- standard population) was 186.53

SUMMARY ageal cancer, and stomach cancer per 100 000. The crude mortality
and more than one third of the new rate was 167.89 per 100 000, and
●● In China, cancer incidence is cases of lung cancer worldwide oc- the age-standardized mortality rate
lower in rural areas than in urban cur in China [1]. (by world standard population) was
areas, whereas cancer mortality In recent decades the cancer 106.09 per 100 000 [2].
is higher in rural areas, indicating burden in China has been increas- The most common cancer types
lower survival in rural areas. ing, posing a serious threat to public in the whole population were cancers
●● Incidence rates of colorectal health and imposing a heavy eco- of the lung, stomach, colorectum,
cancer, breast cancer, prostate nomic burden. In 2014, there were liver, breast, oesophagus, thyroid,
cancer, and bladder cancer are more than 3.8 million new cancer cervix, brain and central nervous
higher in urban areas than in cases (2.3 million in urban areas system, and pancreas. Together,
rural areas, whereas incidence and 1.5 million in rural areas) and these accounted for about 77% of all
rates of oesophageal cancer, 2.3 million cancer deaths (1.3 mil- new cancer cases. Cancers of the
stomach cancer, liver cancer, lion in urban areas and 1.0 million lung, liver, stomach, oesophagus,
in rural areas) in China [2]. The colorectum, pancreas, and breast,
and cervical cancer are high-
crude incidence rate was 278.07 collectively, accounted for about
er in rural areas than in urban
per 100 000, and the age-stan 70% of all cancer deaths [2]. The
areas.
dardized incidence rate (by world direct economic burden attributable
●● Differences in lifestyles and di-
etary patterns between urban
Fig. 4.3.1. The Shanghai skyline. Differences in lifestyles and dietary patterns between
and rural communities are be- urban and rural communities in China are becoming more pronounced along with rapid
coming more pronounced along economic development, urbanization, and the ageing of the population.
with rapid economic develop-
ment, urbanization, and the
ageing of the population. These
could partly explain the urban–
rural difference in the spectrum
of cancer types.
●● There is an urgent need to im-
plement cancer prevention and
control strategies that are cus-
tomized for different regions of
the country.
As the world’s most populous coun-

try, China accounts for more than
23% of new cancer cases and
about 30% of cancer deaths world-
wide [1]. Moreover, about half of the
252
Fig. 4.3.2. A woman preparing rice in rural China.

FUNDAMENTALS
■■ As the world’s most populous
country, China accounts for
more than 23% of new cancer
cases, about half of the new
cases of liver cancer, oesopha-
geal cancer, and stomach can-
cer, and about 30% of cancer
deaths worldwide.
■■ Along with rapid economic
development, urbanization,
and the ageing of the
population, the cancer burden
and the spectrum of cancer
types show considerable
variation between urban and
rural areas.
■■ China is urbanizing rapidly;
the percentage of the
population living in urban
areas increased from 18%
in 1978 to 56% in 2015 – an
increase of 311.1% in about
40 years – and is expected to
reach 71% by 2030.
to cancer in 2015 was estimated to eas in the spectrum of the major ■■ Obesity and physical inactivity,
be ¥221.4 billion, which was 5.4% cancer types [6]. which are the leading risk
of the total health expenditure and These differences in cancer factors for both colorectal
17.7% of the government health ex- patterns could be related mainly to cancer and breast cancer, are
penditure [3]. comprehensive determinants, such more prevalent in urban areas
CHAPTER 4.3
SECTION 4
as demographic and socioeconomic than in rural areas, contributing
determinants (e.g. age, sex, educa-
Cancer burden in urban to the rural–urban disparity
tion level), as well as to lifestyle fac- in the incidence of these two
and rural communities tors and inequalities in health-related cancer types.
Along with rapid economic develop- issues (e.g. allocation of health-care
ment, urbanization, and the ageing resources, health outcomes). ■■ Although differences in cancer
of the population, the cancer burden Quality of life and provision of incidence between urban and
and the spectrum of cancer types health-care services have improved rural areas still exist in China,
show considerable variation be- greatly in China with the rapid so- the gap has been narrowing
tween urban and rural areas [4,5]. cioeconomic development during every year.
In 2014, the age-standardized the past decades. However, ur-
incidence rate (by world standard ban–rural inequalities in health care
population) for all cancers com- are still striking [7,8]. According to cy, China has a lower birth rate and
bined was higher in urban areas the National Bureau of Statistics of a lower death rate, especially in ur-
(191.6 per 100 000) than in rural ar- China, in 2015 the average per cap- ban areas [11]. This has led to a rapid
eas (179.2 per 100 000), whereas ita disposable income of urban resi- ageing of the population, especially
the age-standardized mortality rate dents was ¥31 790, almost 3 times in urban areas, thus increasing the
(by world standard population) for that of rural residents (¥10 772) [9]. pool of older adults, who are more
all cancers combined was higher in The average life expectancy for susceptible to cancer [2,11].
rural areas (110.3 per 100 000) than male and female urban residents In rural areas, there was inad-
in urban areas (102.5 per 100 000) was estimated to be 7.09 years and equate allocation of basic educa-
[2], indicating that cancer survival 6.64 years longer, respectively, than tional resources, and teachers were
was lower in rural areas than in ur- that of their rural counterparts [10]. less highly trained than in urban
ban areas. Differences were also Mainly as a result of the one-child areas [12]. The education level of
seen between urban and rural ar- policy and increases in life expectan- rural residents was also generally
Chapter 4.3 • Cancer in urban and rural communities in China 253

lower than that of their urban coun- than in urban areas, as a result of etables (see Chapter 2.6), are more
terparts [13]. In addition, utilization differences in health-care services, prevalent in rural areas than in ur-
of health-care services of all types socioeconomic inequalities, and ban areas [2,16,17]. Higher rates of
was lower in rural areas than in ur- lack of awareness about cancer Helicobacter pylori and hepatitis B
ban areas [14], as a result of the prevention and early detection, as virus infection also contribute to the
unbalanced development between well as the unbalanced allocation of high incidence of stomach cancer
urban and rural areas in the provi- health-care resources, with lower and liver cancer, especially in ru-
sion of health-care services. government health expenditure per ral areas (see Chapter 2.2) [22,23].
These differences in socioeco- capita and less advanced health- The lower quality of medical treat-
nomic status between urban and ru- care facilities in rural areas [6,10]. ment and limited health-care re-
ral areas could lead to differences in Age-standardized incidence rates sources led to lower survival in rural
lifestyles and dietary patterns. For of oesophageal cancer, stomach areas [24].
example, the prevalence of smok- cancer, liver cancer, and cervi-
ing (see Chapter 2.1) and alcohol cal cancer were higher in rural ar-
consumption (see Chapter 2.3) was eas than in urban areas, and were
Cancer patterns and
still higher in rural residents, where- higher in areas with low GDP per trends in urban and
as in urban residents the level of capita and low urbanization. Strong rural areas
physical activity was relatively low risk factors for cancer, including In recent decades, the overall can-
(see Chapter 2.7), as a result of in- smoking, alcohol consumption, cer incidence in China has been
creasingly sedentary occupations and low intake of fruits and veg- relatively stable, with a total annual
[15]. Surveys also showed that the
intake of animal products is sig-
nificantly higher in urban residents Fig. 4.3.3. Rapid increases in (a) the numbers of hospital beds and of registered doc-
than in rural residents; this may tors and (b) health expenditures in China, during the period 1980–2014.
contribute to differences in energy a
intake [15]. Problems associated
with rapid urbanization, including
large-scale migration, ageing of the
population, and pollution in both ur-
ban and rural areas (see Chapter
2.9), have also emerged [10].
Age-standardized incidence rates
of colorectal cancer, breast cancer,
prostate cancer, kidney cancer,
and bladder cancer were higher
in urban areas than in rural areas,
and were higher in areas with high
gross domestic product (GDP)
per capita and high urbanization
[16,17]. Obesity and physical inac-
tivity, which are the leading attribu-
table risk factors for both colorec- b
tal cancer and breast cancer, are
more prevalent in urban areas than
in rural areas, not only in China but
also worldwide; differences in the
prevalence of obesity and physical
inactivity are partly responsible for
the rural–urban disparity in the in-
cidence of these two cancer types
[18–20]. Changes in reproductive
factors, such as increasing expo-
sure to xeno-estrogens and oral
contraceptives, may also lead to a
higher incidence of breast cancer
in urban areas [21]. For colorectal
cancer and breast cancer, cancer
survival was lower in rural areas
254
change of 4% in the crude incidence the prevalence of cigarette smoking ciated with air pollution, radiation,
rate, whereas cancer mortality has will emerge in the future [26]. and other cancer risk factors, such
decreased [4,5]. During the past 20 years, there as exposure to asbestos, which
From 2003–2005 to 2012–2015, has been a rapid upward trend in could lead to the development
age-standardized 5-year relative the incidence of breast cancer and of cancer [35]. According to the
survival increased significantly for colorectal cancer, especially in ur- Hukou policy in China, when a mi-
all cancers combined, from 30.9% ban areas [4,5]. From the 1970s to grant is diagnosed with cancer, the
to 40.5%; age-standardized 5-year the 1990s, liver cancer, stomach case will be registered in the rural
relative survival also increased for cancer, and oesophageal cancer cancer registry where the person
most cancer types, including can- were the most common cancers in was born [36]. Another explanation
cers of the oesophagus, stomach, both urban and rural areas [4,5]. for the high cancer burden in rural
larynx, bone, cervix, uterus, blad- Oesophageal cancer, stomach areas could be the lack of aware-
der, and thyroid [6]. This reflected cancer, and liver cancer are still the ness among rural residents about
the overall improvement in the qual- major cancer types in rural residents health care and cancer prevention
ity of cancer care in China, which [5]. Declining trends in age-stan- [6]. As a result, the willingness to
could be shown partly by an annual dardized incidence rates and mor- participate in cancer screening pro-
increase in health-care resources, tality rates were observed for these grammes and the subsequent fol-
including the numbers of hospital three cancer types in both sexes in low-up is lower in rural areas than
beds and of registered doctors, as 2000–2013. These declines are a in urban areas, even if the screen-
well as increases in health expendi- result of socioeconomic developing is provided free of charge.
tures (Fig. 4.3.3) [25]. ment and a series of cancer pre-
During the past 40 years, the vention and control programmes,
such as comprehensive interven-
Conclusions
lung cancer mortality rate in China
has increased 4-fold. Consequently, tion and control strategies imple- Global experience in alleviating the
lung cancer has replaced stomach mented in high-risk rural ar eas cancer burden has demonstrated
cancer as the leading cause of since the 1990s and early detection the importance of comprehensive
cancer death [4,5], accounting for programmes initiated in rural ar- strategies such as tobacco control
27.3% of all cancer deaths in China. eas and aimed at specific high-risk campaigns, vaccination, targeted
Although the prevalence of tobacco cancer types [27–30]. Control of in- cancer screening programmes, and
smoking is slowly decreasing in fections, including hepatitis B virus appropriate and efficient diagnostic
China, the development of lung can- and hepatitis C virus for liver cancer and treatment technology. In China,
cer may take decades. Therefore, and H. pylori for stomach cancer, although some cancer preven-
the new cases of lung cancer may may also contribute to these tem- tion and control programmes have
CHAPTER 4.3
SECTION 4
be the result of a high prevalence of poral patterns [22,23]. Studies have yielded significant benefits, chal-
smoking in the past. The effects of shown that the food policy reforms lenges still remain because of the
current anti-smoking campaigns on in China dramatically decreased heavy cancer burden, the compli-
exposure to aflatoxin and reduced cated cancer patterns, and the un-
overall liver cancer risk in Qidong, balanced allocation of health-care
a city in Jiangsu Province, even resources and primary health care
Fig. 4.3.4. Colourized scanning electron
micrograph of Helicobacter pylori and hu- before universal hepatitis B virus between urban and rural areas.
man gastric epithelium cells. vaccination of newborns was imple- The distinct differences in can-
mented [31,32]. cer patterns between urban and
Although differences in cancer rural communities emphasize an
incidence between urban and rural urgent need to implement cancer
areas still exist in China, the gap prevention and control strategies
has been narrowing every year. that are customized for different re-
Cancer incidence in rural areas is gions of the country. For example,
predicted to surpass that in urban the hazards associated with smok-
areas in the future [33,34]. As a re- ing were previously more severe in
sult of rapid urbanization, a large- urban areas, because of the lim-
scale migration from rural to urban ited availability and affordability of
areas is occurring [10]. Although cigarettes in rural areas. However,
migrants move to cities seeking a this difference is diminishing and
better life, most of them can only the situation is even likely to be
find jobs in areas like construction, reversed, because rural residents
manufacture, or mining, because of start smoking at a younger age and
their comparatively lower education with a somewhat higher prevalence
level. Most of these jobs are asso- than urban residents [37].

It is also important to further public awareness of cancer preven- mental and psychological health.
improve the primary health-care tion, and develop programmes for The effective implementation of
system in rural areas, including a the early detection and treatment targeted early diagnosis and treat-
more comprehensive design and of major cancer types that focus on ment programmes is also crucial in
implementation of the health insur- rural residents. urban areas.
ance system, which can effectively For urban residents, the points In addition, international cooper-
serve low-income residents of ru- of focus are (i) to promote healthy ation should be enhanced, to learn
ral areas. Moreover, it is of great lifestyles and dietary habits, (ii) to from useful experiences and ap-
importance to improve basic living control smoking, alcohol consump- proaches and to avoid common pit-
and sanitary conditions, strengthen tion, and obesity, and (iii) to improve falls and unnecessary expenditures.
References
1. Ferlay J, Ervik M, Lam F, Colombet M, Mery 9. National Bureau of Statistics of China 17. Chen W, Zheng R, Zhang S, Zeng H,
L, Piñeros M, Znaor A, Soerjomataram I, (2016). Statistical Communiqué of the Peo Zuo T, Xia C, et al. (2017). Cancer inci-
Bray F (2018). Global Cancer Observatory: ple’s Republic of China on the 2015 Na dence and mortality in China in 2013:
Cancer Today. Lyon, France: International tional Economic and Social Development. an analysis based on urbanization
Available from: http://www.stats.gov.cn/ level. Chin J Cancer Res. 29(1):1–10.
Agency for Research on Cancer. Available
english / Pres s Release / 2016 02 / t 2016
from: https://gco.iarc.fr/today. ht t p s: //d o i.o r g /10. 21147/ j.i s sn.10 0 0 -
0229_1324019.html.
9604.2017.01.01 PMID:28373748
2. Chen W, Sun K, Zheng R, Zeng H, 10. Yang J, Siri JG, Remais JV, Cheng
Zhang S, Xia C, et al. (2018). Cancer inci- Q, Zhang H, Chan KKY, et al. (2018). 18. Fang C, Liang Y (2017). Social dispari-
dence and mortality in China, 2014. Chin The Tsinghua-Lancet Commission on ties in body mass index (BMI) trajecto-
J Cancer Res. 30(1):1–12. https://doi. Healthy Cities in China: unlocking the ries among Chinese adults in 1991-2011.
org/10.21147/j.issn.1000-9604.2018.01.01 power of cities for a healthy China. Int J Equity Health. 16(1):146. https://
PMID:29545714 Lancet. 391(10135):2140–84. https://doi. d o i.o r g /10 .118 6 /s12 9 3 9 - 017- 0 6 3 6 - 5
org/10.1016/S0140 - 673 6(18) 3 0 48 6 - 0 PMID:28814339
3. Cai Y, Xue M, Chen W, Hu M, Miao Z, PMID:29678340
Lan L, et al. (2017). Expenditure of hospi- 19. Attard SM, Howard AG, Herring AH,
tal care on cancer in China, from 2011 to 11. Song Y (2014). Losing an only child: the Zhang B, Du S, Aiello AE, et al. (2015).
one-child policy and elderly care in China. Differential associations of urbanicity and
2015. Chin J Cancer Res. 29(3):253–62.
Reprod Health Matters. 22(43):113–24. income with physical activity in adults in
ht t p s: //d o i.o r g /10. 21147/ j.i s sn.10 0 0 -
https://doi.org/10.1016/S0968-8080(14)
9604.2017.03.11 PMID:28729776 urbanizing China: findings from the pop-
43755-8 PMID:24908462
ulation-based China Health and Nutrition
4. Chen Z (2008). Report on Third National 12. Zhang JY, Chen T (2009). Empirical Survey 1991-2009. Int J Behav Nutr Phys
Retrospective Sampling Survey on Causes analysis on the allocation of the urban Act. 12(1):152. https://doi.org/10.1186/
of Death in China. Beijing, China: Chinese and rural educational resource in China. s12966-015-0321-2 PMID:26653097
Union Medical University Press. Zhongguo Nong-Jihua. (6):115–9.
20. Islami F, Chen W, Yu XQ, Lortet-Tieulent
5. Chen W, Zheng R, Zhang S, Zeng H, Xia 13. National Bureau of Statistics of China J, Zheng R, Flanders WD, et al. (2017).
C, Zuo T, et al. (2017). Cancer incidence (2010). Tabulation on the 2010 Population Cancer deaths and cases attributable to
and mortality in China, 2013. Cancer Lett. Census of the People’s Republic of China.
lifestyle factors and infections in China,
401:63–71. https://doi.org/10.1016/j.canlet. Available from: http://www.stats.gov.cn/
2013. Ann Oncol. 28(10):2567–74. https://
tjsj/pcsj/rkpc/6rp/indexch.htm.
2017.04.024 PMID:28476483 d o i . o r g / 1 0 .1 0 9 3 / a n n o n c / m d x 3 4 2
14. Haggerty JL, Roberge D, Lévesque JF, PMID:28961829
6. Zeng H, Chen W, Zheng R, Zhang S, Ji
Gauthier J, Loignon C (2014). An explora-
JS, Zou X, et al. (2018). Changing cancer tion of rural-urban differences in health- 21. Fan L, Strasser-Weippl K, Li JJ, St Louis
survival in China during 2003-15: a pooled care-seeking trajectories: implications for J, Finkelstein DM, Yu KD, et al. (2014).
analysis of 17 population-based cancer measures of accessibility. Health Place. Breast cancer in China. Lancet Oncol.
registries. Lancet Glob Health. 6(5):e555– 28:92–8. https://doi.org/10.1016/j.health 15(7):e279 –89. https://doi.org/10.1016/
67. https://doi.org/10.1016/S2214-109X(18) place.2014.03.005 PMID:24793139 S1470-2045(13)70567-9 PMID:24872111
30127-X PMID:29653628
15. Zeng Q, Zeng Y (2018). Eating out and get- 22. Nagy P, Johansson S, Molloy-Bland M
7. Chen Y, Yin Z, Xie Q (2014). Suggestions ting fat? A comparative study between ur- (2016). Systematic review of time trends
to ameliorate the inequity in urban/ruban and rural China. Appetite. 120:409–15. in the prevalence of Helicobacter py-
ral allocation of healthcare resources https://doi.org/10.1016/j.appet.2017.09.027 lori infection in China and the USA. Gut
PMID:28964905
in China. Int J Equity Health. 13(1):34. Pathog. 8(1):8. https://doi.org/10.1186/
https://doi.org/10.1186/1475-9276-13-34 16. Yang Z, Zheng R, Zhang S, Zeng H, s13099-016-0091-7 PMID:26981156
PMID:24884614 Xia C, Li H, et al. (2017). Comparison
of cancer incidence and mortality in 23. Cui Y, Jia J (2013). Update on epidemi-
8. Shi L (1993). Health care in China: a ru- three GDP per capita levels in China, ology of hepatitis B and C in China. J
ral-urban comparison after the socioeco- 2013. Chin J Cancer Res. 29(5):385–94. Gastroenterol Hepatol. 28(Suppl 1):7–
nomic reforms. Bull World Health Organ. ht t p s: //d o i.o r g /10. 21147/ j.i s sn.10 0 0 - 10. https://doi.org/10.1111/jgh.12220
71(6):723–36. PMID:8313490 9604.2017.05.02 PMID:29142457 PMID:23855289
256
24. Miao J, Wu X (2016). Urbanization, socio- 29. Chen JG, Parkin DM, Chen QG, Lu 34. Zheng RS, Gu XY, Li XT, Zhang SW,
economic status and health disparity in JH, Shen QJ, Zhang BC, et al. (2003). Zeng HM, Sun KX, et al. (2018). Analysis
China. Health Place. 42:87–95. https://doi. Screening for liver cancer: results of a on the trend of cancer incidence and age
org/10.1016/j.healthplace.2016.09.008 randomised controlled trial in Qidong, change in cancer registry areas of China,
PMID:27750075 China. J Med Screen. 10(4):204–9. https:// 2000 to 2014. [in Chinese] Zhonghua
doi.org/10.1258/096914103771773320 Yu Fang Yi Xue Za Zhi. 52(6):593–600.
25. National Bureau of Statistics of China PMID:14738659 PMID:29886680
(2015). China Statistical Yearbook.
Beijing, China: China Statistics Press. 30. Zhao P, Dai M, Chen W, Li N (2010). 35. Mou J, Griffiths SM, Fong H, Dawes MG
Available from: http://www.stats.gov.cn/ Cancer trends in China. Jpn J Clin Oncol.
(2013). Health of China’s rural-urban mi-
tjsj/ndsj/2015/indexeh.htm. 40(4):281–5. https://doi.org/10.1093/jjco/
grants and their families: a review of lit-
hyp187 PMID:20085904
erature from 2000 to 2012. Br Med Bull.
26. WHO (2018). WHO global report on
trends in prevalence of tobacco smok- 31. Chen JG, Egner PA, Ng D, Jacobson 106(1):19–43. https://doi.org/10.1093/bmb/
ing 2000–2025. 2nd ed. Geneva, LP, Muñoz A, Zhu YR, et al. (2013). ldt016 PMID:23690451
Reduced aflatoxin exposure presages
Switzerland: World Health Organization. 36. Qian Y, Ge D, Zhang L, Sun L, Li J, Zhou
decline in liver cancer mortality in an
Available from: http://www.who.int/to C (2018). Does Hukou origin affect estab-
endemic region of China. Cancer Prev
bacco/publications/surveillance/trends- lishment of health records in migrant in-
Res (Phila). 6(10):1038–45. https://doi.
tobacco-smoking-second-edition/en/. flow communities? A nation-wide empiri-
org/10.1158/1940 - 6207.CAPR-13- 0168
27. Chen TJ, Lian SY, Liu ZC, Lu JB, Sun PMID:23963804 cal study in China. BMC Health Serv Res.
XB, Wei WQ (2010). Retrospection and 18(1):704. https://doi.org/10.1186/s12913-
32. Sun Z, Chen T, Thorgeirsson SS, Zhan Q,
prospection on esophageal cancer scene 018-3519-6 PMID:30200941
Chen J, Park JH, et al. (2013). Dramatic re-
in Linzhou City, Henan Province. China duction of liver cancer incidence in young 37. Chen Z, Peto R, Zhou M, Iona A, Smith
Cancer Chinese Journal. 19(1):24–8. adults: 28 year follow-up of etiological in- M, Yang L, et al.; China Kadoorie
terventions in an endemic area of China. Biobank (CKB) collaborative group
28. Qiao YL, Dawsey SM, Kamangar F, Fan
Carcinogenesis. 34(8):1800–5. https://doi. (2015). Contrasting male and female
JH, Abnet CC, Sun XD, et al. (2009). Total
org/10.1093/carcin/bgt007 PMID:23322152 trends in tobacco-attributed mortali-
and cancer mortality after supplementa-
tion with vitamins and minerals: follow- 33. Chen W (2016). Discussion on the clinical ty in China: evidence from successive
up of the Linxian General Population characteristics and trends of cancers in nationwide prospective cohort studies.
Nutrition Intervention Trial. J Natl China according to cancer registry data. Lancet. 386(10002):1447–56. https://doi.
Cancer Inst. 101(7):507–18. https://doi. Zhonghua Jiankang Guanlixue Zazhi. org/10.1016/S0140 - 673 6(15) 0 03 40 -2
org/10.1093/jnci/djp037 PMID:19318634 10(4):249–52. PMID:26466050
CHAPTER 4.3
SECTION 4

4.4
Socioeconomic factors and cancer

prevention in India
Diverse interventions are needed
Rengaswamy Sankaranarayanan Cindy L. Gauvreau (reviewer) Aswathy Sreedevi (reviewer)
Kunnambath Ramadas Mohandas K. Mallath (reviewer)
Filip Meheus (reviewer)
●● As the reduction of socioeco- an increasing adoption of sedentary

SUMMARY nomic inequalities in population lifestyles and lower levels of physi-
groups in India is addressed, cal activity (see Chapter 2.7), and an
●● Cancer incidence rates differ highly focused and tailored increasing risk of noncommunicable
markedly within India. In the public health interventions are diseases, including cancer [2].
north-eastern state of Mizoram, needed to target different socio- Socioeconomic factors such as
1 in 5 men and women will de- economic groups to reduce the education level, income, occupation,
velop cancer during their life- disparities in cancer prevention. and standard of living determine the
times, compared with 1 in 22
social standing of an individual or
men and 1 in 18 women in the
a population in terms of low, mid-
Barshi region.
During the past two decades, India dle, and high socioeconomic status.
●● There are currently 164 million has had one of the world’s best Compared with people with high so-
users of smokeless tobacco, performing and most stable econ- cioeconomic status, those with low
69 million smokers, and 42 mil- omies, which has grown by more socioeconomic status are resource-
lion smokers and chewers in than 7% annually in most years, de- constrained. The vast differences
India. More than 90% of patients spite a global economic slowdown. in socioeconomic factors within a
with oral cancer have low or low- This economic development has country can lead to significant dis-
er-middle socioeconomic status. given rise to vast socioeconomic parities in access to cancer preven-
●● Among people with lower socio- changes, with improvements in life tion and control services.
economic status, non-aware- expectancy and education and re- Cancer disparities refer to dif-
ness of the harms of tobacco ductions in rates of poverty, hunger, ferences in cancer occurrence, the
use in any form and of chewing and malnutrition. Between 1990 availability of and access to cancer
products that contain areca nut and 2017, the value of the Human health services, cancer survival,
is common, as is inadequate Development Index (HDI) for India cancer deaths, quality of life, and
comprehension of the associ- increased from 0.427 to 0.640, an the adverse economic impact of
ated health risks. increase of about 50%, and the cancer in populations. There is con-
country’s gross national income per vincing evidence that the striking
●● Urbanization appears to be capita increased by 267% [1]. socioeconomic differences among
associated with an increas- However, in a large country like various regions and states in India
ing incidence of breast can- India, consideration of aggregate are a major responsible factor for
cer. Similarly, the incidence of economic indicators may hide in- the cancer disparities observed in
colorectal cancer is increasing equalities of socioeconomic prog- the country [3]. Cancer control initia-
in the most developed states in ress and of HDI. For instance, four tives can reduce disparities across
India and in urban populations. of the five most developed states are the country only if such initiatives go
●● Given the focus of primary pre- in southern India, and all nine states hand in hand with policies and pro-
vention on health literacy, aware- with HDI values less than the nation- grammes directed towards the rapid
ness, and behaviour change, al average are in northern and east- elimination of poverty and illiteracy,
addressing the socioeconom- ern India. Unfortunately, the prog- an increase in purchasing power to
ic determinants that influence ress in economic development is improve the affordability and acces-
these factors is critical to ad- associated with an increasing prev- sibility of healthy foods, and the al-
vance cancer prevention in India. alence of overweight and obesity, leviation of social inequalities.
258
Cancer prevention aims to re- Cancer burden and

duce the burden of cancer (i) by de- patterns in India FUNDAMENTALS
creasing the frequency of new cases
For the age-standardized incidence
of cancer, by avoiding or reducing ■■ During the past two decades,
rate of all cancers observed during
exposure to cancer risk factors, and
2012–2014, there was an almost India has had one of the
(ii) by detecting and treating precan-
cerous lesions through screening 7-fold difference between the lowest world’s best performing and
programmes linked with diagnosis and highest reported rates in men most stable economies,
and treatment. Socioeconomic fac- (40.9 per 100 000 in the Barshi ex- which has grown by more
tors play a major role in determining panded rural registry vs 270.7 per
than 7% annually in most
the exposure of an individual and 100 000 in Aizawl district in Mizoram
years. This economic
a population to cancer risk factors. state) and an almost 5-fold differ-
ence in women (52.0 per 100 000 in development has given
Socioeconomic factors also affect
the Barshi expanded rural registry rise to vast socioeconomic
the behaviour patterns of the popu-
lation, in adopting lifestyles condu- vs 249.0 per 100 000 in Papumpare changes, with an increasing
cive to cancer prevention, including district in the state of Arunachal risk of noncommunicable
a healthy diet and adequate physical Pradesh) [4]. These rates indicate diseases, including cancer,
activity, among others, and in ac- that in the north-eastern state of
and significant disparities in
cessing cancer prevention services, Mizoram, 1 in 5 men and women will
access to cancer prevention
such as vaccination, screening, and develop cancer during their lifetimes,
compared with 1 in 22 men and 1 in and control services.
treatment of cancer precursor le-
sions (see Chapter 6.1). 18 women in the Barshi region. ■■ Cancer patterns in India are
The inherent differences in socio- The estimated cancer burden in
dominated by a high burden
economic development and cultural India in 2018 is given in Box 4.4.1
[6]. Six cancer types – breast can- of tobacco-related head and
practices across India are reflected
in the major differences observed in cer, oral cancer, cervical cancer, neck cancers, particularly
cancer incidence and patterns, as lung cancer, stomach cancer, and oral cancer, in men and of
documented by data provided by the colorectal cancer – together account cervical cancer in women;
29 population-based cancer registries for almost half of the new cancer both of these cancer types
under the National Cancer Registry cases occurring in India. Whereas are associated with lower
Programme of the Indian government tobacco-related cancers account for
socioeconomic status.
[4]. Given the focus of primary preven- 34–69% of all cancers in men, they
tion on health literacy, awareness, constitute 10–27% of all cancers in ■■ The burden of cancer types
CHAPTER 4.4
SECTION 4
and behaviour change, addressing women in most regions in India. associated with overweight
the socioeconomic determinants that Increasing trends (e.g. breast
and obesity, lower levels of
influence these factors is critical to ad- cancer, colorectal cancer) or de-
physical activity, and seden-
vance cancer prevention in India [5]. creasing trends (e.g. cervical can-
tary lifestyles, such as breast
cancer and colorectal cancer,
Box 4.4.1. Cancer burden and patterns in India in 2018. is increasing, and these can-
cer types are associated with
• There are an estimated 1.16 mil- lung cancer (49 000), stomach higher socioeconomic status.
lion new cancer cases, 784 800 cancer (39 000), colorectal can-
cancer deaths, and 2.26 million cer (37 000), and oesophageal
5-year prevalent cases in India’s cancer (34 000) account for 45%
population of 1.35 billion. of cases. cer) in the incidence of the major
• The six most common cancer • Of the 587  000 new cancer cancer types over time (since the
types are breast cancer (162 500 cases in women, breast can- documentation of incidence began
cases), oral cancer (120 000 cas- cer (162 500), cervical cancer in different cancer registries) are
es), cervical cancer (97 000 cas- (97 000), ovarian cancer (36 000), evident with the socioeconomic
es), lung cancer (68 000 cases), oral cancer (28 000), and colorec- changes that are occurring in differ-
stomach cancer (57 000 cases), tal cancer (20 000) account for ent regions and states in India [4,7].
and colorectal cancer (57 000); 60% of cases. Recently, an increasing trend in the
together, these account for 49% • 1 in 10 Indians will develop can- incidence of oral cancer has been
of all new cancer cases. cer during their lifetimes, and 1 observed among men in the fourth
• Of the 570 000 new cancer cas- in 15 Indians will die of cancer. to seventh decades of life [4], possi-
es in men, oral cancer (92 000), bly as a result of the increasing con-
sumption of unregulated flavoured
Chapter 4.4 • Socioeconomic factors and cancer prevention in India 259

chewing products that contain are- Fig. 4.4.1. Trends in age-standardized incidence rates (per 100 000 women) of breast
ca nut, such as paan masala [8]. cancer in selected populations in India, 1983–2015.
There is a clear increasing trend
in the incidence rates of breast can- 45
41.0
cer across the country, with an an- 40 40.8
nual percentage increase that ranges
from 1.4% to 2.8% and is more pro- 35 34.4
nounced in urban areas than in rural 28.7 33.6
30
areas (Fig. 4.4.1). Incidence rates are
also increasing for cancer types as- 24.6
25
sociated with overweight and obesity
20 18.9
and lower levels of physical activ-
ity, such as colorectal cancer (an- 15 18.2
nual percentage change, 1.0–3.9%), 12.4
uterine cancer (annual percentage 10
change, 2.7–5.5%), ovarian cancer 8.7
5
(annual percentage change, 0.8–
2.4%), and prostate cancer (annual 0
percentage change, 1.2–4.1%). 1983-87 1988-92 1993-97 1998-02 2003-07 2008-11 2012-15
There is a clear decreasing Bangalore Chennai Mumbai Delhi Barshi
trend in the incidence rates of cer-
vical cancer in most regions in
India (annual percentage change, Fig. 4.4.2. Trends in age-standardized incidence rates (per 100 000 women) of cervical
−2.0% to −3.5%), with age-stan- cancer in selected populations in India, 1983–2015.
dardized incidence rates as low as
6 per 100 000 in women in Kerala 50
[4] (Fig. 4.4.2). However, rates of 44.5
45
cervical cancer are still high in less
educated women with low socioeco- 40
nomic status [7]. 35
The underlying socioeconomic
factors and changes that influence 30
31.1 25.8
risk factors, exposure patterns, pat- 25 27.4
terns of health beliefs, health-seek- 19.3
ing behaviours, and the availability
20 16.0
15.5
of and access to health-care ser- 15 15.3
vices are largely responsible for the 10 14.9
observed cancer patterns in India. 9.0
5
Socioeconomic factors 0
and cancer prevention 1983-87 1988-92 1993-97 1998-02 2003-07 2008-11 2012-15
Bangalore Chennai Mumbai Delhi Barshi
Prevention of lung cancer,
oral cancer, and other
tobacco-related cancers
Socioeconomic determinants of to- Recent studies indicate that be- has increased in all socioeconom-
bacco use patterns have a major tween 2000 and 2012, the preva- ic groups, with a greater increase in
impact on the prevention of cancer lence of any form of tobacco use households with higher income and
types associated with tobacco use, decreased in the richest house- higher education levels, and the
such as lung cancer, oral cancer, and holds (from 43.8% to 36.8%) and re- volume of smokeless tobacco and
other head and neck cancers (see mained stable in the poorest house- areca nut products used is increas-
Chapter 2.1). There are currently holds (from 61.5% to 62.7%) [9]. ing [10]. The reported prevalence
164 million users of smokeless tobac- Despite the implementation of pre- of tobacco use in tribal populations
co, 69 million smokers, and 42 million ventive interventions, in India there exceeded 80%.
smokers and chewers in India, and is a distinct and unique pattern of Because inadequate attention
tobacco-related cancers constitute a tobacco use; the use of smokeless has been paid to curtailing the use
major burden in the country. tobacco and areca nut products of smokeless tobacco and areca
260
Fig. 4.4.3. A woman in India rolling bidis. wrapped in dried temburni leaf), as

use of smokeless tobacco in the
form of chewing paan (a mixture of
lime, pieces of areca nut, cured to-
bacco, and spices wrapped in betel
leaf) and many other forms, such as
tobacco-containing paan masala,
gutka (tobacco with crushed areca
nut, wax, catechu, slaked lime, and
sweet flavourings), khaini, mishri
(burned tobacco), zarda (boiled to-
bacco), mawa (tobacco, lime, and
areca nut), or as dual use (both
smoking and chewing).
The prevalence of tobacco use
in any form exceeds 60% in adult
men (age 15 years and older) in the
north-eastern states in India and in
the less-developed states, such as
Bihar, Jharkhand, Chhattisgarh, and
Madhya Pradesh, and exceeds 45%
in West Bengal, Uttar Pra desh,
Rajasthan, Uttarakhand, Odisha, and
Gujarat [14]. The prevalence of to-
bacco use (mostly as chewing) in
adult women exceeds 40% in the
nut products, the anti-tobacco poli- cioeconomic status; use of various north-eastern states and in Bihar,
cies need to be reviewed to address forms of tobacco and chewing of Chhattisgarh, and Odisha [15].
inequalities in their use. Although flavoured products that contain are- Paan masala is packed in at-
11 states in India have banned all ca nut, such as paan masala, are tractive, user-friendly packets and
forms of smokeless tobacco, vari- more common among people with containers. Increasing disposable
ous tobacco chewing products are lower socioeconomic status [14]. incomes, convenient packaging,
still clandestinely sold. In India, tobacco use occurs aggressive advertising campaigns
CHAPTER 4.4
SECTION 4
Oral cancer is the major tobac- as smoking of cigarettes and bidis by manufacturers, and the large-
co-related cancer type in India, and (made of shredded tobacco leaves scale switching by consumers from
low socioeconomic status is associ-
ated with a high risk of oral cancer
and precancerous lesions such as Fig. 4.4.4. A man in West Bengal, India, holding gutka in his hand.
leukoplakia, erythroplakia, and oral
submucous fibrosis (see Chapter
5.2) [11–13]. Alcohol consumption
is an independent risk factor and
substantially increases the risk of
oral cancer when combined with
tobacco use. In India, substantial
differences exist in the sociodem-
ographic correlates of alcohol con-
sumption and types of alcoholic
beverages.
Socioeconomic disadvantages
appear to have a cumulative effect
over the life course and are associ-
ated with a high risk of oral cancer.
Early-life socioeconomic disadvan-
tages have a lasting effect on oral
cancer risk in adulthood [12]. More
than 90% of patients with oral can-
cer have low or lower-middle so-

tobacco products to paan masa- Fig. 4.4.5. Mobile oral cancer screening in India.
la are currently encouraging the
growth of the paan masala market.
The Indian paan masala market
was valued at about US$ 5 billion in
2017 and is expected to increase to
US$ 8 billion by 2023.
In 2016, after a Supreme Court
order, the central government issued
a complete ban across India on the
production, promotion, and sale of
food products containing tobacco
and nicotine as ingredients, includ-
ing gutka, paan masala, zarda, and
tobacco-based flavoured mouth
fresheners. However, several states
have yet to follow suit, and illegal
sales continue (see Chapter 6.8).
Among people with lower socio-
economic status, non-awareness
of the harms of tobacco use in any
form and of chewing products that
contain areca nut is common, as is improvements in the socioeconom- economic status were found to be
inadequate comprehension of the ic status of women, as indicated by more likely to participate in screen-
associated health risks. The use of higher education levels, increasing compared with other women
hookah (water pipes) and e-ciga- ing household incomes, later ages [19]. In a recent study in Mumbai,
rettes is increasing among young at marriage and at first birth, lower women with higher socioeconomic
people, and this is creating a new parity, and increasing adoption of status were found to have higher
problem. There is an urgent need to sedentary lifestyles, dietary pat- breast awareness than women with
create comprehensive awareness terns typical of industrialized coun- lower socioeconomic status [20].
about the health hazards of all forms tries, and lower levels of physical Two large randomized trials
of tobacco and areca nut use among activity in successive generations of screening by clinical breast ex-
every subsection of society and to of women (see Chapter 5.9). amination in India have shown that
regulate the availability, affordabil- The most developed states clinical breast examination screen-
ity, and accessibility of tobacco and report the highest breast cancer ing is followed by early diagnosis of
areca nut products, to prevent all rates in the country [4]. In India, breast cancer [21,22]. Findings from
tobacco-related cancers. high socioeconomic status is as-
a randomized trial in Kerala indi-
In a randomized trial of oral can- sociated with a higher prevalence
cated that women who had a higher
cer screening with oral visual inspec- of overweight and obesity and with
education level and a higher house-
tion in Kerala, which demonstrated a shift towards sedentary lifestyles
hold income, were employed in
a significant reduction in oral cancer and dietary patterns typical of in-
non-manual occupations, and were
mortality in users of tobacco or al- dustrialized countries, which are
living in better housing were more
cohol or both, participation was sig- established risk factors for breast
nificantly higher among people with likely to have breast awareness and
cancer; households with high so-
higher socioeconomic status than cioeconomic status spend less on to practice breast self-examination
among those with lower socioeco- cereals, millets, and vegetables but less likely to participate in clini-
nomic status [16,17]. and more on beverages, processed cal breast examination screening,
foods, dairy products, meat, eggs, which was offered in the trial by the
Breast cancer control and fish [18]. public health services [23]. A pos-
In India, the incidence of breast The most effective intervention sible explanation for these para-
cancer is consistently increas- for breast cancer control is early doxical findings is that women with
ing and the incidence of cervical detection and prompt treatment. higher socioeconomic status have
cancer is decreasing with time, as Breast awareness and participation less faith in public health services,
shown by data from several popu- in screening are conducive to early can afford private health care, and
lation-based cancer registries [4]. detection and completion of treat- seek mammography screening
The diverging incidence trends for ment. In a cross-sectional study of elsewhere. Similar findings were re-
breast cancer and cervical cancer breast cancer screening practices ported in a breast cancer screening
in India may be partly explained by in Kerala, women with higher socio- trial in Mumbai [22].
262
Cervical cancer prevention similar protection against persistent in India [4]. Colorectal cancer, for
India accounts for about one fifth of HPV16 and 18 infections as three which incidence rates in India were
the global burden of cervical cancer, doses and has shown that even a previously low, is already the sixth
despite decreasing incidence rates single dose is immunogenic and most common cancer (Box 4.4.1),
in several regions of the country (see provides lasting protection against and increasing trends are evident in
Chapter 5.10). Thus, elimination of HPV16 and 18 infections, similar to the most developed states in India
the three-dose and two-dose vac- and in urban populations [4,6]. To
cervical cancer in India will have a
cine schedules [30,31]. Currently, curtail the future burden of these
major impact on global elimination
Punjab is implementing two doses lifestyle-related cancer types, in-
of the disease as a public health
of HPV vaccination in an incremen- cluding breast cancer, it is critical
problem. Cervical cancer dispropor-
tal fashion, and Sikkim has imple- to reverse the emerging trends in
tionately affects women with lower
mented a statewide HPV vaccina- risk factors and to preserve the
socioeconomic status, who are at
tion programme targeting girls aged lifestyles that kept the incidence of
a considerable disadvantage in the
11–12 years, with high vaccination these cancer types low.
availability of and access to public
coverage and an excellent safety
health services for prevention and
profile. Delhi state is implementing
early detection, and therefore this is
opportunistic HPV vaccination sup-
Conclusions
an equity issue. Low socioeconomic Because cancer is not one disease
ported by the state government.
status is a major risk factor for cervi- but a group of many diseases that
Despite the decreasing inci-
cal cancer [24]. differ in their etiology and biology, it
dence of cervical cancer, there is a
It is well established that persis- is not surprising that socioeconom-
6-fold difference in age-standardized
tent infection with one of the high- rates, ranging from 5 per 100 000 ic determinants of cancer risk are
risk human papillomavirus (HPV) women to 30 per 100 000 women, variable for different cancer types,
types is the necessary cause of reflecting the underlying differences reflecting the underlying complex
cervical cancer. HPV types 16 and in socioeconomic factors and HPV relationships. There is a positive
18 are detected in about 80% of all prevalence, among other risk fac- association of low socioeconomic
cervical cancers in India [25]. Low tors [4]. Incidence rates are about 6 status with the incidence of tobac-
socioeconomic status is associated per 100 000 women in Kerala, which co-related cancer types. However,
with a high prevalence of HPV in- has achieved 100% literacy and has improvements in education, in-
fection in India [26,27]. Cervical the highest HDI value (0.784) of any creasing disposable incomes, and
cancer is an eminently preventable state in the country [4]. Because higher overall socioeconomic status
disease, by HPV vaccination and cervical cancer disproportionately are associated with an increasing
screening. The decreasing inci- affects women with low socioeco- risk of breast cancer and colorectal
CHAPTER 4.4
SECTION 4
dence rates of cervical cancer nomic status, the lack of effective cancer, among other lifestyle-relat-
provide an exciting opportunity to interventions such as HPV vaccina- ed cancer types.
rapidly decrease risk and eliminate tion and screening in public health The limited available data in-
cervical cancer by implementing services will widen the disparities dicate disparities in participation
an integrated HPV vaccination and and increase the inequities in the in cancer screening by socioeco-
screening programme. cervical cancer burden in India. nomic status. Good participation by
A large randomized trial in India people with low socioeconomic sta-
has shown a 50% reduction in cer- Prevention of other cancer tus in the cervical cancer screening
vical cancer mortality after a single types related to lifestyle studies and the high participation
round of HPV screening; in another factors of girls in all socioeconomic groups
trial, a 35% reduction in cervical Given the association between diet, in HPV vaccination programmes
cancer mortality was seen after a overweight, obesity, and physical in Punjab and Sikkim indicate the
single round of screening by visual activity and cancer types such as importance of appropriate educa-
inspection of the cervix with acetic colorectal cancer, ovarian cancer, tional initiatives.
acid [28,29]. An HPV vaccination endometrial cancer, and prostate As the reduction of socioeconom-
study that is under way in India to cancer, among others, and the ic inequalities in population groups in
assess the effectiveness of fewer emerging trends in the prevalence India is addressed, highly focused
than three doses of HPV vaccine of these lifestyle factors accompa- and tailored public health interven-
has demonstrated that two doses nying socioeconomic changes, the tions are needed to target different
of quadrivalent vaccine offer an incidence of these cancer types socioeconomic groups to reduce the
equivalent immune response and is increasing in various regions disparities in cancer prevention.

References
1. UNDP (2018). Human development in- 10. Misra PJ, Mini GK, Thankappan KR 19. Sreedevi A, Quereshi MA, Kurian B,
dices and indicators, 2018 statistical up- (2014). Risk factor profile for non-commu- Kamalamma L (2014). Screening for
date. New York (NY), USA: United Nations nicable diseases among Mishing tribes in breast cancer in a low middle income
Development Programme. Available from: Assam, India: results from a WHO STEPs country: predictors in a rural area of
http://hdr.undp.org/sites/default /f iles/ survey. Indian J Med Res. 140(3):370–8. Kerala, India. Asian Pac J Cancer Prev.
2018_human_development_statistical_ PMID:25366204 15(5):1919–24. https://doi.org/10.7314/
update.pdf.
APJCP.2014.15.5.1919 PMID:24716912
11. Hashibe M, Jacob BJ, Thomas G, Ramadas
2. Dandona L, Dandona R, Kumar GA, K, Mathew B, Sankaranarayanan R, et al. 20. Gadgil A, Sauvaget C, Roy N, Grosse Frie
Shukla DK, Paul VK, Balakrishnan K, et al.; (2003). Socioeconomic status, lifestyle K, Chakraborty A, Lucas E, et al. (2015).
India State-Level Disease Burden Initiative factors and oral premalignant lesions.
Collaborators (2017). Nations within a na- Breast cancer awareness among middle
Oral Oncol. 39(7):664–71. https://doi. class urban women – a community-based
tion: variations in epidemiological transi-
org /10.1016/S13 6 8 - 8 375(03) 0 0 074 - 5 study from Mumbai, India. Asian Pac J
tion across the states of India, 1990-2016
PMID:12907205 Cancer Prev. 16(15):6249–54. https://
in the Global Burden of Disease Study.
Lancet. 390(10111):2437–60. https://doi. 12. Rao SVK (2014). Epidemiology of oral doi.org/10.7314/APJCP.2015.16.15.6249
org /10.1016/S0140 - 673 6(17) 328 0 4 - 0 cancer in India – a life course study [the- PMID:26434824
PMID:29150201 sis]. Adelaide, Australia: The University
21. Sankaranarayanan R, Ramadas K, Thara
of Adelaide. Available from: https://
3. Goss PE, Strasser-Weippl K, Lee- S, Muwonge R, Prabhakar J, Augustine
Bychkovsky BL, Fan L, Li J, Chavarri- digital.library.adelaide.edu.au/dspace/
P, et al. (2011). Clinical breast examina-
Guerra Y, et al. (2014). Challenges to bitstream/2440/91305/3/02whole.pdf.
tion: preliminary results from a cluster
effective cancer control in China, India, 13. Gupta B, Ariyawardana A, Johnson randomized controlled trial in India. J Natl
and Russia. Lancet Oncol. 15(5):489– NW (2013). Oral cancer in India contin- Cancer Inst. 103(19):1476–80. https://doi.
538. https://doi.org/10.1016/S1470-2045 org/10.1093/jnci/djr304 PMID:21862730
ues in epidemic proportions: evidence
(14)70029-4 PMID:24731404
base and policy initiatives. Int Dent J.
63(1):12–25. https://doi.org/10.1111/j.1875- 22. Mittra I, Mishra GA, Singh S, Aranke S,
4. National Cancer Registry Programme
595x.2012.00131.x PMID:23410017 Notani P, Badwe R, et al. (2010). A cluster
(2016). Three-year report of population-
randomized, controlled trial of breast and
based cancer registries: 2012–2014. Ben
14. Sharma S, Satyanarayana L, Asthana cervix cancer screening in Mumbai, India:
galuru, India: Indian Council of Medical
S, Shivalingesh KK, Goutham BS, methodology and interim results after
Research.
Ramachandra S (2018). Oral cancer sta- three rounds of screening. Int J Cancer.
5. Mathur MR, Singh A, Dhillon PK, Dey tistics in India on the basis of first report 126(4):976–84. https://doi.org/10.1002/
S, Sullivan R, Jain KK, et al. (2014). of 29 population-based cancer registries. ijc.24840 PMID:19697326
Strategies for cancer prevention in India – J Oral Maxillofac Pathol. 22(1):18–26.
catching the ‘low hanging fruits’. J Can PMID:29731552 23. Grosse Frie K, Ramadas K, Anju GA,
cer Policy. 2(4):105–6. https://doi.org/10. Mathew BS, Muwonge R, Sauvaget CS,
1016/j.jcpo.2014.07.001 15. Singh A, Ladusingh L (2014). Prevalence et al. (2013). Determinants of participa-
and determinants of tobacco use in tion in a breast cancer screening trial
6. Ferlay J, Ervik M, Lam F, Colombet M, Mery India: evidence from recent Global in Trivandrum district, India. Asian Pac
L, Piñeros M, Znaor A, Soerjomataram I, Adult Tobacco Survey data. PLoS One.
Bray F (2018). Global Cancer Observatory: J Cancer Prev. 14(12):7301–7. https://
9(12):e114073. https://doi.org/10.1371/ doi.org/10.7314/APJCP.2013.14.12.7301
Cancer Today. Lyon, France: International journal.pone.0114073 PMID:25474196
Agency for Research on Cancer. Available PMID:24460292
from: https://gco.iarc.fr/today. 16. Sankaranarayanan R, Ramadas K, Thara
24. Thulaseedharan JV, Malila N, Hakama
S, Muwonge R, Thomas G, Anju G, et al.
7. Gandhi AK, Kumar P, Bhandari M, M, Esmy PO, Cheriyan M, Swaminathan
(2013). Long term effect of visual screen-
Devnani B, Rath GK (2017). Burden of R, et al. (2012). Socio demographic and
ing on oral cancer incidence and mortal-
preventable cancers in India: time to strike reproductive risk factors for cervical can-
ity in a randomized trial in Kerala, India.
the cancer epidemic. J Egypt Natl Canc cer – a large prospective cohort study
Oral Oncol. 49(4):314–21. https://doi.
Inst. 29(1):11–8. https://doi.org/10.1016/j. from rural India. Asian Pac J Cancer Prev.
org/10.1016/j.oraloncology.2012.11.004
jnci.2016.08.002 PMID:27591115 13(6):2991–5. https://doi.org/10.7314/
PMID:23265945
APJCP.2012.13.6.2991 PMID:22938495
8. Shridhar K, Rajaraman P, Koyande S,
Parikh PM, Chaturvedi P, Dhillon PK, et al. 17. Ramadas K, Arrossi S, Thara S, Thomas
G, Jissa V, Fayette JM, et al. (2008). 25. Deodhar K, Gheit T, Vaccarella S, Romao
(2016). Trends in mouth cancer incidence CC, Tenet V, Nene BM, et al. (2012).
in Mumbai, India (1995-2009): an age-pe- Which socio-demographic factors are
associated with participation in oral can- Prevalence of human papillomavirus
riod-cohort analysis. Cancer Epidemiol.
cer screening in the developing world? types in cervical lesions from women
42:66–71. https://doi.org/10.1016/j.canep.
Results from a population-based screen- in rural Western India. J Med Virol.
2016.03.007 PMID:27043865
ing project in India. Cancer Detect Prev. 84(7):1054–60. https://doi.org/10.1002/
9. Bhan N, Karan A, Srivastava S, Selvaraj 32(2):109–15. https://doi.org/10.1016/j. jmv.23310 PMID:22585722
S, Subramanian SV, Millett C (2016). cdp.2008.02.008 PMID:18632218
Have socioeconomic inequalities in 26. Franceschi S, Rajkumar R, Snijders PJ,
tobacco use in India increased over 18. Varadharajan KS, Thomas T, Kurpad Arslan A, Mahé C, Plummer M, et al.
time? Trends from the National Sample AV (2013). Poverty and the state of nu- (2005). Papillomavirus infection in rural
Surveys (2000–2012). Nicotine Tob Res. trition in India. Asia Pac J Clin Nutr. women in southern India. Br J Cancer.
18(8):1711–8. https://doi.org/10.1093/ntr/ 22(3):326–39. https://doi.org/10.6133%2f 92(3):601–6. https://doi.org/10.1038/sj.
ntw092 PMID:27048274 apjcn.2013.22.3.19 PMID:23945402 bjc.6602348 PMID:15668709
264
27. Gupta S, Sodhani P, Sharma A, Sharma 29. Sankaranarayanan R, Esmy PO, Rajkumar 31. Sankaranarayanan R, Joshi S, Muwonge
JK, Halder K, Charchra KL, et al. (2009). R, Muwonge R, Swaminathan R, Shan R, Esmy PO, Basu P, Prabhu P, et al.;
Prevalence of high-risk human papilloma- thakumari S, et al. (2007). Effect of visual Indian HPV vaccine study group (2018).
virus type 16/18 infection among women screening on cervical cancer incidence and Can a single dose of human papillomavi-
with normal cytology: risk factor analysis mortality in Tamil Nadu, India: a cluster- rus (HPV) vaccine prevent cervical can-
and implications for screening and pro- randomised trial. Lancet. 370(9585):398– cer? Early findings from an Indian study.
406. https://doi.org/10.1016/S0140-6736 Vaccine. 36(32 Pt A):4783–91. https://
phylaxis. Cytopathology. 20(4):249–55.
(07)61195-7 PMID:17679017 doi.org/10.1016/j.vaccine.2018.02.087
https://doi.org/10.1111/j.1365-2303.2008.
00611.x PMID:19018810 30. Bhatla N, Nene BM, Joshi S, Esmy PO, PMID:29551226
Poli URR, Joshi G, et al.; Indian HPV vac-
28. Sankaranarayanan R, Nene BM, Shastri cine study group (2018). Are two doses of
SS, Jayant K, Muwonge R, Budukh AM, human papillomavirus vaccine sufficient
et al. (2009). HPV screening for cervi- for girls aged 15-18 years? Results from
cal cancer in rural India. N Engl J Med. a cohort study in India. Papillomavirus
360(14):1385–94. https://doi.org/10.1056/ Res. 5:163–71. https://doi.org/10.1016/j.
NEJMoa0808516 PMID:19339719 pvr.2018.03.008 PMID:29578097
CHAPTER 4.4
SECTION 4

4.5
Variations in implementation of cancer

screening in European countries
Striving for best practice
Harry J. de Koning Partha Basu (reviewer)
Nereo Segnan (reviewer)
rates, provision of informed cal cancer and 17 000 fewer deaths

SUMMARY choice, and fair designs, spe- from breast cancer [8].
cifically with respect to benefits
●● Basic differences are evident and harms, and taking equity
between screening practices fol- Screening programmes
into account.
lowed in European Union coun- Breast cancer, cervical cancer, and
tries, including the target age colorectal cancer are currently the
ranges for screening, the interval only three cancer types for which
Cancer is the second leading cause
between screening tests, and the European Council recommends
of death in Europe [1]. Together,
the screening procedures used. screening [9]. Currently, all EU coun-
colorectal cancer, breast cancer,
tries have some form of screening
●● For breast cancer screening, and cervical cancer are responsi-
for breast cancer and cervical can-
there is a nearly 2-fold difference ble for 20% of cancer mortality and
cer, and most countries have started
in the coverage by invitations for approximately 250 000 deaths in
to implement screening for colorec-
and a more than 5-fold differ- the European Union (EU) per year tal cancer (see Chapter 6.6).
ence in the attendance reported. [2–5]. Each year more than 1 mil- It has been estimated that
lion people in the EU are diagnosed 125 million people in the EU could
●● For cervical cancer screening, with one of these three cancer
both the number of tests that have been screened in 2007 if the
types. The burden of disease is un- screening tests had been available
are offered over a woman’s evenly distributed across countries
lifetime and the actual num- to and utilized by all EU citizens in
in the EU, and it is estimated that the target age ranges. However,
ber of screens received differ
by 2050 the burden will grow by in 2007 approximately 55 million
tremendously, leading to high
up to 50% as a result of population screening tests were actually per-
health inequalities across the
growth and ageing [4–6]. formed in the EU [10]. Therefore,
European Union.
Substantial progress has been successfully improving screening
●● Research shows that achiev- made in the early detection and coverage would potentially have
ing relatively high participation treatment of breast cancer, cervi- an impact on the lives of millions
rates in cancer screening will re- cal cancer, and colorectal cancer; of people, but would also put fur-
duce health inequalities. In pain many countries, mortality has ther pressure on the available clini-
tients with breast cancer, screen decreased by 1–2% per year since cal and economic resources. The
detection is an independent fa- the early 1990s [4,7]. However, 55 million screening tests alone
vourable prognostic factor. great inequity persists in mortal- are estimated to cost more than
ity trends [8]. In addition, there is €500 million per year [11]. In the
●● There appears to be a lack
considerable debate about whether light of the current economic crisis,
of quantified country-specific
this decline in mortality can be at- it is especially important to ensure
knowledge on the expected
tributed to screening or to improve- that this money is well spent and
benefits and harms of the
ments in treatment. Some have that people benefit optimally and
screening policies.
estimated that if all countries in the equally well, if possible.
●● Much effort is needed to ensure EU could reduce mortality rates to In December 2003 the European
the implementation of high- those in the best-performing coun- Council recommended mammog-
quality organized screening pro- try, each year there would be more raphy screening for breast cancer,
grammes with fair attendance than 4000 fewer deaths from cervi- Pap smear (cytology) screening
266
for cervical cancer, and faecal oc- across countries throughout Europe
cult blood test (FOBT) screening (Tables 4.5.1, 4.5.2, and 4.5.3).
FUNDAMENTALS
for colorectal cancer. The latest
revision of the EU code reconfirms Breast cancer
■■ Currently, based on the
the appropriateness of population- There is wide agreement within recommendations of the
based screening programmes for the EU on different aspects of the
European Council, all European
these three cancer types, and not policy for breast cancer screening,
Union countries have some
yet for other cancer types [12]. In such as the screening test based on
most EU countries, organized or mammography, the minimum tar- form of screening for breast
opportunistic screening is available get age range of 50–69 years, and cancer and cervical cancer,
for these cancer types. the screening interval of 2 years and most countries have
The total target population in (Table 4.5.1) [10,13–16]. started to implement screening
the EU is massive: almost 68 mil- However, there are substantial for colorectal cancer.
lion women in the EU are eligible differences within the EU in the
■■ It would not be appropriate to
for breast cancer screening (age extent to which target populations
range, 50–69 years), and more than are actually exposed to screening implement a single, uniform
100 million women can participate [13]. Among the EU countries, there screening programme per
in Pap smear screening (age range, is a nearly 2-fold difference in the cancer type for all countries;
30–59 years). Although the potential coverage by invitations and a more however, in many instances,
target population for colorectal can- than 5-fold difference in the atten- there is no plausible reason for
cer screening is even larger (more dance reported. the huge variations in the three
than 150 million people; age range, cancer screening programmes
50–74 years), approximately 25% of Cervical cancer
across the European Union.
this population had not yet been tar- Cervical cancer screening usually
geted by a screening programme. starts at age 20–30 years and stops ■■ Successfully improving
The number of screening tests that at age 60–70 years. Some coun- screening coverage would po-
are actually performed in the EU is tries recommend starting screening tentially have an impact on the
much lower. In addition, the existing before age 20 years (Table 4.5.2) lives of millions of people, but
screening programmes for breast [10,13,17,18]. would also put further pressure
cancer, cervical cancer, and colorec- For the screening interval, nine on the available clinical and
tal cancer vary in terms of their ap- countries recommend an interval economic resources.
plication, both within countries and of 5 years, and six countries rec-
■■ Organized population-based
CHAPTER 4.5
SECTION 4
screening programmes could
Fig. 4.5.1. A woman undergoing breast cancer screening in Moscow, Russian Federation. be very effective in reducing
health inequalities.
■■ Although nearly all countries
make some degree of
national recommendations for
screening policy, the decision-
making and implementation
are often delegated to lower-
level health authorities.
ommend an interval of 1 year; most

countries recommend a screening
interval of 3 years. As a result, the
number of tests that women in the
EU have over their lifetimes rang-
es from 6 to more than 40. The
proportion of the target popula-
tion covered by the screening test
ranges from 10% to approximately
80%, and for several countries this
proportion is unknown.
Chapter 4.5 • Variations in implementation of cancer screening in European countries 267

Table 4.5.1. Breast cancer screening practices in countries in the European Union
Starting age Stopping age Interval Attendance

Country Primary test
(years) (years) (years) (%) a
Austria 45 69 2 57 Mammography/US
Belgium 50 69 2 33 b
Mammography
Bulgaria 50 69 – ND Mammography
Croatia 50 69 2 45 Mammography
Cyprus 50 69 2 17c Mammography/CBE
Czechia 45 69 d
2 70 Mammography
Denmark 50 69 2 72 Mammography
Estonia 50 64 2 46 Mammography
Finland 50 69 2 76 Mammography
France 50 74 2 53 Mammography/CBE
Germany 50 69 2 53 Mammography
Greece 40 49 2 1 Mammography/CBE
> 50 1 – Mammography/CBE
Hungary 45 64 2 56 Mammography
Ireland 50 69 2 74 Mammography
Italy 50 69 2 ND e
Mammography
Piedmont and Emilia- 45f 49 f 1 ND Mammography
Romagna 50 74 2 ND Mammography
Latvia 50 69 2 34 Mammography
Lithuania 50 69 2 45 Mammography
Luxembourg 50 69 2 60 Mammography
Malta 50 69 3 36 Mammography
Netherlands 50 75 2 80 Mammography
Poland 50 69 2 44 Mammography
Portugal 60 Mammography
Algarve 50 69 2 56 Mammography
Azores 45 74 2 ND Mammography
Other regions 45 69 2 ND Mammography
Romania 50 69 – 0.2g Mammography
Slovakia – – – ND Mammography/US
Slovenia 50 69 2 19 Mammography
Spain 50 h 64h 2 67 Mammography

Some regions 45 69 2 ND Mammography
Sweden 40 74 1.5–2 70 Mammography
United Kingdom 50 70 3 84 i
Mammography
CBE clinical breast examination; ND, no data available; US, ultrasound.

a
The attendance (%) represents the proportion of the target population that has been screened.
b
In Belgium, large regional differences are seen in attendance: Flemish Region, 50%; Brussels, 10%; Wallonia, 8%.
c
In Cyprus, large regional differences are seen in attendance: Nicosia, 42%; other regions, 0%.
d
In Czechia, the invitations are sent only to women up to age 70 years.
e
For Italy, no data about national attendance were found. Regional attendance was: North, 61%; Centre, 56%; South and Islands, 40%.
f
In Italy, the target age range is 45–74 years only in Piedmont and Emilia-Romagna. In other regions, the target age range is 50–69 years.
g
In Romania, large regional differences are seen in attendance: Cluj, 49%; other regions, 0%.
h
In Spain, the standard target age range is 50–64 years, but in some regions the target age range is 45–69 years.
i
In the United Kingdom, regional differences are seen in attendance: England, 86%; Northern Ireland, 80%; Scotland, 73%; Wales, 74%.
268
Table 4.5.2. Cervical cancer screening practices in countries in the European Union
Starting age Stopping age Interval Coverage

Country Triage test
(years) (years) (years) (%) a
Austria ≥ 18 1 ND Cytology
Belgium 25 64 3 37 b
Cytology/HPV
Bulgaria 30 59 3 47
Croatia 25 64 3 105 Cytology/HPV
Cyprus 24 65 3 67 Cytology
Czechia ≥ 15 1 53 Cytology/HPV
Denmark 23 59 3 74 (total) Cytology/HPV

60 65 5
Estonia 30 59 5 77 Cytology/HPV
Finland 30 c
64 c
5 98 Cytology/HPVc
France 25 64 3 8d Cytology/HPV
Germany ≥ 20 1 53 Cytology/HPV
Greece ≥ Age of sexual 1 69 Cytology

onset
Hungary 25 65 3 15 Cytology
Ireland 25 44 3 70 Cytology
45 60 5
Italy 25 64 3 67e Cytology/HPV
Latvia 25 69 3 94 Cytology
Lithuania 25 59 3 78 Cytology/HPV
Luxembourg ≥ 18 1 55 Cytology/HPV
Malta f
25 35 3 49 Cytology/HPV
Netherlands 30 64 5 95 Cytology/HPV
Poland 25 29 3 98 Cytology/HPV
CHAPTER 4.5
SECTION 4
Co-test 30 59 3
Portugal 25 59 3 19 g Cytology/HPV
Azores 25 64 3 ND Cytology/HPV
Lisbon/Madeira – – – – No programme
Romania 25 64 5 65 Cytology
Slovakia 23 24 1 48 (total) Cytology

25 64 3
Slovenia 20 21 1 71 Cytology/HPV
22 64 3
Spain 25 64 3 73 Cytology/HPV
Sweden 23 50 3 81 Cytology/HPV
51 60 5
United Kingdom 25 49 3 101h Cytology/HPV

50 64 5
HPV, human papillomavirus; ND, no data available.

a
The coverage exceeds 100% in some cases. Using a single index year to estimate coverage for screening with intervals of 3–5 years entails some
imprecision because of variability between years, and may lead to estimates exceeding 100%.
b
In Belgium, large regional differences can be seen in attendance: Flemish Region, 65%.
c
In Finland, some municipalities target women younger than 30 years and older than 60 years. The screening test can be either cytology or HPV.
d
In France, an attendance of 89% was found in the 13 departments.
e
In Italy, large regional differences can be seen in attendance: North, 65%; Centre, 83%; South, 60%.
In Malta, the screening programme is being piloted.
f
g
Azores excluded from attendance. In Portugal, large regional differences can be seen in attendance: North, 34%; Centre, 100%; Alentejo, 57%; Algarve, 13%.
h
In the United Kingdom, regional differences can be seen in attendance: England, 104%; Northern Ireland, 91%; Scotland, 93%; Wales, 104%.

Fig. 4.5.2. A biomedical scientist in England making an assessment in relation to variations in the three cancer
cellular characteristics in the context of cervical cancer screening. screening programmes.
These substantial differences
may result in inappropriate inter-
ventions, excessive screening,
and overtreatment, or in delayed
provision of appropriate treatment.
The differences certainly result in
a higher disease burden, a lower
quality of life, health inequities,
and increased costs for health and
care systems. For example, there
are countries where cervical can-
cer screening is performed in a
non-organized manner and where,
even though very large numbers of
tests are performed, no appropriate
benefit has been seen in terms of
reduced incidence of and mortal-
ity from cervical cancer [22,23].
Major modifiable barriers to effec-
tive screening programmes are re-
sponsible for the observed differ-
ences [24], and there appears to be
Cytology is the most commonly ences. For example, in France, the
a lack of quantified country-specific
recommended primary screening target population is invited to gFOBT
knowledge on the expected ben-
test in Europe, with human papillo- screening; in Italy, FIT screening is
efits and harms of the policies.
mavirus (HPV)-based follow-up for used, except in some areas in the
In 2014, an international com-
women with minor cytological abnor- north of the country, where sigmoi-
parison was made of screening
malities (atypical squamous cells of doscopy is offered once in a lifetime
policy-making in Europe and glob-
undetermined significance [ASCUS] at age 58–60 years. The target age
ally [25], with a focus on comparing
and low-grade squamous intraepithe- groups also differ substantially: in
these processes with, for example,
lial lesion [LSIL] cytology). However, some countries, screening is con-
those used in the United Kingdom.
there is no consensus on the use of fined to people aged 60–69 years,
The authors found some impor-
cytology or HPV testing as a triage whereas in others it covers a much
tant differences: (i) Although all of
test for a given cytological diagnosis. larger range of at-risk individuals
the countries considered except
Currently, the Netherlands and some (aged 50–74 years).
Spain made some degree of na-
regions of Italy are the only parts of Attendance rates for screen-
tional recommendations for screen-
Europe where HPV-based screening ing programmes based on FOBT
ing policy, the decision-making
is offered [5,10]. range from 8% to 71% in different
and implementation were often
EU countries. Because colorectal
delegated to lower-level health au-
Colorectal cancer cancer screening is currently still
thorities. (ii) Although in the United
For colorectal cancer, the most widely being implemented in many coun-
Kingdom proposals for new screen-
used FOBT is guaiac FOBT (gFOBT), tries, clear guidance on reducing
ing programmes from stakeholder
which is based on a biochemical test inequities is crucial now.
organizations would generally be
that detects haemoglobin in the stool reviewed, considerations for decid-
(Table 4.5.3) [10,13,19–21]. For a Variation in programmes ing which topics to work on varied
gFOBT, dietary restrictions are re- The underlying risk of cancer varies across the countries to a very large
quired before testing, to reduce the across the EU – and, in the case of extent. (iii) Required measures of
number of false positives. For a fae- colorectal cancer, between the sex- effectiveness varied across coun-
cal immunochemical test (FIT), which es. The countries also vary in terms tries, ranging from high-quality evi-
is based on human haemoglobin an- of capacity and organizational re- dence from randomized controlled
tibodies, a special diet is not required sources. Therefore, it would not trials (in the United Kingdom) to
before testing. be appropriate to implement a sin- Grading of Recommendations As
Assessment of the colorectal gle, uniform screening programme sessment, Development and Evalu
cancer screening strategies cur- per cancer type for all countries. ation (GRADE) working groups (in
rently adopted by the 28 EU coun- However, in many instances, there Sweden) to including international
tries reveals remarkable differ- is no plausible reason for the huge consensus (in France); the United
270
Fig. 4.5.3. Elements of a community-based campaign to encourage colorectal cancer screening, from the Institut Paoli-Calmettes in
Marseille, France.
CHAPTER 4.5
SECTION 4
Kingdom explicitly required con- inequalities. In both study areas, in study in the Netherlands among pa-
sideration of the public pressure the period before the introduction of tients with breast cancer showed that
for widening the inclusion criteria. screening, overall survival was signif- screen detection was a significant in-
(iv) Differences were found in the icantly lower in women with a lower dependent prognostic variable, after
methods for appraising the quality education level than in those with a adjustment for all well-known predic-
of evidence and in the methodolo- higher education level, in both the tive variables, including tumour size,
gies for synthesizing the evidence. younger and older age groups. After lymph node status, and other stage
(v) Differences were found in the the screening programme was fully characteristics [28].
decision-making process itself implemented, the differences in sur- A cross-sectional study in 22
(ranging from voting to decision vival decreased in both age groups European countries using individu-
support systems). and then disappeared completely al-level data from the WHO World
among women in the age group in- Health Survey showed substan-
Health inequalities research vited to screening. These findings tial socioeconomic inequalities in
related to screening suggest that an organized popula- countries with opportunistic screen-
Two studies in Italy showed that the tion-based mammography screening for cervical cancer (comparing
introduction of an organized breast ing programme could be effective highest with lowest education level,
cancer screening programme can in reducing differences in survival relative index of inequality [RII],
have an impact in reducing health in the target population [26,27]. A 1.28; 95% confidence interval [CI],

Table 4.5.3. Colorectal cancer screening practices in countries in the European Union, and in European Council countries outside
of the European Union
Starting age Stopping age

Country Interval (years) Attendance (%) Primary test
(years) (years)
European Union countries
Austriaa 40 80 1 61 gFOBT
> 50 10 2 TC
Burgenland 40 80 1 ND FIT
> 50 10 ND TC
Belgium 28
Wallonia–Brussels 50 74 2 6–7 FIT or gFOBT
50 74 10 ND TC
Flemish Region 56 74 2 47–49 FIT
56 74 10 ND TC
Bulgaria 40 60 1 ND FOBT
Croatia 50 74 2 15 gFOBT
Cyprus 50 69 2 ND FIT
Czechia 50 54 1 21–26 FIT

≥ 55 2 (total FIT) FIT
≥ 55 10 1–2 TC
Denmark 50 74 2 ND FIT
Estonia b
60 69 2 ND FIT
Finland 60 69 2 14–17 gFOBT
France 50 74 2 25–28 gFOBT

Calvados 22–27 FIT
Germany 50 54 1 19 gFOBT/FIT
≥ 55 2 ND gFOBT/FIT
≥ 55 10 3–4 TC
Greece 50 70 2 8 FOBT/gFOBT
50 70 5 ND TC
Hungary 50 70 2 1 FIT
Ireland 60 c 69 c 2 12 FIT
Italy 50 69 2 29 d
FIT
Piedmont 58 60 Once in a lifetime ND FSe
59 69 2 ND FIT
Latvia 50 74 1 11 gFOBT
Lithuania 50 74 2 47–58 FIT
Luxembourg 55 74 2 ND FIT/TC
Malta 55 66 2 45 FIT
Netherlands 55 75 2 27–28 FIT
Poland 55 64 ≥ 10 2 TC
Portugal 50 70 2 1 FIT/gFOBT
Romania – – ND –
Slovakia > 50 ND TC
Slovenia 50 74 2 43–52 FIT
Spain 50 69 2 8–9 FIT
Sweden 60 69 2 11–13 gFOBT
United Kingdom 60 74 2 56 f
gFOBT
England 60 74 2 50–60 gFOBT
ND FS
Scotland 50 74 2 61–65 gFOBT
272
Table 4.5.3. Colorectal cancer screening practices in countries in the European Union, and in European Council countries outside
of the European Union (continued)
Starting age Stopping age

Country Interval (years) Attendance (%) Primary test
(years) (years)
Non-European Union countries
Bosnia and Herzegovina > 50 – ND FOBT
Georgia 50 69 2 53 gFOBT
Iceland 55 75 2 84 FOBT
50 59 – ND TC
Monaco 50 80 2 60 FIT
Montenegro 50 74 – 33 FIT
Norway 55 64 2 ND FIT
– 65 FOBT + FS
Russian Federation
Saint Petersburg 48 75 – ND FIT
Kazan/Tatarstan ND FOBT + DRE
San Marino 50 79 2 65 FIT
Serbia 50 74 2 58 FIT
Switzerland 50 80 2/10 22 FOBT or TC

50 80 – ND FOBT and/or TC
50 69 – ND FIT or TC
Turkey 50 69 – 30 FOBT
Ukraine – – – ND ND
DRE, digital rectal examination; FIT, faecal immunochemical test; FOBT, faecal occult blood test; FS, sigmoidoscopy; gFOBT, guaiac faecal occult blood
test; ND, no data available; TC, colonoscopy.
a
In Austria, a population-based screening programme has been implemented only in the state of Burgenland. In the rest of the country, screening is opportunistic.
b

In Estonia, the population-based pilot programme started in 2016 among a cohort aged 60 years, with an intended target group of age 60–69 years.
c

Ireland is planning to extend the target age range to 55–74 years.

CHAPTER 4.5
SECTION 4
d
In Italy, large regional differences can be seen in attendance: North, 48–52%; Centre, 21–24%; South, 8%; Piedmont (FS + FIT), 17–20%.
e

In Piedmont, Italy, FIT is offered to individuals aged 59–69 years if they are unwilling to undergo FS. For both FIT and FS together, the attendance is
17–20%.
f
In the United Kingdom, regional differences can be seen in gFOBT attendance: England, 50–60%; Northern Ireland, 54%; Scotland, 61–65%; Wales, 52–56%.
1.12–1.48) and for breast cancer and the disappearance of this effect compared with breast cancer and
(RII, 3.11; 95% CI, 1.78–5.42) [29], in screened women. cervical cancer screening, may lead
as well as in countries with regional European data on colorectal to substantial inequalities.
programmes. In countries with or- cancer screening are even more lim- The possible reasons for so-
ganized programmes (limited to ited, but in the first 2.6 million invita- cioeconomic differences in partici-
Denmark, Finland, the Netherlands, tions in England, there was a clear pation in cancer screening are not
Sweden, and the United Kingdom gradient in screening participation well known. In the United Kingdom
for cervical cancer, and those rates across quintiles of deprivation, Flexible Sigmoidoscopy Screening
countries plus Luxembourg for ranging from 35% in the most de- Trial, at the Scottish centre, 6383
breast cancer), such inequalities prived quintile to 61% in the least de- people responded to a question-
were not found for cervical cancer prived quintile (with an average rate naire about psychosocial and cogni-
(RII, 1.13; 95% CI, 0.92–1.40) or for of 54%) [30]. Multivariate analyses tive factors and interest in screening
breast cancer (RII, 1.03; 95% CI, confirmed an independent effect of [31]. The results showed the predict-
0.88–1.20). An early study in the deprivation, with stronger effects in ed gradient in interest with socio-
Netherlands had found the same women, in older people, and in the economic status, but also showed
positive and unfavourable asso- most ethnically diverse areas. It is that the groups with lower socioeco-
ciation in women not screened for possible that the lower participation nomic status felt at high risk of can-
breast cancer or cervical cancer, rates in colorectal cancer screening, cer and were more worried about

cancer. Therefore, the lesser inter- experience the same inequalities, health inequalities. In patients with
est did not derive from complacen- although evidence is sparse. In breast cancer, it has been shown
cy or lack of concern about cancer. southern Italy, attendance rates for that screen detection is an inde-
In contrast, in the groups with high- breast cancer and cervical cancer pendent favourable prognostic fac-
er socioeconomic status, perceived screening were about 40% for im- tor. Therefore, much effort is still
benefits were higher and perceived migrants [34], and in Norway, regis- needed in the EU to ensure the im-
barriers, fear, and fatalism were try data showed that in immigrants, plementation of high-quality orga-
lower. The authors described these rates of non-adherence to the cer- nized screening programmes with
findings as being consistent with vical cancer screening programme fair attendance rates, provision of
evidence that groups with lower so- were 1.7 times those in the autoch- informed choice, and fair designs,
cioeconomic status are less hope- thonous population [35]. specifically with respect to benefits
ful that behaviour change will yield and harms. Equity should be taken
health gains [32] and more fatalistic Reducing health inequalities into account in all the decision-mak-
about the future [33]. Research shows that achieving ing and implementation processes.
It is likely that immigrant sub- relatively high participation rates
groups in many European countries in cancer screening will reduce
References
1. Lozano R, Naghavi M, Foreman K, Lim 7. Bosetti C, Bertuccio P, Malvezzi M, 12. Vale DB, Anttila A, Ponti A, Senore C,
S, Shibuya K, Aboyans V, et al. (2012). Levi F, Chatenoud L, Negri E, et al. Sankaranaryanan R, Ronco G, et al.
Global and regional mortality from 235 (2013). Cancer mortality in Europe, (2019). Invitation strategies and coverage
causes of death for 20 age groups in in the population-based cancer screening
2005-2009, and an overview of trends
1990 and 2010: a systematic analysis for programmes in the European Union. Eur
the Global Burden of Disease Study 2010. since 1980. Ann Oncol. 24(10):2657–71.
https://doi.org/10.1093/annonc/mdt301 J Cancer Prev. 28(2):131–40. https://doi.
Lancet. 380(9859):2095–128. https://doi.
PMID:23921790 org/10.1097/CEJ.0 0 0 0 0 0 0 0 0 0 0 0 0426
org /10.1016/S 0140 - 673 6 (12) 61728 - 0
PMID:29570103
PMID:23245604
8. Mackenbach J, McKee M, editors (2013).
Successes and failures of health pol- 13. Ponti A, Anttila A, Ronco G, Senore C,
2. Ferlay J, Ervik M, Lam F, Colombet M,
Mery L, Piñeros M, et al. (2018). Global icy in Europe: four decades of diverg- Basu P, Segnan N, et al. (2017). Against
Cancer Observatory: Cancer Today. Lyon, ing trends and converging challenges. Cancer. Cancer screening in the European
France: International Agency for Research Buckingham, UK: Open University Press. Union. Report on the implementation of
on Cancer. Available from: https://gco.iarc. Available from: http://www.euro.who.int/ the Council Recommendation on cancer
fr/today. en/publications/abstracts/successes- screening. Brussels, Belgium: European
and-failures-of-health-policy-in-europe.- Commission. Available from: https://
3. Mackenbach JP, editor (2011). Successen ec.europa.eu/health/sites/health/files/
van preventie 1970–2010. Rotterdam, four-decades-of-divergent-trends-and-
converging-challenges-2013. major_chronic_diseases/docs/2017_can
Netherlands: Erasmus Publishing.
c e r s c r e e n i n g _ 2n d r e p o r t i m p l e m e nt a
4. LEBA (2011). Landelijke Evalu atie 9. European Council (2003). Council rec- tion_en.pdf.
Bevolkingsonderzoek Baarmoederhals ommendation of 2 December 2003 on
kanker. LEBA rapportage tot en met cancer screening (2003/878/EC). Off J 14. Altobelli E, Rapacchietta L, Angeletti
2011. Available from: https://www.rivm. Eur Union. L 327/34–38. http://eur-lex. PM, Barbante L, Profeta FV, Fagnano
nl/landelijke-evaluatie-van-bevolkings europa.eu/LexUriServ/LexUriServ.do?uri R (2017). Breast cancer screening pro-
onderzoek-baarmoederhalskanker-leba. =OJ:L:2003:327:0034:0038:EN:PDF grammes across the WHO European
Region: differences among countries
5. Gezondheidsraad (2011). Screening 10. von Karsa L, Anttila A, Ronco G, Ponti A,
op baarmoederhalskanker. The Hague, based on national income level. Int J
Malila N, Arbyn M, et al. (2008). Cancer
Netherlands: Health Council of the Environ Res Public Health. 14(4):452.
screening in the European Union. Report
Netherlands. Available from: https://www. https://doi.org/10.3390/ijerph14040452
on the implementation of the Council
gezondheidsraad.nl/documenten/ PMID:28441745
Recommendation on cancer screen-
a d v i e z e n / 2 011/ 0 5 / 24 /s c r e e n i n g - o p -
ing. First report. Luxembourg: European 15. Simou E, Tsimitselis D, Tsopanlioti
baarmoederhalskanker.
Communities. Available from: https:// M, Anastasakis I, Papatheodorou D,
6. Modig K, Drefahl S, Andersson T, Ahlbom ec.europa.eu/health/ph_determinants/ge Kourlaba G, et al. (2011). Early evalu-
A (2012). The aging population in Sweden: netics/documents/cancer_screening.pdf.
ation of an organised mammography
can declining incidence rates in MI, stroke
and cancer counterbalance the future dem 11. EU-topia (2018). EU-topia: towards im- screening program in Greece 2004-2009.
ographic challenges? Eur J Epidemiol. proved cancer screening in all of Europe. Cancer Epidemiol. 35(4):375–80. https://
27(2):139–45. https://doi.org/10.1007/ Available from: http://eu-topia.org/about- doi.o rg /10.1016 / j.c anep. 2011.02.013
s10654-012-9653-2 PMID:22350145 eu-topia/background/. PMID:21474412
274
16. Ventura L, Giorgi D, Giordano L, Frigerio 23. Arbyn M, Rebolj M, De Kok IM, Fender 29. Palència L, Espelt A, Rodríguez-Sanz M,
A, Mantellini P, Zappa M; Italian breast M, Becker N, O’Reilly M, et al. (2009). Puigpinós R, Pons-Vigués M, Pasarín MI,
cancer screening survey group (2015). The challenges of organising cervi- et al. (2010). Socio-economic inequalities
Mammographic breast cancer screening cal screening programmes in the 15 in breast and cervical cancer screening
in Italy: 2011-2012 survey. Epidemiol Prev. old member states of the European practices in Europe: influence of the type
39(3 Suppl 1):21–9. PMID:26405773 Union. Eur J Cancer. 45(15):2671–8. of screening program. Int J Epidemiol.
17. Elfström KM, Arnheim-Dahlström L, von https://doi.org/10.1016/j.ejca.2009.07.016 39(3):757–65. https://doi.org/10.1093/ije/
Karsa L, Dillner J (2015). Cervical can- PMID:19695867 dyq003 PMID:20176587
cer screening in Europe: quality assur-
24. Priaulx J, de Koning HJ, de Kok IMCM, 30. von Wagner C, Baio G, Raine R, Snowball
ance and organisation of programmes.
Széles G, McKee M (2018). Identifying the J, Morris S, Atkin W, et al. (2011).
Eur J Cancer. 51(8):950–68. https://
d o i . o r g /10 .1016 / j . e j c a . 2 015 . 0 3 . 0 0 8 barriers to effective breast, cervical and Inequalities in participation in an orga-
PMID:25817010 colorectal cancer screening in thirty one nized national colorectal cancer screening
European countries using the Barriers programme: results from the first 2.6 mil-
18. Bruni L, Albero G, Serrano B, Mena M, to Effective Screening Tool (BEST). lion invitations in England. Int J Epidemiol.
Gómez D, Muñoz J, et al. (2019). Human 40(3):712–8. https://doi.org/10.1093/ije/
Health Policy. 122(11):1190–7. https://
papillomavirus and related diseases in dyr008 PMID:21330344
doi.org/10.1016/j.healthpol.2018.08.004
the world: summary report. ICO/IARC
PMID:30177278
Information Centre on HPV and Cancer 31. Wardle J, McCaffery K, Nadel M, Atkin
(HPV Information Centre). Available from: 25. Seedat F, Cooper J, Cameron L, W (2004). Socioeconomic differences in
http://hpvcentre.net /statistics/repor ts/ Stranges S, Kandala NB, Burton H, et cancer screening participation: compar-
XWX.pdf. ing cognitive and psychosocial explana-
al. (2014). International comparisons of
19. Altobelli E, Lattanzi A, Paduano R, screening policy-making: a systematic tions. Soc Sci Med. 59(2):249–61. https://
Varassi G, di Orio F (2014). Colorectal review. Available from: https://assets. doi.org/10.1016/j.socscimed.2003.10.030
cancer prevention in Europe: burden publishing.service.gov.uk /government / PMID:15110417
of disease and status of screening pro- uploads/system/uploads/at tachment _ 32. Clark DO, Patrick DL, Grembowski D,
grams. Prev Med. 62:132–41. https:// d a t a / f i l e / 4 4 4 2 2 7/ F I N A L _ R E P O R T_ Durham ML (1995). Socioeconomic sta-
doi.org/10.1016/j.ypmed. 2014.02.010 International_Screening.pdf.
PMID:24530610 tus and exercise self-efficacy in late life.
26. Puliti D, Miccinesi G, Manneschi G, J Behav Med. 18(4):355–76. https://doi.
20. Navarro M, Nicolas A, Ferrandez A, org/10.1007/BF01857660 PMID:7500327
Buzzoni C, Crocetti E, Paci E, et al.
Lanas A (2017). Colorectal cancer popu-
(2012). Does an organised screen- 33. Wardle J, Steptoe A (2003). Socioeconomic
lation screening programs worldwide in
2016: an update. World J Gastroenterol. ing programme reduce the inequalities differences in attitudes and beliefs about
23(20):3632–42. https://doi.org/10.3748/ in breast cancer survival? Ann Oncol. healthy lifestyles. J Epidemiol Community
wjg.v23.i20.3632 PMID:28611516 23(2):319–23. https://doi.org/10.1093/ann Health. 57(6):440–3. https://doi.org/
onc/mdr121 PMID:21515663 10.1136/jech.57.6.440 PMID:12775791
21. Altobelli E, D’Aloisio F, Angeletti PM
(2016). Colorectal cancer screening in 27. Pacelli B, Carretta E, Spadea T, Caranci 34. Bianco A, Larosa E, Pileggi C, Nobile
countries of European Council outside N, Di Felice E, Stivanello E, et al. CGA, Pavia M (2017). Cervical and
of the EU-28. World J Gastroenterol. (2014). Does breast cancer screen- breast cancer screening participation and
22(20):4946–57. https://doi.org/10.3748/ ing level health inequalities out? A utilisation of maternal health services: a
CHAPTER 4.5
SECTION 4
wjg.v22.i20.4946 PMID:27239121 population-based study in an Italian re- cross-sectional study among immigrant
22. Anttila A, Ronco G; Working Group on the gion. Eur J Public Health. 24(2):280–5. women in Southern Italy. BMJ Open.
Registration and Monitoring of Cervical ht tps://doi.org/10.1093/eurpub/ckt119 7(10):e016306. https://doi.org/10.1136/
Cancer Screening Programmes in the PMID:24008553 bmjopen-2017-016306 PMID:29038177
European Union; within the European
28. Mook S, Van ’t Veer LJ, Rutgers EJ, 35. Leinonen MK, Campbell S, Ursin G, Tropé
Network for Information on Cancer
(EUNICE) (2009). Description of the na- Ravdin PM, van de Velde AO, van A, Nygård M (2017). Barriers to cervical
tional situation of cervical cancer screen- Leeuwen FE, et al. (2011). Independent cancer screening faced by immigrants:
ing in the member states of the European prognostic value of screen detection in in- a registry-based study of 1.4 million
Union. Eur J Cancer. 45(15):2685–708. vasive breast cancer. J Natl Cancer Inst. women in Norway. Eur J Public Health.
https://doi.org/10.1016/j.ejca.2009.07.017 103(7):585–97. https://doi.org/10.1093/ 27(5):873–9. https://doi.org/10.1093/eur
PMID:19744852 jnci/djr043 PMID:21350218 pub/ckx093 PMID:28957477

4.6
Disparities in cancer prevention

services in the USA
A long-standing, persistent cause of inequity
Robert A. Smith Graham A. Colditz (reviewer)
Electra D. Paskett Karen M. Emmons (reviewer)
Carol E. DeSantis
Health disparities are not simply groups), and to geographical differ-

SUMMARY differences between groups, but ences in availability of and access
rather differences that are avoid- to high-quality care in rural versus
●● In the USA, overall cancer mor- able, unfair, unjust, and result from suburban and urban areas, and in
tality has declined among men “systemic and potentially remediable urban areas that have high poverty
and women in all racial and differences in one or more aspects rates. However, these predominant,
ethnic groups, but disparities in of health across socially, demo- more apparent categories do not
cancer mortality persist between graphically, or geographically de- cover the full spectrum of dispari-
non-Hispanic Whites and racial fined populations or population sub- ties, which may also be experienced
and ethnic minority groups for groups” [1]. Broadly defined, health according to age, disability, obesity,
many cancer types. disparities may be evident in any mental health, sexual identity, and
●● Persistent disparities in health, group of people who systematically other characteristics linked to sys-
health services, and health experience social and/or economic tematic discrimination. In 2016,
outcomes are associated with obstacles to health and health care. the United States National Institute
race and ethnicity, sexual and In the USA, social, economic, on Minority Health and Health
gender minority status, lower and geographical inequalities have Disparities announced the formal
education level, lower income, long been associated with persis- designation of sexual and gender
lack of health insurance, lower tent inequity in health outcomes. minorities – an all-encompassing
health literacy, lower access Disparities in cancer outcomes in umbrella term to ensure inclusion
to health services, low-quality the USA are largely attributable of all sexual orientations and gen-
health services, distance from to the lack of a national system of der identities, including those who
health services, rural residence, universal health care, and to an op- may not self-identify as lesbian,
and racial segregation. portunistic model of access to can- gay, bisexual, or transgender – as
cer prevention and early detection, a specific health disparity popula-
●● Low-quality care also may be
which poorly serves both advan- tion for National Institutes of Health
influenced by implicit racial and
taged and disadvantaged groups. research (https://www.edi.nih.gov/
class bias, which reflects auto-
This health-care model results in sites/default/files/EDI_Public_files/
matic and unconscious negative
unequal access to health care, be- sgm-strategic-plan.pdf).
attitudes towards low-income and
cause of differences in health insur- Morris et al. [2] conceptualized
minority groups and has been
ance coverage, quality of care, and that cancer outcomes could be best
shown to negatively influence
health literacy (i.e. a person’s ability understood as a function of three un-
patient communication, clinical
to obtain, process, and understand derlying mechanistic domains: pa-
care, and cancer outcomes.
basic health education), as well as tient factors, utilization of care, and
●● Disparities in access to cancer the lack of a usual source of care provider factors (Fig. 4.6.1). Patient
prevention and early detection and barriers to accessing care factors also include behaviours that
and in cancer incidence and when it is needed. increase risk of cancer or comorbid
mortality can be reduced by a These disparities are predomi- conditions, each of which may dif-
combination of national policies nantly linked to race and ethnicity, ferentially have its roots in social
and local initiatives that remove to socioeconomic status (which ac- inequality, and each of which may
barriers to care. counts for most of the inequality in also contribute to inequity in out-
outcomes between racial and ethnic comes. Low-quality care, regardless
276
Fig. 4.6.1. A conceptual model of mechanisms underlying disparities in cancer care

and outcomes. SES, socioeconomic status. FUNDAMENTALS
■■ At the core of cancer disparities
in the USA is an opportunistic
Sociodemographic Health system model of access to cancer
context context
o Neighbourhood Hospital factors prevention and early detection
o Race versus a national system of
o Income
o SES universal health care, which is
Patient factors Provider factors
o Tumour biology o Knowledge/beliefs
deeply rooted in societal beliefs
o Comorbid disease o Technical skill about access to health care as a
o Beliefs/preferences o Resources
basic human right.
■■ Enduring disparities in cancer
incidence and mortality are
Utilization of care Quality of care attributable mostly to persistent,
systemic racial, ethnic, and class
bias, geographical location, and
inequality in education, income,
Appropriate cancer care
geography, and access to high-
o Screening quality health services.
o Surgery
o Adjuvant therapy ■■ Factors associated with
o Surveillance
disparities in access to high-
quality care and in cancer
outcomes are interrelated and
Outcomes interdependent, have historical
o Overall survival and systemic antecedents, and
o Cancer-specific survival
reflect a combination of patient
factors, provider factors, the
availability and quality of health
of health insurance coverage, also rates varied considerably by race care, and for some cancer types,
may be influenced by structural in- and ethnicity; the adjusted relative differences in tumour biology.
equality and by implicit racial and risk of cancer death was 33% high-
■■ Increasing access to health
er in non-Hispanic Blacks and 51%
CHAPTER 4.6
SECTION 4
class bias, which reflects automatic insurance has been shown to
and unconscious negative attitudes higher in non-Hispanic American
be an effective, low-intensity
towards low-income and minority Indians/Alaska Natives than in non- intervention to reduce cancer
groups and has been shown to neg- Hispanic Whites [4]. disparities in disadvantaged
atively influence patient communica- This chapter focuses on both groups; however, interventions
tion and clinical care [3]. the descriptive epidemiology of that only improve insurance
From 2009 to 2013, the trends cancer disparities in the USA and coverage without ensuring direct
in overall cancer incidence in the the structural and systemic factors pathways to high-quality care will
USA for all cancers combined in that contribute to their persistence. not reduce disparities.
men and women in each racial and ■■ Patient navigation has been
ethnic group were similar in direc- Racial and ethnic shown to improve disease out-
tion to those in the overall popula- disparities comes by overcoming institu-
tion [4]. Also, from 2010 to 2014, The United States Census Bureau tional barriers attributable to the
overall cancer death rates declined defines race as an individual’s difficulty of manoeuvring through
in men and women in all racial and self-identification as Asian, Black, complex and often unresponsive
ethnic groups [4]. These trends Native Hawaiian or another Pacific health-care institutions.
were attributed mostly to reductions Islander, American Indian, Alaska ■■ Recent evidence has shown that
in tobacco use, the contribution of Native, and White. Hispanic origin the most effective interventions
screening to early detection of inva- is considered an ethnicity, and a to reduce disparities in cancer
sive cancer and precursor lesions, person of any race may also iden- outcomes occur when key institu-
and improvements in therapy. tify themselves as Hispanic or tions and leaders in local settings
However, Black men and women Latino. Health disparities research commit to the implementation of
still had the highest cancer mortali- consistently shows racial inequali- multicomponent interventions that
ty rates among all racial and ethnic ties across most health outcomes. target specific barriers to care.
groups, and 5-year relative survival Socioeconomic status contributes
Chapter 4.6 • Disparities in cancer prevention services in the USA 277

to racial inequalities, but generally enduring inequities that continue cer screening rates, because of recall
residual disparities by race and eth- to be enabled by health systems, bias and social desirability, which has
nicity remain after adjustment for their administrations, and health- been shown to be highest in Blacks
socioeconomic status [5]. care professionals. This direct and and lowest in Hispanics [8].
In 2003, the Institute of Medicine indirect discrimination also leads
published a landmark report on ra- to patient-level attributes that fur-
cial and ethnic disparities in health ther contribute to disparities, such
Socioeconomic
care in the USA [6]. The report’s as refusing recommended services disparities
conclusions were direct and un- because of mistrust, prior adverse
Income
hesitant. In the USA, racial and eth- experiences, and so on [6].
nic minorities receive less and low- Racial and ethnic disparities in re- In 2017, the United States federal
er-quality health care, for reasons cent cancer screening are shown in government’s poverty level was an
that go beyond lower socioeco- Table 4.6.1. In general, reported can- annual income of US$ 12 140 for a
nomic status and being uninsured cer screening rates are similar be- single individual or US$ 25 100 for
or underinsured. These disparities tween Blacks and Whites but lower in a family of four. In the USA, recent
are attributable to structural racism, Hispanics and Asians [7]. However, cancer screening is strongly asso-
which has its roots in historical and these data overestimate recent can- ciated with a usual source of care,
Table 4.6.1. Prevalence (%) of recent cancer screening examinations among adults in the USA by race and ethnicity, health
insurance coverage, and education level, from the 2015 National Health Interview Survey
Screening Race and ethnicitya Health insurance b Education level

examination
Some High Some
high school college/ College
White Black Hispanic Asian Yes No
school or diploma associate graduate
less or GED degree
% SE % SE % SE % SE % SE % SE % SE % SE % SE % SE
Colorectal cancer (adults aged ≥ 50 years)
Endoscopyc 63.3 0.7 59.3 1.4 47.6 1.5 44.8 2.6 56.8 0.9 24.0 2.2 45.3 1.4 56.4 1.0 61.6 0.9 68.9 1.0
Stool-based
6.9 0.3 8.0 0.9 7.3 0.8 9.2 1.4 6.2 0.4 4.0 1.1 6.3 0.7 7.1 0.6 7.2 0.6 7.7 0.5
testd
Stool-based
test or 65.4 0.7 61.8 1.4 49.9 1.5 49.4 2.7 59.6 0.9 25.1 2.2 47.4 1.4 58.6 1.0 64.3 0.9 71.3 1.0
endoscopye
Breast cancer (women aged ≥ 40 years)
Mammogram
within the
50.3 0.8 55.4 1.8 45.7 1.5 47.1 2.6 52.5 0.9 20.9 2.3 38.9 1.8 45.0 1.5 51.2 1.3 57.9 1.1
preceding
year
Mammogram
within the
64.8 0.8 68.8 1.6 60.8 1.6 59.4 2.5 67.8 0.8 30.7 2.4 50.8 1.9 58.0 1.4 65.9 1.2 73.2 1.0
preceding
2 years
Cervical cancer (women aged 21–64 years)
Pap test f 83.3 0.7 84.8 1.1 77.5 1.2 73.3 2.0 84.4 0.5 60.8 1.7 70.1 1.8 75.4 1.4 84.0 0.9 88.8 0.6
GED, General Educational Development test; SE, standard error.

a
Estimates for Whites, Blacks, and Asians are among non-Hispanics.
b
Health insurance status was analysed among adults aged ≤ 64 years.
c
Endoscopy included sigmoidoscopy within the preceding 5 years or colonoscopy within the preceding 10 years.
d
Stool-based tests included faecal occult blood test (FOBT) or faecal immunochemical test (FIT) using a home test kit performed within the preceding year.
The 2015 data include FIT; data for prior years do not.
e
Stool-based test within the preceding year or sigmoidoscopy within the preceding 5 years or colonoscopy within the preceding 10 years.
f
Women with intact uteri who had a Pap test within the preceding 3 years. Estimates by education level are among women aged 25–64 years.
278
a personal health-care provider, Low educational attainment, low units of analysis, and developing
a recommendation from a health- health literacy, and limited English geospatial epidemiological meth-
care professional, and a recent proficiency have been shown to ods to explore the interplay be-
health maintenance visit, each of be negatively correlated with rates tween population characteristics,
which is strongly associated with of recent cancer screening [13]. health resources, social and envi-
having health insurance. Low- Similar to the associations between ronmental barriers, and the influ-
income groups have higher rates of income and recent cancer screen- ence of spatial patterning on social
being uninsured. ing, there is a significant linear re- inequality and disparities [15].
Access to health insurance has lationship between educational at- Modern approaches to medical
improved as a result of the Patient tainment and being adherent with geography recognize that there are
Protection and Affordable Care all cancer screening recommenda- independent and interdependent fac-
Act of 2010 [9,10], which expanded tions (Table 4.6.1) [7]. tors associated with context (place)
eligibility for Medicaid coverage and composition (people) that con-
to those with incomes at or below Health insurance coverage tribute to health disparities [16]. For
138% of the federal poverty level Some of the largest gaps that are example, a review of research on
and provided tax subsidies to low- observed in cancer prevention, ear- the association between segregation
income populations with incomes ly detection, and cancer outcomes and Black–White cancer disparities
too high to qualify for Medicaid. are those between insured and showed a common association be-
However, in 2018 25% of those with uninsured populations. Preliminary tween racial segregation and high-
incomes of 100% to less than 200% data from the 2018 National Health er rates of late-stage diagnosis of
of the poverty level still reported Interview Survey showed that breast cancer and lung cancer after
lacking health insurance [10]. among adults aged 18–64 years, adjustment for socioeconomic status
Lower socioeconomic status is 12.5% had no health insurance, and health insurance coverage [17].
associated with lower rates of cancer 20.0% had public insurance (includ-
screening. Compared with people ing Medicaid), and 69.2% had pri-
Rural–urban disparities
who have incomes above 400% of vate insurance [10]. In the USA, about 46 million people
the federal poverty level, women with Under the Patient Protection and (~14% of the population) live in rural
incomes of less than 139% of the Affordable Care Act, individuals with areas. According to the Pew Research
federal poverty level are less likely private insurance may receive pre- Center (https://www.pewsocialtrends.
to have had a recent mammogram ventive services recommended by org/2018/05/22/what-unites-and-
(58.7% vs 78.8%) or Pap test (75.2% the United States Preventive Services divides-urban-suburban-and-rural-
vs 89.7%), and among both men and Task Force at no cost to the patient, communities/), rural counties are pre-
women, those with incomes of less and this also applies to public insur- dominantly White (79%); compared
CHAPTER 4.6
SECTION 4
than 139% of the federal poverty ance in the 37 states that expanded with cities, rural areas have a higher
level are less likely to have recently access to Medicaid to low-income in- proportion of adults with a high school
been screened for colorectal cancer dividuals. The expansion of Medicaid education or less (51% vs 38%) and
(46.9% vs 70.0%) [11]. eligibility has been associated with a substantially higher proportion of
higher rates of screening for cervical counties in which the poverty rate ex-
Education level ceeds 20% (31% vs 19%), and nearly
cancer and colorectal cancer for low-
In the USA, data on individual and income adults [14]. Adults with health twice as many rural residents (63% vs
family incomes are difficult to obtain insurance report significantly higher 36%) report that access to health care
in research studies on health-care rates of cancer screening compared is a problem. Compared with people
utilization. Given the strong corre- with adults who report that they are who live in metropolitan areas, rural
lation between educational attain- uninsured (Table 4.6.1) [7]. However, residents have higher rates of being
ment, unemployment, occupation, health insurance coverage alone uninsured, have higher rates of smok-
and income, education level has does not guarantee access to high- ing, obesity, and physical inactivity,
been used as a surrogate mea- quality care. and have lower rates of human papil-
sure for an individual’s socioeco- lomavirus (HPV) vaccination and can-
nomic status. Education level also cer screening (Fig. 4.6.2) [18].
is strongly associated with health Geographical disparities
literacy [12]. An assessment of the Geographical disparities in cancer Disparities by state and region
health literacy of adults in the USA outcomes have been documented States and regions of the USA
found that 49% of adults who did since the mid-20th century. More vary in the proportions of men and
not complete high school had a recently, greater attention has women who have incomes below
below basic level of health literacy, been focused on improving the the poverty level, have health in-
compared with 15% of adults with measurement of health disparities surance, have convenient access
a high school diploma and 3% of by examining data from smaller, to health services, have been vac-
adults with a bachelor’s degree [12]. more homogeneous geographical cinated against HPV infection, and

Fig. 4.6.2. Healthy People 2020 objectives related to cancer, including social determinants of health, health risk behaviours,
recommended vaccination against cancer-causing viruses, and recommended cancer screening, by rural versus urban residence.
have access to cancer screening tion of tobacco control policies, such ability in the decline in the breast
and to specialty care if they are diag- as Tobacco 21 (banning the sale of cancer mortality rates in states. In
nosed with cancer [19]. States also tobacco products to people younger the USA, from 1988–1990 to 2013–
vary in the prevalence of obesity and than 21 years) and excise taxes [20]. 2015, the breast cancer mortality
physical activity, in the proportion of Taken together, these factors rate declined by 39% overall, but by
adults who use tobacco and who contribute to considerable variation only 20–29% in 10 states (Fig. 4.6.3)
have access to cessation treatment in cancer incidence and mortality [21]. Similar variability is evident for
coverage, and in spending on to- rates across states and in trends colorectal cancer mortality: from
bacco control and the implementa- over time, as is evident in the vari- 1980–1982 to 2013–2015, the rate
280
declined by 49% overall, but by only the Centers for Disease Control to National Breast and Cervical
12–31% in eight states, of which six collaborate on a fact-finding mis- Cancer Early Detection
also had the smallest reductions in sion in which Dr Harold Freeman of Program
breast cancer mortality [21]. Harlem Hospital Center convened In 1990, the United States Congress
Siegel et al. [22] examined colo- seven fact-finding hearings across passed the Breast and Cervical
rectal cancer mortality rates in the the USA to gather testimony from Cancer Mortality Prevention Act,
USA to assess trends over time low-income people affected by can- which directed the Centers for
from 1970 to 2011 and to identify cer and from clinicians who served Disease Control and Prevention to es-
clusters of significantly higher mor- low-income populations [23,24]. tablish a programme to provide breast
tality rates, designated as hotspots. In its 1989 Report to the Nation, cancer and cervical cancer screen-
The regions with the highest colo- the American Cancer Society de- ing services to low-income women
rectal cancer mortality rates shifted scribed the disproportionate pain, in all states, the District of Columbia,
over the 40-year period from 1970 to suffering, institutional indifference, United States territories, and tribes
2009 (Fig. 4.6.4). Before 1990, the and obstacles faced by low-income or tribal organizations (https://www.
rates were high in the mid-central cancer patients and their families cdc.gov/cancer/nbccedp/index.htm).
and north-eastern parts of the USA and issued 10 broad recommenda- Uninsured and underinsured women
and low in the south of the country. tions to reduce inequities in cancer who have incomes at or below 250%
By 2000–2009, there was a more prevention, early detection, and of the federal poverty level and who
homogeneous pattern of similar meet the recommended age require-
treatment, and to reform health-
rates across most of the country, ments (~1 in 10 women) are eligible
care services [23].
with the exception of three distinct for the programme. However, the
There are now annual reports
hotspots: the Lower Mississippi federal government only appropri-
on cancer disparities, and in the de-
Delta, west central Appalachia, and ates enough funding to cover services
cades since 1989, there have been
eastern Virginia/North Carolina. In for a small fraction of eligible women
investments in research, imple-
these three hotspots, the mortality (6.5% for Pap testing and 10.5% for
mentation of interventions such as
rates in 2009–2011 were respec- mammography) [25]. Since 1991, the
patient navigation (Fig. 4.6.5; see
tively 40%, 18%, and 9% higher National Breast and Cervical Cancer
also “Patient navigation”), special
than those in non-hotspot counties. Early Detection Program has served
programmes to increase access
more than 5.4 million women [25]. A
to screening, and policy changes,
Interventions to reduce similar programme exists to increase
such as legislation to increase ac- colorectal cancer screening rates (the
disparities cess to health insurance. Although Colorectal Cancer Control Program;
By the late 1980s, the accumula- these interventions have been ben- https://www.cdc.gov/cancer/crccp/
CHAPTER 4.6
SECTION 4
tion of evidence of broad dispari- eficial, they are unable to overcome index.htm), but it covers even fewer
ties in cancer care and outcomes the core underpinnings of systemic eligible people.
led the American Cancer Society, inequality and the lack of universal
the National Cancer Institute, and access to health care in the USA. Patient Protection and
Affordable Care Act of 2010
The Patient Protection and Affordable
Fig. 4.6.3. Decline in breast cancer mortality rates from 1988–1990 to 2013–2015, Care Act of 2010 has improved the
by state. quality of health insurance, eliminat-
ed patient costs for recommended
preventive services, and increased
the availability of affordable health
care to millions of Americans [9].
The insurance coverage provisions
went into effect in 2014. The origi-
nal legislation intended that states
would expand Medicaid eligibility
to all individuals with incomes at or
below 138% of the federal poverty
level. However, in 2012, the United
States Supreme Court ruled that
states could reject Medicaid expan-
sion, and as of 2018, 17 states have
not expanded their public insurance
programmes, leaving 4.2 million
non-elderly adults uninsured.

Fig. 4.6.4. Hotspot analysis of county-level colorectal cancer mortality rates during the four decades from 1970 to 2009. Three hotspots
are indicated for 2000–2009: (1) the Lower Mississippi Delta, (2) west central Appalachia, and (3) eastern Virginia/North Carolina.
1970-1979 1980- 1989
Significant clusters of Significant clusters of

low mortality rates high mortality rates
1990-1999 2000-2009
2
3
Although insurance coverage in- groups for which disparities have cancer mortality rate in Blacks, re-
creased substantially, the short pe- persisted are just another inequity sulting in an annual mortality rate in
riod since the beginning of coverage added to the others. A growing Blacks that was nearly the same as
in 2014 and the lags in data avail- emphasis on genetics and person- that in Whites [27]. In San Francisco,
ability limit the ability to measure alized medicine overshadows the California (www.sfcancer.org), and
the impact of new coverage on use persistent failure to ensure that in Chicago, Illinois (www.chicago
of cancer preventive services and disadvantaged groups have equal breastcancer.org; see also “The
outcomes. However, a review of 14 access to long-standing, conven- enduring disparity in breast cancer
studies concluded that the Patient tional, evidence-based cancer pre- mortality between Black and White
Protection and Affordable Care Act vention, early detection, and state- women in the USA”), there is a city-
had improved access to cancer of-the-art treatment services. wide commitment to reduce cancer
screening, and especially colorec- National policies can contribute disparities by engaging local health
tal cancer screening, among adults to reducing disparities, but there is systems, local government, com-
who had faced the highest cost bar- an increasing realization that major, munity leaders, and the population.
riers before its passage [26]. enduring change can occur only The knowledge needed to eliminate
when community leaders commit cancer disparities exists; what must
to removing barriers in access to also exist is the national and local
Conclusions high-quality care. A statewide effort commitment to do so.
Despite progress in cancer control in Delaware to eliminate disparities
in most population demographics in colorectal cancer outcomes led
in the USA, smaller gains in the to a 42% reduction in the colorectal
282
Fig. 4.6.5. The patient navigation (PN) model of the New Hampshire Colorectal Cancer Screening Program (NHCRCSP), showing
inputs, activities, and outcomes. CRC, colorectal cancer; RN, registered nurse; USMSTF, United States Multi-Society Task Force
on Colorectal Cancer; USPSTF, United States Preventive Services Task Force.
Intermediate
Inputs Activities Short-term Outcomes Long-term Outcomes
Outcomes
• Program • Deliver Six Topic Navigation • Reduced missed • Improve coordination • Decrease colorectal
infrastructure & Protocol appointments and continuity of cancer mortality
resources care for primary care
− Engagement, CRC Screening • Reduced late providers and patients • Decrease colorectal
• Trained RN Navigators Education, and Barrier cancellations of cancer incidence
Assessment appointments • Increase clinic-level
• Contracts with health screening rates • Improve state’s
systems and other − Prep Education and Barrier • Improved quality of colorectal cancer
partners Resolution bowel prep • Enhance access to screening rates
− Prep Review and Re-addressing screening and other
• Eligible patients Barriers • Improved completion clinic services • Increase early-stage
enrolled in NHCRCSP of colonoscopy detection
− Assessment of Prep and • Provide complete
Confirmation of Test Day Details • Improved receipt of and timely diagnostic • Reduce colorectal
− Day of Colonoscopy colonoscopy results follow up cancer-related health
by patients disparities
− Follow-up and Patient • Create timely access
Understanding of Results • Improved receipt of to medical treatment
colonoscopy results for persons diagnosed
• Facilitate needed services by primary care with CRC
providers
• Document PN services delivered
• Increase adherence to
• Improved accuracy recall and surveillance
• Track patients
of rescreening/ intervals
• Verify receipt of colonoscopy surveillance intervals
results by patients and primary
care providers
• Assess concordance of rescreening

interval recommended by
endoscopist with USPSTF/USMSTF
guidelines
Patient navigation
CHAPTER 4.6
SECTION 4
The first patient navigation pro- Patient navigation has been cancer types included in navigation
grammes in the USA were devel- shown to overcome common barri- programmes. Therefore, a range of
oped by Dr Harold Freeman and ers attributable to poverty, low edu- remaining and new questions are
established at Harlem Hospital cation level and health literacy, lack being addressed.
Center in New York City to reduce of English fluency, poor clinical com- • Which patients need naviga-
disparities in breast cancer care munication, lack of knowledge and tion services? At the National
for low-income Black and Hispanic confidence required to manoeuvre Academies of Sciences, Engi
women [1]. Patient navigation was in a complex health system, lack of neering, and Medicine work-
initially designed to ensure timely insurance and need to access finan- shop on Establishing Effective
follow-up of abnormal screening cial aid, and lack of transportation Patient Navigation Programs in
findings and eliminate delays in di- [2]. A skilled navigator can recog- Oncology [2], there was agree-
agnosis and initiation of treatment. nize and address barriers that may ment that all patients would
The substantial investment in re- exist at the system level, with the probably benefit from some
search funding to further develop clinician, or with the patients them- degree of navigation; how-
this concept has extended naviga- selves, and thus prevent delays in ever, because of the limited
tion programmes to improve rates the receipt of care. resources available to support
of cancer screening; to ensure Although the benefits of patient navigation, it was suggested
timely progress through follow-up navigation are well documented, that programmes should target
of abnormal screening findings, di- there are still some areas where those patients at greatest risk
agnostic evaluation, and initiation the benefit of navigation has yet to for delays in care, and expand
of treatment; and to build trust be- be determined, such as accrual to to cancer types that are not so
tween patients and families and the clinical trials, cost–effectiveness, commonly studied, for example
health-care system. and the expansion of the range of types other than breast cancer.

• What background is needed developed a toolkit to support the References

to be a navigator? Experience efforts of navigation programmes 1. Vargas RB, Ryan GW, Jackson CA,
has shown that the answer to to make the financial case for in- Rodriguez R, Freeman HP (2008).
this question lies in the princi- stitutional support for navigation Characteristics of the original patient
navigation programs to reduce dispari-
pal needs of the patients be- services [3].
ties in the diagnosis and treatment of
ing served. Navigators include The American Cancer Society breast cancer. Cancer. 113(2):426–33.
nurses, social workers, and non- supports the National Navigation h t t p s: //d o i .o r g /10 .10 0 2 /c n c r. 2 3 5 47
clinical community workers with Roundtable (https://navigation PMID:18470906
the same racial or ethnic and roundtable.org), a coalition of 2. National Academies of Sciences,
religious backgrounds as the leading oncology, public health, Engineering, and Medicine (2018).
Establishing effective patient naviga-
populations they serve. social work, and advocacy or-
tion programs in oncology: proceedings
• How can support for patient navi- ganizations to address evidence- of a workshop. Washington (DC), USA:
gation programmes be acquired? based practices, training and National Academies Press. https://doi.
Currently, patient navigation is not certification criteria, and policy is- org/10.17226/25073
covered by health insurance, so sues to enhance and promote the 3. National Colorectal Cancer Roundtable
patient navigation programmes effectiveness of patient naviga- (2019). Paying for colorectal cancer
screening patient navigation toolkit:
commonly depend on grants, tion programmes across all areas
strategies for payment and sustain-
institutional resources, and vol- of the cancer control continuum ability. Available from: http://nccrt.org/
unteer efforts. The National Co and in all populations at risk for or intervention/navigation.
lorectal Cancer Roundtable has diagnosed with cancer.
Fig. B4.6.1. Patient navigator model.
Patient navigator roles

Provide disease-specific health education
Facilitate shared decision-making
Provide informal emotional support and
refer for formal psychosocial support
Educate patient about health-care system processes
Promote
self-
efficacy
Reinforced over
time by the patient
Patient
Sustain
Enhance
engage-
access to
ment with
care
care
Patient navigator roles Patient navigator roles
Assist with paperwork, insurance approval, Coordinate timely access to recommended

and financial counselling testing/procedures
Identify appropriate care settings based Arrange referrals to specialists
on evolving patient needs or ancillary care
Help arrange appointment reminders, Facilitate communication
transportation, childcare among multiple providers
284
The enduring disparity in breast cancer mortality between Black and White
women in the USA
In the USA, there has persistently likely to receive therapy that adheres been associated with higher rates of
been a significantly higher breast to practice guidelines, and are more obesity, diabetes, and hypertension
cancer mortality rate in Black likely to discontinue hormone ther- in Black women, although variation
women than in White women [1]. apy early. The higher rate of being between states is attributable mainly
Past efforts to understand this dis- uninsured and underinsured is as- to differences in access to high-
parity focused on differences in sociated with these health services quality health care. Differences in
socioeconomic status or inherent disparities, as is well-documented these disparities across states prob-
differences in tumour biology; to- poor communication with health- ably account for much of the range
day, the disparity in breast cancer care providers, especially among in breast cancer mortality rate ratios
mortality is better understood as African immigrants. Differences in between Black women and White
complex and multifactorial. Daly breast cancer mortality have also women (Fig. B4.6.2) [1].
and Olopade [2] described racial
disparities in cancer mortality as a
“perfect storm” (in which a combi- Fig. B4.6.2. Mortality rate ratios comparing breast cancer mortality rates in Black
women versus White women in the USA, by state, in 2012–2016. Lighter shaded
nation of circumstances aggravates
bars indicate that mortality rates in Black women and in White women were not
the situation) resulting from the col- statistically different.
lision of tumour biology, genomics,
and health-care delivery patterns.
Differences in tumour biol-
ogy are well documented, includ-
ing higher percentages of hormone
receptor-negative tumours in Black
women, intratumour genetic hetero-
geneity, and a higher rate of triple-
negative disease in Black women
(approximately double the rate in
White women).
Health services research in vari-
CHAPTER 4.6
SECTION 4
ous communities in the USA has re-
vealed disparities in standards of
breast cancer-related care. Among
younger women diagnosed with
breast cancer, Black women are less
likely to report a discussion about
BRCA testing and less likely to un-
dergo BRCA testing compared with
White women, and among carriers
of BRCA mutations, Black women
are significantly less likely to under-
go risk-reducing surgery compared
with White women. Black women
are less likely to have undergone re-
cent mammography screening, are
less likely to have access to high-
quality mammography screening,
and are more likely to experience
a longer duration from abnormal
mammography results to diagnosis,
and from diagnosis to treatment.
Compared with White women, Black
women are more likely to be under-
Black:White mortality rate ratio
treated for breast cancer, are less

The racial disparity in breast 1999–2013, Chicago was the only 2. Daly B, Olopade OI (2015). A perfect
storm: how tumor biology, genomics, and
cancer mortality rates in the USA United States city among 10 stud- health care delivery patterns collide to
will only be overcome through local, ied in which the breast cancer create a racial survival disparity in breast
multilevel interventions, such as mortality rate in Black women de- cancer and proposed interventions for
change. CA Cancer J Clin. 65(3):221–
those initiated by the Metropolitan creased more (by 13.9%) than the 38. https://doi.org/10.3322/caac.21271
Chicago Breast Cancer Task Force, rate in White women (which de- PMID:25960198
which established a partnership creased by 7.7%) [3]. 3. Sighoko D, Murphy AM, Irizarry B,
between community organizations, Rauscher G, Ferrans C, Ansell D (2017).
medical providers, and government References Changes in the racial disparity in breast
cancer mortality in the ten US cities with
leaders to improve the quality of 1. DeSantis CE, Ma J, Goding Sauer A, the largest African American populations
mammography and follow-up of Newman LA, Jemal A (2017). Breast from 1999 to 2013: the reduction in breast
cancer statistics, 2017, racial disparity cancer mortality disparity in Chicago.
abnormal findings in Black women
in mortality by state. CA Cancer J Clin. Cancer Causes Control. 28(6):563–8.
living in low-income, segregated 67(6):439–48. https://doi.org/10.3322/ ht tps: //doi.org /10.10 07/s10 5 52- 017-
neighbourhoods [3]. For the period caac.21412 PMID:28972651 0878-y PMID:28275936
286
References
1. Whitehead M (1992). The concepts and 10. Martinez ME, Zammitti EP, Cohen RA 19. American Cancer Society (2017). Colo
principles of equity and health. Int J (2018). Health insurance coverage: early rectal cancer facts & figures, 2017–2019.
Health Serv. 22(3):429–45. https://doi. release of estimates from the National Atlanta (GA), USA: American Cancer
org/10.219 0/986L- LHQ6 -2V TE-YRRN Health Interview Survey, January– Society. Available from: https://www.can
PMID:1644507 June 2018. National Health Interview cer.org/research/cancer-facts-statistics/
Survey Early Release Program. Atlanta colorectal-cancer-facts-figures.html.
2. Morris AM, Rhoads KF, Stain SC,
(GA), USA: National Center for Health
Birkmeyer JD (2010). Understanding ra- 20. American Cancer Society (2017). Cancer
Statistics. Available from: https://www.
cial disparities in cancer treatment and prevention & early detection facts and
c dc.gov/nchs/data /nhis/ear lyrelease/
outcomes. J Am Coll Surg. 211(1):105– figures, 2017–2018. Atlanta (GA), USA:
insur201811.pdf. American Cancer Society. Available from:
13. https://doi.org/10.1016/j.jamcollsurg.
2010.02.051 PMID:20610256 11. White A, Thompson TD, White MC, https://www.cancer.org/research/cancer-
Sabatino SA, de Moor J, Doria-Rose facts-statistics/cancer-prevention-early-
3. Paradies Y, Truong M, Priest N (2014). A detection.html.
PV, et al. (2017). Cancer screening test
systematic review of the extent and mea-
use – United States, 2015. MMWR Morb 21. Siegel RL, Jemal A, Wender RC, Gansler
surement of healthcare provider racism.
Mortal Wkly Rep. 66(8):201–6. https:// T, Ma J, Brawley OW (2018). An assess-
J Gen Intern Med. 29(2):364–87. https://
d o i . o r g / 10 .15 5 8 5 / m m w r. m m 6 6 0 8 a1 ment of progress in cancer control. CA
d o i .o r g /10 .10 07/s116 0 6 - 013 - 2 5 8 3 -1
PMID:28253225 Cancer J Clin. 68(5):329–39. https://doi.
PMID:24002624
org/10.3322/caac.21460 PMID:30191964
12. Kutner M, Greenberg ER, Jin Y, Paulsen
4. Jemal A, Ward EM, Johnson CJ, Cronin
C (2006). The health literacy of America’s 22. Siegel RL, Sahar L, Robbins A, Jemal
KA, Ma J, Ryerson B, et al. (2017). Annual
adults: results from the 2003 National A (2015). Where can colorectal can-
report to the nation on the status of can-
Assessment of Adult Literacy. Washington cer screening interventions have
cer, 1975-2014, featuring survival. J Natl
(DC), USA: National Center for Education the most impact? Cancer Epidemiol
Cancer Inst. 109(9):djx030. https://doi.
Statistics. Available from: https://nces. Biomarkers Prev. 24(8):1151–6. https://
org/10.1093/jnci/djx030 PMID:28376154
ed.gov/pubs2006/2006483.pdf. doi.org/10.1158/1055-9965.EPI-15-0082
5. Nuru-Jeter AM, Michaels EK, Thomas PMID:26156973
13. Oldach BR, Katz ML (2014). Health litera-
MD, Reeves AN, Thorpe RJ Jr, LaVeist TA 23. American Cancer Society (1989). A sum-
cy and cancer screening: a systematic re-
(2018). Relative roles of race versus so- mary of the American Cancer Society
view. Patient Educ Couns. 94(2):149–57.
cioeconomic position in studies of health Report to the Nation: cancer in the
https://doi.org/10.1016/j.pec.2013.10.001
inequalities: a matter of interpretation. poor. CA Cancer J Clin. 39(5):263–5.
PMID:24207115
Annu Rev Public Health. 39(1):169–88. https://doi.org/10.3322/canjclin.39.5.263
ht t ps: //doi.o rg /10.114 6 /annurev- publ 14. Hendryx M, Luo J (2018). Increased can- PMID:2513098
health-040617-014230 PMID:29328880 cer screening for low-income adults under
the Affordable Care Act Medicaid expan- 24. Freeman HP (1989). Cancer in the so-
CHAPTER 4.6
SECTION 4
6. Institute of Medicine of the National cioeconomically disadvantaged. CA
sion. Med Care. 56(11):944–9. https://doi.
Academies (2003). Unequal treatment: Cancer J Clin. 39(5):266–88. https://
org/10.1097/MLR.0000000000000984
confronting racial and ethnic disparities doi.org /10. 3 32 2 /c anjc lin. 3 9. 5. 26 6
PMID:30199428
in health care. Washington (DC), USA: PMID:2513099
National Academies Press. https://doi. 15. Schootman M, Gomez SL, Henry KA,
org/10.17226/12875 Paskett ED, Ellison GL, Oh A, et al. 25. Centers for Disease Control and
(2017). Geospatial approaches to cancer Prevention (2016). National Breast and
7. Smith RA, Andrews KS, Brooks D, Cervical Cancer Early Detection Program
control and population sciences. Cancer
Fedewa SA, Manassaram-Baptiste D, (NBCCEDP). Available from: https://www.
Epidemiol Biomarkers Prev. 26(4):472–5.
Saslow D, et al. (2017). Cancer screen- cdc.gov/cancer/nbccedp/about.htm.
https://doi.org/10.1158/1055 -9965.EPI-
ing in the United States, 2017: a review
17-0104 PMID:28325736 26. Sabik LM, Adunlin G (2017). The ACA
of current American Cancer Society
and cancer screening and diagnosis.
guidelines and current issues in cancer 16. Smyth F (2008). Medical geography:
Cancer J. 23(3):151–62. https://doi.
screening. CA Cancer J Clin. 67(2):100– understanding health inequalities. Prog
org/10.1097/PPO.0000000000000261
21. https://doi.org/10.3322/caac.21392 Hum Geogr. 32(1):119–27. https://doi.
PMID:28537960
PMID:28170086 org/10.1177/0309132507080628
27. Grubbs SS, Polite BN, Carney J Jr,
8. Cronin KA, Miglioretti DL, Krapcho 17. Landrine H, Corral I, Lee JGL, Efird JT,
Bowser W, Rogers J, Katurakes N, et al.
M, Yu B, Geller BM, Carney PA, et al. Hall MB, Bess JJ (2017). Residential
(2013). Eliminating racial disparities in
(2009). Bias associated with self-report segregation and racial cancer dispari-
colorectal cancer in the real world: it took
of prior screening mammography. Cancer ties: a systematic review. J Racial Ethn a village. J Clin Oncol. 31(16):1928–30.
Epidemiol Biomarkers Prev. 18(6):1699– Health Disparities. 4(6):1195–205. https:// https://doi.org/10.1200/JCO.2012.47.8412
705. https://doi.org/10.1158/1055-9965. d o i.o r g /10.10 07/s 4 0 615 - 016 - 0 32 6 - 9 PMID:23589553
EPI-09-0020 PMID:19505902 PMID:28039602
9. Compilation of Patient Protection and 18. Henley SJ, Jemal A (2018). Rural can-
Affordable Care Act, in Title X (2010). cer control: bridging the chasm in geo-
Senate and House of Representatives of graphic health inequity. Cancer Epidemiol
the United States of America: Washington Biomarkers Prev. 27(11):1248–51. https://
(DC), USA. Available from: https://www. doi.org/10.1158/1055-9965.EPI-18-0807
hhs.gov/sites/default/files/ppacacon.pdf. PMID:30385497

4.7
Cancer in Indigenous populations

Focusing on inequalities that are
sometimes invisible
Diana Sarfati Malcolm King (reviewer)
Bridget H. Robson Diana R. Withrow (reviewer)
Gail Garvey
In 2018, WHO Director-General as “inheritors and practitioners of

SUMMARY Dr Tedros Adhanom Ghebreyesus unique cultures and ways of relat-
wrote, in an article on improving ing to people and the environment.
●● Cancer data relating to Indige the health of Indigenous people They have retained social, cultural,
nous people tend to be absent globally, “Health equity for the cur- economic, and political character-
or of poor quality, making many rent generation cannot wait, and istics that are distinct from those
Indigenous peoples statistically
we cannot fail future generations of of the dominant societies in which
invisible.
Indigenous people” [1]. they live” [3]. (For more details, see
●● Indigenous peoples tend to Indigenous peoples live in all “Who are Indigenous peoples?”.)
have higher rates of cancers re- regions of the world. There are esti- Indigenous paradigms common-
lated to tobacco exposure, alco- mated to be 370 million Indigenous ly embrace a holistic worldview that
hol consumption, poor diet, and people worldwide, living in more understand lands, waterways, seas,
high body mass index. than 90 countries and represent- the people, and all living things as vi-
●● These are all expected relation- ing 90% of the world’s cultural di- tally connected. Indigenous models
ships given the higher exposure versity [2]. The United Nations, ac- emphasize the importance of keep-
of Indigenous peoples to these knowledging that some countries ing social and economic activity in
risk factors; however, these pat- use different terms – such as First balance with the natural environ-
terns of exposure are in turn Peoples, First Nations, Nations, ment, thereby ensuring sustainabil-
related to societal and systemic Tribal, Aboriginal, Native, and eth- ity for generations to come.
determinants that can be traced nic groups – and that self-identifi- Colonization disrupts systems
to colonialism and racism. cation is a fundamental principle, of kinship between peoples and
recognizes Indigenous peoples with the natural world, intrudes on
●● Rates of chronic oncogenic in-
fections, particularly those that
are related to poverty and over-
crowding, tend to be higher in Who are Indigenous peoples?
Indigenous populations; exam-
ples are Helicobacter pylori, and • distinct language, culture, and
The United Nations Permanent Fo
hepatitis B virus in regions where
rum on Indigenous Issues uses the beliefs;
vaccination is not occurring.
following criteria to identify Indig • form non-dominant groups of
●● Toxic contamination of the en- enous peoples: society;
vironment has been linked to • self-identification as Indigenous • resolve to maintain and repro-
high cancer rates in some In peoples at the individual level, duce their ancestral environ-
digenous populations, such as and accepted by the community ments and systems as distinctive
those living near nuclear test as their member; peoples and communities.
sites in the Pacific.
• historical continuity with pre-colo-
●● Comprehensive, sustained ef- nial and/or pre-settler societies; Reference
forts are needed to improve can- • strong link to territories and sur- 1. United Nations Permanent Forum on
cer outcomes for Indigenous peo- Indigenous Issues (2018). Who are
rounding natural resources;
Indigenous peoples? Available from:
ples, centred around Indigenous • distinct social, economic, or po- https://www.un.org/esa/socdev/unpfii/
leadership and participation. litical systems; documents/5session_factsheet1.pdf.
288
traditional social and legal structures, groups tend to have the worst health
and imposes new cultural values, of identifiable ethnic groups; the FUNDAMENTALS
languages, and economic and politi- United Nations report on the state of
cal systems that serve to advantage Indigenous peoples concluded that ■■ There are estimated to be
the colonizing populations [4]. “discrimination against Indigenous 370 million Indigenous people
Through historical and cur- peoples, based on language, race, worldwide, living in more than
rent colonialism, the health of 90 countries.
culture, and identity, is rampant
Indigenous peoples is adversely af- across the Asian states” [2]. ■■ Indigenous peoples are the first
fected by destruction of their lands, Where data are available, Indig peoples of a country or region.
resources, and cultures, typically enous peoples tend to have high They have traditions and social,
resulting in marginalization, loss rates of preventable cancers, re- cultural, economic, and political
of autonomy, lower income levels, lated to tobacco exposure, alcohol characteristics that are distinct
worse living conditions, greater consumption, poor diet, and infec- from those of the new arrivals
food insecurity, and poorer access who later became dominant
tions [2,9,10].
through invasion, occupation,
to health, education, and other ser- The relationships between over- settlement, or other means.
vices [2,5]. These factors are ex- arching historical and contemporary
acerbated by health systems and forces shape the social determi- ■■ Indigenous peoples have a
other systems that generally do not nants of health, in turn influencing special relationship to their
reflect the worldview or practices both factors that enhance health and ancestral lands, seas, and
of Indigenous peoples. Indigenous waterways, and holistic
prevent cancer and those that af-
people may experience discrimina- understandings of health that
fect access to effective health care
are fundamentally important
tion and racism in their everyday (Fig. 4.7.1). These interacting ele- for their cultural and physical
lives and in their encounters with ments all affect cancer outcomes, survival and well-being.
the health system (Fig. 4.7.1) [2,6]. both positively and negatively, in
There is a lack of data relating Indigenous peoples globally. ■■ Colonization has taken differ-
to Indigenous peoples in almost event forms, involving varying
ery country in which they live; this degrees of violence, dispos-
greatly limits the extent to which in- Preventing cancer in session, dislocation, cultural
equalities in health and in upstream Indigenous peoples oppression, and discrimination.
Each has had impacts on the
determinants of health can be de-
Tobacco exposure social determinants of health
fined, measured, and addressed and on disparities in conditions
[2,7]. The United Nations estimates In several regions of the world,
of daily life experienced by
that about 80% of Indigenous peo- Indigenous populations have high
Indigenous peoples.
SECTION 4
CHAPTER 4.7
ples live in Africa, Asia, and Latin rates of exposure to tobacco [10–
America, but very little detailed 15]. In Australia, Canada, and New ■■ Colonization and systemic
information is available about the Zealand, the prevalence of smoking racism drive health inequities
is declining in all populations, but by the establishment of, and
health status of these peoples. For
despite that, Indigenous peoples perpetuation of, forces and
Indigenous and minority peoples in
systems, social norms, social
high-income countries like Australia, still have smoking rates that are
policies, and political systems
Canada, and the USA, there is not 2–3 times those of non-Indigenous
that serve to advantage the
necessarily better reporting or mea- peoples [11,12,14]. This pattern is colonizing populations.
surement of health outcomes. For also seen in the USA and in some
example, in Canada, authors have regions of Latin America, although ■■ The cancer burden and,
the differences between Indigenous more generally, the health
described “the absence of relevant,
and non-Indigenous populations of Indigenous peoples are
consistent, and inclusive Indigenous
significantly affected by the
identifiers in core population health vary [10,13].
broader social, political, and
data sources” [8]. As a result, rates of tobacco-re- economic environments as well
Despite this lack of data, it is lated cancers, particularly lung can- as by the legacy of colonization
clear from the existing literature cer, tend to be higher in Indigenous and racism.
that Indigenous peoples frequent- peoples [9,16]. For example, in New
ly face the double burden of high Zealand, the risk of lung cancer in ■■ Indigenous peoples must
be involved in the design,
rates of infectious diseases and a Māori is 3–4 times that in non-Māori
implementation, monitoring, and
rapidly increasing burden of non- [16]. However, it is also worth noting
quality improvement processes
communicable diseases, including that tobacco holds a sacred place of all policies related to health
cancer, as well as poor access to in the culture of some Indigenous (including the determinants of
health services [2]. For example, in populations and is used in traditional health) and to the elimination of
Asia, where there are massively di- rituals and ceremonies, although it is inequities in health care.
verse Indigenous populations, these not necessarily smoked or inhaled.
Chapter 4.7 • Cancer in Indigenous populations 289

Fig. 4.7.1. Drivers of equitable cancer outcomes among Indigenous peoples.
In the USA, some tobacco compa- colonization, which has had serious Indigenous populations. In some re-
nies historically appealed to these long-term effects on the health of gions, marginalized people in gen-
cultural connections to encourage Indigenous Australians [18]. eral, and Indigenous peoples in par-
the use of tobacco among Native ticular, tend to have higher or more
Americans [17]. In Australia, to- Alcohol consumption hazardous alcohol consumption; ex-
bacco was used by early colonists Alcohol consumption is related to amples are the Scheduled Tribes in
as payment for labour or as govern- several cancer types, including some regions of India and Indigenous
ment-funded rations – along with breast cancer, liver cancer, colorec- peoples in Australia and Canada
flour, tea, and sugar – to encour- tal cancer, oral cancer, and stomach [12,14,19,20]. In New Zealand and the
age Indigenous people to remain in cancer (see Chapter 2.3). Patterns of USA, Indigenous people and non-
White settlements. The underlying alcohol consumption vary marked- Indigenous people are similarly likely
sentiment of that time was one of ly around the world, including in to consume alcohol, but Indigenous
people are more likely to have a con-
sumption pattern that is hazardous to
Fig. 4.7.2. A woman from the Tupi–Guarani tribe in Brazil smoking tobacco in a pipe. their health [13,21].
Diet, physical activity, and

body mass index
Commonly, traditional diets of In
digenous people were high in fruits
and vegetables. As Indigenous peo-
ple have lost access to their tradi-
tional foods and land, and societies
have become more urbanized, food
insecurity has been cited as a ma-
jor contributor to the health inequal-
ities faced by Indigenous people.
For example, in New Zealand, 29%
of Māori reported food insecurity
compared with 14% of New Zealand
Europeans [22]. In Africa and Asia,
Indigenous people are more likely
to be poorly nourished compared
with non-Indigenous people [2].
290
Fig. 4.7.3. A Mayan woman selling fruits and vegetables at a market in San Cris B virus infection, which increases
tób al, Mexico. the risk of primary liver cancer, re-
main higher in Indigenous people in
Australia and New Zealand, and in
the Inuit of Canada, although infec-
tion rates are generally declining as
a result of successful vaccination
programmes [33–36]. In general,
rates of infections including HIV, zo-
onotic infections, and tuberculosis
tend to be high in Indigenous popu-
lations in Africa and Asia [2].
In parts of Africa and Asia,
Indigenous peoples have higher
rates of HIV infection than other
groups because of a range of fac-
tors, which are compounded by the
fact that many of the Indigenous
peoples live in remote and hard-
to-reach places, making access to
health care extremely difficult. HIV
infection is associated with sev-
eral cancer types, including Kaposi
sarcoma and B-cell lymphomas.
Although very few data exist on
Patterns of physical activity Chronic infections these populations, it is likely that
are highly variable, and few coun- Infection with human papillomavirus the rates of these associated can-
tries measure the physical activity (HPV) is common in many coun- cer types are also high in these
of their Indigenous populations. In tries, and generally does not seem Indigenous populations [2].
those countries that do report this, to occur with substantially greater
the picture is a mixed one, with frequency in Indigenous popula- Environmental degradation
some countries reporting similar tions, although the specific patterns Loss and degradation of land and
or mixed levels of physical activ-
SECTION 4
CHAPTER 4.7
vary between countries [10,24–26]. resources are critical determinants
ity between Indigenous and non- Despite this, rates of cervical can- of health for Indigenous populations
Indigenous peoples [14,21], and cer are often higher in Indigenous around the globe. These factors re-
some countries suggesting that people, probably reflecting poorer sult in disempowerment, political
Indigenous peoples may be more access to screening and other marginalization, and loss of au-
likely to be sedentary [12,13]. health services [9,10,27,28]. tonomy, which have impacts on all
Consistent with patterns globally, In contrast, rates of oncogenic aspects of health and well-being. In
rates of overweight and obesity are infections that are strongly related addition to these broad considera-
tending to increase in Indigenous to poverty and overcrowding tend to tions, there are many examples of
populations; however, the increases be substantially higher in Indigenous environmental damage that poten-
are tending to occur more rapidly people. An example is Helicobacter tially has a direct impact on cancer
and more severely in Indigenous pylori, an important cause of stom- risk in Indigenous peoples.
populations in many countries, in- ach cancer (see Chapter 5.4). Environmental contamination
cluding Canada, the USA, Australia, Infection with H. pylori is strongly has been associated with concerns
New Zealand, and countries in related to overcrowding, particu- about increased risk of cancer in
several regions of Latin America larly in childhood. Rates of H. pylori some Indigenous groups in the west-
[10,12–14,23]. A recent study in New infection in Indigenous people are ern USA, through contamination of
Zealand showed that although to- 2–3 times those in non-Indigenous water and soil with cadmium, arse-
bacco-related cancers remained the people in both Australia and New nic, uranium, and other heavy met-
main driver of inequalities in cancer Zealand, and very high prevalence als [37]. Similarly, oil drilling in the
incidence between Māori and non- rates of H. pylori infection have been Amazon basin of Ecuador caused
Māori, rates of obesity-related can- found in Indigenous populations in continuous contamination, which
cers, including breast cancer and Canada, the USA, the circumpolar may have resulted in higher cancer
endometrial cancer, were increasing region, and Latin America [29–32]. incidence in local Indigenous popu-
the most rapidly [16]. Similarly, rates of chronic hepatitis lations [38]. However, the starkest

Fig. 4.7.4. Two members of the Nunukul Yuggera Aboriginal Dance Company perform ally. They often face political and
at the opening ceremony of the inaugural World Indigenous Cancer Conference, held social isolation, prejudice, and
in Australia in 2016. poverty. These influence their
health and quality of life, and are
reflected in issues across the can-
cer continuum. The current state
of Indigenous health is the direct
result of past policies related to
colonization [2,4,5,41]. Data relating
to Indigenous people are scarce.
Indigenous people are more likely
to be exposed to risk factors for
many cancer types, and for many
Indigenous groups there are sub-
stantial barriers to accessing cancer
services and other health services.
Indigenous peoples have rich,
holistic, complex, and heteroge-
neous worldviews, which are cen-
tral to their health and well-being.
Article 24 of the United Nations
Declaration on the Rights of
Indigenous Peoples clearly articu-
lates that Indigenous peoples have
the right to the highest attainable
example of environmental contami- was 26 percentage points lower for
standard of health. Signatories are
nation was seen after the nuclear Indigenous women than for non-In-
obliged to take action to improve
testing in the Pacific. Testing by the digenous women (41.8% vs 68.3%)
the health of Indigenous peoples
USA on Bikini Atoll in the Marshall [27]. In New Zealand, participation
within their countries. This means
Islands in 1954 was “the most se- rates in screening for breast can-
actively identifying and address-
rious episode of radioactive con- cer, colorectal cancer, and cervical
cancer have improved for Māori over ing social, economic, and political
tamination in the history of nuclear
time but still remain lower than rates structural barriers, which hinder the
weapons testing” [39]. It resulted attainment of equitable health for
in a continuing excess of thyroid for non-Māori [16]. In Canada and
the USA, there are smaller differ- Indigenous peoples.
cancer and other cancer types in Improving cancer outcomes for
the local Indigenous population, ences in rates for breast cancer and
cervical cancer screening between Indigenous peoples requires that
as well as massive pollution of the achieving equity is a central prior-
marine ecosystem. Nuclear test- Indigenous and non-Indigenous
women, and in general screen- ity and that all action must have
ing by France in the Moruroa and Indigenous leadership, participa-
ing rates are improving [13,14]. In
Fangataufa atolls has also resulted tion, and decision-making at its
low- and middle-income countries
in continuing high rates of thyroid core [2,5]. It must include improve-
throughout Africa, Asia, and the
cancer in the Indigenous popula- ment of data related to Indigenous
Pacific region, screening services
tions of French Polynesia [40]. peoples, including Indigenous iden-
are frequently poorly coordinated, of
low quality, or completely absent for tifiers, which will enable Indigenous
Cancer screening peoples to identify and prioritize
many Indigenous people.
Effective cancer screening can re- their health needs [7,42]. There is
duce both the incidence and the im- an urgent need for comprehensive,
pact of cancer (see Chapter 6.6), but How cancer outcomes in sustained efforts to improve can-
services may not meet the needs Indigenous peoples may cer outcomes for Indigenous peo-
of Indigenous peoples. In Australia, be improved ples, grounded in the principles of
Whop et al. found that 3-year partici- Indigenous people are among the Indigenous autonomy and empow-
pation in cervical cancer screening most marginalized peoples glob- erment (Fig. 4.7.5).
292
Fig. 4.7.5. Framework for intervention to improve cancer outcomes among Indigenous peoples.
References
SECTION 4
CHAPTER 4.7
1. Ghebreyesus TA (2018). Improving the 6. Harris R, Tobias M, Jeffreys M, Waldegrave 10. Moore SP, Forman D, Piñeros M,
health of Indigenous people globally. Lancet K, Karlsen S, Nazroo J (2006). Effects of Fernández SM, de Oliveira Santos M, Bray
Oncol. 19(6):e277. https://doi.org/10.1016/ self-reported racial discrimination and F (2014). Cancer in indigenous people in
S1470-2045(18)30375-9 PMID:29893252 deprivation on Māori health and inequal- Latin America and the Caribbean: a re-
ities in New Zealand: cross-sectional view. Cancer Med. 3(1):70–80. https://doi.
2. United Nations Department of Economic study. Lancet. 367(9527):2005–9. https:// org/10.1002/cam4.134 PMID:24403278
and Social Affairs (2015). State of the doi.org/10.1016/S0140-6736(06)68890-9
11. Ball J, Stanley J, Wilson N, Blakely T,
world’s Indigenous peoples, 2nd volume: PMID:16782491
Edwards R (2016). Smoking prevalence
Indigenous peoples’ access to health
7. Sarfati D, Garvey G, Robson B, Moore S, in New Zealand from 1996-2015: a criti-
services. United Nations Publications. cal review of national data sources to
Cunningham R, Withrow D, et al. (2018).
Available from: https://www.un.org/esa/ inform progress toward the Smokefree
Measuring cancer in indigenous popu-
socdev/unpf ii/documents/2016/Doc s- 2025 goal. N Z Med J. 129(1439):11–22.
lations. Ann Epidemiol. 28(5):335–42.
updates/SOWIP_Health.pdf. PMID:27507718
ht t ps: //do i.o r g /10.1016 / j. annepi dem.
3. United Nations (2018). Indigenous peo- 2018.02.005 PMID:29503062 12. AIHW (2016). Australia’s Health 2016. 4.8:
ples at the UN. https://www.un.org/devel 8. Smylie J, Firestone M (2015). Back to Health behaviours and biomedical risks of
opment/desa/indigenouspeoples/about- the basics: identifying and addressing Indigenous Australians. Canberra, Australia:
us.html underlying challenges in achieving high Australian Institute of Health and Welfare.
quality and relevant health statistics for 13. Cobb N, Espey D, King J (2014). Health be-
4. Reid P, Robson B (2007). Understanding
indigenous populations in Canada. Stat J haviors and risk factors among American
health inequities. In: Robson B, Harris
IAOS. 31(1):67–87. http://doi.org/10.3233/ Indians and Alaska Natives, 2000-
R, editors. Haoura: Māori standards of SJI-150864 PMID:26793283 2010. Am J Public Health. 104(Suppl
health IV: a study of the years 2000–
3):S481–9. https://doi.org/10.2105/AJPH.
2005. Wellington, New Zealand: Te Rōpū 9. Moore SP, Antoni S, Colquhoun A, Healy
2014.301879 PMID:24754662
Rangahau Hauora a Eru Pōmare; pp. 3–10. B, Ellison-Loschmann L, Potter JD, et
al. (2015). Cancer incidence in indig- 14. Withrow DR, Amartey A, Marrett LD
5. King M, Smith A, Gracey M (2009). enous people in Australia, New Zealand, (2014). Cancer risk factors and screening
Indigenous health part 2: the underly- Canada, and the USA: a comparative in the off-reserve First Nations, Métis and
ing causes of the health gap. Lancet. population-based study. Lancet Oncol. non-Aboriginal populations of Ontario.
374(9683):76–85. https://doi.org/10.1016/ 16(15):1483–92. https://doi.org/10.1016/ Chronic Dis Inj Can. 34(2–3):103–12.
S0140-6736(09)60827-8 PMID:19577696 S1470-2045(15)00232-6 PMID:26476758 PMID:24991773

15. Janakiram C, Joseph J, Vasudevan S, 25. Schmidt-Grimminger DC, Bell MC, Muller 34. Graham S, Guy RJ, Cowie B, Wand
Taha F, Deepan Kumar C, Venkitachalam CJ, Maher DM, Chauhan SC, Buchwald HC, Donovan B, Akre SP, et al. (2013).
R, et al. (2016). Prevalence and de- DS (2011). HPV infection among rural Chronic hepatitis B prevalence among
pendency of tobacco use in an American Indian women and urban White Aboriginal and Torres Strait Islander
Indigenous population of Kerala, India. women in South Dakota: an HPV preva- Australians since universal vaccination:
J Oral Hyg Health. 4(1):198. https://doi. lence study. BMC Infect Dis. 11(1):252. a systematic review and meta-analysis.
org/10.4172/2332-0702.1000198 https://doi.org/10.1186/1471-2334-11-252 BMC Infect Dis. 13(1):403–403. https://
16. Teng AM, Atkinson J, Disney G, Wilson PMID:21943050 d o i . o r g / 1 0 .118 6 / 1471 - 2 3 3 4 -13 - 4 0 3
N, Sarfati D, McLeod M, et al. (2016). PMID:24004727
26. Zehbe I, Moeller H, Severini A, Weaver B,
Ethnic inequalities in cancer incidence
Escott N, Bell C, et al. (2011). Feasibility 35. Kim WR (2009). Epidemiology of hepatitis
and mortality: census-linked cohort stud-
of self-sampling and human papillomavi- B in the United States. Hepatology. 49(5
ies with 87 million years of person-time
rus testing for cervical cancer screening Suppl):S28–34. https://doi.org/10.1002/
follow-up. BMC Cancer. 16(1):755. https://
in First Nation women from Northwest hep.22975 PMID:19399791
d o i .o r g /10 .118 6 /s12 8 8 5 - 016 - 2 781- 4
PMID:27669745 Ontario, Canada: a pilot study. BMJ Open.
36. Minuk GY, Uhanova J (2003). Viral hepatitis
1(1):e000030. https://doi.org/10.1136/
17. Maron DF (2018). The fight to keep tobac- in the Canadian Inuit and First Nations pop-
bmjopen-2010-000030 PMID:22021733
co sacred. Scientific American. 29 March ulations. Can J Gastroenterol. 17(12):707–
2018. https://www.scientificamerican.com/ 27. Whop LJ, Garvey G, Baade P, 12. https://doi.org/10.1155/2003/350175
article/the-fight-to-keep-tobacco-sacred/ Cunningham J, Lokuge K, Brotherton PMID:14679425
JML, et al. (2016). The first comprehen-
18. Gracey M, King M (2009). Indigenous 37. Lewis J, Hoover J, MacKenzie D (2017).
sive report on Indigenous Australian
health part 1: determinants and disease Mining and environmental health dis-
women’s inequalities in cervical screen-
patterns. Lancet. 374(9683):65–75. https:// parities in Native American communities.
ing: a retrospective registry cohort study
doi.org/10.1016/S0140-6736(09)60914-4 Curr Environ Health Rep. 4(2):130–41.
PMID:19577695 in Queensland, Australia (2000-2011).
https://doi.org/10.1007/s40572-017-0140-5
Cancer. 122(10):1560–9. https://doi.
PMID:28447316
19. Kumar RK, Tiwari R (2016). A cross org/10.1002/cncr.29954 PMID:27149550
sectional study of alcohol consump- 38. Hurtig A-K, San Sebastián M (2002).
tion among tribal and non-tribal adults of 28. McLeod M, Harris R, Purdie G, Cormack Geographical differences in can-
Narayanganj block in Mandla district of D, Robson B, Sykes P, et al. (2010). cer incidence in the Amazon basin of
Madhya Pradesh, India. Int J Community Improving survival disparities in cervical Ecuador in relation to residence near
Med Public Health. 3(4):791–5. https://doi. cancer between Māori and non-Māori oil fields. Int J Epidemiol. 31(5):1021–7.
org/10.18203/2394-6040.ijcmph20160737 women in New Zealand: a national ret- ht t ps: //d o i.o r g /10.10 9 3 / ij e / 31. 5 .10 21
rospective cohort study. Aust N Z J PMID:12435778
20. Mohindra KS, Narayana D, Anushreedha
Public Health. 34(2):193–9. https://doi.
SS, Haddad S (2011). Alcohol use and its
org/10.1111/j.1753 - 6405.2010.00506.x 39. Prăvălie R (2014). Nuclear weapons tests
consequences in South India: views from
PMID:23331365 and environmental consequences: a
a marginalised tribal population. Drug
global perspective. Ambio. 43(6):729–44.
Alcohol Depend. 117(1):70–3. https://doi. 29. Goodman KJ, Jacobson K, Veldhuyzen https://doi.org/10.1007/s13280-014-0491-1
org/10.1016/j.drugalcdep.2010.12.021 van Zanten S (2008). Helicobacter
PMID:21282019 PMID:24563393
pylori infection in Canadian and re-
21. Ministry of Health (2016). Annual update of lated Arctic Aboriginal populations. 40. Simon SL, Bouville A, Land CE (2006).
key results 2015/16: New Zealand Health Can J Gastroenterol. 22(3):289–95. Fallout from nuclear weapons tests and
Survey. Wellington, New Zealand: Ministry ht t ps: //do i.o r g /10.115 5 / 2 0 0 8 / 25 8 610 cancer risks. Am Sci. 94(1):48–57. https://
of Health. Available from: https://www. PMID:18354758 doi.org/10.1511/2006.57.982
health.gov t.nz /public ation /annual - up
30. McDonald AM, Sarfati D, Baker MG, 41. Truth and Reconciliation Commission of
date-key-results-2015-16-new-zealand-
Blakely T (2015). Trends in Helicobacter Canada (2015). Calls to action. Winnipeg,
health-survey.
pylori infection among Māori, Pacific, and Canada: Truth and Reconciliation
22. Carter KN, Lanumata T, Kruse K, Gorton European birth cohorts in New Zealand. Commission of Canada. Available from:
D (2010). What are the determinants of Helicobacter. 20(2):139–45. https://doi. https://nctr.ca/assets/reports/Calls_to_
food insecurity in New Zealand and does org/10.1111/hel.12186 PMID:25403622 Action_English2.pdf.
this differ for males and females? Aust N Z
J Public Health. 34(6):602–8. https://doi. 31. Robinson L-GE, Black FL, Lee FK, Sousa 42. Coleman C, Elias B, Lee V, Smylie J,
org/10.1111/j.1753 - 6405.2010.00615.x AO, Owens M, Danielsson D, et al. (2002). Waldon J, Schanche Hodge F, et al. (2016).
PMID:21134063 Helicobacter pylori prevalence among International Group for Indigenous Health
indigenous peoples of South America. Measurement: recommendations for best
23. Ministry of Health (2011). A focus on Māori practice for estimation of Indigenous mor-
nutrition: findings from the 2008/09 New J Infect Dis. 186(8):1131–7. https://doi.
org/10.1086/343803 PMID:12355364 tality. Stat J IAOS. 32(4):729–38. https://
Zealand Adult Nutrition Survey. Wellington, doi.org/10.3233/SJI-161023
New Zealand: Ministry of Health. Available 32. Windsor HM, Abioye-Kuteyi EA, Leber
from: https://www.health.govt.nz/publication/
JM, Morrow SD, Bulsara MK, Marshall BJ
focus-nutrition-key-findings-2008-09-nz-
(2005). Prevalence of Helicobacter pylori
adult-nutrition-survey.
in Indigenous Western Australians: com-
24. Garland SM, Brotherton JML, Condon JR, parison between urban and remote rural
McIntyre PB, Stevens MP, Smith DW, et populations. Med J Aust. 182(5):210–3.
al.; WHINURS study group (2011). Human PMID:15748129
papillomavirus prevalence among indig-
enous and non-indigenous Australian 33. Addidle M (2011). Impact of universal hep-
women prior to a national HPV vaccina- atitis B vaccination on antenatal hepatitis
tion program. BMC Med. 9(1):104–104. B prevalence in the Midlands region of the
https://doi.org/10.1186/1741-7015-9-104 North Island, New Zealand. N Z Med J.
PMID:21910918 124(1332):40–4. PMID:21747422
294
Towards the World Code Against Cancer
Carolina Espina and Joachim Schüz
Prevention offers the greatest public been enthusiastically promoted by sociocultural norms, risk factor pat-
health potential and the most cost- the European cancer associations. terns, cancer burden, and the state
effective long-term cancer control. The ECAC also acts as a guide to of development of health systems.
However, with today’s multiple media aid in the development of national These differences underscore the
streams, the general public is often health policies in cancer preven- importance of an in-depth appraisal
overwhelmed by an abundance of tion and provides an important ba- of the recommendations on primary
confusing, ambiguous, or apparently sis for health promotion. However, and secondary prevention of cancer
contradictory messages on disease for the ECAC to achieve its full im- in other regions of the world.
prevention. It has been estimated pact, wider dissemination among The adapted Codes Against
that at least 40% of cancer cases both the general public and policy- Cancer will offer exceptional public
could be prevented through actions makers is needed, as well as period- health tools to support governments
targeted towards risk prevention at ic updates. The ECAC emphasizes in the implementation of cancer con-
the individual or population level. that its 12 recommendations need trol strategies adapted to the local
What can we recommend to people to be aligned with population-level needs, priorities, and resources.
to reduce their risk of cancer? preventive actions, either supported Consideration of such an adapted
The European Code Against by policies aimed at minimizing ex- model illustrates why a simple trans-
Cancer (ECAC) is an integrated posures that are beyond the control lation of the ECAC would not be suf-
multirisk instrument for cancer pre- of individuals or by empowering in- ficient to promote cancer prevention
vention that informs the general dividuals to enable them to comply globally. In addition, support from
public about how to avoid or reduce with the recommendations. authoritative regional leaders in can-
exposures to established causes of The experience of developing cer prevention and in cancer control
cancers, to adopt behaviours to re- and promoting the ECAC has gen- enables regional ownership of the
duce cancer risk, and to participate erated interest in developing such recommendations, and may help to
in vaccination programmes and or- a set of recommendations for oth- secure the highest acceptance and
ganized screening programmes ac- er regions of the world. Under the uptake, both by the general public
cording to the respective national overall umbrella of a World Code and by those working in the health
guidelines, by following 12 recom- Against Cancer using the same system. Broad involvement of the
mendations [1]. The ECAC carries IARC methodology, regional Codes scientific community and of civil so-
the authority of the leading expert Against Cancer would be devel- ciety networks to ensure the most
scientists, who worked under the co- oped. They would focus on regions suitable dissemination and advoca-
ordination of IARC to develop a rig- sufficiently large but also distinct cy is key for the successful imple-
orous evidence-based methodology enough to merit the development of mentation of the recommendations.
to synthesize the scientific evidence, versions adapted to differences in
leading to the update of the ECAC risk factors and cancer patterns, as References
(4th edition) in 2014. Several work- well as economic, social, and cul-
1. Schüz J, Espina C, Villain P, Herrero R,
ing groups of cancer experts and, tural conditions [2]. Leon ME, Minozzi S, et al.; Working Groups
importantly, experts in the commu- The main goal of developing re- of Scientific Experts (2015). European
nication of health messages worked gional Codes Against Cancer would Code Against Cancer 4th edition: 12
together to revise the previous rec- be to raise awareness about risk ways to reduce your cancer risk. Cancer
Epidemiol. 39(Suppl 1):S1–10. https://
ommendations. As a result, the factors and the available preven-
doi.org /10.1016/ j.c anep. 2015.0 5.0 0 9
ECAC stands out among other initiation measures by effectively com- PMID:26164654
tives for its clarity and accessibility municating the current state of the
as a short set of recommendations science and, as a consequence, 2. Espina C, Herrero R, Sankaranarayanan R,
Krug E, Wild CP, Schüz J (2018). Toward
for the general public. empowering individuals and commu- the World Code Against Cancer. J Glob
The messages of the ECAC nities. Other world regions differ from Oncol. (4):1–8. https://doi.org/10.1200/
are aimed at individuals and have the European context in terms of JGO.17.00145 PMID:30241265
Towards the World Code Against Cancer 295

EUROPEAN CODE AGAINST CANCER

12
ways to reduce your cancer risk
1 Do not smoke. Do not use any form of tobacco.
2 Make your home smoke free. Support smoke-free policies in your workplace.
3 Take action to be a healthy body weight.
4 Be physically active in everyday life. Limit the time you spend sitting.
5 Have a healthy diet:

 Eat plenty of whole grains, pulses, vegetables and fruits.
 Limit high-calorie foods (foods high in sugar or fat) and avoid sugary drinks.
 Avoid processed meat; limit red meat and foods high in salt.
6 If you drink alcohol of any type, limit your intake. Not drinking alcohol is better
for cancer prevention.
7 Avoid too much sun, especially for children. Use sun protection. Do not use
sunbeds.
8 In the workplace, protect yourself against cancer-causing substances by following

health and safety instructions.
9 Find out if you are exposed to radiation from naturally high radon levels in your
home. Take action to reduce high radon levels.
10 For women:
 Breastfeeding reduces the mother’s cancer risk. If you can, breastfeed your baby.
 Hormone replacement therapy (HRT) increases the risk of certain cancers.
Limit use of HRT.
11 Ensure your children take part in vaccination programmes for:

 Hepatitis B (for newborns)
 Human papillomavirus (HPV) (for girls).
12 Take part in organized cancer screening programmes for:

 Bowel cancer (men and women)
 Breast cancer (women)
 Cervical cancer (women).
The European Code Against Cancer focuses on actions that individual citizens can take to help prevent cancer.
Successful cancer prevention requires these individual actions to be supported by governmental policies and actions.
Find out more about the European Code Against Cancer at: http://cancer-code-europe.iarc.fr
This project is co-financed by the European Union and coordinated by the specialized cancer agency of the
World Health Organization, the International Agency for Research on Cancer.
296
5 Preventing particular
tumour types
Cancer is not a single disease but a multi- sporadic disease and detection of precan-
plicity of variously related diseases. This un- cerous lesions. Screening procedures can
derstanding is as applicable and relevant to be meaningfully explored only with respect
cancer prevention as it is to the clinical man- to particular cancer sites. For many cancer
agement of cancer. Broad knowledge about types, there are no recognized population-
cancer causation, development, detection, based screening procedures. However, suc-
and avenues to prevention must be qualified cess with respect to any research aspect of
according to the tumour type or subtype be- tumour development or a preventive measure
ing considered. Descriptions of causation and for one tumour type often indicates a possi-
prevention cannot be given uniformly for all ble way to approach the same challenge for
cancer types. For example, exogenous caus- at least one other tumour type and perhaps
es of prostate cancer are not evident; for now, many other tumour types.
prevention of prostate cancer must focus on
A guide to the epidemiology data in Section 5: Preventing particular tumour types
Typically, epidemiology is dealt with Data source Standardization is necessary when

at the beginning of each chapter. The incidence and mortality data comparing several populations (or
Unless otherwise stated, all of the are based on national incidence the same population at different time
incidence and mortality data are and mortality estimates from the points); age has a powerful influence
from the GLOBOCAN 2018 data- GLOBOCAN 2018 database [1]. This on the risk of cancer, and popula-
base. Further information about the provides estimates of incidence and tions differ with respect to their age
epidemiology data is provided here. mortality for 36 site-specific cancer distribution. Here, the ASR uses the
types and for all cancer sites com- World Standard Population (of Segi
Incidence
bined for 185 countries or territories [3], as modified by Doll et al. [4]).
Cancer incidence is defined as the of the world in 2018, by sex and age The calculated incidence or mortal-
number of new cancer cases arising group. The underlying principle in ity rate is then called the age-stan-
in a specified population over a given the estimation process is a reliance dardized incidence or mortality rate
period of time (typically 1 year). It can on the best available data on cancer (World) and is conventionally ex-
be expressed as an absolute number incidence and/or mortality within a pressed per 100 000 person-years.
of cases within the entire population country to build up the global pic-
per year or as a rate per 100 000 ture. The results are more accurate References
persons per year. The incidence rate or less accurate for different coun- 1. Bray F, Ferlay J, Soerjomataram I, Siegel
provides an approximation of the av- tries, depending on the extent and RL, Torre LA, Jemal A (2018). Global
erage risk of developing a cancer. accuracy of locally available data. cancer statistics 2018: GLOBOCAN es-
Incidence information is collected timates of incidence and mortality world-
routinely by cancer registries. Data visualization tools wide for 36 cancers in 185 countries. CA
Cancer J Clin. 68(6):394–424. https://doi.
The Cancer Today subsection of the org/10.3322/caac.21492 PMID:30207593
Mortality Global Cancer Observatory [2] pro-
Cancer mortality is defined as the vides data visualization tools to ex- 2. Ferlay J, Ervik M, Lam F, Colombet M,
number of deaths due to cancer Mery L, Piñeros M, et al. (2018). Global
plore the current scale and profile of
Cancer Observatory: Cancer Today. Lyon,
occurring in a specified population cancer worldwide using incidence, France: International Agency for Research
over a given period of time (typi- mortality, and prevalence estimates on Cancer. Available from: https://gco.iarc.
cally 1 year). It can be expressed from the GLOBOCAN 2018 database. fr/today.
as an absolute number of deaths
within the entire population per Age standardization 3. Segi M (1960). Cancer mortality for se-
lected sites in 24 countries (1950–57).
year or as a rate per 100 000 per- All incidence and mortality rates pro- Department of Public Health, Tohoku
sons per year. The mortality rate vided in the chapters are age-stan University of Medicine, Sendai, Japan.
provides an approximation of the dardized. An age-standardized rate
4. Doll R, Payne P, Waterhouse JAH, edi-
average risk of death from a can- (ASR) is a summary measure of the tors (1966). Cancer incidence in five con-
cer. Mortality data are provided by rate that a population would have tinents. Volume I. Geneva, Switzerland:
national statistical offices. if it had a standard age structure. International Union Against Cancer.
298
5.1
Lung cancer
Continues to be the leading cause
of cancer death
Rayjean J. Hung Joanna Didkowska (reviewer)
Adi F. Gazdar Mattias Johansson (reviewer)
There are four main histological hydrocarbons, tobacco-specific ni-

SUMMARY types of lung cancer: adenocarci- trosamines, and benzene [4,5].
noma, squamous cell carcinoma, Tobacco smoking is known to
●● Lung cancer continues to be small cell carcinoma, and large cell have a stronger effect on squamous
the most common cancer type carcinoma, each with different mor- cell carcinoma and small cell lung
and the leading cause of cancer phological features, molecular char- carcinoma (SCLC) than on adeno-
death worldwide. acterization, and etiology; the most carcinoma [6]. In addition, the effect
common types are adenocarcinoma of smoking on risk of squamous cell
●● Relative to the hazards of smok-
and squamous cell carcinoma [1]. carcinoma and SCLC increases with
ing tobacco cigarettes, the haz-
increased smoking duration and de-
ards presented by e-cigarettes
Epidemiology creases rapidly after smoking ces-
and by cannabis smoking are
sation. The effect of smoking on risk
largely unknown. Lung cancer continues to be the of adenocarcinoma decreases less
leading cause of cancer death rapidly after smoking cessation; this
●● The role of lung diseases, includ-
worldwide, accounting for about partly explains the increasing per-
ing chronic obstructive pulmo-
18% of all cancer deaths [2]. The centage of adenocarcinoma in coun-
nary disease and emphysema,
highest incidence rates of lung tries that are in a late stage of the to-
in lung cancer is now clearer.
cancer are observed in parts of bacco epidemic. Another contributor
●● Several lung cancer susceptibil- North America, in East Asia, and in to the increase in lung adenocarci-
ity loci have been identified in parts of central and eastern Europe noma in smokers is the introduction
the past decade, and more con- (Fig. 5.1.1) [2]. Incidence rates in of filtered and low-tar or low-nicotine
tinue to be discovered through men have declined during the past cigarettes [7].
large-scale collaborations. four decades in most countries, Apart from tobacco smoking,
whereas incidence rates in women about 29 agents have been recog-
●● Comprehensive molecular pro- continue to rise, with a few excep- nized to cause lung cancer, with vary-
filing of adenocarcinoma, squa- tions (Fig. 5.1.2) [3]. Because lung ing degrees of risk and prevalences
mous cell carcinoma, and small cancer survival is low globally, in of exposure. These include asbestos,
cell carcinoma has been carried general the trends in mortality rates
SECTION 5
CHAPTER 5.1
silica, several heavy metals, and ra-
out. Some molecular changes over time correspond to the trends don (see Chapter 2.10). In addition,
provide druggable targets. in incidence rates. indoor air pollution from household
●● Lung cancer in never-smokers combustion of solid fuel and cooking
is a specific disease entity. Etiology fumes in poorly ventilated homes was
established as a lung carcinogen,
●● Lung cancer screening by low- Carcinogens predominantly on the basis of studies
dose computed tomography The major cause of lung cancer is in female never-smokers in Asia (see
in high-risk populations repre- tobacco smoking (see Chapter 2.1), Chapter 4.3). More recently, outdoor
sents an opportunity for mortal- which is responsible for 80–85% of air pollution, particulate matter in out-
ity reduction, but its efficiency lung cancer cases worldwide; tobac- door air pollution, and one specific
will be improved by individual co smoke contains more than 7000 pollutant – diesel engine exhaust –
risk prediction. chemicals and at least 69 carcino- have each been classified by the
gens, including polycyclic aromatic IARC Monographs as carcinogenic
Chapter 5.1 • Lung cancer 299

to humans (Group 1), on the basis therefore the possibility of an as-

of consistency in large pooled anal- sociation in heavy users cannot be
yses and prospective cohort stud- excluded [14].
FUNDAMENTALS
ies (see Chapter 2.9). These agents
Previous lung disease ■■ There are four main
can have increasing importance as
histological types of lung
causes of lung cancer, especially in In addition to the known lung car-
never-smokers. The established lung cinogens, previous lung diseases cancer: adenocarcinoma,
cancer carcinogens are included in were shown to be associated with squamous cell carcinoma,
the list of IARC Monographs classi- risk of lung cancer. In particular, it small cell carcinoma, and
fications [8] (see “Known causes of is well established that chronic ob- large cell carcinoma, each
human cancer by organ site”, p. 45). structive pulmonary disease is asso- with different morphological
The prevalence of tobacco ciated with risk of lung cancer [15]; features, molecular charac
smoking has declined in most high- this association can be explained at
terization, and etiology;
income countries during the past least partly by shared etiology, such
the most common types
few decades [9]. Recently, alter- as tobacco smoking and chronic in-
native smoking products have be- flammation [16]. The International are adenocarcinoma and
come popular. In addition, the use Lung Cancer Consortium conducted squamous cell carcinoma.
of cannabis has been legalized in a series of pooled analyses based
■■ The major cause of lung cancer
some countries. Therefore, recent on 17 studies with a total of 24 607
research efforts related to puta- is tobacco smoking. Other
lung cancer cases and 81 829 con-
tive lung cancer risk factors have trols. Although a history of chronic causes of lung cancer include
focused on electronic nicotine de- obstructive pulmonary disease was asbestos, silica, several heavy
livery systems (also called e-ciga- shown to be associated with lung metals, radon, and indoor and
rettes) and cannabis smoking. cancer risk, only emphysema was outdoor air pollution.
To date, studies on e-cigarettes associated with risk of lung cancer in
have been based predominantly never-smokers, and this association ■■ Risk of lung cancer is also
on cell culture or animal studies, persisted even when considering a affected by an individual’s
which have demonstrated that e- history of emphysema 5–10 years genetic susceptibility.
cigarettes have pulmonary toxic- before the diagnosis of lung cancer
ity, although to a much smaller [17]. A similar association was found ■■ Lung cancer survival remains
extent than tobacco smoking [10]. for pneumonia, based on a pooled dismal, with 5-year survival
Therefore, e-cigarettes are con- analysis of 12 studies [17]. rates of 10–20%, because
sidered by some to be an effective most patients are diagnosed at
tool for harm reduction. However, Genetic susceptibility late stages of the disease.
because very limited data are avail- Although tobacco smoking is the
able in humans, much effort will be main risk factor for lung cancer,
required to fully monitor the effect only about 15% of smokers even-
of e-cigarettes on lung cancer risk tually develop lung cancer [18]. A in EGFR and HER2 in predomi-
and nicotine addiction, given the in- genetic component of lung cancer nantly never-smokers have recently
creasing popularity of e-cigarettes etiology is recognized on the basis been described [22,23]. However,
as an alternative to tobacco ciga- of familial studies, and the analyses these high-penetrance mutations
rettes, particularly among young either accounted for smoking or fo- only account for perhaps 1% of lung
people [10,11]. cused on never-smokers [18]. The cancer cases.
Cannabis has been legalized familial relative risk of lung cancer is In the past decade, genome-
in Canada, in 28 states of the USA consistently estimated to be about wide association studies (GWAS)
for medicinal use, and in several 2-fold across several large cancer (see Chapter 3.2) have identified
European countries. Cannabis registries [19], and the heritability several lung cancer susceptibility
smoke has some of the same car- of lung cancer has been estimated loci, including CHRNA3/5, TERT-
cinogenic constituents as tobacco as 18% [20]. Having a first-degree CLPTM1L, the HLA/MHC region,
smoke, such as selected polycy- relative with lung cancer increases RAD52, BRCA2, and CHEK2, with
clic aromatic hydrocarbons [12]. the risk of lung cancer by 1.25–1.5- extensive validations [24,25]. A list
Therefore, several studies have fold in never-smokers [21]. of major lung cancer susceptibility
been conducted to evaluate its High-penetrance genetic syn- loci for European descendants was
potential association with risk of dromes, such as Li–Fraumeni syn- reported in the most recent and larg-
lung cancer [13,14]. However, most drome and hereditary retinoblas- est GWAS analysis to date [25]. The
studies are limited by either poten- toma, are associated with increased loci identified so far accounted for
tial underreporting or sparse data risk of lung cancer [18]. In addition, about 12% of the familial relative risk
among heavy cannabis users, and high-penetrance germline mutations of lung cancer.
300
Fig. 5.1.1. Global distribution of estimated age-standardized (World) incidence rates (ASR) per 100 000 person-years for lung
cancer (A) in men and (B) in women, 2018.
Several large-scale GWAS anal- Somatic characteristics in never-smokers (Fig. 5.1.3B), and

yses conducted in Asian popula- of histological types KRAS mutations were the predomi-
SECTION 5
CHAPTER 5.1
tions have identified multiple Asian- nant mutations in adenocarcinomas
specific lung cancer susceptibility Comprehensive genomic charac- arising in patients in Europe and
loci, such as ROS1, along with sev- terizations were conducted by the North America, of which about 85%
eral loci in common with those found Cancer Genome Atlas Research were ever-smokers (Fig.  1.5.3A).
in European descendants, such as Network for lung adenocarcino- TP53 mutations occurred in 46%
TERT-CLPTM1L. A detailed list of ma and squamous cell carcinoma of adenocarcinomas [29] and in al-
lung cancer susceptibility loci in both [28,29]. Both tumour types showed most all squamous cell carcinomas,
European and Asian populations is a very high average tumour muta- along with a variety of activating mu-
included in a recent review [26]. tion burden of about 8–9 somatic tations, although none at very high
Data on the African American popu- mutations per megabase. In ade- frequencies [28]. Biallelic inactiva-
lation are currently limited to a single nocarcinoma, mutations in KRAS tion of TP53 and RB1 appears to be
study, which confirmed the asso- were mutually exclusive with those a universal feature of SCLC [30]. All
ciation of the CHRNA5 and TERT- in EGFR. EGFR mutations were the three of these types of lung tumours
CLPTM1 loci with lung cancer [27]. main mutations in adenocarcinoma have marked genomic complexity,

Fig. 5.1.2. Age-standardized (World) incidence rates per 100 000 person-years by calendar year in selected countries for lung
cancer (left) in men and (right) in women, circa 1978–2012.
including rearrangements and copy are classified by the most common subtype, which is less well differenti-
number variations. subtype present [31]. ated, is the presence or absence of
The mutation spectra shown in The adenocarcinoma in situ visible keratin on histological exami-
Fig. 5.1.3 are markedly different by subtype is characterized by lepidic nation. No other molecular features
histological type, suggesting that (scale-like) growth along existing al- have been described that separate
they may arise via very different veolar walls without underlying tissue these two common subtypes. The
molecular pathways. In addition, invasion. The papillary subtype has basaloid subtype has cells that are
spatial and temporal intratumour fibrovascular cores, which distinguish morphologically similar to those
heterogeneity in the processes of it from the micropapillary subtype. found in the basal layer of the large
genomic instability is an active new The acinar subtype is frequent and airways and that have a specific
area of research, with potential val- has gland formation as its hallmark mRNA expression profile [33]. The
ue as a prognostic predictor. The feature. These three subtypes have mutation spectrum for squamous cell
morphological and molecular fea- frequent EGFR mutations. The solid
carcinoma is shown in Fig. 5.1.3C.
tures of the main histological sub- with mucin subtype is poorly differen-
types are described below. tiated and is associated with KRAS Small cell lung carcinoma
mutations or translocations in ALK,
Adenocarcinoma SCLCs are aggressive carcinomas
ROS, RET, and NTRK. The recently
that originate from neuroendocrine
Adenocarcinomas have more mor- recognized micropapillary subtype
phological heterogeneity than other lacks fibrovascular cores and may cells in the bronchial epithelium. Only
types of lung cancer; a uniform ter- contain ALK or HER2 mutations. two SCLC subtypes are recognized:
minology was recently proposed Mucinous carcinomas, although not pure SCLCs and combined SCLCs.
and has been widely accepted an official subtype, are relatively rare Combined SCLCs have a non-SCLC
[31,32]. The new subtypes, along and have frequent KRAS mutations. (NSCLC) component that consists of
with their major morphological fea- at least 10% of the tumour [34].
tures and the presence of frequent Squamous cell carcinoma
gene mutations, are summarized Squamous cell carcinoma has three
Epigenetics of lung cancer
in Table 5.1.1 and illustrated in subtypes: keratinizing, non-keratiniz- The epigenetic landscape of lung
Fig. 5.1.4. However, most adeno- ing, and basaloid. The morphological cancer commences early during
carcinomas are composed of more difference between the keratinizing pathogenesis and consists of two
than one subtype, and the tumours subtype and the non-keratinizing major components: methylation and
302
Fig. 5.1.3. Mutation spectra by histological type of lung cancer, showing the percent- histone modifications (see Chapter
age of samples with a mutation detected by automated analysis. “Unknown” refers 3.8) [35].
to potentially druggable mutations and excludes tumour recessive genes including Global hypomethylation is a com-
TP53. (A) Mutation pattern of adenocarcinomas arising in patients in Europe and North
America, of which about 85% were ever-smokers. (B) Mutation pattern of adenocar-
mon feature of cancer. Smoking-
cinomas in Asian never-smokers. (C) Mutation pattern of squamous cell carcinomas. related hypomethylation measured
(D) Mutation pattern of small cell lung carcinomas. in pre-diagnostic blood samples
was shown to be associated with in-
creased risk of lung cancer, and the
A TP53 46
most consistently replicated change
KRAS 33
KEAP1 17 was in the AHRR gene [36]. DNA
EGFR 14 hypermethylation, mainly in the pro-
NF1 11
SETD2 9
moter region, is a major mechanism
RBM10 8 for the silencing of tumour suppres-
ARID1A 7 sor genes, although it may also oc-
MET 7
PIK3CA 7
cur in the body of the gene, where
SMARCA4 6 it may result in gene activation.
RB1 4 Several hundred genes are methyl-
CDKN2A 4
UNKNOWN 42 ated in lung cancers, and the best
studied and most frequently methyl-
0 5 10 15 20 25 30 35 40 45 50
ated genes are listed in Table 5.1.2.
Methylation results in inactivation
B EGFR 74
of one allele, and the other allele is
TP53 30 usually deleted.
ALK 6 In addition to methylation, many
HER2 4 covalent modifications can occur
KRAS 3 on the N-terminal tail that protrudes
ROS 1 from each of the four histone pro-
RET 1
teins. Histone modifications target
METTX14
many key tumour suppressor genes.
1
The major histone changes that
UNKNOWN 9
characterize NSCLC are listed in
0 10 20 30 40 50 60 70 80 Table 5.1.2.
Although most epigenetic studies
C TP53 81 of lung cancer focus on NSCLC, the
MLL2 20 epigenetics of SCLC has both simi-
PIK3CA 16 larities and differences with NSCLC.
CDKN2A 15 In particular, EZH2, a master regula-
NFE2L2 15 tor of transcription that affects DNA
KEAP1 12 methylation via upregulation of DNA
NOTCH1 8 methyltransferases, is upregulated
PTEN 8
in many cancer types, including
NOTCH 1 8
SCLC, where it plays a major role
RB1 7
in tumour progression and is asso-
UNKNOWN 22
ciated with poor prognosis. These
SECTION 5
CHAPTER 5.1
0 10 20 30 40 50 60 70 80 90 findings have led to widespread ef-
forts to therapeutically target EZH2.
D TP53 98
The genetic and epigenetic somatic
RB1 91
KIAA1211 26 alterations of lung cancer have re-
NOTCH GENES 25 cently been reviewed [37].
PRE4DIP 16
CREBBP 15
COLO22A1 14 Lung cancer in
TP73
EP300
13
13
never-smokers
RBL1 6 Lung cancer in never-smokers is
KIT 6 a specific disease entity, because
CDKN2A 5
UNKNOWN 11
there are significant differences in eti-
ology and clinical characteristics be-
0 20 40 60 80 100 120
tween lung cancer in never-smokers

Fig. 5.1.4. Morphological features of adenocarcinoma subtypes: (A) adenocarcinoma in situ, (B) acinar, (C) solid with mucin,
(D) papillary, (E) micropapillary, and (F) mucinous.
A B C
D E F
Table 5.1.1. Lung adenocarcinoma subtypes, growth patterns, and frequent gene mutations
Subtype a Major growth features Frequent gene mutationsb
Adenocarcinoma in situ Lepidic growth EGFR
Acinar Gland-forming, occasional cribriform pattern EGFR
Solid with mucin Predominantly solid with focal mucin production KRAS, ALK
Papillary Papillary invasive, with fibrovascular cores EGFR
Micropapillary Small papillary-like structures without fibrovascular cores ALK, HER2
Mucinous Predominantly lepidic with abundant intracellular and occasionally KRAS

extracellular mucin
a
Most adenocarcinomas consist of mixtures of more than one subtype. The major component should be identified and its percentage estimated. The other
subtypes present should also be identified.
b
Certain mutations are more common in certain subtypes, but there is considerable heterogeneity. Very little information is available for BRAF and ROS1
mutations and their associations with adenocarcinoma subtypes.
versus ever-smokers. For example, mutation landscape, burden, and smokers. However, in populations
adenocarcinomas are more preva- affected genes; TP53 is the most where the prevalence of smoking is
lent in never-smoker patients with extensively documented gene [39]. low, an increasing proportion of lung
lung cancer [38]. In addition, lung Other features that distinguish lung cancer occurs in never-smokers and
cancers in never-smokers have cancer in never-smokers and ever- former smokers.
different somatic characteristics smokers, such as methylation pat-
(Fig. 5.1.3). Most notably, never- terns, have also been reported [39]. Screening
smoker patients with lung cancer The National Lung Screening Trial
have a lower prevalence of KRAS in the USA reported that the low-
mutations and a higher prevalence
Prevention and mortality dose computed tomography (LDCT)
of EGFR mutations and show longer reduction screening reduced the lung cancer
survival after treatment with EGFR Currently, the best hopes for reduc- mortality by 20% in former and cur-
inhibitors than do ever-smokers. ing lung cancer mortality are pre- rent smokers who were eligible to
Overall, there are extensive differ- venting smoking through effective be screened, based on age (age
ences between smokers and never- tobacco control and promoting suc- 55 years to 74 or 80 years) and his-
smokers with regard to the tumour cessful smoking cessation in current tory of tobacco smoking (at least
304
Table 5.1.2. Frequent abnormalities of genes involved in the epigenetic regulation of lung cancers
Frequently methylated genes Mutations or dysregulation of epigenetic regulators
SHOX2 Histone acetyltransferase EP300
TCF21 Histone acetyltransferase CREBBP
APC Histone deacetylase HDAC4
EPBH1K3 Histone deacetylase HDAC9
PYCARD Lysine methyltransferase KMT2A–D
FHIT Lysine methyltransferase PRDM9
TSLC1 Lysine methyltransferase SETD2
RAR Lysine methyltransferase NSD1
CDH1 Lysine methyltransferase EZH2
RASSF1A Lysine demethylase KDM5A–C
CDKN2A DNA methyltransferase DNMT1, 3A, and 3B
DAPK H3K9 methyltransferase SETDB9
CDH13 H3K36 demethylase KDM2A
PTEN SWI/SNF complex SMARCA4
RUNX3 SWI/SNF complex ARID1A
30 pack-years of smoking, or have are now considering implement- approve LDCT screening for insur-
smoked within the past 15 years). ing LDCT lung cancer screening at ance coverage.
This presented an appealing comple- the population level, and the United Currently, in the USA most of the
mentary strategy for reducing lung States Preventive Services Task screening recommendations provid-
cancer mortality through detection of ed by health agencies are derived
Force has issued the Grade B rec-
early-stage lung cancer, which is still from the National Lung Screening
potentially curable by surgical resec- ommendation for LDCT screening.
Trial eligibility criteria based on age
tion [40]. Since 2015, several major health and history of tobacco smoking, and
As a result, many public health insurance programmes in the USA, a recent National Comprehensive
agencies and medical institutions including Medicare, have started to Cancer Network Category 2 recom-
mendation also included family his-
tory and non-tobacco risk factors to
Fig. 5.1.5. A man undergoing lung cancer screening at the University of Connecticut
Health Center, USA. improve the screening criteria [41].
However, studies have shown that
applying individual risk probabili-
ty-based screening criteria could
prevent more lung cancer deaths
SECTION 5
CHAPTER 5.1
and reduce the number needed to
screen to prevent one lung cancer
death [42]. Although substantial ef-
forts have been made to establish
lung cancer risk prediction models
based on personal health and ex-
posure history [43], lung cancer re-
searchers are now working towards
integrating individual molecular pro-
files to improve risk prediction.
Biomarkers
The development of biomarkers for
early detection of lung cancer is an

active research area, which encom- reported biomarker panels was re- better identify individuals who are at
passes a wide range of biomarker cently published [46]. high risk of lung cancer and should
research, including markers and me- In terms of epigenetic mark- be recommended for LDCT screen-
tabolites that could be found in the ers, in addition to methylation and ing. To yield an optimal predictive
various biological fluids, particularly histone modification as mentioned performance for early detection of
circulating blood, urine, or sputum. above, microRNAs and long non- lung cancer, one can consider mul-
The main types of circulating bio- coding RNAs are also potential epi- tiple layers of data, including epide-
markers are protein-based markers, genomic biomarkers. In particular, miological and clinical information
metabolites, autoantibodies from hu- several previous studies have shown and an individual’s molecular pro-
moral immune response, epigenetic a promising predictive performance files; this aligns with the concept of
markers, and circulating tumour DNA. of multi-microRNA panels [47], al- precision medicine (Fig. 5.1.6) [52].
Although most of the biomarkers though the sample sizes tend to be It is anticipated that biomarkers
have failed to be replicated in inde- limited and external validation in in- may also help to differentiate ma-
pendent studies, several promising dependent studies is still required. lignant nodules from benign ones.
biomarkers have been established In addition to blood-based bio- The challenge is to establish a pan-
across multiple prospective cohort markers, another type of biomarker el that would be applicable in the
studies. For example, plasma level of for early detection of lung cancer clinical setting and remain cost-ef-
pro-surfactant protein B was shown focuses on the gene expression fective for the health-care system.
to be an independent predictor of profile of the airway epithelium,
lung cancer risk based on a pan-Ca- based on the theory of field of in- Nodule malignancy
nadian screening programme and jury and field cancerization [48,49]. For individuals who undergo LDCT
the Carotene and Retinol Efficacy Finally, given the known association screening, about 15–20% of chest
Trial, after adjusting for demograph- between chronic obstructive pulmo- scans detect non-calcified pulmo-
ic factors and lung cancer risk fac- nary disease and risk of lung cancer, nary nodules. However, the National
tors [44]. It has become clear that a previous studies have evaluated the Lung Screening Trial reported that
panel of multiple biomarkers, rather added predictive performance of only 1 in 20 nodules detected by
than any single marker, would be lung function [50,51]. LDCT screening are actually lung
needed to improve risk prediction Biomarker research for early de- cancers [40]. To address this issue,
[45]. A succinct review of various tection of lung cancer can help to several clinical probability models
Fig. 5.1.6. A concept schema of biomarker integration for precision medicine.
Information Commons
Exposome
Genome
Transcriptome
Biocomputing
Integration
Informed Epigenome Observational

mechanistic
Microbiome clinical studies
studies
Metabolome
Clinical information
Epidemiological data
Knowledge Network
Target
identification Treatment
New molecular taxonomic classification
of patients by biomarkers
Molecular Guide Guide Improve Health

Diagnosis mechanisms Diagnosis outcomes
Biomedical research Prevention research Clinical medicine
306
were proposed to improve the as- ture extraction, the analytical meth- sis, perhaps because the NSCLC
sessment of nodules, and use of the ods vary greatly across studies, and component is resistant to cytotoxic
Lung CT Screening Reporting and proper validation is still required for therapies. Adenocarcinoma in situ
Data System (Lung-RADS) classi- robust reproducibility. Therefore, it usually has an excellent prognosis
fication of the American College of is currently considered premature if it is completely resected, even if
Radiology was shown to substantial- to implement radiomics as part of small foci of invasion are present
ly decrease the false-positive rate, the routine diagnostic process. (microinvasive carcinomas).
with a moderate reduction in sensi- Despite the growing number of
tivity [53]. However, currently there mutations that continue to be iden-
Prognosis and targeted
is still a wide range of clinical proto- tified, only a few somatic mutations
cols for how patients with pulmonary
treatment can be used for targeted therapy,
nodules detected on LDCT screen- Lung cancer survival remains dis- such as EGFR mutation, ROS1 fu-
ing are managed, and the diagnostic mal, with 5-year survival rates of sion, and ALK translocation; more
evaluation of suspicious abnormali- only 10–20% in most parts of the recently, immunotherapy agents
ties can range from watchful waiting world [55,56]. The stage at diagnosis have been developed that target
and monitoring to needle biopsy and is a major determinant of lung can- programmed cell death 1 (PD-1) pro-
pulmonary resection. cer prognosis; 5-year survival rates tein. A range of other targeted and
In response to the need to dif- range from 50–70% for diagnosis immunotherapy trials are currently in
ferentiate between benign and at stage I to 1–5% for diagnosis at progress, with the hope of improving
malignant nodules, radiomics has stage IV, because surgical resec- treatment response based on the
emerged as a field of study. Radio tion at an early stage is still the most principle of precision medicine. A
mics is the analysis of high-dimen- effective treatment [55]. However, complete review has been provided
sional imaging data, focusing on fewer than 20% of patients are diag- by the International Association for
the extraction of quantitative vari- nosed at stage I, and most are diag- the Study of Lung Cancer [57].
ables from radiographic features for nosed at stage IIIB or IV [55]; hence, In summary, lung cancer com-
subsequent agnostic data mining early detection is important. prises very different types and sub-
[54]. This field has shown promise The clinical outcome varies by types, which affect the strategies
to better differentiate nodules with histological type. SCLC is the most for prevention, early detection, di-
malignant potential. However, there aggressive type, and combined agnosis, and clinical management.
is no standardized process of fea- SCLC may have a worse progno-
SECTION 5
CHAPTER 5.1

References
1. Travis WD, Brambilla E, Nicholson AG, 9. GBD 2015 Tobacco Collaborators (2017). 18. Thun MJ, Henley SJ, Travis WD (2018).
Yatabe Y, Austin JHM, Beasley MB, et al.; Smoking prevalence and attributable dis- Lung cancer. In: Thun MJ, Linet MS,
WHO Panel (2015). The 2015 World Health ease burden in 195 countries and territo- Cerhan JR, Haiman CA, Schottenfeld D,
Organization classification of lung tumors: ries, 1990–2015: a systematic analysis from editors. Cancer epidemiology and pre-
impact of genetic, clinical and radiologic the Global Burden of Disease Study 2015. vention. 4th ed. New York (NY), USA:
advances since the 2004 classification. J Lancet. 389(10082):1885–906. https:// Oxford University Press; pp. 519–52.
Thorac Oncol. 10(9):1243–60. https://doi. doi.org/10.1016/S0140-6736(17)30819-X
19. Penney KL, Michailidou K, Carere DA,
org/10.1097/JTO.0 0 0 0 0 0 0 0 0 0 0 0 0 630 PMID:28390697
Zhang C, Pierce B, Lindstrom S, et al.
PMID:26291008 10. Shields PG, Berman M, Brasky TM, (2018). Genetic epidemiology of can-
Freudenheim JL, Mathe E, McElroy JP, cer. In: Thun MJ, Linet MS, Cerhan JR,
2. Bray F, Ferlay J, Soerjomataram I, Siegel
et al. (2017). A review of pulmonary toxic- Haiman CA, Schottenfeld D, editors.
RL, Torre LA, Jemal A (2018). Global Cancer epidemiology and prevention.
ity of electronic cigarettes in the context of
cancer statistics 2018: GLOBOCAN es- 4th ed. New York (NY), USA: Oxford
smoking: a focus on inflammation. Cancer
timates of incidence and mortality world- University Press; pp. 53–76.
Epidemiol Biomarkers Prev. 26(8):1175–91.
wide for 36 cancers in 185 countries. CA https://doi.org/10.1158/1055-9965.EPI-17-
Cancer J Clin. 68(6):394–424. https://doi. 20. Mucci LA, Hjelmborg JB, Harris JR, Czene
0358 PMID:28642230
org/10.3322/caac.21492 PMID:30207593 K, Havelick DJ, Scheike T, et al.; Nordic
11. Dautzenberg B, Garelik D (2017). Patients Twin Study of Cancer (NorTwinCan)
3. Ferlay J, Colombet M, Bray F (2018). with lung cancer: are electronic cigarettes Collaboration (2016). Familial risk and
Cancer incidence in five continents, harmful or useful? Lung Cancer. 105: heritability of cancer among twins in
CI5plus. IARC CancerBase No. 9 [Internet]. 42–8. https://doi.org/10.1016/j.lungcan. Nordic countries. JAMA. 315(1):68–76.
Lyon, France: International Agency for 2016.05.011 PMID:27241679 https://doi.org/10.1001/jama.2015.17703
Research on Cancer. Available from: http:// PMID:26746459
ci5.iarc.fr. 12. Moir D, Rickert WS, Levasseur G, Larose
Y, Maertens R, White P, et al. (2008). A 21. Coté ML, Liu M, Bonassi S, Neri M,
4. Parkin DM, Boyd L, Walker LC (2011). 16. comparison of mainstream and side- Schwartz AG, Christiani DC, et al. (2012).
The fraction of cancer attributable to life- stream marijuana and tobacco cigarette Increased risk of lung cancer in individuals
smoke produced under two machine with a family history of the disease: a pooled
style and environmental factors in the UK
smoking conditions. Chem Res Toxicol. analysis from the International Lung Cancer
in 2010. Br J Cancer. 105(Suppl 2):S77–
Consortium. Eur J Cancer. 48(13):1957–68.
81. https://doi.org/10.1038/bjc.2011.489 21(2):494–502. https://doi.org/10.1021/
PMID:22158327 tx700275p PMID:18062674
PMID:22436981
5. U.S. Department of Health and Human 13. Callaghan RC, Allebeck P, Sidorchuk A
22. Gazdar A, Robinson L, Oliver D, Xing C,
Services (2014). The health consequences (2013). Marijuana use and risk of lung
Travis WD, Soh J, et al. (2014). Hereditary
of smoking – 50 years of progress: a re- cancer: a 40-year cohort study. Cancer
lung cancer syndrome targets never smok-
port of the Surgeon General. Atlanta (GA), Causes Control. 24(10):1811–20. https://
ers with germline EGFR gene T790M muta-
USA: U.S. Department of Health and d o i.o r g /10.10 07/s10 5 52- 013 - 0 2 5 9 - 0
tions. J Thorac Oncol. 9(4):456–63. https://
PMID:23846283
Human Services, Centers for Disease doi.org/10.1097/JTO.0000000000000130
Control and Prevention, National Center 14. Zhang LR, Morgenstern H, Greenland PMID:24736066
for Chronic Disease Prevention and Health S, Chang SC, Lazarus P, Teare MD,
23. Yamamoto H, Higasa K, Sakaguchi M,
Promotion, Office on Smoking and Health. et al. (2015). Cannabis smoking and
Shien K, Soh J, Ichimura K, et al. (2014).
Available from: https://www.ncbi.nlm.nih. lung cancer risk: pooled analysis in the Novel germline mutation in the transmem-
gov/books/NBK179276/. International Lung Cancer Consortium. brane domain of HER2 in familial lung
Int J Cancer. 136(4):894–903. https://doi. adenocarcinomas. J Natl Cancer Inst.
6. Freedman ND, Abnet CC, Caporaso NE, org/10.1002/ijc.29036 PMID:24947688
Fraumeni JF Jr, Murphy G, Hartge P, et al. 106(1):djt338. https://doi.org/10.1093/jnci/
15. Skillrud DM, Offord KP, Miller RD (1986). djt338 PMID:24317180
(2016). Impact of changing US cigarette
smoking patterns on incident cancer: risks Higher risk of lung cancer in chronic 24. Hung RJ, McKay JD, Gaborieau V,
of 20 smoking-related cancers among the obstructive pulmonary disease. A pro- Boffetta P, Hashibe M, Zaridze D, et al.
women and men of the NIH-AARP cohort. spective, matched, controlled study. (2008). A susceptibility locus for lung
Int J Epidemiol. 45(3):846–56. https://doi. Ann Intern Med. 105(4):503–7. https:// cancer maps to nicotinic acetylcholine
org/10.1093/ije/dyv175 PMID:26411408 doi.org/10.7326/0003- 4819 -105 - 4-503 receptor subunit genes on 15q25. Nature.
7. Song MA, Benowitz NL, Berman M, nature06885 PMID:18385738
16. Punturieri A, Szabo E, Croxton TL,
Brasky TM, Cummings KM, Hatsukami
Shapiro SD, Dubinett SM (2009). Lung 25. McKay JD, Hung RJ, Han Y, Zong X,
DK, et al. (2017). Cigarette filter ventila-
cancer and chronic obstructive pulmo- Carreras-Torres R, Christiani DC, et al.;
tion and its relationship to increasing
nary disease: needs and opportunities for SpiroMeta Consortium (2017). Large-
rates of lung adenocarcinoma. J Natl
integrated research. J Natl Cancer Inst. scale association analysis identifies new
Cancer Inst. 109(12):djx075. https://doi.
101(8):554–9. https://doi.org/10.1093/jnci/ lung cancer susceptibility loci and het-
org/10.1093/jnci/djx075 PMID:28525914 djp023 PMID:19351920 erogeneity in genetic susceptibility across
8. IARC (2019). List of classifications by histological subtypes. Nat Genet. 49(7):
17. Brenner DR, Boffetta P, Duell EJ,
cancer sites with sufficient or limited 1126–32. https://doi.org/10.1038/ng.3892
Bickeböller H, Rosenberger A, Muscat
evidence in humans, Volumes 1 to 124. PMID:28604730
JE, et al.; International Lung Cancer
Lyon, France: International Agency for Consortium (2012). Previous lung diseases 26. Bossé Y, Amos CI (2018). A decade of
Research on Cancer. Available from: and lung cancer risk: a pooled analysis from GWAS results in lung cancer. Cancer
https://monographs.iarc.fr/wp-content / the International Lung Cancer Consortium. Epidemiol Biomarkers Prev. 27(4):363–79.
uploads/2019/07/Classif ic ations _ by _ Am J Epidemiol. 176(7):573–85. https://doi. https://doi.org/10.1158/1055-9965.EPI-16-
cancer_site.pdf. org/10.1093/aje/kws151 PMID:22986146 0794 PMID:28615365
308
27. Zanetti KA, Wang Z, Aldrich M, Amos 37. Swanton C, Govindan R (2016). Clinical 48. Perez-Rogers JF, Gerrein J, Anderlind C,
CI, Blot WJ, Bowman ED, et al. (2016). implications of genomic discoveries in lung Liu G, Zhang S, Alekseyev Y, et al.; AEGIS
Genome-wide association study confirms cancer. N Engl J Med. 374(19):1864–73. Study Team (2017). Shared gene expres-
lung cancer susceptibility loci on chro- https://doi.org/10.1056/NEJMra1504688 sion alterations in nasal and bronchial ep-
mosomes 5p15 and 15q25 in an African- PMID:27168435 ithelium for lung cancer detection. J Natl
American population. Lung Cancer. 98: Cancer Inst. 109(7):djw327. https://doi.
38. Sun S, Schiller JH, Gazdar AF (2007).
33–42. https://doi.org/10.1016/j.lungcan. Lung cancer in never smokers – a dif- org/10.1093/jnci/djw327 PMID:28376173
2016.05.008 PMID:27393504 ferent disease. Nat Rev Cancer. 7(10): 49. Billatos E, Vick JL, Lenburg ME, Spira
778–90. https://doi.org/10.1038/nrc2190 AE (2018). The airway transcriptome as a
28. Cancer Genome Atlas Research Network
PMID:17882278
(2012). Comprehensive genomic charac- biomarker for early lung cancer detection.
terization of squamous cell lung cancers. 39. Govindan R, Ding L, Griffith M, Clin Cancer Res. 24(13):2984–92. https://
Nature. 489(7417):519–25. https://doi. Subramanian J, Dees ND, Kanchi KL, et doi.org/10.1158/1078-0432.CCR-16-3187
org/10.1038/nature11404 PMID:22960745 al. (2012). Genomic landscape of non- PMID:29463557
small cell lung cancer in smokers and
29. Cancer Genome Atlas Research Network never-smokers. Cell. 150(6):1121–34. 50. Tammemagi MC, Lam SC, McWilliams
(2014). Comprehensive molecular pro- https://doi.org/10.1016/j.cell.2012.08.024 AM, Sin DD (2011). Incremental value
filing of lung adenocarcinoma. Nature. PMID:22980976 of pulmonary function and sputum DNA
511(7511):543–50. https://doi.org/10.1038/ image cytometry in lung cancer risk pre-
40. Aberle DR, Adams AM, Berg CD, Black diction. Cancer Prev Res (Phila). 4(4):
nature13385 PMID:25079552
WC, Clapp JD, Fagerstrom RM, et al.; 552–61. https://doi.org/10.1158/1940-6207.
30. Gazdar AF, Bunn PA, Minna JD (2017). National Lung Screening Trial Research CAPR-10-0183 PMID:21411501
Small-cell lung cancer: what we know, Team (2011). Reduced lung-cancer mor-
tality with low-dose computed tomographic 51. Muller DC, Johansson M, Brennan P (2017).
what we need to know and the path for-
screening. N Engl J Med. 365(5):395–409. Lung cancer risk prediction model incor-
ward. Nat Rev Cancer. 17(12):725–37.
https://doi.org/10.1056/NEJMoa1102873 porating lung function: development and
h t t p s : //d o i . o r g /10 .10 3 8 / n r c . 2 017. 8 7
PMID:21714641 validation in the UK Biobank Prospective
PMID:29077690
Cohort Study. J Clin Oncol. 35(8):861–9.
41. Wood DE, Kazerooni EA, Baum SL, Eapen
31. Travis WD, Brambilla E, Noguchi M, https://doi.org/10.1200/JCO.2016.69.2467
GA, Ettinger DS, Hou L, et al. (2018). Lung
Nicholson AG, Geisinger K, Yatabe Y, PMID:28095156
cancer screening, version 3.2018, NCCN
et al. (2011). International Association Clinical Practice Guidelines in Oncology. 52. Vargas AJ, Harris CC (2016). Biomarker
for the Study of Lung Cancer/American J Natl Compr Canc Netw. 16(4):412–41. development in the precision medicine
Thoracic Society/European Respiratory https://doi.org/10.6004/jnccn.2018.0020 era: lung cancer as a case study. Nat
Society international multidisciplinary PMID:29632061 Rev Cancer. 16(8):525–37. https://doi.
classification of lung adenocarcinoma. org/10.1038/nrc.2016.56 PMID:27388699
42. Katki HA, Kovalchik SA, Berg CD,
J Thorac Oncol. 6(2):244–285. https://
Cheung LC, Chaturvedi AK (2016).
doi.org/10.1097/JTO.0b013e318206a221 53. Oudkerk M, Devaraj A, Vliegenthart R,
Development and validation of risk models
PMID:21252716 Henzler T, Prosch H, Heussel CP, et al.
to select ever-smokers for CT lung can-
cer screening. JAMA. 315(21):2300–11. (2017). European position statement on
32. Austin JH, Garg K, Aberle D, Yankelevitz D, lung cancer screening. Lancet Oncol.
https://doi.org/10.1001/jama.2016.6255
Kuriyama K, Lee HJ, et al. (2013). Radiologic 18(12):e754–66. https://doi.org/10.1016/
PMID:27179989
implications of the 2011 classification of S1470-2045(17)30861-6 PMID:29208441
adenocarcinoma of the lung. Radiology. 43. Tammemagi MC, Lam S (2014). Screening
266(1):62–71. https://doi.org/10.1148/radiol. for lung cancer using low dose computed 54. Parmar C, Grossmann P, Bussink J,
12120240 PMID:23070271 tomography. BMJ. 348:g2253. https://doi. Lambin P, Aerts HJWL (2015). Machine
org/10.1136/bmj.g2253 PMID:24865600 learning methods for quantitative radi-
33. Brambilla C, Laffaire J, Lantuejoul S, omic biomarkers. Sci Rep. 5(1):13087.
Moro-Sibilot D, Mignotte H, Arbib F, et 44. Sin DD, Tammemagi CM, Lam S, Barnett h t t p s : / / d o i . o r g / 1 0 .1 0 3 8 / s r e p13 0 8 7
al. (2014). Lung squamous cell carci- MJ, Duan X, Tam A, et al. (2013). Pro- PMID:26278466
nomas with basaloid histology repre- surfactant protein B as a biomarker for
sent a specific molecular entity. Clin lung cancer prediction. J Clin Oncol. 55. Noone AM, Howlader N, Krapcho M,
31(36):4536–43. https://doi.org/10.1200/ Miller D, Brest A, Yu M, et al., editors
Cancer Res. 20(22):5777–86. https://doi.
JCO.2013.50.6105 PMID:24248694 (2018). SEER cancer statistics review,
org/10.1158/1078 - 0 432.CCR-14 - 0 459
1975–2015. Bethesda (MD), USA:
PMID:25189482 45. Guida F, Sun N, Bantis LE, Muller DC, Li
National Cancer Institute. Available from:
P, Taguchi A, et al.; Integrative Analysis of
34. Gazdar AF (2018). Morphologic and Lung Cancer Etiology and Risk (INTEGRAL) https://seer.cancer.gov/csr/1975_2015/.
other forms of heterogeneity in small Consortium for Early Detection of Lung 56. Allemani C, Weir HK, Carreira H,
cell lung cancer: what can we learn from Cancer (2018). Assessment of lung can-
SECTION 5
CHAPTER 5.1
Harewood R, Spika D, Wang XS, et al.;
them? J Thorac Oncol. 13(2):148–50. cer risk on the basis of a biomarker panel CONCORD Working Group (2015). Global
https://doi.org/10.1016/j.jtho.2017.11.004 of circulating proteins. JAMA Oncol.
surveillance of cancer survival 1995–2009:
PMID:29425612 4(10):e182078. https://doi.org/10.1001/jama
analysis of individual data for 25 676 887
oncol.2018.2078 PMID:30003238
35. Langevin SM, Kelsey KT (2017). Clinical patients from 279 population-based reg-
epigenetics of lung cancer. In: Laurence J, 46. Hanash SM, Ostrin EJ, Fahrmann JF istries in 67 countries (CONCORD-2).
Van Beusekom M, editors. Translating epi- (2018). Blood based biomarkers beyond Lancet. 385(9972):977–1010. https://doi.
genetics to the clinic. London, UK: Elsevier; genomics for lung cancer screening. org /10.1016/S0140 - 673 6(14) 6203 8 - 9
pp. 97–133. https://doi.org/10.1016/B978- Transl Lung Cancer Res. 7(3):327–35. PMID:25467588
https://doi.org/10.21037/tlcr.2018.05.13
0-12-800802-7.00005-8 57. Soo RA, Stone ECA, Cummings KM, Jett
PMID:30050770
36. Bojesen SE, Timpson N, Relton C, Davey JR, Field JK, Groen HJM, et al. (2017).
47. Moretti F, D’Antona P, Finardi E, Barbetta Scientific advances in thoracic oncology
Smith G, Nordestgaard BG (2017). AHRR M, Dominioni L, Poli A, et al. (2017). 2016. J Thorac Oncol. 12(8):1183–209.
(cg05575921) hypomethylation marks Systematic review and critique of circu- https://doi.org/10.1016/j.jtho.2017.05.019
smoking behaviour, morbidity and mor- lating miRNAs as biomarkers of stage I-II PMID:28579481
tality. Thorax. 72(7):646–53. https:// non-small cell lung cancer. Oncotarget.
doi.org/10.1136/thoraxjnl-2016 -208789 8(55):94980–96. https://doi.org/10.18632/
PMID:28100713 oncotarget.21739 PMID:29212284

5.2
Head and neck cancers

New etiological insights
Laia Alemany Vilches Devasena Anantharaman (reviewer)

Paul Brennan (reviewer)
C. René Leemans (reviewer)
Head and neck cancers originate ynx (upper part), oropharynx (middle
SUMMARY from squamous cells located in the part, including the soft palate, uvula,
mucosal epithelium inside the head the base of the tongue, the tonsils,
●● Worldwide, head and neck can- and neck. They can also begin in tonsillar pillars, and oropharyngeal
cer is the seventh most com- the salivary glands, but cancers of wall), and hypopharynx (lower part);
mon cancer overall (the fifth the salivary glands are relatively (iii) larynx: located below the pharynx,
most common in men and the uncommon [1]. including the supraglottic and infra-
12th most common in women), Head and neck cancers are fur- glottic areas, with the vocal cords in
accounting for an estimated ther classified by the anatomical the middle; (iv) nasal cavity and para-
888 000 new cases in 2018. area in which they arise (Fig. 5.2.1): nasal sinuses; and (v) salivary glands.
(i) oral cavity: lips, front two thirds of Within these major anatomical
●● In the past 15 years, strong
the tongue, hard palate, mucosa in- areas, the head and neck can be fur-
evidence has accumulated that
side the cheeks, gums, and floor of ther subdivided into at least 14 sub-
infection with certain human
the mouth; (ii) pharynx: nasophar- sites, according to the International
papillomaviruses (HPVs) is eti-
ologically involved in a subset
of head and neck cancers, par- Fig. 5.2.1. Major anatomical areas within the head and neck.
ticularly oropharyngeal cancer.
●● HPV-related oropharyngeal
cancers differ from those that
are non-HPV-related, in terms
of epidemiological, clinical, and
molecular characteristics. HPV-
related cases of oropharyngeal
cancer have better survival than
non-HPV-related cases.
●● The main carcinogenic process
in HPV-related head and neck
cancers is through the action
of viral oncoproteins: E6 affects
p53 and E7 affects retinoblasto-
ma, disrupting these pathways.
●● HPV vaccination is a potential
tool for prophylaxis of HPV-
related head and neck cancers.
There are promising new po-
tential screening and monitor-
ing biomarkers, such as HPV16
E6 serology.
310
Statistical Classification of Diseases male-to-female incidence ratio is

and Related Health Problems, 10th 2:1 (Fig. 5.2.3).
revision (ICD-10). These numerous
FUNDAMENTALS
locations give rise to tumours that Pharyngeal cancer
■■ Most head and neck cancers
exhibit heterogeneous pathology. Cancers of the pharynx (nasophar- are squamous cell carcinomas
ynx, oropharynx, and hypopharynx)
of the upper aerodigestive
Epidemiology together accounted for an esti-
tract: predominantly cancers
mated 302 000 new cancer cases
Worldwide, head and neck cancer of the oral cavity, pharynx,
worldwide in 2018, of which about
is the seventh most common can- 40% were nasopharyngeal cancer, and larynx. Other tumours
cer overall (the fifth most common in about 30% were oropharyngeal that occur in this anatomical
men and the 12th most common in cancer, and about 30% were hy- area, such as brain cancer,
women), accounting for an estimat- popharyngeal cancer [2]. thyroid cancer, and some
ed 888 000 new cases in 2018 [2]. Globally, age-standardized in- melanomas, are not usually
The male-to-female incidence ratio cidence rates for both sexes com- included in this category.
is 3:1, and about 70% of new cases bined are 1.5 per 100 000 for na-
occur in low- and middle-income ■■ The male-to-female incidence
sopharyngeal cancer and 2.0 per
countries. In 2018, there were an es- ratio for head and neck
100 000 for other pharyngeal can-
timated 453 000 deaths from head cers (Fig. 5.2.2). The burden of na- cancers is 3:1. These tumours
and neck cancer globally. About sopharyngeal cancer falls predomi- are typically caused by
75% of those deaths occurred in nantly on low- and middle-income tobacco smoking, alone or
low- and middle-income countries. countries (93% of the worldwide bur- in combination with alcohol
den), such as countries in East Asia, consumption. In some
Oral cavity cancer
where almost 50% of the global countries, such as India, oral
Almost 50% of head and neck can- cases of nasopharyngeal cancer oc- cavity cancer is mainly caused
cers arise in the oral cavity. In 2018, cur. For other pharyngeal cancers, by betel quid chewing.
there were an estimated 355 000 the difference is smaller: 60% of
new cases and 177 000 deaths the cases occur in low- and middle- ■■ Infection with human
worldwide for oral cavity cancer [2]. income countries. The male-to-fe- papillomaviruses was initially
Of the cancers of anatomical ar- male incidence ratio is 3:1 for naso- recognized as causing cancers
eas in the head and neck, cancer pharyngeal cancer and 5:1 for other of the oropharynx and the
of the oral cavity has the highest pharyngeal cancers (Fig. 5.2.3). In base of the tongue.
age-standardized incidence rate 2018, there were an estimated
globally for both sexes combined: 4 ■■ Nasopharyngeal cancers are
73 000 deaths from nasopharyngeal
per 100 000 (Fig. 5.2.2). The high- common in parts of South-
cancer and 86 000 deaths from oth-
est age-standardized incidence er pharyngeal cancers. East Asia and North Africa;
rates (per 100 000) are observed in In 1970–2007, the age-standard- their etiology worldwide
Papua New Guinea (20.4), Pakistan ized incidence rates of nasopharyn- involves Epstein–Barr virus,
(12.2), Bangladesh (9.5), India (9.1), geal cancer decreased significantly wood dust, formaldehyde, and
Sri Lanka (7.6), and Hungary (7.5). in South and East Asia, North genetic factors.
The burden in South and Central America, and the Nordic countries.
■■ Early-stage tumours of the
Asia (160 000) is more than one The declines in the age-standard-
third of the global burden of oral upper aerodigestive tract
ized mortality rates in 1970–2013
can be cured; for late-stage
CHAPTER 5.2
SECTION 5
cavity cancer. In 2018, India was were even more remarkable and
the country with the highest burden, extensive. Decreasing trends in in- disease, prognosis is poor.
with 120 000 new cases. cidence are probably due to tobac-
Trends in incidence rates of co control, changes in dietary pat-
oral cavity cancer were evaluated terns, and economic development.
for 23 countries across four conti- Declines in mortality rates are the Laryngeal cancer
nents in 1983–2002. In men, inci- results of advances in diagnostic Laryngeal cancer is the 16th most
dence rates increased significantly and radiotherapy techniques, as common cancer in men and is rare
in Denmark, the Netherlands, the well as decreased incidence rates in women; the male-to-female inci-
United Kingdom, Brazil, and India. [4]. In 1983–2002, incidence rates dence ratio is 7:1 (Fig. 5.2.3). In 2018,
In women, the burden is much low- of oropharyngeal cancer increased there were an estimated 177 000
er, and incidence rates of oral cav- significantly, predominantly in high- new cases of laryngeal cancer world-
ity cancer increased significantly income countries and at younger wide [2]. About 66% of the new cases
only in European countries [3]. The ages [3]. occurred in low- and middle-income
Chapter 5.2 • Head and neck cancers 311

Fig. 5.2.2. Global distribution of estimated age-standardized (World) incidence rates (ASR) per 100 000 person-years for head
and neck cancers in both sexes, 2018: (A) lip and oral cavity, (B) larynx, (C) nasopharynx, (D) oropharynx, and (E) hypopharynx.
A Lip and oral cavity
B Larynx
C Nasopharynx
312
Fig. 5.2.2. Global distribution of estimated age-standardized (World) incidence rates (ASR) per 100 000 person-years for head
and neck cancers in both sexes, 2018: (A) lip and oral cavity, (B) larynx, (C) nasopharynx, (D) oropharynx, and (E) hypopharynx.
D Oropharynx
E Hypopharynx
countries, and about half of the cases Etiology cause of their multifactorial etiology,
occurred in Asia. which is largely attributed to tobacco
CHAPTER 5.2
SECTION 5
Age-standardized incidence rates Human papillomaviruses use and alcohol consumption [6,8].
tend to be higher in the Caribbean In the past 15 years, strong evidence The mere presence of HPV DNA
and in some countries in eastern has accumulated that infection with is not sufficient to prove viral causa-
Europe (Fig. 5.2.2). In 2018, laryn- certain human papillomaviruses tion, because it may reflect only a tran-
geal cancer accounted for an es- (HPVs) is etiologically involved in a sient infection unrelated to the carci-
timated 95 000 deaths worldwide. subset of head and neck cancers, nogenic process (see Chapter 2.2)
In some countries, such as in most particularly oropharyngeal cancer [9,10]. Most early studies and meta-
of Europe, a declining trend in inci- [6]. Although almost all squamous analyses assessing the quantitative
dence and mortality was observed cell carcinomas of the cervix are contribution of HPVs in head and neck
over the past few decades, after considered to be HPV-driven [7], cancer used the presence and detec-
favourable changes in tobacco use quantitative assessment of the etio- tion of HPV DNA in the tumour as the
and, mostly for Mediterranean coun- logical involvement of HPVs in head sole criterion. To accurately classify a
tries, alcohol consumption [5]. and neck cancer is challenging be- tumour as HPV-driven, it is crucial to

Fig. 5.2.3. Estimated age-standardized (World) incidence and mortality rates (ASR) per 100 000 person-years for head and neck
cancers, by sex and region, 2018: (A) lip and oral cavity, (B) larynx, (C) nasopharynx, (D) oropharynx, and (E) hypopharynx.
A Lip and oral cavity
B Larynx
314
C Nasopharynx
D Oropharynx
CHAPTER 5.2
SECTION 5

E Hypopharynx
include other markers related to HPV- Cancer Incidence in Five Continents cavity cancer and laryngeal cancer
induced carcinogenesis, such as database (Table 5.2.1) [11]. were overestimated. Currently, the
p16INK4a and messenger RNA (mRNA) HPV-related cases of head and attributable fractions are estimated
of the viral oncoproteins E6 and E7. neck cancer arise more often in the as 2.2% for oral cavity cancer and
A recent systematic review reported oropharynx (for which 30.8% of cas- 2.4% for laryngeal cancer [11]; these
on the attributable fractions in head es are HPV-related), and particu- figures have been confirmed with re-
and neck cancers, on the basis of larly in the tonsils. Recent estimates cent comprehensive studies [10].
HPV DNA and viral E6/E7 mRNA and showed that previous figures based Globally, approximately 38 000
the numbers of new cases from the on HPV DNA for HPV-related oral cases of head and neck cancer are
Table 5.2.1. Numbers of new cases of head and neck cancer attributable to human papillomavirus (HPV) infection and corresponding
attributable fractions by cancer site, worldwide, 2012a
Number or fraction Cancer site (ICD-10 code)
Oral cavity Oropharynx Other pharynx Larynx

(C02–06) (C01, C09–10) (C12–14) (C32)
Number of incident cases 200  000 9 6  000 78  000 160  000
Attributable fraction (%) 2.2 30.8 0 2.4
Number attributable to HPV 4 400 2 9  000 0 3 800
Number attributable to HPV by sex

Male 2 900 24  000 0 3 300
Female 1 500 5 500 0 460
ICD-10, International Statistical Classification of Diseases and Related Health Problems, 10th revision.
a
Numbers are rounded to two significant digits.
316
Fig. 5.2.4. Global distribution of age-standardized (World) incidence rates (ASR) per 100 000 of head and neck cancer cases
(oropharynx, oral cavity, and larynx) attributable to human papillomavirus (HPV) in both sexes, 2012.
HPV-related (Fig. 5.2.4). The bur- of tumours in locations such as the alcohol consumption are still poorly
den of HPV-related oropharyngeal oral cavity, larynx, and hypopharynx understood. Studies have produced
cancer is higher in high-income and are responsible for a different diverse results [8].
regions such as North America fraction of oropharyngeal cancers
and northern Europe, where HPV- across regions, with a higher attribu- Other risk factors
related cancers make up about 70– table fraction in regions with a lower Other risk factors include poor oral
80% of cases [10–12]. rate of HPV-related cancers. The risk hygiene, smoking marijuana, drink-
In several countries, particularly of cancer is higher in heavy smoking hot beverages such as maté,
in high-income regions, there has ers, as identified by a high product and some occupational exposures,
been an increasing trend in oro- of smoking rate in packs per day and such as metal smelting and tex-
pharyngeal cancer, attributed to an duration of smoking in years (“pack- tile production. These etiological
increase in HPV-related cases [8,12]. years”), and is higher for longer du- agents cause a field cancerization
This increasing trend could be ex- ration of smoking and in smokers of background that produces a high
plained partly by changes in the prev- black tobacco (see Chapter 2.1). probability of developing second
alence of risk factors, with decreases Use of chewing tobacco, other primary cancers at different sites in
in tobacco use and changes in sexual smokeless tobacco products, and the head and neck. This is not the
behaviour resulting in an increase in other substances, such as through case for HPV-related cancers of the
CHAPTER 5.2
SECTION 5
the likelihood of oral HPV infection. betel quid and areca nut chewing, head and neck, for which the inci-
There is a greater predominance is associated with risk of oral cav- dence of second primary cancers is
of HPV16 in head and neck cancers ity cancer, particularly in India and lower than that for non-HPV-relat-
compared with other HPV-related China, and specifically affects the ed cancers. HPV-related cancers
cancers. Globally, 84.9% of HPV- floor of the mouth and the pharynx. result from a persistent localized
related head and neck cancers The interaction between to- epithelial infection, which – if not
are attributable to HPV16/18; for bacco use and alcohol consump- resolved – may evolve by a trans-
HPV6/11/16/18/31/33/45/52/58, the tion is greater than additive. For formation process.
proportion is 89.7% (Table 5.2.2) [11]. alcohol consumption (see Chapter Epstein–Barr virus (EBV) is clas-
2.3), the risk is related to the du- sified by the IARC Monographs as
Tobacco use and alcohol ration of heavy drinking more than carcinogenic to humans (Group 1) for
consumption to the quantity consumed per day. nasopharyngeal cancers, consider-
Tobacco use and alcohol consump- The types of interactions between ing that almost all tumours harbour
tion are the most important causes HPV infection and tobacco use and the EBV genome and express certain

EBV gene products [13]. Other risk Table 5.2.2. Numbers of cases of head and neck cancer attributable to human
factors for nasopharyngeal cancers papillomavirus (HPV) infection by region and sex, and relative contributions by specific
include genetic susceptibility (the fa- HPV types, worldwide, 2012a
milial relative risk of nasopharyngeal Number or proportion Male Female
cancer is estimated to be greater
than 4–fold) [14]; consumption of pre- Number attributable to HPV
Africa 600 230
served foods, particularly Chinese- Asia 9 810 2  200
type salted fish, probably because Americas 7 980 2  180
of their high content of nitrosamines Europe 11  000 2  800
Oceania 320 90
[15]; and, less consistently, other ex- Less-developed countries 8 600 2  100
posures such as tobacco use or, per- More-developed countries 2 2  000 5  500
haps, alcohol consumption [16,17].
Number of cases and proportion among HPV-related cases
HPV16/18 32 000; 84.9%
Genetics HPV6/11/16/18/31/33/45/52/58 34 000; 89.7%
HPV-related head and neck can-

a
Numbers are rounded to two significant digits.
cers are a distinct entity, compared

with those that are non-HPV-re- oncogene [21]. For some additional main carcinogen of a particular tu-
lated, in terms of epidemiological, alterations, the crucial role as driver mour is crucial, because of the new
clinical, and molecular character- events is not yet clear: the losses proposals on de-intensification of
istics. The epidemiology has been of chromosomal loci 14q32 and 9q, treatments for HPV-related cancers,
described above. which contain the tumour necrosis which are undergoing evaluation.
Clinically, HPV-related cases factor receptor-associated factor 3 The reference standard in as-
of oropharyngeal cancer have bet- (TRAF3) and ataxia telangiectasia signing HPV causality is detec-
ter survival than non-HPV-related mutated (ATM) genes, respectively tion of E6/E7 mRNA (E6*I mRNA
cases [18]. In HPV-related cancers, [22]. Finally, APOBEC has a specif- by reverse transcriptase polymer-
p16INK4a is overexpressed through ic mutational profile in HPV-related ase chain reaction); however, this
disruption by E7 of the retinoblas- tumours, with high cytosine deami- is a rather complicated technique
toma pathway. The eighth edition of nase activity [22]. for routine clinical laboratories [9].
the American Joint Committee on The driver genes and pathways Other alternatives considered are
Cancer and Union for International most affected in non-HPV-related in situ hybridization, which is spe-
Cancer Control tumour–node–me- tumours have been reported in the cific but lacks sensitivity; p16INK4a,
tastasis (TNM) classification pre- published genomic data, involving which has high sensitivity but mod-
sented a different staging system for 279 cases of head and neck can- erate specificity; and double testing
p16INK4a-positive tumours, resulting in cer and available data for more than of HPV DNA and p16INK4a, which is
a lower stage of these tumours com- 500 cases from the Cancer Genome emerging as the most suitable and
pared with the previous edition [19]. Atlas [21]. These data have recently reliable strategy for HPV-driven oro-
With respect to molecular differ- been summarized in a review on pharyngeal cancers [24]. In addition,
ences, the main carcinogenic pro- genomics in head and neck can- HPV16 E6 serology has recently
cess in HPV-related head and neck cers [23]. The main driver genes been proposed as a potential bio-
cancers is through the action of the implicated in the carcinogenesis of marker for diagnosis of HPV-driven
viral oncoproteins. E6 binds to and non-HPV-related tumours are sum- oropharyngeal cancers, with good
degrades p53, preventing apopto- marized in Table 5.2.3. sensitivity and specificity reported,
sis, whereas E7 binds to and de- Genomic profiling is not regu- and also as a potential biomarker for
grades retinoblastoma, promoting larly used at clinics for the man- prevention and follow-up.
cell proliferation [20]. The genes agement of patients with head and In addition to being useful for
that are most affected in non-HPV- neck cancer. However, such clas- HPV diagnosis in HPV-related can-
related head and neck cancers, sifications will be more relevant in cers, genomic profiling reveals in-
TP53 and cyclin-dependent kinase the future, with increasing informa- teresting patterns. A recent system-
inhibitor 2A (CDKN2A), are unaf- tion on genetics and potential drug- atic review of the available literature
fected in HPV-related tumours. In gable targets and differential man- reported the following potential ge-
addition to the actions of the viral agement of patients. nomic progression models and ge-
oncoproteins, the most common The management of HPV-related nomic profiles (Fig. 5.2.5) [23].
genetic changes in HPV-related tu- oropharyngeal cancers is not modi-
mours are in the phosphoinositide fied by the HPV diagnosis, but this HPV-related head and neck
3-kinase (PI3K) pathway, particu- information is used for prognostic cancers
larly involving activating mutations purposes [24]. In these cancers, an HPV infection in oral squamous epi-
and amplifications of the PIK3CA accurate diagnosis of HPV as the thelium leads mainly to productive
318
Table 5.2.3. Genes with frequent and highly significant somatic genetic changes in human papillomavirus (HPV)-negative head
and neck cancers
Cellular Gene Protein Type of gene Mutation Frequency of copy

process frequency (%) number alterations (%)
Cell cycle TP53 p53 Tumour 7 2 1.4
suppressor
CDKN2 p16 INK4a Tumour 2 2 32
suppressor
CCND1 Cyclin D1 Oncogene 0.6 25
Growth signals EGFR Epidermal growth factor Oncogene 4 11
receptor
Survival PIK3CA Catalytic p110α subunit of Oncogene 18 21
class 1 PI3Ks
PTEN PTEN Tumour 3 4
suppressor
WNT signalling FAT1 Protocadherin FAT1 Tumour 23 8
suppressor
AJUBA LIM domain-containing Tumour 7a 1
protein AJUBA suppressor
NOTCH1 NOTCH1 Tumour 18 4
suppressor
Epigenetic KMT2D Histone-lysine Tumour 16 0.4
regulation N-methyltransferase KMT2D suppressor
NSD1 Histone-lysine Tumour 12a 0.8
N-methyltransferase NSD1 suppressor
a
Putative passenger mutation that requires further functional studies.
infections, whereas the viral trans- invisible pre-malignant lesions are Prevention and
formation process more commonly also identified microscopically as monitoring biomarkers
arises from the epithelium of the dysplastic mucosal epithelium.
Early-stage tumours of the upper
tonsillar crypts. The tonsillar epi- Two potential genomic profiles
thelium may be a non-permissive can be identified for non-HPV-re- aerodigestive tract can be cured;
productive medium in which HPV lated cancers: (i) a profile presum- for late-stage disease, prognosis is
infection progresses at a higher ably related to ageing, with CNA- poor. For non-HPV-related cancers,
frequency directly to a transforma- silent tumours, wild-type TP53, prevention strategies could include
tion process, without a clear pre- and HRAS and CASP8 mutations; oral cancer screening through visu-
neoplastic lesion. and (ii) a tobacco-related profile, in al oral examination, which has been
Two genomic profiles can be which deregulation of the cell cycle demonstrated to result in a lower
described for HPV-related cancers by abrogation of the retinoblastoma mortality rate in a randomized con-
on the basis of expression profil- and p53 pathways seems to occur trolled trial setting.
ing: (i) immune response and mes- at the very beginning of the carci- Options for prevention of head
enchymal cell differentiation, indi- nogenic process. The first profile and neck cancer depend on the
cated by enrichment of 16q losses; seems to have better prognosis type of etiological factor involved in
and (ii) keratinocyte differentiation than the second. Within the second various situations and the type of
CHAPTER 5.2
SECTION 5
and oxidative reduction process, group, at least three subgroups can prevention.
indicated by enrichment of 3q copy be identified on the basis of expres-
number alterations (CNA) and sion profiling: classical, basal, and
Primary prevention
PIK3CA mutations [25]. There is no mesenchymal. The classical sub- The aim of primary prevention is
evidence that these two groups be- group is characterized by mutations to intervene before health effects
have differently in terms of survival. of the nuclear factor erythroid 2-re- occur. Globally, the burden of can-
lated factor 2 (NFE2L2) pathway. cers attributable to tobacco use has
Non-HPV-related head and More subgroups may exist, and been reduced in some world regions
neck cancers further research is required, also on as a result of intensive campaigns
Although a high proportion of cas- the clinical implications [26–28]. to prevent and reduce tobacco use.
es present with tumours de novo, Finally, it is noteworthy that im- Reduced alcohol consumption may
there are precancerous lesions that mune checkpoint inhibitors have also be a consideration.
are visible, such as leukoplakia and recently emerged as novel potential An increasing proportion of cas-
erythroplakia lesions, and many therapeutic options [29,30]. es, such as those of oropharyngeal

Fig. 5.2.5. Genomic carcinogenesis models of head and neck squamous cell carcinoma. CNA, copy number alterations; HPV,
human papillomavirus; IMU, immune response and mesenchymal cell differentiation; KRT, keratinocyte differentiation and oxidative
reduction process.
HPV+ve tumours Normal mucosal epithelium
HRAS and HPV–ve,

CASP8 CNA-silent tumours
HPV
Age ?
HPV–ve, CNA-high tumours
TP53 and Basal

CDKN2A
HPV E6 and E7
Field cancerization
FAT1 and
NOTCH1
NFE2L2,
CUL3 or
HPV-IMU HPV-KRT KEAP1
Classical
PIK3CA
Mesenchymal
cancer, are caused by other agents, infections compared with non- be successful for head and neck
such as HPV. Therefore, HPV vacci- vaccinated individuals (1.99% vs cancer. However, repeated visual
nation is a potential tool for primary 3.52%; P = 0.04). The second study oral examination has been demon-
prevention (see Chapter 6.3). Only analysed data from 2627 people strated to have long-term effects
one study has reported on the ef- aged 18–33 years who participated in reducing oral cancer incidence
ficacy of the bivalent HPV vaccine in NHANES in 2011–2014 [33]. The and mortality in a randomized trial
as prophylaxis against oral infection prevalence of oral HPV6/11/16/18 in India; this result supports the
[31]. In the context of this vaccine infections was significantly lower in introduction of visual oral screen-
clinical trial in women aged 18– vaccinated individuals than in non- ing, particularly targeting users of
25 years, the estimated efficacy of vaccinated individuals (0.11% vs smoking or chewing tobacco, al-
the vaccine in reducing oral HPV in- 1.61%; P = 0.008), corresponding cohol drinkers, or both in high-inci-
fection was 93.3% (95% confidence to an estimated reduction in preva- dence countries [34].
interval, 63–100%) [31]. lence of 88.2% (95% confidence in- Serological detection of antibod-
Two recently published stud- terval, 5.7–98.5%) after adjustment ies against HPV (anti-HPV16 E6) has
ies have assessed the effective- for age, sex, and race. recently been postulated as a poten-
ness of the quadrivalent HPV vac- tial biomarker for HPV-related oro-
cine in reducing oral HPV infection Screening pharyngeal cancer. In cohort studies,
[32,33]. The first study included Because a considerable proportion seroconversion has been detected
3040 people aged 18–30 years of cases are diagnosed at locally ad- up to 10 years before the diagnosis of
who participated in the National vanced stages, screening (second- oropharyngeal cancer [35]. This ob-
Health and Nutrition Examination ary prevention) for early detection of servation is very relevant given that
Survey (NHANES) in the USA in disease is of great importance. it is not yet known what the pre-neo-
2009–2014 [32]. Vaccinated indi- Early detection strategies based plastic lesion of oropharyngeal can-
viduals had a significantly lower on cytology, such as for cervical cer is. However, there are still many
prevalence of oral HPV6/11/16/18 cancer, have not been proven to gaps in knowledge that must be filled,
320
such as determining the best clinical increased risk of recurrence, and able tumour DNA was 100% for
triage algorithm for identifying po- the third study did not observe dif- early-stage disease and 95% for
tential lesions once an HPV-positive ferences [36–38]. In relation to late-stage disease. Saliva was ob-
case has been detected, among oth- the persistence of viral HPV DNA served to be preferentially enriched
er considerations. in oral rinses after treatment, one for tumour DNA from the oral cavity,
study reported that persistence whereas plasma was preferentially
Monitoring biomarkers could be linked with the incidence enriched for tumour DNA from the
The prognostic value of monitoring of recurrence [39]. other sites. Tumour DNA in saliva
anti-HPV antibody titres throughout In 2015, a study explored tu- was found after surgery in three
a patient’s treatment in the survival- mour-specific DNA as a biomarker patients before clinical diagnosis of
free period of disease is not well detected in saliva or plasma for recurrence, but in none of the five
characterized. So far, only three head and neck cancer by search- patients without recurrence. These
studies have provided information ing for somatic mutations or HPV findings, if confirmed, have direct
on this topic. Two of the studies ob- genes (collectively referred to as implications in the follow-up and
served an association of increased tumour DNA) in 93 cases [40]. The clinical management of patients.
levels of antibodies in blood with fraction of patients with detect-
References
1. NCI (2018). Head and neck cancer. United 8. Gillison ML, Alemany L, Snijders PJ, 14. Liu Z, Chang ET, Liu Q, Cai Y, Zhang Z,
States National Cancer Institute, National Chaturvedi A, Steinberg BM, Schwartz S, Chen G, et al. (2017). Quantification of
Institutes of Health. Available from: https:// et al. (2012). Human papillomavirus and familial risk of nasopharyngeal carci-
www.cancer.gov/types/head-and-neck/ diseases of the upper airway: head and noma in a high-incidence area. Cancer.
head-neck-fact-sheet. neck cancer and respiratory papillomato- 123(14):2716–25. https://doi.org/10.1002/
sis. Vaccine. 30(Suppl 5):F34–54. https:// cncr.30643 PMID:28241094
2. Bray F, Ferlay J, Soerjomataram I, Siegel doi.org/10.1016/j.vaccine.2012.05.070
RL, Torre LA, Jemal A (2018). Global PMID:23199965 15. IARC (2012). Personal habits and in-
cancer statistics 2018: GLOBOCAN es- door combustions. IARC Monogr Eval
timates of incidence and mortality world- 9. Holzinger D, Schmitt M, Dyckhoff G, Benner Carcinog Risks Hum. 100E:1–575.
wide for 36 cancers in 185 countries. CA A, Pawlita M, Bosch FX (2012). Viral RNA Available from: http://publications.iarc.
Cancer J Clin. 68(6):394–424. https://doi. patterns and high viral load reliably define fr/122 PMID:23193840
org/10.3322/caac.21492 PMID:30207593 oropharynx carcinomas with active HPV16
involvement. Cancer Res. 72(19):4993– 16. Chua MLK, Wee JTS, Hui EP, Chan ATC
3. Chaturvedi AK, Anderson WF, Lortet- 5003. https://doi.org/10.1158/0008-5472. (2016). Nasopharyngeal carcinoma.
Tieulent J, Curado MP, Ferlay J, Franceschi CAN-11-3934 PMID:22991302 Lancet. 387(10022):1012–24. https://doi.
S, et al. (2013). Worldwide trends in inci- org/10.1016/S0140 - 673 6(15) 0 0 055 - 0
dence rates for oral cavity and oropharyn- 10. Castellsagué X, Alemany L, Quer M, PMID:26321262
geal cancers. J Clin Oncol. 31(36):4550–9. Halec G, Quirós B, Tous S, et al.; ICO
https://doi.org/10.1200/JCO.2013.50.3870 International HPV in Head and Neck 17. Tsao SW, Yip YL, Tsang CM, Pang PS, Lau
PMID:24248688 Cancer Study Group (2016). HPV in- VMY, Zhang G, et al. (2014). Etiological
volvement in head and neck cancers: factors of nasopharyngeal carcinoma.
4. Tang LL, Chen WQ, Xue WQ, He YQ, Zheng comprehensive assessment of biomark- Oral Oncol. 50(5):330–8. https://doi.
RS, Zeng YX, et al. (2016). Global trends ers in 3680 patients. J Natl Cancer Inst. org/10.1016/j.oraloncology.2014.02.006
in incidence and mortality of nasopharyn- 108(6):djv403. https://doi.org/10.1093/jnci/ PMID:24630258
geal carcinoma. Cancer Lett. 374(1):22–30. djv403 PMID:26823521
https://doi.org/10.1016/j.canlet.2016.01.040 18. Ang KK, Harris J, Wheeler R, Weber
PMID:26828135 11. de Martel C, Plummer M, Vignat J, R, Rosenthal DI, Nguyen-Tân PF, et al.
Franceschi S (2017). Worldwide burden (2010). Human papillomavirus and sur-
5. Chatenoud L, Garavello W, Pagan E, of cancer attributable to HPV by site, vival of patients with oropharyngeal
CHAPTER 5.2
SECTION 5
Bertuccio P, Gallus S, La Vecchia C, et al. country and HPV type. Int J Cancer. cancer. N Engl J Med. 363(1):24–35.
(2016). Laryngeal cancer mortality trends 141(4):664–70. https://doi.org/10.1002/ https://doi.org/10.1056/NEJMoa0912217
in European countries. Int J Cancer. ijc.30716 PMID:28369882 PMID:20530316
138(4):833–42. https://doi.org/10.1002/ijc.
29833 PMID:26335030 12. Chaturvedi AK, Engels EA, Pfeiffer RM, 19. Brierly JD, Gospodarowicz MK, Wittekind
Hernandez BY, Xiao W, Kim E, et al. (2011). C, editors (2017). TNM classification of
6. IARC (2007). Human papillomaviruses. Human papillomavirus and rising oro- malignant tumours. 8th ed. Oxford, UK:
IARC Monogr Eval Carcinog Risks Hum. pharyngeal cancer incidence in the United Wiley Blackwell.
90:1–636. Available from: http://publica States. J Clin Oncol. 29(32):4294–301.
tions.iarc.fr/108 PMID:18354839 https://doi.org/10.1200/JCO.2011.36.4596 20. Doorbar J, Quint W, Banks L, Bravo IG,
7. Walboomers JM, Jacobs MV, Manos PMID:21969503 Stoler M, Broker TR, et al. (2012). The
MM, Bosch FX, Kummer JA, Shah biology and life-cycle of human papillo-
13. IARC (2012). Biological agents. IARC maviruses. Vaccine. 30(Suppl 5):F55–70.
KV, et al. (1999). Human papillomavi- Monogr Eval Carcinog Risks Hum. 100B:1–
rus is a necessary cause of invasive https://doi.org/10.1016/j.vaccine.2012.
441. Available from: http://publications.iarc. 06.083 PMID:23199966
cervical cancer worldwide. J Pathol. fr/119 PMID:23189750
189(1):12–9. https://doi.org/10.1002/(SICI)
10 9 6 - 9 8 9 6 (19 9 9 0 9)18 9 :1<12 :: A I D -
PATH431>3.0.CO;2-F PMID:10451482

21. The Cancer Genome Atlas Network 29. Ferris RL, Blumenschein G Jr, Fayette 36. Koslabova E, Hamsikova E, Salakova
(2015). Comprehensive genomic char- J, Guigay J, Colevas AD, Licitra L, et al. M, Klozar J, Foltynova E, Salkova E, et
acterization of head and neck squamous (2016). Nivolumab for recurrent squa- al. (2013). Markers of HPV infection and
cell carcinomas. Nature. 517(7536):576– mous-cell carcinoma of the head and survival in patients with head and neck tu-
82. https://doi.org/10.1038/nature14129 neck. N Engl J Med. 375(19):1856–67. mors. Int J Cancer. 133(8):1832–9. https://
PMID:25631445 https://doi.org/10.1056/NEJMoa1602252 doi.org/10.1002/ijc.28194 PMID:23564321
PMID:27718784
22. Hayes DN, Van Waes C, Seiwert TY 37. Fakhry C, Qualliotine JR, Zhang Z,
(2015). Genetic landscape of human 30. Seiwert TY, Burtness B, Mehra R, Weiss Agrawal N, Gaykalova DA, Bishop JA, et
papillomavirus-associated head and neck J, Berger R, Eder JP, et al. (2016). Safety al. (2016). Serum antibodies to HPV16
cancer and comparison to tobacco-related and clinical activity of pembrolizumab early proteins warrant investigation as po-
tumors. J Clin Oncol. 33(29):3227–34. for treatment of recurrent or metastatic tential biomarkers for risk stratification and
https://doi.org/10.1200/JCO.2015.62.1086 squamous cell carcinoma of the head and recurrence of HPV-associated oropharyn-
PMID:26351353 neck (KEYNOTE-012): an open-label, geal cancer. Cancer Prev Res (Phila).
multicentre, phase 1b trial. Lancet Oncol. 9(2):135–41. https://doi.org/10.1158/1940-
23. Leemans CR, Snijders PJF, Brakenhoff
17(7):956–65. https://doi.org/10.1016/ 6207.CAPR-15-0299 PMID:26701665
RH (2018). The molecular landscape of
head and neck cancer. Nat Rev Cancer. S1470-2045(16)30066-3 PMID:27247226
38. Lang Kuhs KA, Kreimer AR, Trivedi S,
18(5):269–82. https://doi.org/10.1038/nrc. Holzinger D, Pawlita M, Pfeiffer RM, et al.
31. Herrero R, Quint W, Hildesheim A,
2018.11 PMID:29497144 (2017). Human papillomavirus 16 E6 anti-
Gonzalez P, Struijk L, Katki HA, et al.;
24. Prigge ES, Arbyn M, von Knebel Doeberitz CVT Vaccine Group (2013). Reduced bodies are sensitive for human papilloma-
M, Reuschenbach M (2017). Diagnostic ac- prevalence of oral human papillomavirus virus-driven oropharyngeal cancer and
curacy of p16INK4a immunohistochemistry in (HPV) 4 years after bivalent HPV vaccina- are associated with recurrence. Cancer.
oropharyngeal squamous cell carcinomas: tion in a randomized clinical trial in Costa 123(22):4382–90. https://doi.org/10.1002/
a systematic review and meta-analysis. Rica. PLoS One. 8(7):e68329. https:// cncr.30966 PMID:28950407
Int J Cancer. 140(5):1186–98. https://doi. doi.org/10.1371/journal.pone.0068329 39. Rettig EM, Wentz A, Posner MR, Gross
org/10.1002/ijc.30516 PMID:27859245 PMID:23873171 ND, Haddad RI, Gillison ML, et al. (2015).
25. Zhang Y, Koneva LA, Virani S, Arthur AE, 32. Hirth JM, Chang M, Resto VA, Guo F, Prognostic implication of persistent hu-
Virani A, Hall PB, et al. (2016). Subtypes Berenson AB; HPV Study Group (2017). man papillomavirus type 16 DNA detection
of HPV-positive head and neck cancers Prevalence of oral human papillomavi- in oral rinses for human papillomavirus-
are associated with HPV characteris- rus by vaccination status among young related oropharyngeal carcinoma. JAMA
tics, copy number alterations, PIK3CA adults (18–30 years old). Vaccine. Oncol. 1(7):907–15. https://doi.org/10.1001/
mutation, and pathway signatures. Clin 35(27):3446–51. https://doi.org/10.1016/j. jamaoncol.2015.2524 PMID:26226294
Cancer Res. 22(18):4735–45. https://doi. vaccine.2017.05.025 PMID:28526331 40. Wang Y, Springer S, Mulvey CL, Silliman
org/10.1158/1078 - 0 432.CCR-16 - 0323
33. Chaturvedi AK, Graubard BI, Broutian N, Schaefer J, Sausen M, et al. (2015).
PMID:27091409
Detection of somatic mutations and HPV
T, Pickard RKL, Tong ZY, Xiao W, et al.
26. Chung CH, Parker JS, Karaca G, Wu J, in the saliva and plasma of patients with
(2018). Effect of prophylactic human pap-
Funkhouser WK, Moore D, et al. (2004). head and neck squamous cell carcino-
illomavirus (HPV) vaccination on oral
Molecular classification of head and neck mas. Sci Transl Med. 7(293):293ra104.
HPV infections among young adults in the
squamous cell carcinomas using pat- https://doi.org/10.1126/scitranslmed.aaa
United States. J Clin Oncol. 36(3):262–7.
terns of gene expression. Cancer Cell. 8507 PMID:26109104
https://doi.org/10.1200/JCO.2017.75.0141
5(5):489–500. https://doi.org/10.1016/
PMID:29182497
S1535-6108(04)00112-6 PMID:15144956
34. Sankaranarayanan R, Ramadas K, Thara
27. Walter V, Yin X, Wilkerson MD, Cabanski
CR, Zhao N, Du Y, et al. (2013). Molecular S, Muwonge R, Thomas G, Anju G, et al.
subtypes in head and neck cancer exhibit (2013). Long term effect of visual screen-
distinct patterns of chromosomal gain and ing on oral cancer incidence and mortal-
loss of canonical cancer genes. PLoS ity in a randomized trial in Kerala, India.
One. 8(2):e56823. https://doi.org/10.1371/ Oral Oncol. 49(4):314–21. https://doi.
journal.pone.0056823 PMID:23451093 org/10.1016/j.oraloncology.2012.11.004
PMID:23265945
28. Keck MK, Zuo Z, Khattri A, Stricker TP,
Brown CD, Imanguli M, et al. (2015). 35. Kreimer AR, Johansson M, Waterboer
Integrative analysis of head and neck T, Kaaks R, Chang-Claude J, Drogen D,
cancer identifies two biologically distinct et al. (2013). Evaluation of human papil-
HPV and three non-HPV subtypes. Clin lomavirus antibodies and risk of subse-
Cancer Res. 21(4):870–81. https://doi. quent head and neck cancer. J Clin Oncol.
org/10.1158/1078 - 0 432.CCR-14 -2481 31(21):2708–15. https://doi.org/10.1200/
PMID:25492084 JCO.2012.47.2738 PMID:23775966
322
5.3
Oesophageal cancer
A tale of two malignancies
Reza Malekzadeh Valerie McCormack (reviewer) David Whiteman (reviewer)

Christian C. Abnet Aoife Ryan (reviewer)
Sanford M. Dawsey Katherine Van Loon (reviewer)
may be more important in high- early detection and treatment (sec-

SUMMARY incidence regions of Asia and ondary prevention) are the most
Africa, including poor diet, in- important strategies to reduce the
●● Oesophageal cancer is the sixth door air pollution, consump- burden of this fatal cancer [1].
most common cause of cancer tion of hot beverages, poor oral
death worldwide, and it is an im- health, use of non-piped water,
portant global health challenge.
Molecular characteristics
and opium use.
Oesophageal cancer has two
●● Oesophageal squamous cell ●● Evaluations of preventive strat- main histological types: oesopha-
carcinoma and oesophageal egies are under way for both geal squamous cell carcinoma
adenocarcinoma are very dif- types of oesophageal cancer, (Fig. 5.3.1) and oesophageal ade-
ferent diseases that occur in including efforts to reduce ex- nocarcinoma. There are molecular
the same organ; they have dis- posure to known carcinogens, similarities between squamous cell
tinct biological characteristics, chemoprevention trials, and carcinoma of the oesophagus and
geographical distributions, risk development of effective early squamous cancers of other organs,
factors, and time trends. detection and treatment proto- and between oesophageal adeno-
cols for populations at high risk. carcinoma and stomach adenocar-
●● The eastern coast of Africa is
a recognized area of high risk cinoma, but there are significant
for oesophageal squamous cell molecular differences at both the
Oesophageal cancer is the eighth genomic and epigenomic levels
carcinoma. Unique to this high-
most common cancer and the sixth between oesophageal squamous
risk corridor is that up to 20% of
most common cause of cancer cell carcinoma and oesophageal
cases occur in people younger death worldwide, and it is an impor-
than 40 years. adenocarcinoma [4]. These two
tant global health challenge [1]. The cancer types have different sets
●● Genome-wide association stud- two histological types of oesopha- of driver genes, mutational signa-
ies of both types of oesophageal geal cancer differ in the populations tures, and prognostic biomarkers,
cancer have identified a modest that are affected and have com- which are almost mutually exclusive
number of germline polymor- pletely distinct biological charac- [4]. Recently, several mutations and
CHAPTER 5.3
SECTION 5
phisms associated with risk of teristics, geographical distributions, mutational signatures have been
these tumours, but genetic pre- risk factors, and time trends [2,3]. correlated with the overall survival
disposition has not been defini- Five-year survival rates for oe- of patients with oesophageal can-
tively characterized. High rates sophageal cancer are about 20% in cer; in the future, these may serve
of TP53 mutations occur in both Europe and the USA and less than as prognostic biomarkers [4].
tumour types in most, but not 5% in low- and middle-income coun-
all, populations. tries [1], mainly because of the late
occurrence of symptoms and the Epidemiology
●● Tobacco use and alcohol con- consequent usually advanced stage In 2012, there were an estimated
sumption are the known and at diagnosis. Therefore, identifying 398 000 new cases of oesopha-
primary causes of oesopha- and reducing exposure to modifiable geal squamous cell carcinoma and
geal squamous cell carcinoma, risk factors (primary prevention) and 52 000 new cases of oesophageal
particularly in low-incidence development and implementation of adenocarcinoma worldwide, corre-
countries. Other risk factors practical and accurate methods for sponding to global incidence rates
Chapter 5.3 • Oesophageal cancer 323

Fig. 5.3.1. Oesophageal squamous cell 20%) of cases occur in people youn-

carcinoma. Macroscopic appearance of a ger than 40 years [5]. Worldwide,
mid-oesophageal mass. FUNDAMENTALS
the male-to-female incidence ratio
is 2.7:1 for oesophageal squamous ■■ Oesophageal cancer has
cell carcinoma and 4.4:1 for oesoph- two main histological types:
ageal adenocarcinoma [2]. oesophageal squamous cell
carcinoma and oesophageal
Genetics and genomics adenocarcinoma.
A moderate familial susceptibility ■■ Oesophageal squamous cell
to oesophageal cancer has been carcinoma makes up about
reported for both oesophageal 87% of all cases of oesopha-
squamous cell carcinoma and oe- geal cancer worldwide. The
sophageal adenocarcinoma; this is
incidence is very uneven geo-
thought to be at least partially due
graphically, with large propor-
to the inheritance of susceptibility
tions of cases occurring in a
alleles [3].
few populations at high risk.
Genome-wide association stud-
ies for oesophageal squamous cell ■■ Oesophageal adenocarcino-
carcinoma [6] and oesophageal ad- ma makes up the majority of
enocarcinoma [7] have identified a oesophageal cancer cases
modest number of germline poly- in North America, western
morphisms associated with risk of Europe, Australia, and New
these tumours, but neither disease Zealand.
has been studied in large enough
numbers to comprehensively define ■■ Although incidence rates of
genetic predisposition overall or oesophageal squamous cell
in different ethnic groups. Several carcinoma are declining, inci-
of 5.2 per 100 000 for oesophageal
rare, high-penetrance genetic de- dence rates of oesophageal
squamous cell carcinoma and 0.7
fects, such as tylosis and Fanconi adenocarcinoma are increas-
per 100 000 for oesophageal ad-
anaemia, have been linked to high ing in many regions.
enocarcinoma [2]. Oesophageal
squamous cell carcinoma makes up risk of oesophageal squamous cell
■■ Oesophageal cancer has very
about 87% of all cases of oesopha- carcinoma, but they explain only a
poor survival, with mortality
geal cancer globally; more than half small fraction of cases.
rates (7.7 per 100 000) that are
of the cases occur in China, and Whole-genome and whole-ex-
close to the incidence rates
25% occur in India, South-East Asia, ome sequencing of paired tumour
(9.0 per 100 000).
and Central Asia [2]. For oesopha- and normal tissues from Chinese
geal adenocarcinoma, about 44% patients with oesophageal squa- ■■ The low survival rates for
of the global burden occurs in North mous cell carcinoma has revealed oesophageal cancer are due to
America and western Europe [2]. eight genes with frequent somatic the advanced stage at diagno-
The global distribution of mutations, including six known sis, but practical, cost-effective
age-standardized incidence rates tumour-associated genes (TP53, population-based screening
for oesophageal cancer is shown in RB1, CDKN2A, PIK3CA, NOTCH1, has not yet been developed.
Fig. 5.3.2. The incidence of oesoph- and NFE2L2) and two novel
ageal squamous cell carcinoma is genes (ADAM29 and FAM135B)
remarkably uneven geographically, [3]. Whole-exome sequencing of
with a 21-fold difference between paired tumour and normal tissues suggesting a potential role for thera-
the countries with the lowest and from patients with oesophageal py with trastuzumab (an anti-HER2
the highest incidence rates. The adenocarcinoma found mutations monoclonal antibody) in these tu-
incidence of oesophageal squa- in 28 genes, of which five (TP53, mours [9]. Studies are under way to
mous cell carcinoma is very high CDKN2A, SMAD4, ARID1A, and find mutational signatures associ-
within sharply defined regions in PIK3CA) are relevant to the patho- ated with both tumour types.
the north-eastern Islamic Republic genesis of adenocarcinoma [3]. A Oesophageal squamous cell
of Iran, Central Asia, north-central minority (15–29%) of oesophageal carcinoma tumours in people from
China, East Africa, southern Africa, adenocarcinomas show overex- Golestan Province, Islamic Republic
and southern South America [1,2]. pression or amplification of human of Iran, have the highest rate of
Unique to the African high-risk cor- epidermal growth factor receptor 2 TP53 mutations ever reported in
ridor is that large numbers (up to (HER2; also known as ERBB2) [8], any cancer [10]. The heterogeneous
324
Fig. 5.3.2. Global distribution of estimated age-standardized (World) incidence rates (ASR) per 100 000 person-years for oesopha-
geal cancer (A) in men and (B) in women, 2018.
CHAPTER 5.3
SECTION 5
mutation pattern is highly sugges- Etiology New Zealand, almost 90% of cases
tive of a causative role for multiple of oesophageal squamous cell car-
Risk factors for oesophageal squa-
environmental carcinogens, includ- cinoma are attributable to tobacco
mous cell carcinoma and oesopha-
ing polycyclic aromatic hydrocar- use and heavy alcohol consump-
bons (PAHs) [10]. In contrast, a geal adenocarcinoma are listed in
tion, and the incidence rate in men
substantial fraction of oesophageal Table 5.3.1. is 3–4 times that in women [1,3,12].
squamous cell carcinoma tumours However, in the oesophageal squa-
Oesophageal squamous cell
in East Africa do not appear to have mous cell carcinoma hotspots in
carcinoma
TP53 mutations, and a novel muta- Asia, Africa, and South America,
tional signature suggests that an- Oesophageal squamous cell carci- where the incidence rates in men
other, as-yet-unknown carcinogen noma is well known for its marked and women can be nearly equal,
could be important in this high-inci- etiological heterogeneity [1,12]. In multiple additional risk factors have
dence area [11]. the USA, Europe, Australia, and been implicated, including a poor

diet deficient in vitamins (especially Table 5.3.1. Risk factors for squamous cell carcinoma and adenocarcinoma of
riboflavin), indoor air pollution, con- the oesophagus a
sumption of hot beverages, poor
Risk factor Oesophageal Oesophageal
oral health, use of non-piped water, squamous cell adenocarcinoma
and opium use [1,12–14], with dif- carcinoma
ferent profiles of attributable risks
Sex Male > female Male > female
in different hotspot regions.
Low socioeconomic status is also Race Black > White White > Black
a consistent risk factor for oesopha- Genetic susceptibility ++ +
geal squamous cell carcinoma, even
Gastro-oesophageal reflux disease No data ++++
after comprehensive adjustment
for tobacco use, alcohol consump- Obesity Limited data ++++
tion, age, and many other potential Tobacco use ++++ ++
risk factors (see Chapter 4.3) [15]. In
addition, as suggested by the novel Alcohol consumption ++++ No association
mutational signature seen in the ge- Very hot beverages +++ No data
nomic study of tumours in East Africa
Diet low in fruits and vegetables +++ +
mentioned above [11], there may also
be as-yet-unknown risk factors that Low socioeconomic status +++ Limited data
may be important for the carcinogen- Helicobacter pylori infection No association Protective
esis of oesophageal squamous cell
carcinoma in the high-risk regions Poor oral health ++ Limited data
of the world. Recent epidemiologi- Opium use ++ No data

cal studies have shown no evidence
Indoor air pollution + No data
for a role of human papillomavirus
(HPV) in the etiology of oesophageal Non-piped water + No data
squamous cell carcinoma [1,12], and a
+, positive association (the number of + signs is based on the amount of evidence).
tumour sequencing has not revealed
any viral sequences incorporated
into the host DNA [11,12]. Fig. 5.3.3. A woman in Jamkhed, India, cooks indoors in smoky conditions. Exposure
to polycyclic aromatic hydrocarbons from combustion is a suspected carcinogen for
In most populations at high oesophageal squamous cell carcinoma.
risk, many of the above-mentioned
risk factors occur together. It is not
known how they interact to increase
risk, but a recent prospective analy-
sis estimated the combined effects
of multiple risk factors. Low socio-
economic status, opium smoking,
drinking hot tea, low intake of fruits
and vegetables, excessive tooth
loss, drinking non-piped water, and
exposure to indoor air pollution had
a combined population attributable
risk of 76% for oesophageal squa-
mous cell carcinoma [16].
Polycyclic aromatic
hydrocarbons and nitrosamines
One of the main suspected carcino-
gens for oesophageal squamous
cell carcinoma is PAHs. PAHs are
important carcinogens in tobacco
smoke (see Chapter 2.1) as well as
in the combustion products of other to high incidence rates in certain in poorly ventilated rooms may be a
organic materials, such as opium, regions [17]. In populations at high major factor for both the high rates
automobile and industrial fuels, risk, exposure to PAHs from indoor of oesophageal squamous cell car-
coal, and wood; exposure to PAHs air pollution caused by heating and cinoma and the nearly equal rates in
from both sources could contribute cooking with open coal or wood fires men and women [12,18].
326
Opium use
The cultivation of opium and the of opium and showed dose–re- interval, 1.18–2.90) for developing
consumption of raw opium take sponse trends for intensity, dura- oesophageal squamous cell carci-
place mainly in West and Central tion, and cumulative use [2]. Since noma, and there was a significant
Asia. These regions have a rela- the 1970s, opium use has also dose–response trend [3]. In an-
tively high incidence of oesopha- been shown to increase the risk of other analysis of total mortality in
geal cancer. In these areas, opium other malignancies, including can- the Golestan Cohort Study, 40%
has traditionally been used for rec- cers of the stomach, larynx, lung, of deaths among opium users and
reational purposes – in lieu of al- and bladder [1]. 10% of all deaths were attributable
cohol, which is strictly forbidden in The Golestan Cohort Study to opium use.
Islam – and as a medication to reis the only long-term prospec- There are at least two mecha-
lieve pain from chronic conditions. tive study that has detailed infor- nisms by which opium could cause
The first evidence that opium mation on opium use from large oesophageal squamous cell carci-
use may increase the risk of oe- numbers of participants. Of the noma [1]. Opium smoke and opium
sophageal squamous cell carci- cohort participants, 17% reported dross – the material left in the pipe
noma came from ecological and opium use, which is largely without after opium is smoked, which is
case–control studies of urinary me- negative social stigma. Over a me- sometimes eaten – contain car-
tabolites in north-eastern Islamic dian of 11 years of follow-up, 317 cinogenic pyrolysis products,
Republic of Iran in the early 1970s cases of oesophageal squamous including polycyclic aromatic hy-
[1]. A more recent case–control cell carcinoma were diagnosed. drocarbons, heterocyclic amines,
study of 300 cases of oesophageal Compared with participants who and N-nitrosamines. Some opium
squamous cell carcinoma and 571 had never smoked opium, those constituents can prolong exposure
neighbourhood controls found an in the highest tertile of cumula- of the oesophagus to ingested
odds ratio of 2.00 (95% confidence tive opium smoking had a haz- carcinogens: papaverine reduces
interval, 1.39–2.88) for ever use ard ratio of 1.85 (95% confidence oesophageal peristalsis, and mor-
phine inhibits relaxation of the low-
er oesophageal sphincter.
Fig. B5.3.1. Opium use in the Central Asia oesophageal cancer belt. Global map
showing ranking of opiates in order of prevalence among most commonly used References
drugs, in 2004. Inset: “tears” of the opium poppy. 1. Kamangar F, Shakeri R, Malekzadeh R,
Islami F (2014). Opium use: an emerging
risk factor for cancer? Lancet Oncol. 15(2):
e69–77. https://doi.org/10.1016/S1470-
2045(13)70550-3 PMID:24480557
2. Nasrollahzadeh D, Kamangar F, Aghcheli

K, Sotoudeh M, Islami F, Abnet CC, et
al. (2008). Opium, tobacco, and alcohol
use in relation to oesophageal squa-
mous cell carcinoma in a high-risk area
of Iran. Br J Cancer. 98(11):1857–63.
https://doi.org/10.1038/sj.bjc.6604369
PMID:18475303
3. Sheikh M, Poustchi H, Pourshams A,

Etemadi A, Islami F, Khoshnia M, et
CHAPTER 5.3
SECTION 5
al. (2019). Individual and combined
effects of environmental risk factors
for esophageal cancer based on re-
sults from the Golestan Cohort Study.
Gastroenterology. 156(5):1416–27. https://
doi.org/10.1053/j.gastro.2018.12.024
PMID:30611753
A case–control study in the trols, reported odds ratios of more carcinoma, but confirmation in pro-
Islamic Republic of Iran, which than 25 for the most exposed quin- spective studies is required.
measured exposure to PAHs in en- tile compared with the least exposed Nitrosamines are probably anoth-
doscopically normal oesophageal quintile [19]. This finding strongly er important carcinogen for oesoph-
tissues from cases of oesophageal implicates PAHs in the carcinogen- ageal squamous cell carcinoma.
squamous cell carcinoma and con- esis of oesophageal squamous cell They are an important carcinogen in

tobacco smoke, and they are thought low selenium status is also known agus are the established precursor
to be the main factor contributing to to combine with other exposures lesions for oesophageal squamous
the increased risk of oesophageal (especially viral infections) to cause cell carcinoma and oesophageal
squamous cell carcinoma associat- novel diseases that require both ex- adenocarcinoma, respectively, but
ed with poor oral health and the con- posures, as in Keshan disease [20], most of these tumours are diag-
sumption of non-piped water [12,13]. so it may also be important for the nosed in patients without a prior di-
Further studies to identify sources of oesophageal carcinogenicity of oth- agnosis of these precursor lesions
and routes of exposure to PAHs and er exposures. Further studies are [1,3]. Endoscopic screening for pre-
nitrosamines are needed to confirm needed to confirm the association cursor lesions and endoscopic re-
their role in the etiology of oesopha- of low selenium status and oesoph- section or ablation of the dysplastic
geal squamous cell carcinoma and ageal squamous cell carcinoma in lesions have been shown to reduce
to translate the knowledge of these Africa, and to explore how low se- the risk of developing oesophageal
associations into strategies for pri- lenium status and other risk factors squamous cell carcinoma and dy-
mary prevention in regions with high interact to increase risk of oesopha- ing from the disease [23]. A large
incidence of oesophageal squamous geal squamous cell carcinoma. trial is now under way.
cell carcinoma. Screening for Barrett oesopha-
Oesophageal gus has been used in clinics on
Low selenium status adenocarcinoma an individual basis in high-income
Another risk factor for oesophageal The main etiological factors for countries, but no randomized con-
squamous cell carcinoma that de- oesophageal adenocarcinoma are trolled trials have shown a signifi-
serves special attention is low sele- similar across the world and in- cant benefit [3]. Population-based
nium status. The selenium content clude gastro-oesophageal reflux endoscopic screening will require
of soil is variable worldwide, and disease, obesity (especially vis- well-trained health workers with di-
soil selenium levels are reflected in ceral obesity), tobacco use, and verse skills as well as considerable
local plants and animals as well as genetic risk factors [3,21]. People infrastructure; these are not widely
in people, assuming that they eat who have never been infected with available, especially in low- and
local foods. Helicobacter pylori also appear to middle-income countries, where
In both China and Africa, there be at elevated risk of oesophageal most cases of oesophageal squa-
are suggestive similarities in the dis- adenocarcinoma. Several recent mous cell carcinoma occur.
tribution of low selenium availability studies have suggested that sex Non-endoscopic screening of
(low soil selenium levels in China hormones, physical activity, certain oesophageal cells obtained with
and low dietary intake of selenium medications, and diet may also play balloon or sponge samplers and
in Africa) and the high-risk areas for a role in altering the risk of oesoph- molecular biomarker identification
oesophageal squamous cell carci- ageal adenocarcinoma [21]. of precursor lesion cells are now
noma [1]. In addition, cohort studies The markedly higher risk in men being evaluated for early detection
in both China and the Netherlands compared with women (up to 6-fold) of Barrett oesophagus and oesoph-
have shown significant inverse as- and in Whites compared with Blacks ageal adenocarcinoma in Europe
sociations between low serum or (up to 8-fold) cannot be explained [24] and for early detection of
toenail selenium levels and risk of by any of the confirmed risk factors, squamous dysplasia and oesopha-
oesophageal squamous cell carci- although visceral obesity, which is geal squamous cell carcinoma in
noma [12], and two intervention trials more common in men and is more the Islamic Republic of Iran [25],
in China have reported results sug- strongly associated with oesopha- with promising preliminary results.
gesting that selenium supplementa- geal adenocarcinoma, may contrib- However, further randomized con-
tion may be able to prevent oesoph- ute to the sex difference. Age–peri- trolled trials or well-conducted, ac-
ageal squamous cell carcinoma in od–cohort analyses suggest that a curate studies are required before
populations with low selenium status change in exposures in about 1950 these procedures can be recom-
when it is given early in the course of may have started the subsequent mended for implementation outside
the disease (see Chapter 6.4) [12]. rapid increase in oesophageal ad- of research studies.
Low selenium status is not an enocarcinoma rates in high-income Sampling of blood or other body
important risk factor for oesopha- countries [22]. fluids (referred to as liquid biopsies)
geal squamous cell carcinoma in to measure tumour-derived material
all high-risk populations, and spe- is also being evaluated for its poten-
cifically it is not a risk factor in the Early detection tial in early detection (see Chapter
Islamic Republic of Iran, but it is the Detection of oesophageal cancer at 6.7), including interrogation of cir-
only suspected risk factor in China an earlier, potentially curable stage culating cell-free tumour DNA, cir-
and Africa that is not commonly pres- of disease is critical to improve pa- culating tumour cells, exosomes,
ent in the low-risk populations of tient survival. Oesophageal squa- and microRNAs [26]. For example,
these regions as well. Furthermore, mous dysplasia and Barrett oesoph- one recent study investigated the
328
Drinking hot beverages
Many observational studies have was asked to sip the tea and say Tanzania, in the African corridor of
found an association between whether that was the temperature high risk of oesophageal squamous
drinking hot beverages and the at which they usually drank tea. If cell carcinoma, 62% of the par-
development of oesophageal squa- not, the tea was allowed to cool ticipants drank milky tea (or chai),
mous cell carcinoma [1]. The IARC further and the question was asked which is common in East Africa and
Monographs classified drinking again at 5 °C intervals until the rel- is made by boiling black tea leaves
very hot beverages at above 65 °C evant temperature was reached. and equal amounts of cow’s milk and
as probably carcinogenic to hu- At baseline, the cohort drank water, and 37% drank black tea. The
mans (Group 2A), especially for a mean tea volume of 1179 mL/ same protocol as in the Golestan
oesophageal squamous cell car- day, at a mean temperature of Cohort Study was used. Participants
cinoma. However, nearly all of the 62.4 °C. After a median follow- started drinking tea at a mean tem-
relevant studies were question- up of 10 years, 328 cases of oe- perature of 70.6 °C, and those who
naire-based studies that analysed sophageal cancer (96% of them consumed milky tea drank their tea
only subjective estimates of bever- oesophageal squamous cell carci- an average of 1.9 °C hotter than
age temperatures. noma) were diagnosed. Compared those who drank black tea [3].
The first large study to mea- with drinking less than 700 mL/day Thermal injury may increase
sure actual beverage tempera- of tea at less than 60 °C, drinking risk of oesophageal cancer by
tures was the Golestan Cohort 700 mL/day or more of tea at 60 °C inducing inflammatory process-
Study of 50 000 adults in Golestan or above was associated with a es. Formation of carcinogenic
Province, in north-eastern Islamic 75% higher risk of oesophageal N-nitroso compounds may be rele-
Republic of Iran. In this study, a cancer; drinking any amount of tea vant. Thermal injury may impair the
fresh cup of tea was prepared for at less than 60 °C was not associ- barrier function of the oesophageal
each participant, and the tempera- ated with risk [2]. mucosa, thereby increasing expo-
ture was measured. When the tem- In a cross-sectional study of 188 sure to intraluminal carcinogens
perature was 75 °C, the participant villagers in rural United Republic of such as N-nitroso compounds and
polycyclic aromatic hydrocarbons.
Fig. B5.3.2. Measurement of the temperature of tea, in the Golestan Cohort Study. References
Inset: close-up of the measuring device. 1. Islami F, Boffetta P, Ren JS, Pedoeim
L, Khatib D, Kamangar F (2009). High-
temperature beverages and foods and
esophageal cancer risk – a systematic
review. Int J Cancer. 125(3):491–524.
h t t p s : / / d o i . o r g / 1 0 .1 0 0 2 / i j c . 2 4 4 4 5
PMID:19415743
2. Islami F, Poustchi H, Pourshams A,

Khoshnia M, Gharavi A, Kamangar F,
et al. (2020). A prospective study of
tea drinking temperature and risk of
esophageal squamous cell carcinoma.
146(1):18–25. https://doi.org/10.1002/ijc.
32220 PMID:30891750
CHAPTER 5.3
SECTION 5
3. Munishi MO, Hanisch R, Mapunda O,
Ndyetabura T, Ndaro A, Schüz J, et al.
(2015). Africa’s oesophageal cancer
corridor: do hot beverages contribute?
Cancer Causes Control. 26(10):1477–
86. https://doi.org/10.1007/s10552-015-
0646-9 PMID:26245249
use of next-generation sequencing of this malignancy in a population Prevention

of a multigene panel to detect mu- at high risk [27]. These evaluations
tations in cell-free DNA in plasma of body fluids are only beginning, Reduced exposure to
samples to identify biomarkers of and many studies will be needed carcinogens
oesophageal cancer that could be to identify markers and to develop The translation of epidemiological
clinically useful for early detection protocols that have clinical utility. studies into preventive strategies –

such as prevention of tobacco use, housing, use of natural gas instead or oesophageal adenocarcinoma
smoking cessation, moderation of of biomass for cooking and heating [32]. However, emerging data sug-
alcohol consumption, weight loss, (resulting in the elimination of indoor gest that a comprehensive assess-
and modification of diet – is promis- air pollution from biomass smoke), ment of the health effects of proton-
ing but is difficult to accomplish [3]. and use of piped water instead of pump inhibitors is critical to assess
However, it should be possible to non-piped cistern water (preventing the overall effects of these agents.
reduce exposure to several risk fac- exposure to high concentrations of Aspirin and other non-steroidal
tors for oesophageal squamous cell nitrosamines) [29]. In 1970, fewer anti-inflammatory drugs have also
carcinoma by relatively straight- than 5% of people in the rural ar- been shown in observational stud-
forward interventions. Finland was eas had refrigerators; this propor- ies to be associated with reduced
able to eliminate the low selenium tion has now increased to more risk, by up to 50%, of oesopha-
status of its population by inexpen- than 98%, enabling better food geal squamous cell carcinoma
sive supplementation of chemical storage and decreased consump- and oesophageal adenocarcinoma
fertilizers [28]. Indoor air pollution tion of salted and smoked foods. [33]. A meta-analysis of 13 stud-
from coal or wood fires can be In addition, electricity, telephone ies showed a reduction of 28%
reduced by improving room ven- communication, and transporta- overall in the risk of oesophageal
tilation, replacing open fires with tion networks are now available to adenocarcinoma among users of
stoves, and adding chimneys to 98% of the population in the urban statins, compared with non-users,
stoves. Exposure to nitrosamines areas and 92% in the rural areas
and a reduction of 41% in the risk
can probably be reduced by cam- [29]. These dramatic changes in
of oesophageal adenocarcinoma in
paigns to encourage tooth brushing living standards in Golestan are
patients with Barrett oesophagus
and by increasing the availability of probably the main reasons for the
who took statins [3]. Given the ad-
treated water. sharp decrease in incidence rates
ditional preventive benefits of use
A comprehensive way to reduce of oesophageal squamous cell car-
of aspirin and statins for other can-
many of these harmful exposures, cinoma [29].
cer types and for cardiovascular
and hence rates of oesophageal
Cancer management in disease, these drugs may be good
squamous cell carcinoma, may be
groups at high risk candidates for chemoprevention in
to improve living standards and the
groups at high risk.
socioeconomic status of the popu- In high-risk regions in the Islamic
lation. This appears to be what has Republic of Iran and China, the avail- Several large trials examining
happened in north-eastern Islamic ability of free endoscopy services for the effects of proton-pump inhibi-
Republic of Iran over the past sev- early diagnosis and of therapeutic tors, aspirin, and statins for preven-
eral decades. In 1968–1971, the capabilities including endoscopic tion of oesophageal cancer are in
age-standardized incidence of oe- therapy, surgery, radiotherapy, and progress [3]. Recent results from
sophageal cancer in what is now chemotherapy have resulted in a randomized trial of protein-pump
Golestan Province was estimated much better care for patients with inhibitors and aspirin in Barrett
to be 80 per 100 000 in both sexes oesophageal squamous cell carci- oesophagus patients without high-
[29]. A retrospective study of cases noma, including improved survival grade dysplasia showed a sig-
in the same area in 1996–2000 re- and better quality of life after treat- nificant reduction in a combined
ported rates of 44 per 100 000 in ment. In more resource-limited set- end-point of death, oesophageal
men and 36 per 100 000 in women tings, oesophageal stents can pro- adenocarcinoma, or high-grade
[29], and the prospective Golestan vide significant palliation [31]. dysplasia in patients taking high-
Population-Based Cancer Registry dose proton-pump inhibitors, com-
reported rates in 2004–2008 of Chemoprevention pared with those taking low-dose
24 per 100 000 in men and 19 per Several clinical cohort studies have proton-pump inhibitors, and there
100 000 in women [30]. shown that use of proton-pump in- was some evidence that adding
During the 40 years between hibitors can significantly reduce the aspirin improved the beneficial ef-
1968 and 2008, living standards risk of progression from Barrett oe- fect of the high-dose proton-pump
improved significantly, with better sophagus to high-grade dysplasia inhibitors regimen [34].
330
References
1. Murphy G, McCormack V, Abedi-Ardekani 10. Abedi-Ardekani B, Kamangar F, Sotoudeh 18. Kayamba V, Heimburger DC, Morgan
B, Arnold M, Camargo MC, Dar NA, et al. M, Villar S, Islami F, Aghcheli K, et al. DR, Atadzhanov M, Kelly P (2017).
(2017). International cancer seminars: (2011). Extremely high Tp53 mutation load Exposure to biomass smoke as a risk
a focus on esophageal squamous cell in esophageal squamous cell carcinoma factor for oesophageal and gastric can-
carcinoma. Ann Oncol. 28(9):2086–93. in Golestan Province, Iran. PLoS One. cer in low-income populations: a system-
https://doi.org/10.1093/annonc/mdx279 6(12):e29488. https://doi.org/10.1371/jour atic review. Malawi Med J. 29(2):212–7.
PMID:28911061 nal.pone.0029488 PMID:22216294 ht tps://doi.org /10.4314/mmj.v 29 i2. 25
2. Arnold M, Soerjomataram I, Ferlay J, PMID:28955435
11. Liu W, Snell JM, Jeck WR, Hoadley KA,
Forman D (2015). Global incidence of
Wilkerson MD, Parker JS, et al. (2016). 19. Abedi-Ardekani B, Kamangar F, Hewitt SM,
oesophageal cancer by histological sub-
Subtyping sub-Saharan esophageal Hainaut P, Sotoudeh M, Abnet CC, et al.
type in 2012. Gut. 64(3):381–7. https://
squamous cell carcinoma by compre- (2010). Polycyclic aromatic hydrocarbon
d o i .o r g /10 .113 6 /g u t j n l - 2 014 - 3 0 8124
PMID:25320104 hensive molecular analysis. JCI Insight. exposure in oesophageal tissue and risk
1(16):e88755. https://doi.org/10.1172/jci. of oesophageal squamous cell carcinoma
3. Rustgi AK, El-Serag HB (2014). insight.88755 PMID:27734031 in north-eastern Iran. Gut. 59(9):1178–83.
Esophageal carcinoma. N Engl J Med. https://doi.org/10.1136/gut.2010.210609
371(26):2499–509. https://doi.org/10.1056/ 12. Abnet CC, Arnold M, Wei WQ (2018). PMID:20584779
NEJMra1314530 PMID:25539106 Epidemiology of esophageal squa-
mous cell carcinoma. Gastroenterology. 20. Beck MA, Levander OA, Handy J (2003).
4. Lin DC, Dinh HQ, Xie JJ, Mayakonda Selenium deficiency and viral infec-
154(2):360–73. https://doi.org/10.1053/j.
A, Silva TC, Jiang YY, et al. (2018). tion. J Nutr. 133(5 Suppl 1):1463S–7S.
gastro.2017.08.023 PMID:28823862
Identification of distinct mutational pat- https://doi.org/10.1093/jn/133.5.1463S
terns and new driver genes in oesopha- 13. Golozar A, Etemadi A, Kamangar F, PMID:12730444
geal squamous cell carcinomas and Fazeltabar Malekshah A, Islami F,
adenocarcinomas. Gut. 67(10):1769–79. Nasrollahzadeh D, et al. (2016). Food 21. Coleman HG, Xie SH, Lagergren J
https://doi.org/10.1136/gutjnl-2017-314607 preparation methods, drinking water (2018). The epidemiology of esopha-
PMID:28860350 geal adenocarcinoma. Gastroenterology.
source, and esophageal squamous
cell carcinoma in the high-risk area of 154(2):390–405. https://doi.org/10.1053/j.
5. Parker RK, Dawsey SM, Abnet CC,
White RE (2010). Frequent occurrence Golestan, Northeast Iran. Eur J Cancer gastro.2017.07.046 PMID:28780073
of esophageal cancer in young peo- Prev. 25(2):123–9. https://doi.org/10.1097/ 22. Edgren G, Adami HO, Weiderpass E,
ple in western Kenya. Dis Esophagus. CEJ.0000000000000156 PMID:25851181 Nyrén O (2013). A global assessment
23(2):128–35. https://doi.org/10.1111/j.1442-
14. Islami F, Poustchi H, Pourshams A, of the oesophageal adenocarcinoma
2050.2009.00977.x PMID:19473205
Khoshnia M, Gharavi A, Kamangar F, et al. epidemic. Gut. 62(10):1406–14. https://
6. Wu C, Wang Z, Song X, Feng XS, Abnet (2020). A prospective study of tea drink- d o i . o r g /10 .113 6 /g u t j n l - 2 012- 3 0 2 412
CC, He J, et al. (2014). Joint analysis of ing temperature and risk of esophageal PMID:22917659
three genome-wide association stud- squamous cell carcinoma. Int J Cancer.
ies of esophageal squamous cell carci- 23. Wei WQ, Chen ZF, He YT, Feng H, Hou J,
146(1):18–25. https://doi.org/10.1002/ijc. Lin DM, et al. (2015). Long-term follow-up of
noma in Chinese populations. Nat Genet.
32220 PMID:30891750 a community assignment, one-time endo-
46(9):1001–6. https://doi.org/10.1038/ng.
3064 PMID:25129146 15. Islami F, Kamangar F, Nasrollahzadeh D, scopic screening study of esophageal can-
Aghcheli K, Sotoudeh M, Abedi-Ardekani cer in China. J Clin Oncol. 33(17):1951–7.
7. Levine DM, Ek WE, Zhang R, Liu X, https://doi.org/10.1200/JCO.2014.58.0423
Onstad L, Sather C, et al. (2013). A ge- B, et al. (2009). Socio-economic status
PMID:25940715
nome-wide association study identifies and oesophageal cancer: results from
new susceptibility loci for esophageal ad- a population-based case-control study 24. Ross-Innes CS, Debiram-Beecham I,
enocarcinoma and Barrett’s esophagus. in a high-risk area. Int J Epidemiol. O’Donovan M, Walker E, Varghese S,
Nat Genet. 45(12):1487–93. https://doi. 38(4):978–88. https://doi.org/10.1093/ije/ Lao-Sirieix P, et al.; BEST2 Study Group
org/10.1038/ng.2796 PMID:24121790 dyp195 PMID:19416955 (2015). Evaluation of a minimally invasive
cell sampling device coupled with as-
8. Plum PS, Gebauer F, Krämer M, Alakus 16. Sheikh M, Poustchi H, Pourshams A,
H, Berlth F, Chon SH, et al. (2019). HER2/ sessment of trefoil factor 3 expression for
Etemadi A, Islami F, Khoshnia M, et al.
neu (ERBB2) expression and gene am- diagnosing Barrett’s esophagus: a multi-
(2019). Individual and combined effects
CHAPTER 5.3
SECTION 5
plification correlates with better survival center case-control study. PLoS Med.
of environmental risk factors for esopha-
in esophageal adenocarcinoma. BMC 12(1):e1001780. https://doi.org/10.1371/
geal cancer based on results from the
Cancer. 19(1):38. https://doi.org/10.1186/ journal.pmed.1001780 PMID:25634542
Golestan Cohort Study. Gastroenterology.
s12885-018-5242-4 PMID:30621632 156(5):1416–27. https://doi.org/10.1053/j. 25. Roshandel G, Merat S, Sotoudeh M,
9. Bang YJ, Van Cutsem E, Feyereislova A, gastro.2018.12.024 PMID:30611753 Khoshnia M, Poustchi H, Lao-Sirieix P, et
Chung HC, Shen L, Sawaki A, et al.; ToGA al. (2014). Pilot study of cytological testing
17. Etemadi A, Poustchi H, Chang CM, Blount
Trial Investigators (2010). Trastuzumab for oesophageal squamous cell dyspla-
in combination with chemotherapy ver- BC, Calafat AM, Wang L, et al. (2019).
sia in a high-risk area in Northern Iran.
sus chemotherapy alone for treatment of Urinary biomarkers of carcinogenic ex-
Br J Cancer. 111(12):2235–41. https://doi.
HER2-positive advanced gastric or gastro- posure among cigarette, waterpipe, and
org/10.1038/bjc.2014.506 PMID:25247319
oesophageal junction cancer (ToGA): a smokeless tobacco users and never us-
phase 3, open-label, randomised controlled ers of tobacco in the Golestan Cohort 26. Perakis S, Speicher MR (2017). Emerging
trial. Lancet. 376(9742):687–97. https:// Study. Cancer Epidemiol Biomarkers Prev. concepts in liquid biopsies. BMC Med.
doi.org/10.1016/S0140-6736(10)61121-X 28(2):337–47. https://doi.org/10.1158/1055- 15(1):75. https://doi.org/10.1186/s12916-
PMID:20728210 9965.EPI-18-0743 PMID:30622099 017-0840-6 PMID:28381299

27. Lan YT, Chen MH, Fang WL, Hsieh CC, 30. Roshandel G, Sadjadi A, Aarabi M, 33. Qiao Y, Yang T, Gan Y, Li W, Wang C,
Lin CH, Jhang FY, et al. (2017). Clinical Keshtkar A, Sedaghat SM, Nouraie Gong Y, et al. (2018). Associations be-
relevance of cell-free DNA in gastroin- SM, et al. (2012). Cancer incidence in tween aspirin use and the risk of can-
testinal tract malignancy. Oncotarget. Golestan Province: report of an ongoing cers: a meta-analysis of observational
8(2):3009–17. https://doi.org/10.18632/ population-based cancer registry in Iran studies. BMC Cancer. 18(1):288. https://
oncotarget.13821 PMID:27936467 d o i.o r g /10 .118 6 /s12 8 8 5 - 018 - 415 6 - 5
between 2004 and 2008. Arch Iran Med.
28. Alfthan G, Eurola M, Ekholm P, 15(4):196–200. PMID:22424034 PMID:29534696
Venäläinen ER, Root T, Korkalainen K,
31. White RE, Parker RK, Fitzwater JW, 34. Jankowski JAZ, de Caestecker J, Love
et al.; Selenium Working Group (2015).
Kasepoi Z, Topazian M (2009). Stents as SB, Reilly G, Watson P, Sanders S, et al.;
Effects of nationwide addition of selenium
to fertilizers on foods, and animal and hu- sole therapy for oesophageal cancer: a pro- AspECT Trial Team (2018). Esomeprazole
man health in Finland: from deficiency to spective analysis of outcomes after place- and aspirin in Barrett’s oesophagus
optimal selenium status of the population. ment. Lancet Oncol. 10(3):240–6. https:// (AspECT): a randomised factorial trial.
J Trace Elem Med Biol. 31:142–7. https:// doi.org/10.1016/S1470-2045(09)70004-X Lancet. 392(10145):400–8. https://doi.
d o i.o r g /10.1016 / j.jte mb. 2 014.0 4.0 0 9 PMID:19230771 org /10.1016/S0140 - 673 6(18) 313 8 8 - 6
PMID:24908353 PMID:30057104
32. Singh S, Garg SK, Singh PP, Iyer PG,
29. Semnani S, Sadjadi A, Fahimi S, Nouraie M,
El-Serag HB (2014). Acid-suppressive
Naeimi M, Kabir J, et al. (2006). Declining
medications and risk of oesophageal ad-
incidence of esophageal cancer in the
Turkmen Plain, eastern part of the Caspian enocarcinoma in patients with Barrett’s
Littoral of Iran: a retrospective cancer sur- oesophagus: a systematic review and
veillance. Cancer Detect Prev. 30(1):14–9. meta-analysis. Gut. 63(8):1229–37. https://
https://doi.org/10.1016/j.cdp.2005.11.002 d o i .o r g /10 .113 6 /g u t j n l - 2 013 - 3 0 5 9 97
PMID:16495018 PMID:24221456
332
5.4
Stomach cancer
Still one of the main cancer types
worldwide
Christine Varon Rolando Herrero (reviewer)
Francis Mégraud Wenbo Meng (reviewer)
Liang Qiao (reviewer)
immune checkpoint inhibitors in (in the proximal or distal part of the

SUMMARY resectable gastric cancer. New stomach), and histological subtype.
cellular markers are putative The 2010 WHO classification of
●● Two systematic reviews and biomarkers for diagnosis and gastric cancer specifies five main
meta-analyses have been per- therapeutic targets. histological subtypes: tubular, pa-
formed of the worldwide preva-
pillary, mucinous, poorly cohesive
lence of Helicobacter pylori infec-
(including signet ring cell carcino-
tion, the main (necessary but not
In the 19th century, stomach can- ma), and mixed. Tubular, papillary,
sufficient) risk factor for gastric
cer was one of the major causes and mucinous adenocarcinomas
cancer. The global prevalence in
of cancer-related death. The situ- correspond to the intestinal type
adults is close to 50%, with large
ation changed in the 20th century described by Laurén in 1965, and
differences between continents
in high-income countries after an poorly cohesive carcinomas corre-
and a trend towards a decrease
over the years. improvement in the socioeconom- spond to the diffuse type of Laurén
ic status of the populations and the (Table 5.4.1).
●● A recent emergence of gastric introduction of antibiotics. However, Although both the intestinal and
cancer possibly not related to stomach cancer is still an important diffuse types of gastric cancer are
H. pylori in younger patients cause of death in many countries. related mainly to H. pylori infec-
should be explored. The breakthrough in under- tion, the intestinal type is often re-
●● The Stomach Cancer Pooling standing the causation of stomach lated to environmental factors, diet,
Project, by using individual data, cancer was the discovery that a and lifestyle, and the diffuse type
confirmed the role of additional bacterium – Helicobacter pylori –
risk factors such as tobacco was the main causal agent of this
smoking and alcohol consump- disease. The role of H. pylori was Fig. 5.4.1. Scanning electron micrograph
tion but at a lower magnitude determined by Warren and Marshall of Helicobacter pylori bacterium.
than previously established. in 1982, and they subsequently de-
scribed its role in the development
● ● Among emerging risk fac-
of peptic ulcer disease. For this
tors, a modified composition
CHAPTER 5.4
SECTION 5
discovery, Warren and Marshall
of the gastric microbiota may
contribute to gastric carcino- were awarded the Nobel Prize in
genesis by increasing inflam- Physiology or Medicine 2005. The
mation and producing carci- IARC Monographs classified infec-
nogenic compounds. tion with H. pylori as carcinogenic
to humans (Group 1) in 1994, on the
●● The molecular profiles of gastric basis of epidemiological evidence
cancer were recently identified, [1], and this classification was con-
and two molecular classifica- firmed in 2009 [2].
tions are based on sequencing; Stomach cancers, often referred
these will provide a roadmap for to as gastric cancers, are mostly
trials of targeted therapies.
gastric adenocarcinomas. They are
●● New treatments are being pro- classified according to stage (early
posed, especially those using or advanced), anatomical location
Chapter 5.4 • Stomach cancer 333

is more often associated with ge- the intestinal and diffuse types [4].
netic abnormalities. The molecular In addition, in African populations,
FUNDAMENTALS
profiles of gastric cancer were re- parasitic infections that drive the
cently identified and classified by immune response appear to be ■■ Stomach cancer is the third
the Cancer Genome Atlas (TCGA) beneficial (i.e. Th2 response rather most common cause of cancer
Research Network and the Asian than Th1 response), leading to less death worldwide.
Cancer Research Group (ACRG). inflammation [5]. Because gastric
cancer typically occurs later in life, ■■ Stomach cancers, often
the shorter life expectancy of popu- referred to as gastric
Epidemiology cancers, are mostly gastric
lations in many African countries
The incidence of gastric cancer is also contributes to the low rate of adenocarcinomas. They are
still high, and it is the third most gastric cancer in these populations. classified according to stage
common cause of cancer death Two systematic reviews and me- (early or advanced), anatomical
worldwide, responsible for an es- ta-analyses of the worldwide preva- location (in the proximal or
timated 783 000 deaths in 2018 lence of H. pylori infection were distal part of the stomach), and
[3]. However, there is consider- published in 2017 and 2018. Hooi et histological subtype.
able geographical heterogeneity. al. covered the period 1970–2016
The countries with the highest inci- and 62 countries (531 880 subjects) ■■ The principal cause of gastric
dence rates are in East Asia, and [6], whereas Zamani et al. analysed cancer is infection with the
incidence rates in men are much the period 2000–2017 and 73 coun- bacterium Helicobacter pylori,
higher than those in women. tries (410 879 subjects) (Fig. 5.4.2) which is particularly prevalent
Infection with H. pylori is a nec- [7]. Both studies showed the same in Africa, Latin America, and
essary but not sufficient cause; global prevalence of H. pylori infec- Asia. Infection with H. pylori is
this explains why the incidence of tion in adults (48.5% and 48.6%, a necessary but not sufficient
gastric cancer does not mirror the respectively). The prevalence was cause.
prevalence of H. pylori infection. It highest in Africa, followed by Latin
is now well known that the impor- ■■ Decreases in the incidence
America and Asia, and the preva-
tant risk factors are the host’s ge- of stomach cancer over the
lence was lowest in Australia, North
netic makeup, the characteristics of decades before the role of H.
America, and western Europe.
H. pylori strains, and environmental pylori was known have been
However, large differences were
factors, notably diet. correlated with environmental
observed between countries on the
People in East Asia harbour ag- same continent and between areas factors such as type of diet,
gressive strains of H. pylori, have within large countries. There was a i.e. decreased consumption of
a diet that is high in salt, and may trend towards a decrease in preva- salt-preserved food, avoidance
have genetic elements that favour lence in 2009–2016 compared with of a diet that is high in salt, and
the development of gastric cancer, 2000–2009 [7]. availability of fresh fruits and
whereas people in Africa harbour Several relevant studies have vegetables throughout the year.
less aggressive strains of H. pylori been performed in East Asia. In the ■■ Patients with stomach cancer
and generally have a diet that in- Republic of Korea, the prevalence are often diagnosed with
cludes more vitamins and less salt. of H. pylori infection, determined in advanced disease, and
Recently, a dietary inflammatory 4920 asymptomatic subjects by se-
survival is poor.
index was calculated for partici- rology, was 51.0%. The prevalence
pants in the European Prospective decreased progressively from 1998
Investigation into Cancer and to 2005, 2011, and 2015–2016.
Nutrition (EPIC) study. The results Interestingly, the prevalence was the urea breath test found a 34.4%
showed that the inflammatory po- lower in urban areas than in rural prevalence of H. pylori infection,
tential of the diet was associated areas [8]. In south-western China, and an association was noted with
with the risk of gastric cancer, but a cross-sectional study carried out low albumin levels and hyperglycae-
no differences were seen between in 2014 on 10 912 subjects using mia [9]. In Viet Nam, the observed
Table 5.4.1. Histological subtypes of gastric adenocarcinoma according to the Laurén classification and the WHO classification
Classification Histological subtype of gastric adenocarcinoma
Laurén (1965) Intestinal Diffuse Mixed Indeterminate
WHO (2010) Tubular Papillary Mucinous Poorly cohesive Mixed Uncommon variants
334
Fig. 5.4.2. Prevalence of Helicobacter pylori infection for adults (A) and children (C) than 50 years (Fig. 5.4.3). This in-
across six continents. The reference line represents the overall global prevalence crease was more marked in women
(44.3%).
than in men; the incidence in wom-
en born in 1983 was double that in
Africa (A)
those born 30 years earlier. These
Africa (C) data were collected from registries
Northern America (A) where there was no information on
H. pylori infection status, but given
Northern America (C)
the socioeconomic status of these
Latin America & the Caribbean (A) cases and the predominant locali-
Continent & age group
Latin America & the Caribbean (C) zation of the tumours to the corpus
Asia(A) of the stomach, H. pylori infection is
unlikely to have played a role. One
Asia (C)
hypothesis is that gastric cancer in
Europe (A) these patients is the consequence
Europe (C) of autoimmunity related to dysbio-
sis of the gastric microbiome [16].
Oceania (A) No data
In an evaluation of trends in gas-
Oceania (C) tric cancer incidence, an increased
Overall (A) risk was also noted in recent birth
Overall (C)
cohorts in several countries in
South America and Europe, for both
0 20 40 60 80 100 men and women [17]. This change,
Infection rate (%) and 95% CI which is most likely to be related
to lifestyle and environmental risk
factors, needs to be explored fur-
prevalence was similar (38.1%), but tion in Japan was estimated to be ther. In a systematic review of the
it varied according to ethnicity [10]. 27% in 2016 [12], and the sponta- prevalence of H. pylori infection in
Mortality from gastric cancer was neous decrease has been boosted Europe, the prevalence was lowest
also studied in China (see Chapter by the eradication policy. The in- in northern Europe and highest in
4.3). When mortality rates were cidence of gastric cancer is also eastern and southern Europe. Two
standardized by the age scale of decreasing more rapidly since this countries still had a high preva-
the population in 2010, a 17.8% de- policy was implemented [13]. H. py- lence (84%): Poland and Portugal.
crease was observed between 2006 lori eradication reduces the cumu- Studies on the impact of lifestyle
and 2013, which is in line with the lative incidence of gastric cancer in indicated the usual risk factors for
global decrease in the prevalence a healthy asymptomatic population, gastric cancer [18].
of H. pylori infection during that pe- and the effect on the prevention of
riod. The age-standardized mortality gastric cancer is observed in all age
rate was higher in rural areas than in groups [14]. Genetics and genomics
urban areas. However, a surprising In the USA, a study of 11 mil-
Genetic susceptibility
finding was an increasing trend in lion patients investigated the preva-
mortality rates in young age groups lence of H. pylori infection in people Hereditary gastric cancer makes
(0–29 years) between 2006 and 2013 of five ethnic groups who had up- up about 1–3% of cases of gastric
[11]. In Mongolia, which has high per gastrointestinal symptoms. The cancer. It includes mainly heredi-
CHAPTER 5.4
SECTION 5
gastric cancer incidence and mortal- relative risk of gastric diseases as- tary diffuse gastric cancer, gastric
ity rates, the prevalence of H. pylori sociated with H. pylori infection was adenocarcinoma and proximal pol-
infection was 80.0%. Dyspepsia is highest in Blacks and Asian Pacific yposis of the stomach, and familial
common in this population, and the Islanders, and the prevalence of H. intestinal gastric cancer [18].
salty diet was considered to worsen pylori infection was highest in Native About 30–40% of cases of he-
the atrophy observed. Americans and Alaska Natives [15]. reditary diffuse gastric cancer are
In Japan, insurance coverage A study assessed incidence linked to a dominant germline patho-
for H. pylori eradication began in trends in 1995–2013 in the USA. genic mutation in CDH1, which en-
2000 for peptic ulcer disease and in There were 137  447 non-cardia codes E-cadherin. In whole-exome
2013 for gastritis, leading to eradi- gastric cancers in 4.4 billion person- sequencing studies, germline muta-
cation in about 650 000 patients per years of observation. An overall de- tions in the tumour suppressor genes
year from 2001 to 2012, and double cline in incidence rates was seen, CTNNA1, STK11, and SDHB and the
that number annually since 2013. but a slight increase was observed DNA repair-related genes PALB2,
The prevalence of H. pylori infec- in non-Hispanic Whites younger BRCA2, and ATM were identified

Fig. 5.4.3. Age-specific incidence trends of non-cardia gastric cancer among non- virus (EBV) (8.8%), which display
Hispanic White men (A) and women (B). The symbols represent the observed incidence recurrent PIK3CA mutations, ex-
rates in 15 4-year age groups over four 4-year time periods. The shaded areas denote treme DNA hypermethylation, and
95% confidence intervals from the age–period–cohort models. The modelled 95%
confidence intervals provide a good fit to the observed data for every age group except
amplification of JAK2, PD-L1, and
women aged 25–28 years. PD-L2; (ii) tumours with microsatel-
lite instability (MSI) (21.7%), which
A B have elevated mutation rates in on-
cogenes such as human epidermal
growth factor receptor 2 (HER2),
epidermal growth factor receptor 1
(EGFR1), and HER3 (also known as
ERBB3); (iii) genomically stable tu-
mours (19.7%), which are enriched
for the diffuse type and mutations
of CDH1, RHOA, and genes as-
sociated with the cytoskeleton and
cell junctions; and (iv) tumours with
chromosomal instability (49.8%),
which are of the intestinal type and
show marked aneuploidy, TP53 mu-
tations, and focal amplification of
RAS and receptor tyrosine kinases.
The EBV-positive subtype was as-
sociated with the most favourable
prognosis, followed by the MSI and
chromosomal instability subtypes.
ACRG reported a similar classi-
fication of gastric cancer and distin-
guished the following four molecular
subtypes (Fig. 5.4.4): (i) MSI hyper-
mutated tumours, which are of the
intestinal type and are mostly local-
ized to the antrum, and microsatellite
stable (MSS) tumours, subdivided
into (ii) those that exhibit features of
epithelial–mesenchymal transition
(MSS/EMT), which occur at a youn-
ger age and are mostly of the diffuse
type; (iii) those that lose p53 activity
(MSS/TP53−) and show amplification
of HER2 (ERBB2); and (iv) those with
wild-type TP53 (MSS/TP53+), which
are associated with EBV. The MSS/
EMT and MSS/TP53− gastric can-
cers had the poorest survival [23].
A recent meta-analysis con-
in hereditary diffuse gastric cancer Genomics firmed the prognostic value of histo-
without CDH1 mutation [19]. In 2014, by integrating whole-genome logical subtyping of gastric cancer,
Hereditary gastric cancer also sequencing, genomic data, and pro- showing that the diffuse subtype is
develops in patients with Lynch syn- teomic data, TCGA [21] and ACRG associated with younger patients and
drome (mutations in the mismatch [22] each defined four molecular sub- poorer prognosis than the intestinal
repair genes MSH2, MSH6, PMS2, types of gastric cancer, to provide a type [24]. According to the TCGA
or MLH1) and, more rarely, in pa- roadmap for patient stratification and and ACRG molecular classifications,
tients with Li–Fraumeni syndrome trials of targeted therapies. the genomically stable and MSS/
(TP53 germline mutation), Peutz– TCGA distinguished the fol- EMT subtypes, which are composed
Jeghers syndrome (STK11 muta- lowing four molecular subtypes of mostly of tumours of the diffuse type,
tion), and familial adenomatous pol- gastric cancer (Fig. 5.4.4): (i) tu- have the worst prognosis and overall
yposis (APC mutation) [20]. mours positive for Epstein–Barr survival (Fig. 5.4.4) [25,26,27].
336
Fig. 5.4.4. The main subtypes of gastric adenocarcinoma defined according to the Laurén histological classification and the
Cancer Genome Atlas (TCGA) Research Network and Asian Cancer Research Group (ACRG) molecular classifications. The global
distribution frequencies of gastric cancer subtypes are indicated as percentages. TCGA subtypes: MSI, microsatellite instability;
CIN, chromosomal instability; EBV+, positive for Epstein–Barr virus; GS, genomically stable. ACRG subtypes: MSI, microsatellite
instability; MSS/TP53−, microsatellite stable with inactive TP53; MSS/TP53+, microsatellite stable with active TP53; MSS/EMT,
microsatellite stable with features of epithelial–mesenchymal transition. For each subtype, the clinical characteristics and the main
genetic and molecular alterations are listed. mTOR, mammalian target of rapamycin; RTKs, receptor tyrosine kinases.
Laurén Diffuse type

histological Intestinal type (54%)
(32%, worst prognosis)
classification
CIN (49.8%)
EBV+ (8.8%)
MSI (21.7%) Moderate prognosis
Best prognosis and GS
Moderate prognosis Mostly intestinal
overall survival
Mostly intestinal Marked aneuploidy with
Younger patients,
(19.7%)
Older patients focal amplification Worst prognosis
mostly males
Elevated mutation (RTKs, KRAS, MYC, Mostly diffuse
Extreme DNA
TCGA molecular rate (PIK3CA, TP53, CCND1, PDL-2, Younger patients
hypermethylation
PTEN, RNF43, KRAS, PIK3CA, etc.) Fewer genomic
classification ERBB3, ARID1A, and deletion (CDH1,
(CDKN2A, etc.)
Mutations (PIK3CA, alterations
etc.) and RB, etc.) Low mutation rates
ARID1A, etc.)
hypermethylation TP53 mutations (71%) Mutations in ARID1A,
Amplification of
(MLH1, etc.) Less homogeneous RHOA, CDH1, etc.
JAK2, PD-L1, and
Activation of subgroup Alteration of cell
PD-L2
mitotic pathways Activation of RTKs and adhesion
Activation of
RAS pathways, immune
immune signalling
evasion
MSI MSS/EMT
MSS/TP53 MSS/TP53+
(22.7%) (15.3%)
(35.7%) (26.3%)
Best prognosis Worst prognosis
Intermediate Intermediate
(survival, 77.8 (survival, 42.6
prognosis (survival, prognosis (survival,
months) months)
59.8 months) 66.9 months)
ACRG molecular Mostly intestinal Mostly diffuse
Mostly intestinal Enriched in EBV+
classification Diagnosis at an early
Low mutation rate tumours (66%)
Younger patients
stage (50%) Lowest mutation rate
TP53 mutations Mutations in PIK3CA,
MLH1 loss Mutations in ARID1A
(60%) ARID1A, KRAS,
High mutation rate and CDH1 (loss)
Amplification of APC, etc.
(ARID1A, PIK3CA, Low cell adhesion
ERBB2, CCNE1, Amplification of
KRAS, mTOR pathway, and mesenchymal
MYC, and EGFR CCNE1
etc.) phenotype
Recent studies using integrated gastric adenocarcinomas – via the gastric biopsies by 16S ribosomal
bioinformatics analyses have led Correa cascade of multistep gastric DNA sequencing and compared the
to the proposal of a panel of genes carcinogenesis for the intestinal type microbiota of patients with gastritis,
CHAPTER 5.4
SECTION 5
that are associated with the patho- and by other mechanisms for the dif- precancerous lesions, and gastric
genesis of gastric cancer, the value fuse type – and that about 10% of cancer. A study in Singapore and
of adjuvant therapy, and the prog- gastric cancers are the consequence Malaysia compared cases of gastric
nosis of resectable gastric cancer of EBV infection. However, since cancer and controls with functional
[28,29]. There is a need for further the development of new molecular dyspepsia (n = 32) and found that
validation in prospective studies methods to study the microbiota (see patients with gastric cancer had
and for standardization of tools that Chapter 3.10), it has been shown that higher relative abundances of bac-
can be used in clinical practice to H. pylori is not the only bacterium terial species that are commonly
screen gene expression in tumours. that is found in the stomach, and the found in the oral cavity [30]. A study
question of the newly recognized role in Taiwan, China, compared patients
of the microbiota in gastric carcino- with gastritis, intestinal metaplasia,
Etiology genesis has emerged. and gastric cancer (n = 27) and found
It is now agreed that H. pylori infec- Recent studies, mainly in Asia, a gastric cancer-specific bacterial
tion is responsible for about 90% of have identified the microbiota from signature consisting of Clostridium

(mainly C. colicanis), Fusobacterium that there was a significant positive More studies are needed on pa-
(F. nucleatum), and Lactobacillus (L. correlation between expression of tient cohorts, on humanized animal
gasseri and L. reuteri) [31]. A study Deinococcus, Sulfurospirillum, and models, and on different populations;
in Xi'an, China, observed significant Campylobacter and H. pylori genes, also, elements of the microbiota oth-
microbial dysbiosis in cases of intes- especially those involved in pH regu- er than bacteria should be included,
tinal metaplasia and gastric cancer lation and nickel transport [34]. such as fungi, archaea, and viruses
compared with cases of superficial The main limitation of these [35]. A more in-depth knowledge of
gastritis only (n = 81) and highlighted studies is that they are cross-sec- the gastric microbiota in relation to
a group of five species of oral bac- tional and cannot reveal whether gastric cancer should help research-
teria that are associated with gastric the gastric microbiota described ers to develop strategies for reducing
cancer [32]. In contrast, a study in corresponds to bacteria that the burden of this disease.
Portugal of patients with chronic are resident or only transitory.
gastritis and with gastric carcinoma However, because high pH is an
Biological characteristics
(n = 135) found an enrichment of important determinant of bacterial
intestinal bacteria rather than oral colonization, it is logical to imagine
and early detection
bacteria, and these results were that these bacteria can colonize the Biomarkers
confirmed in validation cohorts in stomach in the case of atrophy and
Many biomarkers for gastric can-
China and Mexico (Fig. 5.4.5) [33]. intestinal metaplasia, which leads
cer diagnosis have been described,
A study in Nicaragua determined the to decreased acid production and is
including CA72-4, CA12-5, SLE,
presence of viable bacteria by meta- the outcome of long-term H. pylori
BCA-225, hCG, and the ratio be-
transcriptomic analysis of stomach infection. Once established, these
tween the levels of pepsinogen I
biopsy specimens from patients bacteria could contribute to carcino-
and II; the most frequently used
undergoing endoscopy (n = 25) genesis by increasing inflammation,
biomarkers in clinical practice are
and found that the gastric micro- producing N-nitroso compounds or
biota did not change in relation to CEA and CA19-9 [36]. Cellular het-
acetaldehyde, and also modifying
the level of atrophy in the tissue but erogeneity must be considered in
the physiology of the stomach.
research on biomarkers for early
detection, prognosis, and targeted
therapy. Cancer stem cells are a
Fig. 5.4.5. The influence of Helicobacter pylori in the microbiota composition of chronic
gastritis and gastric carcinoma. Relative abundance of the different bacterial phyla rare subpopulation of gastric can-
overall (i.e. in all patients), in patients with chronic gastritis only, and in patients with cer cells at the origin of tumour initi-
gastric carcinoma. NS, not significant. ation and progression [37]. Several
cell surface markers of gastric can-
100 Other cer stem cells have been identified
Fusobacteria using mouse models of patient-
80
Actinobacteria derived tumour xenografts, gastric
Relative abundance (%)
organoid culture, and transgenic

Bacteroidetes
60
mouse models. These markers in-
Firmicutes clude CD44, CD133, Lgr5, CD24,
Non-Helicobacter Proteobacteria CD166, and ALDH, all of which are
40 putative biomarkers for diagnosis
Helicobacter spp.
and therapeutic targets [25].
20 The pathogenesis of gastric can-
cer also involves epigenetic mecha-
nisms (see Chapter 3.8). Infection
0
Overall Chronic Gastric with H. pylori and EBV and the sub-
gastritis carcinoma sequent chronic inflammation all
participate in aberrant DNA methyla-
Taxa Chronic Gastric P value
carcinoma (%)
tion and more generally in this epi-
gastritis (%)
genetic dysregulation. The detection
Proteobacteria 68.8 70.2 NS
of CDH1 promoter methylation in
Helicobacter spp. 41.7 5.9 < 0.001 blood samples has been proposed
Non-Helicobacter Proteobacteria 27.1 64.3 < 0.001 as a diagnostic tool [38]. Other
Firmicutes 13.6 16.4 0.040 non-invasive biomarkers have been
Bacteroidetes 10.6 6.6 0.003
proposed for gastric cancer diag-
5.9
nosis and follow-up, including long
Actinobacteria 3.3 < 0.001
non-coding RNAs and small non-
Fusobacteria 1.8 0.5 < 0.001
coding RNAs such as microRNAs,
338
Table 5.4.2. Current topics of molecular markers associated with diagnosis, prognosis, and prediction of therapeutic response of
gastric cancer
Marker Alteration Clinical purpose Detection method

Metastasis-related genes
Growth factors
HER2, FGFR, PI3K/Akt/mTOR (PIK3CA), MET, VEGF Overexpression Diagnostic, prognostic, Tissue
(VEGFR2, VEGFD) therapeutic
Cell-cycle regulation
TP53 Mutation Diagnostic Tissue
Adhesion molecule
E-cadherin (CDH1) Mutation, epigenetic Diagnostic, prognostic Tissue, blood
alteration
Immune checkpoint
PD-L1 Mutation Prognostic, therapeutic Tissue
Comprehensive gene analysis
CEACEM6, APOC1, YF13H12, CDH17, REG4, OLFM4, Upregulation Diagnostic, prognostic, Tissue
HOXA10, DSC2, TSPAN8, TM9SF3, FUS, COLIA1, therapeutic
COLIA2, APOE
ATP4B, S100A9, CYP20A1, ARPC3, DDX5, CLDN18 Downregulation Diagnostic, prognostic, Tissue
therapeutic
Microsatellite instability High level Prognostic, therapeutic Tissue
Epigenetic alterations
CDH1, CHFR, DAPK, GSTP1, p15, p16, RARβ, RASSF1A, Hypermethylation Diagnostic Tissue
RUNX3, TFPI2
Genetic polymorphism
IL-1β, IL-1RN, CD44 Single-nucleotide Prognostic Tissue
polymorphism
TP53, SYNE1, CSMD3, LRP1B, CDH1, PIK3CA, ARID1A, Copy number variations, Diagnostic, prognostic, Tissue
PKHD, KRAS, JAK2, CD274, PDCD1LG2 mutations therapeutic
Circulating tumour cells
CD44, N-cadherin, vimentin Overexpression Diagnostic, therapeutic Blood
pan-CK, E-cadherin Decreased expression EMT process Blood
HER2 Overexpression Therapeutic Blood
Circulating cell-free DNA
APC promotor 1, RASSF1A Hypermethylation Diagnostic Blood, plasma
ERBB2 Copy number variations Therapeutic Plasma
MicroRNAs
miR-21, miR-23a, miR-27a, miR-106b-25, miR-130b, Upregulation Diagnostic, prognostic, Blood, plasma
miR-199a, miR-215, miR-222-221, miR-370 therapeutic
CHAPTER 5.4
SECTION 5
miR-29a, miR-101, miR-125a, miR-129, miR-148b, Upregulation Diagnostic, prognostic, Blood, plasma
miR-181c, miR-212, miR-218, miR-335, miR-375, therapeutic
miR-449, miR-486, miR-512
Cell-free microRNAs
miR-331 and miR-21 Upregulation Diagnostic, prognostic Blood
miR-20b, miR-125a, miR-137, miR-141, miR-146a, Upregulation Prognostic Blood, plasma
miR-196a, miR-206, miR-218, miR-486-5p
miR-10b-5p, miR-132-3p, miR-185-5p, miR-195-5p, Upregulation Prognostic Plasma
miR-20a-3p, miR-296-5p
Long non-coding RNAs
ncRuPAR Downregulation Diagnostic, prognostic Tissue
AI364715, GACAT1, GACAT2 Downregulation Prognostic Tissue
PVT1 Upregulation Prognostic Tissue

Table 5.4.2. Current topics of molecular markers associated with diagnosis, prognosis, and prediction of therapeutic response of
gastric cancer (continued)
Marker Alteration Clinical purpose Detection method

Exosomes
miR-19b, miR-106a Upregulation Diagnostic, prognostic Plasma
miR-21, miR-1225-5p Upregulation Diagnostic, therapeutic Peritoneal lavage fluid
Stomach-specific biomarkers
ADAM23, GDNF, MINT25, MLF1, PRDM5, RORA Hypermethylation Diagnostic Gastric wash
BARHL2 Hypermethylation Diagnostic, therapeutic Gastric wash, gastric
juice
PVT1 Upregulation Diagnostic, prognostic Gastric juice
miR-421, miR-21, miR-106a, miR-129 Upregulation Diagnostic Gastric juice
CagA Upregulation Diagnostic Tissue
VacA Upregulation Diagnostic Tissue
Gastrokine 1 Inactivation Prognostic Tissue
CagA, cytotoxin-associated gene A; EMT, epithelial–mesenchymal transition; FGFR, fibroblast growth factor receptor; HER2, human epidermal growth
factor receptor 2; MSI, microsatellite instability; mTOR, mammalian target of rapamycin; PI3K, phosphoinositide 3-kinase; PD-L1, programmed death-ligand
1; VacA, vacuolating toxin A; VEGF, vascular endothelial growth factor; VEGFD, vascular endothelial growth factor D; VEGFR2, vascular endothelial growth
factor receptor 2.
which are abnormally expressed in ness to chemotherapy and on prog- Carcinogens from red meat in-
tumour tissue and can be detected nosis in resectable gastric cancer. clude heterocyclic aromatic amines
by sensitive molecular methods in In two clinical trials, patients with and polycyclic aromatic hydrocar-
body fluids including serum, plasma, either MSI-high or mismatch repair- bons produced by cooking meat at
gastric juice, and urine of patients. deficient tumours (6.6%) had better high temperatures. N-nitroso com-
Additional studies are required to overall survival than patients with pounds and polycyclic aromatic hy-
improve their diagnostic and prog- neither MSI-high nor mismatch re- drocarbons are found in processed
nostic accuracy (Table 5.4.2) [36]. pair-deficient tumours when treated meat after curing and smoking.
with surgery alone [41,42]. Inhibition Red meat and processed meat also
Targeted therapies of anti-tumour immune cell activity, contain salt; high dietary salt intake,
Trastuzumab therapy for patients mediated by programmed death- low intake of fresh fruits and veg-
with HER2-positive tumours was ligand 1 (PD-L1) or PD-L2, is par- etables, and tobacco smoking are
the first example of molecular ticularly upregulated in EBV-positive behavioural factors that increase
targeted therapy for gastric can- tumours [21]. The successful out- the risk of gastric cancer [2].
cer. The Trastuzumab for Gastric comes of multicentre trials of the im-
Cancer international randomized mune checkpoint inhibitor pembroli-
clinical trial demonstrated that treat- zumab support the use of tumour Fig. 5.4.6. The consumption of processed
ment with trastuzumab (a mono- PD-L1 and MSI status as a guide to meat has been associated with increased
clonal antibody targeting HER2) therapy and prognosis in resectable risk of gastric cancer.
plus chemotherapy significantly gastric cancer [43,44].
improved survival of patients with
HER2-positive advanced disease Prevention
[39]. HER2 amplification is routine-
ly detected in resected tumours by Reduced exposure to
standard immunohistochemistry. In carcinogens
an international randomized mul- The consumption of processed
ticentre trial, ramucirumab, which meat has been associated with gas-
targets vascular endothelial growth tric cancer in several case–control
factor receptor 2 (VEGFR2), has and cohort studies in many coun-
also shown efficacy as anti-angio- tries worldwide. For gastric cancer
genic therapy for previously treated specifically, the IARC Monographs
advanced gastric cancer [40]. found the evidence to be limited for
The MSI and mismatch repair processed meat and inadequate for
status has an impact on responsive- red meat (see Chapter 2.6) [45].
340
The Stomach Cancer Pooling (OLGA) and the Operative Link gastric pre-neoplastic lesions is the
Project, a consortium that included on Gastric Intestinal Metaplasia use of both serum pepsinogen lev-
23 epidemiological studies with Assessment (OLGIM) systems is els and H. pylori serology [53]. A low
10 290 cases and 26 145 controls recommended [48,49]. CDH1 test- serum pepsinogen I level or a low
from Europe, North America, and ing is recommended for patients pepsinogen I/II ratio is associated
Asia, evaluated the risk factors for with a family history of hereditary with gastric atrophy and is the best
gastric cancer using individual data diffuse gastric cancer and those available marker, despite its limited
rather than conventional meta- with precursor lesions for signet sensitivity for predicting risk of gas-
analysis. Tobacco smoking was ring cell carcinoma [50]. Guidelines tric cancer. A recent meta-analysis
confirmed as an important risk fac- were also developed for follow-up of 27 studies including 8654 pa-
tor. The risk was higher for cardia of individuals at risk [51]. tients from different geographical
tumours than for non-cardia tu- The development of new endo- regions confirmed the potential use
mours, both with and without H. scopy imaging technologies will help of serum pepsinogen I and II levels
pylori infection. In addition, the risk health professionals to diagnose in- in combination with gastrin-17 and
increased with the intensity and du- testinal metaplasia and early gastric anti-H. pylori antibodies for the non-
ration of smoking and decreased cancer [52]. Another strategy, in ad- invasive diagnosis and screening of
after smoking cessation [46]. dition to upper digestive endoscopy, atrophic gastritis of the corpus and
Alcohol consumption was also a for the diagnosis and surveillance of the antrum [54].
risk factor for both cardia and non-
cardia gastric cancer and for both
the intestinal and diffuse histologi- Fig. 5.4.7. A patient undergoing endoscopy.
cal subtypes, but at a lower magni-
tude than that found in conventional
meta-analysis [47].
Screening and improved

methods of detection and
diagnosis
In countries with low or medium in-
cidence of gastric carcinoma, and
in subjects at increased risk on the
basis of family history, H. pylori in-
fection history, ethnic background,
or immigration from a geographi-
cal location where risk of gastric
cancer is high, endoscopic surveil-
lance with multiple biopsies for a
topographical mapping of the entire
stomach and staging of gastric his-
tology according to the Operative
Link on Gastritis Assessment
CHAPTER 5.4
SECTION 5
References
1. IARC (1994). Schistosomes, liver flukes 3. Bray F, Ferlay J, Soerjomataram I, Siegel 4. Agudo A, Cayssials V, Bonet C,
and Helicobacter pylori. IARC Monogr RL, Torre LA, Jemal A (2018). Global Tjønneland A, Overvad K, Boutron-
Eval Carcinog Risks Hum. 61:1–241. cancer statistics 2018: GLOBOCAN es- Ruault M-C, et al. (2018). Inflammatory
Available from: http://publications.iarc. timates of incidence and mortality world- potential of the diet and risk of gastric
fr/79 PMID:7715068 wide for 36 cancers in 185 countries. CA cancer in the European Prospective
Cancer J Clin. 68(6):394–424. https://doi. Investigation into Cancer and Nutrition
2. IARC (2012). Biological agents. IARC org/10.3322/caac.21492 PMID:30207593 (EPIC) study. Am J Clin Nutr. 107(4):607–
Monogr Eval Carcinog Risks Hum. 16. https://doi.org/10.1093/ajcn/nqy002
100B:1–441. Available from: http://publi PMID:29635497
cations.iarc.fr/119 PMID:23189750

5. Whary MT, Sundina N, Bravo LE, Correa P, 14. Bae SE, Choi KD, Choe J, Kim SO, Na 23. Cristescu R, Lee J, Nebozhyn M, Kim K-M,
Quinones F, Caro F, et al. (2005). Intestinal HK, Choi JY, et al. (2018). The effect of Ting JC, Wong SS, et al. (2015). Molecular
helminthiasis in Colombian children pro- eradication of Helicobacter pylori on analysis of gastric cancer identifies sub-
motes a Th2 response to Helicobacter gastric cancer prevention in healthy types associated with distinct clinical out-
pylori: possible implications for gas- asymptomatic populations. Helicobacter. comes. Nat Med. 21(5):449–56. https://
tric carcinogenesis. Cancer Epidemiol 23(2):e12464. https://doi.org/10.1111/ doi.org/10.1038/nm.3850 PMID:25894828
Biomarkers Prev. 14(6):1464–9. https:// hel.12464 PMID:29345408
doi.org/10.1158/1055-9965.EPI-05-0095 24. Petrelli F, Berenato R, Turati L, Mennitto A,
15. Huerta-Franco MR, Banderas JW, Steccanella F, Caporale M, et al. (2017).
PMID:15941957
Allsworth JE (2018). Ethnic/racial dif- Prognostic value of diffuse versus intesti-
6. Hooi JKY, Lai WY, Ng WK, Suen MMY, ferences in gastrointestinal symptoms nal histotype in patients with gastric can-
Underwood FE, Tanyingoh D, et al. and diagnosis associated with the risk cer: a systematic review and meta-anal-
(2017). Global prevalence of Helicobacter of Helicobacter pylori infection in the ysis. J Gastrointest Oncol. 8(1):148–63.
pylori infection: systematic review US. Clin Exp Gastroenterol. 11:39–49. https://doi.org/10.21037/jgo.2017.01.10
and meta-analysis. Gastroenterology. ht tps://doi.org/10. 2147/CEG.S14 49 67 PMID:28280619
153(2):420–9. https://doi.org/10.1053/j. PMID:29403299
gastro.2017.04.022 PMID:28456631 25. Carrasco-Garcia E, García-Puga M,
16. Anderson WF, Rabkin CS, Turner N, Arevalo S, Matheu A (2018). Towards pre-
7. Zamani M, Ebrahimtabar F, Zamani V, Miller Fraumeni JF Jr, Rosenberg PS, Camargo cision medicine: linking genetic and cel-
WH, Alizadeh-Navaei R, Shokri-Shirvani MC (2018). The changing face of noncar- lular heterogeneity in gastric cancer. Ther
J, et al. (2018). Systematic review with dia gastric cancer incidence among US Adv Med Oncol. 10:1758835918794628.
meta-analysis: the worldwide prevalence non-Hispanic whites. J Natl Cancer Inst. https://doi.org/10.1177/1758835918794628
of Helicobacter pylori infection. Aliment 110(6):608–15. https://doi.org/10.1093/ PMID:30181784
Pharmacol Ther. 47(7):868–76. https://doi. jnci/djx262 PMID:29361173
org/10.1111/apt.14561 PMID:29430669 26. Sohn BH, Hwang J-E, Jang H-J, Lee
17. Luo G, Zhang Y, Guo P, Wang L, Huang H-S, Oh SC, Shim J-J, et al. (2017).
8. Lee JH, Choi KD, Jung H-Y, Baik GH, Y, Li K (2017). Global patterns and trends Clinical significance of four molecular
Park JK, Kim SS, et al.; Korean College of in stomach cancer incidence: age, period subtypes of gastric cancer identified by
Helicobacter and Upper Gastrointestinal and birth cohort analysis. Int J Cancer. The Cancer Genome Atlas project. Clin
Research (2018). Seroprevalence of 141(7):1333–44. https://doi.org/10.1002/ Cancer Res. 23(15):4441–9. https://doi.
Helicobacter pylori in Korea: a multicen- ijc.30835 PMID:28614909 org/10.1158/1078 - 0 432.CCR-16 -2211
ter, nationwide study conducted in 2015 18. Venneman K, Huybrechts I, Gunter MJ, PMID:28747339
and 2016. Helicobacter. 23(2):e12463. Vandendaele L, Herrero R, Van Herck K
h t t p s : / / d o i . o r g / 1 0 .1111 / h e l .1 2 4 6 3 27. Cisło M, Filip AA, Arnold Offerhaus GJ,
(2018). The epidemiology of Helicobacter
PMID:29345022 Ciseł B, Rawicz-Pruszyński K, Skierucha
pylori infection in Europe and the impact
M, et al. (2018). Distinct molecular subtypes
9. Liu J, Wang Y, Zhao Q, Luo R, Xiao M, of lifestyle on its natural evolution toward
of gastric cancer: from Laurén to molecular
Zhang M, et al. (2017). Prevalence and stomach cancer after infection: a system-
pathology. Oncotarget. 9(27):19427–42.
risk factors for Helicobacter pylori in- atic review. Helicobacter. 23(3):e12483.
https://doi.org/10.18632/oncotarget.24827
fection in southwest China: a study of h t t p s : / / d o i . o r g / 1 0 .1111 / h e l .1 2 4 8 3
PMID:29721214
health examination participants based PMID:29635869
on 13C-urea breath test. Turk J Med Sci. 28. Li X, Wu WKK, Xing R, Wong SH, Liu Y,
19. Fewings E, Larionov A, Redman J,
47(5):1456–62. https://doi.org/10.3906/ Fang X, et al. (2016). Distinct subtypes
Goldgraben MA, Scarth J, Richardson S,
sag-1605-149 PMID:29151317 of gastric cancer defined by molecular
et al. (2018). Germline pathogenic variants
characterization include novel mutational
10. Binh TT, Tuan VP, Dung HDQ, Tung PH, Tri in PALB2 and other cancer-predisposing
signatures with prognostic capability.
TD, Thuan NPM, et al. (2018). Molecular ep- genes in families with hereditary diffuse
gastric cancer without CDH1 mutation: a Cancer Res. 76(7):1724–32. https://doi.
idemiology of Helicobacter pylori infection org/10.1158/0 0 08 - 5 472.CAN -15 -24 43
whole-exome sequencing study. Lancet
in a minor ethnic group of Vietnam: a mul- PMID:26857262
Gastroenterol Hepatol. 3(7):489–98. https://
tiethnic, population-based study. Int J Mol
doi.org/10.1016/S2468-1253(18)30079-7 29. Liu X, Meltzer SJ (2017). Gastric can-
Sci. 19(3):E708. https://doi.org/10.3390/
PMID:29706558 cer in the era of precision medicine. Cell
ijms19030708 PMID:29494554
20. Oliveira C, Pinheiro H, Figueiredo J, Mol Gastroenterol Hepatol. 3(3):348–58.
11. Yin J, Song JN, Bai ZG, Cai J, Zhang J, https://doi.org/10.1016/j.jcmgh.2017.02.003
Seruca R, Carneiro F (2015). Familial gas-
Zheng Z, et al. (2017). Gastric cancer PMID:28462377
tric cancer: genetic susceptibility, pathol-
mortality trends in China (2006–2013)
ogy, and implications for management. 30. Castaño-Rodríguez N, Goh K-L, Fock
reveal increasing mortality in young
Lancet Oncol. 16(2):e60–70. https://doi. KM, Mitchell HM, Kaakoush NO (2017).
subjects. Anticancer Res. 37(8):4671–9.
o r g /10.1016 /S1470 -2 0 4 5 (14)71016 -2
https://doi.org/10.21873/anticanres.11871 Dysbiosis of the microbiome in gastric car-
PMID:25638682
PMID:28739770 cinogenesis. Sci Rep. 7(1):15957. https://
21. The Cancer Genome Atlas Research doi.org /10.103 8/s 4159 8 - 017-1628 9 -2
12. Hiroi S, Sugano K, Tanaka S, Kawakami PMID:29162924
Network (2014). Comprehensive mo-
K (2017). Impact of health insurance
lecular characterization of gastric ad-
coverage for Helicobacter pylori gastritis 31. Hsieh YY, Tung S-Y, Pan H-Y, Yen
enocarcinoma. Nature. 513(7517):202–9.
on the trends in eradication therapy in C-W, Xu H-W, Lin Y-J, et al. (2018).
ht t p s: //d o i.o r g /10.10 3 8 /n at u r e13 4 8 0
Japan: retrospective observational study Increased abundance of Clostridium and
PMID:25079317
and simulation study based on real-world Fusobacterium in gastric microbiota of
data. BMJ Open. 7(7):e015855. https:// 22. Wang K, Yuen ST, Xu J, Lee SP, Yan patients with gastric cancer in Taiwan.
doi.org/10.1136/bmjopen-2017- 015855 HHN, Shi ST, et al. (2014). Whole- Sci Rep. 8(1):158. https://doi.org/10.1038/
PMID:28760790 genome sequencing and comprehensive s41598-017-18596-0 PMID:29317709
molecular profiling identify new driver
13. Tsuda M, Asaka M, Kato M, Matsushima 32. Coker OO, Dai Z, Nie Y, Zhao G, Cao
mutations in gastric cancer. Nat Genet.
R, Fujimori K, Akino K, et al. (2017). 46(6):573–82. https://doi.org/10.1038/ L, Nakatsu G, et al. (2018). Mucosal mi-
Effect on Helicobacter pylori eradication ng.2983 PMID:24816253 crobiome dysbiosis in gastric carcino-
therapy against gastric cancer in Japan. genesis. Gut. 67(6):1024–32. https://
Helicobacter. 22(5):e12415. https://doi. d o i . o r g /10 .113 6 /g u t j n l - 2 017- 3142 81
org/10.1111/hel.12415 PMID:28771894 PMID:28765474
342
33. Ferreira RM, Pereira-Marques J, Pinto- 41. Choi Y, Kim H, Yang H, Kim WH, Kim 49. den Hollander WJ, Holster IL, den Hoed
Ribeiro I, Costa JL, Carneiro F, Machado YW, Kook M-C, et al. (2017). Clinical CM, Capelle LG, Tang TJ, Anten M-P, et
JC, et al. (2018). Gastric microbial com- impact of microsatellite instability in pa- al. (2019). Surveillance of premalignant
munity profiling reveals a dysbiotic tients with stage II and III gastric cancer: gastric lesions: a multicentre prospec-
cancer-associated microbiota. Gut. results from the CLASSIC trial. J Clin tive cohort study from low incidence
67(2):226–36. https://doi.org/10.1136/ Oncol. 35(15_suppl):4022. https://doi. regions. Gut. 68(4):585–93. https://
gutjnl-2017-314205 PMID:29102920 org/10.1200/JCO.2017.35.15_suppl.4022 d o i . o r g /10 .113 6 /g u t j n l - 2 017- 314 4 9 8
PMID:29875257
34. Thorell K, Bengtsson-Palme J, Liu OH- 42. Smyth EC, Wotherspoon A, Peckitt C,
F, Palacios Gonzales RV, Nookaew I, Gonzalez D, Hulkki-Wilson S, Eltahir Z, 50. van der Post RS, Vogelaar IP, Carneiro F,
Rabeneck L, et al. (2017). In vivo analysis Guilford P, Huntsman D, Hoogerbrugge
et al. (2017). Mismatch repair deficiency,
of the viable microbiota and Helicobacter N, et al. (2015). Hereditary diffuse gastric
microsatellite instability, and survival:
pylori transcriptome in gastric infec- cancer: updated clinical guidelines with
an exploratory analysis of the Medical
tion and early stages of carcinogen- an emphasis on germline CDH1 muta-
esis. Infect Immun. 85(10):e00031–17. Research Council Adjuvant Gastric
tion carriers. J Med Genet. 52(6):361–74.
ht t p s: //d o i.o r g /10.1128 / I A I . 0 0 0 31-17 Infusional Chemotherapy (MAGIC) trial.
https://doi.org/10.1136/jmedgenet-2015-
PMID:28694295 JAMA Oncol. 3(9):1197–203. https://
103094 PMID:25979631
doi.org/10.1001/jamaoncol.2016.6762
35. Noto JM, Peek RM Jr (2017). The gas- PMID:28241187 51. Dinis-Ribeiro M, Areia M, de Vries AC,
tric microbiome, its interaction with Marcos-Pinto R, Monteiro-Soares M,
Helicobacter pylori, and its potential role 43. Le DT, Uram JN, Wang H, Bartlett BR,
O’Connor A, et al.; European Society of
in the progression to stomach cancer. Kemberling H, Eyring AD, et al. (2015).
Gastrointestinal Endoscopy; European
PLoS Pathog. 13(10):e1006573. https:// PD-1 blockade in tumors with mis- Helicobacter Study Group; European
doi.org/10.1371/journal.ppat.10 0 6573 match-repair deficiency. N Engl J Med. Society of Pathology; Sociedade
PMID:28982167 372(26):2509–20. https://doi.org/10.1056/ Portuguesa de Endoscopia Digestiva
NEJMoa1500596 PMID:26028255 (2012). Management of precancerous
36. Matsuoka T, Yashiro M (2018). Biomarkers
of gastric cancer: current topics and fu- conditions and lesions in the stomach
44. Muro K, Chung HC, Shankaran V, Geva
ture perspective. World J Gastroenterol. (MAPS): guideline from the European
R, Catenacci D, Gupta S, et al. (2016).
24(26):2818–32. https://doi.org/10.3748/ Society of Gastrointestinal Endoscopy
Pembrolizumab for patients with PD-
wjg.v24.i26.2818 PMID:30018477 (ESGE), European Helicobacter Study
L1-positive advanced gastric cancer
Group (EHSG), European Society of
37. Nguyen PH, Giraud J, Chambonnier L, (KEYNOTE-012): a multicentre, open-
Pathology (ESP), and the Sociedade
Dubus P, Wittkop L, Belleannée G, et al. label, phase 1b trial. Lancet Oncol.
Portuguesa de Endoscopia Digestiva
(2017). Characterization of biomarkers of 17(6):717–26. https://doi.org/10.1016/ (SPED). Endoscopy. 44(1):74–94. https://
tumorigenic and chemoresistant cancer S1470-2045(16)00175-3 PMID:27157491 d o i . o r g / 1 0 .1 0 5 5 / s - 0 0 3 1 - 1 2 9 1 4 9 1
stem cells in human gastric carcinoma. PMID:22198778
45. Bouvard V, Loomis D, Guyton KZ, Grosse
Clin Cancer Res. 23(6):1586–97. https://
Y, Ghissassi FE, Benbrahim-Tallaa L, et 52. Sugano K, Tack J, Kuipers EJ, Graham
doi.org/10.1158/1078-0432.CCR-15-2157
PMID:27620279 al.; International Agency for Research DY, El-Omar EM, Miura S, et al.; faculty
on Cancer Monograph Working Group members of Kyoto Global Consensus
38. Wen J, Zheng T, Hu K, Zhu C, Guo L, Ye G (2015). Carcinogenicity of consumption of Conference (2015). Kyoto global con-
(2017). Promoter methylation of tumor- red and processed meat. Lancet Oncol. sensus report on Helicobacter pylori
related genes as a potential biomarker 16(16):1599–600. https://doi.org/10.1016/ gastritis. Gut. 64(9):1353–67. https://
using blood samples for gastric cancer S1470-2045(15)00444-1 PMID:26514947 d o i .o r g /10 .113 6 /g u t j n l - 2 015 - 3 0 9 2 5 2
detection. Oncotarget. 8(44):77783–93. PMID:26187502
https://doi.org/10.18632/oncotarget.20782 46. Praud D, Rota M, Pelucchi C, Bertuccio
PMID:29100425 P, Rosso T, Galeone C, et al. (2018). 53. Miki K (2011). Gastric cancer screening
Cigarette smoking and gastric cancer in the by combined assay for serum anti-Heli-
39. Bang YJ, Van Cutsem E, Feyereislova A, Stomach Cancer Pooling (StoP) Project. cobacter pylori IgG antibody and serum
Chung HC, Shen L, Sawaki A, et al.; ToGA pepsinogen levels - “ABC method”. Proc
Eur J Cancer Prev. 27(2):124–33. https://
Trial Investigators (2010). Trastuzumab Jpn Acad Ser B Phys Biol Sci. 87(7):405–
doi.org/10.1097/CEJ.0000000000000290
in combination with chemotherapy ver- 14. https://doi.org/10.2183/pjab.87.405
PMID:27560662
sus chemotherapy alone for treatment of
PMID:21785258
HER2-positive advanced gastric or gas- 47. Ferro A, Morais S, Pelucchi C, Aragonés
tro-oesophageal junction cancer (ToGA): N, Kogevinas M, López-Carrillo L, et 54. Syrjänen K (2016). A panel of serum bio-
a phase 3, open-label, randomised con- al. (2019). Smoking and Helicobacter markers (GastroPanel®) in non-invasive
trolled trial. Lancet. 376(9742):687–97. diagnosis of atrophic gastritis. Systematic
pylori infection: an individual partici-
https://doi.org/10.1016/S0140- 6736(10) review and meta-analysis. Anticancer Res.
pant pooled analysis (Stomach Cancer
61121-X PMID:20728210 36(10):5133–44. https://doi.org/10.21873/
CHAPTER 5.4
SECTION 5
Pooling- StoP Project). Eur J Cancer
Prev. 28(5):390–6. https://doi.org/10.1097/ anticanres.11083 PMID:27798873
40. Fuchs CS, Tomasek J, Yong CJ, Dumitru
F, Passalacqua R, Goswami C, et al.; CEJ.0000000000000471 PMID:30272597
REGARD Trial Investigators (2014).
Ramucirumab monotherapy for previ- 48. Rugge M, Genta RM, Fassan M, Valentini
ously treated advanced gastric or gastro- E, Coati I, Guzzinati S, et al. (2018).
oesophageal junction adenocarcinoma OLGA gastritis staging for the predic-
(REGARD): an international, randomised, tion of gastric cancer risk: a long-term
multicentre, placebo-controlled, phase follow-up study of 7436 patients. Am J
3 trial. Lancet. 383(9911):31–9. https:// Gastroenterol. 113(11):1621–8. https://
doi.org/10.1016/S0140-6736(13)61719-5 d o i.o r g /10.10 3 8 /s 413 9 5 - 018 - 0 3 5 3 - 8
PMID:24094768 PMID:30333540

5.5
Colorectal cancer
Decreasing disparities and promoting
prevention are policy priorities
Carlo Senore Colinda Simons (reviewer)
Nereo Segnan Piet van den Brandt (reviewer)
Marc Gunter
for about 25–30% of sporadic The global disease burden in 2016

SUMMARY colorectal cancers. was estimated as 17.2 million (95%
confidence interval, 6.5–17.9 mil-
●● The estimated age-standard- ●● Dietary patterns characterized
lion) disability-adjusted life years,
ized incidence rates of colo- by high intakes of fruits and
of which 97% came from years of
rectal cancer in countries with vegetables, whole grains, nuts
life lost due to premature mortality
higher Human Development and legumes, fish and other
and 3% came from years of healthy
Index are about 5 times those seafood, and milk and other
life lost due to disability. Colorectal
in countries with lower Human dairy products are associated
cancer survivors diagnosed with
Development Index. In Australia with a lower risk of colorectal
the disease during the previous
and Europe, the rates are 35– cancer. Dietary patterns char-
5 years made up about 11% of all
42 per 100 000 in men and 24– acterized by high intakes of red
5-year cancer survivors estimated
32 per 100 000 in women, com- meat, processed meat, sugar-
to be alive at the end of 2018 [1].
pared with 7 per 100 000 in men sweetened beverages, refined
In general, colorectal cancer in-
and 6 per 100 000 in women in grains, desserts, and potatoes
cidence rates are now considered
West Africa and 6 per 100 000 are associated with a higher
to be one of the clearest indicators
in men and 4 per 100 000 in risk of colorectal cancer.
of disease transition within coun-
women in South Asia. ●● There is convincing evidence tries that are undergoing socioeco-
●● Sporadic colorectal cancers that physical activity decreases nomic development, which is asso-
have traditionally been de- the risk of colon cancer. ciated with shifts to lifestyles more
scribed as developing along typical of industrialized countries,
●● Screening, with stool-based
two molecular pathways: (i) the because colorectal cancer rates
tests for occult blood or with en-
conventional adenoma–carci- show a strong positive gradient with
doscopic methods, is associat-
noma, or chromosomal insta- Human Development Index (HDI) or
ed with a reduction in colorectal
bility, pathway, and (ii) the ser- Sociodemographic Index (SDI) (see
cancer incidence and mortality.
rated pathway. Chapter 1.3) [2].
●● Use of aspirin appeared to re- The estimated age-standard-
●● The chromosomal instability duce colorectal cancer incidence ized incidence rates of colorectal
pathway, which involves Wnt and mortality, after a latency of cancer in countries with higher
signalling and KRAS mutation, about 10 years. HDI (e.g. Australia, New Zealand,
accounts for about 70–75% of
and European countries) are about
sporadic colorectal cancers.
5 times those in countries with
●● The serrated pathway involves Epidemiology lower HDI (e.g. countries in Africa
BRAF mutation and the ac- and South Asia). In Australia and
cumulation of epigenetic al- Global burden Europe, the rates are 35–42 per
terations, which cause silenc- Worldwide, colorectal cancer is 100 000 in men and 24–32 per
ing of regulatory genes, often the third most common cancer in 100 000 in women, compared with
including MLH1 (CpG island men and the second most com- 7 per 100 000 in men and 6 per
methylator phenotype and mi- mon in women, accounting for an 100 000 in women in West Africa
crosatellite instability-high phe- estimated 1.85 million new cases and 6 per 100 000 in men and 4 per
notype); this pathway accounts and 881 000 deaths in 2018 [1]. 100 000 in women in South Asia [1].
344
Colorectal cancer tends to oc- nostic assessment or to appropriate

cur more frequently in men than high-quality oncological care after FUNDAMENTALS
in women, although the male-to- diagnosis, have emerged as the
female ratio decreases from 1.6 in most likely determinants of the lower ■■ Colorectal cancer is a
countries with high SDI to 1.0 in survival in disadvantaged groups. highly preventable disease.
countries with low SDI [3]. The in- Indeed, a more advanced stage at A substantial proportion
cidence rates increase with age: of diagnosis, a lower chance of receiv- of the colorectal cancer
the estimated 1.85 million new cas- ing curative treatment, and a higher
burden is attributable to
es worldwide in 2018, about 10% risk of having permanent stoma have
modifiable lifestyle factors
were estimated to occur in people been observed in patients with low
and environmental factors.
younger than 50 years, 59% in peo- socioeconomic status, as well as in
low-income countries [7,8,12,13]. Effective screening methods
ple aged 50–74 years, and 31% in
people aged 75 years and older [1]. are available.
Those countries with the high- Time trends ■■ An improved understanding
est incidence rates tend to have Independent analyses of trends in of the biology and the natural
relatively low mortality rates, com- colorectal cancer incidence and history of colorectal cancer
pared with the regions of Africa, mortality rates by SDI quintile re- has been associated with a
Asia, and South America, which vealed three distinct patterns [3,4].
trend towards more favourable
have considerably higher mortali- The first pattern, characterized
outcomes in more recent years.
ty-to-incidence ratios [1,4,5]. The by increases in both incidence rates
observed association of colorectal and mortality rates, was observed ■■ As long as the stage at
cancer mortality-to-incidence ratios in rapidly transitioning countries, diagnosis remains the main
with health system ranking sug- i.e. in countries in the low-middle determinant of survival, access
gests that health-care organization, and low SDI quintiles, in which the to appropriate, high-quality
including cancer-related screening economic growth was often associ- screening can make a crucial
and care, has a substantial impact ated with a shift towards unhealthy contribution to improving
on colorectal cancer mortality [6]. dietary habits, together with reduc- colorectal cancer outcomes.
Geographical patterns of colo- tions in levels of physical activity
rectal cancer incidence and mortal- and increases in the prevalence of ■■ Public health strategies aimed
ity are related to indexes of develop- overweight and obesity. In coun- at reducing the prevalence of
ment. In addition, colorectal cancer tries in the low SDI quintile, there obesity, promoting physical
mortality is strongly associated with was a larger increase in mortality activity, and discouraging
indexes of socioeconomic status, rates than in incidence rates. the consumption of high-
also within high-income countries. The second pattern was char- energy, obesogenic foods are
Most reports have documented acterized by a decrease in mortality gradually being implemented
higher colorectal cancer mortality rates and an increase in incidence in many regions of the world.
in people with lower socioeconom- rates. The decrease in mortality
ic status; this is consistent with the rates is probably related to an in- ■■ Together with policies
observed association of lower colo- creased availability of health-care that promote prevention,
rectal cancer survival with lower so- resources, which favour the dis- policies aimed at decreasing
cioeconomic status [7,8]. semination of best practices in can- disparities in timely access to
Over the past decades, evolv- cer management. The increase in diagnostic assessment and to
ing cancer treatment, as well as incidence rates is probably related high-quality oncological care
the more recent availability of in- to the recent introduction of screen- are priorities, to reduce the
CHAPTER 5.5
SECTION 5
novative drugs and chemotherapy ing and/or to persisting unfavour- colorectal cancer burden.
regimens, has resulted in a trend to- able lifestyle patterns. This pattern
wards improved stage-specific sur- was observed in countries in the
vival outcomes, in particular for pa- high-middle and middle SDI quin-
tients with stage II and III colorectal tiles, as well as in some countries explained by the early introduction
cancer. Improvement in patient with high HDI and high SDI, reflect- of screening as well as changes in
management and closer adherence ing the observed variability in the profiles of risk factors and protec-
to treatment guidelines – reflected implementation of screening and in tive factors, together with the avail-
in a higher use of curative surgery, the patterns of risk factors. ability of high-quality cancer care.
chemotherapy, and radiotherapy – The third pattern, character- On the basis of currently esti-
have contributed to the increasing ized by decreases in both incidence mated incidence and mortality rates,
trends in survival [9–11]. rates and mortality rates, was ob- the projected demographic changes
Financial and cultural barriers, served in countries with high HDI in the global population alone will re-
which delay or limit access to diag- and high SDI. This pattern may be sult in increases of about 80% both
Chapter 5.5 • Colorectal cancer 345

Fig. 5.5.1. Time trends in colorectal cancer incidence and mortality for countries by quintile of the Sociodemographic Index (SDI).
Average annual percentage change in age-standardized death rate (ASDR) and age-standardized incidence rate (ASIR) between
2006 and 2016 for both sexes, by SDI quintile.
25%
Average annual percentage change in

20%
age-standardized rate 15%
10%
5%
0%
-5%
-10%
-15%
High SDI High-middle SDI Middle SDI Low-middle SDI Low SDI
Change in ASDR 2006–2016 (%) -11% -10% -3% 3% 8%
Change in ASIR 2006–2016 (%) -2% 5% 20% 9% 4%
SDI quintile
in the annual incidence of colorectal Chromosomal instability growth, they are likely to be detected
cancer (from 1.2 million new cases pathway at screening [17].
in 2008 to 2.2 million in 2030) and The chromosomal instability path- Different mechanisms contribute
in the mortality from colorectal can- way is thought to be driven by the to chromosomal instability, resulting
cer (from 0.6 million deaths in 2008 accumulation of mutational events in in karyotypic abnormalities, such as
to 1.1 million in 2030). Most of this oncogenes and tumour suppressor chromosome number alterations,
additional disease burden will oc- genes during the progression from telomere dysfunction or overexpres-
cur in countries with lower HDI, as small adenoma to invasive carcino- sion, or loss of heterozygosity, which
a result of the demographic transi- ma [15]. The earliest genetic event has been reported in more than 70%
tion and the adoption of lifestyles of colorectal cancers at chromosome
is the activation of Wnt signalling
more typical of industrialized coun- 18q. The stage of colorectal carcino-
by an inactivating mutation of the
tries. Although the number of new genesis at which the chromosomal
adenomatous polyposis coli (APC)
cases per year will remain higher in instability phenotype arises is still
tumour suppressor gene. Sporadic
countries with high HDI, by 2035 the uncertain. A role of APC mutation in
APC mutations are detected in 5%
number of deaths from colorectal favouring the initiation of chromoso-
of aberrant crypt foci, in 30–70% of
cancer will be greatest in countries mal instability has been proposed,
adenomas, and in more than 70%
with low HDI [14]. although chromosomal abnormali-
of colorectal cancers. Mutation of
the KRAS oncogene occurs prefer- ties have also been observed at very
Pathogenesis entially in early phases of the ade- early stages of tumorigenesis [16].
Colorectal cancer is a heteroge- noma–carcinoma sequence. KRAS
Serrated pathway
neous disease. The majority of mutations are detected in about
cases are sporadic tumours, which 50% of large polyps and colorectal Sessile serrated adenoma
have traditionally been described cancers and result in promotion of The initiating event in the develop-
as developing along two molecular adenomatous growth. Mutations of ment of sessile serrated adenoma is
pathways: (i) the conventional ade- TP53, SMAD4, PIK3C, and PTEN thought to be activation of the mito-
noma–carcinoma, or chromosomal are late events in colorectal carcinogen-activated protein kinase (MAPK)
instability, pathway, and (ii) the ser- genesis [16]. The dwell time of these pathway through mutation of the
rated pathway. These two pathways lesions (i.e. the period of time for a BRAF oncogene; this triggers down-
account for about 70–75% and benign polyp to evolve into cancer) regulation of apoptosis and enables
25–30%, respectively, of sporadic is thought to be about 10–15 years, cell proliferation. In the serrated path-
colorectal cancers. and because of their regular, slow way, BRAF mutation is associated
346
Trends by age, sex, site, and stage
Data from the United States people younger than 40 years, re- 2. Peterse EFP, Meester RGS, Siegel RL,
Chen JC, Dwyer A, Ahnen DJ, et al.
Surveillance, Epidemiology, and sulting in an increase in the lifetime (2018). The impact of the rising colorectal
End Results (SEER) programme risk of colorectal cancer. However, cancer incidence in young adults on the
have shown that colorectal can- colorectal cancer mortality rates optimal age to start screening: microsim-
cer incidence and mortality rates have remained stable since the ulation analysis I to inform the American
Cancer Society colorectal cancer screen-
have declined in people older than mid-1990s in adults younger than ing guideline. Cancer. 124(14):2964–73.
50 years since the late 1980s. This 50 years, with a modest increase ht t p s: //d o i.o r g /10.10 0 2 /c n c r. 315 4 3
decline is probably related to the only in White people in the most PMID:29846933
implementation of screening. In recent years, and the large rela- 3. Murphy CC, Lund JL, Sandler RS (2017).
contrast, there has been a continu- tive increase in colorectal cancer Young-onset colorectal cancer: earlier
diagnoses or increasing disease bur-
ous increase in colorectal cancer incidence was based on a small den? Gastroenterology. 152(8):1809–
incidence rates in adults younger increase in the absolute risk [3]. 1812.e3. https://doi.org/10.1053/j.gas-
than 50 years, from the 1990s until Also, a similar trend towards an tro.2017.04.030 PMID:28461196
2014 [1]. increase in colorectal cancer inci- 4. Senore C, Bellisario C, Segnan N (2017).
Although an increase in the dence in younger cohorts has not Distribution of colorectal polyps: impli-
screening of adults younger than yet been reported in other high- cations for screening. Best Pract Res
Clin Gastroenterol. 31(4):481–8. https://
50 years may have contributed to income countries. Therefore, the d o i.o r g /10.1016 / j. b p g. 2 017. 0 4. 0 0 8
the observed increase in colorectal net benefit of starting screening at PMID:28842058
cancer incidence in people aged a younger age remains uncertain. 5. Koo JH, Leong RW (2010). Sex dif-
40–49 years [1], age–period–co- The proportion of colorectal ferences in epidemiological, clinical
hort modelling of the colorectal cancers and adenomas located in and pathological characteristics of
colorectal cancer. J Gastroenterol
cancer incidence data indicated the proximal colon has an increas-
Hepatol. 25(1):33–42. https://doi.
only a modest period effect. A ing trend with age [4]. The shift to org/10.1111/j.1440-1746.2009.05992.x
trend towards an increase in age- a higher proportion of colorectal PMID:19874446
specific risk of colon and rectal cancers located in the proximal 6. Massat NJ, Moss SM, Halloran SP, Duffy
cancer was observed in more re- colon occurs at a younger age in SW (2013). Screening and primary pre-
cent birth cohorts. This supports women than in men [5,6]. This may vention of colorectal cancer: a review of
sex-specific and site-specific differences.
the hypothesis of a strong birth also suggest a need to design sex- J Med Screen. 20(3):125–48. https://
cohort effect that began in people specific screening strategies. doi.org /10.1177/0 9 6 9141313501292
born in the 1950s [1]. The stage distribution of colo- PMID:24197771
The results of a decision-ana- rectal cancer at diagnosis has 7. Goodyear SJ, Leung E, Menon A,
lytic modelling analysis suggest- remained stable over the past de- Pedamallu S, Williams N, Wong LS
cades in several high-income coun- (2008). The effects of population-based
ed that starting screening at age
faecal occult blood test screening upon
45 years instead of age 50 years tries, with a shift towards a more fa- emergency colorectal cancer admissions
may have a favourable balance vourable stage distribution after the in Coventry and north Warwickshire. Gut.
between benefits and costs for all introduction of screening [7,8]. 57(2):218–22. https://doi.org/10.1136/
gut.2007.120253 PMID:18048571
people at average risk in the USA
[2]. This result holds only under References 8. Zorzi M, Mangone L, Sassatelli R,
Baracco S, Budroni M, Castaing M, et al.;
the assumption of an increase in 1. Siegel RL, Fedewa SA, Anderson WF, IMPATTO COLONRETTO working group
the age-specific risks of colorec- Miller KD, Ma J, Rosenberg PS, et al. (2015). Screening for colorectal cancer in
(2017). Colorectal cancer incidence pat-
tal cancer for all ages older than Italy: 2011–2012 survey. Epidemiol Prev.
CHAPTER 5.5
SECTION 5
terns in the United States, 1974–2013. J 39(3 Suppl 1):115–25. PMID:26405783
40 years that is proportional to Natl Cancer Inst. 109(8):109. https://doi.
the observed incidence trends for org/10.1093/jnci/djw322 PMID:28376186
with the accumulation over time of ized by high microsatellite instability adenomas with these features have
epigenetic alterations, in the form of (MSI-high), are frequently associated a BRAF-mutated/CIMP-high/MSI-
global methylation of CpG islands with the development of cytological high molecular profile, it was sug-
(the CpG island methylator pheno- dysplasia [17]. Epigenetic silencing gested that a subset of sessile ser-
type [CIMP]), which cause silencing of p16 is associated with the devel- rated adenomas, with methylation of
of regulatory genes. Methylation of opment of high-grade dysplasia or the DNA repair gene MGMT, may be
the promoter region and suppression invasive carcinoma [18]. precursors of BRAF-mutated/CIMP-
of the mismatch repair gene MLH1, Although serrated colorectal high/microsatellite stable serrated
resulting in a phenotype character- cancers arising in sessile serrated colorectal cancers [18].

Sessile serrated adenomas may rarely methylated in traditional ser- rectal cancers generally have a fa-
have an indolent course in the early rated adenomas; this supports the vourable prognosis; this may relate
phase after BRAF mutation, with a hypothesis that traditional serrated to an immune response, because
rapid progression to invasive colo- adenomas are precursors of micro- these tumours are strongly infil-
rectal cancer after the development satellite stable or MSI-low colorec- trated by T lymphocytes, opening
of cytological dysplasia, which is tal cancers. Inactivation of p53 has up opportunities for immunotherapy
associated with the development been associated with the develop- [24]. BRAF mutation is also associ-
of MSI [17]. This hypothesis is sup- ment of high-grade dysplasia and ated with poorer survival within the
ported by the observation of a very invasive carcinoma [18]. MSI group [25]. MSI has been as-
low risk of BRAF-mutated colorec- Recent efforts using data on sociated with resistance to 5-fluoro-
tal cancers in people younger than RNA expression and immune re- uracil chemotherapy [24].
60 years, who, however, have a sponse have led to new classifi-
similar prevalence of sessile serrat- cations associated with survival,
which are undergoing validation
Risk factors
ed adenomas to older people [19].
The prevalence of sessile ser- [20]. Also, the detection of tumour Of 17.2 million disability-adjusted
rated adenomas in people at average mutational signatures on the basis life years due to colorectal cancer,
risk who undergo colonoscopy or of genome-wide data may yield 6.8 million (39.4%) are attributable
stool-based tests for occult blood – possible targets for prevention, to lifestyle factors [26]. This fraction
guaiac faecal occult blood test because specific signatures have appears to be fairly constant across
(gFOBT) or faecal immunochemi- been associated with particular different countries, irrespective of the
cal test (FIT) – has been reported to exposures [21]. However, tumour– large differences in colorectal cancer
be 2–7%. Sessile serrated adeno- node–metastasis (TNM) stage and risk. The available evidence supports
mas are located predominantly in markers associated with the chro- the association of diet, physical activ-
the proximal colon; they have a flat mosomal instability pathway and ity, and smoking with risk of colorec-
or sessile morphology, and they are the serrated pathway remain the tal cancer (Table 5.5.1) [27–31].
often covered by a mucus cap. These guides in clinical decision-making.
KRAS mutations have been as- Dietary and nutrient patterns
features interfere with their detection,
both by endoscopy (their subtle en- sociated with reduced survival and The analysis of dietary or nutrient
doscopic appearance and indistinct with treatment failure in patients with patterns has been developed as a
borders are associated with a higher advanced colorectal cancer who un- complementary approach to analy-
miss rate and a higher proportion of dergo targeted treatment with anti- ses of single foods or nutrients, to
incomplete excisions) and by gFOBT epidermal growth factor receptor adequately account for the interac-
or FIT (their morphology and the mu- (anti-EGFR) antibodies [22]. tion between food components and
cus cap are associated with a lower CIMP-high status and BRAF to characterize specific dietary hab-
likelihood to bleed) [17]. mutation have been associated with its in a more comparable way across
poor prognosis [23]. MSI-high colo- populations (see Chapter 2.6).
Traditional serrated adenoma
Traditional serrated adenomas Fig. 5.5.2. In high-income countries, a dietary pattern characterized by, among other
make up less than 1% of all serrated things, high intakes of fruits and vegetables, whole grains, and nuts and legumes is
lesions. Therefore, limited evidence associated with a lower risk of colorectal cancer.
is available about their epidemiol-
ogy and natural history. Traditional
serrated adenomas are located
predominantly in the distal colon
and have a polypoid morphology
and a villous component, similar to
advanced conventional adenomas.
Activation of the MAPK path-
way is more frequently associated
with mutation of the KRAS onco-
gene, although traditional serrated
adenomas may also have BRAF
mutation. Both CIMP-high and
CIMP-low phenotypes have been
described in different series; the
variance is probably related also to
differences in the panel of markers
used to define CIMP [18]. MLH1 is
348
Genetic susceptibility
Germline mutations or epimuta- Much of the heritable risk is offers the prospect of tailoring
tions of genes involved in colorectal probably explained by co-inher- colorectal cancer screening to an
carcinogenesis, which are also in- itance of low-penetrance genetic individual’s level of risk, thereby
volved in sporadic colorectal can- variants. Genome-wide associa- optimizing the use of screening
cer pathways, are associated with tion studies have so far identified resources. Assessments of the
hereditary syndromes. These syn- about 60 common single-nucleo- feasibility and cost–effectiveness
dromes can be divided into three tide polymorphisms that influence of this approach in the setting of
broad categories: (i) non-polyposis individual susceptibility to colorec- population-based screening are
syndromes, (ii) adenomatous pol- tal cancer [2]. Although the risk being planned.
yposis syndromes, and (iii) non-ad- associated with variation at each A recent report from a large
enomatous polyposis syndromes. locus is modest, risk genotypes prospective cohort study showed
They collectively account only for a are common in the population. It that a genetic risk score composed
small fraction of colorectal cancer has been suggested that develop- of 41 published, genome-wide sig-
risk attributable to genetic factors. ing genome-wide polygenic scores nificant single-nucleotide polymor-
These syndromes are character- may enable the identification of in- phisms for colorectal cancer did
ized by an increased risk of colo- dividuals with risk levels compara- not meaningfully improve model
rectal cancer during the individual’s ble to those of people with heredi- discrimination of two previously
lifetime. The estimated cumulative tary syndromes. validated risk prediction models
probability of developing the dis- Accounting for the interaction for colorectal cancer, and did not
ease by age 70 years ranges from between genetic and lifestyle-re- substantially influence the predict-
90% in familial adenomatous polyp- lated factors may present a chal- ed probabilities for 95% of partici-
osis to almost 0% in some variants lenge. However, the development pants [3]. These findings suggest
of Lynch syndrome [1]. A summary of risk prediction models that incor- that a genetic risk score for colo-
of these syndromes is presented in porate genetic risk scores together rectal cancer risk prediction may
Table B5.5.1. with other risk factor information have some additional practical
benefit only if it is applied to people
who are already predicted to be at
Table B5.5.1. Genetic syndromes associated with increased risk of colorectal cancer high risk, on the basis of existing
Syndrome Gene mutations Inheritance pattern models, rather than to people at
average risk.
Non-polyposis syndromes Implementing such an ap-
Lynch syndrome MLH1, MSH2, MSH6, Autosomal dominant proach also requires taking into ac-
PMS2, and EPCAM count the confidentiality and ethi-
Familial colorectal cancer Not defined
cal implications of genetic testing,
(previously known as familial and this consideration influences
colorectal cancer type X) the acceptability of this approach.
Adenomatous polyposis syndromes
References
Familial adenomatous polyposis APC Autosomal dominant 1. IARC (2019). Colorectal cancer screen-
Attenuated familial adenomatous
ing. IARC Handb Cancer Prev. 17:1–
polyposis
CHAPTER 5.5
SECTION 5
300. Available from: http://publications.
iarc.fr/573.
MUTYH-associated polyposisa MUTYH Autosomal recessive
2. Schmit SL, Edlund CK, Schumacher FR,
Non-adenomatous polyposis syndromes Gong J, Harrison TA, Huyghe JR, et al.
(2019). Novel common genetic suscep-
Peutz–Jeghers syndromeb SKT11 Autosomal dominant
tibility loci for colorectal cancer. J Natl
Cancer Inst. 111(2):146–57. https://doi.
Cowden syndrome (PTEN PTEN Autosomal dominant
org/10.1093/jnci/djy099 PMID:29917119
hamartoma tumour syndrome)
3. Smith T, Gunter MJ, Tzoulaki I, Muller
Juvenile polyposis syndrome SMAD4 and BMPR1A Autosomal dominant DC (2018). The added value of genetic
information in colorectal cancer risk pre-
Serrated polyposis syndromec GREM1 and MUTYH
diction models: development and evalu-
a
The phenotype is highly variable, presenting also with both adenomatous and hyperplastic polyps. ation in the UK Biobank prospective co-
hort study. Br J Cancer. 119(8):1036–9.
b
Genetic testing may be negative in up to 50% of the cases that meet the clinical criteria.
ht tps://doi.org/10.1038/s 41416 - 018 -
c
Not universal. Associated with increased risk of sporadic mismatch repair-deficient colorectal cancer. 0282-8 PMID:30323197

Table 5.5.1. Risk factors and protective factors for colorectal cancera
Evidence grade b Reduced risk Increment/contrast Increased risk Increment/contrast
Strong – convincing Physical activityc,d Higher versus lower Processed meat per 50 g/day
levels
Alcoholic beverages > about 2 drinks a day
Body fatness
Adult attained heighte
Strong – probable Whole grains per 90 g/day Red meat > 100 g/day
Foods containing dietary per 10 g/day

fibre
Dairy products per 400 g/day overall

(milk: 200 g/day)
Calcium supplements > 200 mg/day
Limited – suggestive Foods containing Low intake of starchy

vitamin C vegetables
Vitamin D f Low intake of fruits < 100 g/day
Fish per 100 g/day Foods containing haem

iron
Multivitamin supplements
Sufficient evidence Smoking g Never-smoker/former

smoker/current smoker
a
Risk factors and protective factors for colorectal adenomas are generally consistent with those identified for colorectal cancer. Also, risk factors and
protective factors for conventional adenomas and for sessile serrated adenomas generally overlap; the main difference is the higher risk of sessile serrated
adenoma in women, as opposed to the higher prevalence of conventional adenomas in men [29].
b
Evidence grade is based on the classification used in the WCRF/AICR report [27] and, for smoking only, in the IARC Monographs [28].
c
Protective effect observed for colon cancer only [30].
d
Measured as: active versus sedentary; vigorous or high versus low; < 10 metabolic equivalent (MET) hours/week, or < 1 hour/week, or < 10 minutes per
day, compared with higher levels.
e
Probably as a marker of factors that influence growth in early life.
f
Epidemiological studies have consistently shown an inverse association between circulating concentrations of vitamin D and risk of colorectal cancer,
although findings from recent Mendelian randomization analyses do not necessarily support a causal relationship [31].
g
Current smoking was associated with risk of rectal cancer and proximal colorectal cancer, but not of distal colorectal cancer [30].
Fig. 5.5.3. In high-income countries, a dietary pattern characterized by high intakes of Two distinct dietary patterns have
red meat, processed meat, sugar-sweetened beverages, refined grains, desserts, and been associated with risk of colo-
potatoes is associated with a higher risk of colorectal cancer. rectal cancer, and the association is
stronger for men than for women. A
“healthy” pattern, which is associated
with a lower risk of colorectal cancer,
is characterized by high intakes of
fruits and vegetables, whole grains,
nuts and legumes, fish and other
seafood, and milk and other dairy
products. In contrast, an “unhealthy”
pattern, which is associated with a
higher risk of colorectal cancer, is
characterized by high intakes of red
meat, processed meat, sugar-sweet-
ened beverages, refined grains, des-
serts, and potatoes [32].
In the European Prospective
Investigation into Cancer and Nutrition
(EPIC) cohort, higher scores on two
nutrient patterns have been associat-
ed with a reduction in risk of colorectal
350
cancer (mainly for lesions located in tal cancer burden, the scale of the metastases, and are thought to be
the proximal colon) [33]. One of the obesity epidemic and the high in- pro-inflammatory [38]. Atopobium, a
nutrient patterns is characterized by a cidence of colorectal cancer may gram-positive anaerobic bacterium,
high variety of vitamins and minerals, necessitate more direct preventive is associated with Crohn disease and
and the other is characterized by vi- interventions that target people at was reported to inhibit colonocyte
tamin B12, calcium, phosphorus, ribo- higher risk. apoptosis in vitro [39]. These studies
flavin, cholesterol, and total proteins. are consistent with microbiotic imbal-
Microbiota ance (known as dysbiosis) leading
Adiposity and body fatness There is a growing body of experi- to a pro-inflammatory microenviron-
Overweight, obesity, and type 2 mental and observational evidence ment, which is conducive to colorec-
diabetes (see Chapter 2.7) are es- implicating the gut microbiome in the tal tumorigenesis. However, caution
tablished risk factors for colorectal development of colorectal cancer is required in the interpretation of
cancer, and it has been estimated (see Chapter 3.10). However, human case–control and cross-sectional
that they may account for more than studies linking variation in the gut studies, because of the potential of
10% of cases worldwide [34]. Given microbiome with colorectal cancer reverse causality [40].
the worldwide rising prevalence of are limited, and more are needed.
obesity and type 2 diabetes, these A small case–control study with Prevention and screening
diseases are likely to have significant available faecal samples demon-
impacts on colorectal cancer inci- strated differences between colorec- Screening
dence in the future [35]. tal cancer cases and controls in the The available evidence suggests that
Public health strategies aimed relative abundance of bacterial taxa, screening, with stool-based tests for
at reducing the prevalence of obe- with enrichment of Bacteroidetes and occult blood (gFOBT or FIT) or with
sity, promoting physical activity, depletion of Firmicutes in cases [36]. endoscopic methods, is associated
and discouraging the consumption In addition, increased carriage of the with a reduction in colorectal cancer
of high-energy, obesogenic foods genera Fusobacterium, Atopobium, incidence and mortality (Table 5.5.2)
are gradually being implemented in and Porphyromonas has been asso- [41,42].
many regions of the world. Although ciated with colorectal cancer [36,37]. Colorectal cancer incidence
such strategies could, if successful, Fusobacterium are prevalent in co- and mortality have been observed
lead to a reduction in the colorec- lon tissue, are maintained in distal to decline in countries where the
Table 5.5.2. Evidence supporting colorectal cancer screening methods
Screening method a Evidence for reduction in mortality/incidence Benefit–harm Screening

ratio interval
Target age range
Guaiac faecal occult blood test (gFOBT) Sufficient evidence for reduction in mortality Sufficient 2 years
Evidence suggestive of a lack of effect for evidence 50–60 to 75 years
reduction in incidence
Higher-sensitivity guaiac faecal occult Sufficient evidence for reduction in mortality Sufficient 1 or 2 years
blood test (gFOBT) (with rehydration) Limited evidence for reduction in incidence evidence 50–60 to 75 years
Faecal immunochemical test for Sufficient evidence for reduction in mortality Sufficient 2 years
haemoglobin (FIT) Limited evidence for reduction in incidence evidence 50–60 to 75 years
CHAPTER 5.5
SECTION 5
Sigmoidoscopy Sufficient evidence for reduction in mortality Sufficient Once in lifetimeb
Sufficient evidence for reduction in incidence evidence
Colonoscopy Sufficient evidence for reduction in mortality Sufficient Once in lifetimec

Sufficient evidence for reduction in incidence evidence
Computed tomography (CT) Limited evidence for reduction in mortality Inadequate Once in lifetimed
colonography Limited evidence for reduction in incidence evidence
a
Evidence on newer techniques that have emerged recently was deemed insufficient. In particular, only one study was available assessing the accuracy
of a multitarget stool DNA test combined with the faecal immunochemical test (FIT); it showed an increased sensitivity for sessile serrated adenoma,
compared with FIT alone. Similarly, only one study has been conducted to assess the accuracy of a blood biomarker (methylated septin 9 DNA); it showed
a low sensitivity for advanced adenomas.
b
Screening trial included people 55 years or older, and current population-based programmes offer screening between age 55 years and age 59 years.
c
Available evidence supporting the colonoscopy screening test refers to people aged 50 years and older and suggests that the impact is lower in elderly
people (aged > 75 years).
d
Evidence about the optimal target age is limited.

implementation of interventions for with higher socioeconomic status creased risk of colorectal cancer (see
early detection started in the 1990s (see Chapter 4.5) [46]. In the USA, Chapter 6.4) [51]. In individuals at
already [43]. In addition, preliminary use of screening remains consis- average risk, calcium supplementa-
reports show a reduction in colorec- tently lower, independent of ethnic- tion (> 200 mg/day) was associated
tal cancer incidence, mortality, and ity and education level, in uninsured with a reduction in risk of colorectal
surgery rates after the introduction people and in those without access cancer [32], and use of aspirin (daily
of population-based screening pro- to primary care, because of eco- or alternate-day dose, ≥ 75 mg) ap-
grammes [42,44,45]. These find- nomic and organizational barriers peared to reduce colorectal cancer
ings confirm the beneficial impact of (see Chapter 4.6) [47]. incidence and mortality, after a la-
screening on the colorectal cancer Reports from organized gFOBT- tency of about 10 years, with a small
burden at the population level. based or FIT-based screening pro- reduction in all-cause mortality
However, screening rates in grammes still document lower par- within 10 years of initiating use [52].
adults aged 50–75 years remain ticipation in screening in the most In a recent network meta-analysis,
low, and non-adherence to recom- disadvantaged groups [48]. However, low-dose aspirin appeared to be as
mended protocols is an important screening rates are higher and the effective as gFOBT or sigmoidos-
attributable factor for colorectal can- gap by socioeconomic status is copy in reducing colorectal cancer
cer mortality, in particular in disad- smaller in settings with organized incidence and mortality, and more
vantaged groups. programmes than in settings with op- effective for cancers located in the
Trends in colorectal cancer mor- portunistic screening; this suggests proximal colon [53]. The cost–effec-
tality in the USA have been observed that implementing population-based tiveness of an approach combining
to be associated with socioeconom- screening can ensure the organiza- screening and chemoprevention still
ic status. This association, together tional framework for enhancing par- needs to be assessed.
with the timing of the implementation ticipation, while reducing inequities in
Primary prevention
of screening in the USA, is consistent access [49,50].
with the hypothesis that the gradient Preventive interventions aimed at
in screening uptake with socioeco- Chemoprevention promoting healthier lifestyles may
reduce the risk of colorectal can-
nomic status and the later adoption of There is some evidence that aspirin
cer, or may maintain the low risk in
screening in disadvantaged groups and cyclooxygenase 2 (COX-2) inhib-
those countries where industrial-
resulted in widening disparities in itors may reduce recurrence of ad-
ized lifestyles are not yet common.
colorectal cancer mortality – a dis- enomas and incidence of advanced
Such preventive measures may be
parity that is now in favour of groups adenomas in individuals at an in-
implemented at the population level
and/or at the individual level.
Fig. 5.5.4. This food from Ethiopia exemplifies a vegetable-based diet. Incidence Regular cancer screening of-
rates of colorectal cancer continue to be low in many countries, and in such countries fers the opportunity to convey health
industrialized lifestyles are typically not yet common. education messages, and the over-
all impact of primary prevention and
screening could reduce the inci-
dence of colorectal cancer by up to
60% in screenees who comply with
health education recommendations.
Studies assessing the impact of life-
style interventions proposed in the
screening setting showed that coun-
selling can be effective in encourag-
ing the adoption of healthier dietary
patterns, but not in promoting an
increase in physical activity or in
prompting smoking cessation [54].
In evaluating trends of colorec-
tal cancer risk, the impact of inter-
ventions not specifically designed
for colorectal cancer prevention,
but targeting multiple chronic dis-
eases sharing the same risk fac-
tors, should also be considered.
352
References
1. Bray F, Ferlay J, Soerjomataram I, Siegel 9. Jemal A, Ward EM, Johnson CJ, Cronin 19. Bettington M, Brown I, Rosty C, Walker N,
RL, Torre LA, Jemal A (2018). Global KA, Ma J, Ryerson B, et al. (2017). Annual Liu C, Croese J, et al. (2019). Sessile serrat-
cancer statistics 2018: GLOBOCAN es- report to the nation on the status of can- ed adenomas in young patients may have
timates of incidence and mortality world- cer, 1975–2014, featuring survival. J limited risk of malignant progression. J Clin
wide for 36 cancers in 185 countries. CA Natl Cancer Inst. 109(9):109. https://doi. Gastroenterol. 53(3):e113–6. https://doi.
Cancer J Clin. 68(6):394–424. https://doi. org/10.1093/jnci/djx030 PMID:28376154 org/10.1097/MCG.0000000000001014
org/10.3322/caac.21492 PMID:30207593 PMID:29570172
10. Guren MG, Kørner H, Pfeffer F, Myklebust
2. Fidler MM, Bray F, Vaccarella S, TA, Eriksen MT, Edna TH, et al. (2015). 20. Guinney J, Dienstmann R, Wang X, de
Soerjomataram I (2017). Assessing glob- Nationwide improvement of rectal cancer Reyniès A, Schlicker A, Soneson C, et
al transitions in human development and treatment outcomes in Norway, 1993– al. (2015). The consensus molecular
colorectal cancer incidence. Int J Cancer. 2010. Acta Oncol. 54(10):1714–22. https:// subtypes of colorectal cancer. Nat Med.
140(12):2709–15. https://doi.org/10.1002/ doi.org/10.3109/0284186X.2015.1034876 21(11):1350–6. https://doi.org/10.1038/
ijc.30686 PMID:28281292 PMID:25924970 nm.3967 PMID:26457759
3. Fitzmaurice C, Akinyemiju TF, Al Lami 11. Brouwer NPM, Bos ACRK, Lemmens 21. Alexandrov LB, Nik-Zainal S, Wedge DC,
FH, Alam T, Alizadeh-Navaei R, Allen C, VEPP, Tanis PJ, Hugen N, Nagtegaal ID, Aparicio SA, Behjati S, Biankin AV, et al.;
et al.; Global Burden of Disease Cancer et al. (2018). An overview of 25 years of Australian Pancreatic Cancer Genome
Collaboration (2018). Global, regional, incidence, treatment and outcome of Initiative; ICGC Breast Cancer Consortium;
and national cancer incidence, mortality, ICGC MMML-Seq Consortium; ICGC
colorectal cancer patients. Int J Cancer.
years of life lost, years lived with disabil- PedBrain (2013). Signatures of mutation-
143(11):2758–66. https://doi.org/10.1002/
ity, and disability-adjusted life-years for 29 al processes in human cancer. Nature.
ijc.31785 PMID:30095162
cancer groups, 1990 to 2016: a systematic 500(7463):415–21. https://doi.org/10.1038/
analysis for the Global Burden of Disease 12. Goss PE, Lee BL, Badovinac-Crnjevic T, nature12477 PMID:23945592
Study. JAMA Oncol. 4(11):1553–68. https:// Strasser-Weippl K, Chavarri-Guerra Y, St
22. Dahabreh IJ, Terasawa T, Castaldi PJ,
doi.org/10.1001/jamaoncol.2018.270 6 Louis J, et al. (2013). Planning cancer con-
Trikalinos TA (2011). Systematic review:
PMID:29860482 trol in Latin America and the Caribbean.
anti-epidermal growth factor receptor
4. Arnold M, Sierra MS, Laversanne M, treatment effect modification by KRAS
doi.org/10.1016/S1470-2045(13)70048-2
Soerjomataram I, Jemal A, Bray F (2017). mutations in advanced colorectal cancer.
PMID:23628188 Ann Intern Med. 154(1):37–49. https://
Global patterns and trends in colorectal
cancer incidence and mortality. Gut. 66(4): 13. Kingham TP, Alatise OI, Vanderpuye doi.org/10.7326/0003- 4819-154-1-2011
683–91. https://doi.org/10.1136/gutjnl-2015- V, Casper C, Abantanga FA, Kamara 01040-00006 PMID:21200037
310912 PMID:26818619 TB, et al. (2013). Treatment of cancer 23. Alwers E, Jia M, Kloor M, Bläker H,
in sub-Saharan Africa. Lancet Oncol. Brenner H, Hoffmeister M (2019).
5. Allemani C, Matsuda T, Di Carlo V, 14(4):e158–67. https://doi.org/10.1016/
Harewood R, Matz M, Nikšić M, et al.; Associations between molecular clas-
S1470-2045(12)70472-2 PMID:23561747 sifications of colorectal cancer and pa-
CONCORD Working Group (2018). Global
surveillance of trends in cancer survival tient survival: a systematic review. Clin
14. Karsa LV, Lignini TA, Patnick J, Lambert
2000-14 (CONCORD-3): analysis of in- Gastroenterol Hepatol. 17(3):402–410.e2.
R, Sauvaget C (2010). The dimensions
https://doi.org/10.1016/j.cgh.2017.12.038
dividual records for 37 513 025 patients of the CRC problem. Best Pract Res
PMID:29306042
diagnosed with one of 18 cancers from Clin Gastroenterol. 24(4):381–96. https://
322 population-based registries in 71 d o i . o r g / 10 .10 1 6 / j . b p g . 2 0 10 . 0 6 . 0 0 4 24. Gupta R, Sinha S, Paul RN (2018).
countries. Lancet. 391(10125):1023–75. PMID:20833343 The impact of microsatellite stabil-
https://doi.org/10.1016/S0140- 6736(17) ity status in colorectal cancer. Curr
33326-3 PMID:29395269 15. Fearon ER, Vogelstein B (1990). A genetic
Probl Cancer. 42(6):548–59. https://doi.
model for colorectal tumorigenesis. Cell.
org/10.1016/j.currproblcancer.2018.06.010
6. Sunkara V, Hébert JR (2015). The colo- 61(5):759–67. https://doi.org/10.1016/0092- PMID:30119911
rectal cancer mortality-to-incidence ratio 8674(90)90186-I PMID:2188735
as an indicator of global cancer screen- 25. Rosty C, Williamson EJ, Clendenning M,
ing and care. Cancer. 121(10):1563–9. 16. Zoratto F, Rossi L, Verrico M, Papa A, Walters RJ, Walsh MD, Win AK, et al.
h t t p s : / / d o i . o r g / 10 .10 0 2 / c n c r. 2 9 2 2 8 Basso E, Zullo A, et al. (2014). Focus on (2014). Re: Microsatellite instability and
CHAPTER 5.5
SECTION 5
PMID:25572676 genetic and epigenetic events of colo- BRAF mutation testing in colorectal can-
rectal cancer pathogenesis: implications cer prognostication. J Natl Cancer Inst.
7. Aarts MJ, Lemmens VE, Louwman for molecular diagnosis. Tumour Biol. 106(8):106. https://doi.org/10.1093/jnci/
MW, Kunst AE, Coebergh JW (2010). 35(7):6195–206. https://doi.org/10.1007/ dju180 PMID:25114271
Socioeconomic status and chang- s13277-014-1845-9 PMID:25051912
ing inequalities in colorectal cancer? 26. Institute for Health Metrics and Evaluation
A review of the associations with risk, 17. Haque T, Greene KG, Crockett SD (2014). (2018). Global Burden of Disease (GBD)
treatment and outcome. Eur J Cancer. Serrated neoplasia of the colon: what 2017 data. Available from: http://www.
46(15):2681–95. https://doi.org/10.1016/j. do we really know? Curr Gastroenterol healthdata.org/gbd.
ejca.2010.04.026 PMID:20570136 Rep. 16(4):380. https://doi.org/10.1007/
s11894-014-0380-6 PMID:24595617 27. WCRF/AICR (2018). Diet, nutrition,
8. Manser CN, Bauerfeind P (2014). Impact physical activity and colorectal cancer.
of socioeconomic status on incidence, 18. Bettington M, Walker N, Clouston A, Continuous Update Project Expert Report
mortality, and survival of colorectal Brown I, Leggett B, Whitehall V (2013). 2018. World Cancer Research Fund/
cancer patients: a systematic review. The serrated pathway to colorectal carci- American Institute for Cancer Research.
Gastrointest Endosc. 80(1):42–60.e9. noma: current concepts and challenges. Available from: https://www.aicr.org/con
https://doi.org/10.1016/j.gie.2014.03.011 Histopathology. 62(3):367–86. https://doi. tinuous-update-project/reports/colorec
PMID:24950641 org/10.1111/his.12055 PMID:23339363 tal-cancer-2017-report.pdf.

28. IARC (2012). Personal habits and in- 38. Bullman S, Pedamallu CS, Sicinska E, 48. de Klerk CM, Gupta S, Dekker E, Essink-
door combustions. IARC Monogr Eval Clancy TE, Zhang X, Cai D, et al. (2017). Bot ML; Expert Working Group ‘Coalition
Carcinog Risks Hum. 100E:1–575. Analysis of Fusobacterium persistence to reduce inequities in colorectal can-
Available from: http://publications.iarc. and antibiotic response in colorectal cer screening’ of the World Endoscopy
fr/122 PMID:23193840 cancer. Science. 358(6369):1443–8. Organization (2018). Socioeconomic and
https://doi.org/10.1126/science.aal5240 ethnic inequities within organised colorectal
29. Haque TR, Bradshaw PT, Crockett SD PMID:29170280 cancer screening programmes worldwide.
(2014). Risk factors for serrated polyps of Gut. 67(4):679–87. https://doi.org/10.1136/
the colorectum. Dig Dis Sci. 59(12):2874– 39. Chen W, Liu F, Ling Z, Tong X, Xiang gutjnl-2016-313311 PMID:28073892
89. https://doi.org/10.1007/s10620-014- C (2012). Human intestinal lumen and
3277-1 PMID:25030942 mucosa-associated microbiota in pa- 49. Carrozzi G, Sampaolo L, Bolognesi L,
tients with colorectal cancer. PLoS One. Sardonini L, Bertozzi N, Giorgi Rossi P,
30. Murphy N, Ward HA, Jenab M, Rothwell
7(6):e39743. https://doi.org/10.1371/jour et al.; Regional and local PASSI coordina-
JA, Boutron-Ruault MC, Carbonnel F, et
nal.pone.0039743 PMID:22761885 tors (2015). Cancer screening uptake: as-
al. (2019). Heterogeneity of colorectal
sociation with individual characteristics,
cancer risk factors by anatomical sub- 40. Amitay EL, Werner S, Vital M, Pieper geographic distribution, and time trends
site in 10 European countries: a multi- DH, Höfler D, Gierse IJ, et al. (2017). in Italy. Epidemiol Prev. 39(3 Suppl 1):9–
national cohort study. Clin Gastroenterol Fusobacterium and colorectal cancer: 18. PMID:26405772
Hepatol. 17(7):1323–1331.e6. https:// causal factor or passenger? Results
d o i . o r g / 1 0 .1 0 1 6 / j . c g h . 2 0 1 8 . 0 7. 0 3 0 from a large colorectal cancer screen- 50. Levin TR, Corley DA, Jensen CD,
PMID:30056182 ing study. Carcinogenesis. 38(8):781–8. Schottinger JE, Quinn VP, Zauber AG, et
https://doi.org/10.1093/carcin/bgx053 al. (2018). Effects of organized colorectal
31. Dimitrakopoulou VI, Tsilidis KK, Haycock
PMID:28582482 cancer screening on cancer incidence and
PC, Dimou NL, Al-Dabhani K, Martin RM,
et al.; GECCO Consortium; PRACTICAL mortality in a large community-based pop-
41. Armaroli P, Villain P, Suonio E, Almonte M, ulation. Gastroenterology. 155(5):1383–
Consortium; GAME-ON Network Anttila A, Atkin WS, et al. (2015). European
(CORECT, DRIVE, ELLIPSE, FOCI- 1391.e5. https://doi.org/10.1053/j.gastro.
Code Against Cancer, 4th edition: cancer 2018.07.017 PMID:30031768
OCAC, TRICL-ILCCO) (2017). Circulating screening. Cancer Epidemiol. 39(Suppl
vitamin D concentration and risk of seven 1):S139–52. https://doi.org/10.1016/j. 51. Lang M, Gasche C (2015). Chemopreven
cancers: Mendelian randomisation study. canep.2015.10.021 PMID:26596722 tion of colorectal cancer. Dig Dis. 33(1):
BMJ. 359:j4761. https://doi.org/10.1136/ 58–67. https://doi.org/10.1159/000366037
bmj.j4761 PMID:29089348 42. IARC (2019). Colorectal cancer screen- PMID:25531498
ing. IARC Handb Cancer Prev. 17:1–300.
32. Tabung FK, Brown LS, Fung TT (2017). 52. Chubak J, Whitlock EP, Williams SB,
Available from: http://publications.iarc.
Dietary patterns and colorectal cancer Kamineni A, Burda BU, Buist DS, et al.
fr/573.
risk: a review of 17 years of evidence (2016). Aspirin for the prevention of can-
(2000–2016). Curr Colorectal Cancer Rep. 43. Yang DX, Gross CP, Soulos PR, Yu cer incidence and mortality: systematic
13(6):440–54. https://doi.org/10.1007/ JB (2014). Estimating the magnitude evidence reviews for the U.S. Preventive
s11888-017-0390-5 PMID:29399003 of colorectal cancers prevented dur- Services Task Force. Ann Intern Med.
ing the era of screening: 1976 to 2009. 164(12):814–25. https://doi.org/10.7326/
33. Moskal A, Freisling H, Byrnes G, Assi
Cancer. 120(18):2893–901. https://doi. M15-2117 PMID:27064482
N, Fahey MT, Jenab M, et al. (2016).
org/10.1002/cncr.28794 PMID:24894740
Main nutrient patterns and colorectal
53. Emilsson L, Holme Ø, Bretthauer M, Cook
cancer risk in the European Prospective 44. Fedeli U, Zorzi M, Urso ED, Gennaro N, NR, Buring JE, Løberg M, et al. (2017).
Investigation into Cancer and Nutrition Dei Tos AP, Saugo M (2015). Impact of fe- Systematic review with meta-analysis: the
study. Br J Cancer. 115(11):1430–40. cal immunochemical test-based screen- comparative effectiveness of aspirin vs.
ht t ps: //doi.org /10.10 3 8 / bjc. 2016. 3 3 4 ing programs on proximal and distal screening for colorectal cancer preven-
PMID:27764841 colorectal cancer surgery rates: a natural tion. Aliment Pharmacol Ther. 45(2):193–
34. Pearson-Stuttard J, Zhou B, Kontis V, multiple-baseline experiment. Cancer. 204. https://doi.org/10.1111/apt.13857
Bentham J, Gunter MJ, Ezzati M (2018). 121(22):3982–9. https://doi.org/10.1002/ PMID:27859394
Worldwide burden of cancer attributable cncr.29623 PMID:26264471
to diabetes and high body-mass index: 54. Senore C, Giordano L, Bellisario C, Di
45. McClements PL, Madurasinghe V, Stefano F, Segnan N (2012). Population
a comparative risk assessment. Lancet Thomson CS, Fraser CG, Carey FA,
Diabetes Endocrinol. 6(6):e6–15. https:// based cancer screening programmes as a
Steele RJ, et al. (2012). Impact of the teachable moment for primary prevention
doi.org/10.1016/S2213-8587(18)30150-5 UK colorectal cancer screening pilot
PMID:29803268 interventions. A review of the literature.
studies on incidence, stage distribution Front Oncol. 2:45. https://doi.org/10.3389/
35. Murphy N, Jenab M, Gunter MJ (2018). and mortality trends. Cancer Epidemiol. fonc.2012.00045 PMID:22649789
Adiposity and gastrointestinal cancers: 36(4):e232–42. https://doi.org/10.1016/j.
epidemiology, mechanisms and future di- canep.2012.02.006 PMID:22425027
rections. Nat Rev Gastroenterol Hepatol.
46. Breen N, Lewis DR, Gibson JT, Yu M,
15(11):659–70. https://doi.org/10.1038/
Harper S (2017). Assessing disparities in
s41575-018-0038-1 PMID:29970888
colorectal cancer mortality by socioeco-
36. Ahn J, Sinha R, Pei Z, Dominianni C, Wu nomic status using new tools: health dis-
J, Shi J, et al. (2013). Human gut microbi- parities calculator and socioeconomic quin-
ome and risk for colorectal cancer. J Natl tiles. Cancer Causes Control. 28(2):117–25.
Cancer Inst. 105(24):1907–11. https://doi. https://doi.org/10.1007/s10552-016-0842-2
org/10.1093/jnci/djt300 PMID:24316595 PMID:28083800
37. Arthur JC, Perez-Chanona E, Mühlbauer 47. White A, Thompson TD, White MC,
M, Tomkovich S, Uronis JM, Fan TJ, et Sabatino SA, de Moor J, Doria-Rose
al. (2012). Intestinal inflammation tar- PV, et al. (2017). Cancer screening test
gets cancer-inducing activity of the mi- use – United States, 2015. MMWR Morb
crobiota. Science. 338(6103):120–3. Mortal Wkly Rep. 66(8):201–6. https://
https://doi.org/10.1126/science.1224820 d o i . o r g / 10 .15 5 8 5 / m m w r. m m 6 6 0 8 a1
PMID:22903521 PMID:28253225
354
5.6
Liver cancer
An infectious disease
for many communities
Chien-Jen Chen Isabelle Chemin (reviewer)
Zdenko Herceg (reviewer)
Tatsuhiro Shibata (reviewer)
measure of aflatoxin B1 expo- The Global Burden of Disease

SUMMARY sure over several months. Study reported that from 1990 to
2015, there was a 75% increase in
●● From 1990 to 2015, there was ●● Viral hepatitis control is includ-
global cases of incident liver cancer,
a 75% increase in global cas- ed within the United Nations
of which 47% could be attributed
es of incident liver cancer, of Sustainable Development Goals.
to changing population age struc-
which 47% could be attributed The hepatitis B virus vaccine has
tures, 35% to population growth,
to changing population age high efficacy and cost–effective-
and −8% to decreasing age-specif-
structures, 35% to population ness to prevent hepatocellular ic incidence rates. Globally, hepa-
growth, and −8% to decreasing carcinoma. titis B virus (HBV) infection was
age-specific incidence rates. responsible for 33% of deaths from
liver cancer, alcohol consumption
●● Genetic modifications observed Primary liver cancer is a group of for 30%, hepatitis C virus (HCV) in-
in liver cancer development in- pathologically heterogeneous ma- fection for 21%, and other causes
clude alterations at TP53, MYC, lignancies [1]. It includes mainly for 16%, with significant variation
WNT, CTNNB1 (β-catenin), and (~80%) hepatocellular carcinoma in the underlying etiologies among
other genes that mediate cell- (HCC), as well as intrahepatic chol- regions and countries [3].
cycle regulation, telomere sta- angiocarcinoma, mucinous cystic A recent review documented that
bility, epigenetic regulation, and neoplasms, intraductal papillary both the incidence of and mortality
chromatin remodelling. biliary neoplasms, hepatoblastoma from liver cancer have declined sig-
●● The incidence of liver cancer in children, angiosarcoma, and nificantly in the past two decades
and the prevalence of infec- other types, with different under- after the launch of the first HBV im-
lying etiologies and carcinogenic munization programme in the world
tion with hepatitis B virus and
mechanisms. in 1984 and the first chronic viral
hepatitis C virus are consistent-
ly high in East and South-East hepatitis therapy programme in the
Asia and sub-Saharan Africa. Epidemiology world in 2003 [4].
Most of the burden of disease
●● Ethanol-induced liver injury re- In 2018, liver cancer was the sixth
from HBV infection comes from in-
CHAPTER 5.6
SECTION 5
sults in fibrosis and cirrhosis, most common cancer and the fourth
fections acquired before age 5 years.
which predisposes to the devel- most common cause of cancer
A significant decrease in the global
opment of liver cancer. Alcohol death worldwide [2]. The cumulative
incidence of liver cancer is expected
acts synergistically with chron- incidence of liver cancer from birth
in the future, because the worldwide
to age 75 years was 1.6% for males
ic viral hepatitis and tobacco prevalence of chronic HBV infection
and 0.6% for females, and the cu-
use in causing hepatocellular in children younger than 5 years has
mulative mortality from liver cancer
carcinoma. been reduced dramatically by HBV
was 1.5% for males and 0.5% for
vaccination programmes.
●● There is a dose–response rela- females. There is substantial geo-
tionship between risk of hepa- graphical variation in liver cancer
tocellular carcinoma and in- incidence and mortality globally. Genetics and genomics
creasing serum level of aflatoxin Age-standardized rates in Africa Hereditary diseases that are as-
B1–albumin adducts, a biomark- and Asia are 2–3 times those in the sociated with an increased risk of
er that provides a cumulative Americas, Europe, and Oceania. HCC include haemochromatosis,
Chapter 5.6 • Liver cancer 355

α-1-antitrypsin deficiency, acute in- tion, alcohol consumption, aflatoxin

termittent porphyria, and porphyria exposure, liver fluke infection, and FUNDAMENTALS
cutanea tarda. Although the familial obesity (Table 5.6.1). The incidence
tendency of liver cancer may be at- of liver cancer and the prevalence ■■ Liver cancer includes mainly
tributable to common environmental of HBV and HCV infection are con- hepatocellular carcinoma,
factors shared by family members, sistently high in East and South- as well as intrahepatic
such as HBV infection, HCV infec- East Asia and sub-Saharan Africa. cholangiocarcinoma and
tion, liver fluke infection, alcohol However, the relative etiological hepatoblastoma.
consumption, and aflatoxin expo- proportion of HBV and HCV in HCC
sure, the familial tendency remains varies in different countries. For ■■ Major etiological factors for
significant after adjustment for these example, HBV is more important in liver cancer include hepatitis B
environmental factors, suggesting China and the Republic of Korea, virus infection, hepatitis C virus
that common genes shared by fam- HCV is more important in Japan, infection, alcohol consumption,
ily members also play an important and both HBV and HCV are impor- aflatoxin exposure, liver
role. For example, genetic polymor- tant in Mongolia. fluke infection, obesity, and
phisms of the sodium taurocholate Alcohol consumption (see Chap several genetic diseases. The
co-transporting peptide (NTCP, an ter 2.3) is the most prevalent risk global variation in liver cancer
HBV receptor), human leukocyte factor for HCC in eastern Europe, incidence rates coincides with
antigen (HLA), interferon lambda central Europe, southern Latin
the geographical distribution of
(IFNL) genes, metabolism enzymes, America, southern sub-Saharan
its major causes.
oncogenes, tumour suppressor Africa, Australia, and North America.
genes, and the androgen receptor Aflatoxin exposure (see Chapter 2.8) ■■ Hepatocarcinogenesis is a
are associated with risk of HCC [4]. is ubiquitous in many of the lowest-in- multistage process with a
Numerous somatic genetic alter- come populations worldwide, show- multifactorial etiology of host–
ations have been observed in HCC, ing a synergistic effect with HBV on environment interactions.
including mutations, copy number al- HCC. Liver fluke infection, a major
terations, and intra- and inter-chromo- risk factor for intrahepatic cholangio- ■■ Hepatitis B immunization,
somal rearrangements [5]. Frequent carcinoma, is prevalent only in parts antiviral therapy for chronic
alterations are at genes that play key of East Asia, including Thailand, viral hepatitis, reduction
roles in cancer development (TP53, Lao People’s Democratic Republic, in aflatoxin exposure, and
MYC, and CTNNB1 [β-catenin]), cell- China, the Republic of Korea, the elimination of liver fluke
cycle regulation (CCND1, CDKN2A, Russian Federation, and Viet Nam. infection have been well
and RB1), telomere stability (TERT), documented to lower the risk
Hepatitis virus infection
epigenetic regulation (IDH1 and of liver cancer.
IDH2), and chromatin remodelling Both HBV infection and HCV infec-
(ARID1, ARID2, MLL, BAP, and tion have been classified by the ■■ When the relevant options are
EZH2). Alterations are frequent in the IARC Monographs as carcinogenic available and affordable, liver
following 11 pathways: telomerase to humans; they cause HCC and cancer can be detected early,
reverse transcriptase (TERT), WNT/ intrahepatic cholangiocarcinoma. by seromarkers and imaging
β-catenin, PI3K/AKT/mTOR, TP53/ The estimated global number of technology, and can be treated
cell cycle, mitogen-activated protein chronic infections in 2015 was promptly, by surgical resec-
kinase (MAPK), hepatic differentia- 257 million for HBV and 71 million tion, transplantation, ablation,
tion, epigenetic regulation, chromatin for HCV [4]. In the natural history embolization, radiotherapy, tar-
remodelling, oxidative stress, inter- of HBV infection, about 10–20% of
geted therapy, chemotherapy,
leukin 6 (IL-6)/JAK/STAT, and trans- people with HBV infection will be-
and immunotherapy.
forming growth factor β (TGF-β). The come chronic carriers of HBV, de-
total mutation burden is moderate, pending on the age at infection.
and hypermutated cases, which are Spontaneous seroclearance of
expected to respond to immunothera- HBV e antigen (HBeAg), HBV DNA, genotype C or basal core promoter
py, are not common [6,7]. Epigenetic and even HBV surface antigen (BCP) A1762T/G1764A double mu-
silencing of CDKN2A, HHIP, CPS1, (HBsAg) may occur sequentially tations, and co-infection with HCV
and other tumour suppressor genes in patients with chronic HBV infec- or HIV. In the Risk Evaluation of
has also been reported [8,9]. tion (Fig. 5.6.2). Seroclearance of Viral Load Elevation and Associated
HBeAg, HBV DNA, and HBsAg may Liver Disease/Cancer (REVEAL)
lead to a decreased risk of HCC [10]. study, for patients with chronic
Etiology HCC occurs mostly in patients hepatitis B the lifetime (ages 30–
The major etiological factors for liver with HBeAg-seropositive status or 75 years) cumulative incidence
cancer are HBV infection, HCV infec- high viral load, infection with HBV of HCC was 27% for men and 8%
356
Table 5.6.1. Major etiological factors for liver cancer with their biomarkers and related major genes
Etiological factor Cancer type Biomarkers Related major genes

Hepatitis B virus infection Hepatocellular carcinoma HBsAg/HBeAg serostatus NTCP
Intrahepatic cholangiocarcinoma Viral load (HBV DNA) HLA
Genotypes/mutant types
Serum HBsAg level
Hepatitis C virus infection Hepatocellular carcinoma Anti-HCV IFNL3
Intrahepatic cholangiocarcinoma Viral load (HCV RNA) HLA
Genotypes/mutant types
Alcohol consumption Hepatocellular carcinoma Frequency ADH
Quantity ALDH
Duration/starting age
Aflatoxin exposure Hepatocellular carcinoma Metabolites in urine TP53
Guanine adducts GST M1/T1
Albumin adducts
Liver fluke infection Intrahepatic cholangiocarcinoma Eggs in faeces –
Obesity Hepatocellular carcinoma Body mass index Adiponectin
Waist circumference
Anti-HCV, hepatitis C antibody; HBeAg, HBV e antigen; HBsAg, HBV surface antigen; HBV, hepatitis B virus; HCV, hepatitis C virus.
for women. The AA genotype of cause most people who are infected the REVEAL study, for patients with
the S267F (rs2296651) variant on have no or mild symptoms. Most chronic hepatitis C the lifetime (ages
NTCP was found to be associated asymptomatic infections progress to 30–75 years) cumulative incidence
with HBsAg-seropositive status, and chronic hepatitis, with the patient not of HCC was 24% for men and 17%
the GA or AA genotype was associ- being aware of this until end-stage for women. Co-infection with HBV
ated with a low risk of progression liver diseases, including cirrhosis may increase the lifetime cumulative
to cirrhotic and non-cirrhotic HCC in and HCC, occur. Spontaneous clear- risk of HCC to 38% for men and 27%
patients with chronic hepatitis B [11]. ance of HCV RNA occurs in about for women. Polymorphisms near the
HCV infection is infrequently di- 8–36% of patients with chronic hep- IFNL3 gene (formerly known as
agnosed during the acute phase, be- atitis C without antiviral treatment. In IL28B) are associated with sponta-
neous clearance of HCV RNA and
reduced risk of HCC [12]. In particu-
Fig. 5.6.1. Chinese liver fluke. Human liver fluke infection, a major risk factor for lar, the TT variant of rs8099917 near
intrahepatic cholangiocarcinoma, is prevalent in parts of East Asia. IFNL3 is significantly associated
with increased spontaneous clear-
ance of HCV RNA and decreased
risk of HCC.
HLA also plays an important
role in the progression of chronic
CHAPTER 5.6
SECTION 5
hepatitis C. For example, eight sin-
gle-nucleotide polymorphisms near
HLA-DQB1 are associated with risk
of HCC in patients with HCV geno-
type 1 infection. DQB1*03:01 has a
protective effect, and DQB1*06:02
increases the risk of HCC [13].
Hepatocarcinogenesis caused
by infection with HBV or HCV is a
multistage process with a multifac-
torial etiology (Fig. 5.6.3). Infection
with HBV or HCV also causes in-
trahepatic cholangiocarcinoma, at
a much lower incidence than HCC.

Fig. 5.6.2. In the natural history of chronic hepatitis B, there are milestone transitions of three hepatitis B virus (HBV) seromarkers.
The serum HBV DNA level (viral load) remains high during the immune tolerance phase. Both HBV e antigen (HBeAg) seroclearance
and anti-HBe seropositivity occur after a period of elevation of serum level of alanine aminotransferase (ALT), a seromarker of liver
inflammation, during the immune clearance phase. The serum HBV DNA level (viral load) may remain high or may gradually decline
in patients with HBeAg-seronegative status. Seroclearance of HBV surface antigen (HBsAg) may occur after the serum HBV DNA
level becomes undetectable during the residual phase. HCC, hepatocellular carcinoma.
Alcohol consumption stantially increased risk of HCC in developed to estimate exposure to

Alcohol consumption has been clas- obese alcohol drinkers compared aflatoxins, particularly aflatoxin B1.
sified by the IARC Monographs as with non-obese never-drinkers [14]. Aflatoxin exposure increases the
carcinogenic to humans; it causes Polymorphisms of enzymes in- risk of cirrhosis and HCC in pa-
HCC. Ethanol as a solvent may involved in alcohol metabolism (see tients with chronic hepatitis B [16].
crease the exposure of hepatocytes Chapter 3.3), including alcohol de- Aflatoxin exposure also increases
to carcinogens such as 4-amino- hydrogenase 1B (ADH1B) and al- the risk of HCC in patients with
biphenyl and polycyclic aromatic dehyde dehydrogenase 2 (ALDH2), chronic hepatitis C and in habitual
hydrocarbons in tobacco smoke. were found to have significant effects alcohol drinkers without chronic vi-
Ethanol may also be converted by on risk of HCC, mediated through al- ral hepatitis [17].
alcohol dehydrogenase into carcino- cohol consumption [15]. Genotypes There is a dose–response re-
genic acetaldehyde. of both enzymes were associated lationship between risk of HCC and
Ethanol-induced liver injury re- with the frequency and quantity of al- increasing serum level of aflatoxin
sults in fibrosis and cirrhosis, which cohol consumption, and with the de- B1–albumin adducts, a biomarker
predisposes to the development velopment of subsequent HCC. that provides a cumulative measure
of HCC [1]. Alcohol acts synergis- of aflatoxin B1 exposure over several
tically with chronic viral hepatitis Aflatoxin months. Glutathione S-transferase
and tobacco use in causing HCC. A Aflatoxin has been classified by the (GST) M1 and T1 are the enzymes in-
synergistic effect on HCC between IARC Monographs as carcinogenic volved in the detoxification of aflatox-
alcohol consumption and obesity to humans; it causes HCC. Urinary ins. The increasing risk of HCC with
has been reported, showing a sub- and serum biomarkers have been aflatoxin exposure is significant in
358
Fig. 5.6.3. The progression from self-limited hepatitis B virus (HBV) and hepatitis C virus (HCV) infection through chronic hepatitis
and cirrhosis to hepatocellular carcinoma (HCC) is a multistage pathogenic process driven by the interaction among viral, host, and
environmental factors. Viral co-factors include viral load and genotype of HBV and HCV. Environmental co-factors for virus-related
HCC include alcohol consumption, aflatoxin exposure, cigarette smoking, low intake of carotenes and selenium, and others. Host
co-factors include age, sex, family history of HCC, obesity, serum level of alanine aminotransferase (ALT, a seromarker of hepatic
inflammation), fibrosis score, serum testosterone level, and genetic polymorphisms of metabolism enzymes, oncogenes, tumour
suppressor genes, hormone-related genes, immunity-related genes, inflammation-related genes, and others. F, female; HBeAg,
HBV e antigen; M, male.
Elevated viral load
HBeAg
HBV
status
Genotype
Viral factors
Host characteristics
and
environmental exposure Gender
Age
(M > F)
Hepatic
Alcohol
inflammation
consumption
/fibrosis
Family history
Fig. 5.6.4. A liver with cirrhosis. Cirrhosis induced by alcohol consumption predisposes patients with chronic hepatitis B with
to the development of hepatocellular carcinoma. null genotypes of GST M1 or T1 (i.e.
without detoxification capability), but
not in those with non-null genotypes.
The TP53 tumour suppressor
CHAPTER 5.6
SECTION 5
gene is critically important for the
regulation of the cell cycle and the
maintenance of genomic integrity.
A specific mutation at codon 249 in
exon 7 of TP53 has been associ-
ated with aflatoxin B1-induced HCC.
Liver flukes
Both Opisthorchis viverrini infection
and Clonorchis sinensis infection
have been classified by the IARC
Monographs as carcinogenic to
humans; they cause intrahepatic
cholangiocarcinoma. Globally, the

estimated number of infections with er HBsAg seroclearance rate and a recent study, long-term exposure to
O. viverrini is at least 10 million and persistently higher serum levels of PM2.5 was found to increase the risk
with C. sinensis is at least 35 mil- HBV DNA [19]. There is a dose– of liver cancer mediated by serum
lion [1]. The spread of these flukes response relationship between in- ALT level, after adjustment for age,
is restricted by the distribution of two creasing adiponectin levels and risk sex, alcohol consumption, cigarette
definitive hosts other than humans – of cirrhosis and HCC in patients smoking, and HBV and HCV infec-
particular species of snails and cypri- with chronic hepatitis B. tion [21]. However, this finding needs
nid fish – and by the cultural practice Diabetes increases risk of HCC further scrutiny.
of eating raw fish. The transmission with or without chronic viral hepa-
cycle requires eggs from fish-eating titis. Patients with HCV infection
hosts, which emerge in faeces to have a significantly increased inci-
Risk prediction
contaminate the freshwater inhab- dence of diabetes, with a multivar- Because several risk factors in-
ited by snails and fish. iate-adjusted hazard ratio of 1.5 in teract to cause liver cancer, it is
a long-term prospective study [20]. important to integrate them into a
Obesity and diabetes risk prediction model to derive one
Both obesity (see Chapter 2.7) and Fine particulate matter measure of absolute risk, for the ap-
diabetes are associated with the Exposure to fine particulate mat- propriate identification of people at
development of HCC. Obesity may ter (particulate matter with particles high risk who require clinical inter-
influence HCC through non-alco- of aerodynamic diameter less than vention. Risk prediction is very im-
holic fatty liver disease and non-al- 2.5 µm [PM2.5 ]) is associated with portant for the personalized health
coholic steatohepatitis, which pro systematic inflammation markers care of those who are susceptible
gress through fibrosis and cirrhosis and serum levels of liver enzymes, to liver cancer.
to liver cancer [18]. including alanine aminotransferase Easy-to-use nomograms have
Higher plasma levels of adi- (ALT), aspartate aminotransferase, been developed for predicting long-
ponectin are associated with a low- and gamma-glutamyl transferase. In term risk of HCC in patients with
Fig. 5.6.5. Nomograms from the Risk Evaluation of Viral Load Elevation and Associated Liver Disease/Cancer of Hepatitis B Virus
(REVEAL-HBV) study are some of the earliest risk calculators for predicting risk of cirrhosis or hepatocellular carcinoma (HCC) in
patients with chronic hepatitis B. Integer risk scores are assigned to various groups of eight predictors: sex, age, family history of
HCC, alcohol intake, serum alanine aminotransferase (ALT) level, HBV e antigen (HBeAg) serostatus, serum HBV DNA level, and
HBV genotype. Both 5-year and 10-year risks of HCC by summed risk score are depicted in the nomogram. It is easy to identify the
long-term HCC risk by summing the risk scores. These nomograms have high internal validity and discriminatory ability to triage
patients with chronic hepatitis B into different risk groups.
100
5-year risk
10-year risk
10 Risk Factor Adjustment Scores

HBV DNA level HBV
HBeAg
(copies/mL) genotype
Negative < 300 0
Negative 300–9999 1
Risk of HCC (%)
Negative 10 000–99 999 B or B+C 3

Negative 10 000–99 999 C 4
Negative 100 000–999 999 B or B+C 3
1 Negative 100 000–999 999 C 7
Negative ≥ 10 6 B or B+C 4
Negative ≥ 10 6 C 7
Positive B or B+C 6
Positive C 6
Risk Factor Adjustment Scores

Family
Age Alcohol Serum ALT
Gender history of
0.1 (years) intake level (U/L)
HCC
Female 0 30–34 0 No 0 No 0 < 15 0
Male 2 35–39 1 Yes 2 Yes 2 15–44 1
40–44 2 ≥ 45 1
45–49 3
50–54 4
55–59 5
60–65 6
0.01
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Risk Score
360
Hepatocellular carcinoma risk calculators for patients with chronic viral hepatitis
In the era of precision medicine, biomarkers during long-term fol- and 1313 anti-HCV-seropositive
it is important to classify patients low-up are essential. For exam- participants were enrolled, and
with viral hepatitis into subgroups ple, the Risk Evaluation of Viral among them 384 new cases of
that differ in their susceptibility to Load Elevation and Associated HCC occurred until 30 June 2008.
liver cancer, their prognosis, and Liver Disease/Cancer (REVEAL) A series of HCC risk calculators
their response to clinical manage- study recruited a cohort of 23 820 were developed and validated for
ment. Preventive or therapeutic adult male and female residents patients with chronic hepatitis B
interventions can then be concen- of seven townships in 1991–1992. and those with chronic hepatitis C,
trated on those who will benefit, The health examination at study from the REVEAL study [2,3].
thus sparing expense and side-ef- entry and follow-up visit included The Risk Estimation for Hepa
fects for those who will not. In the abdominal ultrasonography and tocellular Carcinoma in Chronic
past decade, risk calculators for serological tests of (i) hepatitis B Hepatitis B (REACH-B) scoring
predicting long-term risk of hepa- biomarkers, including HBV surface systems were derived from the
tocellular carcinoma (HCC) in pa- antigen (HBsAg), HBV e antigen community cohort of the REVEAL-
tients with chronic hepatitis B and (HBeAg), genotype, mutant types, HBV study and validated inter-
C have been derived and validated and DNA (HBV DNA); (ii) hepati- nationally in hospital cohorts.
internationally [1]. tis C biomarkers, including HCV Important risk predictors includ-
For the derivation and vali- antibody (anti-HCV), genotype, ing age, sex, HBeAg serostatus,
dation of the risk of liver cancer, and RNA (HCV RNA); and (iii) liver and serum levels of ALT, HBV
well-designed prospective co- function biomarkers, including ala- DNA, and HBsAg were incorpo-
hort studies on a large cohort of nine aminotransferase (ALT) and rated into the REACH-B scores.
patients with viral hepatitis with aspartate aminotransferase (AST). These scores have high validity for
comprehensive collection of serial A total of 4155 HBsAg-seropositive HCC risk prediction. Table B5.6.1
Table B5.6.1. Projected risk of developing hepatocellular carcinoma in patients with chronic hepatitis B, from the Risk
Estimation for Hepatocellular Carcinoma in Chronic Hepatitis B (REACH-B) IIa prediction model
Predictor Risk score Cumulative Projected risk of developing

(summed) hepatocellular carcinoma (%)
risk score
3-year 5-year 10-year
Sex 0 0.002 0.007 0.02
Female 0 1 0.003 0.01 0.03
Male 2 2 0.006 0.02 0.06
Age, 5-year increment 1 3 0.01 0.03 0.09
Serum ALT level (U/L) 4 0.02 0.05 0.15
< 15 0 5 0.03 0.08 0.25
15–44 1 6 0.05 0.13 0.42
≥ 45 2 7 0.08 0.22 0.69
CHAPTER 5.6
SECTION 5
HBeAg/HBV DNA (copies/mL)/HBsAg (IU/mL) 8 0.13 0.37 1.13
Negative/< 10 4/< 100 0 9 0.21 0.61 1.87
Negative/< 10 4/< 100–999 2 10 0.35 1.01 3.08
Negative/< 10 /≥ 1000
4
3 11 0.59 1.66 5.04
Negative/10 4 –10 6/< 100 2 12 0.97 2.74 8.21
Negative/10 4 –10 6/100–999 3 13 1.60 4.49 13.21
Negative/10 4 –10 6/≥ 1000 4 14 2.63 7.32 20.91
Negative/≥ 10 6/any 6 15 4.32 11.82 32.18
Positive 7 16 7.04 18.80 47.42
17 11.39 29.15 65.49
ALT, alanine aminotransferase; HBeAg, HBV e antigen; HBsAg, HBV surface antigen; HBV, hepatitis B virus.

shows the risk scores assigned serum HCV RNA level, and HCV 2. Yang HI, Tseng TC, Liu J, Lee MH, Liu
CJ, Su TH, et al. (2016). Incorporating
to different groups of risk predic- genotype, were incorporated into serum level of hepatitis B surface anti-
tors, together with the projected the risk score. The risk score has gen or omitting level of hepatitis B virus
risk of developing HCC for poten- satisfactory to high validity and DNA does not affect calculation of risk
for hepatocellular carcinoma in patients
tial cumulative risk scores in the discriminatory ability for HCC risk without cirrhosis. Clin Gastroenterol
REACH-B IIa prediction model [2]. prediction. However, it needs to be Hepatol. 14(3):461–468.e2. https://
An HCC risk score for anti- validated internationally for its ap- d o i .o r g /10 .1016 / j .c g h . 2 015 .10 . 0 3 3
PMID:26598229
HCV-seropositive patients was plication in other countries.
derived from the REVEAL study 3. Lee MH, Lu SN, Yuan Y, Yang HI, Jen CL,
References You SL, et al. (2014). Development and
and validated in another commu- validation of a clinical scoring system for
nity-based high-risk cohort [3]. 1. Chen CJ, Lee M-H, Liu J, Yang H-I predicting risk of HCC in asymptomatic
(2015). Hepatocellular carcinoma risk individuals seropositive for anti-HCV an-
Important risk predictors, includ-
scores: ready to use in 2015? Hepat tibodies. PLoS One. 9(5):e94760. https://
ing age, serum ALT level, serum Oncol. 2(1):1–4. https://doi.org/10.2217/ doi.org/10.1371/journal.pone.0094760
AST/ALT ratio, cirrhosis status, hep.14.32 PMID:30190979 PMID:24801353
chronic viral hepatitis (Fig. 5.6.5). therapy. Lamivudine was first ap- variants were found to be associated
These risk calculators are helpful proved in 1998 for the treatment of with the efficacy of interferon-based
for the triage of patients with viral chronic hepatitis B. It significantly therapy for chronic hepatitis C, and
hepatitis who need intensive liver decreases the risk of HCC in treat- ethnicities in the Asia-Pacific region
surveillance and/or antiviral ther- ed patients but has the disadvan- were shown to have a high frequency
apy, and for the evaluation of the tage of inducing antiviral-resistant of favourable genotypes.
efficacy of clinical management of YMDD mutants. Newly developed Direct-acting antiviral agents
patients with chronic viral hepatitis antiviral drugs for chronic hepatitis are highly effective for all HCV ge-
in South and East Asia. Risk calcu- B have higher genetic barriers, to notypes, without any ethnic vari-
lators for predicting cirrhosis, the limit the development of antiviral- ation. They are convenient oral
most important predisposing factor resistant strains. A recent cohort agents with a low side-effect profile.
for HCC, in patients with chronic study reported a significant de- In a recent study of 62 354 patients
hepatitis B have also been derived crease in the incidence of HCC in with chronic hepatitis C treated with
and validated internally [22]. 973 patients with chronic hepatitis interferon and/or direct-acting anti-
B treated with pegylated interferon viral agents, sustained virological
or any anti-HBV nucleoside/nucleo- response (versus non-sustained vi-
Prevention tide analogue. The study found a rological response) was associated
Liver cancer may be prevented 77% reduction in HCC incidence with a significant reduction in risk of
through interventions related in treated patients, compared with HCC in patients treated with direct-
to its major etiological factors 4935 untreated patients, after ad- acting antiviral agents only (71%
(Table 5.6.2). Viral hepatitis con- justment for the Risk Estimation reduction), with both direct-acting
trol is included within the United for Hepatocellular Carcinoma in antiviral agents and interferon (52%
Nations Sustainable Development Chronic Hepatitis B (REACH-B) reduction), and with interferon only
Goals. The HBV vaccine has high predictive risk score [24]. In a (68% reduction), after adjustment for
efficacy and cost–effectiveness to European study of 1951 adult multiple risk factors [26]. In a study
prevent HCC. It is the first vaccine Caucasian patients with chronic of 4639 patients with chronic hepati-
to prevent a cancer type in humans hepatitis B treated with entecavir or tis C treated with pegylated interfer-
(see Chapter 6.3). The HBV vaccine tenofovir, there was a significant de- on and ribavirin, sustained virologi-
has been incorporated into the na- cline in the annual HCC incidence cal response (versus non-sustained
tional immunization programmes of rate in patients with cirrhosis, from virological response) was associat-
187 countries. The worldwide per- 3.22% within the first 5 years of ed with a significant decline in HCC
centage of children younger than therapy to 1.57% within 5–10 years incidence in patients with cirrhosis
5 years living with chronic HBV after enrolment [25]. (46% reduction) and in those without
infection fell from 4.7% in the pre- The standard treatment for chron- cirrhosis (63% reduction) [27].
vaccine era to 1.3% in 2015. HBV ic hepatitis C was interferon-based The global targets for 2030 set
vaccination prevents an estimated therapy until the advent of direct-act- by WHO include 90% HBV vaccina-
4.5 million HBV infections per year ing antiviral agents in 2013. HCV ge- tion coverage, 90% prevention of
in children [23]. notypes 1 and 4 are less responsive mother-to-child HBV transmission,
Several antiviral drugs have to interferon-based therapy com- 100% blood transfusion safety and
been approved for viral hepatitis pared with other genotypes. IFNL3 injection safety, diagnosis of 90% of
362
Table 5.6.2. Prevention of liver cancer through interventions related to its major etiological factors
Etiological factor Cancer type Intervention
Hepatitis B virus infection Hepatocellular carcinoma Immunization with HBIg and vaccine
Intrahepatic cholangiocarcinoma Interruption of mother-to-child transmission
Early diagnosis of HBV infection
Treatment of eligible patients with HBV infection
Hepatitis C virus infection Hepatocellular carcinoma Injection safety using engineered devices
Intrahepatic cholangiocarcinoma Blood safety by donation screening
Harm reduction for people who inject drugs
Early diagnosis of HCV infection
Treatment of eligible patients with HCV infection
Alcohol consumption Hepatocellular carcinoma Increase in alcohol taxes

Limitation on days and/or hours of sale
Enforcement of laws against privatizing retail sale of alcohol
Regulation of density of alcohol outlets
Enhancement of prohibiting sales to minors
Behavioural intervention
Aflatoxin exposure Hepatocellular carcinoma Pre-harvest good agricultural practices to reduce crop stress
Post-harvest sorting, storing, and drying
Improvement in grain storage
Introduction of fungus-resistant strains
Avoidance or reduction of consumption of contaminated foods
Biocontrol to reduce aflatoxin-producing fungi
Liver fluke infection Intrahepatic cholangiocarcinoma Stopping the consumption of raw fish
Cooking fish before eating
Screening and treatment with single-dose praziquantel
Practising hygienic defecation
Obesity Hepatocellular carcinoma Diet control

Exercise
HBIg, hepatitis B immunoglobulin; HBV, hepatitis B virus; HCV, hepatitis C virus.
HBV and HCV infections, and treat- Sustainable financing and innova- aflatoxin exposure, liver fluke infec-
ment of 80% of eligible patients [23]. tion are also required for the devel- tion, and obesity (Table 5.6.2). [1].
To reach these targets, concerted opment and delivery of vaccines, Basic improvements in sorting, dry-
national and international efforts diagnostics, and treatments to trans- ing, and storing the groundnut crop
are urgently needed. The coverage form the global hepatitis response. in West Africa resulted in a marked
of diagnosis and treatment should Several effective interventions reduction in aflatoxin contamination,
be rapidly scaled up through a pub- are recommended to reduce the in a feasible and cost-effective ap-
lic health approach to benefit all. prevalence of alcohol consumption, proach [28]. Reductions in aflatoxin
biomarkers over time in China, linked
to changes in consumption of afla-
Table 5.6.3. Methods for early detection, diagnosis, and treatment of liver cancer toxin-contaminated foods, were also
associated with reduced incidence
Clinical strategy Methods
of HCC [29]. Concerted efforts to
control liver fluke infection have been
CHAPTER 5.6
SECTION 5
Early detection and diagnosis α-Fetoprotein (low sensitivity for small tumours)
Serum M2BPGi level implemented in Thailand and have
Ultrasonography (< 1 cm)
High-resolution computed tomography (CT) scan
resulted in a large reduction in the
Contrast magnetic resonance imaging (MRI) scan prevalence of infection [1].
Angiogram
Laparoscopy
Biopsy (not required for diagnosis) Detection
Treatment Surgical resection (partial hepatectomy) Methods for screening, diagnosis,
Liver transplantation and treatment of liver cancer are
Radiofrequency ablation shown in Table 5.6.3. Both serum
Radiotherapy
Chemoembolization
α-fetoprotein (AFP) level and ab-
Radioembolization dominal ultrasonography are used
Targeted therapy for the screening of HCC in high-
Immunotherapy
risk patients: those with chronic viral
M2BPGi, Mac-2-binding protein glycosylation isomer. hepatitis and those with cirrhosis.

Efficacy of viral hepatitis control to reduce risk of liver cancer
Because chronic hepatitis B virus (HBeAg)-seropositive status or in people aged 5–29 years have
(HBV) infection and chronic hep- with a high HBsAg titre received decreased significantly from birth
atitis C virus (HCV) infection are hepatitis B immunoglobulin with cohorts born in 1977–1980 to those
major etiological factors for liver the first dose of vaccine at birth. born in 1997–2000. The age- and
cancer, their effective control may From July 1987, previously unim- sex-adjusted rate ratio for HCC
significantly reduce the burden of munized preschool children were incidence was 0.37 and for HCC
disease globally. Hepatitis B may also vaccinated, which means that mortality was 0.21, for the 1997–
be prevented by immunization, and birth cohorts born in 1981–1984 2000 birth cohorts compared with
both hepatitis B and hepatitis C were vaccinated after age 1 year. the 1977–1980 birth cohorts. From
may be treated with antiviral drugs. The immunization rate of eligi- a study of 3.8 million vaccinees,
Successful reduction of liver can- ble infants was more than 90%. incomplete immunization and ma-
cer incidence and mortality has The rate of HBsAg-seropositive ternal serostatus of HBsAg and
been demonstrated through sev- status at age 6 years decreased HBeAg are important predictors of
eral national programmes of viral significantly, from more than 10% HCC risk for the vaccinees [2].
hepatitis control. in unimmunized birth cohorts to For patients with chronic viral
The first national immuniza- less than 1% in immunized birth hepatitis, prompt treatment is the
tion programme in the world was cohorts. only strategy to prevent liver can-
launched in July 1984 [1]. From This immunization programme cer. The first national programme
July 1984 to June 1986, only ba- has been well documented to to treat patients with chronic viral
bies born to HBV surface anti- prevent hepatocellular carci- hepatitis in the world was launched
gen (HBsAg)-positive mothers noma (HCC) in immunized birth in October 2003 [3]. Available
were immunized; after July 1986, cohorts, showing a very high ef- treatments for chronic HBV infec-
all newborns were immunized. ficacy 30 years after the launch tion include interferon-α, pegylated
Although all newborns received of the immunization programme interferon-α, lamivudine, adefovir,
vaccines, only babies born to high- (Table B5.6.2) [1]. Both the inci- entecavir, telbivudine, and tenofo-
risk mothers with HBV e antigen dence of and mortality from HCC vir. Available treatments for chronic
Table B5.6.2. Significant reductions in hepatocellular carcinoma (HCC) incidence and mortality through national programmes
of hepatitis B virus immunization and chronic viral hepatitis therapy
Hepatitis B virus immunization programme

Birth year HCC mortality (ages 5–29 years) HCC incidence (ages 5–29 years)
Rate per 100 000 Age- and sex-adjusted Rate per 100 000 Age- and sex-adjusted
person-years rate ratio (95% CI) person-years rate ratio (95% CI)
1977–1980 0.81 1.00 (reference) 1.14 1.00 (reference)
1981–1984 0.56 0.70 (0.59–0.83) 0.77 0.73 (0.63–0.85)
1985–1988 0.30 0.43 (0.33–0.55) 0.37 0.48 (0.38–0.60)
1989–1992 0.17 0.27 (0.19–0.39) 0.23 0.37 (0.27–0.51)
1993–1996 0.12 0.21 (0.13–0.34) 0.22 0.43 (0.30–0.62)
1997–2000 0.12 0.21 (0.12–0.38) 0.17 0.37 (0.21–0.62)
Chronic viral hepatitis therapy programme
Calendar year HCC mortality (ages 30–69 years) HCC incidence (ages 30–69 years)
Rate per 100 000 Age- and sex-adjusted Rate per 100 000 Age- and sex-adjusted
person-years rate ratio (95%CI) person-years rate ratio (95% CI)
2000–2003 36.59 1.00 (reference) 54.12 1.00 (reference)
2004–2007 35.77 0.95 (0.93–0.97) 54.79 0.98 (0.96–0.99)
2008–2011 30.21 0.76 (0.75–0.78) 50.77 0.86 (0.85–0.88)
2012–2015 27.44 0.64 (0.62–0.65) 47.55 0.76 (0.74–0.77)
CI, confidence interval; HCC, hepatocellular carcinoma.
364
HCV infection include ribavirin, launch of the chronic viral hepatitis 2. Chien YC, Jan CF, Chiang CJ, Kuo HS,
You SL, Chen CJ (2014). Incomplete hep-
pegylated interferon, and direct- therapy programme. Further diag-
atitis B immunization, maternal carrier
acting antiviral agents. nosis and treatment of more eligi- status, and increased risk of liver diseas-
From 2000–2003 to 2012– ble patients with viral hepatitis are es: a 20-year cohort study of 3.8 million
still needed. vaccinees. Hepatology. 60(1):125–32.
2015, there was a significant reduc-
h t t p s: //d o i .o r g /10 .10 0 2 / h e p. 2 70 4 8
tion in the incidence of and mortali- PMID:24497203
ty from liver cancer (Table B5.6.2). References
3. Chiang CJ, Yang YW, Chen JD, You
The age- and sex-adjusted rate ra- 1. Chiang CJ, Yang YW, You SL, Lai SL, Yang HI, Lee MH, et al. (2015).
tio for HCC incidence was 0.76 and MS, Chen CJ (2013). Thirty-year out- Significant reduction in end-stage liver
comes of the national hepatitis B im- diseases burden through the national vi-
for HCC mortality was 0.64, for
munization program in Taiwan. JAMA. ral hepatitis therapy program in Taiwan.
2012–2015 compared with 2000– 310(9):974–6. https://doi.org/10.1001/ Hepatology. 61(4):1154–62. https://doi.
2003, the 4-year period before the jama.2013.276701 PMID:24002285 org/10.1002/hep.27630 PMID:25476749
Ultrasonography may detect HCC sylation isomer (M2BPGi) is able to ods of choice depend on the tumour
tumours smaller than 1 cm. AFP accurately distinguish between stag- size, lymph node involvement, me-
level has a screening sensitivity of es of fibrosis in patients with chronic tastasis, liver function and cirrhosis
about 70% for detecting early-stage, viral hepatitis. It has been reported status, overall health condition, and
small HCC tumours. However, AFP to be a seromarker that is as good patient preference. However, detec-
level remains a useful seromarker as AFP level for short-term prediction and treatment options are very
for short-term prediction of HCC af- tion of HCC in patients with chronic limited in low- and middle-income
ter antiviral treatment in patients with hepatitis B [30]. countries, where liver cancer is a
chronic hepatitis C [27]. The serum There are several options for the major health problem.
level of Mac-2-binding protein glyco- treatment of liver cancer. The meth-
References
1. Gelband H, Chen CJ, Chen W, Franceschi 5. Castelli G, Pelosi E, Testa U (2017). 9. The Cancer Genome Atlas Research
S, Hall A, London WT, et al. (2015). Liver cancer: molecular characteriza- Network (2017). Comprehensive and inte-
Liver cancer. In: Gelband H, Jha P, tion, clonal evolution and cancer stem grative genomic characterization of hepa-
Sankaranarayanan R, Horton S, editors. cells. Cancers (Basel). 9(9):127–74. tocellular carcinoma. Cell. 169(7):1327–
Disease control priorities. 3rd ed. Vol. 3, https://doi.org/10.3390/cancers9090127 1341.e23. https://doi.org/10.1016/j.cell.
Cancer. Washington (DC), USA: World PMID:28930164 2017.05.046 PMID:28622513
Bank; pp. 147–164.
6. Alexandrov LB, Nik-Zainal S, Wedge 10. Liu J, Yang HI, Lee MH, Lu SN, Jen CL,
2. Bray F, Ferlay J, Soerjomataram I, Siegel DC, Aparicio SAJR, Behjati S, Biankin Batrla-Utermann R, et al.; R.E.V.E.A.L.-
RL, Torre LA, Jemal A (2018). Global AV, et al.; Australian Pancreatic Cancer HBV Study Group (2014). Spontaneous
cancer statistics 2018: GLOBOCAN es- Genome Initiative; ICGC Breast seroclearance of hepatitis B seromark-
timates of incidence and mortality world- Cancer Consortium; ICGC MMML-Seq ers and subsequent risk of hepatocellular
wide for 36 cancers in 185 countries. CA Consortium; ICGC PedBrain (2013). carcinoma. Gut. 63(10):1648–57. https://
Cancer J Clin. 68(6):394–424. https://doi. d o i .o r g /10 .113 6 /g u t j n l - 2 013 - 3 0 5 78 5
Signatures of mutational processes in
PMID:24225939
CHAPTER 5.6
SECTION 5
org/10.3322/caac.21492 PMID:30207593 human cancer. Nature. 500(7463):415–
21. https://doi.org/10.1038/nature12477 11. Hu HH, Liu J, Lin YL, Luo WS, Chu YJ,
3. Akinyemiju T, Abera S, Ahmed M, PMID:23945592 Chang CL, et al.; REVEAL-HBV Study
Alam N, Alemayohu MA, Allen C, et al.;
Group (2016). The rs2296651 (S267F)
Global Burden of Disease Liver Cancer 7. Totoki Y, Tatsuno K, Covington KR, Ueda
variant on NTCP (SLC10A1) is inversely
Collaboration (2017). The burden of pri- H, Creighton CJ, Kato M, et al. (2014). associated with chronic hepatitis B and
mary liver cancer and underlying etiolo- Trans-ancestry mutational landscape progression to cirrhosis and hepatocel-
gies from 1990 to 2015 at global, region- of hepatocellular carcinoma genomes. lular carcinoma in patients with chronic
al, and national level: results from the Nat Genet. 46(12):1267–73. https://doi. hepatitis B. Gut. 65(9):1514–21. https://
Global Burden of Disease Study 2015. org/10.1038/ng.3126 PMID:25362482 d o i .o r g /10 .113 6 /g u t j n l - 2 015 - 310 6 8 6
JAMA Oncol. 3(12):1683–91. https:// PMID:26642861
doi.org/10.10 01/jamaoncol.2017.3055 8. Shibata T, Aburatani H (2014). Exploration
PMID:28983565 of liver cancer genomes. Nat Rev 12. Lee MH, Yang HI, Lu SN, Lin YJ, Jen CL,
Gastroenterol Hepatol. 11(6):340–9. Wong KH, et al. (2015). Polymorphisms
4. Chen CJ (2018). Global elimination of viral https://doi.org/10.1038/nrgastro.2014.6 near the IFNL3 gene associated with
hepatitis and hepatocellular carcinoma: PMID:24473361 HCV RNA spontaneous clearance and
opportunities and challenges. Gut. 67(4): hepatocellular carcinoma risk. Sci Rep.
595–8. https://doi.org/10.1136/gutjnl-2017- 5(1):17030. https://doi.org/10.1038/srep
315407 PMID:29367206 17030 PMID:26602024

13. Lee MH, Huang YH, Chen HY, Khor SS, 19. Chen CL, Yang W-S, Yang H-I, Chen 25. Papatheodoridis GV, Idilman R, Dalekos
Chang YH, Lin YJ, et al.; REVEAL-HCV C-F, You S-L, Wang L-Y, et al. (2014). GN, Buti M, Chi H, van Boemmel F, et
Cohort Study Group (2017). Human leu- Plasma adipokines and risk of hepato- al. (2017). The risk of hepatocellular car-
kocyte antigen variants and risk of hepa- cellular carcinoma in chronic hepatitis cinoma decreases after the first 5 years
tocellular carcinoma modified by hepatitis B virus-infected carriers: a prospec- of entecavir or tenofovir in Caucasians
C virus genotypes: a genome-wide as- tive study in Taiwan. Cancer Epidemiol with chronic hepatitis B. Hepatology.
66(5):1444–53. https://doi.org/10.1002/
sociation study. Hepatology. 67(2):651– Biomarkers Prev. 23(8):1659–71. https://
hep.29320 PMID:28622419
61. https://doi.org/10.1002/hep.29531 doi.org/10.1158/1055-9965.EPI-14-0161
PMID:28921602 PMID:24895413 26. Ioannou GN, Green PK, Berry K (2018).
HCV eradication induced by direct-acting
14. Loomba R, Yang HI, Su J, Brenner D, 20. Lin YJ, Shaw TG, Yang HI, Lu SN, Jen CL, antiviral agents reduces the risk of hepato-
Barrett-Connor E, Iloeje U, et al. (2013). Wang LY, et al.; R.E.V.E.A.L.-HCV Study cellular carcinoma. J Hepatol. 68(1):25–32.
Synergism between obesity and alcohol in Group (2017). Chronic hepatitis C virus https://doi.org/10.1016/j.jhep.2017.08.030
increasing the risk of hepatocellular carci- infection and the risk for diabetes: a com- PMID:28887168
noma: a prospective cohort study. Am J munity-based prospective study. Liver
Epidemiol. 177(4):333–42. https://doi. Int. 37(2):179–86. https://doi.org/10.1111/ 27. Lee MH, Huang CF, Lai HC, Lin CY, Dai
org/10.1093/aje/kws252 PMID:23355498 liv.13194 PMID:27363856 CY, Liu CJ, et al. (2017). Clinical effi-
cacy and post-treatment seromarkers
15. Liu J, Yang HI, Lee MH, Jen CL, Hu HH, 21. Pan WC, Wu CD, Chen MJ, Huang YT, associated with the risk of hepatocel-
Lu SN, et al. (2016). Alcohol drinking Chen CJ, Su HJ, et al. (2015). Fine par- lular carcinoma among chronic hepatitis
mediates the association between poly- ticle pollution, alanine transaminase, C patients. Sci Rep. 7(1):3718. https://
morphisms of ADH1B and ALDH2 and and liver cancer: a Taiwanese prospec- doi.org /10.103 8/s 4159 8 - 017- 02313 -y
PMID:28623331
hepatitis B-related hepatocellular carci- tive cohort study (REVEAL-HBV). J Natl
noma. Cancer Epidemiol Biomarkers Prev. Cancer Inst. 108(3):djv341. https://doi. 28. Turner PC, Sylla A, Gong YY, Diallo
25(4):693–9. https://doi.org/10.1158/1055- org/10.1093/jnci/djv341 PMID:26561636 MS, Sutcliffe AE, Hall AJ, et al. (2005).
9965.EPI-15-0961 PMID:26827895 Reduction in exposure to carcino-
22. Lee MH, Yang HI, Liu J, Batrla-Utermann R, genic aflatoxins by postharvest inter-
16. Chu YJ, Yang HI, Wu HC, Liu J, Wang Jen CL, Iloeje UH, et al.; R.E.V.E.A.L.-HBV vention measures in west Africa: a
LY, Lu SN, et al. (2017). Aflatoxin B1 ex- Study Group (2013). Prediction models of community-based intervention study.
posure increases the risk of cirrhosis long-term cirrhosis and hepatocellular car- Lancet. 365(9475):1950–6. https://doi.
and hepatocellular carcinoma in chronic cinoma risk in chronic hepatitis B patients: org/10.1016/S0140 - 673 6(05) 6 6 6 61- 5
hepatitis B virus carriers. Int J Cancer. risk scores integrating host and virus pro- PMID:15936422
141(4):711–20. https://doi.org/10.1002/ijc. files. Hepatology. 58(2):546–54. https://doi.
29. Chen JG, Egner PA, Ng D, Jacobson
30782 PMID:28509392 org/10.1002/hep.26385 PMID:23504622
LP, Muñoz A, Zhu YR, et al. (2013).
17. Chu YJ, Yang HI, Wu HC, Lee MH, Liu J, 23. WHO (2017). Global hepatitis report, Reduced aflatoxin exposure presages
Wang LY, et al. (2018). Aflatoxin B1 exposure 2017. Geneva, Switzerland: World Health decline in liver cancer mortality in an
increases the risk of hepatocellular carcino- Organization. Available from: https:// endemic region of China. Cancer Prev
ma associated with hepatitis C virus infec- Res (Phila). 6(10):1038–45. https://doi.
w w w.w h o. i nt / h e p a t i t i s /p u b l i c a t i o n s /
org/10.1158/1940 - 6207.CAPR-13- 0168
tion or alcohol consumption. Eur J Cancer. global-hepatitis-report2017/en/.
PMID:23963804
94(5):37–46. https://doi.org/10.1016/j.ejca.
2018.02.010 PMID:29533866 24. Lin D, Yang HI, Nguyen N, Hoang J, Kim
30. Liu J, Hu HH, Lee MH, Jen CL, Batrla-
Y, Vu V, et al. (2016). Reduction of chronic Utermann R, Lu SN, et al. (2017). Serum
18. Michelotti GA, Machado MV, Diehl AM hepatitis B-related hepatocellular carci- levels of M2BPGi as short-term predic-
(2013). NAFLD, NASH and liver can- noma with anti-viral therapy, including low tors of hepatocellular carcinoma in un-
cer. Nat Rev Gastroenterol Hepatol. risk patients. Aliment Pharmacol Ther. treated chronic hepatitis B patients. Sci
10(11):656–65. https://doi.org/10.1038/ 44(8):846–55. https://doi.org/10.1111/apt. Rep. 7(1):14352. https://doi.org/10.1038/
nrgastro.2013.183 PMID:24080776 13774 PMID:27549411 s41598-017-14747-5 PMID:29085039
366
5.7
Pancreatic cancer
Many risk factors too poorly
characterized to enable prevention
Jessica N. Everett Eric J. Duell (reviewer)
Diane M. Simeone Donghui Li (reviewer)
Núria Malats (reviewer)
Pancreatic cancer is the seventh tion options has been challenging.

SUMMARY most common cause of cancer- In the 15% of patients who present
related mortality worldwide, with an with resectable tumours, the 5-year
●● Pancreatic cancer is the sev- overall 5-year survival rate of 9%. survival rate of 30% remains much
enth most common cause of The most common type of pancre- lower than that for many other can-
cancer-related mortality world- atic cancer (> 90%) is infiltrating cer types; this highlights the unique
wide, with an overall 5-year sur- pancreatic ductal adenocarcinoma. propensity for pancreatic cancer to
vival rate of 9%. The most com- The epidemiological study of metastasize early in the course of
mon type of pancreatic cancer pancreatic ductal adenocarcinoma the disease. Biomarkers for early
(> 90%) is infiltrating pancreatic is complicated by significant geo- detection are lacking for clinical
ductal adenocarcinoma. graphical and temporal variations in use, and established modifiable
●● Smoking, obesity, and long- the sensitivity and specificity of clin- risk factors remain inadequately
standing type 2 diabetes are ical diagnosis and in the proportion characterized to enable an impact-
known risk factors for pancre- of cases that are histologically veri- ful plan for primary prevention of
atic cancer development. New- fied. Differences in access to health pancreatic cancer.
onset diabetes can be an early care, such as differences related to
sign of pancreatic cancer. social classes or age groups, can
affect the reported incidence and
Epidemiology
●● More than 90% of cases of pan- mortality rates. Pancreatic cancer is among the
creatic cancer are sporadic (i.e. In 2018 an estimated 459 000 deadliest types of cancer. In 2018,
due to spontaneous rather than new cases of pancreatic ductal there were an estimated 459 000
inherited mutations), although adenocarcinoma were diagnosed new cases of pancreatic cancer
a family history increases risk, worldwide. Incidence rates of pan-
worldwide, with age-standardized
particularly where more than creatic cancer in 2018 were high-
incidence rates in both sexes of
one first-degree family member est in western Europe (8.3 per
6.2 per 100 000 in more-developed
is involved. The presence of 100 000) and North America (7.6
countries and 1.5 per 100 000 in
pathogenic germline mutations per 100 000). The lowest incidence
less-developed countries. In the
in patients with sporadic pan-
SECTION 5
CHAPTER 5.7
USA, there were projected to be rates of pancreatic cancer (~1.0 per
creatic cancer, even in the ab- 100 000) were observed in East
55  440 new cases and 44  310
sence of a positive family his- Africa and South-Central Asia.
deaths from pancreatic cancer in
tory, is increasingly recognized. Global differences in pancreatic
2018. The USA has one of the high-
●● Activating mutations in the est pancreatic cancer incidence cancer incidence rates have been
KRAS oncogene and loss-of- rates in the world, and it is still ris- attributed largely to exposure to
function mutations in the tu- ing. Pancreatic cancer is projected known or suspected risk factors re-
mour suppressor genes TP53, to become the second most com- lated to lifestyle or the environment,
SMAD4, and CDKN2A are prev- mon cause of cancer death in the although heritable factors may con-
alent in pancreatic adenocarci- USA by 2030 [1]. tribute. The contributions of inter-
noma. None of these genetic Despite advances in the under- national differences in diagnostic
alterations can be targeted with standing of the biology of pancre- capacity or registry quality to ob-
current chemotherapeutics. atic cancer, clinical translation into served pancreatic cancer incidence
effective treatment and early detec- rates are not known.
Chapter 5.7 • Pancreatic cancer 367

Etiology sumption (three or more drinks per

day) has been linked to risk of pan- FUNDAMENTALS
Several non-modifiable factors are
creatic cancer (see Chapter 2.3).
associated with risk of pancreatic
This association may be related to ■■ Pancreatic ductal adeno-
cancer. Increasing age correlates
an increased incidence in this popu- carcinoma is an aggressive
with risk of pancreatic cancer; most
lation of chronic pancreatitis, which disease with innate resistance
patients are diagnosed at ages 60–
is known to increase the risk of pan- to standard chemotherapy and
80 years, and pancreatic cancer
creatic cancer 2-fold. There is no radiotherapy regimens.
is unusual in people younger than
link with moderate alcohol consump-
45 years. Pancreatic cancer affects ■■ Most patients with pancreatic
tion. A low level of physical activity
men and women equally. Studies in cancer present with advanced
has also been associated with risk
the USA have shown that pancre- disease. No reliable screening
of pancreatic cancer [7].
atic cancer is more common in the test is currently available
Large case–control and cohort
African American population than it for the early detection of
studies have identified obesity and pancreatic cancer.
is in the White population, but the
long-standing type 2 diabetes as
potential confounding contribution ■■ In the minority of patients
risk factors for pancreatic cancer
of socioeconomic factors, smoking [2]. There is a complex relationship who present with early-stage,
status, and the presence of type between obesity and type 2 dia- localized disease, the 5-year
2 diabetes and obesity has not betes, because they often coexist. survival rate is 30%, even
been calculated (see Chapter 4.6). Several large studies have consis- with surgical resection; this
Higher attained adult height and tently shown that obesity is a dose- highlights that pancreatic
non-O blood group are also associ- dependent risk factor for pancreatic cancer metastasizes early in
ated with increased risk. cancer, independent of the pres- the course of the disease.
Among the known modifiable ence of type 2 diabetes. For exam- ■■ Most pancreatic cancers
risk factors, smoking is the best ple, in a pooled cohort of more than harbour oncogenic KRAS
documented and is thought to be 900 000 people in whom 2454 pan- mutations, which occur early
responsible for about 25% of cases creatic cancers were diagnosed, in the tumorigenic process.
of pancreatic cancer (see Chapter the incidence of pancreatic cancer Secondary events – either ge-
2.1). Smokers have a relative risk was increased by 19% in the group netic changes, such as acqui-
of 1.5–1.9 of developing pancre- with body mass index 30–35 kg/m2 sition of loss-of-function muta-
atic cancer [2], with a documented (compared with the group with nor- tions in TP53, SMAD4, and
dose–risk relationship and a posi- mal weight; body mass index 18.5– CDKN2A, or tissue damage or
tive benefit identified with smoking 25 kg/m2), independent of the pres- inflammation – are required,
cessation. Use of smokeless tobac- ence of type 2 diabetes [8]. along with KRAS mutations,
co products is also associated with Paradoxically, diabetes has been for formation of pancreatic
increased risk of pancreatic cancer. established as both a risk factor for intraepithelial neoplasia and
Certain dietary habits, including pancreatic cancer (long-standing tumour progression.
high intake of saturated fats, fructose, type 2 diabetes) and a manifesta- ■■ Pancreatic cancer is charac-
and red meat and low intake of fruits tion of early-stage pancreatic cancer terized by an intense desmo-
and vegetables, have been associat- (new-onset type 3c diabetes). Long- plastic stromal reaction, which
ed with higher risk of pancreatic can- standing type 2 diabetes increases contributes to the biology of
cer. Very few studies – notably the the risk of pancreatic cancer devel- the disease and challenges
European Prospective Investigation opment about 2-fold [9]. Diabetes medical treatment.
into Cancer and Nutrition (EPIC) can also be caused by the presence
study [3], the Nurses’ Health Study of pancreatic cancer (type 3c diabe-
[4], and the Health Professionals tes). New-onset diabetes can be an
Follow-Up Study [5] – have compre- early sign of pancreatic cancer, and it atic inflammation, acceleration of
hensively investigated the effects of is being explored as a biomarker for tumour progression, and resistance
individual nutrition components on early detection (as discussed below). to chemotherapy [11,12]. Targeting
risk of pancreatic cancer. Obesity and type 2 diabetes obesity by calorie restriction de-
Current evidence on diet, nutri- are increasingly recognized as creased inflammation and reduced
tion, and physical activity related systemic, low-grade inflammatory pancreatic cancer incidence and
to reduction of higher risk of pan- conditions with increased expres- progression [13]. Similarly, type
creatic cancer is available as part sion of pro-inflammatory cytokines, 2 diabetes and hyperinsulinaemia
of the Continuous Update Project adipokines, and reactive oxygen have been shown to lead to chronic
of the World Cancer Research species [10]. In mouse models, obe- inflammation and increased cancer
Fund/American Institute for Cancer sity has been demonstrated to be risk and progression in mouse mod-
Research [6]. Heavy alcohol con- associated with increased pancre- els, and inhibition of inflammatory
368
signalling pathways reduced tumour rare cases of hereditary pancreati- sortium [21], the Pancreatic Cancer
growth in an animal model [14]. tis, caused by mutations in the cat- Cohort Consortium, and the Pancre
Oral antidiabetic medications ionic trypsinogen (PRSS1) gene. In atic Cancer Case-Control Con sor
have significant potential to decrease people with hereditary pancreatitis, tium [22] – have identified loci asso-
risk of pancreatic cancer. In a meta- the lifetime risk of pancreatic can- ciated with risk of pancreatic cancer.
analysis, use of metformin was asso- cer is about 40%. Further studies will be needed to
ciated with reduced risk of pancreatic Family history and genetic risk understand the functional conse-
cancer in patients with type 2 dia- factors also play a role in risk of quences of the identified common
betes [15], and metformin has been pancreatic cancer. Up to 8–10% variants. Risk models could poten-
shown to inhibit pancreatic tumour of patients with pancreatic cancer tially be developed to estimate risk
growth in mouse models [16]. carry a pathogenic germline vari- using validated SNPs and the pres-
The inflammatory microenviron- ant in a known cancer risk gene ence of other modifiable and non-
ment is also thought to be a major (including ATM, BRCA1, BRCA2, modifiable risk factors to identify
mechanism by which chronic pan- CDKN2A, EPCAM, MLH1, MSH2, patients at higher risk [23].
creatitis leads to the development MSH6, PALB2, PMS2, STK11, and
of pancreatic cancer (see Chapter TP53) [18–20]; these confer a life-
3.5). Although the population attri- time risk of pancreatic cancer that
Pathology
butable fraction is less than 3% [2], ranges from 3% to 58%. An addi- Infiltrating pancreatic ductal ade-
chronic pancreatitis has been as- tional group of patients with two or nocarcinoma is characterized by
sociated with pancreatic cancer in more family members with pancre- glandular neoplastic epithelial cells
multiple independent epidemiologi- atic cancer have familial pancre- typically surrounded by an intense
cal studies. A recent systematic re- atic cancer without an identifiable desmoplastic stromal reaction
view of 17 587 cases of pancreatitis genetic risk factor; this is associ- (Fig. 5.7.1). Therefore, the bulk of a
confirmed a strong association be- ated with a lifetime risk of 3–32%, pancreatic cancer is composed of
tween chronic pancreatitis and risk depending on the number of close stromal cells and collagen, with in-
of pancreatic cancer [17]. In that relatives affected. Patients with flammatory cells and blood vessels.
study, the risk of pancreatic cancer symptomatic pancreatitis who carry Pancreatic cancers are known
was associated with the duration of a pathogenic germline variant in to contain a high interstitial pres-
pancreatitis, with the highest risk in PRSS1 or have a documented fam- sure, and blood vessels within the
pancreatitis cases diagnosed within ily history of chronic pancreatitis tumour are compressed, creating
1 year. It is possible that the very also have an elevated lifetime risk, a hypoxic environment with de-
strong association in this group of up to 44%. Data on risk of pan- creased perfusion, as evidenced
could be ascribed to pre-existing creatic cancer associated with in- by the presence of a hypodense
pancreatic cancer that presented as herited syndromes are summarized mass on cross-sectional imaging
pancreatitis; however, the high risk in Table 5.7.1. (Fig. 5.7.2). The desmoplastic stro-
of pancreatic cancer in the groups Common single-nucleotide poly- mal reaction has been proposed
with pancreatitis duration of 2, 5, morphisms (SNPs) in the popula- to limit effective delivery of thera-
and 10 years highlights the clear tion may account for an additional peutic agents within the tumour.
association. Further evidence of the portion of pancreatic cancer cases. Therapeutic strategies that target
link between pancreatitis and risk of Large-scale efforts – including the the stroma are being developed.
pancreatic cancer is evident in the Pancreatic Disease Research Con Perineural tumour invasion is also
Table 5.7.1. Inherited syndromes associated with risk of pancreatic cancer
SECTION 5
CHAPTER 5.7
Syndrome Genes mutated Published risk estimates
Peutz–Jeghers syndrome STK11 Cumulative risk: 32–36% by age 70 years
Familial atypical multiple mole melanoma CDKN2A Cumulative risk: 17% by age 75 years
(FAMMM) syndrome
Familial pancreatic cancer Unknown Overall: SIR = 9.0
Three affected first-degree relatives: SIR = 32
Hereditary pancreatitis PRSS1 Cumulative risk: 44% by age 70 years
Hereditary breast and ovarian cancer syndrome BRCA1 Relative risk: 2.6
BRCA2 Relative risk: 3.5–5.9
Lynch syndrome MLH1, MSH2, MSH6, PMS2 Cumulative risk: 3–4% by age 70 years
ATM, PALB2 Unknown
SIR, standardized incidence ratio.

Fig. 5.7.1. Histopathology of infiltrating pancreatic ductal adenocarcinoma, highlighting Fig. 5.7.2. A hypodense mass in the pan-
desmoplastic stroma. creas, indicated by a white arrow, shows
the characteristic imaging appearance of
a pancreatic ductal adenocarcinoma.
pancreatic cancer have been identi-

fied, although most have been stud-
ied in small retrospective cohorts
using samples collected from late-
stage disease, with relatively small
numbers of control samples from
common and causes pain in many (Fig. 5.7.3). Beyond these common patients with chronic pancreatitis,
patients with pancreatic cancer. mutations, deeper whole-genome diabetes, or non-cancerous biliary
Pancreatic adenocarcinoma can analyses have identified potential obstruction. Some recently identi-
develop from any of at least three subtypes of pancreatic cancer [25]. fied biomarkers that are being ac-
histologically distinct precursor le- In an analysis of 150 samples of tively studied include single mark-
sions. Pancreatic intraepithelial pancreatic ductal adenocarcinoma, ers [26], multi-analyte panels [27],
neoplasia lesions are microscopic including samples with the low cel- and immune-based proteomic pan-
proliferations that can progress to lularity that is characteristic of many els [28]. Specific phylotypes in oral
pancreatic cancer. However, they tumours, a subset of tumours har- flora have been associated with
are not detectable with current im- boured multiple KRAS mutations, risk of pancreatic cancer in a large
aging modalities. Intraductal pa- with some evidence of biallelic mu- prospective cohort study of the oral
pillary mucinous neoplasms are tations [24]. The contribution of this microbiome, suggesting that micro-
relatively common cystic lesions finding to tumour biology remains to biome signatures also hold promise
of the pancreatic ducts. They are be discerned. as biomarkers for early detection
often identified incidentally on ab- Next-generation sequencing for [29]. Prospective studies in a large-
dominal imaging, and they can have patients with pancreatic cancer iden- scale high-risk cohort are needed
dysplasia and malignant potential. tifies alterations in about 40% of se- to validate the clinical utility of bio-
Mucinous cystic neoplasms are quenced patients. This information markers for early detection, sepa-
recognized by the unique presence is currently used in clinical research rately and in combination.
of ovarian-type stroma. They oc- to inform enrolment in a genotype- Recently, detailed work has shed
cur more commonly in women and directed clinical trial. For example, light on the potential role of new-
have a higher associated risk, with germline or somatic alterations in onset diabetes as a biomarker for
a chance of about 30% of progress- DNA repair genes such as BRCA1, early pancreatic cancer. In a study in
ing to adenocarcinoma. BRCA2, PALB2, or ATM give rise Olmsted County, Minnesota, USA,
to genomic instability in a subset
which had near-complete clinical
Genetics of pancreatic ductal adenocarcino-
data capture of the entire popula-
mas; this could make them more
Extensive studies to characterize tion of the county, fasting blood glu-
sensitive to platinum-based che-
the genomic landscape of pancre- cose level was associated with time
motherapy and/or poly(ADP-ribose)
atic cancer have improved the un- to diagnosis of pancreatic cancer,
polymerase (PARP) inhibitors. It is
derstanding of intertumour hetero- and the data showed that patients
not currently recommended in clini-
geneity in patients with pancreatic diagnosed with pancreatic cancer
cal practice.
cancer. The most commonly mutat- were hyperglycaemic for a mean
ed genes in pancreatic adenocarci- of 30–36 months before diagnosis
noma include the KRAS oncogene Biomarkers [30] (Fig. 5.7.4). From this work, a
and the tumour suppressor genes Several putative biomarkers that risk prediction model was developed
TP53, SMAD4, and CDKN2A [24] may play a role in early detection of that incorporated change in weight,
370
Fig. 5.7.3. Overview of the molecular genomic features of pancreatic ductal adeno- eral population. In individuals with
carcinoma, from the Cancer Genome Atlas Research Network [24]. lncRNA, long non- significantly increased risk of pan-
coding RNA; mRNA, messenger RNA; miRNA, microRNA; WT, wild-type.
creatic cancer on the basis of fam-
ily history and genetic risk factors,
imaging of the pancreas is per-
formed for screening. Endoscopic
ultrasonography and magnetic res-
onance imaging (MRI) or magnetic
resonance cholangiopancreatogra-
phy (MRCP) are used in the clinical
setting. However, clear definitions
of who should be screened and at
what age screening should com-
mence have not been formalized.
The potential benefit of screen-
ing of high-risk individuals has been
demonstrated in a study in Europe,
which noted that CDKN2A mutation
carriers were more likely to be di-
agnosed with a resectable pancre-
atic cancer and had a higher 5-year
survival rate [32]. Recent data from
the International Cancer of the
Pancreas Screening Consortium
showed that 9 of 10 screen-detect-
ed pancreatic cancers were resect-
able, suggesting a benefit of screen-
ing in individuals at high risk [33].
An effort to engage in larger-scale,
collaborative consortia is needed to
provide more rigorous evidence of
the value of screening of high-risk
change in blood glucose level, and Screening and
age at onset of diabetes. The model individuals. Patients with new-on-
identification of set diabetes and intraductal papil-
identified patients who developed
pancreatic cancer within 3 years of
high-risk groups lary mucinous neoplasms are also
onset of diabetes with an area under No reliable screening test is cur- groups with elevated risk in which
the receiver operating characteristic rently available for the early detec- studies of the benefits of screening
curve value of 0.87 [31]. tion of pancreatic cancer in the gen- are under way.
Fig. 5.7.4. The elevation of fasting blood glucose (FBG) levels beginning 30–36 months before diagnosis of pancreatic ductal
adenocarcinoma (PDAC) is an area of interest for early detection strategies.
SECTION 5
CHAPTER 5.7

Prevention a diet high in fruits and vegetables, for vitamin D the results have been
regular physical activity, and avoid- mixed. Non-aspirin non-steroidal
Risk factors such as age, attained
adult height, race, and family his- ing heavy alcohol consumption. anti-inflammatory drugs do not
tory cannot be modified, but primary In the absence of effective appear to have an effect on risk
prevention by the alteration of modi- screening methods, options for pri- [34]. Metformin appears to protect
fiable risk factors has the potential mary prevention of pancreatic can- against genomic instability through
to decrease the overall risk of pan- cer are of significant importance, various mechanisms in vitro, and
creatic cancer and warrants further and chemoprevention for pancre- metformin in combination with as-
study. Potentially modifiable risk fac- atic cancer is a high priority for pirin has been shown to inhibit tu-
tors include smoking, obesity, diabe- translational research. A review of mour growth in a mouse model of
tes, diet, and alcohol consumption. epidemiological data performed by pancreatic cancer [35]. These stud-
The best strategy for risk reduction is a working group in 2015 suggested ies have provided some insights for
lifestyle modification: smoking ces- that aspirin and statins may provide planning future prospective preven-
sation, maintaining a healthy weight, some protective effect, whereas tion trials.
References
1. Rahib L, Smith BD, Aizenberg R, 7. Behrens G, Jochem C, Schmid D, 12. Incio J, Liu H, Suboj P, Chin SM, Chen
Rosenzweig AB, Fleshman JM, Matrisian Keimling M, Ricci C, Leitzmann MF (2015). IX, Pinter M, et al. (2016). Obesity-
LM (2014). Projecting cancer incidence and Physical activity and risk of pancreatic induced inflammation and desmoplasia
deaths to 2030: the unexpected burden of cancer: a systematic review and meta- promote pancreatic cancer progres-
thyroid, liver, and pancreas cancers in the analysis. Eur J Epidemiol. 30(4):279–98. sion and resistance to chemotherapy.
United States. Cancer Res. 74(11):2913–21. https://doi.org/10.1007/s10654-015-0014-9 Cancer Discov. 6(8):852–69. https://doi.
https://doi.org/10.1158/0008-5472.CAN- PMID:25773752 o r g /10 .115 8 / 215 9 - 8 2 9 0 .C D -15 -117 7
14-0155 PMID:24840647 PMID:27246539
8. Jiao L, Berrington de Gonzalez A, Hartge
P, Pfeiffer RM, Park Y, Freedman DM, 13. Harvey AE, Lashinger LM, Hays D,
2. Maisonneuve P, Lowenfels AB (2015).
et al. (2010). Body mass index, effect Harrison LM, Lewis K, Fischer SM, et
Risk factors for pancreatic cancer: a sum- modifiers, and risk of pancreatic cancer: a
mary review of meta-analytical studies. al. (2014). Calorie restriction decreases
pooled study of seven prospective cohorts. murine and human pancreatic tumor cell
Int J Epidemiol. 44(1):186–98. https://doi. Cancer Causes Control. 21(8):1305–14. growth, nuclear factor-κB activation, and
org/10.1093/ije/dyu240 PMID:25502106 https://doi.org/10.1007/s10552-010-9558-x inflammation-related gene expression in
PMID:20383573 an insulin-like growth factor-1-dependent
3. Riboli E, Kaaks R (1997). The EPIC proj-
ect: rationale and study design. European 9. Bosetti C, Rosato V, Li D, Silverman D, manner. PLoS One. 9(5):e94151. https://
Prospective Investigation into Cancer and Petersen GM, Bracci PM, et al. (2014). doi.org/10.1371/journal.pone.0094151
Nutrition. Int J Epidemiol. 26(Suppl 1):S6– Diabetes, antidiabetic medications, and PMID:24804677
14. https://doi.org/10.1093/ije/26.suppl_1. pancreatic cancer risk: an analysis from
14. Wang L, Bai Y-Y, Yang Y, Hu F, Wang
S6 PMID:9126529 the International Pancreatic Cancer
Y, Yu Z, et al. (2016). Diabetes mellitus
Case-Control Consortium. Ann Oncol.
4. Belanger CF, Hennekens CH, Rosner B, stimulates pancreatic cancer growth
25(10):2065–72. https://doi.org/10.1093/
annonc/mdu276 PMID:25057164 and epithelial-mesenchymal transition-
Speizer FE (1978). The Nurses’ Health
mediated metastasis via a p38 MAPK
Study. Am J Nurs. 78(6):1039–40.
10. Eibl G, Cruz-Monserrate Z, Korc M, pathway. Oncotarget. 7(25):38539–50.
PMID:248266
Petrov MS, Goodarzi MO, Fisher WE, et https://doi.org/10.18632/oncotarget.9533
5. Rimm EB, Stampfer MJ, Colditz GA, al.; Consortium for the Study of Chronic PMID:27413117
Giovannucci E, Willett WC (1990). Pancreatitis, Diabetes, and Pancreatic
Cancer (2018). Diabetes mellitus and 15. Wang Z, Lai ST, Xie L, Zhao JD, Ma
Effectiveness of various mailing strategies NY, Zhu J, et al. (2014). Metformin is
among nonrespondents in a prospective obesity as risk factors for pancreatic
cancer. J Acad Nutr Diet. 118(4):555–67. associated with reduced risk of pancre-
cohort study. Am J Epidemiol. 131(6):1068– atic cancer in patients with type 2 dia-
https://doi.org/10.1016/j.jand.2017.07.005
71. https://doi.org/10.1093/oxfordjournals. betes mellitus: a systematic review and
PMID:28919082
aje.a115598 PMID:2343859 meta-analysis. Diabetes Res Clin Pract.
11. Dawson DW, Hertzer K, Moro A, Donald 106(1):19–26. https://doi.org/10.1016/j.
6. WCRF/AICR (2018). Continuous Update G, Chang H-H, Go VL, et al. (2013). diabres.2014.04.007 PMID:24837144
Project Expert Report 2018. Diet, nutri- High-fat, high-calorie diet promotes early
tion, physical activity and pancreatic can- pancreatic neoplasia in the conditional
cer risk. World Cancer Research Fund/ KrasG12D mouse model. Cancer Prev
American Institute for Cancer Research. Res (Phila). 6(10):1064–73. https://doi.
Available from: https://www.wcrf.org/sites/ org/10.1158/1940 - 6207.CAPR-13- 0065
default/files/Pancreatic-cancer-report.pdf. PMID:23943783
372
16. Kato K, Iwama H, Yamashita T, Kobayashi 23. Nakatochi M, Lin Y, Ito H, Hara K, 30. Sharma A, Smyrk TC, Levy MJ, Topazian
K, Fujihara S, Fujimori T, et al. (2016). Kinoshita F, Kobayashi Y, et al. (2018). MA, Chari ST (2018). Fasting blood glucose
The anti-diabetic drug metformin inhib- Prediction model for pancreatic cancer levels provide estimate of duration and
its pancreatic cancer cell proliferation in risk in the general Japanese popula- progression of pancreatic cancer before
vitro and in vivo: study of the microRNAs tion. PLoS One. 13(9):e0203386. https:// diagnosis. Gastroenterology. 155(2):490–
associated with the antitumor effect of doi.org/10.1371/journal.pone.0203386 500.e2. https://doi.org/10.1053/j.gastro.
metformin. Oncol Rep. 35(3):1582–92. PMID:30192808 2018.04.025 PMID:29723506
ht tps://doi.org /10.3 8 92 /or. 2015.4 49 6
PMID:26708419 24. The Cancer Genome Atlas Research 31. Sharma A, Kandlakunta H, Nagpal SJS,
Network (2017). Integrated genomic char- Feng Z, Hoos W, Petersen GM, et al.
17. Tong GX, Geng Q-Q, Chai J, Cheng acterization of pancreatic ductal adenocar- (2018). Model to determine risk of pancreat-
J, Chen P-L, Liang H, et al. (2014). cinoma. Cancer Cell. 32(2):185–203.e13. ic cancer in patients with new-onset diabe-
Association between pancreatitis and https://doi.org/10.1016/j.ccell.2017.07.007 tes. Gastroenterology. 155(3):730–739.e3.
subsequent risk of pancreatic cancer: PMID:28810144 https://doi.org/10.1053/j.gastro.2018.05.
a systematic review of epidemiologi- 023 PMID:29775599
cal studies. Asian Pac J Cancer Prev. 25. Bailey P, Chang DK, Nones K, Johns AL,
15(12):5029–34. https://doi.org/10.7314/ Patch A-M, Gingras M-C, et al.; Australian 32. Vasen H, Ibrahim I, Ponce CG, Slater EP,
APJCP.2014.15.12.5029 PMID:24998582 Pancreatic Cancer Genome Initiative Matthäi E, Carrato A, et al. (2016). Benefit
(2016). Genomic analyses identify molecu- of surveillance for pancreatic cancer in
18. Yurgelun MB, Chittenden AB, Morales- lar subtypes of pancreatic cancer. Nature. high-risk individuals: outcome of long-term
Oyarvide V, Rubinson DA, Dunne RF, 531(7592):47–52. https://doi.org/10.1038/ prospective follow-up studies from three
Kozak MM, et al. (2019). Germline can- nature16965 PMID:26909576 European expert centers. J Clin Oncol.
cer susceptibility gene variants, somatic 34(17):2010–9. https://doi.org/10.1200/
second hits, and survival outcomes in 26. Kim J, Bamlet WR, Oberg AL, Chaffee JCO.2015.64.0730 PMID:27114589
patients with resected pancreatic can- KG, Donahue G, Cao X-J, et al. (2017).
cer. Genet Med. 21(1):213–23. https:// Detection of early pancreatic ductal ade- 33. Canto MI, Almario JA, Schulick RD, Yeo
d o i.o r g /10.10 3 8 /s 414 3 6 - 018 - 0 0 0 9 - 5 nocarcinoma with thrombospondin-2 and CJ, Klein A, Blackford A, et al. (2018).
PMID:29961768 CA19-9 blood markers. Sci Transl Med. Risk of neoplastic progression in indi-
9(398):eaah5583. https://doi.org/10.1126/ viduals at high risk for pancreatic can-
19. Hu C, Hart SN, Polley EC, Gnanaolivu cer undergoing long-term surveillance.
scitranslmed.aah5583 PMID:28701476
R, Shimelis H, Lee KY, et al. (2018). Gastroenterology. 155(3):740 –751.e2.
Association between inherited germline 27. Cohen JD, Javed AA, Thoburn C, Wong ht tps://doi.org/10.1053/j.gastro. 2018.
mutations in cancer predisposition genes F, Tie J, Gibbs P, et al. (2017). Combined 05.035 PMID:29803839
and risk of pancreatic cancer. JAMA. circulating tumor DNA and protein bio-
319(23):2401–9. https://doi.org/10.1001/ marker-based liquid biopsy for the ear- 34. Miller MS, Allen P, Brentnall TA, Goggins
jama.2018.6228 PMID:29922827 lier detection of pancreatic cancers. Proc M, Hruban RH, Petersen GM, et al. (2016).
Natl Acad Sci U S A. 114(38):10202–7. Pancreatic Cancer Chemoprevention
20. Brand R, Borazanci E, Speare V, Translational Workshop: meeting report.
Dudley B, Karloski E, Peters MLB, et https://doi.org/10.1073/pnas.1704961114
PMID:28874546 Pancreas. 45(8):1080–91. https://doi.
al. (2018). Prospective study of germ-
org/10.1097/MPA.0000000000000705
line genetic testing in incident cases of 28. Mellby LD, Nyberg AP, Johansen JS, PMID:27518363
pancreatic adenocarcinoma. Cancer. Wingren C, Nordestgaard BG, Bojesen SE,
124(17):3520–7. https://doi.org/10.1002/ et al. (2018). Serum biomarker signature- 35. Yue W, Zheng X, Lin Y, Yang CS, Xu Q,
cncr.31628 PMID:30067863 based liquid biopsy for diagnosis of early- Carpizo D, et al. (2015). Metformin com-
stage pancreatic cancer. J Clin Oncol. bined with aspirin significantly inhibit
21. Campa D, Matarazzi M, Greenhalf W,
36(28):2887–94. https://doi.org/10.1200/ pancreatic cancer cell growth in vitro and
Bijlsma M, Saum KU, Pasquali C, et al.
JCO.2017.77.6658 PMID:30106639 in vivo by suppressing anti-apoptotic
(2019). Genetic determinants of telo-
proteins Mcl-1 and Bcl-2. Oncotarget.
mere length and risk of pancreatic can-
29. Fan X, Alekseyenko AV, Wu J, Peters BA, 6(25):21208–24. https://doi.org/10.18632/
cer: a PANDoRA study. Int J Cancer.
Jacobs EJ, Gapstur SM, et al. (2018). oncotarget.4126 PMID:26056043
144(6):1275–83. PMID:30325019
Human oral microbiome and prospec-
22. Klein AP, Wolpin BM, Risch HA, tive risk for pancreatic cancer: a popu-
Stolzenberg-Solomon RZ, Mocci E, Zhang lation-based nested case-control study.
M, et al. (2018). Genome-wide meta-analy- Gut. 67(1):120–7. https://doi.org/10.1136/
sis identifies five new susceptibility loci for gutjnl-2016-312580 PMID:27742762
pancreatic cancer. Nat Commun. 9(1):556.
ht t p s: //d o i .o r g /10 .10 3 8 /s 414 6 7- 018 -
02942-5 PMID:29422604
SECTION 5
CHAPTER 5.7

5.8
Skin cancer
A focus on primary prevention
David Whiteman Bruce K. Armstrong (reviewer)

Rüdiger Greinert (reviewer)
Massimo Tommasino (reviewer)
genes as well as through local- dermis and dermis. The most com-
SUMMARY ized immunosuppression. High monly described subtypes are
mutational burdens have been superficial spreading melanomas
●● The highest incidence rates of identified in both tumour types, (with an initial radial growth phase
skin cancer are observed in consistent with extensive ultra- in the epidermis, followed by dermal
the predominantly fair-skinned violet radiation-induced dam- invasion) and nodular melanomas
populations living in areas with age, but the driver genes differ (with early vertical growth and little
very high ambient levels of so- between the two. or no radial growth). Lentigo malig-
lar radiation, such as Australia na melanomas occur on chronically
and New Zealand. sun-damaged skin, and acral len-
●● Genes associated with pigmen- Cancers of the skin are the most tiginous melanomas are distinctive
tation or with naevi, together common cancer type in humans. tumours that arise on palmar and
with DNA repair genes and The term “skin cancer” covers a plantar surfaces.
other genes of unknown func- range of pathological entities that Histological characteristics of
arise from different cells of the epi- melanomas, notably tumour thick-
tion, have been confirmed to in-
dermis and dermis. This chapter ness and presence of ulceration,
crease heritable melanoma risk.
is restricted to cutaneous melano- correlate strongly with mortality.
●● Genes critical for melanoma mas and the keratinocyte cancers The American Joint Committee on
development, which often have (basal cell carcinomas and squa- Cancer incorporates these prog-
ultraviolet radiation-induced mu- mous cell carcinomas). nostic features into its staging sys-
tation, include genes that control tem. Recent analyses of long-term
cell proliferation (e.g. BRAF), Melanoma survival have led to changes in mel-
cell cycle and replication (e.g. anoma staging criteria, particularly
TP53), and metabolic pathways. Pathology for thinner lesions [1]. The eighth
●● Cutaneous melanomas may Melanoma, the most aggressive (2017) edition of the American Joint
arise from a pre-existing be- type of skin cancer, arises from Committee on Cancer staging sys-
nign naevus or occur on chroni- melanocytes – pigment-producing tem recognizes a new threshold
cells in the skin. Most melanomas for melanoma thickness (0.8 mm),
cally sun-damaged skin. Since
(> 95%) are cutaneous tumours that which now separates T1a from T1b
2007, the incidence of melano-
arise on skin surfaces exposed to melanomas. Also, whereas earlier
ma has been declining overall
the sun, but melanomas also oc- staging criteria incorporated both
in Australia, driven largely by
cur on skin of the palms and soles. ulceration and tumour mitotic rate
significant reductions in recent
Melanomas also occur in the eye, as prognostic features, in the eighth
birth cohorts, consistent with a
in the meninges, and on mucous edition only ulceration has been re-
successful intervention to re-
membranes of the gastrointestinal tained (Table 5.8.1).
duce sun exposure.
and genital tracts; these types of The most recent (2018) edition
●● Sunlight is the principal environ- melanoma are not discussed here. of the WHO classification of skin tu-
mental cause of basal cell carci- Various histological subtypes of mours introduced a pathway-based
noma and squamous cell carci- cutaneous melanomas are recog- classification of melanoma, which
noma, mediated through direct nized, reflecting patterns of growth explains many of the differences
mutagenic effects on regulatory and attendant changes in the epi- in pathology and clinical behaviour
374
Table 5.8.1. Categorization of primary cutaneous melanoma on the basis of histological

characteristics of the primary tumour, according to the eighth (2017) edition of the
American Joint Committee on Cancer staging system for melanoma FUNDAMENTALS
T category Thickness (mm) Ulceration ■■ Melanoma is a potentially
aggressive cancer that arises
Tis (melanoma in situ) Not applicable Not applicable
from pigment-producing cells
T1 ≤ 1.0 Unknown or unspecified
T1a < 0.8 Ulceration absent in the skin. The incidence of
T1b < 0.8 Ulceration present melanoma has been rising in
T1b 0.8–1.0 Ulceration present or absent
most populations with predom-
T2 > 1.0–2.0 Unknown or unspecified inantly European ancestry.
T2a > 1.0–2.0 Ulceration absent
T2b > 1.0–2.0 Ulceration present ■■ Recent studies have docu-
T3 > 2.0–4.0 Unknown or unspecified mented the extremely high
T3a > 2.0–4.0 Ulceration absent burden of mutations in the
T3b > 2.0–4.0 Ulceration present
melanoma genome induced
T4 > 4.0 Unknown or unspecified
T4a > 4.0 Ulceration absent
by ultraviolet radiation. This
T4b > 4.0 Ulceration present confirms earlier epidemiologi-
cal observations that the inci-
between the different types. The New Zealand (~50 per 100 000 per- dence of melanoma is strongly
primary diagnostic tool remains his- son-years). In those populations, correlated with ambient levels
topathology, and the histopathologi- melanomas are the most common of solar radiation.
cal patterns recognized by patholo- cancer type in people younger than
■■ The constitutional genes that
gists have now very clearly been 40 years, and are among the most
shown to correspond to distinct common cancer type overall. The in- confer susceptibility to mela-
genetic profiles. The classification cidence of melanoma is also high in noma include those associated
of melanoma is divided into nine low-latitude parts of North America with pigmentation characteris-
pathways. The tumours included in (~30 per 100 000 person-years), tics as well as telomere length
three of these pathways are com- and there is an overall inverse gra- and cell-cycle control.
mon at sun-exposed sites, and the dient of incidence with increasing
■■ Immunotherapies and targeted
remainder are tumours that are less latitude. At higher latitudes in both
common (although important be- North America and Europe, the in- therapies have recently shown
cause of their global occurrence) cidence of melanoma has been ris- enormous promise in treating
and arise in sun-shielded skin, in ing steadily in recent decades; this metastatic melanoma; this
mucosae, and in the eye. The mela- trend is probably due to the advent area of research is developing
nomas that occur at sun-exposed of inexpensive leisure travel and the very quickly and will change
sites are subdivided according to widespread use of tanning devices rapidly in the next few years.
whether they are associated with a (sunlamps and sunbeds).
low degree or a high degree of cu- Melanoma remains an uncom- ■■ Basal cell carcinomas and
mulative sun damage [2]. mon cancer in Central and South squamous cell carcinomas
America, Asia, Africa, and the are the most common cancer
Epidemiology Pacific (< 3 per 100 000 person- types in humans. They are
In 2018, there were estimated to years). In recent years, the inci- caused by sunlight and are
be almost 290 000 new cases of dence of melanoma has been fall- largely preventable through
CHAPTER 5.8
SECTION 5
melanoma and about 61 000 deaths ing in Australia, particularly in more control programmes.
from melanoma worldwide [3]. The recent birth cohorts; this is consis-
global range of population incidence tent with the impact of prolonged
of melanoma is the greatest of any public health campaigns (as dis-
cancer type. The incidence in a giv- cussed below). of cutaneous melanomas, manifest-
en region is determined largely by ing as a very high mutational burden
the pigmentation characteristics of Risk factors in key regulatory genes that bear
individuals in that population and the Observational epidemiological stud- UVB signature mutations [7] (see
ambient levels of solar radiation. ies long ago identified both solar ul- Chapter 2.4). In addition to solar UV
The highest incidence is ob- traviolet (UV) radiation [4] and host radiation, there is strong evidence
served in the predominantly fair- factors [5,6] as causes of melanoma. that repeated exposures to artificial
skinned populations living in areas Recent genomic sequencing studies sources of UV radiation from tan-
with very high ambient levels of so- have confirmed the causal role of ning devices and phototherapy also
lar radiation, such as Australia and UVB radiation for the vast majority increase risk of melanoma.
Chapter 5.8 • Skin cancer 375

Fig. 5.8.1. Malignant melanoma. At the TERT), and the remainder are pre- highest rate observed among all
left is a plaque of early-stage superficial sumed to carry private mutations solid tumours); this is largely due to
spreading melanoma in the radial growth [9]. However, for most patients damage from UV radiation. The very
phase. At the right, contiguous with the
genetic susceptibility is conferred high rate of mutations in melanoma
plaque, is a pink (amelanotic) nodule of
deeply invasive melanoma in the vertical through multiple polymorphisms presented an analytical challenge
growth phase. Melanomas diagnosed at in low-risk genes that act through when attempting to identify which
this stage have a poor prognosis. many different pathways. of the mutations were “drivers” (i.e.
The genes first linked to mela- those occurring in key genes at criti-
noma were candidates identified cal points in the evolution of melano-
through their association with pig- ma) and which were “passengers”.
mentation characteristics. Of these, Using sophisticated bioinfor-
the highly polymorphic gene that matics techniques that control for
encodes the melanocortin 1 recep- patient-specific and gene-specific
tor (MC1R) is the most prevalent parameters, investigators have con-
and the most strongly associated verged on a core group of genes
with melanoma. A large and grow- that are critical for melanoma de-
ing number of genes associated velopment. These include genes
either with pigmentation (ASIP, that control cell proliferation (BRAF,
TYR, and SLC45A2) or with naevi NRAS, and NF1), cell cycle and
(CDKN2A-MTAP, PLA2G6, and replication (CDKN2A, TP53, and
TERT) have also been confirmed to TERT), and metabolic pathways
increase risk of melanoma. Large (PTEN and KIT). Other genes that
meta-analyses of genome-wide as- have been shown to be important in
sociation studies have extended subsets of cutaneous melanomas
the list of confirmed gene variants include RAC1, MAP2K1, PPP6C,
associated with melanoma to at ARID1, IDH1, and RB1.
Host factors that confer an in-
least 20, including several genes The mutational spectrum for cu-
creased risk of melanoma relate to
not associated with pigmentation or taneous melanomas differs accord-
the function or number of melano- ing to anatomical site, as predicted
with naevi [10]. Other variants that
cytes. Overall, the strongest pheno- by earlier epidemiological studies.
have been confirmed are for genes
typic risk factor for melanoma is the Melanomas that occur at habitually
involved in DNA repair (PARP1 and
propensity to develop large numbers
ATM), as well as genes for which
of melanocytic naevi (moles) on the
the functional relevance remains Fig. 5.8.2. A woman applying sunscreen.
skin. People with very large num-
unclear (ARNT-SETDB1, CASP8, Use of sunscreen has a recognized role
bers of naevi (> 100) have risks of
FTO, and MX2). To date, no sus- in reducing the burden of skin cancer in
melanoma up to 7 times those in
ceptibility loci have been identified fair-skinned populations.
people with very few naevi (< 15)
for acral melanomas.
[5]. The pigmentation characteris-
tics consistently associated with in- Somatic mutations
creased risks of melanoma include With the advent of high-throughput
fair skin that burns and does not genomic sequencing (see Chapter
tan, red or light hair, blue eyes, and 3.2), in the past few years there has
the propensity to develop freckles; been an explosion in knowledge
therefore, melanoma is rare in popu- about the cascade of mutations that
lations with non-European ancestry lead to melanoma. The first report
[6]. Immunosuppression increases described the mutational burden
the risk of melanoma 2–3-fold [8]. in a cell line derived from a meta-
static deposit in one patient [11].
Constitutional genetics Subsequent investigations expand-
About 5–10% of patients with cu- ed the catalogue; hundreds of mel-
taneous melanoma have a strong anomas have now been sequenced
family history of the disease. About [7,12], including growing numbers
half of these patients are found to of acral, desmoplastic, and uveal
carry a highly penetrant germline melanomas [13].
mutation in one of a small number All sequencing studies have re-
of genes (in descending order of ported exceptionally high mutational
frequency: CDKN2A, CDK4, BAP1, burdens in cutaneous melanomas
MITF, POT1, ACD, TERF2IP, and (> 10 mutations per megabase, the
376
sun-exposed sites have markedly existing benign naevus (the naevus tural rearrangements. Melanomas
higher overall mutational loads than pathway). Other cutaneous melano- that arise through the chronic sun
those that occur at sun-shielded mas, particularly those that occur exposure pathway exhibit a differ-
sites. Thus, mucosal and acral mel- on chronically sun-damaged skin, ent sequence of driver mutations,
anomas exhibit strikingly different do not arise from pre-existing naevi often harbouring mutations in
mutational spectra from other cuta- but rather arise through a variety NRAS and NF1, as well as muta-
neous melanomas, with much lower of intermediate lesions (e.g. lentigo tions in TERT promoter sites and
mutation frequencies overall and dif- maligna) or frankly invasive tumours heterozygous CDKN2A mutations
ferent driver genes implicated [13]. (nodular melanoma), which are as- [17] (Fig. 5.8.4).
Mutations in TP53, PTEN, or RB1 sociated epidemiologically with high
are infrequent in acral melanomas, levels of cumulative sun exposure. Prevention
but a diverse range of triple wild- For tumours that arise through Primary prevention
type mutations are evident, includ- the naevus pathway, the initial mu- Despite exciting progress in new
ing mutations in KIT and GNAQ, as tation is in BRAF. In the absence therapies to treat melanoma, pre-
well as notably higher occurrence of of any further mutations, the nae- ventive strategies remain of para-
breakpoints and structural variants. vus enters a senescence-like state mount importance to deliver cost-
and eventually involutes in middle effective melanoma control. The
Pathogenesis life. However, a very small fraction population attributable fraction esti-
Recent studies have sought to over- of naevi acquire additional muta- mates the proportion of melanoma
lay the sequence order in which driv- tions in targets such as TERT pro- that would, in theory, be prevented if
er mutations are acquired onto the moter sites (probably due to ad- exposure to sunlight was reduced to
histologically discernible stages of ditional exposure to UV radiation, historical lows. For populations with
progression from benign melanocyt- although other mutagens are also predominantly European ancestry,
ic tumours to metastatic melanoma possible), followed by biallelic loss the population attributable fraction
[14]. Findings from epidemiological of CDKN2A. Combinations of muta- for exposure to solar UV radiation
studies about 30 years ago and sub- tional events of this type allow the has been variously estimated at 65–
sequent genetic studies led to and naevus to escape senescence and 90%, with most estimates closer to
elaborated the hypothesis that cuta- acquire proliferative and invasive the upper bound, underscoring the
neous melanomas can arise through characteristics, eventually leading potential gains to be had from pri-
multiple pathways, depending on the to metastasis. At this later stage, as mary prevention [18]. Encouraging
anatomical site of the target cell, the the cancers are becoming invasive, behaviours that minimize hazard-
age and constitutional characteris- it appears that they acquire addi- ous exposure to sunlight remains
tics of the host, and the pattern of tional mutations in TP53 and PTEN, the mainstay of primary prevention
sun exposure [15,16]. Many cutane- as well as increasing frequencies of efforts, supported by evidence that
ous melanomas arise from a pre- copy number alterations and struc- regularly applying sunscreen signifi-
cantly reduces the risk of melanoma
[19]. In many jurisdictions, the use of
Fig. 5.8.3. Children playing on the beach wearing sun-protective clothing. Sun protec- tanning devices is being restricted
tion at an early age and avoidance of sunburn are key goals in programmes aimed at
through regulation.
reducing the incidence of skin cancer.
Primary prevention campaigns
have been running in Australia
since the 1980s and have focused
on reducing sun exposure through
rescheduling outdoor activities,
CHAPTER 5.8
SECTION 5
seeking shade, using clothing to
protect the skin, and applying sun-
screen to exposed body sites. There
is moderately strong evidence from
controlled trials that sun protection
including use of sunscreen reduces
development of naevi and risk of
melanoma [19,20]. Since 2007, the
incidence of melanoma has been
declining overall in Australia, driven
largely by significant reductions in
recent birth cohorts, consistent with
a successful intervention to reduce
sun exposure [21].

Fig. 5.8.4. Schematic pathogenesis of cutaneous melanomas. Cutaneous melanomas arise on a background of susceptibility
conferred by a large number of genetic variants. Most cutaneous melanomas appear to be initiated by exposure to ultraviolet B
radiation in early life, which causes mutations in BRAF in melanocytes. Further progression depends on the site of the target cell
and the genetic background of the host, but several key driver genes appear to be important in all pathways.
Early detection and screening Keratinocyte cancers cy and the attendant costs of diag-
Currently, no national or international Keratinocyte cancers of the skin – nosis and surgery.
authorities (except in Germany) rec- basal cell carcinomas (BCCs) and
Etiology
ommend population-based screen- squamous cell carcinomas (SCCs) –
ing for melanoma, based on the as- are the most common cancer types BCCs are slow-growing tumours
sessment that there is insufficient in humans. Although mortality rates that occur most frequently on the
evidence of mortality benefit. In from these cancer types are very face, neck, shoulders, and chest
most jurisdictions where melanoma low, they impose a heavy financial of fair-skinned people who are ex-
is prevalent, people deemed at high burden on health systems in many posed to high levels of solar radia-
risk are advised to engage in early countries, because of their frequen- tion. BCCs can be locally invasive
detection strategies. Several predic-
tion algorithms have been developed Fig. 5.8.5. Deliberate sun exposure by fair-skinned people to attain a tanned appear-
to identify those at high risk, incorpo- ance is at odds with cancer prevention.
rating information on demographic,
phenotypic, and clinical factors [22]
and, in some instances, genetic data
as well. The performance of these
tools varies and is influenced by set-
ting-specific characteristics including
ambient insolation and population
diversity, but discrimination indices
of 0.65–0.75 are typical, which is
indicative of moderate accuracy. In
Germany, a biannual skin cancer
screening programme was intro-
duced nationwide in 2008 for insured
people 35 years and older. As yet,
there is no evidence of a sustained
change in mortality from melanoma
after the introduction of the screening
programme [23].
378
but rarely metastasize. The cell of Fig. 5.8.6. The shade structure above this basketball court in San Antonio, Texas,
origin remains an open question, USA, provides sun protection.
but emerging consensus points to
cells in the hair follicle.
SCCs are epidermal cancers
that grow more rapidly and are
much more likely to invade and me-
tastasize. SCCs arise on habitually
sun-exposed sites, particularly the
face, ears, neck, and exposed sur-
faces of the limbs. Precancerous
skin lesions that have similar mor-
phology to SCCs include actinic
keratoses (sunspots), intraepider-
mal or in situ SCCs, and Bowen
disease. There is debate about
whether these are true precursors
of SCC or concomitant actinic le-
sions, and about whether the term
“Bowen disease” encompasses all
intraepidermal SCCs [24].
Epidemiology with organ transplantation, have the Genetics

Because BCC and SCC primar- highest SCC multiplicity rates [27]. Several very rare but highly pen-
ily (although not exclusively) affect
etrant gene loci have been identified
populations of European ances- Risk factors
in families with clinical syndromes
try, incidence correlates strongly Sunlight is the principal environ- characterized by very high incidence
with ambient insolation. Therefore, mental cause of BCC and SCC, of BCC. Mutation or deletion of the
these cancer types occur most fre- mediated through direct mutagenic PTCH1 gene is the cause of Gorlin
quently among the fair-skinned res- effects on key regulatory genes as syndrome, an autosomal dominantly
idents of Australia, New Zealand, well as through localized immu- inherited disease characterized by a
and low-latitude states of the USA. nosuppression. As noted, people very high risk of BCC, an increased
However, the incidence of BCC who are immunosuppressed, either risk of some other (mainly benign)
and SCC has been rising rapidly in therapeutically (e.g. after organ neoplasms, and some non-neo-
most European countries in recent transplantation) or as a result of plastic manifestations. Families with
decades; currently, the incidence disease (e.g. HIV/AIDS), may have germline mutations in several DNA
in Scandinavian countries is ap- markedly increased incidence of repair genes (XPA1, XPA2, XRCC2,
proaching that in the USA [25]. SCC, and to a lesser extent BCC. and XRCC3) exhibit several different
For both BCC and SCC, the inci- Other environmental factors that traits, including extreme sensitivity
dence increases with age, although are known to increase the risks of to UV radiation. Such patients mani-
BCCs tend to present at earlier cutaneous SCC include exposure fest with multiple, early-onset SCCs.
ages than SCCs and the age effect to arsenic, polycyclic aromatic hy- In the general population, host
is much stronger for SCCs than for drocarbons, and ionizing radiation susceptibility is conferred by poly-
CHAPTER 5.8
SECTION 5
BCCs. Consequently, the ratio of (particularly for BCC). morphisms in many genes, all with
BCC to SCC changes rapidly, from Cutaneous infection with human small effect. Genome-wide associa-
about 10:1 at age 40 years to about papillomavirus (HPV), specifically tion studies have confirmed a suite
3:1 at age 60 years. the beta types, has been repeat- of previously identified pigmenta-
Both BCC and SCC are prone edly implicated as a cause of SCC, tion genes as risk loci for BCC and
to multiplicity. Data from Australia although the connection is not SCC, including MC1R, ASIP, TYR,
suggest that most people who de- completely certain and the precise SLC45A2, OCA2, IRF4, and BNC2.
velop one lesion will develop more mechanism remains open to ques- At least 31 loci have now been impli-
within 3 years; a small proportion tion [28]. A suite of phenotypic char- cated in BCC [29]. Recently, four loci
will develop more than 20 can- acteristics confers increased risks not known to be associated with pig-
cers, and this has important conse- for both BCC and SCC, including mentation were identified as putative
quences for detection and control fair skin that does not tan, light or risk loci exclusively for SCC: 2p22.3,
[26]. People who are immunosup- red hair, propensity to freckling, and AHR, SEC16A, and CADM1-BUD13
pressed, particularly in connection blue eyes. [30]. The mechanisms enabled by

polymorphism of these loci remain Fig. 5.8.7. Genetic makeup corresponding to a Celtic complexion – as characterized by
to be elucidated. blue eyes, red hair, and fair skin – contributes markedly to increased risk of skin cancer.
Sequencing studies have iden-
tified extremely high mutational
burdens in both BCC and SCC,
consistent with extensive UV ra-
diation-induced damage, but the
lists of driver genes differ for BCC
and SCC. Mutations in genes in
the hedgehog pathway appear to
be critical for BCC development,
particularly PTCH1 and SMO [31].
TP53 is also very often mutated
in BCC. Recurrent mutations have
also been reported in MYCN,
PPP6C, STK19, LATS1, ERBB2,
PIK3CA, and the RAS family.
For SCC, NOTCH1 appears to
be a gatekeeper, although muta-
tions in other key genes such as
TP53, CDKN2A, and HRAS (some-
times within the same tumour) sug-
gest that tumours arise through candidates, CDKN2A may be the use of sunscreen for preventing
key suppressor of SCC formation. SCC and actinic keratoses, but not
multiple pathways and may be poly-
BCC. It is possible that the lack of
clonal in origin. NOTCH1 plays a Prospects any observed effect for BCC was
key role in cell–cell signalling and
Although mortality from BCC and because the intervention was deliv-
serves to regulate the switch be- ered to adults, and not earlier in life.
SCC is very low, these cancer
tween proliferation and differen- types exact a sizeable toll in terms Encouraging behaviours that mini-
tiation of keratinocytes; hence, it of morbidity and costs. The recent mize hazardous exposure to sun-
is a highly credible candidate [32]. steady rises in incidence reported light remains the mainstay of prima-
Notably, many of the driver muta- across Europe and North America ry prevention efforts, complemented
tions in SCC, except for CDKN2A, are likely to continue in the absence by regulating against the use of tan-
are also readily detectable in mac- of systematic primary prevention ning devices and other sources of
roscopically normal photo-exposed campaigns. Randomized trials have artificial UV radiation.
skin [33], suggesting that of all the demonstrated the benefit of daily
References
1. Gershenwald JE, Scolyer RA, Hess KR, 3. Bray F, Ferlay J, Soerjomataram I, Siegel 5. Gandini S, Sera F, Cattaruzza MS,
Sondak VK, Long GV, Ross MI, et al. RL, Torre LA, Jemal A (2018). Global Pasquini P, Abeni D, Boyle P, et al.
(2017). Melanoma staging: evidence- cancer statistics 2018: GLOBOCAN es- (2005). Meta-analysis of risk factors for
based changes in the American Joint timates of incidence and mortality world- cutaneous melanoma: I. Common and
Committee on Cancer eighth edition can- wide for 36 cancers in 185 countries. CA atypical naevi. Eur J Cancer. 41(1):28–44.
cer staging manual. CA Cancer J Clin. Cancer J Clin. 68(6):394–424. https://doi. https://doi.org/10.1016/j.ejca.2004.10.015
67(6):472–92. https://doi.org/10.3322/ org/10.3322/caac.21492 PMID:30207593
PMID:15617989
caac.21409 PMID:29028110
4. Armstrong BK, Kricker A (1993). How
6. Gandini S, Sera F, Cattaruzza MS, Pasquini
2. Elder DE, Massi D, Scolyer RA, Willemze much melanoma is caused by sun ex-
R, editors (2018). WHO classification of P, Zanetti R, Masini C, et al. (2005). Meta-
posure? Melanoma Res. 3(6):395–401.
skin tumours. Lyon, France: International h t t p s : / / d o i . o r g / 1 0 .1 0 9 7/ 0 0 0 0 8 3 9 0 - analysis of risk factors for cutaneous mela-
Agency for Research on Cancer (WHO 199311000-00002 PMID:8161879 noma: III. Family history, actinic damage
Classification of Tumours series, 4th edi- and phenotypic factors. Eur J Cancer.
tion). Available from: http://publications. 41(14):2040–59. https://doi.org/10.1016/j.
iarc.fr/560. ejca.2005.03.034 PMID:16125929
380
7. Akbani R, Akdemir KC, Aksoy BA, Albert 16. Armstrong BK, Cust AE (2017). Sun ex- 25. Lomas A, Leonardi-Bee J, Bath-Hextall F
M, Ally A, Amin SB, et al.; Cancer Genome posure and skin cancer, and the puzzle (2012). A systematic review of worldwide
Atlas Network (2015). Genomic clas- of cutaneous melanoma: a perspective on incidence of nonmelanoma skin cancer. Br
sification of cutaneous melanoma. Cell. Fears et al. Mathematical models of age J Dermatol. 166(5):1069–80. https://doi.
161(7):1681–96. https://doi.org/10.1016/j. and ultraviolet effects on the incidence of org/10.1111/j.1365 -2133.2012.10830.x
cell.2015.05.044 PMID:26091043 skin cancer among whites in the United PMID:22251204
States. American Journal of Epidemiology
8. Green AC, Olsen CM (2015). Increased 26. Pandeya N, Olsen CM, Whiteman DC
1977; 105: 420–427. Cancer Epidemiol.
risk of melanoma in organ transplant (2017). The incidence and multiplic-
48:147–56. https://doi.org/10.1016/j.canep.
recipients: systematic review and ity rates of keratinocyte cancers in
2017.04.004 PMID:28478931
meta-analysis of cohort studies. Acta Australia. Med J Aust. 207(8):339–43.
Derm Venereol. 95(8):923–7. https:// 17. Shain AH, Bastian BC (2016). From mela- ht t ps: //do i.o r g /10. 5 6 9 4 /mja17.0 0 28 4
do i.o r g /10. 2 3 4 0 / 0 0 015 5 5 5 -214 8 nocytes to melanomas. Nat Rev Cancer. PMID:29020905
PMID:26012553 16(6):345–58. https://doi.org/10.1038/nrc.
27. Grulich AE, Vajdic CM (2015). The epidemi-
2016.37 PMID:27125352
9. Aoude LG, Wadt KA, Pritchard AL, Hayward ology of cancers in human immunodeficien-
NK (2015). Genetics of familial melanoma: 18. Arnold M, de Vries E, Whiteman DC, cy virus infection and after organ transplan-
20 years after CDKN2A. Pigment Cell Jemal A, Bray F, Parkin DM, et al. (2018). tation. Semin Oncol. 42(2):247–57. https://
Melanoma Res. 28(2):148–60. https://doi. Global burden of cutaneous melanoma doi.org/10.1053/j.seminoncol.2014.12.029
org/10.1111/pcmr.12333 PMID:25431349 attributable to ultraviolet radiation in 2012. PMID:25843729
10. Law MH, Bishop DT, Lee JE, Brossard Int J Cancer. 143(6):1305–14. https://doi.
28. Hasche D, Vinzón SE, Rösl F (2018).
M, Martin NG, Moses EK, et al.; org/10.1002/ijc.31527 PMID:29659012
Cutaneous papillomaviruses and non-mel-
GenoMEL Consortium; Essen-Heidelberg 19. Green AC, Williams GM, Logan V, anoma skin cancer: causal agents or in-
Investigators; SDH Study Group; Q-MEGA Strutton GM (2011). Reduced melanoma nocent bystanders? Front Microbiol. 9:874.
and QTWIN Investigators; AMFS after regular sunscreen use: randomized https://doi.org/10.3389/fmicb.2018.00874
Investigators; ATHENS Melanoma Study trial follow-up. J Clin Oncol. 29(3):257–63. PMID:29770129
Group (2015). Genome-wide meta-analysis
https://doi.org/10.1200/JCO.2010.28.7078 29. Chahal HS, Wu W, Ransohoff KJ, Yang L,
identifies five new susceptibility loci for cu-
PMID:21135266 Hedlin H, Desai M, et al. (2016). Genome-
taneous malignant melanoma. Nat Genet.
47(9):987–95. https://doi.org/10.1038/ng. 20. Milne E, Johnston R, Cross D, Giles-Corti wide association study identifies 14 novel
3373 PMID:26237428 B, English DR (2002). Effect of a school- risk alleles associated with basal cell
based sun-protection intervention on the carcinoma. Nat Commun. 7(1):12510.
11. Pleasance ED, Cheetham RK, Stephens ht tps://doi.org/10.1038/ncomms12510
development of melanocytic nevi in chil-
PJ, McBride DJ, Humphray SJ, Greenman PMID:27539887
dren. Am J Epidemiol. 155(8):739–45.
CD, et al. (2010). A comprehensive cata-
ht tps://doi.org/10.10 93/aje/155.8.739 30. Chahal HS, Lin Y, Ransohoff KJ, Hinds
logue of somatic mutations from a human
PMID:11943692 DA, Wu W, Dai HJ, et al. (2016). Genome-
cancer genome. Nature. 463(7278):191–
6. https://doi.org/10.1038/nature08658 wide association study identifies novel
21. Whiteman DC, Green AC, Olsen CM (2016).
PMID:20016485 susceptibility loci for cutaneous squamous
The growing burden of invasive melanoma:
cell carcinoma. Nat Commun. 7(1):12048.
12. Hodis E, Watson IR, Kryukov GV, Arold projections of incidence rates and numbers
https://doi.org/10.1038/ncomms12048
ST, Imielinski M, Theurillat JP, et al. of new cases in six susceptible popula-
PMID:27424798
(2012). A landscape of driver mutations through 2031. J Invest Dermatol.
tions in melanoma. Cell. 150(2):251–63. 136(6):1161–71. https://doi.org/10.1016/j.jid. 31. Jayaraman SS, Rayhan DJ, Hazany S,
https://doi.org/10.1016/j.cell.2012.06.024 2016.01.035 PMID:26902923 Kolodney MS (2014). Mutational land-
PMID:22817889 scape of basal cell carcinomas by whole-
22. Olsen CM, Neale RE, Green AC, Webb
exome sequencing. J Invest Dermatol.
13. Hayward NK, Wilmott JS, Waddell PM, The QSkin Study, The Epigene Study,
134(1):213–20. https://doi.org/10.1038/jid.
N, Johansson PA, Field MA, Nones Whiteman DC (2015). Independent valida-
2013.276 PMID:23774526
K, et al. (2017). Whole-genome land- tion of six melanoma risk prediction mod-
scapes of major melanoma subtypes. els. J Invest Dermatol. 135(5):1377–84. 32. South AP, Purdie KJ, Watt SA, Haldenby
Nature. 545(7653):175–80. https://doi. ht t p s: //d o i.o r g /10.10 3 8 / j i d. 2 014. 5 3 3 S, den Breems N, Dimon M, et al. (2014).
org/10.1038/nature22071 PMID:28467829 PMID:25548858 NOTCH1 mutations occur early during
cutaneous squamous cell carcinogen-
14. Shain AH, Yeh I, Kovalyshyn I, Sriharan 23. Stang A, Jöckel KH (2016). Does skin esis. J Invest Dermatol. 134(10):2630–8.
A, Talevich E, Gagnon A, et al. (2015). cancer screening save lives? A de- h t t p s: //d o i .o r g /10 .10 3 8 / j i d. 2 014.15 4
The genetic evolution of melanoma tailed analysis of mortality time trends in PMID:24662767
from precursor lesions. N Engl J Med. Schleswig-Holstein and Germany. Cancer.
CHAPTER 5.8
SECTION 5
373(20):1926–36. https://doi.org/10.1056/ 122(3):432–7. https://doi.org/10.1002/cncr. 33. Martincorena I, Roshan A, Gerstung M,
NEJMoa1502583 PMID:26559571 29755 PMID:26480048 Ellis P, Van Loo P, McLaren S, et al. (2015).
High burden and pervasive positive se-
15. Whiteman DC, Watt P, Purdie DM, Hughes 24. Martin AA, Hudgins EM, McMullan FH lection of somatic mutations in normal
MC, Hayward NK, Green AC (2003). (2010). All Bowen’s disease is squamous human skin. Science. 348(6237):880–6.
Melanocytic nevi, solar keratoses, and cell carcinoma in situ, but all squamous cell https://doi.org/10.1126/science.aaa6806
divergent pathways to cutaneous mela- carcinoma in situ is not Bowen’s disease. PMID:25999502
noma. J Natl Cancer Inst. 95(11):806–12. J Cutan Pathol. 37(11):1186–7. https://doi.
ht tps://doi.org/10.1093/jnci/95.11.80 6 org/10.1111/j.1600 - 0560.2010.01580.x
PMID:12783935 PMID:20579212

5.9
Breast cancer
Multiple, often complex, risk factors
Susan E. Hankinson Benjamin O. Anderson (reviewer)

Kornelia Polyak Valerie McCormack (reviewer)
Judy E. Garber
●● Emerging data indicate that many have been identified [1], but these
SUMMARY risk factors directly influence the have yet to be considered in epide-
numbers and/or properties of miological studies.
●● Exposures occurring in utero breast epithelial progenitors.
and until menopause can in-
fluence breast cancer risk. Epidemiology
Therefore, prevention efforts Breast cancer is the most commonly
Breast cancer is a heterogeneous
should be considered through- diagnosed cancer type and the lead-
disease, with wide variation in tu-
out a woman’s life. ing cause of cancer death in women
mour morphology, molecular char-
worldwide. In 2018, there were an
●● Some breast cancer risk fac- acteristics, and clinical response.
estimated 2.1 million new cases of
tors (e.g. mammographic den- Invasive ductal carcinoma is the
breast cancer and 627 000 deaths
sity) are similarly associated most common type of breast cancer,
from breast cancer worldwide [2].
with most currently recognized making up about 70% of tumours,
The incidence and mortality rates
breast cancer subtypes, where- and about 15–20% of tumours are
show marked international variation
as for others (e.g. parity) the invasive lobular carcinomas.
(Fig. 5.9.1 and Fig. 5.9.2). However,
relationships vary significantly Assessment of the estrogen re-
incidence and mortality data remain
by subtype; reliable estimates ceptor (ER), progesterone receptor
extremely limited for several world
of these differences have only (PR), and human epidermal growth
recently begun to emerge. regions, such as Africa.
factor receptor 2 (HER2) expres-
More than half of breast can-
●● Tumour subtypes should be sion status of tumours has been
cer cases are now diagnosed in
considered when evaluating used in clinical decision-making
low- and middle-income countries
etiology and in developing pre- for many years. Tumour molecular
[3], where a greater proportion of
vention strategies. subtypes have subsequently been
cases (and sometimes a markedly
identified, for example on the ba-
●● Breast cancer risk conferred greater proportion) are diagnosed
sis of prognostic multigene classi-
by an increasing number of at later stages, which are linked to
fiers, to derive at least the luminal
high-penetrance predisposition poorer survival (see Chapter 1.3)
A, luminal B, HER2-enriched, and
genes has been better quanti- (Fig. 5.9.3). Continuing reductions
basal-like classifications.
fied and characterized. Panels in the prevalence of infectious dis-
The importance of distinguish-
of single-nucleotide polymor- eases and associated increases in
ing between ER-positive and ER-
phisms both modify penetrance life expectancy, along with changes
negative breast cancer in epide-
of the strong susceptibility in population reproductive patterns
miological studies of etiology and
genes and confer quantifiable (e.g. later age at first birth) and life-
prevention is now established.
breast cancer risk themselves. style factors (e.g. increasing obe-
Studies linking risk factors with spe-
sity) portend an ever-increasing
●● Large population studies and cific molecular subtypes of breast burden of breast cancer in low- and
major international collabora- cancer are more recent, and sev- middle-income countries [3].
tions, particularly those integrat- eral consistent findings, noted be-
ing new technologies and basic low, have emerged. Most recently,
science discoveries, are provid- several subtypes of triple-negative Genetics and genomics
ing novel insights into breast (i.e. ER-negative, PR-negative, An inherited component to breast
cancer etiology and prevention. and HER2-negative) breast cancer cancer susceptibility has long been
382
recognized. Progress in recent years Susceptibility loci

has included the identification of Recent GWAS analyses (see Chap FUNDAMENTALS
multiple breast cancer susceptibility ter 3.2) have increased in size [10]
genes, improved estimates of their ■■ Breast cancer is the most
and have yielded multiple new sus-
penetrance, the identification of mod- commonly diagnosed cancer
ceptibility loci both for breast cancer
ifier genes, and increases in the yield type and the leading cause
overall and for specific breast cancer
of genome-wide association studies of cancer death in women
subtypes, especially triple-negative
(GWAS) for breast cancer both over- worldwide.
breast cancer [11]. A group of SNPs
all (i.e. all subtypes of breast cancer has been included in a personalized ■■ Reproductive factors, including
combined) and by subtype [4]. risk score that shows increased risk late age at menarche, early
of breast cancer in women with and age at menopause, parity,
High-penetrance gene
without a family history of breast can- early age at first birth, and
mutations
cer [12]. One cluster of SNPs has breastfeeding, all decrease
The most common high-pene- been shown to improve the perfor- risk of breast cancer overall
trance susceptibility alleles re- mance of the Tyrer–Cuzick breast (i.e. all subtypes of breast
main BRCA1 and BRCA2, both cancer risk prediction model, with the cancer combined).
of which are critical for repair of incorporation of mammographic den-
DNA double-strand breaks and ■■ Family history of breast cancer,
sity as well. These loci are entering
remodelling of stalled replication personal history of proliferative
clinical use, but most have been sub-
forks. Data and specimens from benign breast disease, dense
jected to only limited validation [13].
large cohorts of well-characterized breasts on mammogram,
germline mutation carriers, such radiation exposure, alcohol
as the Consortium of Investigators Etiology consumption, low physical
of Modifiers of BRCA1/2 (CIMBA), Several reproductive and lifestyle activity, being lean before
have permitted stable estimates of factors are confirmed contribu- menopause, postmenopausal
breast cancer risk [5]. tors to breast cancer risk. In recent obesity, recent use of
Other genes involved in DNA re- years, the understanding of the postmenopausal hormone
pair (see Chapter 3.4) were identified impact of these exposures on risk therapy (particularly estrogen
through mechanistic studies elucidat- has been improved largely through plus progestin), and recent use
ing DNA repair pathways, Fanconi assessment of these exposures of oral contraceptives are all
anaemia complementation groups, over a woman’s lifetime, accord- associated with increases in
and interacting genes associated ing to breast tumour subtype, and overall breast cancer risk.
with novel functions of known genes through detailed assessments in ■■ Inherited mutations in breast
[4,6]. The widespread adoption of large consortia. cancer predisposition genes
next-generation sequencing tech- confer increased risk of breast
nologies has led to the identification Lifestyle and environmental
cancer, often preferentially by
of germline mutations in individuals exposures
tumour subtype. Elucidation
and families without classic pheno- A notable aspect of breast cancer of the mechanisms of action
typic characteristics of a syndrome etiology is the long-term influence of these genes provides clues
or syndromes associated with spe- of exposures experienced over the to breast cancer etiology,
cific gene mutations, suggesting im- life-course. The best current exam- treatment, and prevention.
portant selection bias in early studies ple is body size (see Chapter 2.7):
(e.g. TP53, CDH1) [6,7]. birth weight is positively associated
BRCA1 and BRCA2 have been with breast cancer risk; childhood, ably causal. Multiple studies also
CHAPTER 5.9
SECTION 5
studied in the greatest detail in adolescent, and premenopaus- have assessed childhood and ado-
large collaborative cohorts (e.g. al body size are inversely related lescent body size and have noted
CIMBA), from which the available to risk; and postmenopausal body similar inverse associations [14].
data include genotype–phenotype size is positively related to risk [14]. Mechanistic understanding of these
correlations and the identification of On the basis of recent data from 19 inverse associations may offer fu-
modifier single-nucleotide polymor- prospective cohorts, the inverse as- ture targets for prevention.
phisms (SNPs) [8], although none sociation with larger adult body size A consortium analysis with more
are yet used clinically to improve in- in premenopausal women is strong than 36 000 breast cancer cases
dividual risk prediction. Examination and linear [15] and is apparent for reported that long duration of smok-
of somatic and germline mutational both ER-positive and ER-negative ing before a first pregnancy was as-
signatures (Fig. 5.9.4) may provide disease and across race and eth- sociated with a significant 18% (95%
clues to breast cancer etiology nicity [15]; furthermore, on the basis confidence interval [CI], 12–24%)
based on specific patterns of ac- of a large Mendelian randomization increase in breast cancer risk; the
quired DNA alteration [9]. study [16], the association is prob- associations were not confounded
Chapter 5.9 • Breast cancer 383

Fig. 5.9.1. Global distribution of estimated age-standardized (World) incidence rates (ASR) per 100 000 person-years for breast cancer
in women, 2018.
Fig. 5.9.2. Global distribution of estimated age-standardized (World) mortality rates (ASR) per 100 000 person-years for breast cancer in
women, 2018.
by current alcohol consumption and breast cancer risk [14], particularly ited the conclusions. Increasingly,
were observed predominantly for for ER-negative disease; similarly, efforts have focused on evaluating
ER-positive tumours [17]. These data several studies have observed exposure during windows of sus-
support a causal link of smoking with an inverse association between a ceptibility, by assessing links be-
breast cancer risk and re-emphasize Mediterranean diet score and ER- tween contaminants and intermedi-
the importance of smoking prevention negative breast cancer [18]. ate markers of risk such as breast
and cessation programmes in adoles- The potential role of environ- density, and by increasing transdis-
cents and young adults (see “Tobacco mental and occupational exposures ciplinary research efforts. Such ef-
cessation: the WHO perspective”). in breast carcinogenesis has re- forts are providing new insights into
Studies suggest that caroten- mained a major interest, although the potential for exposures such as
oids, or other constituents in ca- challenges in exposure assess- endocrine disrupters to influence
rotenoid-rich foods, may decrease ment and study design have lim- breast cancer risk [19].
384
Fig. 5.9.3. Percentage of breast cancer cases diagnosed at a late stage (stages III and IV combined), by country or region and by
time period or population group.
Reproductive factors breast cancer. Women who breast- observed more recently include di-
The inverse association observed fed versus never breastfed had etary factors [14,18]. Furthermore,
between parity and risk of breast lower risk of ER-negative and triple- associations of parity and breast-
negative disease (Fig. 5.9.5) [22]. feeding with risk appear to vary by
cancer overall is consistently seen
Importantly from a prevention per- molecular subtype [20,22,23]. Such
for ER-positive disease, whereas no
spective, breastfeeding appears to analyses require both large sample
association or a positive association
reduce risk of these breast cancer sizes and the availability of tumour
has been observed for ER-negative
subtypes that have poorer progno- tissue; hence, reliable estimates of
and triple-negative disease [20]. In
sis. (For a discussion of reproduc- these differences have only recent-
addition, breastfeeding has been
tive factors such as age at menar- ly begun to emerge.
associated with lower risk of hor-
che, age at first birth, and age at
CHAPTER 5.9
SECTION 5
mone receptor-negative (including menopause, see Chapter 3.6.) Population attributable risks
ER-negative, triple-negative, and Several recent efforts have evalu-
basal-like) breast cancer; weaker Breast tumour subtypes ated the population attributable
and less consistent associations Studies have increasingly focused risks for breast cancer. In a study
have been observed for ER-positive on evaluating risk factors by mo- that combined data from two large
tumour subtypes [21]. These stud- lecular characteristics of breast cohorts and assessed a range of
ies have been conducted largely in tumours, to provide causal insight well-established breast cancer risk
populations of European ancestry. for observed associations and to factors in relation to breast cancer in
Recently, across four studies of better inform prevention strate- postmenopausal women, the popu-
African American women, parity gies. The differential associations lation attributable risk was 70.0%
was observed to significantly in- of postmenopausal obesity and (95% CI, 55.0–80.7%) overall [24].
crease risk of ER-negative and tri- use of hormone therapy with ER- For modifiable risk factors only,
ple-negative breast cancer, and to positive but not ER-negative breast the population attributable risk was
modestly lower risk of ER-positive cancer are established; differences 34.6% overall and was higher for

Fig. 5.9.4. Pathway enrichment map for susceptibility loci based on summary association statistics for 65 new breast cancer loci.
Each coloured circle (node) represents a pathway (gene set), coloured by enrichment score, where redder nodes indicate lower
false discovery rates. Larger nodes indicate pathways with more genes. Green lines connect pathways with overlapping genes
(minimum overlap, 0.55). Pathways are grouped by similarity and organized into major themes (large labelled circles).
Fig. 5.9.5. Relative risks (with 95% confidence intervals) for number of births in relation to estrogen receptor (ER) status, according
to history of breastfeeding, from the African American Breast Cancer Epidemiology and Risk (AMBER) Consortium. The reference
for both ER-positive (ER+) and ER-negative (ER-) analyses is women who had only one birth and had breastfed.
Never breastfed, ER-
Ever breastfed, ER-

2.00
Never breastfed, ER+
Ever breastfed, ER+

Odds ratio
1.00
0.50
1 2 3 4+
Number of births
386
Clues to the biology of breast cancer risk
Breast epithelial stem cells and as a marker of quiescent cells with (COX-2) also upregulate the ex-
progenitors are the cells of origin proliferative potential. pression of aromatase, an enzyme
of breast carcinomas; therefore, A follow-up study analysed the that is key for estrogen biosynthe-
cancer risk factors are expected frequencies of cells with expres- sis, resulting in higher local estro-
to affect the numbers and/or prop- sion of the proliferative marker Ki- gen levels. The presence of crown-
erties of these cells [1]. Despite 67 and the quiescent marker p27 like structures was associated with
the importance of this issue, the in normal breast biopsies of wom- poor clinical outcome independent
knowledge of cancer risk-associat- en in the Nurses’ Health Study. of body mass index and in all breast
ed differences in the normal breast Premenopausal women with high cancer subtypes, suggesting that
is rather limited. Among the best- Ki-67-positive and low p27-positive inflammation is a general inducer
understood risk factors are early cell frequencies had a 5-fold high- of cancer risk. In addition to local
full-term pregnancy and obesity. A er risk of breast cancer compared effects, obesity also increases cir-
full-term pregnancy in young adult- with women with low Ki-67-positive culating levels of leptin and IL-6,
hood (age < 20 years) decreases and low p27-positive cell frequen- which can promote tumour initia-
the risk of estrogen receptor (ER)- cies. These results suggest that tion via direct effects on the mam-
positive postmenopausal breast the higher number of cycling cells mary epithelial cells and by chang-
cancer. In contrast, the risk of in the normal mammary epithelium ing the microenvironment.
ER-negative breast tumours is not increases the probability of muta-
decreased by pregnancy, and mul- tions; thus, the fraction of these References
tiple early pregnancies, coupled cells may be a biomarker of breast
1. Visvader JE, Stingl J (2014). Mammary
with lack of breastfeeding, is one cancer risk [3].
stem cells and the differentiation hi-
of the most significant risk factors One potential mechanism by erarchy: current status and perspec-
for triple-negative breast cancer. which obesity influences breast tives. Genes Dev. 28(11):1143–58.
Comprehensive comparative cancer risk is via alterations in the https://doi.org/10.1101/gad.242511.114
analysis of normal human breast local and systemic microenviron- PMID:24888586
tissues from nulliparous and par- ments [4]. Obesity is associated 2. Meier-Abt F, Bentires-Alj M, Rochlitz C
ous women, including BRCA1 and with inflammation in white adipose (2015). Breast cancer prevention: les-
BRCA2 germline mutation car- tissue, which is characterized by sons to be learned from mechanisms of
early pregnancy-mediated breast can-
riers, determined that the most crown-like structures formed by
cer protection. Cancer Res. 75(5):803–7.
significant gene expression and macrophages surrounding dead or https://doi.org/10.1158/0008-5472.CAN-
epigenetic changes occur in line- dying adipocytes. Such structures 14-2717 PMID:25660950
age-negative progenitor-enriched lead to the release of free fatty
3. Atashgaran V, Wrin J, Barry SC, Dasari
cells and that the numbers of these acids that trigger Toll-like recep- P, Ingman WV (2016). Dissecting the
cells are higher in nulliparous tor signalling and nuclear factor biology of menstrual cycle-associated
women and even higher in BRCA1 kappa-light-chain-enhancer of ac- breast cancer risk. Front Oncol. 6:267.
and BRCA2 mutation carriers tivated B cells (NF-κB)-mediated https://doi.org/10.3389/fonc.2016.00267
PMID:28083513
[2]. Transforming growth factor upregulation of pro-inflammatory
β (TGF-β), WNT, and insulin-like cytokines (e.g. tumour necrosis 4. Iyengar NM, Gucalp A, Dannenberg
growth factor 1 (IGF-1) signalling factor α [TNF-α] and interleukin-1β AJ, Hudis CA (2016). Obesity and can-
cer mechanisms: tumor microenviron-
were identified as candidate regu- [IL-1β]). Besides creating a pro- ment and inflammation. J Clin Oncol.
lators of hormone-responsive pro- inflammatory environment, these 34(35):4270–6. https://doi.org/10.1200/
genitors, and p27 was identified cytokines and cyclooxygenase 2 JCO.2016.67.4283 PMID:27903155
CHAPTER 5.9
ER-positive tumours (39.7%) than for breast cancer, and personal history or urine, and risk of breast cancer. SECTION 5
ER-negative tumours (27.9%) [24]. of benign breast disease; the great- Postmenopausal levels of estradiol
The Breast Cancer Surveillance est contributors to these estimates and testosterone are established
Consortium reported that 52.7% and were BMI and breast density [25]. risk factors, with relative risks of
54.7% of breast cancers in premen- breast cancer of 1.5–3.0 when com-
opausal and postmenopausal wom- Biological characteristics paring women in the top versus the
en, respectively, could potentially be bottom 20–25% of hormone levels.
attributed to six risk factors: Breast Endogenous hormones Data in premenopausal women
Imaging Reporting and Data System Substantial progress has been made have been sparse, largely because
(BI-RADS) breast density, parity, to further the understanding of the of complexities in measuring estro-
age at first birth, body mass index link between endogenous hormone gen levels during the menstrual cy-
(BMI), first-degree family history of concentrations, measured in blood cle. In a recent pooled analysis of

Fig. 5.9.6. Relative risks (with 95% confidence intervals) of premenopausal breast prospective studies with 767 cases
cancer by quantile of circulating hormone concentrations. and 1699 controls, a modest but
significant positive association was
noted for estradiol and testosterone
levels in premenopausal women,
with comparable relative risks of
1.41 (P trend = 0.01) for estradiol and
1.32 (P trend = 0.02) for testosterone
(Fig. 5.9.6); no association was
observed for plasma progesterone
levels [26]. A positive association
between prolactin levels and risk
of breast cancer, primarily in post-
menopausal women, also has in-
creasingly been documented [27].
Estrogen metabolites have been
hypothesized to independently influ-
ence risk via effects on proliferation or
by inducing oxidative damage. With
an improved assay technology [28]
Mammographic density
Mammographic density represents Increasing efforts have focused on neural networks or autoencoders,
the relative amounts of dense (epi- delineating the biological mecha- for mammogram-based breast
thelial and stromal) tissue versus nisms underlying mammographic cancer risk assessment has also
non-dense (adipose) tissue in the density and its strong association been promising [1]. Although these
breast as seen on mammogram. with breast carcinogenesis. Recent new approaches have shown
Mammographic density varies findings suggest roles for multiple great potential, there is currently a
widely between women. Both quali- factors, including those that influ- lack of evaluations of multiple ap-
tative measures (e.g. the four-cate- ence the composition and stiffness proaches simultaneously in large
gory Breast Imaging Reporting and of the extracellular matrix, and and diverse populations to deter-
Data System [BI-RADS]) and quan- genes associated with increased mine the optimal combination of
titative measures (e.g. quantitative cellular proliferation, although fur- tissue features and the strength
thresholding using the Cumulus ther work is needed. of their association with future risk
software) of mammographic den- In the past few years, several across different tumour subtypes.
sity are strongly predictive of breast novel approaches have begun to
cancer risk, with relative risks of be evaluated to fully automate as- References
4–6 (when comparing women with sessments of mammographic den-
1. Gastounioti A, Conant EF, Kontos D
high versus low percentage density (e.g. using Volpara, a program (2016). Beyond breast density: a review
sity) that do not vary by tumour es- that provides an automated volu- on the advancing role of parenchymal
trogen receptor status [1,2]. metric measure of density) and to texture analysis in breast cancer risk as-
Recent large genome-wide as- assess additional parenchymal sessment. Breast Cancer Res. 18(1):91.
https://doi.org/10.1186/s13058 - 016 -
sociation studies (GWAS) have textural features (e.g. run-length
0755-8 PMID:27645219
pointed to a shared genetic basis and structural features) that may
between mammographic density better characterize tissue com- 2. Sherratt MJ, McConnell JC, Streuli CH
(2016). Raised mammographic density:
and breast cancer [2]. Newer studies plexity on mammogram. These
causative mechanisms and biological
also have suggested that assess- studies have generally shown simi- consequences. Breast Cancer Res.
ment of mammographic density can lar or stronger associations with 18(1):45. https://doi.org/10.1186/s13058-
add substantially to current breast breast cancer risk, indicating that, 016-0701-9 PMID:27142210
cancer risk prediction models, and relative to traditional density as- 3. Shawky MS, Martin H, Hugo HJ, Lloyd
that change in mammographic den- sessment, some of the new mea- T, Britt KL, Redfern A, et al. (2017).
sity (e.g. > 10% decrease in density) sures are likely to substantially Mammographic density: a potential mon-
can be used as a surrogate marker improve upon or add independent itoring biomarker for adjuvant and pre-
ventative breast cancer endocrine thera-
for breast cancer to indicate who will new information in risk prediction pies. Oncotarget. 8(3):5578–91. https://
most benefit from chemopreven- [1]. Emerging research with a deep doi.org /10.18 6 32 /onc ot arget.13 4 8 4
tion or other prevention efforts [3]. learning approach, using either PMID:27894075
388
used across five studies in postmen- Risk stratification Given the increasing proportion
opausal women, a relative increase Breast cancer risk prediction models of breast cancer cases in low- and
in levels of 2-hydroxylation pathway have been developed to estimate middle-income countries, as well
metabolites versus 16-hydroxylation the risk of carrying a high-risk germ- as the changing patterns of risk fac-
pathway metabolites was associ- line mutation, the risk of developing tors in these countries, it is critical
ated with a 34% decrease (95% CI, breast cancer, or both [32]. Until re- to identify feasible strategies to im-
16–48%) in breast cancer risk inde- cently, existing models, such as the prove prevention and early detec-
pendent of total estrogen levels [29]. Breast Cancer Risk Assessment tion in these settings.
Data in premenopausal women are Tool (also known as the Gail mod-
limited but are suggestive of similar Racial and ethnic variations
el) and the Rosner–Colditz model,
associations [28]. generally included reproductive fac- Racial differences in breast cancer
Anti-Müllerian hormone is pro- tors, family history of breast cancer, incidence and mortality exist, and it
duced by the ovaries, is measurable and a subset of lifestyle factors. has become increasingly clear that
only before menopause, reflects the Recent work has suggested signifi- differences in the distribution of
size of the ovarian follicular pool, cant improvements in model perfor- both individual risk factors and soci-
and is strongly correlated with age at mance with the addition of several etal and contextual factors, as well
menopause [30]. In a large consor- biological markers, including mam- as tumour biology, all contribute to
tium analysis of 10 prospective stud- mographic breast density, genetic this variation.
ies, a significant positive association risk scores, and plasma endoge- For example, from the United
was observed, with a multivariable nous hormone levels (e.g. [12,33]). States National Cancer Institute’s
relative risk comparing the top ver- Further enhancements are needed, Surveillance, Epidemiology, and
sus the bottom quartile categories End Results (SEER) programme, the
including incorporation of newly
of 1.60 (95% CI, 1.31–1.94; Ptrend 2007–2011 age-adjusted incidence
confirmed risk factors (e.g. anti-
< 0.001) (Fig. 5.9.6) [30]. The find- rate (per 100 000) for breast can-
Müllerian hormone), more specific
ings were unchanged after account- cer was 128 for non-Hispanic White
disease definitions, and develop-
ing for testosterone concentrations, women and 123 for African American
ment and validation in a wider range
were similar regardless of meno- women, but the age-adjusted mor-
of study populations. Other priorities
pausal status at diagnosis, and were tality rate (per 100 000) was 21.7 for
are assessment of clinical utility and
observed primarily for ER-positive non-Hispanic White women and 30.6
strategies to successfully implement
tumours. Anti-Müllerian hormone is for African American women. African
these models in clinical practice.
one of the few hormones assessed American women have a higher
The current Women Informed to
in premenopausal women that is prevalence of triple-negative breast
Screen Depending on Measures of
now confirmed to predict later risk cancers [35], for which outcomes are
Risk (WISDOM) clinical trial exam-
of breast cancer. Additional facets of poorer, and this is a likely contributor
ining risk-stratified mammographic
this association, as well as the bio- to the higher SEER mortality rates.
screening, and the work by the group
logical mechanisms underlying the However, even among the subset of
in Manchester, United Kingdom, in-
association, require further study. women diagnosed with similar early-
corporating risk SNPs and mammo-
graphic density into the Tyrer–Cuzick stage disease, mortality rates were
Novel technologies
multivariable model, among others, higher for African American women,
New analytical technologies such indicating that other factors, such as
will provide data with which to assess
as metabolomics and proteomics differences in patterns of care [35],
the impact of these approaches.
(see Chapter 3.7) can be used in contribute as well (see “The enduring
population-based studies and are disparity in breast cancer mortality
beginning to provide new insights Social inequalities in risk between Black and White women in
CHAPTER 5.9
SECTION 5
into the biological mechanisms un- and burden the USA” in Chapter 4.6).
derlying known breast cancer risk
factors, as well as offering the po- Socioeconomic differences
tential to identify new biomarkers of In epidemiological studies, a positive Prevention
risk or early detection. For example, association between socioeconomic Prevention trials require large study
several diet-related metabolites (re- status and breast cancer risk is well populations and long follow-up
lated to alcohol, vitamin E, and ani- established. This is due in large part periods, which makes them both
mal fat) were associated with risk to different distributions by socioeco- costly and challenging to conduct.
of breast cancer, particularly for nomic status of breast cancer risk Therefore, preliminary data for pre-
ER-positive disease, thus suggest- factors such as parity, age at first vention trials often come from bio-
ing additional factors that may play birth, and use of hormone therapy. marker modulation studies, or from
a mechanistic role underlying these Other possible contributors include evaluation of the effects of interven-
dietary exposures and modulation differences in screening practices tions on contralateral breast cancer
of risk [31]. across socioeconomic status [34]. events in breast cancer treatment

trials. Colditz and Bohlke recently tion to prevent breast cancer recur- marker of breast tissue proliferation,
reviewed the evidence that acting rence [37]. If BWEL is successful, in women with insulin resistance
on already established information weight loss would probably be fur- [38], but in a meta-analysis on met-
about modifiable risk factors could ther targeted in a trial for breast formin and cancer risk, after adjust-
substantially reduce breast cancer cancer risk reduction. ment for BMI, no significant reduc-
incidence in high-income countries tion in breast cancer incidence was
(Table 5.9.1) [36]. Metformin observed [38].
Metformin, which is used for treat-
Weight loss ment of metabolic syndrome and Familial or other high-risk
There have not been compelling diabetes, has been linked with lower groups
new data for weight loss, but Breast risk of breast cancer in observa- Other medical interventions gener-
Cancer Weight Loss (BWEL) is a tional studies. In a pre-surgical trial ally target women who have sub-
current randomized trial addressing in Italy, metformin taken before sur- stantial risk of breast cancer. The
the ability of a weight-loss interven- gery decreased levels of Ki-67, a duration of the effects of selective
Table 5.9.1. Current strategies to prevent breast cancer
Health message Risk group Estimated proportion Possible Time until

of female population in reduction benefit
the USA aged < 50 years in risk (%) b (years)
affected (%)a
Premenopausal women
Alcohol intake: none Youth (aged 12–17 years), drinking 13 20–30 10–20

≥ 1 drink in the past 30 days
Alcohol intake: none or Young adults (aged 18–24 years) 15 20–30 10–20

≤ 4 servings/day drinking ≥ 4 drinks/week
Adults (aged ≥ 18 years) drinking 13 35 10–20

≥ 4 drinks/week
Healthy weight: avoid weight gain All women 100 50 (after 10–30
menopause)
Physical activity: ≥ 30 minutes/day Women not meeting physical activity 54 20 10–30

guidelines
Healthy diet: fruits, vegetables, Youth eating few fruits and 5–11 20–50 5–20
and whole grains vegetables
Breastfeed: 1 year total across all Women who have given birth 81 18 5
children
Prophylactic bilateral oophorectomy BRCA1 and BRCA2 mutation carriers < 1 50 ≥ 2
Tamoxifen High-risk women aged ≥ 35 years 3 50 2

(≥ the risk for an average woman
aged 60 years)
Postmenopausal women
Alcohol intake: none or Women drinking ≥ 4 drinks/week 13 35 5–10

< 1 serving/day
Healthy weight: weight loss Overweight and obese women 64 50 2–5
Physical activity: ≥ 30 minutes/day Women not meeting physical activity 54 20 10–20

guidelines
Estrogen plus progestin postmeno- Current users 1.7 10 1

pausal hormone therapy: avoid
Long-term current users 1 50 2
Tamoxifen and raloxifene c

High-risk women (≥ the risk for an 30 50 2
average woman aged 60 years)
a
Estimates are from nationally representative samples of women in the USA.
b
Risk factors in the table are not necessarily biologically independent of each other.
c
Exemestane is not listed for prevention, because the United States Food and Drug Administration has not approved this agent for primary breast cancer
risk reduction.
390
ER modulators, such as tamox- 0.91) (Fig. 5.9.7) [40]. The effect was and of ovarian cancer. The timing
ifen and raloxifene, on breast can- observed for both ER-positive breast and advisability may be considered
cer prevention was estimated in a cancer and ductal carcinoma in situ, in a framework put forward by Tung
meta-analysis, which demonstrated but not for triple-negative breast can- et al. (see Chapter 6.5) [42].
a measurable reduction in breast cer. Aromatase inhibitors, both anas- Recent data indicating that
cancer incidence that was greatest trozole and exemestane, have been RANK ligand is an essential mol-
in the first 5 years of follow-up but shown to reduce breast cancer risk ecule in the development of breast
also extended into years 5–10 of fol- by about half [41]. There is a lack of cancer in BRCA1 mutation carriers
low–up [39]. proven strategies for reducing the have led to an international che-
In the follow-up of the Internation risk of HER2-positive and triple-neg- moprevention trial evaluating the
al Breast Cancer Intervention Study ative breast cancers. RANK ligand inhibitor denosumab
(IBIS) trial (tamoxifen vs placebo), the The management of women at in BRCA1 mutation carriers, led by
hazard ratio for the occurrence of all high risk based on predisposing the Austrian Breast and Colorectal
breast cancers in the tamoxifen group mutations in cancer susceptibility Cancer Study Group (ABCSG). The
versus the placebo group in the first genes includes risk-reducing mas- next phase of trials will focus on
10 years of follow-up was 0.72 (95% tectomies and premenopausal oo- bringing progress in cancer immu-
CI, 0.59–0.88) and after 10 years of phorectomies, which may reduce nology to prevention.
follow-up was 0.69 (95% CI, 0.53– risk of ER-positive breast cancer
Fig. 5.9.7. Cumulative incidence of breast cancer over time in the International Breast Cancer Intervention Study I (IBIS-I) trial,
according to treatment group (tamoxifen or placebo) and duration of follow-up. Solid lines indicate all breast cancers, and dashed
lines indicate invasive estrogen receptor (ER)-positive breast cancers.
CHAPTER 5.9
SECTION 5

References
1. Lehmann BD, Jovanović B, Chen X, Estrada 10. Michailidou K, Beesley J, Lindstrom S, 18. Du M, Liu SH, Mitchell C, Fung TT (2018).
MV, Johnson KN, Shyr Y, et al. (2016). Canisius S, Dennis J, Lush MJ, et al.; Associations between diet quality scores
Refinement of triple-negative breast cancer BOCS; kConFab Investigators; AOCS and risk of postmenopausal estrogen re-
molecular subtypes: implications for neoad- Group; NBCS; GENICA Network (2015). ceptor-negative breast cancer: a systemat-
juvant chemotherapy selection. PLoS One. Genome-wide association analysis of ic review. J Nutr. 148(1):100–8. https://doi.
11(6):e0157368. https://doi.org/10.1371/ more than 120,000 individuals identifies org/10.1093/jn/nxx015 PMID:29378048
journal.pone.0157368 PMID:27310713 15 new susceptibility loci for breast can-
cer. Nat Genet. 47(4):373–80. https://doi. 19. Forman MR, Winn DM, Collman GW, Rizzo
2. Bray F, Ferlay J, Soerjomataram I, Siegel org/10.1038/ng.3242 PMID:25751625 J, Birnbaum LS (2015). Environmental
RL, Torre LA, Jemal A (2018). Global exposures, breast development and can-
cancer statistics 2018: GLOBOCAN es- 11. Milne RL, Kuchenbaecker KB, Michailidou cer risk: through the looking glass of
timates of incidence and mortality world- K, Beesley J, Kar S, Lindström S, et al.; breast cancer prevention. Reprod Toxicol.
wide for 36 cancers in 185 countries. CA ABCTB Investigators; EMBRACE; GEMO 54:6–10. https://doi.org/10.1016/j.reprotox.
Cancer J Clin. 68(6):394–424. https://doi. Study Collaborators; HEBON; kConFab/ 2014.10.019 PMID:25499721
org/10.3322/caac.21492 PMID:30207593 AOCS Investigators; NBSC Collaborators
(2017). Identification of ten variants as- 20. Anderson KN, Schwab RB, Martinez ME
3. Ginsburg O, Bray F, Coleman MP, sociated with risk of estrogen-receptor- (2014). Reproductive risk factors and breast
Vanderpuye V, Eniu A, Kotha SR, et al. negative breast cancer. Nat Genet. cancer subtypes: a review of the literature.
(2017). The global burden of women’s can- 49(12):1767–78. https://doi.org/10.1038/ Breast Cancer Res Treat. 144(1):1–10.
cers: a grand challenge in global health. ng.3785 PMID:29058716 https://doi.org/10.1007/s10549-014-2852-7
Lancet. 389(10071):847–60. https://doi. PMID:24477977
org /10.1016 /S 014 0 - 673 6 (16) 313 92-7 12. Mavaddat N, Pharoah PD, Michailidou K,
PMID:27814965 Tyrer J, Brook MN, Bolla MK, et al. (2015). 21. Islami F, Liu Y, Jemal A, Zhou J, Weiderpass
Prediction of breast cancer risk based E, Colditz G, et al. (2015). Breastfeeding
4. Nielsen FC, van Overeem Hansen T, on profiling with common genetic vari- and breast cancer risk by receptor sta-
Sørensen CS (2016). Hereditary breast ants. J Natl Cancer Inst. 107(5):djv036. tus – a systematic review and meta-
and ovarian cancer: new genes in confined h t t p s : / / d o i . o r g / 10 .10 9 3 / j n c i / d j v 0 3 6 analysis. Ann Oncol. 26(12):2398–407.
pathways. Nat Rev Cancer. 16(9):599– PMID:25855707 https://doi.org/10.1093/annonc/mdv379
612. https://doi.org/10.1038/nrc.2016.72 PMID:26504151
PMID:27515922 13. Evans DG, Brentnall A, Byers H,
Harkness E, Stavrinos P, Howell A, et 22. Palmer JR, Viscidi E, Troester MA, Hong
5. Kuchenbaecker KB, Hopper JL, Barnes al.; FH-Risk Study Group (2017). The CC, Schedin P, Bethea TN, et al. (2014).
DR, Phillips KA, Mooij TM, Roos-Blom impact of a panel of 18 SNPs on breast Parity, lactation, and breast cancer sub-
MJ, et al.; BRCA1 and BRCA2 Cohort cancer risk in women attending a UK types in African American women: re-
Consortium (2017). Risks of breast, ovarian, familial screening clinic: a case-control sults from the AMBER Consortium. J Natl
and contralateral breast cancer for BRCA1 study. J Med Genet. 54(2):111–3. https:// Cancer Inst. 106(10):dju237. https://doi.
and BRCA2 mutation carriers. JAMA. doi.org/10.1136/jmedgenet-2016-104125 org/10.1093/jnci/dju237 PMID:25224496
317(23):2402–16. https://doi.org/10.1001/ PMID:27794048
jama.2017.7112 PMID:28632866 23. Brouckaert O, Rudolph A, Laenen A,
14. WCRF/AICR (2018). Diet, nutrition, physi- Keeman R, Bolla MK, Wang Q, et al.;
6. Easton DF, Pharoah PDP, Antoniou AC, cal activity and cancer: a global perspec- kConFab (2017). Reproductive profiles
Tischkowitz M, Tavtigian SV, Nathanson tive. Continuous Update Project Expert and risk of breast cancer subtypes: a mul-
KL, et al. (2015). Gene-panel sequenc- Report 2018. World Cancer Research ti-center case-only study. Breast Cancer
ing and the prediction of breast-cancer Fund/American Institute for Cancer Res. 19(1):119. https://doi.org/10.1186/
risk. N Engl J Med. 372(23):2243–57. Research. Available from: https://www. s13058-017-0909-3 PMID:29116004
https://doi.org/10.1056/NEJMsr1501341 wcrf.org/dietandcancer.
PMID:26014596 24. Tamimi RM, Spiegelman D, Smith-Warner
15. Schoemaker MJ, Nichols HB, Wright SA, Wang M, Pazaris M, Willett WC, et
7. Rana HQ, Gelman R, LaDuca H, LB, Brook MN, Jones ME, O’Brien KM, al. (2016). Population attributable risk of
McFarland R, Dalton E, Thompson J, et al.; Premenopausal Breast Cancer modifiable and nonmodifiable breast can-
et al. (2018). Differences in TP53 muta- Collaborative Group (2018). Association cer risk factors in postmenopausal breast
tion carrier phenotypes emerge from of body mass index and age with subse- cancer. Am J Epidemiol. 184(12):884–
panel-based testing. J Natl Cancer Inst. quent breast cancer risk in premenopaus- 93. https://doi.org/10.1093/aje/kww145
110(8):863–70. https://doi.org/10.1093/ al women. JAMA Oncol. 4(11):e181771. PMID:27923781
jnci/djy001 PMID:29529297 h t t p s : / / d o i . o r g / 1 0 .1 0 0 1/ j a m a o n c o l .
2018.1771 PMID:29931120 25. Engmann NJ, Golmakani MK, Miglioretti
8. Rebbeck TR, Mitra N, Wan F, Sinilnikova DL, Sprague BL, Kerlikowske K; Breast
OM, Healey S, McGuffog L, et al.; CIMBA 16. Guo Y, Warren Andersen S, Shu XO, Cancer Surveillance Consortium (2017).
Consortium (2015). Association of type Michailidou K, Bolla MK, Wang Q, et al. Population-attributable risk proportion
and location of BRCA1 and BRCA2 (2016). Genetically predicted body mass of clinical risk factors for breast can-
mutations with risk of breast and ovar- index and breast cancer risk: Mendelian cer. JAMA Oncol. 3(9):1228–36. https://
ian cancer. JAMA. 313(13):1347–61. randomization analyses of data from doi.org/10.1001/jamaoncol.2016.6326
https://doi.org/10.1001/jama.2014.5985 145,000 women of European descent. PMID:28152151
PMID:25849179 PLoS Med. 13(8):e1002105. https://
doi.org/10.1371/journal.pmed.1002105 26. Key TJ, Appleby PN, Reeves GK, Travis RC,
9. Alexandrov LB, Nik-Zainal S, Wedge DC, PMID:27551723 Alberg AJ, Barricarte A, et al.; Endogenous
Aparicio SA, Behjati S, Biankin AV, et al.; Hormones and Breast Cancer Collaborative
Australian Pancreatic Cancer Genome 17. Gaudet MM, Carter BD, Brinton LA, Falk Group (2013). Sex hormones and risk of
Initiative; ICGC Breast Cancer Consortium; RT, Gram IT, Luo J, et al. (2017). Pooled breast cancer in premenopausal women:
ICGC MMML-Seq Consortium; ICGC analysis of active cigarette smoking and a collaborative reanalysis of individual par-
PedBrain (2013). Signatures of mutation- invasive breast cancer risk in 14 cohort ticipant data from seven prospective stud-
al processes in human cancer. Nature. studies. Int J Epidemiol. 46(3):881–93. ies. Lancet Oncol. 14(10):1009–19. https://
500(7463):415–21. https://doi.org/10.1038/ h t t p s : / / d o i . o r g / 1 0 .1 0 9 3 / i j e / d y w 2 8 8 doi.org/10.1016/S1470-2045(13)70301-2
nature12477 PMID:23945592 PMID:28031315 PMID:23890780
392
27. Tworoger SS, Eliassen AH, Zhang X, Qian 33. Zhang X, Rice M, Tworoger SS, Rosner 39. Cuzick J, Sestak I, Bonanni B, Costantino
J, Sluss PM, Rosner BA, et al. (2013). A BA, Eliassen AH, Tamimi RM, et al. JP, Cummings S, DeCensi A, et al.;
20-year prospective study of plasma pro- (2018). Addition of a polygenic risk score, SERM Chemoprevention of Breast
lactin as a risk marker of breast cancer mammographic density, and endogenous Cancer Overview Group (2013). Selective
development. Cancer Res. 73(15):4810–9. hormones to existing breast cancer risk oestrogen receptor modulators in preven-
https://doi.org/10.1158/0008-5472.CAN- prediction models: a nested case-control tion of breast cancer: an updated meta-
13-0665 PMID:23783576 study. PLoS Med. 15(9):e1002644. https:// analysis of individual participant data.
doi.org/10.1371/journal.pmed.1002644 Lancet. 381(9880):1827–34. https://doi.
28. Ziegler RG, Fuhrman BJ, Moore SC, PMID:30180161
Matthews CE (2015). Epidemiologic stud- org /10.1016/S 0140 - 673 6(13) 6 0140 - 3
ies of estrogen metabolism and breast 34. Verdial FC, Etzioni R, Duggan C, PMID:23639488
cancer. Steroids. 99(Pt A):67–75. https:// Anderson BO (2017). Demographic 40. Cuzick J, Sestak I, Cawthorn S, Hamed H,
doi.org/10.1016/j.steroids.2015.02.015 changes in breast cancer incidence, stage
Holli K, Howell A, et al.; IBIS-I Investigators
PMID:25725255 at diagnosis and age associated with pop-
(2015). Tamoxifen for prevention of breast
ulation-based mammographic screening.
29. Sampson JN, Falk RT, Schairer C, Moore cancer: extended long-term follow-up
J Surg Oncol. 115(5):517–22. https://doi.
SC, Fuhrman BJ, Dallal CM, et al. (2017). of the IBIS-I breast cancer prevention
org/10.1002/jso.24579 PMID:28194807
Association of estrogen metabolism with trial. Lancet Oncol. 16(1):67–75. https://
breast cancer risk in different cohorts of 35. Daly B, Olopade OI (2015). A perfect storm: doi.org/10.1016/S1470-2045(14)71171-4
postmenopausal women. Cancer Res. how tumor biology, genomics, and health PMID:25497694
77(4):918–25. https://doi.org/10.1158/0008- care delivery patterns collide to create a
5472.CAN-16-1717 PMID:28011624 racial survival disparity in breast cancer 41. Cuzick J, Sestak I, Forbes JF, Dowsett M,
and proposed interventions for change. CA Knox J, Cawthorn S, et al.; IBIS-II investi-
30. Ge W, Clendenen TV, Afanasyeva Y, Cancer J Clin. 65(3):221–38. https://doi. gators (2014). Anastrozole for prevention of
Koenig KL, Agnoli C, Brinton LA, et al. org/10.3322/caac.21271 PMID:25960198 breast cancer in high-risk postmenopausal
(2018). Circulating anti-Müllerian hor- women (IBIS-II): an international, double-
mone and breast cancer risk: a study in 36. Colditz GA, Bohlke K (2014). Priorities for blind, randomised placebo-controlled
ten prospective cohorts. Int J Cancer. the primary prevention of breast cancer. CA trial. Lancet. 383(9922):1041–8. https://
142(11):2215–26. https://doi.org/10.1002/ Cancer J Clin. 64(3):186–94. https://doi. doi.org/10.1016/S0140-6736(13)62292-8
ijc.31249 PMID:29315564 org/10.3322/caac.21225 PMID:24647877 PMID:24333009
31. Playdon MC, Ziegler RG, Sampson JN, 37. Demark-Wahnefried W, Schmitz KH, 42. Tung N, Domchek SM, Stadler Z,
Stolzenberg-Solomon R, Thompson Alfano CM, Bail JR, Goodwin PJ,
Nathanson KL, Couch F, Garber JE, et al.
HJ, Irwin ML, et al. (2017). Nutritional Thomson CA, et al. (2018). Weight man-
(2016). Counselling framework for moder-
metabolomics and breast cancer risk agement and physical activity through-
ate-penetrance cancer-susceptibility mu-
in a prospective study. Am J Clin Nutr. out the cancer care continuum. CA
tations. Nat Rev Clin Oncol. 13(9):581–8.
106(2):637–49. https://doi.org/10.3945/ Cancer J Clin. 68(1):64–89. https://doi.
ajcn.116.150912 PMID:28659298 org/10.3322/caac.21441 PMID:29165798 https://doi.org/10.1038/nrclinonc.2016.90
PMID:27296296
32. Cintolo-Gonzalez JA, Braun D, Blackford 38. Gandini S, Puntoni M, Heckman-Stoddard
AL, Mazzola E, Acar A, Plichta JK, et BM, Dunn BK, Ford L, DeCensi A, et al.
al. (2017). Breast cancer risk models: a (2014). Metformin and cancer risk and
comprehensive overview of existing mod- mortality: a systematic review and meta-
els, validation, and clinical applications. analysis taking into account biases and
Breast Cancer Res Treat. 164(2):263–84. confounders. Cancer Prev Res (Phila).
h t t p s: //d o i .o r g /10 .10 07/s10 5 4 9 - 017- 7(9):867–85. https://doi.org/10.1158/1940-
4247-z PMID:28444533 6207.CAPR-13-0424 PMID:24985407
CHAPTER 5.9
SECTION 5

5.10
Cervical cancer
Successes in some communities
to be extended worldwide
Zvavahera Mike Chirenje Maribel Almonte (reviewer)
Bothwell Takaingofa Guzha Karen Canfell (reviewer)
Lynette Denny (reviewer)
guidelines were introduced in such as Colombia, India, and the

SUMMARY 2018 that incorporate imaging Philippines; this is probably be-
and pathology results. Lymph cause of improving socioeconomic
●● In 2018, there were an estimat- node involvement, an important conditions, and possibly because
ed 570 000 new cases of cervi- adverse prognostic factor, is of associated changes in behaviour
cal cancer and 311 000 deaths now included in the staging. and lifestyle. However, an increas-
from the disease worldwide. ing trend in incidence has been
●● Remarkable progress has been observed in countries in sub-Sa-
●● WHO has issued a call to ac- made worldwide to scale up HPV
tion for the elimination of cer- haran Africa, such as Uganda and
vaccination, especially in high- Zimbabwe, and in some countries
vical cancer as a public health income countries.
problem. in eastern Europe [3].
The elimination of cervical can-
●● Some changes to the WHO clas- cer as a public health problem is
sification of neoplasms of the Epidemiology considered a priority under the WHO
cervix were introduced in 2014. Cervical cancer constitutes 80% of 13th General Programme of Work.
●● The essential molecular interac- all cancers attributable to human In some high-income countries, the
tions of the different human pap- papillomavirus (HPV) infection [1]. combined approach of implementa-
illomavirus (HPV) oncoproteins The global disparity in cervical can- tion of wide-scale HPV vaccination
cer incidence and mortality rates is with adequate population coverage,
to induce cervical carcinogen-
an indicator of the enormous ineq- improved primary screening for high-
esis are now better understood.
uities in access to health services. risk HPV, and treatment of cervical
●● HPV infection causes almost Cervical cancer is the fourth cancer makes the elimination of cer-
all cervical squamous cell car- most common cancer type diag- vical cancer a possibility in the fore-
cinomas. About 5–10% of cer- nosed in women and the fourth most seeable future [4].
vical adenocarcinomas are uncommon cause of cancer death in
related to HPV infection. women. In 2018, there were an esti- Pathology
●● Primary HPV testing is a more mated 570 000 new cases of cervi- The most recent (2014) edition of
effective screening modality cal cancer and 311 000 deaths from the WHO classification of tumours
than cytology. It is now being the disease worldwide [2]. Cervical of the female reproductive organs
introduced in many high-income cancer remains the most common introduced changes to the classi-
countries, with an increasing fo- cause of cancer death in many fication of neoplasms of the cervix
cus on effective delivery in low- countries in Africa and South-East [5] (Box 5.10.1). These include the
and middle-income countries. Asia, where the incidence and mor- introduction of a stratified mucin-
tality rates are about 10 times those producing intraepithelial lesion as
●● There is increased interest in
in North America, Australia and a variant of adenocarcinoma in situ,
the use of biomarkers in cervi-
New Zealand, and West Asia [2] restructuring of the nomenclature of
cal cancer screening to better
(Fig. 5.10.1 and Fig. 5.10.2). adenocarcinomas, and the classifi-
triage women with high-risk
In regions with a high bur- cation of cervical precursor lesions
HPV infection.
den of cervical cancer, the inci- into a two-tiered system in line with
●● Because of the limitations of dence of cervical cancer has been the Bethesda classification for cy-
clinical staging, new staging decreasing in some countries, tology [5] (Table 5.10.1).
394
Box 5.10.1. Significant changes in the 2014 WHO classification of neoplasms of the
cervix.
FUNDAMENTALS
• Two-tiered subdivision of precursor lesions of squamous cell carcinoma ■■ Infection with high-risk human
(according to the Bethesda classification for cytology) papillomavirus (HPV) types
causes almost all cases of
• Stratified mucin-producing intraepithelial lesion (SMILE) as a variant cervical cancer.
of adenocarcinoma in situ (AIS)
■■ Cytology-based screening
• Subdivision of adenocarcinomas programmes have demon-
strated remarkable success in
• Relation of the individual carcinoma types to human papillomavirus (HPV) reducing the incidence of and
• Neuroendocrine tumours mortality from cervical cancer
in high-income countries. The
main limitation of cytology is its
The Lower Anogenital Squa include most squamous cell carci- relatively low sensitivity, espe-
cially if comprehensive quality
mous Terminology Standardization nomas without any specific differ- assurance processes are not in
Project also recommended the use entiation or cornification. place. Because of the complexi-
of a two-tiered classification system ties and cost involved in setting
Invasive glandular cell up cytology-based screening
for cervical precursor lesions, as programmes, most low- and
well as the use of p16 immunohisto- carcinomas middle-income countries have
chemical staining as a biomarker to Invasive cervical adenocarcinomas either opportunistic screening
differentiate between cervical pre- or no screening at all.
(Fig. 5.10.4) constitute 10–25% of
cancerous lesions and their mimics, cervical carcinomas. About 5–10% ■■ Advances in molecular technol-
and in the stratification of cervical in- ogy have made testing for
of cervical adenocarcinomas are high-risk HPV widely available,
traepithelial neoplasia grade 2 (CIN2) unrelated to HPV infection. albeit mostly in high-income
lesions [6]. Low-grade squamous The histological subtypes of cer- countries, and the increasing
intraepithelial lesions encompass focus is now on demonstrating
vical adenocarcinoma are shown its broader applicability to low-
HPV infection and CIN1, whereas
in Table 5.10.3. The most frequent and middle-income countries.
high-grade squamous intraepithe- Testing for high-risk HPV types
histological variant is HPV-related
lial lesions include CIN2 and CIN3. is currently being used for
adenocarcinoma of the usual type. primary screening, to triage
Invasive squamous cell Other types include the various women with atypical squamous
subtypes of mucinous adenocarci- cells of undetermined signifi-
carcinomas cance and low-grade squamous
nomas and clear cell carcinomas, intraepithelial lesion cytology
Invasive squamous cell carcinoma
of the cervix (Fig. 5.10.3) accounts which occur more commonly in results, for co-testing with cytol-
younger women. Primary serous ogy, and as a test of cure.
for 80–85% of cervical carcinomas.
HPV infection causes almost 100% adenocarcinomas are uncommon. ■■ The high negative predictive
Immunohistochemistry aids in the value of high-risk HPV DNA
of cases of cervical squamous cell testing has enabled screening
carcinoma, and in most cases an diagnosis of mesonephric tumours intervals to be safely increased.
underestimation of HPV prevalence and mixed adenocarcinoma and
■■ Previously, it was considered
is due to the limitations of relevant neuroendocrine carcinoma [8]. that a limitation of high-risk
studies [7]. HPV testing was its lower
The histological subtypes of Rare epithelial cervical specificity for detection of high-
tumours grade squamous intraepithelial
cervical squamous cell carcinoma lesions. However, this is effec-
are shown in Table 5.10.2. The term Rare epithelial neoplasms of the tively managed through the use
“squamous cell carcinoma, not oth- cervix (Table 5.10.4) include ade- of clinically validated tests, by
CHAPTER 5.10
SECTION 5
limiting the age range of testing,
erwise specified” was introduced to nosquamous carcinomas (1–2%), and – in some settings and in
some countries – by the effect
of HPV vaccination, which
Table 5.10.1. Comparison of classifications of precursor lesions of squamous cell enables HPV-based screening
carcinoma of the cervix to be done in women younger
than 30 years.
1975/1994 WHO 2003 WHO classification 2014 WHO classification
■■ There is much interest in
classification biomarkers to predict which
cervical precancerous lesions
Low (mild) dysplasia CIN grade 1 (CIN1) Low-grade squamous
are likely to progress in women
intraepithelial lesion (LSIL)
with high-risk HPV infection and
Moderate dysplasia CIN grade 2 (CIN2) High-grade squamous
normal, atypical squamous cells
of undetermined significance, or
Severe dysplasia CIN grade 3 (CIN3) intraepithelial lesion (HSIL)
low-grade squamous intraepi-
Carcinoma in situ
thelial lesion cytology results.
CIN, cervical intraepithelial neoplasia.
Chapter 5.10 • Cervical cancer 395

Fig. 5.10.1. Global distribution of estimated age-standardized (World) incidence rates (ASR) per 100 000 person-years for cervical
cancer, 2018.
Fig. 5.10.2. Global map of estimated age-standardized (World) mortality rates (ASR) per 100 000 person-years for cervical
cancer, 2018.
glassy cell carcinomas, and neu- Genetics and genomics loci within 6p21.3 were associated
roendocrine tumours. with increased susceptibility to cer-
Cervical cancer is a rare outcome
vical cancer [9,10].
in women with HPV infection. The
Rare non-epithelial cervical Persistent HPV infection is due
biological underpinnings of this pro-
tumours to both viral and host immune sys-
cess are not yet clearly understood. tem factors. Several factors attribu-
Rare non-epithelial neoplasms of There is renewed interest in the role table to HPV contribute to the ability
the cervix include mesenchymal of host genetics in the development of the infection to evade the host im-
types and other tumorous changes, of cervical cancer. mune system. Host genetic variants
such as postoperative spindle cell In the Han Chinese population, influence the ability of the immune
nodules. The occurrence of a sec- loci at 4q12 and 17q12 were asso- system to clear HPV infection.
ondary malignancy in the cervix is ciated with a higher risk of cervical New data from genome-wide
clinically important but is very rare. cancer; in the Swedish population, association studies have shown
396
Fig. 5.10.3. Photomicrograph of invasive Table 5.10.2. Histological types of squamous cell carcinoma of the cervix (2014 WHO
cervical squamous cell carcinoma, show- classification)
ing invasive squamous cells (bottom right)
budding off from a focus of cervical high- Histological type ICD-O code
grade squamous intraepithelial lesion (cen- Squamous cell carcinoma, NOS 8070/3
tre). Normal non-dysplastic cervical squa-
mous epithelium is present at the periphery Keratinizing squamous cell carcinoma 8071/3
(top right and top left). Haematoxylin and Non-keratinizing squamous cell carcinoma 8072/3
eosin stain, 40× magnification.
Papillary squamous cell carcinoma 8052/3
Basaloid squamous cell carcinoma 8083/3
Warty squamous cell carcinoma 8051/3
Verrucous squamous cell carcinoma 8051/3
Squamotransitional carcinoma 8120/3
Lymphoepithelioma-like carcinoma 8082/3
ICD-O, International Classification of Diseases for Oncology; NOS, not otherwise specified.
association was seen with HPV18- Etiology

associated cervical cancers [11].
Persistent infection with high-risk
Genetic analysis of 80 tumours of
that the amino acids carried at HPV is necessary for the devel-
the cervix for 1250 known mutations
positions 13 and 71 in pocket 4 of opment of cervical cancer (see
in 139 genes found the highest mu-
human leukocyte antigen (HLA)- tation rates in the PIK3CA (31.3%), Chapter 2.2). Co-factors associ-
DRB1 and at position 156 in HLA-B KRAS (8.8%), and EGFR (3.8%) ated with disease progression are
control whether HLA haplotypes in- genes. PIK3CA mutation rates did well established.
crease the risk of cervical neoplasia not differ significantly between ad- HPV DNA encodes for six early
or protect against cervical cancer enocarcinomas and squamous cell genes and two late genes. In the
[11]. Three HLA haplotypes were carcinomas. KRAS mutations were most recent evaluation by the IARC
identified that are associated with identified only in adenocarcinomas, Monographs programme, the follow-
an increased risk of both HPV16- and a new EGFR mutation was de- ing 12 HPV types were classified as
and HPV18-associated cervical tected only in squamous cell carci- carcinogenic to humans: 16, 18, 31,
cancer, and for the development of nomas [12]. PIK3CA mutations may 33, 35, 39, 45, 51, 52, 56, 58, and 59
both cervical squamous cell carci- be associated with shorter survival. [13]. The most common oncogenic
noma and adenocarcinoma. The
HLA-B*15 haplotype was associ-
ated with a lower risk of squamous Table 5.10.3. Histological types of adenocarcinoma of the cervix (2014 WHO classifi-
cell carcinomas and other HPV16- cation) related to human papillomavirus infection
associated cervical cancers, but no
Histological type Related HPV ICD-O code
Endocervical adenocarcinoma, usual type HR-HPV 8140/3
Fig. 5.10.4. Photomicrograph of invasive
Mucinous adenocarcinoma, NOS – 8480/3
cervical adenocarcinoma, showing fused
glandular structures (top) as well as small Mucinous adenocarcinoma, stomach type No 8482/3
nests and cords of invasive glandular Mucinous adenocarcinoma, intestinal type HR-HPV 8144/3
CHAPTER 5.10
SECTION 5
cells (bottom centre and bottom left)
surrounded by pale pink, inflamed mu- Mucinous adenocarcinoma, signet ring cell type Partial HR-HPV 8490/3
cus. Haematoxylin and eosin stain, 40× Villoglandular carcinoma HPV16, HPV18, 8263/3
magnification. HPV45
Endometrioid carcinoma Noa 8380/3
Clear cell carcinoma No or HR-HPV b 8310/3
Serous carcinoma Noa 8441/3
Mesonephric carcinoma No 9110/3
Mixed adenocarcinoma and neuroendocrine carcinoma HR-HPV 8574/3
HR-HPV, high-risk human papillomavirus; ICD-O, International Classification of Diseases for

Oncology; NOS, not otherwise specified.
a
If these tumour types contain HPV DNA, they are considered a morphological variant of
endocervical adenocarcinoma, usual type.
b
There is conflicting information in the literature on the HPV reference of the clear cell type.

Table 5.10.4. Other rare epithelial neoplasms of the cervix (2014 WHO classification) mal residual disease after treatment
related to human papillomavirus infection and subclinical recurrence [16].
Squamous cell carcinoma an-
Histological type Related HPV ICD-O code
tigen (SCC-Ag) is a protein-based
Adenosquamous carcinoma HPV18, HPV16 8560/3 biomarker with a good correlation
Glassy cell carcinoma HPV18 8015/3 between its levels before treatment
and tumour burden. It can poten-
Adenoid basal carcinoma HPV16, HPV33 8098/3
tially be used to provide prognos-
Adenoid cystic carcinoma HPV16 8200/3 tic information, as well as to detect
Undifferentiated carcinoma HPV16 8020/3 recurrences early. SCC-Ag was
Neuroendocrine tumours – –
also shown to be a useful adjunct
to imaging in detecting lymph node
“Low-grade” neuroendocrine tumour HR-HPV –
metastasis, an important adverse
Carcinoid tumour – 8240/3 prognostic factor [17]. There is an
Atypical carcinoid tumour – 8249/3 association between SCC-Ag lev-
els and disease recurrence and
“High-grade” neuroendocrine carcinoma HR-HPV (HPV18) –
mortality in women with newly diag-
Small cell neuroendocrine carcinoma – 8041/3 nosed cervical cancer [18].
Large cell neuroendocrine carcinoma – 8013/3 Normal epithelium and carcino-
mas of the uterine cervix produce
HR-HPV, high-risk human papillomavirus; ICD-O, International Classification of Diseases for Oncology.
serum cytokeratin 19 fragments
(CYFRA 21.1). CYFRA 21.1 was
HPV types identified in cervical can- shown to be a useful biomarker
nosuppressive conditions, smoking
in predicting parametrial invasion,
cer include HPV16 (53%), HPV18 (in squamous cell carcinomas only),
another important adverse prog-
(15%), HPV45 (9%), HPV31 (6%), multiparity, and long-term use of
nostic factor. A predictive model us-
and HPV33 (3%) [14]. oral contraceptives [15].
ing CYFRA 21.1 levels, tumour size,
The integration of the viral epi-
and SCC-Ag levels demonstrated
some into the host genome is a Prognostic markers for an ability to accurately predict par-
necessary step in the develop-
ment of cervical cancer. The E6
invasive cervical cancer ametrial invasion in patients with
Despite the wide availability of International Federation of Gy ne
and E7 oncoproteins deactivate
screening in high-income countries cology and Obstetrics (FIGO) stage
the protein products of the TP53
and recent advances in radiother- IB cervical cancer [19].
and retinoblastoma (RB) tumour
suppressor genes, respectively. apy techniques, the 5-year overall
Overexpression of the E6 and E7 survival in cervical cancer remains Socioeconomic, racial,
oncogenes results in the loss of about 60–70% in high-income and ethnic differences
cell-cycle control and leads to un- countries and is much lower in In large parts of sub-Saharan Africa,
controlled cellular proliferation, im- low- and middle-income countries. as well as in countries in Melanesia,
mortalization, and reduced apo- Research is under way on potential cervical cancer is the leading cause
ptosis; the result is chromosomal biomarkers that could help to iden- of cancer death in women, whereas
instability and the development of tify the disease in early stages, pre- in countries with high values of the
cervical cancer. The essential mo- dict tumour burden, detect recur- Human Development Index (HDI),
lecular interactions of the different rences early, and offer prognostic cervical cancer incidence and mor-
HPV oncoproteins to induce cervi- information, thus providing a poten- tality rates are declining [20].
cal carcinogenesis are summarized tial way to improve survival. Many In some countries with high HDI,
in Fig. 5.10.5. of these biomarkers are not yet in racial disparities in disease burden
In most women, the activity routine clinical use. and mortality are common. In the
of humoral and cellular-mediated HPV integration mutation was USA, the incidence of and mortal-
immunity helps to clear the HPV shown to be a molecular marker ity from cervical cancer in African
infection within 12–24 months. If of circulating tumour DNA in HPV- American women was shown to
persistent high-grade squamous in- associated tumours. Tumour bur- be twice that in White women [21].
traepithelial lesions are left untreat- den, an adverse prognostic marker, These disparities are caused by un-
ed, the risk of developing cervical correlated well with serum levels of equal access to primary prevention
cancer is about 30%. circulating tumour DNA. Therefore, (see Chapter 4.6), screening, and
Known co-factors associated circulating tumour DNA may provide treatment services. Compared with
with disease progression include important prognostic information and other ethnicities, African American
infection with HIV and other immu- may also play a role in detecting mini- girls were less likely to complete
398
Fig. 5.10.5. The role of promising biomarkers in the molecular mechanisms that lead advanced disease. These interven-
to a transforming infection. Schematic diagram of molecular and cellular processes tions form part of targets and indica-
that are affected during cervical carcinogenesis after infection with high-risk human tors of the WHO Global Action Plan
papillomavirus (HPV). E6 leads to activation of telomerase-related genes as well
as to the ubiquitination of p53. This results in the degradation of p53 and therefore
for the Prevention and Control of
inhibits apoptosis. E7 inactivates pRb and therefore increases the amount of free Noncommunicable Diseases 2013–
E2F in the cell, leading to both an increase in p16 and aberrant proliferation (which 2020 [23].
can be detected by increased levels of Ki-67 expression). In combination with the
inactivation of tumour suppressor genes (CADM1 and MAL), these actions lead to the
immortalization of the cell. This, in turn, leads to genomic instability, which cannot be Prevention
counteracted by DNA repair mechanisms because these mechanisms are inactivated
by the high-risk HPV oncogenes. Whether viral integration should be considered as Primary prevention
an initiator of genomic instability or a result of it is currently unclear. Nevertheless, Remarkable progress has been
the mechanisms shown in this flow chart lead to a transforming infection, causing the made worldwide to scale up HPV
occurrence of severe dysplasia and ultimately resulting in cervical malignancy.
vaccination, especially in high-in-
come countries (see Chapter 6.3). In
↑ p16 the past 5 years, very few low- and
middle-income countries have rolled
out countrywide HPV vaccination
Free
programmes. More countries are
E6 ↓ p53 ↑ ↓ pRb E7 preparing to introduce national vac-
E2F
cination programmes with the sup-
port of Gavi, the Vaccine Alliance.
For both the bivalent and the
quadrivalent HPV vaccine, two dos-
Aberrant Inactivation of tumor
Activation of
Apoptosis suppressor genes
es were shown to be non-inferior to
telomerase proliferation
inhibition (CADM1 and MAL) three doses, and WHO now recom-
related genes (↑ Ki-67)
by methylation mends the use of two doses in girls
younger than 14 years [24]. There
is emerging evidence that one dose
of HPV vaccine may be equally ef-
Immortalization ficacious; this will reduce the cost of
vaccines and make the delivery of
vaccines easier in low- and middle-
income countries. The introduction
HPV
Genomic instability in 2014 of the nonavalent HPV vac-
integration
cine (against HPV types 6, 11, 16,
18, 31, 33, 45, 52, and 58) was a
significant scientific advance that
Transformation
expanded the number of oncogenic
HPV types for which infection is pre-
ventable through vaccination. For
the three doses of the HPV vaccine with low HDI and in countries with the nonavalent vaccine, WHO also
required at the time of the study high HDI [21]. recommends the use of two doses
[22]. In the USA, women in minority These disparities across socio- in girls younger than 14 years.
CHAPTER 5.10
SECTION 5
groups with low socioeconomic sta- economic, racial, and ethnic groups New HPV vaccines are currently
have also been documented both in undergoing clinical trials, and they
tus tend to be underinsured, which
other countries with high HDI and in may become available by 2020.
limits their access to screening and
countries with low HDI (see Chapter However, the insufficient HPV vac-
clinical services. When these wom-
1.3). However, the burden and im- cine supply is a major challenge
en are screened, they are more
pact of cervical cancer on communi- and will remain a constraint in low-
likely to be lost to follow-up and to and middle-income countries for
ties can be mitigated by implement-
later present with advanced dis- ing national HPV vaccination and the foreseeable future.
ease [22]. The geographical loca- screening programmes with effec-
tion may also play a role in these tive treatment of high-grade squa- Secondary prevention
disparities. Women living in rural ar- mous intraepithelial lesions, early Secondary prevention of cervical
eas have the lowest screening rates detection and treatment of cervical cancer with cytology screening has
and the highest incidence rates of cancer, and improvement of pal- reduced the incidence of cervical
cervical cancer, both in countries liative care services for women with cancer in high-income countries.

Large clinical trials have shown mation. The PreTect HPV-Proofer CIN3. In cytology samples positive
that HPV-based screening leads to assay and the NucliSENS EasyQ for high-risk HPV, a sensitivity of
increased detection of precursor le- HPV assay are nucleic acid se- 70% and a specificity of 78% were
sions and decreased rates of inva- quence-based amplification tests demonstrated for the detection of
sive cervical cancer [25]. WHO and designed to detect HPV E6/E7 lesions of CIN3 or worse [32,33].
other organizations have recom- messenger RNA (mRNA) of the five
mended primary HPV testing in set- most common oncogenic high-risk Management of invasive
tings with sufficient resources. The HPV types (16, 18, 31, 33, and 45). cervical cancer
advent of portable point-of-care The APTIMA HPV assay is a target
testing devices will lead to the wide amplification nucleic acid probe test Microinvasive cervical cancer is
availability of this screening modal- typically an incidental histological
that detects the viral mRNA of 14
diagnosis after large loop excision
ity and increase its use in low- and HPV types (16, 18, 31, 33, 35, 39,
of the transformation zone (type 1
middle-income countries. 45, 51, 52, 56, 58, 59, 66, and 68).
and 2 excision) or cone biopsy (type
HPV self-sampling was intro- In women who have negative
3 excision). Macroscopic cervical
duced to overcome known barri- cytology and are positive for high-
cancer is often suspected clini-
ers to screening, which include risk HPV, a positive mRNA test re-
cally, because most of the women
restrictive work schedules as well sult implies an increased risk of pro-
present with a foul-smelling watery
as cultural and religious beliefs. gressive lesions compared with a
and sometimes bloody vaginal dis-
Therefore, self-sampling has the negative mRNA result; mRNA tests
charge, irregular vaginal bleeding,
potential to increase coverage of showed a higher specificity than
and contact bleeding.
cervical cancer screening in non- DNA tests in detecting high-grade
Until recently, FIGO staging
attendees in both high-income and cervical lesions [30]. Clinically,
of cervical cancer was performed
low-income countries. In Argentina, HPV E6/E7 mRNA molecular test-
mainly by clinical examination, with
the uptake of screening improved ing is being incorporated into cervi-
a few other procedures that were al-
from 20% with cytology-based cal cancer screening algorithms to lowed to change the stage. In 2018,
screening to 86% with the imple- triage women who are positive for the FIGO Gynecologic Oncology
mentation of HPV self-sampling high-risk HPV and have negative Committee revised this to include
[26]. The diagnostic accuracy of cytology to either immediate col- imaging and pathology results,
self-collected samples compares poscopy or close follow-up. Testing where available, to assign the stage.
favourably with that of clinician-col- for E6/E7 mRNA of high-risk HPV The revised FIGO staging is shown
lected specimens. types has also been found to be in Table 5.10.5 [34]. FIGO stage IB
useful as a test of cure. has now been subdivided into three
Improved methods of Women with HPV16 or HPV18 (instead of two) substages, and
detection and diagnosis infection have a much higher risk of lymph node involvement, an impor-
Biomarkers are also being exten- developing high-grade squamous tant adverse prognostic factor, is
sively evaluated for incorporation intraepithelial lesions compared now included in FIGO stage IIIC.
into cervical cancer screening pro- with women who are positive for Treatment for early-stage dis-
grammes, to predict which cervical other high-risk HPV types and have ease is surgical. However, concur-
precancerous lesions are likely to negative cytology [31]. This finding rent chemoradiotherapy has simi-
progress. has been incorporated into cervical lar outcomes. In locally advanced
Dual staining with p16INK4a and cancer screening algorithms to tri- disease, concurrent chemoradio-
Ki-67 has shown high sensitivity age women with normal, atypical therapy is the treatment of choice.
in detecting high-grade squamous squamous cells of undetermined Treatment of women with FIGO
intraepithelial lesions in both cyto- significance, and low-grade squa- stage IVB and recurrent disease is
logical and histological specimens mous intraepithelial lesion cytol- highly individualized. Palliative care
[27–29]. Clinically, it can be used to ogy results to either immediate col- remains an important component
differentiate reactive from dysplas- poscopy or repeat co-testing after of management of cervical cancer.
tic cervical lesions and to detect 12 months. This strategy effectively Women with advanced disease
high-grade squamous intraepithe- reduces the number of women re- should have an early referral to a
lial lesions with higher accuracy. ferred for colposcopy. palliative care team. There is a role
Persistent infection with high- Methylation of the cell adhe- for fertility-sparing surgery in young
risk HPV results in overexpression sion molecule 1 (CADM1) and women with early-stage disease
of the E6 and E7 viral oncogenes, T-lymphocyte maturation-associat- who desire to become parents.
which leads to cellular prolifera- ed protein (MAL) genes was associ- Long-term follow-up is required to
tion, immortalization, and transfor- ated with a high risk of developing detect recurrence.
400
Table 5.10.5. 2018 International Federation of Gynecology and Obstetrics (FIGO) staging of cancer of the cervix uteri
Stage a Description
I The carcinoma is strictly confined to the cervix uteri (extension to the corpus should be disregarded)
IA Invasive carcinoma that can be diagnosed only by microscopy, with maximum depth of invasion < 5 mmb
IA1 Measured stromal invasion < 3 mm in depth
IA2 Measured stromal invasion ≥ 3 mm and < 5 mm in depth
IB Invasive carcinoma with measured deepest invasion ≥ 5 mm (greater than stage IA), lesion limited to the cervix uteric
IB1 Invasive carcinoma with ≥ 5 mm depth of stromal invasion and < 2 cm in greatest dimension
IB2 Invasive carcinoma ≥ 2 cm and < 4 cm in greatest dimension
IB3 Invasive carcinoma ≥ 4 cm in greatest dimension
II The carcinoma invades beyond the uterus but has not extended into the lower third of the vagina or to the pelvic wall
IIA Involvement limited to the upper two thirds of the vagina, without parametrial involvement
IIA1 Invasive carcinoma < 4 cm in greatest dimension
IIA2 Invasive carcinoma ≥ 4 cm in greatest dimension
IIB With parametrial involvement but not up to the pelvic wall
III The carcinoma involves the lower third of the vagina and/or extends to the pelvic wall and/or causes hydronephrosis or non-
functioning kidney and/or involves pelvic and/or para-aortic lymph nodesd
IIIA The carcinoma involves the lower third of the vagina, with no extension to the pelvic wall
IIIB Extension to the pelvic wall and/or hydronephrosis or non-functioning kidney (unless known to be due to another cause)
IIIC Involvement of pelvic and/or para-aortic lymph nodes, irrespective of tumour size and extent (with r and p notations) d
IIIC1 Pelvic lymph node metastasis only
IIIC2 Para-aortic lymph node metastasis
IV The carcinoma has extended beyond the true pelvis or has involved (biopsy proven) the mucosa of the bladder or rectum. A
bullous oedema, as such, does not permit a case to be allotted to stage IV
IVA Spread of the growth to adjacent organs
IVB Spread to distant organs
a
When in doubt, the lower staging should be assigned.
b
Imaging and pathology can be used, when available, to supplement clinical findings with respect to tumour size and extent, in all stages.
c
The involvement of vascular/lymphatic spaces does not change the staging. The lateral extent of the lesion is no longer considered.
d
Adding notation of r (imaging) and p (pathology) to indicate the findings that are used to allocate the case to stage IIIC. For example, if imaging indicates
pelvic lymph node metastasis, the stage allocation would be stage IIIC1r, and if confirmed by pathological findings, it would be stage IIIC1p. The type of
imaging modality or pathology technique used should always be documented. When in doubt, the lower staging should be assigned.
References
1. de Martel C, Plummer M, Vignat J, 4. Hall MT, Simms KT, Lew JB, Smith MA, 6. Darragh TM, Colgan TJ, Cox JT, Heller
Franceschi S (2017). Worldwide burden of Brotherton JM, Saville M, et al. (2019). The DS, Henry MR, Luff RD, et al.; Members
cancer attributable to HPV by site, coun- projected timeframe until cervical cancer of LAST Project Work Groups (2012). The
CHAPTER 5.10
Lower Anogenital Squamous Terminology
SECTION 5
try and HPV type. Int J Cancer. 141(4): elimination in Australia: a modelling study.
664–70. https://doi.org/10.1002/ijc.30716 Lancet Public Health. 4(1):e19–27. https:// Standardization Project for HPV-associated
PMID:28369882 doi.org/10.1016/S2468-2667(18)30183-X lesions: background and consensus recom-
PMID:30291040 mendations from the College of American
2. Bray F, Ferlay J, Soerjomataram I, Siegel Pathologists and the American Society
RL, Torre LA, Jemal A (2018). Global 5. Kurman RJ, Carcangiu ML, Herrington for Colposcopy and Cervical Pathology. J
cancer statistics 2018: GLOBOCAN es- CS, Young RH, editors (2014). WHO clas- Low Genit Tract Dis. 16(3):205–42. https://
timates of incidence and mortality world- sification of tumours of the female repro- doi.org/10.1097/LGT.0b013e31825c31dd
wide for 36 cancers in 185 countries. CA ductive organs. Lyon, France: International PMID:22820980
Cancer J Clin. 68(6):394–424. https://doi. Agency for Research on Cancer (WHO
org/10.3322/caac.21492 PMID:30207593 7. de Sanjosé S, Serrano B, Castellsagué X,
Classification of Tumours series, 4th edi-
Brotons M, Muñoz J, Bruni L, et al. (2012).
3. Forman D, de Martel C, Lacey CJ, tion). Available from: https://publications.
Human papillomavirus (HPV) and related
Soerjomataram I, Lortet-Tieulent J, Bruni iarc.fr/16.
cancers in the Global Alliance for Vaccines
L, et al. (2012). Global burden of human and Immunization (GAVI) countries. A
papillomavirus and related diseases. WHO/ICO HPV Information Centre report.
Vaccine. 30(Suppl 5):F12–23. https:// Vaccine. 30(Suppl 4):D1–83, vi. https://
doi.org/10.1016/j.vaccine.2012.07.055 doi.org/10.1016/S0264-410X(12)01435-1
PMID:23199955 PMID:23510764

8. Lax SF, Horn LC, Löning T (2016). 17. Zhou Z, Li W, Zhang F, Hu K (2017). The 27. von Knebel Doeberitz M, Reuschenbach
Categorization of uterine cervix tumors: value of squamous cell carcinoma anti- M, Schmidt D, Bergeron C (2012).
what’s new in the 2014 WHO classifica- gen (SCCa) to determine the lymph nodal Biomarkers for cervical cancer screening:
tion [in German]. Pathologe. 37(6):573–84. metastasis in cervical cancer: a meta- the role of p16INK4a to highlight transforming
https://doi.org/10.1007/s00292-016-0247-8 analysis and literature review. PLoS One. HPV infections. Expert Rev Proteomics.
PMID:27770187 12(12):e0186165. https://doi.org/10.1371/ 9(2):149–63. https://doi.org/10.1586/epr.
journal.pone.0186165 PMID:29227998 12.13 PMID:22462787
9. Chen D, Juko-Pecirep I, Hammer J,
Ivansson E, Enroth S, Gustavsson I, et al. 18. Charakorn C, Thadanipon K, Chaijindara 28. Petry KU, Schmidt D, Scherbring S,
(2013). Genome-wide association study of tana S, Rattanasiri S, Numthavaj P, Luyten A, Reinecke-Lüthge A, Bergeron
susceptibility loci for cervical cancer. J Natl Thakkinstian A (2018). The association C, et al. (2011). Triaging Pap cytology
Cancer Inst. 105(9):624–33. https://doi. between serum squamous cell carci- negative, HPV positive cervical can-
org/10.1093/jnci/djt051 PMID:23482656 noma antigen and recurrence and sur- cer screening results with p16/Ki-67
vival of patients with cervical squamous dual-stained cytology. Gynecol Oncol.
10. Shi Y, Li L, Hu Z, Li S, Wang S, Liu J, et cell carcinoma: a systematic review 121(3):505–9. https://doi.org/10.1016/j.
al. (2013). A genome-wide association and meta-analysis. Gynecol Oncol. ygyno.2011.02.033 PMID:21420158
study identifies two new cervical cancer 150(1):190–200. https://doi.org/10.1016/j.
susceptibility loci at 4q12 and 17q12. ygyno.2018.03.056 PMID:29606483 29. Schmidt D, Bergeron C, Denton KJ,
Nat Genet. 45(8):918–22. https://doi. Ridder R; European CINtec Cytology
org/10.1038/ng.2687 PMID:23817570 19. Kong TW, Piao X, Chang SJ, Paek J, Lee Study Group (2011). p16/Ki-67 dual-stain
Y, Lee EJ, et al. (2016). A predictive mod- cytology in the triage of ASCUS and LSIL
11. Leo PJ, Madeleine MM, Wang S, Schwartz el for parametrial invasion in patients with Papanicolaou cytology: results from the
SM, Newell F, Pettersson-Kymmer U, et FIGO stage IB cervical cancer: individual- European Equivocal or Mildly Abnormal
al. (2017). Defining the genetic suscepti- ized approach for primary treatment. Int J Papanicolaou Cytology Study. Cancer
bility to cervical neoplasia – a genome- Gynecol Cancer. 26(1):184–91. https://doi. Cytopathol. 119(3):158–66. https://doi.
wide association study. PLoS Genet. org/10.1097/IGC.0 0 0 0 0 0 0 0 0 0 0 0 0560 org/10.1002/cncy.20140 PMID:21442767
13(8):e1006866. https://doi.org/10.1371/ PMID:26512782
journal.pgen.1006866 PMID:28806749 30. Cattani P, Zannoni GF, Ricci C, D’Onghia
20. Torre LA, Bray F, Siegel RL, Ferlay J, S, Trivellizzi IN, Di Franco A, et al.
12. Wright AA, Howitt BE, Myers AP, Dahlberg Lortet-Tieulent J, Jemal A (2015). Global (2009). Clinical performance of human
SE, Palescandolo E, Van Hummelen cancer statistics, 2012. CA Cancer J Clin. papillomavirus E6 and E7 mRNA test-
P, et al. (2013). Oncogenic mutations in 65(2):87–108. https://doi.org/10.3322/caac. ing for high-grade lesions of the cer-
cervical cancer: genomic differences be- 21262 PMID:25651787 vix. J Clin Microbiol. 47(12):3895–901.
tween adenocarcinomas and squamous https://doi.org/10.1128/JCM.01275 - 09
cell carcinomas of the cervix. Cancer. 21. Musselwhite LW, Oliveira CM, Kwaramba
PMID:19828739
119(21):3776–83. https://doi.org/10.1002/ T, de Paula Pantano N, Smith JS, Fregnani
cncr.28288 PMID:24037752 JH, et al. (2016). Racial/ethnic disparities in 31. Kitchener HC, Gilham C, Sargent A, Bailey
cervical cancer screening and outcomes. A, Albrow R, Roberts C, et al. (2011). A
13. Bouvard V, Baan R, Straif K, Grosse Y, Acta Cytol. 60(6):518–26. https://doi. comparison of HPV DNA testing and liq-
Secretan B, El Ghissassi F, et al.; WHO org/10.1159/000452240 PMID:27825171 uid based cytology over three rounds of
International Agency for Research on primary cervical screening: extended fol-
22. Chatterjee S, Gupta D, Caputo TA,
Cancer Monograph Working Group low up in the ARTISTIC trial. Eur J Cancer.
Holcomb K (2016). Disparities in gyneco-
(2009). A review of human carcinogens – 47(6):864–71. https://doi.org/10.1016/j.ejca.
logical malignancies. Front Oncol. 6:36.
Part B: biological agents. Lancet Oncol. 2011.01.008 PMID:21334200
https://doi.org/10.3389/fonc.2016.00036
10(4):321–2. https://doi.org/10.1016/S1470-
PMID:26942126 32. Overmeer RM, Louwers JA, Meijer
2045(09)70096-8 PMID:19350698
23. WHO (2013). WHO Global Action Plan CJ, van Kemenade FJ, Hesselink AT,
14. Muñoz N (2000). Human papillomavirus Daalmeijer NF, et al. (2011). Combined
for the Prevention and Control of Non
and cancer: the epidemiological evi- CADM1 and MAL promoter methylation
communicable Diseases 2013–2020.
dence. J Clin Virol. 19(1–2):1–5. https:// analysis to detect (pre-)malignant cervical
Geneva, Switzerland: World Health
doi.org/10.1016/s1386-6532(00)00125-6 lesions in high-risk HPV-positive women.
Organization. Available from: https://
PMID:11091143 Int J Cancer. 129(9):2218–25. https://doi.
a p p s .w h o . i n t / i r i s / b i t s t r e a m / h a n d l e /
10 6 6 5 / 9 4 3 8 4 / 978 9 2415 0 6 2 3 6 _ e n g . org/10.1002/ijc.25890 PMID:21190187
15. International Collaboration of Epidemio
logical Studies of Cervical Cancer (2007). pdf?sequence=1. 33. Hesselink AT, Heideman DA, Steenbergen
Comparison of risk factors for invasive RD, Coupé VM, Overmeer RM, Rijkaart
24. WHO (2014). Meeting of the Strategic
squamous cell carcinoma and adenocar- D, et al. (2011). Combined promoter meth-
Advisory Group of Experts on immu-
cinoma of the cervix: collaborative reanal- ylation analysis of CADM1 and MAL: an
nization, April 2014 – conclusions and
ysis of individual data on 8,097 women recommendations. Wkly Epidemiol Rec. objective triage tool for high-risk human
with squamous cell carcinoma and 1,374 89(21):221–36. PMID:24864348 papillomavirus DNA-positive women. Clin
women with adenocarcinoma from 12 Cancer Res. 17(8):2459–65. https://doi.
epidemiological studies. Int J Cancer. 25. Ogilvie G, Nakisige C, Huh WK, org/10.1158/1078 - 0 432.CCR-10 -25 48
120(4):885–91. https://doi.org/10.1002/ijc. Mehrotra R, Franco EL, Jeronimo J PMID:21389098
22357 PMID:17131323 (2017). Optimizing secondary preven-
tion of cervical cancer: recent advances 34. Bhatla N, Berek J, Cuello M, Denny L,
16. Campitelli M, Jeannot E, Peter M, and future challenges. Int J Gynaecol Grenman S, Karunaratne K, et al. (2018).
Lappartient E, Saada S, de la Rochefordière Obstet. 138(Suppl 1):15–9. https://doi. New revised FIGO staging of cervical can-
A, et al. (2012). Human papillomavirus org/10.1002/ijgo.12187 PMID:28691338 cer (2018). Abstract S020.2. Presented
mutational insertion: specific marker of at the XXII FIGO World Congress of
circulating tumor DNA in cervical cancer 26. Arrossi S, Thouyaret L, Laudi R, Marín Gynecology and Obstetrics, Rio de
patients. PLoS One. 7(8):e43393. https:// O, Ramírez J, Paolino M, et al. (2015). Janeiro, Brazil, 14–19 October 2018. Int J
doi.org/10.1371/journal.pone.0043393 Implementation of HPV-testing for cervi- Gynecol Obstet. 143(S3):43–157. https://
PMID:22937045 cal cancer screening in programmatic doi.org/10.1002/ijgo.12584
contexts: the Jujuy demonstration proj-
ect in Argentina. Int J Cancer. 137(7):
1709–18. https://doi.org/10.1002/ijc.29530
PMID:25807897
402
5.11
Endometrial cancer
Prevention through control of obesity
Penelope M. Webb Dagfinn Aune (reviewer)

Laure Dossus (reviewer)
of these potential chemopre- erally lowest in Africa and Asia and

SUMMARY ventive agents is not yet clear. highest in Europe and North America
[1]. They increase with increasing
●● A new classification system that sociodemographic index (a measure
categorizes endometrial can- It is now more than 40 years since of development based on income,
cers on the basis of their molec- the publication of the first reports education, and fertility rates); almost
ular characteristics – microsat- of an association between use of three quarters of cases occur in the
ellite instability, POLE mutation, estrogen replacement therapy and top two quintiles [2].
no specific molecular features, risk of endometrial cancer, and Incidence rates of endometrial
or TP53 mutation – provides im- 35 years since endometrial cancers cancer are increasing, both over
proved prognostic information, were classified as either estrogen- time and in successive birth cohorts.
but the implications for etiology dependent (type 1) or estrogen-in- Some of the most rapid increases
are not yet known. dependent (type 2). have been seen in countries that
●● Although distinct in terms of During the decades after this have undergone rapid socioeconom-
their histology and clinical out- seminal work, endometrial cancer ic transitions (see Chapter 1.3), such
comes, high-grade type 2 en- attracted less research interest than as Japan and Singapore (Fig. 5.11.1)
dometrial cancers are not estro- cancer types that are more com- [1]. Interpreting these trends is chal-
gen-independent, as previously mon and more deadly. However, lenging, because of the multiple
considered, but share many risk rising incidence rates and a greater external influences on risk and the
factors with the more common focus on the rarer but more aggres- varying hysterectomy rates, but the
type 1 endometrial cancers, in- sive type 2 endometrial cancers increasing prevalence of obesity is
have changed this. In the past 5–10 likely to be a major contributor (see
cluding factors associated with
years, there have been major shifts Chapter 2.7).
estrogen exposure.
in the understanding of both the Some reports suggest that in-
●● Approximately one third of endo- molecular biology and the etiology cidence rates are increasing more
metrial cancers can be attributed of endometrial cancer. rapidly for type 1 cancers; this is
to overweight and obesity and
consistent with the change being
a smaller proportion to physical
CHAPTER 5.11
Epidemiology
SECTION 5
driven by the prevalence of obesity.
inactivity; therefore, effective in-
Globally, uterine cancer is the sev- However, in Denmark increases
terventions to reduce the preva-
enth most common cancer and the have been reported in the inci-
lence of obesity and increase
14th most common cause of cancer dence of type 2 cancers, despite an
physical activity levels are likely
death in women, with an estimated overall decline in the incidence of
to have the greatest impact on
382  000 new cases and 90  000 endometrial cancer [3]. In the USA,
incidence rates.
deaths in 2018. Age-standardized incidence rates of type 2 cancers
●● Progestin-containing intrauter- incidence rates (per 100 000) vary have also increased more rapidly
ine devices, metformin, and, about 12-fold between countries, than those of type 1 cancers, with
possibly, non-steroidal anti-in- from 2 to 24, although in a few coun- marked increases in Asian women
flammatory drugs may reduce tries the reported rates are lower (e.g. and particularly in non-Hispanic
incidence of endometrial can- 1.5 in Guinea and 1.8 in Mongolia) Black women, who now have the
cer in high-risk women, but the or higher (e.g. 24.1 in Lithuania and highest rates of these more aggres-
full range of risks and benefits 24.9 in Belarus). The rates are gen- sive endometrial cancers [4].
Chapter 5.11 • Endometrial cancer 403

Fig. 5.11.1. Age-standardized (World) incidence rates per 100 000 person-years by

calendar year in selected countries for uterine cancer.
FUNDAMENTALS
20 ■■ Endometrial cancers, which
arise in the lining of the uterus
18 (endometrium), comprise
approximately 90% of all
16
uterine cancers.
Age-standardized incidence per 100 000
■■ Although there is no screening

14 test, endometrial cancers
commonly cause abnormal
vaginal bleeding or discharge;
12
as a result, a high proportion
are diagnosed at an early
10 stage, and 10-year survival
rates are about 80%.
8 ■■ Historically, endometrial
cancer is classified into
6 two major types. The more
common type 1 cancers,
4
also described as estrogen-
dependent, are low-grade
endometrioid tumours that
2 arise on a background of
endometrial hyperplasia. The
0 less common type 2 cancers,
1985 1990 1995 2000 2005 initially labelled estrogen-
independent, are typically
USA White Canada Sweden Norway high-grade serous and clear
cell tumours that arise in an
USA Black UK Australia Singapore
atrophic endometrium.
China Japan
■■ Type 1 cancers are strongly
associated with exposure to
estrogen unopposed by a
Genetics and genomics Low-risk genetic variants
progestogen.
Having a first-degree relative with
Single-gene defects endometrial cancer approximately ■■ Well-established risk factors
Endometrial cancer is associated doubles a woman’s risk of the dis- for type 1 cancers include
with Lynch syndrome, a hereditary ease, but the high-risk genetic mu- conditions associated with
tations described above account greater endogenous estrogen
cancer syndrome that is character-
for only a small proportion of this exposure (obesity, early age
ized by mutations in the mismatch at menarche, late age at
repair genes MLH1, MSH2, MSH6, risk, suggesting that other, low-risk
menopause) or exogenous
and PMS2 or a nearby gene, genes also play a role. Until recently,
estrogen exposure (use of
EPCAM, that causes epigenetic few such genes had been identified
menopausal estrogen therapy,
silencing of MSH2. Women with a for endometrial cancer, but large- use of tamoxifen).
scale genome-wide association
germline mutation in one of these
studies (see Chapter 3.2) have now ■■ Factors associated with
genes have a 16–71% increased higher progestogen exposure
identified 16 genetic loci associated
risk of developing endometrial (pregnancy, use of oral
with endometrial cancer [6]. These
cancer before age 70 years, and include HNF1B, CYP19A1 (which contraceptives) are associated
the cancers typically develop at encodes the aromatase enzyme that with reduced risk.
a younger age than in the gen- converts androgens to estrogens),
eral population [5]. Women with and the MYC multicancer locus.
Cowden syndrome, which is char-
acterized by mutations in the PTEN Somatic changes and tations described above, somatic
tumour suppressor gene, are also molecular subtypes mutations in these genes and also
at increased risk of endometrial In addition to the rare germline mis- epigenetic silencing of the MSH2
cancer [5]. match repair gene and PTEN mu- promoter through methylation are
404
common events in endometrial The historical classification of of type 1 endometrial cancer; it is

cancer. Other genes that are fre- endometrial cancer into two types now clear that the major risk factors
quently mutated include PIK3CA, has long been fraught with problems, for type 2 cancers are very similar
KRAS, CTNNB1 (which encodes largely because these groups are (Fig. 5.11.2), although the relation-
β-catenin), ARID1A, and TP53. In defined based on the suspected eti- ship with obesity is somewhat weak-
2013, the Cancer Genome Atlas ology of the cancer and do not clearer [9]. This suggests that, despite
published a comprehensive analy- ly link to its pathological characteris- their initial description as estrogen-
sis of the genomic changes in en- tics or prognosis. Also, although the independent, type 2 cancers are also
dometrial cancers, in which they histology and grade of endometrial hormonally driven, although perhaps
identified four subsets of endome- cancers are used to determine treat- to a lesser extent than type 1 can-
trial cancers with differing molecu- ment, this classification has poor cers. There have not yet been any
lar profiles [7] (Table 5.11.1). reproducibility and does not reliably comprehensive studies comparing
Approximately 25% of endome- predict risk of recurrence, particular- risk factors for the various molecular
trial cancers, including a high pro- ly within the large group of endome- subtypes discussed above.
portion of high-grade endometrioid trioid cancers, for which outcomes
tumours, have defective mismatch can be very variable. Therefore,
Reproductive factors
repair capability, leading to micro- the new molecular classification is In addition to the strong inverse
satellite instability. In addition, about a major step forward, because it is association with increasing parity,
10% have a very high overall muta- reproducible and, importantly, differ- recent large-scale analyses have
tion frequency, including mutations entiates between histologically simi- shown that risk also decreases
in the exonuclease domain of the lar cancers that have very different by 13% for every 5-year increase
POLE gene (which encodes DNA prognosis [8]. However, it is not yet in age at last birth [10] and by 3%
polymerase ε). The third and larg- clear whether it will have any etio- for every 3 months that a woman
est group comprises mainly low- logical relevance. breastfeeds her children [11]. In
grade endometrioid tumours that contrast, a self-reported history of
have no specific molecular features, infertility has been associated with
although PTEN mutations are com- Etiology a 20% increase in risk [12].
mon in this group and in the first two Table 5.11.2 summarizes factors
groups. The fourth group, which in- known or suspected to increase or Exogenous hormones
cludes high-grade serous tumours decrease risk of endometrial cancer. Risk of endometrial cancer is re-
and carcinosarcomas as well as one It has long been recognized that duced by about 24% for every
quarter of high-grade endometrioid factors associated with increased 5 years of using oral contraceptives;
cancers, is characterized by high exposure to estrogen in the absence the effects are seen for both type 1
copy number and TP53 mutation. of a progestogen increase the risk and type 2 cancers, and, notably, the
Table 5.11.1. Molecular subtypes of endometrial cancer
TCGA label MSI (hypermutated) POLE (ultramutated) Copy-number low Copy-number high
(serous-like)
ProMisE label MMR-deficient POLE-EDM p53 wild-type p53-aberrant
Leiden/TransPORTEC label MSI POLE NSMP p53
Defining characteristic Mutation (germline or Mutation in exonucle Microsatellite stable, TP53 mutation
CHAPTER 5.11
SECTION 5
somatic) or epigenetic ase domain of POLE no POLE or TP53
modification of MLH1, DNA polymerase mutation
MSH2, MSH6, or
PMS2, leading to MMR
deficiency and MSI
Common mutations PTEN (~90%) POLE (100%) PTEN (~75%) TP53 (>90%)
PIK3CA (~50%) PTEN (> 90%) PIK3CA (~50%) PIK3CA (~50%)
[PTEN (~10%)]
Proportion of cancers ~25% ~10% ~40% ~25%
Typical histology High-grade High-grade Low-grade Serous,

endometrioid endometrioid endometrioid carcinosarcoma
Prognosis Intermediate Excellent Intermediate Poor
EDM, exonuclease domain mutation; MMR, mismatch repair; MSI, microsatellite instability; NSMP, no specific molecular profile; ProMisE, Proactive
Molecular Risk Classifier for Endometrial Cancer; TCGA, Cancer Genome Atlas Research Network.

Table 5.11.2. Factors associated with risk of endometrial cancer
Strength of evidence Factors that increase risk Factors that decrease risk
Convincing Family history Pregnancy
Use of estrogen replacement therapy Older age at last birth
Use of sequential estrogen plus progestin Use of oral contraceptives

(combination) menopausal hormone therapy
(progestin for < 10 days/month)
Use of tamoxifen
Body fatness
Diabetes
Early age at menarche
Late age at menopause
Infertility
Probable Metabolic syndrome Use of progestin-containing intrauterine devices
Hypertension Use of continuous estrogen plus progestin

(combination) menopausal hormone therapy
(progestin for ≥ 25 days/month)
Polycystic ovary syndrome Breastfeeding
High glycaemic load Physical activity
Adult height Coffee consumption
Possible Sedentary behaviour Use of metformin
Use of aspirin or other non-steroidal anti-

inflammatory drugs
Use of bisphosphonates
Insufficient Treatment for infertility; endometriosis; use of statins; other aspects of diet
benefit persists for at least 30 years apy) and use of sequential estrogen Although there are suggestions
after last use [13]. Despite reduc- plus progestin (combination) meno- that women who use clomiphene
tions in the hormone content of oral pausal hormone therapy (progestin citrate may have an increased risk
contraceptives since their introduc- for < 10–15 days per month) are of endometrial cancer, the current
tion, the effects appear to be similar associated with an increased risk evidence is limited and it is not pos-
for formulations used in the 1960s, of endometrial cancer, and this in- sible to separate any potential risk
1970s, and 1980s [13]. It is too soon crease in risk may be greater in thin associated with use of the medica-
to say whether use of newer formu- women and normal-weight women,
tion from that associated with the
lations, including progestin-only oral who have lower endogenous es-
underlying cause of the infertility
contraceptives, will reduce risk to trogen levels. In contrast, use of
[16] (see Chapter 2.11).
the same extent, but early data sug- continuous estrogen plus progestin
gest that progestin-containing intra- therapy (progestin for ≥ 25 days per
Body size and physical
uterine devices (e.g. the levonorges- month) has been associated with a
activity
trel-releasing intrauterine system) reduced risk of endometrial cancer
also protect against endometrial [15] (see Chapter 3.6). In postmenopausal women, the pri-
cancer [14]. Increasing use of fertility drugs mary source of estrogen is from con-
Use of estrogen replacement has raised concerns about their version of androgens to estrogens
therapy (unopposed estrogen ther- potential effects on cancer risk. by aromatase in adipose tissue. Risk
406
Fig. 5.11.2. A comparison of risk factors for type 1 and type 2 endometrial cancer, show- type 2 cancers. This has been at-
ing odds ratios with 95% confidence intervals, from the Epidemiology of Endometrial tributed to the fact that smokers tend
Cancer Consortium. BMI, body mass index; OC, oral contraceptive. to have lower endogenous estrogen
2 levels than non-smokers [9].
Type 1 Type 2
1.8 Medical conditions and use of
1.6 medication
1.4 Diabetes and metabolic
1.2 syndrome
1 Metabolic syndrome describes a
0.8 cluster of related metabolic condi-
0.6
tions, including abdominal obesity,
high blood pressure, impaired fast-
0.4
ing glucose or diabetes, high lev-
0.2 els of serum triglycerides, and low
Per 2 kg/m2
Yes
< 11
Current
11–12
13–14
15+
0
1
2
3
4+
Ever
Never
Former
levels of high-density lipoprotein;
the presence of three of these con-
ditions is sufficient for a diagnosis.
Of all of these conditions, obesity is
BMI Diabetes Age at Parity OC use Smoking most strongly associated with risk
menarche of endometrial cancer, but metabol-
ic syndrome, impaired glucose tol-
erance or diabetes, and hyperten-
of endometrial cancer increases by Diet sion appear to increase risk by an
about 50% for every increase of 5 kg/ Data from prospective studies sug- additional 20–40%, independently
m2 in body mass index (BMI), with gest that a diet with a high glycaemic of any underlying obesity [19].
stronger associations seen for type 1 load probably increases risk of endo-
cancers than for type 2 cancers Endometriosis, polycystic ovary
metrial cancer by approximately 15% syndrome, and fibroids
(Fig. 5.11.2). However, the relation-
per 50 units per day, whereas con-
ship is nonlinear and risk increases The relationship between a history
sumption of coffee (caffeinated and
more steeply at higher BMI (risks for of endometriosis and risk of endo-
decaffeinated) reduces risk by ap-
BMI of 30, 35, and 40 kg/m2 are ap- metrial cancer is not clear, but there
proximately 7% per cup per day [18].
proximately 2-, 4-, and 13-fold those is significant genetic overlap be-
Although previous reports suggested
for BMI of 20 kg/m2) [17]. The effect tween the two conditions, suggest-
a possible positive association with
is stronger among premenopausal ing that women who are genetically
intake of red meat and an inverse as-
women and those who have not used predisposed to developing endome-
sociation with intake of non-starchy
menopausal hormone therapy. triosis may also be at increased risk
vegetables, the current data do not
Similar patterns are seen for of endometrial cancer (see Chapter
other measures of obesity, includ- support this, and there is little evi-
3.5) [20]. Other conditions, including
ing waist circumference, hip cir- dence that other components of diet, polycystic ovary syndrome and fi-
cumference, waist-to-hip ratio, and including fat, fibre, or soy products, broids, have been more consistently
weight gain in adulthood. Greater which contain phytoestrogens, play associated with risk of endometrial
height has also been associated an independent role in the etiology of cancer, possibly because both con-
with greater risk, but it is unlikely endometrial cancer [18]. ditions are associated with elevated
CHAPTER 5.11
SECTION 5
that this is a causal relationship; estrogen levels.
Alcohol consumption and
rather, adult height is probably a
tobacco smoking Common medications
marker for a range of other genetic
factors and non-genetic factors Although alcohol intake has been There has been much interest in the
(e.g. nutritional status, hormones) associated with higher estrogen potential chemopreventive effects
before and around menarche [18]. levels and with an increased risk of non-steroidal anti-inflammatory
Independent of its effect on of breast cancer (see Chapter 5.9), drugs (NSAIDs) and of medications
obesity, there is now evidence that there is little evidence to suggest used to treat diabetes, specifically
physical activity of all types (rec- that moderate consumption increas- metformin, and hypercholesterol-
reational, occupational, and house- es risk of endometrial cancer [18]. aemia, namely statins. Regular use –
hold) probably reduces risk of en- Endometrial cancer is one of the usually defined as at least once per
dometrial cancer, and a suggestion few conditions that is less common week – of aspirin and, potentially,
that more time spent sedentary may among smokers, with inverse asso- other NSAIDs has been associated
increase risk [18]. ciations reported for both type 1 and with a reduced risk of endometrial

cancer among obese women; little Fig. 5.11.3. Population attributable fractions for endometrial cancer in the United
effect was seen for normal-weight Kingdom and Australia for overweight and obesity, insufficient physical activity (exer-
women [21]. It is less clear whether cise), and use of menopausal hormone therapy (MHT).
any association is restricted to stand-
ard-dose aspirin or whether use of
low-dose formulations may also con-
United
fer a benefit. An effect is plausible be- Kingdom
cause both aspirin and other NSAIDs
inhibit cyclooxygenase (COX) activ-
Obesity
ity, leading to a reduction in prosta-
glandin levels, and COX inhibitors Exercise
also downregulate aromatase activity Australia MHT
in breast cancer cell lines.
Use of metformin has been re-
ported to reverse endometrial hy-
perplasia, the precursor of type 1 0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00
endometrial cancer, but the current Population attributable fraction (%)
data are very heterogeneous [22].
Although use of statins at baseline
was associated with a significantly
contraceptives prevents approxi- loss among women who are already
reduced risk of endometrial cancer
mately 31% of endometrial cancers obese, have the greatest potential to
during follow-up of the Women’s
[24] and that in high-income coun- reduce risk of endometrial cancer. It
Health Initiative cohorts, there was
tries, it prevented approximately is also likely that increasing physical
no association when information
200  000 endometrial cancers in activity levels would have a benefi-
about statin use was updated during
women younger than 75 years in cial effect, both independently and
follow-up [23]. Use of bisphospho-
the 10 years from 2005 to 2014 [13]. through the effects of exercise on
nates, which are used to treat osteo-
body weight (see Chapter 6.1).
porosis, has also been associated
with reduced risk of endometrial Prevention However, preventing the one
cancer. The heterogeneous results, third of endometrial cancers attri-
The most effective way to prevent
the challenges of interpreting obser- butable to obesity would require all
endometrial cancer is surgery to
vational data on use of medications women to achieve and maintain a
remove the uterus (hysterectomy),
because they may be subject to healthy weight – a highly implausi-
and this is an option for women at
confounding by indication, and the ble scenario. Under more plausible
high risk who have completed their
lack of trial data mean that further family. Greater screening for Lynch weight-loss scenarios, the numbers
evidence is required before firm con- syndrome, for example by testing of cases prevented would be much
clusions can be drawn about any po- all those diagnosed with colorectal lower. For example, an study in
tential benefits of these medications. cancer or endometrial cancer and Australia estimated that if the pro-
cascade testing of family members, portion of women who are obese
Population attributable risks would identify more carriers of decreased by 10% every year for
Estimates from the United Kingdom, high-risk mutations. However, there 10 years and the proportion who
Australia, and globally suggest that is currently no screening test for en- are overweight decreased by 5%
between 30% and 40% of endometri- dometrial cancer that could be used every year for 10 years, this would
al cancers can be attributed to poten- in this group (although regular colo- prevent 11–18% of endometrial
tially modifiable factors (Fig. 5.11.3). rectal cancer screening would re- cancers over a 25-year period [26].
The greatest proportion is attribu- duce their risk of colorectal cancer).
table to overweight and obesity
Chemoprevention
(26–34%), and smaller proportions Behaviour change Use of oral contraceptives cannot
are attributable to physical inactivity There is increasing evidence that be widely recommended for preven-
(4–6%) and use of menopausal hor- intentional weight loss greatly re- tion of endometrial cancer, because
mone therapy (1–3%) [24]. duces risk of endometrial cancer current users are at increased risk of
Given that more recent data (Fig. 5.11.4), and benefits are also breast cancer. Progestin-containing
suggest additional protection from seen for those who undergo bar- intrauterine devices (e.g. the levonor-
breastfeeding, it is possible that iatric surgery [25]. Therefore, inter- gestrel-releasing intrauterine system),
more cancers could be prevented ventions that reduce the prevalence which supply hormones directly to
if all parous women breastfed their of obesity, whether by prevent- the gynaecological tract, might pro-
children for at least 6 months. It ing young women from becoming vide similar gynaecological protec-
has been estimated that use of oral obese or by encouraging weight tion without increasing risk of breast
408
Fig. 5.11.4. Hazard ratios and 95% confidence intervals for the association between cancer. An early report suggests
weight change and risk of endometrial cancer in women who gained or lost at least that they also increase risk of breast
10 pounds (4.54 kg) over a 3-year period, from the Women’s Health Initiative observa- cancer [14], but the current data are
tional study. Weight loss is further separated into intentional and unintentional weight loss.
restricted to women younger than 55
1.6 years so they do not capture the full
range of risks and benefits, which, for
1.4 endometrial cancer, are not likely to
be apparent until users reach their
Hazard ratio (95% CI)
1.2 60s and 70s, when endometrial can-

1 cer is more common.
Trials are under way to assess
0.8 whether progestin-containing intra
uterine devices, and also metformin,
0.6
could be used to prevent endometri-
0.4 al cancer in very obese women, who
are at greatest risk. If the inverse as-
0.2 sociation between use of aspirin or
other NSAIDs and risk of endome-
0
Stable Gain Loss (Intentional) (Unintentional) trial cancer in obese women is con-
firmed, this could provide another
Weight change from baseline to 3 years opportunity for prevention.
References
1. Lortet-Tieulent J, Ferlay J, Bray F, Jemal A 6. O’Mara TA, Glubb DM, Amant F, Annibali D, 11. Jordan SJ, Na R, Johnatty SE, Wise
(2018). International patterns and trends in Ashton K, Attia J, et al. (2018). Identification LA, Adami HO, Brinton LA, et al. (2017).
endometrial cancer incidence, 1978–2013. J of nine new susceptibility loci for endo- Breastfeeding and endometrial cancer
Natl Cancer Inst. 110(4):354–61. https://doi. metrial cancer. Nat Commun. 9(1):3166. risk: an analysis from the Epidemiology of
org/10.1093/jnci/djx214 PMID:29045681 ht t p s: //d o i .o r g /10.10 3 8 /s 414 67- 018 - Endometrial Cancer Consortium. Obstet
05427-7 PMID:30093612 Gynecol. 129(6):1059–67. https://doi.
2. Global Burden of Disease Cancer org/10.1097/AOG.0000000000002057
Collaboration (2017). Global, regional, 7. Getz G, Gabriel SB, Cibulskis K, Lander PMID:28486362
and national cancer incidence, mortality, E, Sivachenko A, Sougnez C, et al.;
years of life lost, years lived with disabili- 12. Yang HP, Cook LS, Weiderpass E,
Cancer Genome Atlas Research Network
ty, and disability-adjusted life-years for 32 Adami HO, Anderson KE, Cai H, et al.
(2013). Integrated genomic characteriza-
cancer groups, 1990 to 2015: a systematic (2015). Infertility and incident endo-
tion of endometrial carcinoma. Nature. metrial cancer risk: a pooled analysis
analysis for the Global Burden of Disease 497(7447):67–73. https://doi.org/10.1038/
Study. JAMA Oncol. 3(4):524–48. https:// from the Epidemiology of Endometrial
nature12113 PMID:23636398 Cancer Consortium (E2C2). Br J Cancer.
doi.org/10.1001/jamaoncol.2016.5688
112(5):925–33. https://doi.org/10.1038/
PMID:27918777 8. McAlpine J, Leon-Castillo A, Bosse T
bjc.2015.24 PMID:25688738
(2018). The rise of a novel classification
3. Faber MT, Frederiksen K, Jensen A,
system for endometrial carcinoma; inte- 13. Collaborative Group on Epidemiological
Aarslev PB, Kjaer SK (2017). Time
gration of molecular subclasses. J Pathol. Studies on Endometrial Cancer (2015).
trends in the incidence of hysterectomy-
244(5):538–49. https://doi.org/10.1002/ Endometrial cancer and oral contracep-
corrected overall, type 1 and type 2 en-
CHAPTER 5.11
tives: an individual participant meta-anal-
SECTION 5
path.5034 PMID:29344951
dometrial cancer in Denmark 1978–2014. ysis of 27 276 women with endometrial
Gynecol Oncol. 146(2):359–67. https:// 9. Setiawan VW, Yang HP, Pike MC, McCann cancer from 36 epidemiological studies.
d o i.o r g /10.1016 / j.yg y n o. 2 017.0 5 .015 SE, Yu H, Xiang YB, et al.; Australian Lancet Oncol. 16(9):1061–70. https://doi.
PMID:28545689 National Endometrial Cancer Study Group org /10.1016/S1470 -20 4 5(15) 0 0212- 0
4. Cote ML, Ruterbusch JJ, Olson SH, Lu K, (2013). Type I and II endometrial cancers: PMID:26254030
Ali-Fehmi R (2015). The growing burden of have they different risk factors? J Clin Oncol.
14. Soini T, Hurskainen R, Grénman S,
endometrial cancer: a major racial disparity 31(20):2607–18. https://doi.org/10.1200/
Mäenpää J, Paavonen J, Pukkala E (2014).
affecting Black women. Cancer Epidemiol JCO.2012.48.2596 PMID:23733771
Cancer risk in women using the levonorges-
Biomarkers Prev. 24(9):1407–15. https:// trel-releasing intrauterine system in Finland.
10. Setiawan VW, Pike MC, Karageorgi S,
doi.org/10.1158/1055-9965.EPI-15-0316 Obstet Gynecol. 124(2 Pt 1):292–9. https://
Deming SL, Anderson K, Bernstein L, et
PMID:26290568 doi.org/10.1097/AOG.0000000000000356
al.; Australian National Endometrial Cancer
PMID:25004338
5. Spurdle AB, Bowman MA, Shamsani J, Kirk Study Group (2012). Age at last birth in re-
J (2017). Endometrial cancer gene panels: lation to risk of endometrial cancer: pooled 15. Brinton LA, Felix AS (2014). Menopausal
clinical diagnostic vs research germline analysis in the Epidemiology of Endometrial hormone therapy and risk of endome-
DNA testing. Mod Pathol. 30(8):1048–68. Cancer Consortium. Am J Epidemiol. trial cancer. J Steroid Biochem Mol Biol.
https://doi.org/10.1038/modpathol.2017.20 176(4):269–78. https://doi.org/10.1093/aje/ 142:83–9. https://doi.org/10.1016/j.jsbmb.
PMID:28452373 kws129 PMID:22831825 2013.05.001 PMID:23680641

16. Skalkidou A, Sergentanis TN, Gialamas 20. Painter JN, O’Mara TA, Morris AP, Cheng 24. Whiteman DC, Webb PM, Green AC,
SP, Georgakis MK, Psaltopoulou T, THT, Gorman M, Martin L, et al. (2018). Neale RE, Fritschi L, Bain CJ, et al. (2015).
Trivella M, et al. (2017). Risk of endome- Genetic overlap between endometriosis Cancers in Australia in 2010 attributa-
trial cancer in women treated with ovary- and endometrial cancer: evidence from ble to modifiable factors: summary and
stimulating drugs for subfertility. Cochrane cross-disease genetic correlation and conclusions. Aust N Z J Public Health.
Database Syst Rev. 3:CD010931. https:// GWAS meta-analyses. Cancer Med. 39(5):477–84. https://doi.org/10.1111/1753-
doi.org/10.1002/14651858.CD010931.pub2 7(5):1978–87. https://doi.org/10.1002/ 6405.12471 PMID:26437735
PMID:28349511 cam4.1445 PMID:29608257
25. Luo J, Chlebowski RT, Hendryx M, Rohan
17. Aune D, Navarro Rosenblatt DA, Chan 21. Verdoodt F, Kjaer SK, Friis S (2017). T, Wactawski-Wende J, Thomson CA,
DS, Vingeliene S, Abar L, Vieira AR, et Influence of aspirin and non-aspirin NSAID et al. (2017). Intentional weight loss and
al. (2015). Anthropometric factors and use on ovarian and endometrial cancer: endometrial cancer risk. J Clin Oncol.
endometrial cancer risk: a systematic summary of epidemiologic evidence of 35(11):1189–93. https://doi.org/10.1200/
review and dose-response meta-anal- JCO.2016.70.5822 PMID:28165909
cancer risk and prognosis. Maturitas.
ysis of prospective studies. Ann Oncol.
100:1–7. https://doi.org/10.1016/j.maturitas.
26(8):1635–48. https://doi.org/10.1093/ 26. Wilson LF, Baade PD, Green AC, Jordan
2017.03.001 PMID:28539172
annonc/mdv142 PMID:25791635 SJ, Kendall BJ, Neale RE, et al. (2019). The
22. Chu D, Wu J, Wang K, Zhao M, Wang C, impact of changing the prevalence of over-
18. WCRF/AICR (2018). Diet, nutrition, physical weight/obesity and physical inactivity in
Li L, et al. (2018). Effect of metformin use
activity and endometrial cancer. Continuous Australia: an estimate of the proportion of
Update Project Expert Report 2018. World on the risk and prognosis of endometrial
cancer: a systematic review and meta- potentially avoidable cancers 2013–2037.
Cancer Research Fund/American Institute
analysis. BMC Cancer. 18(1):438. https:// Int J Cancer. 144(9):2088–98. https://doi.
for Cancer Research. Available from:
d o i.o r g /10.118 6 /s12 8 8 5 - 018 - 4 3 3 4 - 5 org/10.1002/ijc.31943 PMID:30357816
https://www.wcrf.org/sites/default/files/
Endometrial-cancer-report.pdf. PMID:29669520
19. Trabert B, Wentzensen N, Felix AS, Yang 23. Desai P, Wallace R, Anderson ML, Howard
HP, Sherman ME, Brinton LA (2015). BV, Ray RM, Wu C, et al. (2018). An anal-
Metabolic syndrome and risk of endome- ysis of the association between statin
trial cancer in the United States: a study use and risk of endometrial and ovarian
in the SEER-Medicare linked database. cancers in the Women’s Health Initiative.
Cancer Epidemiol Biomarkers Prev. Gynecol Oncol. 148(3):540–6. https://
24(1):261–7. https://doi.org/10.1158/1055- doi.org /10.1016 / j.ygyno. 2018 .01.0 0 6
9965.EPI-14-0923 PMID:25587111 PMID:29422345
410
5.12
Ovarian cancer
Complicated etiology and very few
preventive options
Renée Turzanski Fortner Ronny Drapkin (reviewer)
Rudolf Kaaks Anita Koushik (reviewer)
common cause of cancer death clear cell tumours are generally

SUMMARY in women worldwide, and the fifth high-grade but with intermediate
most common cause of cancer prognosis.
●● Accumulating evidence sug- death in women in Australia, North The cell of origin of ovarian car-
gests that the majority of “ovar- America, and western Europe. cinomas has been a topic of long-
ian” carcinomas are of extra- In high-income countries, more standing controversy. It was long
ovarian origin, originating in the than 90% of ovarian cancers are held that ovarian carcinomas devel-
fallopian tube from serous tubal carcinomas (i.e. derived from epi- op through neoplastic transformation
intraepithelial carcinomas for thelial cells), and the remainder are of the ovarian surface epithelium,
tumours with serous histotype, germ cell tumours and sex cord favoured by the repeated rupture
from endometrial cells for endo- stromal tumours. The vast majority and repair of the surface epithelium
metrioid and clear cell tumours,
of ovarian neoplasms are invasive through successive menstrual cycles
and from the gastrointestinal
carcinomas; 10–15% are classified (the “incessant ovulation” hypothesis)
mucosa and tubal-peritoneal
as borderline tumours, which pres- [1]. This theory received support from
junction for mucinous tumours.
ent without invasion into the stroma. epidemiological observations that the
●● To date, there are no effective This chapter focuses on invasive risk of ovarian cancer is significantly
early detection methods, espe- carcinomas. lower in women with a lower cumula-
cially for the aggressive disease On the basis of tumour histol- tive number of lifetime ovulatory cy-
subtypes, although preliminary ogy and grade, epithelial ovarian cles as a result of high parity or use
results with markers based on carcinomas are classified into his- of oral contraceptives.
detection of tumour DNA in tis- topathological subtypes, or histo- The theory implied that all ovar-
sue close to the ovary, for exam- types, with diverse somatic mutation ian carcinomas should have a
ple using Pap or Tao brushes, profiles, responses to chemother- common cell of origin in the ovar-
hold promise. apy, and prognosis (Table 5.12.1) ian surface epithelium, which is
and diverse risk factors (Table mesothelial in origin, but provided
●● Primary prevention of ovarian
cancer remains a challenge, 5.12.2). These histotypes are con- no direct explanation for the his-
given that the disease has rel- sidered distinct diseases. The five tological diversity of ovarian carci-
CHAPTER 5.12
SECTION 5
atively few known modifiable major subtypes are high-grade se- nomas, or for their resemblance to
risk factors, particularly for the rous (~60%), endometrioid (~10%), tumours arising in organs that are
predominant, and lethal, high- clear cell (~10%), mucinous (~3%), embryologically derived from the
grade serous subtype. Bilateral and low-grade serous (< 5%) ovar- Müllerian ducts, such as the fallo-
salpingectomy is of increasing ian carcinomas. High-grade serous pian tubes, endometrium, and va-
interest for prevention, includ- carcinomas are poorly differenti- gina. Furthermore, extensive path-
ing in women at average risk ated tumours with high response ological searches generally failed
who are undergoing steriliza- to chemotherapy but very poor to identify convincing precursor le-
tion or hysterectomy. survival. In contrast, endometrioid, sions within the ovaries.
low-grade serous, and mucinous There is a growing consensus
carcinomas are generally well-dif- that ovarian carcinomas derive large-
Ovarian cancer is frequently ag- ferentiated low-grade tumours with ly from cells originating in extra-ovar-
gressive and is generally detected lower response to chemotherapy ian tissue. This paradigm shift re-
at a late stage. It is the eighth most but favourable prognosis, whereas garding the origins of ovarian cancer
Chapter 5.12 • Ovarian cancer 411

has major implications for strategies associations between both tumour

for prevention and early detection. types and endometriosis as well
Extensive evidence has ac- as similarities in molecular genetic
FUNDAMENTALS
cumulated that a large proportion profiles of endometrioid and clear
■■ Ovarian cancer is the eighth
of high-grade serous carcinomas cell carcinomas and contiguous en-
most common cause of
develop from fallopian tube epithe- dometriotic cysts [12].
The origins of mucinous ovarian cancer death in women
lium, a tissue with morphology and
genetic and immunohistochemical carcinomas are perhaps least well worldwide, with the highest
expression profiles that are similar to understood. These tumours are incidence rates in Europe and
those of high-grade serous tumours. hypothesized to originate from the North America. The disease
Putative precursor lesions for high- gastrointestinal mucosa or transi- occurs predominantly in
grade serous tumours, called serous tional-type epithelium at the tubal- postmenopausal women.
tubal intraepithelial carcinomas, were peritoneal junction [8].
■■ There are multiple major
first identified at the fimbriated end
of fallopian tubes removed prophy- subtypes of ovarian cancer.
Epidemiology Risk factor associations,
lactically from high-risk women car-
Ovarian cancer is a relatively rare responses to therapy,
rying BRCA1 or BRCA2 mutations
cancer, with an estimated 295 414 and survival differ by
[2], whereas similar lesions were not
new cases (3.4% of all incident can-
found in the ovaries. Subsequent histopathological subtype.
cers in women) worldwide in 2018
studies identified serous tubal in-
[13]. However, it is a lethal malig- ■■ The cell type or types that give
traepithelial carcinomas in 50–60%
nancy, because it is predominantly rise to ovarian cancer remain
of women with sporadic ovarian car-
diagnosed at a late stage. Incidence an area of active investigation,
cinomas [3] and showed that tumour-
rates vary by region; the lowest and include cells originating in
specific molecular features such as
rates (4.7 per 100 000) are ob- extra-ovarian tissue. Besides
DNA mutation patterns were mostly
served in the WHO African Region,
shared between serous tubal intraep- germline genetic variants, the
and the highest rates (9.1 per
ithelial carcinomas and concurrent major ovarian cancer subtypes
100 000) are observed in the WHO
high-grade serous carcinomas, im- exhibit distinct sets of recurrent
European Region (Fig. 5.12.1). Mor
plying that serous tubal intraepithelial somatic mutations.
tality rates also vary across the
carcinomas are the origin [4,5].
world (Fig. 5.12.2). ■■ Epidemiological evidence
Although up to 70% of high-
In general, incidence rates have
grade serous carcinomas potential- implicates pharmacological
been stable over recent decades,
ly arise from fallopian tube fimbria, use of steroids and other risk
with slight decreases noted in North
a subset may also derive from tube- factors, but opportunities for the
America and areas of western and
like epithelium found outside the primary prevention of ovarian
northern Europe, and increases ob-
fallopian tube, from small cortical cancer are extremely limited.
served in parts of eastern Europe
inclusion cysts that are found on the
(Latvia and Poland) [14] (Fig. 5.12.3). ■■ Ovarian cancer is frequently
ovarian surface and that are lined
Invasive serous carcinomas are the
with tubal-type epithelium. It is still aggressive and is generally
predominant histotype worldwide.
debated whether the tubal-like epi- detected at a late stage.
However, there is regional variation
thelium in these cysts derives from
in the distribution of ovarian tumours
implantation of tubal epithelium
by histotype, with a higher propor-
(i.e. endosalpingiosis) or is formed
tion of clear cell carcinomas and a than 3-fold [15], with an elevated
through metaplasia of the ovarian
lower proportion of serous carcino- risk of all except the invasive mu-
surface epithelium, or both [6,7].
mas in countries in Asia, relative to cinous histotype [17]. A family his-
Invasive low-grade serous tu-
other regions [14]. tory of breast cancer is also as-
mours are also thought to derive
from fallopian tube tissue, devel- sociated with an increased risk of
oping stepwise from benign hyper- Genetics and genomics ovarian cancer; the cumulative risk
plastic lesions referred to as atypi- Germline BRCA1 and BRCA2 mu- to age 80 years is 44% for carri-
cal proliferative serous tumours tations are observed in up to 15% ers of BRCA1 mutations and 17%
(also known as serous borderline of patients with invasive ovarian for carriers of BRCA2 mutations
tumours) [8]. In contrast, endome- cancers overall, and up to 23% of [18]. For the endometriosis-related
trioid and clear cell carcinomas are patients with high-grade serous (endometrioid and clear cell) carci-
both thought to arise from endome- disease [15,16]. For women with nomas, risk is increased in women
trial tissue cells. Results from clin- a family history of ovarian cancer with Lynch syndrome (also called
icopathological [9] and epidemio- in a first-degree relative, the risk hereditary non-polyposis colorec-
logical studies [10,11] have shown of the disease is increased more tal cancer), which is characterized
412
Table 5.12.1. Characteristics of the main histotypes of invasive ovarian carcinoma
Characteristic Histotype
High-grade serous Low-grade serous Mucinous Endometrioid Clear cell
Possible tissues of Fallopian tube fimbria; Endosalpingiosis; Endometriosis or Endometriosis; endometrioid adenofibroma
origin ovarian cortical papillary tubal tubal-peritoneal
inclusion cysts hyperplasia junction
Possible cells of Fallopian tube secre- Fallopian tube secre- Unknown Endometrial epithelial cell
origin tory or epithelial cell, tory or epithelial cell,
or progenitor cell or progenitor cell
Precursor lesion Serous tubal Serous borderline Mucinous borderline Endometrioid borderline tumour
intraepithelial tumour/atypical tumour; cystadenoma;
carcinoma, p53 proliferative serous Brenner tumour
signature tumour
Familial/genetic risk Germline mutations Not applicable Not applicable Lynch syndrome (germline mutations in MLH1,
in BRCA1, BRCA2, MSH2, MSH6, PMS2)
BRIP1, PALB2,
RAD51C, RAD51D
Frequent somatic TP53, BRCA1, BRAF, KRAS, NRAS, KRAS ARID1A, PTEN, HNF1β, ARID1A,
mutations BRCA2; copy number HRAS, ERBB2 CTNNB1 (β-catenin), PIK3CA
alterations of CCNE1; PIK3CA, KRAS;
PTEN deletion; loss of mismatch repair
RB1, NF1 defects, microsatellite
instability
Proliferation High Low Intermediate Low Low
Prognosis Poor Favourable Favourable Favourable Intermediate
by germline mutations in genes (rarely observed in other histotypes) vided into three subtypes showing
involved in DNA mismatch repair: and also show associations with different mutation load and DNA
MLH1, MSH2, MSH6, and PMS2. activating mutations of PIK3CA or mismatch repair signatures: ultra-
Besides germline genetic vari- loss-of-function alterations in PTEN. mutator, microsatellite instable, and
ants, the major ovarian carcinoma Endometrioid tumours specifically microsatellite stable; the microsat-
histotypes are associated with dis- may also show KRAS mutations. ellite stable group has a high pro-
tinct sets of recurrent somatic mu- More extensive analyses of portion of CTNNB1 (β-catenin) and
tations and defects in DNA repair. ovarian tumour histotypes by whole- KRAS mutations (Table 5.12.1).
High-grade serous tumours are genome sequencing also identified
ubiquitously characterized by inacti- structural genomic alterations re- Etiology
vating mutations in the TP53 gene, flecting specific DNA repair mecha-
often in combination with genomic nisms, which in combination with Established and putative risk
instability due to BRCA1 or BRCA2 mutation patterns form signatures factors
defects. Furthermore, a key mo- that further segregate tumours into In terms of non-genetic and poten-
lecular characteristic of high-grade distinct molecular and biological tially modifiable risk factors, recent
serous tumours is the presence of classes, both within and between studies increasingly have docu-
widespread copy number alterations histotypes [21]. Thus, high-grade mented distinct risk factor profiles
CHAPTER 5.12
SECTION 5
[19], including of CCNE1 (cyclin E1), serous tumours are distinguished by tumour histotype. However, the
and this histotype often also shows from non-serous tumours by loss etiology of sporadic invasive ovar-
defects in genes of the retinoblas- of heterozygosity and homologous ian cancer remains poorly under-
toma and Notch pathways [20]. recombination signatures, and are stood. Studies in large consortia
Invasive low-grade serous tu- further split into a subgroup en- have shown substantial heteroge-
mours are characterized by (mutu- riched in fold-back inversions and neity in the associations between
ally exclusive) sequence mutations a subgroup characterized by other well-established risk factors for
in the KRAS, BRAF, or ERBB2 on- types of genomic rearrangements. ovarian cancer, such as parity and
cogenes, and mucinous tumours are Clear cell tumours are divided use of oral contraceptives, and dis-
characterized by KRAS mutations. into subgroups characterized by ease risk by histotype [11,15,22].
The endometriosis-associated ovar- deamination of the APOBEC fam- For example, being parous is asso-
ian cancer (clear cell and endome- ily of cytidine deaminases or age- ciated with the largest reductions in
trioid) histotypes both show loss- related mutational signatures. The risk for clear cell and endometrioid
of-function mutations in ARID1A endometrioid tumours can be di- carcinoma (~50–65%) and is more

Fig. 5.12.1. Global distribution of estimated age-standardized (World) incidence rates (ASR) per 100 000 person-years for ovarian
cancer, 2018.
Fig. 5.12.2. Global distribution of estimated age-standardized (World) mortality rates (ASR) per 100 000 person-years for ovarian
cancer, 2018.
modestly associated with risk of risk of serous and endometrioid use [23]. Findings on use of meno-
serous disease (~20%) [11]. Longer ovarian carcinomas [11,23]; higher pausal hormone therapy are com-
duration of use of oral contracep- risk is apparent even for short-term plemented by recent studies on
tives is inversely associated with use (< 5 years) and for estrogen- circulating endogenous estrogens
risk of serous, endometrioid, and only and combined estrogen–pro- and androgens, which have shown
clear cell carcinomas (~15–20% gestogen formulations (see Chapter higher risks of non-serous ovarian
lower risk per 5 years of use) but not 2.11) [23]. The increase in risk as- cancer subtypes with higher blood
of mucinous carcinomas [11,22]. sociated with use of menopausal concentrations of both estrogens
Ever or current/recent use of hormone therapy wanes with time and androgens [24,25].
menopausal hormone therapy is since cessation of use; however, Tubal ligation is associated with
associated with a 40–70% higher this may be dependent on length of an approximately 50% reduction in
414
Fig. 5.12.3. Trends in age-adjusted ovarian cancer incidence rates per 100 000 person-years by region and country from 1973–1977
to 2003–2007, from Volumes IV–X of Cancer Incidence in Five Continents.
risk of endometrioid and clear cell body power, sexually transmitted The sexually transmitted infec-
ovarian cancer [11,26]; one pooled infections, and pelvic inflammatory tion Chlamydia trachomatis was
analysis also indicated a more mod- disease, and use of anti-inflamma- recently associated with increased
est 20% reduction in risk of serous tory analgesics may affect risk of risk of ovarian cancer [29], al-
disease and a 32% reduction in ovarian cancer. though the results to date on sexu-
risk of mucinous carcinomas [26], Perineal use of talc-based body ally transmitted infections are not
whereas a subsequent pooled analy- powder has been classified by the consistent. Infection with C. tra-
sis observed no association in these IARC Monographs as possibly car- chomatis may increase risk via
subtypes [11]. The subtype-specific cinogenic to humans (Group 2B). tubal pathologies induced by pelvic
associations are in line with the hy- However, experimental evidence inflammatory disease.
pothesis that tubal ligation reduces supporting an association is limited, Very frequent use of aspirin
disease risk by blocking “retrograde and prospectively collected data on (≥ 6 days per week) has been as-
menstruation” or reflux of endo- perineal talc exposure are sparse sociated with modest reductions
metrial tissue through the fallopian [28]. Prospective consortium-based in risk of ovarian cancer in both
tubes, and with the observation that studies are required to clarify this pooled case–control and prospec-
endometriosis, the result of ectopic association. tive studies [30,31]. However, the
uterine tissue in the peritoneal cav-
ity, increases the risk of endometri-
Fig. 5.12.4. Talcum powder. In relation to ovarian cancer, perineal use of talc-based
oid, clear cell, and low-grade serous body powder by women has been classified by the IARC Monographs as possibly
ovarian cancer [10,11]. carcinogenic to humans, but there is limited evidence of an association and prospective
Relatively few classic lifestyle studies to clarify the association are sparse.
exposures are associated with risk
of ovarian cancer. Higher body
mass index is associated with mod-
CHAPTER 5.12
SECTION 5
est increases in the risk of muci-
nous carcinomas (~8–15% increase
in risk per 5 kg/m2) and endometri-
oid carcinomas (~8% increase per
5 kg/m2) [11,27]. The available data
do not support strong associations
between diet or physical activity
and risk of ovarian cancer.
Emerging and possible risk

factors
Emerging evidence suggests that
inflammation-related exposures,
including perineal use of talc-based

Table 5.12.2. Associations between established and putative risk factors for ovarian cancer and risk of invasive epithelial ovarian
cancer, by histologya
Risk factor Histology
Serous Endometrioid Clear cell Mucinous
Non-modifiable exposures
Family history of ovarian cancer ↑↑↑↑ ↑↑↑↑ ↑↑↑↑ –
Age at menarche, per year increase – – ↓ –
Age at menopause, per year increase ↑ ↑ ↑↑ –
Endometriosis ↑↑↑↑b ↑↑↑↑ ↑↑↑↑ –
Lifestyle and anthropometric exposures
Parity, per child ↓ ↓ ↓↓ ↓
Use of oral contraceptives, per 5-year duration ↓ ↓ ↓ –
Use of menopausal hormone therapy, ever or ↑↑↑ ↑↑↑ – ↓

current/recent versus never
Tubal ligation ↓ ↓↓↓ ↓↓↓ –
Body mass index, per 5 kg/m 2

– ↑ – ↑
Height, per 5 cm ↑c
Smoking, current versus never – ↓ ↓ ↑↑
a
Relative risks: ↑, > 1.0 to 1.25; ↑↑, 1.25–1.5; ↑↑↑, 1.5–2.0; ↑↑↑↑, > 2.0; ↓, 0.80 to < 1.0; ↓↓, 0.70–0.80; ↓↓↓, 0.50–0.70; ↓↓↓↓, < 0.50.
b
Low-grade serous carcinomas only.
c
No significant heterogeneity by histology.
effects of long-term aspirin use (i.e. ty (in the Prostate, Lung, Colorectal als suggest that transvaginal ul-
≥ 10 years) and use of other analge- and Ovarian Cancer Screening trasound (with or without CA125)
sics (e.g. acetaminophen) and dif- study, in the USA [32]) or only a may generally be more effective for
ferential effects by histotype are not small (15%) and statistically non- early detection of more indolent tu-
well described. significant reduction when longitu- mours than for the more aggressive
dinal changes in CA125 were con- high-grade serous carcinomas [35].
sidered (in the Collaborative Trial of A proportion of the less aggressive
Ovarian Cancer Screening, in the tumours detected early by trans-
and early detection United Kingdom) [33]. Detailed anal- vaginal ultrasound may include dis-
In view of the low absolute incidence yses of data from these studies and ease that would not have been clini-
rates of ovarian cancer, screening some smaller trials indicated limited cally diagnosed if screening had not
tools must have very high specific- sensitivity of ovarian cancer detec- taken place (i.e. overdiagnosis), or
ity to avoid unnecessary intervention for both CA125 and transvaginal caused symptoms or mortality.
tions in false-positive cases, while ultrasound [34]. In population cohort To reduce mortality, screening
providing good detection sensitivity studies, analyses of blood samples tools should aim to more specifi-
for early-stage, curable disease. So collected at different lag times be- cally detect aggressive tumours at
far, ovarian cancer screening strat- fore diagnosis under usual care also a localized stage. A novel class of
egies have been based on blood- suggest limited sensitivity of CA125 promising biomarkers is cell-free
based biomarkers combined with and other candidate markers (e.g. tumour DNA or tumour cells from
transvaginal ultrasound imaging. HE4 and CA72-4) for detection of blood or other body fluids and tissue
In randomized trials, multimodal early-stage disease. samples (see Chapter 6.7), referred
screening by transvaginal ultra- Furthermore, CA125 and trans- to as liquid biopsies (see “Liquid bi-
sound and CA125 – the best avail- vaginal ultrasound also have only opsy: a promising approach for early
able blood-based biomarker – re- limited specificity, for ovarian can- detection”). However, it remains un-
sulted in a shift towards an earlier cer in general and more particularly known at present whether liquid bi-
disease stage at diagnosis but pro- for the more aggressive tumour opsy tools, or any other biomarker,
vided either no reduction in mortali- subtypes. Data from screening tri- will be able to detect serous tubal
416
Liquid biopsy: a promising approach for early detection

Ovarian cancer is generally di- ovarian carcinomas but have also combined yielded an overall detec-
agnosed at advanced stages, for been reported in benign and pre- tion sensitivity of 63%.
which 5-year survival is about 30%. neoplastic gynaecological condi- These results show potential
Diagnosis at an earlier stage yields tions. Although the test as applied for mutation-based detection of
improved survival, even for aggres- is limited by a high false-positive gynaecological cancers, with tests
sive high-grade serous disease; by rate, the results from this study tailored to more aggressive, high-
stage at diagnosis, 5-year survival is demonstrate a proof of concept for grade tumours. If these early re-
84% for localized disease and 32% uterine lavage as a sampling meth- sults are confirmed, particularly for
for distant disease. However, only od, and further discovery studies early-stage cases, early detection
about 5% of cases of high-grade are under way. via Pap tests would be particularly
serous disease are diagnosed at an An early proof-of-principle study attractive, given the widespread
early stage [1]. This motivates con- applied massively parallel sequenc- use of this test in standard care for
tinued research into effective meth- ing to liquid Pap test samples from cervical cancer screening.
ods of early detection. patients with ovarian cancer and
Liquid biopsies of blood sam- showed a 41% detection rate (9 References
ples, or of samples collected of 22 cases detected) for known 1. Peres LC, Cushing-Haugen KL, Köbel
closer to the site of a potential ma- tumour-related mutations in 12 dif- M, Harris HR, Berchuck A, Rossing MA,
et al. (2019). Invasive epithelial ovarian
lignancy, are of mounting interest, ferent genes [3]. cancer survival by histotype and disease
with the promise of high specificity. More recently, a new test stage. J Natl Cancer Inst. 111(1):60–8.
The first studies using uterine la- (called PapSEEK) was used that h t t p s: //d o i .o r g /10 .10 9 3 / j n c i /d j y 071
PMID:29718305
vage and the Pap test have yielded incorporates assays for mutations
promising results towards the ear- in 18 genes plus an assay for an- 2. Maritschnegg E, Wang Y, Pecha N, Horvat
R, Van Nieuwenhuysen E, Vergote I, et
lier detection of ovarian cancer. euploidy. With this test, analyses al. (2015). Lavage of the uterine cavity
In a proof-of-concept study, of Pap brush samples detected for molecular detection of Müllerian duct
Maritschnegg et al. applied mas- 33% (81 of 245) of patients with carcinomas: a proof-of-concept study.
J Clin Oncol. 33(36):4293–300. https://
sively parallel sequencing to cell ovarian cancer and 34% (30 of 89)
d o i.o r g /10.12 0 0 /J C O. 2 015 . 61. 3 0 8 3
samples obtained by uterine la- of patients with early-stage dis- PMID:26552420
vage and observed a 60% ovar- ease (stages I and II), with a 1.4%
3. Kinde I, Bettegowda C, Wang Y,
ian cancer detection rate (18 false-positive rate in women with- Wu J, Agrawal N, Shih IeM, et al.
of 30 cases detected) [2]. The out cancer (10 of 714; specificity, (2013). Evaluation of DNA from the
predominant mutation detected ~99%) [4]. Intrauterine sampling Papanicolaou test to detect ovarian and
endometrial cancers. Sci Transl Med.
among cases was in TP53; these with a Tao brush increased the 5(167):167ra4. https://doi.org/10.1126/
mutations are ubiquitous in high- detection sensitivity to 45% (23 of scitranslmed.3004952 PMID:23303603
grade serous ovarian carcinomas. 51) of patients with ovarian cancer, 4. Wang Y, Li L, Douville C, Cohen JD, Yen
Among women with benign condi- with 100% specificity (0 positives TT, Kinde I, et al. (2018). Evaluation of
tions (e.g. ovarian cyst, fibroma, among 125 cancer-free controls). liquid from the Papanicolaou test and
other liquid biopsies for the detection
or secondary infertility), 30% (8 of Finally, in 83 patients with ovar- of endometrial and ovarian cancers.
27) tested positive for mutations, ian cancer for whom plasma was Sci Transl Med. 10(433):10. https://
predominantly in KRAS (6 of 8) [2]. available, circulating tumour DNA doi.org/10.1126/scitranslmed.aap8793
PMID:29563323
KRAS mutations are observed in was found in 43% of patients, and
CHAPTER 5.12
low-grade serous and mucinous plasma and Pap brush samples
SECTION 5
intraepithelial carcinomas or small- Prevention cessation of use [36] and is evident

volume tumours at a very early (i.e. Prevention strategies for ovarian in women at average risk as well as
microscopic) stage, before more cancer in women at average risk re- in high-risk populations (i.e. carriers
widespread dissemination. A paral- main elusive, given the limited num- of BRCA1 or BRCA2 mutations) (see
lel challenge is the development of ber of known modifiable risk factors. Chapter 6.5). Although these de-
radiological imaging methods with Increased use of oral contraceptives creases in the risk of ovarian cancer
higher sensitivity and specificity for represents a potential avenue for are compelling, they must be bal-
small-volume, aggressive tumours prevention, with a protective effect anced against the increased risks of
as complementary diagnostic tools. that persists for up 20 years after breast cancer, cervical cancer, and

cardiovascular events. In addition, menopausal at surgery, are under that study. Although the risk reduc-
further data are required on risk as- way in women at high risk [37]. tion for bilateral salpingectomy is
sociations based on contemporary Opportunistic salpingectomy with more modest than that for bilateral
oral contraceptive formulations. ovarian conservation (in lieu of tub- salpingo-oophorectomy, the proce-
In terms of other opportuni- al ligation) has been suggested as dure may be appropriate for women
ties for chemoprevention (also re- a risk-reducing measure in women at average risk of ovarian cancer, and
ferred to as preventive therapy; see at average risk who are undergoing has rates of operative complications
Chapter 6.4), emerging evidence hysterectomy or sterilization. In a as low as those reported for tubal
suggests an inverse association large registry-based study, a reduc- ligation or hysterectomy without sal-
between daily aspirin use and risk tion of up to about 65% in risk was pingectomy [39]. Furthermore, ovar-
of ovarian cancer [30,31]. However, reported for bilateral salpingectomy, ian conservation in women younger
additional studies are required to compared with 28% for tubal ligation than 65 years yields a survival ben-
weigh potential risks and benefits and 94% for hysterectomy with bilat- efit [40] and prevents early surgical
and to delineate target populations. eral salpingo-oophorectomy [38]; no menopause in women who are pre-
Use of menopausal hormone data on histotype were available in menopausal at surgery.
therapy and perineal use of talc-
based body powder are avoidable
Fig. 5.12.5. Women wait for consultation at a health centre in Buhigwe, United Republic
exposures. Although associations
of Tanzania. Some of the lowest incidence rates of ovarian cancer are observed in the
between body mass index and risk of WHO African Region.
ovarian cancer are modest, maintain-
ing a healthy body weight has well-
documented and widespread health
benefits. Furthermore, the recently
observed association between C.
trachomatis infection and risk of ovar-
ian cancer suggests potential novel
leads for primary prevention.
In carriers of BRCA1 or BRCA2
mutations, prophylactic oophorec-
tomy is recommended at a rela-
tively young age (< 35–40 years
for BRCA1 and < 40–45 years for
BRCA2) to reduce the risk of both
breast cancer and ovarian cancer.
Trials investigating salpingectomy
with delayed oophorectomy, thus
delaying surgical menopause and
its sequelae in women who are pre-
418
References
1. Fathalla MF (1971). Incessant ovulation – 10. Pearce CL, Templeman C, Rossing MA, 19. Cancer Genome Atlas Research
a factor in ovarian neoplasia? Lancet. Lee A, Near AM, Webb PM, et al.; Ovarian Network (2011). Integrated geno-
2(7716):163. https://doi.org/10.1016/S0140- Cancer Association Consortium (2012). mic analyses of ovarian carcinoma.
6736(71)92335-X PMID:4104488 Association between endometriosis and Nature. 474(7353):609–15. https://doi.
risk of histological subtypes of ovarian can- org/10.1038/nature10166 PMID:21720365
2. Piek JM, van Diest PJ, Zweemer RP, cer: a pooled analysis of case-control stud-
Jansen JW, Poort-Keesom RJJ, Menko ies. Lancet Oncol. 13(4):385–94. https:// 20. Bowtell DD, Böhm S, Ahmed AA, Aspuria
FH, et al. (2001). Dysplastic changes in doi.org/10.1016/S1470-2045(11)70404-1 P-J, Bast RC Jr, Beral V, et al. (2015).
prophylactically removed Fallopian tubes of PMID:22361336 Rethinking ovarian cancer II: reducing mor-
women predisposed to developing ovarian tality from high-grade serous ovarian can-
cancer. J Pathol. 195(4):451–6. https://doi. 11. Wentzensen N, Poole EM, Trabert B,
cer. Nat Rev Cancer. 15(11):668–79. https://
org/10.1002/path.1000 PMID:11745677 White E, Arslan AA, Patel AV, et al. (2016).
doi.org/10.1038/nrc4019 PMID:26493647
Ovarian cancer risk factors by histologic
3. Gilks CB, Irving J, Köbel M, Lee C, Singh subtype: an analysis from the Ovarian 21. Wang YK, Bashashati A, Anglesio MS,
N, Wilkinson N, et al. (2015). Incidental Cancer Cohort Consortium. J Clin Oncol. Cochrane DR, Grewal DS, Ha G, et al.
nonuterine high-grade serous carcino- 34(24):2888–98. https://doi.org/10.1200/
mas arise in the fallopian tube in most (2017). Genomic consequences of aber-
JCO.2016.66.8178 PMID:27325851 rant DNA repair mechanisms stratify ovar-
cases: further evidence for the tubal ori-
gin of high-grade serous carcinomas. Am 12. Dawson A, Fernandez ML, Anglesio M, ian cancer histotypes. Nat Genet. 49(6):
J Surg Pathol. 39(3):357–64. https://doi. Yong PJ, Carey MS (2018). Endometriosis 856–65. https://doi.org/10.1038/ng.3849
org/10.1097/PAS.0000000000000353 and endometriosis-associated cancers: PMID:28436987
PMID:25517954 new insights into the molecular mecha-
22. Beral V, Doll R, Hermon C, Peto R, Reeves
nisms of ovarian cancer development.
4. Ducie J, Dao F, Considine M, Olvera G; Collaborative Group on Epidemiological
Ecancermedicalscience. 12:803. https://
N, Shaw PA, Kurman RJ, et al. (2017). Studies of Ovarian Cancer (2008). Ovarian
d o i . o r g / 1 0 . 3 3 3 2 / e c a n c e r. 2 0 18 . 8 0 3
Molecular analysis of high-grade serous PMID:29456620 cancer and oral contraceptives: collabora-
ovarian carcinoma with and without as- tive reanalysis of data from 45 epidemio-
sociated serous tubal intra-epithelial car- 13. Bray F, Ferlay J, Soerjomataram I, Siegel logical studies including 23 257 women
cinoma. Nat Commun. 8(1):990. https:// RL, Torre LA, Jemal A (2018). Global with ovarian cancer and 87 303 controls.
doi.org /10.10 3 8 /s 414 67- 017- 01217- 9 cancer statistics 2018: GLOBOCAN es- Lancet. 371(9609):303–14. https://doi.
PMID:29042553 timates of incidence and mortality world- org /10.1016/S 014 0 - 673 6 (0 8) 6 0167-1
wide for 36 cancers in 185 countries. CA PMID:18294997
5. Labidi-Galy SI, Papp E, Hallberg D, Niknafs Cancer J Clin. 68(6):394–424. https://doi.
N, Adleff V, Noe M, et al. (2017). High org/10.3322/caac.21492 PMID:30207593 23. Beral V, Gaitskell K, Hermon C, Moser K,
grade serous ovarian carcinomas origi- Reeves G, Peto R; Collaborative Group on
nate in the fallopian tube. Nat Commun. 14. Coburn SB, Bray F, Sherman ME, Trabert Epidemiological Studies of Ovarian Cancer
8(1):1093. https://doi.org/10.1038/s41467- B (2017). International patterns and (2015). Menopausal hormone use and
017-00962-1 PMID:29061967 trends in ovarian cancer incidence, over- ovarian cancer risk: individual participant
all and by histologic subtype. Int J Cancer.
6. Auersperg N (2015). Article by Natalie meta-analysis of 52 epidemiological stud-
140(11):2451–60. https://doi.org/10.1002/
Banet and Robert J. Kurman: Two types ies. Lancet. 385(9980):1835–42. https://
ijc.30676 PMID:28257597
of ovarian cortical inclusion cysts: pro- doi.org/10.1016/S0140-6736(14)61687-1
posed origin and possible role in ovar- 15. Reid BM, Permuth JB, Sellers TA (2017). PMID:25684585
ian serous carcinogenesis; Int. J. Gynecol. Epidemiology of ovarian cancer: a review.
Pathol. 2015;34:3–8. Int J Gynecol Pathol. Cancer Biol Med. 14(1):9–32. https://doi. 24. Ose J, Poole EM, Schock H, Lehtinen M,
34(3):303–4. https://doi.org/10.1097/PGP. org/10.20892/j.issn.2095-3941.2016.0084 Arslan AA, Zeleniuch-Jacquotte A, et al.
0000000000000202 PMID:25844551 PMID:28443200 (2017). Androgens are differentially as-
sociated with ovarian cancer subtypes in
7. Park KJ, Patel P, Linkov I, Jotwani A, 16. Alsop K, Fereday S, Meldrum C, deFazio the Ovarian Cancer Cohort Consortium.
Kauff N, Pike MC (2018). Observations A, Emmanuel C, George J, et al. (2012). Cancer Res. 77(14):3951–60. https://doi.
on the origin of ovarian cortical inclusion BRCA mutation frequency and patterns org/10.1158/0008 - 5 472.CAN -16 -3322
cysts in women undergoing risk-reducing of treatment response in BRCA mutation- PMID:28381542
salpingo-oophorectomy. Histopathology. positive women with ovarian cancer: a re-
72(5):766–76. https://doi.org/10.1111/his. port from the Australian Ovarian Cancer 25. Trabert B, Brinton LA, Anderson GL,
13444 PMID:29197096 Study Group. J Clin Oncol. 30(21):2654–63. Pfeiffer RM, Falk RT, Strickler HD, et
CHAPTER 5.12
SECTION 5
https://doi.org/10.1200/JCO.2011.39.8545 al. (2016). Circulating estrogens and
8. Kurman RJ, Carcangiu ML, Herrington PMID:22711857 postmenopausal ovarian cancer risk
CS, Young RH, editors (2014). WHO clas-
in the Women’s Health Initiative ob-
sification of tumours of female reproduc- 17. Zheng G, Yu H, Kanerva A, Försti
servational study. Cancer Epidemiol
tive organs. Lyon, France: International A, Sundquist K, Hemminki K (2018).
Biomarkers Prev. 25(4):648–56. https://
Agency for Research on Cancer (WHO Familial risks of ovarian cancer by age
doi.org/10.1158/1055-9965.EPI-15-1272-T
Classification of Tumours series, 4th edi- at diagnosis, proband type and histol-
tion). Available from: https://publications. PMID:26908437
ogy. PLoS One. 13(10):e0205000. https://
iarc.fr/16. doi.org/10.1371/journal.pone.0205000 26. Sieh W, Salvador S, McGuire V, Weber RP,
PMID:30281663 Terry KL, Rossing MA, et al.; Australian
9. Veras E, Mao TL, Ayhan A, Ueda S, Lai
H, Hayran M, et al. (2009). Cystic and 18. Kuchenbaecker KB, Hopper JL, Barnes Cancer Study (Ovarian Cancer);
adenofibromatous clear cell carcinomas DR, Phillips K-A, Mooij TM, Roos-Blom Australian Ovarian Cancer Study Group;
of the ovary: distinctive tumors that dif- M-J, et al.; BRCA1 and BRCA2 Cohort Ovarian Cancer Association Consortium
fer in their pathogenesis and behavior: a Consortium (2017). Risks of breast, ovarian, (2013). Tubal ligation and risk of ovar-
clinicopathologic analysis of 122 cases. and contralateral breast cancer for BRCA1 ian cancer subtypes: a pooled analysis
Am J Surg Pathol. 33(6):844–53. https:// and BRCA2 mutation carriers. JAMA. of case-control studies. Int J Epidemiol.
doi.org/10.1097/PAS.0b013e31819c4271 317(23):2402–16. https://doi.org/10.1001/ 42(2):579–89. https://doi.org/10.1093/ije/
PMID:19342944 jama.2017.7112 PMID:28632866 dyt042 PMID:23569193

27. Collaborative Group on Epidemiological 32. Buys SS, Partridge E, Black A, Johnson 37. Harmsen MG, Arts-de Jong M,
Studies of Ovarian Cancer (2012). Ovarian CC, Lamerato L, Isaacs C, et al.; PLCO Hoogerbrugge N, Maas AHEM, Prins JB,
cancer and body size: individual participant Project Team (2011). Effect of screening Bulten J, et al. (2015). Early salpingectomy
meta-analysis including 25,157 women with on ovarian cancer mortality: the Prostate, (TUbectomy) with delayed oophorectomy
ovarian cancer from 47 epidemiological Lung, Colorectal and Ovarian (PLCO) to improve quality of life as alternative for
studies. PLoS Med. 9(4):e1001200. https:// Cancer Screening randomized con- risk-reducing salpingo-oophorectomy in
doi.org/10.1371/journal.pmed.1001200 trolled trial. JAMA. 305(22):2295–303. BRCA1/2 mutation carriers (TUBA study):
PMID:22606070 https://doi.org/10.1001/jama.2011.766 a prospective non-randomised multicen-
PMID:21642681 tre study. BMC Cancer. 15(1):593. https://
28. Penninkilampi R, Eslick GD (2018). Perineal
d o i . o r g /10 .118 6 /s12 8 8 5 - 015 -15 9 7- y
talc use and ovarian cancer: a systematic 33. Jacobs IJ, Menon U, Ryan A, Gentry- PMID:26286255
review and meta-analysis. Epidemiology. Maharaj A, Burnell M, Kalsi JK, et
29(1):41–9. https://doi.org/10.1097/EDE. al. (2016). Ovarian cancer screening 38. Falconer H, Yin L, Grönberg H, Altman D
0000000000000745 PMID:28863045 and mortality in the UK Collaborative (2015). Ovarian cancer risk after salpingec-
29. Trabert B, Waterboer T, Idahl A, Brenner Trial of Ovarian Cancer Screening tomy: a nationwide population-based study.
N, Brinton LA, Butt J, et al. (2019). (UKCTOCS): a randomised controlled J Natl Cancer Inst. 107(2):107. https://doi.
Antibodies against Chlamydia trachoma- trial. Lancet. 387(10022):945–56. https:// org/10.1093/jnci/dju410 PMID:25628372
tis and ovarian cancer risk in two inde- doi.org/10.1016/S0140-6736(15)01224-6
39. McAlpine JN, Hanley GE, Woo MM, Tone
pendent populations. J Natl Cancer Inst. PMID:26707054
AA, Rozenberg N, Swenerton KD, et al.;
111(2):129–36. https://doi.org/10.1093/jnci/ Ovarian Cancer Research Program of
34. Mathieu KB, Bedi DG, Thrower SL,
djy084 PMID:29790947
Qayyum A, Bast RC Jr (2018). Screening British Columbia (2014). Opportunistic
30. Trabert B, Poole EM, White E, Visvanathan for ovarian cancer: imaging challenges and salpingectomy: uptake, risks, and compli-
K, Adami HO, Anderson GL, et al.; Ovarian opportunities for improvement. Ultrasound cations of a regional initiative for ovarian
Cancer Cohort Consortium (OC3) (2019). Obstet Gynecol. 51(3):293–303. https://doi. cancer prevention. Am J Obstet Gynecol.
Analgesic use and ovarian cancer org/10.1002/uog.17557 PMID:28639753 210(5):471.e1–11. https://doi.org/10.1016/j.
risk: an analysis in the Ovarian Cancer ajog.2014.01.003 PMID:24412119
Cohort Consortium. J Natl Cancer Inst. 35. Henderson JT, Webber EM, Sawaya GF
(2018). Screening for ovarian cancer: 40. Parker WH, Broder MS, Liu Z, Shoupe D,
111(2):137–45. https://doi.org/10.1093/jnci/
updated evidence report and systematic Farquhar C, Berek JS (2005). Ovarian con-
djy100 PMID:29860330
review for the US Preventive Services servation at the time of hysterectomy for
31. Trabert B, Ness RB, Lo-Ciganic WH, Task Force. JAMA. 319(6):595–606. benign disease. Obstet Gynecol. 106(2):
Murphy MA, Goode EL, Poole EM, et al.; https://doi.org/10.1001/jama.2017.21421 219–26. https://doi.org/10.1097/01.AOG.
Australian Ovarian Cancer Study Group, PMID:29450530 0000167394.38215.56 PMID:16055568
Australian Cancer Study (Ovarian Cancer);
Ovarian Cancer Association Consortium 36. Havrilesky LJ, Moorman PG, Lowery WJ,
(2014). Aspirin, nonaspirin nonsteroidal Gierisch JM, Coeytaux RR, Urrutia RP,
anti-inflammatory drug, and acetamino- et al. (2013). Oral contraceptive pills as
phen use and risk of invasive epithelial primary prevention for ovarian cancer:
ovarian cancer: a pooled analysis in the a systematic review and meta-analysis.
Ovarian Cancer Association Consortium. J Obstet Gynecol. 122(1):139–47. https://
Natl Cancer Inst. 106(2):djt431. https://doi. doi.org/10.1097/AOG.0b013e318291c235
org/10.1093/jnci/djt431 PMID:24503200 PMID:23743450
420
5.13
Prostate cancer
Challenges for prevention, detection,
and treatment
Timothy R. Rebbeck Timothy J. Key (reviewer)
Richard M. Martin (reviewer)
lifestyle, dietary, or other factors classification of tumours in 2016

SUMMARY to decrease risk of prostate can- recommended a grading system
cer are currently unavailable. that was updated to reflect the five
●● Prostate cancer has one of the grade groups of Epstein et al. [3],
largest disparities by race of ●● Prostate tumour markers have
which better reflect disease prog-
any major cancer type. Men of been identified that indicate etio-
nosis and outcomes compared with
African descent (e.g. African logically and phenotypically dis- previous categorizations.
American men) have the high- tinct groups of tumours, some of
est rates of prostate cancer in- which may have different prog-
cidence and mortality. nosis and response to treatment. Epidemiology
In 2018, prostate cancer was the
●● Prostate cancer has the high- ●● Chemoprevention for prostate
second most common non-cutane-
est heritability of any major cancer has been limited, despite
ous cancer in men worldwide (with
cancer type. evidence that some agents (e.g.
an estimated 1.3 million new cases)
5α-reductase inhibitors) may
●● More than 100 low-penetrance and the fifth most common cause
safely reduce the incidence of
of cancer death in men (with about
loci have been identified via ge- prostate cancer. 359 000 deaths) [4]. Incidence rates
nome-wide association studies,
of prostate cancer are highest in
but the use of this information in
North America, Europe, Australia,
predicting prostate cancer risk Prostate cancer is a group of his- and New Zealand (Fig. 5.13.1).
or outcomes remains limited. topathologically distinct tumour These elevated rates may reflect a
●● BRCA2, HOXB13, and DNA subtypes. These include glandu- truly higher incidence of disease as
mismatch repair genes are high- lar neoplasms (acinar adenocarci- well as higher prostate cancer de-
penetrance genes that may noma, intraductal carcinoma, and tection rates compared with other
have clinical utility in predicting ductal adenocarcinoma), urothelial areas. Incidence rates in Central and
prostate cancer risk, outcomes, carcinoma, squamous carcinoma South America appear to be slightly
and treatment options. (adenosquamous carcinoma and lower, and rates in Asia appear to be
squamous cell carcinoma), basal the lowest currently reported.
CHAPTER 5.13
SECTION 5
●● Among studied exposures and cell carcinoma, and neuroendocrine Rates of prostate cancer in
lifestyle, nutritional, and dietary tumours (adenocarcinoma with neu- Africa, particularly in sub-Saharan
factors, only attained adult roendocrine differentiation, small Africa, are very poorly captured by
height and underlying biologi- cell neuroendocrine carcinoma, and population-based tumour registries,
cal factors associated with adult large cell neuroendocrine carcino- and there is limited screening and
height are likely to be associ- ma) [1]. The most common of these early detection of prostate cancer in
ated with risk of prostate can- tumour subtypes is acinar adeno- Africa. Therefore, it is not clear that
cer. Factors related to obesity carcinoma, which accounts for more the apparently low rates of prostate
appear to be associated with than 99% of all prostate tumours [2]. cancer in Africa estimated by IARC
unfavourable outcomes in men There is significant variation and others are accurate.
diagnosed with prostate cancer. across these subtypes by age at Systematic surveys of the prev-
The evidence for other risk fac- diagnosis, race, prostate-specific alence of prostate cancer in Africa
tors is limited. Therefore, inter- antigen (PSA) level at diagnosis, [5] suggest that rates are as high
ventions to reduce exposures to and stage [2]. In addition, the WHO as or higher than those in African
Chapter 5.13 • Prostate cancer 421

Americans, who have among the screening. In Asia, which has lower
highest incidence rates of prostate rates of prostate cancer compared FUNDAMENTALS
cancer in the world. These infer- with other parts of the world, the in-
ences are consistent with findings crease in prostate cancer incidence ■■ Prostate cancer is highly
from autopsy studies that rates of was less profound. prevalent in middle-aged and
latent (prevalent) prostate cancer Trends in prostate cancer mor- older men. The incidence of
are highest in men of African de- tality (Fig. 5.13.4) have been influ- diagnosed prostate cancer is
scent, lower in men of European enced both by patterns of screen-
strongly correlated with the
descent, and lowest in men of Asian ing-associated detection and by
use of screening by prostate-
descent [6]. Therefore, it is likely treatment advances in some parts
that prostate cancer incidence in of the world. Since the advent of specific antigen (PSA) level.
Africa (particularly sub-Saharan PSA screening and the availability ■■ Older age, African ancestry
Africa) may be substantially higher of new surgical, radiotherapeutic, or race, and a family history
than is currently reported. and chemotherapeutic regimens in of prostate cancer are the
In contrast to prostate cancer the past 20 years, prostate cancer
only consistent risk factors for
incidence, prostate cancer mor- mortality has been slowly declin-
tality rates are highest in sub-Sa- prostate cancer.
ing in most parts of the world. Most
haran Africa, somewhat lower in recently, it has been reported that ■■ Sociodemographic inequities in
Central and South America and the mortality rates have levelled off af- prostate cancer screening and
Caribbean, lower in Europe, and still ter a period of decline, and the inci- treatment are likely to affect
lower in North America, Australia, dence of advanced prostate cancer prostate cancer outcomes.
and New Zealand (Fig. 5.13.2). The has increased in the USA since the The specific nature of these
rates are lowest in Asia. United States Preventive Services
This global variation in prostate inequities remains unclear.
Task Force recommended against
cancer mortality rates in part reflects PSA screening [7]. As discussed ■■ Screening for prostate cancer
underlying biological differences in below, screening has a more pro- by PSA level remains con-
risk as well as access to treatment. found impact on incidence for pros- troversial. After recommen-
For example, regions with increased tate cancer than for most other dations against widespread
detection of low-grade cancers cou- cancer types. The relationship of PSA screening, screening
pled with advanced treatment op- screening with prostate cancer
tions (e.g. the USA) tend to have rates declined, followed by an
mortality is more complex.
lower mortality rates compared with increase in rates of higher-
regions with low screening rates stage tumours at diagnosis. An
and the accompanying diagnosis Genetics and genomics improved approach is needed
of aggressive tumours coupled with Prostate cancer has the highest that balances early detection
limited treatment options (e.g. sub- reported heritability of any major of treatable cancers versus
Saharan Africa). cancer type [8]. Unlike the situation overdetection and overtreat-
Secular trends in prostate can- for other cancer types, the ability to ment of prostate cancers.
cer incidence rates (Fig. 5.13.3) re- define hereditary prostate cancer
flect the patterns of prostate cancer syndromes and identify hereditary
screening, including evaluation of cancer genes (see Chapter 3.2) has
PSA level and digital rectal exami- been limited. Family-based linkage ciations of inherited mutations in
nation. In North America, Australia, studies of hereditary prostate can- BRCA2 had implications for risk
New Zealand, and parts of Central cer have focused largely on popula- assessment and treatment. Among
and South America, prostate cancer tions of European descent to identi- carriers of BRCA2 mutations, the
incidence increased dramatically fy a series of genes responsible for risk of prostate cancer is increased
during the late 1980s and the 1990s hereditary prostate cancer [9,10]. 2.5–4.7-fold [12]. Also, prostate tu-
as a result of widespread PSA Although many high-penetrance mours with BRCA2 mutations have
screening. Similar trends were seen prostate cancer loci have been re- less favourable clinical character-
in other countries (e.g. in Europe) ported, very few have been imple- istics, including higher probability
but occurred about 10 years later, mented clinically. of nodal involvement, metastases,
in part because of later adoption Giri et al. reported the recom- high grade, advanced stage, and
of PSA screening compared with mendations of the first consensus lower median survival [13].
North America, Australia, and New conference to assess the value of Giri et al. also identified HOXB13
Zealand. In many countries, in- genetic testing for risk as well as mutations and DNA mismatch re-
cidence rates of prostate cancer clinical management of prostate pair gene mutations (accounting for
reached a peak about 5 years after cancer, held in 2017 [11]. This ex- Lynch syndrome) as potential can-
the widespread introduction of PSA pert group identified that asso- didates for genetic testing [11]. For
422
Fig. 5.13.1. Global distribution of estimated age-standardized (World) incidence rates (ASR) per 100 000 person-years for prostate
cancer, 2018.
Fig. 5.13.2. Global distribution of age-standardized (World) mortality rates (ASR) per 100 000 person-years for prostate cancer, 2018.
CHAPTER 5.13
HOXB13 mutations, relative risks magnitudes of association with The majority of these have reported SECTION 5
were estimated to be greater than prostate cancer have been identi- associations of loci with prostate
3, and for DNA mismatch repair fied through genome-wide asso- cancer risk (n = 659) as well as as-
gene mutations, estimated relative ciation studies (GWAS) and related sociations with prostate cancer me-
risks were 2.1–3.7 [12]. approaches. At least 170 common tastasis, aggressiveness, or survival
These associations suggest that variants associated with prostate (n = 56). Most associations reported
mutations at these loci confer suffi- cancer have been reported [14]. The in populations of European or Asian
ciently large effects that they can be NHGRI-EBI Catalog of published descent have not been replicated
considered in prostate cancer risk GWAS (https://www.ebi.ac.uk/gwas/, in populations of African descent
management and decision-making. accessed 13 October 2018) reported [15]. Few independent GWAS hits
In addition to high-penetrance more than 700 GWAS associations have been identified in populations
genes, loci with low to moderate for 23 prostate cancer-related traits. of African descent [16]. Multiple in-

Fig. 5.13.3. Age-standardized (World) incidence rates per 100 000 person-years by calendar year in selected countries for prostate
cancer, circa 1978–2012.
Denmark China (regional) Australia

Finland India (regional) Canada (except Quebec)
France (regional) Japan (regional) Colombia (regional)
Slovakia Philippines (regional) Costa Rica
Spain (regional) Singapore New Zealand
UK, England Thailand (regional) USA (SEER)
Fig. 5.13.4. Age-standardized (World) mortality rates per 100 000 person-years by calendar year in selected countries for prostate
cancer, circa 1978–2012.
Denmark Israel Australia

Finland Japan Canada
France Republic of Korea Colombia
Slovakia Singapore Costa Rica
Spain New Zealand
UK, England and Wales USA
424
dependent genomic associations at factors that are correlated with at- making about active treatment with
8q24 have been validated as pros- tained adult height. Exogenous ex- curative intent versus active surveil-
tate cancer susceptibility loci in mul- posures, including diet, nutrition, lance depends in part on knowing
tiple racial groups, including African and lifestyle, have not been con- which prostate tumours are likely
Americans [17]. Although no gene sistently associated with prostate to have unfavourable prognosis.
has been designated to be respon- cancer risk or a protective effect Therefore, knowledge of biomark-
sible for this increased risk of pros- [19]. These include no evidence ers that predict the likelihood of ag-
tate cancer, regulation of the down- for prostate cancer risk associated gressive disease may have clinical
stream gene MYC or regulation by with β-carotene, dietary calcium, vi- utility. These biomarkers include
long non-coding RNAs has been tamin D, dairy products, selenium, the TMPRSS2-ERG gene fusion
reported [18]. vitamin E, lycopene, and other fac- [25], Ki-67 expression [26], and bio-
tors that have been widely studied. markers involved in androgen me-
Etiology The limited convincing evidence tabolism [27]. Multigene genomic
for associations of exogenous ex- classifiers have been identified that
In contrast to the high heritability
posures, physical activity, lifestyle, assess prostate tumour aggres-
and large number of genetic as-
or dietary exposures with prostate siveness or prognosis [28].
sociations that influence prostate
cancer risk or outcomes makes it There are substantial differ-
cancer risk and outcomes, con-
difficult to identify modifiable factors ences in the distribution of pros-
firmed environmental factors or
that may be used in prostate can- tate tumour biomarkers by race,
exposures that influence prostate
cer prevention strategies. However, including ERG, AMACR, SPINK1,
cancer risk and outcomes are lim-
factors related to obesity appear to NKX3-1, GOLM1, and androgen
ited [19]. Older age, African ances-
be associated with unfavourable receptor. Dysregulation of AMACR,
try or race, and a family history of
outcomes in men diagnosed with ERG, FOXP1, and GSTP1 as well
prostate cancer are among the few
as loss-of-function mutations in the
uncontested risk factors for pros- prostate cancer, because of bio-
tumour suppressor genes NKX3-1
tate cancer. As summarized by the logical influences or less effective
and RB1 were found to predict the
2014 Continuous Update Project screening or treatment [20].
risk of extraprostatic extension and/
report on associations between
or seminal vesicle invasion in a
food, nutrition, and physical activity Biological characteristics race-dependent manner [29].
and the risk of prostate cancer, the
“convincing” level of evidence was
and early detection Although TMPRSS2-ERG trans-
Molecular signatures found in pros- locations do not seem to correlate
not achieved for any environmental
tate tumours reflect heterogeneity with clinical outcome in most studies
or behavioural risk factors [19].
in tumour etiology [21,22], corre- [30], the frequency of these events
Attained adult height and under-
late with a biological propensity to differs substantially by race [31]. In
lying biological factors associated
exhibit aggressive phenotypes [23], addition, several predictive or prog-
with adult height are probably risk
factors for prostate cancer. These and/or may direct optimal surveil- nostic models have been developed
effects are indirect and involve lance and treatment [24]. Decision- that include molecular biomarkers as
well as clinical and other traits (e.g.
the Stockholm-3 test, the 4Kscore
Fig. 5.13.5. A group of African American men. African ancestry or race is one of a few test, and multiparametric magnetic
uncontested risk factors for prostate cancer. resonance imaging [mpMRI]). These
results suggest that molecular signa-
tures, perhaps in combination with
clinical or other traits, may aid in un-
CHAPTER 5.13
SECTION 5
derstanding the biological underpin-
nings of prostate cancer disparities
and identify precision surveillance
and treatment regimens.
The PAM50 gene expression
classifier (which is used to identify
the major molecular subtypes of
breast cancer) has been used to de-
fine three prostate tumour subtypes –
luminal A, luminal B, and basal –
with significant differences in 10-
year biochemical recurrence-free
survival, distant metastasis-free
survival, prostate cancer-specific

survival, and overall survival [32]. Fig. 5.13.6. Access to health care is associated with prostate cancer outcomes
Luminal B prostate cancers were and disparities.
significantly associated with post-
operative response to androgen
deprivation therapy. The biomark-
ers identified to date may inform
screening for prostate cancer (e.g.
as an alternative or a supplement
to PSA testing), treatment choices,
and prognosis.
Socioeconomic
differences
Rates of prostate cancer are high-
er in African American men than
in men of other races across the
entire spectrum of prostate car-
cinogenesis, including high-grade
prostatic intraepithelial neoplasia,
prevalent (autopsy-detected) pros-
tate cancer, screen-detected can-
cer, incident prostate cancer, and
prostate cancer mortality [6,33]. protocols at a single institution [35]. captures multivariate comorbidity
At almost every point along the Other studies report that the dispar- data [38]. Given that some groups
prostate cancer continuum and for ity by race disappears after equal are more likely than others to have
almost every age group, prostate clinical protocols are applied [36]. co-existent chronic conditions, co-
cancer is more common in African Critically, the impact of access morbid conditions may play a role in
American men than in men of other to health care on outcomes may prostate cancer disparities in treat-
races in the USA. These data sug- vary by the metric used to assess ment and outcomes.
gest that the disparity may have these associations. A systematic
a biological component, because review and meta-analysis of differ-
Prevention
the disparity is evident even before ences in prognosis by race report-
cancer is usually clinically detected. ed no disparity in overall survival by Prevention and early detection of
However, the disparity increases race but found evidence for differ- prostate cancer have been con-
in magnitude in clinically detected ences in prostate cancer-specific troversial, and the source of great
disease and in mortality, suggest- survival and risk of biochemical re- confusion for both patients and cli-
ing that factors related to exposure, currence [37]. Thus, not all studies nicians. PSA screening had been
behaviour, or access to health care have been able to clearly demon- widely used in the USA and other
are also important in prostate can- strate that equal treatment leads to countries since 1992, when profes-
cer disparities (see Chapter 4.6). equal outcomes. The data available sional organizations, including the
Access to health care, and its to date do not completely resolve American Urological Association,
social, economic, and behavioural the question of whether racial dis- recommended annual PSA screen-
correlates, are associated with parities could be eliminated if treating for men aged 50 years and old-
prostate cancer outcomes and ment were optimized for specific er. Subsequently, a large increase
disparities. For example, the care groups on the basis of race and/or in prostate cancer incidence was
received by African American or socioeconomic status. observed, particularly for low-stage
Hispanic men differs in terms of Prediction of prostate cancer prostate tumours [39].
quality from that received by men screening participation, treatment This trend continued until the
of other races, and this, in turn, choices, and outcomes may also United States Preventive Services
affects outcomes and disparities involve the presence of comorbid Task Force recommended against
[34]. Disparities in outcome may conditions, which may influence the widespread PSA testing, in 2008 for
persist even within settings where clinician to assess whether a patient men older than 75 years and in 2012
men of different races have equal- will be able to benefit from a spe- for all men. Since that recommenda-
ity of care, including in the United cific medical intervention, including tion, rates of prostate tumours, par-
States Veterans Administration active surveillance. A variety of in- ticularly early-stage tumours, have
health-care system, within a clinical dices have been developed that at- decreased [40]. Subsequently, there
trial, or with treatment by standard tempt to create a simple metric that has been a trend towards diagnosis
426
Fig. 5.13.7. A man having blood drawn. Screening has a more profound impact on nasteride to increase the risk of
incidence for prostate cancer than for most other cancer types. high-grade tumours despite an
overall reduction in prostate cancer
incidence. The observation of in-
creased high-grade tumours in men
using finasteride has proven to be
incorrect [43], but use of finasteride
as a chemopreventive agent has
not been widespread.
Recently, the findings of the
earlier Prostate Cancer Prevention
Trial were replicated in a large popu-
lation-based non-randomized study
to demonstrate that 5α-reductase
inhibitors reduce risk of prostate
cancer, without an increase in risk
of high-grade disease [44]. These
data suggest that hormonally driv-
en chemopreventive regimens may
have value in reducing risk of pros-
tate cancer in some men.
of prostate tumours of unfavourable and other men at high risk of devel- Trials of micronutrients have
stage/grade [7]. oping prostate cancer [41]. been conducted both in the general
The public health implications of Chemoprevention for prostate population and in men with high-
prostate cancer screening to detect cancer has been of limited util- grade prostatic intraepithelial neo-
cancers at an early, treatable stage ity to date. In the Prostate Cancer plasia [45,46]. These trials ether
versus a desire to limit overdetec- Prevention Trial [42], evidence demonstrated no effect or revealed
tion and overtreatment of prostate was reported for reduction in risk a reduction in risk of prostate can-
cancers need to be resolved, par- of prostate cancer, but a concern cer at the cost of greater toxicities
ticularly for African American men was raised by the potential for fi- in the treatment arm.
CHAPTER 5.13
SECTION 5

References
1. Moch H, Cubilla AL, Humphrey PA, 10. Breyer JP, Avritt TG, McReynolds KM, 19. WCRF Continuous Update Project (2014).
Reuter VE, Ulbright TM (2016). The 2016 Dupont WD, Smith JR (2012). Confirmation World Cancer Research Fund International
WHO classification of tumours of the uri- of the HOXB13 G84E germline mutation in systematic literature review: the associa-
nary system and male genital organs - familial prostate cancer. Cancer Epidemiol tions between food, nutrition and physical
part A: renal, penile, and testicular tu- Biomarkers Prev. 21(8):1348–53. https:// activity and the risk of prostate cancer.
mours. Eur Urol. 70(1):93–105. https:// doi.org/10.1158/1055-9965.EPI-12-0495 Available from: https://www.wcrf.org/sites/
doi.org/10.1016/j.eururo. 2016.02.029 PMID:22714738 default/files/prostate-cancer-slr.pdf.
PMID:26935559
11. Giri VN, Knudsen KE, Kelly WK, Abida 20. Ma J, Li H, Giovannucci E, Mucci L, Qiu W,
2. Marcus DM, Goodman M, Jani AB, W, Andriole GL, Bangma CH, et al. Nguyen PL, et al. (2008). Prediagnostic
Osunkoya AO, Rossi PJ (2012). A compre- (2018). Role of genetic testing for inher- body-mass index, plasma C-peptide con-
hensive review of incidence and survival in ited prostate cancer risk: Philadelphia centration, and prostate cancer-specific
patients with rare histological variants of Prostate Cancer Consensus Conference mortality in men with prostate cancer: a
prostate cancer in the United States from 2017. J Clin Oncol. 36(4):414–24. long-term survival analysis. Lancet Oncol.
1973 to 2008. Prostate Cancer Prostatic https://doi.org/10.1200/JCO.2017.74.1173 9(11):1039–47. https://doi.org/10.1016/
Dis. 15(3):283–8. https://doi.org/10.1038/ PMID:29236593 S1470-2045(08)70235-3 PMID:18835745
pcan.2012.4 PMID:22349984
12. Zhen JT, Syed J, Nguyen KA, Leapman 21. Ahearn TU, Pettersson A, Ebot EM, Gerke
3. Epstein JI, Egevad L, Amin MB, Delahunt MS, Agarwal N, Brierley K, et al. (2018). T, Graff RE, Morais CL, et al. (2015). A
B, Srigley JR, Humphrey PA; Grading Genetic testing for hereditary prostate prospective investigation of PTEN loss
Committee (2016). The 2014 International cancer: current status and limitations. and ERG expression in lethal prostate
Society of Urological Pathology (ISUP) Cancer. 124(15):3105–17. https://doi. cancer. J Natl Cancer Inst. 108(2):djv346.
Consensus Conference on Gleason org/10.1002/cncr.31316 PMID:29669169 h t t p s : / / d o i . o r g / 10 .10 9 3 / j n c i / d j v 3 4 6
Grading of Prostatic Carcinoma: defini- PMID:26615022
tion of grading patterns and proposal for 13. Castro E, Goh C, Olmos D, Saunders E,
a new grading system. Am J Surg Pathol. Leongamornlert D, Tymrakiewicz M, et al. 22. Graff RE, Meisner A, Ahearn TU,
40(2):244–52. PMID:26492179 (2013). Germline BRCA mutations are as- Fiorentino M, Loda M, Giovannucci EL, et
sociated with higher risk of nodal involve- al. (2016). Pre-diagnostic circulating sex
4. Bray F, Ferlay J, Soerjomataram I, Siegel ment, distant metastasis, and poor survival hormone levels and risk of prostate can-
RL, Torre LA, Jemal A (2018). Global outcomes in prostate cancer. J Clin Oncol. cer by ERG tumour protein expression.
cancer statistics 2018: GLOBOCAN es- 31(14):1748–57. https://doi.org/10.1200/ Br J Cancer. 114(8):939–44. https://doi.
timates of incidence and mortality world- JCO.2012.43.1882 PMID:23569316 org/10.1038/bjc.2016.61 PMID:26986253
wide for 36 cancers in 185 countries. CA
Cancer J Clin. 68(6):394–424. https://doi. 14. Benafif S, Kote-Jarai Z, Eeles RA; 23. Fine SW, Gopalan A, Leversha MA, Al-
org/10.3322/caac.21492 PMID:30207593 PRACTICAL Consortium (2018). A review Ahmadie HA, Tickoo SK, Zhou Q, et al.
of prostate cancer genome-wide associa- (2010). TMPRSS2-ERG gene fusion
5. Hsing AW, Yeboah E, Biritwum R, Tettey Y, tion studies (GWAS). Cancer Epidemiol is associated with low Gleason scores
De Marzo AM, Adjei A, et al. (2014). High Biomarkers Prev. 27(8):845–57. https:// and not with high-grade morphological
prevalence of screen detected prostate doi.org/10.1158/1055-9965.EPI-16-1046 features. Mod Pathol. 23(10):1325–33.
cancer in West Africans: implications for PMID:29348298 h t t p s : / / d o i . o r g / 10 .10 3 8 / m o d p a t h o l .
racial disparity of prostate cancer. J Urol. 2010.120 PMID:20562851
192(3):730–5. https://doi.org/10.1016/j. 15. Chang BL, Spangler E, Gallagher S,
juro.2014.04.017 PMID:24747091 Haiman CA, Henderson B, Isaacs W, 24. Netto GJ (2013). Clinical applications of
et al. (2011). Validation of genome-wide recent molecular advances in urologic ma-
6. Rebbeck TR, Haas GP (2014). Temporal prostate cancer associations in men lignancies: no longer chasing a “mirage”?
trends and racial disparities in global of African descent. Cancer Epidemiol Adv Anat Pathol. 20(3):175–203. https://
prostate cancer prevalence. Can J Urol. Biomarkers Prev. 20(1):23–32. https:// doi.org/10.1097/PAP.0b013e3182863f80
21(5):7496–506. PMID:25347377 doi.org/10.1158/1055-9965.EPI-10-0698 PMID:23574774
PMID:21071540
7. Eapen RS, Herlemann A, Washington 25. Demichelis F, Fall K, Perner S, Andrén
SL 3rd, Cooperberg MR (2017). Impact 16. Haiman CA, Chen GK, Blot WJ, Strom O, Schmidt F, Setlur SR, et al. (2007).
of the United States Preventive Services SS, Berndt SI, Kittles RA, et al. (2011). TMPRSS2:ERG gene fusion associated
Task Force ‘D’ recommendation on Genome-wide association study of prostate with lethal prostate cancer in a watchful
prostate cancer screening and staging. cancer in men of African ancestry identifies waiting cohort. Oncogene. 26(31):4596–9.
Curr Opin Urol. 27(3):205–9. https://doi. a susceptibility locus at 17q21. Nat Genet. https://doi.org/10.1038/sj.onc.1210237
org/10.1097/MOU.0000000000000383 43(6):570–3. https://doi.org/10.1038/ng.839 PMID:17237811
PMID:28221220 PMID:21602798
26. Berney DM, Gopalan A, Kudahetti S,
8. Mucci LA, Hjelmborg JB, Harris JR, Czene 17. Freedman ML, Haiman CA, Patterson N, Fisher G, Ambroisine L, Foster CS, et
K, Havelick DJ, Scheike T, et al.; Nordic McDonald GJ, Tandon A, Waliszewska al. (2009). Ki-67 and outcome in clini-
Twin Study of Cancer (NorTwinCan) A, et al. (2006). Admixture mapping cally localised prostate cancer: analysis
Collaboration (2016). Familial risk and identifies 8q24 as a prostate cancer risk of conservatively treated prostate cancer
heritability of cancer among twins in locus in African-American men. Proc patients from the Trans-Atlantic Prostate
Nordic countries. JAMA. 315(1):68–76. Natl Acad Sci U S A. 103(38):14068–73. Group study. Br J Cancer. 100(6):888–93.
https://doi.org/10.1001/jama.2015.17703 https://doi.org/10.1073/pnas.0605832103 https://doi.org/10.1038/sj.bjc.6604951
PMID:26746459 PMID:16945910 PMID:19293807
9. Berry R, Schaid DJ, Smith JR, French 18. Ahmadiyeh N, Pomerantz MM, Grisanzio
AJ, Schroeder JJ, McDonnell SK, et al. C, Herman P, Jia L, Almendro V, et al.
(2000). Linkage analyses at the chromo- (2010). 8q24 prostate, breast, and co-
some 1 loci 1q24-25 (HPC1), 1q42.2-43 lon cancer risk loci show tissue-specific
(PCAP), and 1p36 (CAPB) in families long-range interaction with MYC. Proc
with hereditary prostate cancer. Am J Natl Acad Sci U S A. 107(21):9742–6.
Hum Genet. 66(2):539–46. https://doi. https://doi.org/10.1073/pnas.0910668107
org/10.1086/302771 PMID:10677314 PMID:20453196
428
27. Li R, Wheeler T, Dai H, Frolov A, 33. Cronin KA, Lake AJ, Scott S, Sherman 40. Jemal A, Fedewa SA, Ma J, Siegel R, Lin
Thompson T, Ayala G (2004). High level RL, Noone A-M, Howlader N, et al. (2018). CC, Brawley O, et al. (2015). Prostate can-
of androgen receptor is associated with Annual report to the nation on the status of cer incidence and PSA testing patterns
aggressive clinicopathologic features cancer, part I: national cancer statistics. in relation to USPSTF screening recom-
and decreased biochemical recurrence- Cancer. 124(13):2785–800. https://doi. mendations. JAMA. 314(19):2054–61.
free survival in prostate: cancer patients org/10.1002/cncr.31551 PMID:29786848 https://doi.org/10.1001/jama.2015.14905
treated with radical prostatectomy. Am PMID:26575061
J Surg Pathol. 28(7):928–34. https:// 34. Jayadevappa R, Chhatre S, Johnson JC,
doi.org/10.1097/00000478-200407000- Malkowicz SB (2011). Association be- 41. Loeb S, Bjurlin MA, Nicholson J, Tammela
00013 PMID:15223964 tween ethnicity and prostate cancer out- TL, Penson DF, Carter HB, et al. (2014).
comes across hospital and surgeon volume Overdiagnosis and overtreatment of pros-
28. Erho N, Buerki C, Triche TJ, Davicioni groups. Health Policy. 99(2):97–106. https:// tate cancer. Eur Urol. 65(6):1046–55.
E, Vergara IA (2012). Transcriptome- doi.org/10.1016/j.healthpol.2010.07.014 https://doi.org/10.1016/j.eururo.2013.12.
wide detection of differentially ex- PMID:20708815 062 PMID:24439788
pressed coding and non-coding tran-
scripts and their clinical significance in 35. Yamoah K, Deville C, Vapiwala N, Spangler 42. Thompson IM, Goodman PJ, Tangen CM,
prostate cancer. J Oncol. 2012:541353. E, Zeigler-Johnson CM, Malkowicz B, et Lucia MS, Miller GJ, Ford LG, et al. (2003).
ht t p s: //d o i.o r g /10.115 5 / 2 012 / 5 413 5 3 al. (2015). African American men with low- The influence of finasteride on the devel-
PMID:22956952 grade prostate cancer have increased dis- opment of prostate cancer. N Engl J Med.
ease recurrence after prostatectomy com- 349(3):215–24. https://doi.org/10.1056/
29. Yamoah K, Johnson MH, Choeurng V, NEJMoa030660 PMID:12824459
pared with Caucasian men. Urol Oncol.
Faisal FA, Yousefi K, Haddad Z, et al.
33(2):70.e15–22. https://doi.org/10.1016/j. 43. Redman MW, Tangen CM, Goodman PJ,
(2015). Novel biomarker signature that
urolonc.2014.07.005 PMID:25304288 Lucia MS, Coltman CA Jr, Thompson IM
may predict aggressive disease in African
American men with prostate cancer. (2008). Finasteride does not increase the
36. Powell IJ, Banerjee M, Bianco FJ, Wood
J Clin Oncol. 33(25):2789–96. https:// risk of high-grade prostate cancer: a bi-
DP Jr, Dey J, Lai Z, et al. (2004). The
d o i . o r g / 10 .12 0 0 /J C O . 2 0 14 . 5 9 . 8 9 12 as-adjusted modeling approach. Cancer
effect of race/ethnicity on prostate can-
PMID:26195723 Prev Res (Phila). 1(3):174–81. https://doi.
cer treatment outcome is conditional:
org/10.1158/1940- 6207.CAPR-08-0092
30. Pettersson A, Graff RE, Bauer SR, Pitt a review of Wayne State University
PMID:19138953
MJ, Lis RT, Stack EC, et al. (2012). The data. J Urol. 171(4):1508–12. https://doi.
TMPRSS2:ERG rearrangement, ERG ex- org/10.1097/01.ju.0000118906.16629.8c 44. Wallerstedt A, Strom P, Gronberg H,
pression, and prostate cancer outcomes: PMID:15017209 Nordstrom T, Eklund M (2018). Risk
a cohort study and meta-analysis. Cancer of prostate cancer in men treated with
37. Evans S, Metcalfe C, Ibrahim F, Persad 5α-reductase inhibitors – a large pop-
Epidemiol Biomarkers Prev. 21(9):1497–
R, Ben-Shlomo Y (2008). Investigating ulation-based prospective study. J Natl
509. https://doi.org/10.1158/1055-9965.
Black-White differences in prostate can- Cancer Inst. 110(11):1216–21. https://doi.
EPI-12-0042 PMID:22736790
cer prognosis: a systematic review and org/10.1093/jnci/djy036 PMID:29548030
31. Magi-Galluzzi C, Tsusuki T, Elson P, meta-analysis. Int J Cancer. 123(2):
Simmerman K, LaFargue C, Esgueva R, 430–5. https://doi.org/10.1002/ijc.23500 45. Cui K, Li X, Du Y, Tang X, Arai S, Geng
et al. (2011). TMPRSS2-ERG gene fusion PMID:18452170 Y, et al. (2017). Chemoprevention of pros-
prevalence and class are significantly dif- tate cancer in men with high-grade pro-
ferent in prostate cancer of Caucasian, 38. Sarfati D (2012). Review of methods static intraepithelial neoplasia (HGPIN):
African-American and Japanese pa- used to measure comorbidity in cancer a systematic review and adjusted indi-
tients. Prostate. 71(5):489–97. https://doi. populations: no gold standard exists. J rect treatment comparison. Oncotarget.
org/10.1002/pros.21265 PMID:20878952 Clin Epidemiol. 65(9):924–33. https:// 8(22):36674–84. https://doi.org/10.18632/
doi.org/10.1016/j.jclinepi.2012.02.017 oncotarget.16230 PMID:28415774
32. Zhao SG, Chang SL, Erho N, Yu M, PMID:22739245
Lehrer J, Alshalalfa M, et al. (2017). 46. Klein EA, Thompson IM Jr, Tangen CM,
Associations of luminal and basal subtyp- 39. Potosky AL, Miller BA, Albertsen PC, Crowley JJ, Lucia MS, Goodman PJ, et
ing of prostate cancer with prognosis and Kramer BS (1995). The role of increasing al. (2011). Vitamin E and the risk of pros-
response to androgen deprivation thera- detection in the rising incidence of prostate tate cancer: the Selenium and Vitamin E
py. JAMA Oncol. 3(12):1663–72. https:// cancer. JAMA. 273(7):548–52. https://doi. Cancer Prevention Trial (SELECT). JAMA.
doi.org/10.10 01/jamaoncol. 2017.0751 org/10.1001/jama.1995.03520310046028 306(14):1549–56. https://doi.org/10.1001/
PMID:28494073 PMID:7530782 jama.2011.1437 PMID:21990298
CHAPTER 5.13
SECTION 5

5.14
Testicular cancer
New inroads into early diagnosis
Ewa Rajpert-De Meyts Gemma Gatta (reviewer)

Ariana Znaor Jason Gurney (reviewer)
Niels E. Skakkebæk Joachim Schüz (reviewer)
for early detection of testicular young men (age 15–45 years). The

SUMMARY germ cell neoplasia. tumours that occur in this age group
are distinct from others through the
●● Testicular cancer comprises ●● Because testicular cancer oc-
association with germ cell neo-
mainly germ cell-derived tu- curs predominantly in young
plasia in situ (GCNIS) and testicu-
mours, which according to the men, survivors should be fol-
lar dysgenesis syndrome [1]. The
most recent (2016) WHO clas- lowed up for many years, with
pathogenesis of these tumours has
sification are divided into two attention paid to preservation of
a strong developmental component
groups: tumours derived from reproductive function and pre-
and overlaps with other disorders of
germ cell neoplasia in situ, vention of late effects, such as
the male reproductive system, such
which are the most common, hypogonadism, metabolic syn-
as cryptorchidism, other genital
and rare tumours unrelated to drome, and secondary cancers.
malformations, and some forms of
germ cell neoplasia in situ. male infertility [1].
●● The incidence of testicular can- The histopathology of germ cell
Testicular cancer is an atypical
cer has been rising steeply in tumours is very heterogeneous, be-
type of solid tumour. It is the most
many countries that previously cause of their plasticity and ability
common cancer type in young
had low incidence rates (e.g. to transdifferentiate. Consequently,
men, and its incidence is increas-
Croatia and Finland), whereas there have been frequent changes
ing worldwide. Testicular cancer
in some high-incidence coun- in classification and disagreements
has strong developmental and
tries (e.g. Denmark) the rates about terminology. The most recent
environmental links, but also sub-
have levelled off. (2016) edition of the WHO classi-
stantial genetic susceptibility.
fication is the result of a thorough
Although testicular cancer can
●● Changing incidence trends are revision and update by a panel of
be derived from several cell types
consistent with a major role of experts, who agreed on a new di-
of the testis, germ cell-derived tu-
environmental factors in the vision and nomenclature to better
mours constitute the vast majority
pathogenesis of testicular germ reflect the biological features and
of cases. Testicular tumours known
cell tumours, acting primarily histogenesis of germ cell tumours
as sex cord stromal tumours and
during early development. of the testis [2]. According to this
Leydig cell tumours are derived
from somatic cells present in the classification (Box 5.14.1), testicular
●● Testicular cancer is a polygenic
testis; these tumours are relatively germ cell tumours are divided into
syndrome, without major predis-
rare, so they are not discussed in two main groups: germ cell tumours
posing oncogenic mutations but
this chapter. Malignancies that are derived from GCNIS, and germ cell
with a large number of germline
not specific for the testis, such as tumours unrelated to GCNIS.
susceptibility loci; this renders
genetic screening impossible. lymphoma or sarcoma, are not con-
Germ cell tumours derived
sidered here either.
●● Particular features of germ cell from GCNIS
neoplasia in situ, including high Germ cell tumours derived from
expression of pluripotency fac- General characteristics GCNIS comprise morphologically
tors, low levels of DNA meth- and histopathology homogeneous seminoma and het-
ylation, and a specific micro- Testicular germ cell tumours are erogeneous non-seminomatous tu-
RNA profile, can be exploited most common in adolescents and mours, which can contain pure or
430
Box 5.14.1. Main types of germ cell tumours of the testis, according to the 2016 WHO
classification.
FUNDAMENTALS
Germ cell tumours derived from germ cell neoplasia in situ ■■ Testicular germ cell tumours
• Germ cell neoplasia in situ occur predominantly in
adolescents and young men
• Seminoma, pure
(aged 15–45 years), who
• Non-seminomatous germ cell tumours, pure develop seminoma or non-
-- Embryonal carcinoma seminomatous tumours, which
are derived from germ cell
-- Yolk sac tumour, postpubertal type
neoplasia in situ. Testicular
-- Trophoblastic tumours, including choriocarcinoma tumours in children and older
-- Teratoma, postpubertal type men are relatively rare and
• Non-seminomatous germ cell tumours, mixed have different pathogenesis.
Germ cell tumours unrelated to germ cell neoplasia in situ ■■ Incidence rates of testicular
cancer vary geographically and
• Spermatocytic tumour
ethnically; rates are highest in
• Prepubertal (paediatric) tumours men of European descent and
-- Teratoma, prepubertal type lowest in men of African and
-- Yolk sac tumour, prepubertal type East Asian ancestry.
-- Mixed tumour, prepubertal type ■■ Because of phenotypic similar-

ity to fetal germ cells and a
strong association with distur-
mixed components of embryonal Spermatocytic tumour is thought bances of early development,
carcinoma, yolk sac tumour, cho- to grow from expanding spermato- testicular germ cell tumours
riocarcinoma, and teratoma. The gonial clones, which underwent associated with germ cell neo-
precursor lesion, GCNIS, consists genomic changes that facilitated
plasia in situ are considered a
of gonocyte-like cells that persisted their survival, such as amplification
part of testicular dysgenesis
in the immature stage after the fetal/ of chromosome 9 (DMRT1 locus),
infantile period and then underwent syndrome, together with dis-
activating mutations in FGFR3,
malignant transformation [1]. The orders of sexual development,
HRAS, and NRAS, or whole-chro-
pathogenesis of GCNIS is depicted mosome aneuploidy [4]. Childhood cryptorchidism, and decreased
in Fig. 5.14.1. germ cell tumours are probably spermatogenesis with signs
In individuals with disorders of derived from primordial germ cells, of impaired function of Sertoli
sexual development, a pre-invasive but their etiology remains unknown. cells and Leydig cells.
lesion similar to GCNIS is called
■■ The majority of cases of
gonadoblastoma. Gonadoblastoma
and GCNIS can be present in the Epidemiology testicular cancer cannot
same patient, and intermediate le- be explained by heritability
Global burden and incidence alone and are attributed to
sions are not uncommon in patients
trends still-unknown environmental
with testicular dysgenesis syn-
drome [3]. Because of the rarity of other types factors, which act mainly
CHAPTER 5.14
SECTION 5
of testicular tumours, germ cell tu- during development.
Germ cell tumours unrelated mours that occur in young men,
to GCNIS which are derived from GCNIS, com- ■■ Testicular cancer has a gener-
Germ cell tumours unrelated to prise about 95% of cases and hence ally good prognosis and low
GCNIS include rare spermatocytic are responsible for the global bur- mortality, when managed by
tumour of older men (mean age at den of testicular cancer. Seminomas modern methods that exploit
diagnosis, > 54 years) and child- are most often diagnosed in men the sensitivity of germ cells to
hood testicular tumours (most com- aged 25–45 years, whereas non- cisplatin-based chemotherapy.
mon in infants and children up to seminomatous tumours occur in Non-seminomas, especially
age 4 years). Spermatocytic tumour relatively younger men, mainly in the somatically differentiated
has been renamed from the previ- age group 15–35 years [5]. teratomas, are more resistant
ously used term “spermatocytic Testicular cancer is relatively to treatment compared with
seminoma”, to avoid confusion with rare compared with other cancer seminomas.
seminoma derived from GCNIS [2]. types, with an estimated 71 105 new
Chapter 5.14 • Testicular cancer 431

Fig. 5.14.1. Schematic depiction of the pathogenesis of testicular germ cell tumours cases worldwide in 2018 (< 1% of the
derived from germ cell neoplasia in situ (GCNIS). These tumours are thought to be male cancer burden) [6]. However, it
caused by a combination of adverse environmental and genetic factors (multifactorial is the most common cancer type in
and polygenic), which result in insufficient masculinization of the gonads, mainly
because of impaired function of Sertoli cells and Leydig cells. The insufficient
young men. Germ cell tumours are
stimulation of developing germ cells causes arrest of gonocyte differentiation. The most common in men of European
delayed gonocytes (pre-GCNIS cells) begin to proliferate during pubertal hormonal descent, whereas the incidence is
stimulation of the testis. Increased proliferation results in genomic changes that favour very low in men of African and East
malignant transformation of GCNIS cells into an invasive tumour, either a seminoma or Asian ancestry.
a non-seminoma. Normal germ cell development is shown in the top part of the figure, Age-standardized incidence rates
on the green background. EC, embryonal carcinoma; PGC, primordial germ cells.
range from less than 0.5 per 100 000
in the lowest-incidence areas to more
than 10 per 100 000 in high-risk
populations (Fig. 5.14.2) [6]. In 2018,
the estimated 5-year prevalence
of testicular cancer worldwide was
284 073, of which 107 570 prevalent
cases (38%) were in Europe [6].
The incidence of testicular germ
cell tumours increased markedly
around the world in the second half
of the 20th century, with substan-
tial geographical differences [7,8].
Recent studies have shown dra-
matically increasing trends in some
European countries that previously
had low incidence rates (e.g. Croatia
Fig. 5.14.2. Distribution of estimated age-standardized (World) incidence rates (ASR) per 100 000 person-years for testicular
cancer, 2018: (A) globally, (B) in Europe.
A
432
and Finland), and in Hispanic popu- and 8.6 per 100 000 in Slovakia) ap- men [11]. In most countries in Asia,
lations in the Americas. In contrast, proached the rates in the high-risk the incidence is low and is increas-
incidence rates in Denmark, which countries in northern Europe [10]. As, ing only modestly, whereas in Latin
were previously very high, have for example, in the Nordic countries, America marked increases have
shown signs of levelling off [8,9] incidence rates of testicular cancer been observed [7,8].
(Fig. 5.14.3). can vary widely between neighbour-
These changing trends reflect ge- ing countries while showing smaller Mortality
ographical patterns. In 2008–2012, within-country variations compared In 2018, there were an estimated
incidence rates in western Europe with other cancer types [9]. In the 9507 deaths from testicular cancer
(9.1 per 100 000 in Germany and multiethnic USA, there are large dif- worldwide [6]. Age-standardized
8.8 per 100 000 in Switzerland) and ferences between ethnic groups; a mortality rates for testicular can-
in some countries in south-eastern recent increase in incidence rates cer are low (≤ 1 per 100 000). This
Europe (8.8 per 100 000 in Slovenia has been noted in Hispanic White is due in part to the relative ease of
Fig. 5.14.3. Age-standardized (World) incidence rates per 100 000 person-years by calendar year in selected countries for testicular
cancer, circa 1955–2010. Asterisks indicate regional registries (other registries are national).
Northern Europe The Americas Asia

10 12
10 12
10 12
Norway
Denmark
Age−standardized (World) incidence rate per 100000

UK,
Scotland* New
7
7
7
Ireland USA Zealand

Iceland White* Australia
5
5
5
Finland
Ecuador* Israel
Canada* Costa
Rica
3
3
Estonia
3
Latvia
Colombia*
Lithuania
2
2
2
1.5
1.5
1.5
USA Japan*
Black* China* Singapore
Philippines*
1
1
1
India*
0.7
0.7
0.7
0.5
0.5
0.5
1960 1970 1980 1990 2000 2010 1960 1970 1980 1990 2000 2010 1960 1970 1980 1990 2000 2010
Year Year Year
Eastern, Southern and Western Europe

10 12
10 12
10 12
Switzerland*
Slovenia Germany*
The
Czech Netherlands
7
7
7
Croatia
Republic UK,
Slovakia Italy* England*
CHAPTER 5.14
SECTION 5
Bulgaria France*
5
5
5
Poland*
3
3
3
Spain*
2
2
2
Belarus
1.5
1.5
1.5
Russian
Federation
1
1
1
0.7
0.7
0.7
0.5
0.5
0.5
1960 1970 1980 1990 2000 2010 1960 1970 1980 1990 2000 2010 1960 1970 1980 1990 2000 2010
Year Year Year

Fig. 5.14.4. Global distribution of estimated age-standardized (World) mortality rates (ASR) per 100 000 person-years for testicular
cancer, 2018.
diagnosis and surgical treatment, tween different world regions, such dividuals is variable, but it is great-
but predominantly to the very effi- as between Latin America and North est in those with mixed gonadal
cient cisplatin-based chemothera- America [7]. dysgenesis (e.g. 45,X/46,XY karyo-
py regimens. For the population of type) and with partial androgen in-
the USA in 2009–2015, the 5-year Etiology and risk factors sensitivity syndrome [3,16].
relative survival was 95.2% over- Cryptorchidism is the most sig-
all; for localized testicular cancer The increasing prevalence of crypt- nificant risk factor for sporadic tes-
it was more than 99%, but for dis- orchidism, other genital malforma- ticular cancer, and about 5% of pa-
seminated testicular cancer it was tions, and male infertility synchro- tients with a history of undescended
about 73% [12]. Despite the gener- nous with testicular cancer was the testis develop a testicular germ cell
ally good prognosis, studies have basis for the hypothesis that these tumour. Other repeatedly identified
revealed that the treatment efficacy conditions could be etiologically risk factors include inguinal hernia,
of disseminated and refractory tes- linked within testicular dysgenesis low birth weight, high maternal age,
ticular cancer, especially for cases syndrome [1]. The causal factors being born first, late age at puberty,
that require salvage surgery, is best behind the epidemic increase in and poor spermatogenesis [1].
in high-volume centres with good incidence rates and the rapidly Epidemiological and clinical
experience [13]. changing trends in testicular cancer studies that identified links to early
Although in high-income coun- remain largely unknown, but they development are consistent with
tries early diagnosis and adequate must be related to environment or the biological features of GCNIS,
treatment are available and mortality lifestyle. The primary importance of which is characterized by close
rates have been declining since the environmental factors is also sup- similarity to fetal gonocytes [17]
1970s or 1980s, in low- and middle- ported by studies of migrant popu- (see below for details). However,
income countries access to testicular lations, in which the risk of testicu- except for the rare cases of disor-
cancer control is more limited [14]; lar cancer changed depending on ders of sexual development with
this is reflected in higher mortali- the geographical location during obvious genetic defects that lead to
ty rates in lower-income countries development (reviewed in [1]). germ cell tumours (e.g. mutations in
(Fig. 5.14.4) and large global vari- The etiology is known only in SRY or AR), identification of specif-
ations in incidence-to-mortality ra- a small percentage of genetically ic causal factors that cause delayed
tios [6,7]. The EUROCARE-5 study determined cases. Individuals with maturation of gonocytes has proven
reported age-standardized 5-year developmental abnormalities of the difficult. Among multiple hypotheti-
relative survival for 2005–2007 of gonads and sex differentiation (in- cal environmental factors, prenatal
90%, with survival in eastern Europe cluding testicular dysgenesis syn- maternal lifestyle factors or intra-
about 10% lower than that in northern drome and disorders of sexual de- uterine or perinatal exposures to
and western Europe [15]. Disparities velopment) are at high risk of germ xenobiotics or endocrine disrupters
in mortality have been reported be- cell neoplasia. The risk in these in- have been suggested.
434
Early studies investigated es- patients with testicular germ cell tu- sequencing study of 919 patients and
trogenic compounds, including in mours, no strong associations have 1609 cancer-free controls [25]. This
utero exposure to diethylstilbes- been found; furthermore, the few complex polygenic nature of testicu-
trol, followed by anti-androgenic existing studies on maternal or pa- lar cancer is consistent with a com-
organochlorine compounds, such rental exposures have not yielded plex and multifactorial pathogenesis,
as 4,4′-dichlorodiphenyltrichloro- any consistent results [21]. which renders genetic screening for
ethane (DDT), and phthalates. Few germline mutations impossible in the
conclusive results were obtained, clinical setting.
Genetics
except for weak associations with Several genome-wide associa-
exposure to 4,4′-dichlorodiphenyl- Testicular cancer is among the tion studies (GWAS) (see Chapter
dichloroethylene (DDE) – a metab- cancer types with a relatively high 3.2) performed since 2009 have
olite of DDT that is sometimes used heritability. Familial risk is high, identified a number of possibly pre-
as a biomarker of exposure to DDT – especially for brothers of patients disposing gene variants. The strong-
and chlordane [1,18]. with germ cell tumours, who have est genetic markers for an increased
Larger, well-controlled studies an estimated 8–10-fold increased risk of testicular germ cell tumours
that are based on novel ideas are risk, whereas the sons of cases are located within or near the fol-
needed [19]. Future studies should have a 4–6-fold increased risk [22]. lowing loci: KITLG, SPRY4, DMRT1,
investigate maternal and devel- A greater risk for brothers than for PRDM14, DAZL, and HPGDS (re-
opmental exposures to emerging sons is consistent with a strong en- viewed in [26]). Other informative
endocrine disrupters and their mix- vironmental modulation of the risk markers have been revealed by re-
tures, preferably in combination with during development. cent meta-analytic GWAS that com-
the evaluation of the genetic predis- Specific oncogenic driver muta- bined data from very large cohorts,
position of the studied individuals. tions in a single gene have not been increasing the number of predispos-
With regard to postpubertal or identified in patients with testicular ing loci to 49 and the combined her-
adult exposures, very few risk fac- cancer, except for secondary gain-of- itability to more than one third of the
tors have been identified. Heavy function KIT mutations, which were studied cases [27,28]. The multicen-
use of cannabis (defined as use detected essentially only in pure tre meta-analyses currently being
at least weekly or use for at least seminomas [23], or KRAS mutations, carried out by international consor-
10 years), but not occasional use, which were detected mainly in non- tia will probably identify additional
has been associated with a dou- seminomas, as well as a few other susceptibility genes. Most of the
bling of the risk of developing a secondary passenger mutations [24]. predisposing genetic markers iden-
non-seminoma, compared with The absence of a major single pre- tified so far implicate predominantly
never use [20]. Among occupation- disposition gene has recently been pathways involved in germ cell de-
al exposures of the relatively young confirmed by a large whole-exome velopment, sex differentiation, and
gonadal development, as well as
Fig. 5.14.5. A cannabis plant. The only consistently reported postpubertal risk factor for centrosome cycle, DNA repair, and
testicular cancer (mainly non-seminoma) is heavy use of cannabis. telomere function [26].
Some of the predisposing vari-
ants have different prevalence be-
tween racial groups, thus shedding
some light on the reasons for the
large ethnic differences in the in-
cidence of testicular cancer. One
illustrative example is the KITLG
CHAPTER 5.14
SECTION 5
locus (single-nucleotide polymor-
phism rs995030), which is carried
by most people of European de-
scent but only a minority of people
of African descent.
important for diagnosis
The biological features of tumours
derived from GCNIS differ markedly
from those of the normal germ cells
found in the adult testis; this provides
insights into their pathogenesis and

Fig. 5.14.6. The precursor lesion of the most prevalent testicular germ cell tumours in young men, germ cell neoplasia in situ
(GCNIS), in a testicular biopsy. (A) Haematoxylin and eosin staining showing the difference between tubules with GCNIS (upper
right corner) and preserved tubules with ongoing spermatogenesis. (B) A close-up of a tubule with nuclei of GCNIS cells stained
positive for OCT4, a pluripotency marker that is not present in normal adult germ cells and helps to detect GCNIS.
A B
facilitates detection and diagnosis. mas, respectively [23]. The miRNA- ing risk factors: history of cryptorchi-
GCNIS and seminoma cells re- based tests outperformed the classi- dism, poor semen quality, young age,
semble fetal gonocytes and have a cal serum markers in a large clinical testicular atrophy, and microlithiasis.
similar gene expression profile, char- study [34]. Efforts are under way to develop
acterized by high expression of em- a less invasive method than testicu-
bryonic pluripotency factors, such as lar biopsy for detection of GCNIS
Prevention of invasive or incipient microinvasive tumour.
POU5F1 (OCT4), NANOG, TFAP2C
(AP2-gamma), and LIN28 [17] (re-
cancer by early detection Such a method would preferably
viewed in [5]). This unusual profile is
of GCNIS require only a blood or semen sam-
partly explained by very low levels of Preventive measures are currently ple. An immunocytological detec-
DNA methylation of the genome of very limited, because of the uncer- tion method has been established,
GCNIS and seminoma, in contrast tainty about the causation of tes- using an automated double-stain-
to non-seminomas, which have high ticular cancer in the vast majority of ing assay for alkaline phosphatase
DNA methylation profiles, similar to cases. The most effective preven- and AP2-gamma or OCT4 in the
those of somatic cells [5,23,29–31]. tion strategy for invasive cancer is ejaculate, but further improvement
In addition, GCNIS cells are charac- early diagnosis at the pre-invasive of sensitivity is needed for routine
terized by permissive histone modifi- stage. This is currently possible use of this approach in the clinic
cations, which render their chromatin only in patients in high-risk groups, [35]. Novel serum assays exploiting
including individuals with disorders miRNAs have a very good specific-
accessible to transcription factors;
of sexual development, cryptorchi- ity and sensitivity for overt tumours
this could potentially explain their
dism, infertility, or other signs of [32–34], but it remains unclear
plasticity in response to environmen-
testicular dysgenesis. whether these tests will be sensitive
tal stimuli [1,30,31].
Unequivocal diagnosis of GCNIS enough to detect GCNIS or early
An important recent development
microinvasive tumours.
in the biology of testicular cancer is requires testicular biopsy (usually
the discovery of specific microRNAs bilateral) and immunohistochemi-
(miRNAs) secreted by malignant cal staining for at least one specific Fertility preservation and
germ cells, including GCNIS cells, marker (e.g. PLAP or OCT4) [5] prevention of late effects
both in adult men and in children (re- (Fig. 5.14.6). In about 5–6% of cas- Because testicular cancer occurs
viewed in [32,33]). The miRNA pro- es of seemingly unilateral testicular predominantly in young men and
file of malignant germ cells is char- germ cell tumours, GCNIS is present modern management means that
acterized by particularly high levels in the contralateral testis. Therefore, the prognosis is good, most sur-
of the miR-371-3 cluster, as well as a biopsy of the remaining testis is ad- vivors live for many decades after
miR-302 and miR-367 [32–34]. The vised at the time of orchidectomy for treatment. Therefore, the emphasis
presence of additional clusters, miR- the primary tumour, at least in men at has shifted from saving life to pre-
519 and miR-375, has been reported high risk, who are defined as present- serving quality of life. Even after be-
in embryonal carcinomas and terato- ing with more than one of the following declared cancer-free, survivors
436
should be followed up for many the overt tumour has developed, cers, cardiovascular disease, pe-
years, taking into account not only and in most men the situation wors- ripheral neuropathy, ototoxicity, and
the possibility of a late recurrence ens markedly after orchidectomy hepatotoxicity [38]. Also important
of the malignancy but also health or cytotoxic chemotherapy [1,37]. are quality-of-life issues related to
issues related to the lack of one Andrological follow-up is important, prolonged anxiety and stress. There
or both testes, such as subfertility, with close monitoring of testoster- is a growing consensus that individu-
hypogonadism, sexual dysfunction, one levels, because Leydig cell alized treatment is needed to dimin-
metabolic syndrome, and osteopo- dysfunction is common and the en- ish immediate and late side-effects,
rosis later in life, which result in de- suing hypogonadism is a major risk and attention should be paid to is-
creased life expectancy [36]. factor for metabolic syndrome [37]. sues related to reproductive health
Many patients with testicular In addition, patients treated with and quality of life.
cancer have poor spermatogenesis radiotherapy or chemotherapy have
and decreased fertility even before an increased risk of secondary can-
References
1. Skakkebaek NE, Rajpert-De Meyts E, 6. Bray F, Ferlay J, Soerjomataram I, Siegel 12. National Cancer Institute Surveillance,
Buck Louis GM, Toppari J, Andersson AM, RL, Torre LA, Jemal A (2018). Global Epidemiology, and End Results Program
Eisenberg ML, et al. (2016). Male repro- cancer statistics 2018: GLOBOCAN es- (2018). Cancer stat facts: testicular can-
ductive disorders and fertility trends: influ- timates of incidence and mortality world- cer. Available from: https://seer.cancer.
ences of environment and genetic suscep- wide for 36 cancers in 185 countries. CA gov/statfacts/html/testis.html.
tibility. Physiol Rev. 96(1):55–97. https:// Cancer J Clin. 68(6):394–424. https://doi.
d o i.o r g /10.1152 /p hys r ev. 0 0 017. 2 015 org/10.3322/caac.21492 PMID:30207593 13. Tandstad T, Kollmannsberger CK, Roth
PMID:26582516 BJ, Jeldres C, Gillessen S, Fizazi K, et
7. Znaor A, Lortet-Tieulent J, Jemal A, Bray F al. (2017). Practice makes perfect: the
2. Ulbright TM, Amin MB, Balzer B, Berney (2014). International variations and trends rest of the story in testicular cancer as a
DM, Epstein JI, Guo C, et al. (2016). Germ in testicular cancer incidence and mor- model curable neoplasm. J Clin Oncol.
cell tumours. In: Moch H, Humphrey PA, tality. Eur Urol. 65(6):1095–106. https:// 35(31):3525–8. https://doi.org/10.1200/
Ulbright TM, Reuter VE, editors. WHO clas- doi.org /10.1016/ j.eur uro. 2013.11.0 0 4 JCO.2017.73.4723 PMID:28854068
sification of tumours of the urinary system PMID:24268506
and male genital organs. Lyon, France: 14. Cherny NI, Sullivan R, Torode J, Saar
International Agency for Research on 8. Trabert B, Chen J, Devesa SS, Bray F, M, Eniu A (2017). ESMO International
Cancer (WHO Classification of Tumours McGlynn KA (2015). International pat- Consortium study on the availability, out-
series, 4th edition); pp. 189–226. terns and trends in testicular cancer inci- of-pocket costs and accessibility of anti-
dence, overall and by histologic subtype, neoplastic medicines in countries outside
3. Jørgensen A, Lindhardt Johansen M, Juul 1973-2007. Andrology. 3(1):4–12. https:// of Europe. Ann Oncol. 28(11):2633–47.
A, Skakkebaek NE, Main KM, Rajpert- doi.org/10.1111/andr.293 PMID:25331326 https://doi.org/10.1093/annonc/mdx521
De Meyts E (2015). Pathogenesis of PMID:28950323
germ cell neoplasia in testicular dysgen- 9. Patama T, Engholm G, Larønningen S,
esis and disorders of sex development. Ólafsdóttir E, Khan S, Storm H, et al. 15. Trama A, Foschi R, Larrañaga N, Sant
CHAPTER 5.14
M, Fuentes-Raspall R, Serraino D, et al.;
SECTION 5
Semin Cell Dev Biol. 45:124–37. https:// (2018). Small-area based map animations
doi.org/10.1016/j.semcdb.2015.09.013 of cancer incidence in the Nordic coun- EUROCARE-5 Working Group (2015).
PMID:26410164 tries, 1971–2015. Nordic Cancer Union. Survival of male genital cancers (pros-
Available from: https://astra.cancer.fi/ tate, testis and penis) in Europe 1999-
4. Giannoulatou E, Maher GJ, Ding Z, Gillis cancermaps/Nordic_18/. 2007: results from the EUROCARE-5
AJM, Dorssers LCJ, Hoischen A, et al. study. Eur J Cancer. 51(15):2206–16.
(2017). Whole-genome sequencing of 10. Bray F, Colombet M, Mery L, Piñeros M, https://doi.org/10.1016/j.ejca.2015.07.027
spermatocytic tumors provides insights Znaor A, Zanetti R, et al., editors (2017). PMID:26421823
into the mutational processes operat- Cancer incidence in five continents, Vol.
ing in the male germline. PLoS One. XI (electronic version). Lyon, France: 16. Cools M, Wolffenbuttel KP, Hersmus R,
12(5):e0178169. https://doi.org/10.1371/ International Agency for Research on Mendonca BB, Kaprová J, Drop SLS, et
journal.pone.0178169 PMID:28542371 Cancer. Available from: http://ci5.iarc.fr. al. (2017). Malignant testicular germ cell
tumors in postpubertal individuals with
5. Rajpert-De Meyts E, McGlynn KA, 11. Ghazarian AA, Trabert B, Devesa SS, androgen insensitivity: prevalence, pa-
Okamoto K, Jewett MAS, Bokemeyer McGlynn KA (2015). Recent trends in thology and relevance of single nucleo-
C (2016). Testicular germ cell tumours. the incidence of testicular germ cell tu- tide polymorphism-based susceptibility
Lancet. 387(10029):1762–74. https://doi. mors in the United States. Andrology. profiling. Hum Reprod. 32(12):2561–73.
org /10.1016/S0140 - 673 6(15) 0 0 9 91- 5 3(1):13–8. https://doi.org/10.1111/andr.288 https://doi.org/10.1093/humrep/dex300
PMID:26651223 PMID:25331158 PMID:29121256

17. Sonne SB, Almstrup K, Dalgaard M, 25. Litchfield K, Loveday C, Levy M, Dudakia 33. Nappi L, Nichols C (2019). MicroRNAs as
Juncker AS, Edsgard D, Ruban L, et al. D, Rapley E, Nsengimana J, et al. (2018). biomarkers for germ cell tumors. Urol Clin
(2009). Analysis of gene expression Large-scale sequencing of testicular germ North Am. 46(3):449–57. https://doi.org/10.
profiles of microdissected cell popu- cell tumour (TGCT) cases excludes ma- 1016/j.ucl.2019.04.011 PMID:31277739
lations indicates that testicular carci- jor TGCT predisposition gene. Eur Urol.
noma in situ is an arrested gonocyte. 73(6):828–31. https://doi.org/10.1016/j. 34. Dieckmann KP, Radtke A, Geczi L, Matthies
Cancer Res. 69(12):5241–50. https://doi. eururo.2018.01.021 PMID:29433971 C, Anheuser P, Eckardt U, et al. (2019).
org/10.1158/0008 - 5472.CAN - 08 - 4554 Serum levels of microRNA-371a-3p (M371
26. Litchfield K, Levy M, Huddart RA, Shipley test) as a new biomarker of testicular germ
PMID:19491264 J, Turnbull C (2016). The genomic land- cell tumors: results of a prospective multi-
18. Bonde JP, Flachs EM, Rimborg S, Glazer scape of testicular germ cell tumours: from centric study. J Clin Oncol. 37(16):1412–23.
CH, Giwercman A, Ramlau-Hansen susceptibility to treatment. Nat Rev Urol.
ht tps://doi.org/10.120 0/JCO.18.01480
CH, et al. (2016). The epidemiologic 13(7):409–19. https://doi.org/10.1038/
PMID:30875280
nrurol.2016.107 PMID:27296647
evidence linking prenatal and postnatal
exposure to endocrine disrupting chemi- 35. Almstrup K, Lippert M, Mogensen HO,
27. Litchfield K, Levy M, Orlando G, Loveday
cals with male reproductive disorders: C, Law PJ, Migliorini G, et al.; UK Testicular Nielsen JE, Hansen JD, Daugaard G, et
a systematic review and meta-analysis. Cancer Collaboration; PRACTICAL al. (2011). Screening of subfertile men for
Hum Reprod Update. 23(1):104–25. Consortium (2017). Identification of 19 testicular carcinoma in situ by an auto-
https://doi.org/10.1093/humupd/dmw036 new risk loci and potential regulatory mated image analysis-based cytological
PMID:27655588 mechanisms influencing susceptibility test of the ejaculate. Int J Androl. 34(4
to testicular germ cell tumor. Nat Genet. Pt 2):e21–30, discussion e30–1. https://
19. Stang A, Trabert B, Rusner C, Poole C, 49(7):1133–40. https://doi.org/10.1038/ doi.org/10.1111/j.1365-2605.2011.01192.x
Almstrup K, Rajpert-De Meyts E, et al. ng.3896 PMID:28604728 PMID:21696398
(2015). A survey of etiologic hypoth-
eses among testicular cancer research- 28. Wang Z, McGlynn KA, Rajpert-De Meyts 36. Capocaccia R, Gatta G, Dal Maso
ers. Andrology. 3(1):19–26. https://doi. E, Bishop DT, Chung CC, Dalgaard MD, et L (2015). Life expectancy of colon,
org/10.1111/andr.306 PMID:25538016 al.; Testicular Cancer Consortium (2017). breast, and testicular cancer patients:
Meta-analysis of five genome-wide as- an analysis of US-SEER population-
20. Gurney J, Shaw C, Stanley J, Signal V, sociation studies identifies multiple new based data. Ann Oncol. 26(6):1263–8.
Sarfati D (2015). Cannabis exposure and loci associated with testicular germ cell https://doi.org/10.1093/annonc/mdv131
risk of testicular cancer: a systematic re- tumor. Nat Genet. 49(7):1141–7. https:// PMID:25735314
view and meta-analysis. BMC Cancer. doi.org/10.1038/ng.3879 PMID:28604732
15(1):897. https://doi.org/10.1186/s12885- 37. Bandak M, Jørgensen N, Juul A, Lauritsen
29. Rijlaarsdam MA, Tax DM, Gillis AJ, J, Oturai PS, Mortensen J, et al. (2017).
015-1905-6 PMID:26560314
Dorssers LC, Koestler DC, de Ridder J, Leydig cell dysfunction, systemic inflam-
21. Ylönen O, Jyrkkiö S, Pukkala E, Syvänen et al. (2015). Genome wide DNA meth- mation and metabolic syndrome in long-
K, Boström PJ (2018). Time trends and ylation profiles provide clues to the origin
term testicular cancer survivors. Eur J
occupational variation in the incidence and pathogenesis of germ cell tumors.
Cancer. 84:9–17. https://doi.org/10.1016/j.
of testicular cancer in the Nordic coun- PLoS One. 10(4):e0122146. https://
ejca.2017.07.006 PMID:28772110
tries. BJU Int. 122(3):384–93. https://doi. doi.org/10.1371/journal.pone.0122146
org/10.1111/bju.14148 PMID:29460991 PMID:25859847 38. Chovanec M, Abu Zaid M, Hanna N, El-
30. Lawaetz AC, Almstrup K (2015). Kouri N, Einhorn LH, Albany C (2017).
22. Hemminki K, Chen B (2006). Familial risks Long-term toxicity of cisplatin in germ-cell
in testicular cancer as aetiological clues. Involvement of epigenetic modifiers in
the pathogenesis of testicular dysgenesis tumor survivors. Ann Oncol. 28(11):2670–
Int J Androl. 29(1):205–10. https://doi. 9. https://doi.org/10.1093/annonc/mdx360
and germ cell cancer. Biomol Concepts.
org/10.1111/j.1365 -2605.2005.00599.x PMID:29045502
6(3):219–27. https://doi.org/10.1515/bmc-
PMID:16466541
2015-0006 PMID:26103631
23. Shen H, Shih J, Hollern DP, Wang L,
31. Lobo J, Gillis AJM, Jerónimo C, Henrique
Bowlby R, Tickoo SK, et al.; Cancer R, Looijenga LHJ (2019). Human germ
Genome Atlas Research Network (2018). cell tumors are developmental cancers:
Integrated molecular characterization impact of epigenetics on pathobiology
of testicular germ cell tumors. Cell Rep. and clinic. Int J Mol Sci. 20(2):E258.
23(11):3392–406. https://doi.org/10.1016/j. https://doi.org/10.339 0/ijms20020258
celrep.2018.05.039 PMID:29898407 PMID:30634670
24. Taylor-Weiner A, Zack T, O’Donnell E, 32. Murray MJ, Huddart RA, Coleman N
Guerriero JL, Bernard B, Reddy A, et al. (2016). The present and future of serum
(2016). Genomic evolution and chemore- diagnostic tests for testicular germ cell
sistance in germ-cell tumours. Nature. tumours. Nat Rev Urol. 13(12):715–25.
540(7631):114–8. https://doi.org/10.1038/ https://doi.org/10.1038/nrurol.2016.170
nature20596 PMID:27905446 PMID:27754472
438
5.15
Bladder cancer
A genotoxic causal agent recognized
Joëlle L. Nortier Wolfgang A. Schulz (reviewer)

Thierry Roumeguère Jiri Zavadil (reviewer)
[1]. Like tumours of the renal pelvis sion and mutation patterns, includ-
SUMMARY and ureter, tumours of the bladder ing aggressive histological variants
are derived from transitional epithe- with poor response to existing thera-
●● More than 90% of bladder can- lia. Together, these tumour types ac- pies. Muscle-invasive bladder can-
cers are urothelial carcinomas, count for 10–15% of all primary ma- cers are heterogeneous and can be
which are usually staged as ei- lignancies in adults. These urothelial grouped into the basal and luminal
ther muscle-invasive tumours, carcinomas are multicentric in na- intrinsic subtypes [5].
which have a poorer prognosis, ture and often occur – and recur – Five expression subtypes have
or non-muscle-invasive tumours, at multiple sites in the lower urinary been identified that may stratify re-
which have a better prognosis tract in an affected patient. The wall sponse to different treatments. The
but frequently recur. of the bladder is the most common luminal-papillary subtype is char-
●● In addition to causes includ- site of involvement. acterized by FGFR3 mutations, fu-
ing inhaled tobacco smoke and sions with TACC3, and/or amplifica-
certain occupational exposures, Molecular subtypes tion. The luminal-infiltrated subtype
aristolochic acid is now recog- is characterized by high expression
Significant differences in patient
nized as causing bladder can- characteristics, incidence, and sur- of epithelial–mesenchymal transi-
cer, possibly in association with vival exist, and research is continu- tion and myofibroblast markers, with
renal failure. ing on gene–environment interac- medium expression of PD-L1 and
tions with risk of bladder cancer [2]. CTL4 immune markers. The lumi-
●● Aristolactam–DNA adducts and nal subtype has high expression of
a specific mutational signature Urothelial carcinoma is the most
common type of bladder cancer, but luminal markers, as well as KRT20
(A:T → T:A transversion), initial-
distinct histomorphological pheno- and SNX31. The basal-squamous
ly discovered in the TP53 gene,
types have been reported (10–25%) subtype is characterized by a higher
may serve as biomarkers of ex-
that are associated with more ag- incidence in women, squamous dif-
posure to aristolochic acid.
gressive disease and poor response ferentiation, basal keratin expres-
●● With the increasing use of to existing therapies [3]. These can- sion, and high expression of PD-L1
large-scale genome-wide pro- cers are usually staged as either and CTLA4 immune markers. The
CHAPTER 5.15
SECTION 5
filing studies, the conventional non-muscle-invasive tumours (~75%) neuronal subtype is characterized
two-pathway model of bladder or muscle-invasive tumours (~25%). by expression of both neuroendo-
cancer pathogenesis is being The Cancer Genome Atlas crine and neuronal genes, as well as
superseded by a molecular de- (TCGA) project identified genetic a high cell-cycle signature, reflective
scription of disease pathogen- drivers for muscle-invasive bladder of a proliferative state [6].
esis and clinical behaviour. This cancer as well as clusters associ- The identification of multiple
approach should provide more ated with distinct prognostic factors distinct molecular subtypes of
adequate information for per- and therapeutic responses [4]. The non-muscle-invasive and muscle-
sonalized clinical and therapeu- TCGA Research Network reported invasive bladder cancer suggests
tic management. the major genetic determinants of multiple pathways within each of
muscle-invasive bladder cancer and the major pathways. Development
showed that bladder cancer can be of histopathologically recognizable
Bladder cancer causes an estimated further subclassified at the molecu- urothelial alterations is preceded
199 900 deaths per year worldwide lar level according to gene expres- by clonal expansion of altered cells
Chapter 5.15 • Bladder cancer 439

within the urothelium. Low-grade Exposure to arsenic through

papillary tumours may arise via contaminated groundwater sources FUNDAMENTALS
simple hyperplasia and minimal (see Chapter 2.9) and also through
dysplasia, and these are charac- food (such as rice and seafood) is ■■ Bladder cancer is the 12th
terized at the molecular level by a public health problem in many most common cancer type
loss of heterozygosity of chromo- countries. It is estimated that more worldwide, and urothelial
some 9 and activating mutations than 200 million people in 70 coun- carcinoma is the most common
of FGFR3, PIK3CA, and STAG2. tries are chronically exposed to ar- tumour type. Most patients are
These non-invasive tumours fre- senic at levels at or above the WHO
diagnosed with non-invasive
quently recur but are genetically threshold of 10 µg/L, leading to car-
and low-grade tumours.
stable [7]. Invasive carcinoma is diovascular, pulmonary, and skin
thought to arise via flat dysplasia diseases and also different types ■■ Tobacco smoking is the most
and carcinoma in situ, which com- of cancer, including bladder cancer important cause of bladder
monly show TP53 mutations in ad- and urinary tract cancer [9]. cancer. Arsenic and some
dition to chromosome 9 deletions Arsenic is classified by the occupational exposures also
but no FGFR3 mutations. Invasive IARC Monographs as carcinogenic cause bladder cancer.
tumours are genetically unstable to humans (Group 1). Mechanisms
and accumulate many genomic of arsenic carcinogenesis are com- ■■ Aristolochic acid, a constituent
alterations, such as RB1 loss and plex and are not fully understood. of all Aristolochia plants,
ERBB2 or PTEN mutations [6]. According to cancer studies con- is a powerful nephrotoxin
ducted mainly in endemic areas of and human carcinogen
arsenic contamination (Argentina, and causes, among other
Epidemiology Bangladesh, northern Chile, and diseases, bladder cancer and
Bladder cancer is a highly preva- Taiwan, China), the mechanisms renal cell carcinoma.
lent disease and is associated with involve oxidative stress and DNA
substantial morbidity, mortality, and damage, epigenetic DNA modifica- ■■ Cystoscopy enables a
cost. Tobacco smoking and occu- tion, and genomic instability [10]. definitive diagnosis of bladder
pational exposures to carcinogens Aristolochic acid, a constituent cancer. Prognosis and
remain the factors with the highest of all Aristolochia plants, is a pow- management of bladder cancer
attributable risk. In 2018, there were erful nephrotoxin and human car- depend on histopathology.
an estimated 549 000 new cases of cinogen, which is associated with
bladder cancer and 199 900 deaths chronic kidney disease and upper ■■ Some evidence supports
from bladder cancer globally; blad- urinary tract urothelial carcinoma a genetic predisposition to
der cancer was the 12th most com- as well as bladder cancer. The term bladder cancer, and genome-
mon cancer type and the 12th most “aristolochic acid nephropathy” ac- wide association studies have
common cause of cancer death tually includes any form of toxic in- found sequence variants
worldwide [1]. terstitial nephropathy that is caused that can increase the risk
Classical epidemiological stud- either by the ingestion of plants of bladder cancer and of
ies have confirmed a markedly in- containing aristolochic acid as part chemoresistance.
creased incidence of bladder can- of traditional phytotherapies (for-
cer in workers exposed to various merly known as “Chinese herbs ne- ■■ Immunotherapy with checkpoint
aromatic amines used in the dyeing, phropathy”) or by the environmental inhibitors has revolutionized the
chemical, and rubber industries. contamination of food (known as treatment paradigm of bladder
Besides these occupational expo- “Balkan endemic nephropathy”) [11] cancer; since 2016, five agents
sures, inhaled tobacco smoke is (see Chapter 2.8). have been approved to treat
the most prominent environmental In addition to its nephrotoxic ef- platinum-refractory bladder
carcinogen known to cause bladder fects, possibly leading to end-stage cancer.
cancer (see Chapter 2.1). Additional renal disease, exposure to aristo-
agents include arsenic exposure lochic acid has frequently been as-
from contaminated water in endemic sociated with the development of
areas for blackfoot disease (a type urothelial malignancies. Aristolochic gens” by the United States National
of peripheral vasculitis) in south- acid (and plants containing it) was Toxicology Program in 2014 [13].
western Taiwan, China. Moreover, a classified by the IARC Monographs Since the identification of aris-
high incidence of bladder cancer of as carcinogenic to humans (Group 1) tolochic acid nephropathy in the
the squamous type has been found in 2008, after an earlier evaluation in early 1990s in Belgium, an increas-
in patients with chronic parasitic in- 2002 [12]. This finding is consistent ing number of cases of aristolochic
festation due to Schistosoma hae- with aristolochic acids being listed acid intoxication have been reported
matobium [8]. as “known to be human carcino- around the world [14]. The incidence
440
of upper urinary tract urothelial car- cur in Lynch syndrome, are an active cific biomarker of exposure, even
cinoma is particularly high in Asian area of research. Lynch syndrome more than 10 years after exposure
countries, including specifically in is an inherited condition that in- to aristolochic acid. They are also
Taiwan, China, because traditional creases the risk of cancers, includ- found in the urothelium, where they
medicines are very popular and the ing urothelial carcinoma. Screening give rise to a unique mutational sig-
complexity of the pharmacopoeia of patients known to have Lynch nature in the TP53 gene and gener-
presents a high risk of aristolo- syndrome is important, to evaluate ally (Fig. 5.15.2).
chic acid intoxication, as a result of for the development of primary tu- This A:T → T:A transversion –
some confusion between closely mours. Inherited mutations in DNA also called COSMIC signature 22 –
related species [15]. In the Balkan repair genes confer a greater risk of has frequently been detected in
countries, the causative factor was urothelial carcinoma. Additional re- cases of upper urinary tract urothe-
identified as the environmental phy- search is needed to evaluate the op- lial carcinoma described in the
totoxin aristolochic acid contained in timal frequency and type of screen- Balkans and in Taiwan, China [20],
Aristolochia clematitis, a common ing for individual patients [18]. whereas this mutation rarely occurs
plant growing in the wheat fields, Genome-wide association stud- in tumours that are not related to ex-
which was ingested in home-baked ies (GWAS) (see Chapter 3.2) have posure to aristolochic acid [15,21].
bread [16] (Fig. 5.15.1). found sequence variants that can In Taiwan, China, such mutations
The nephrotoxic effect of aristo- increase the risk of bladder can- were also detected at activating
lochic acid is irreversible. Given that cer. Most of the significant vari- positions in the FGFR3 and HRAS
chronic kidney disease and carci- ants associated with risk of blad- oncogenes. Extensive analyses of
der cancer are located in DNA mutation spectra from bladder can-
nogenic complications may develop
repair genes. Polymorphisms for cer cases in Singapore and Taiwan,
very slowly after the initial exposure,
GSTM1-null, NAT2-slow, APOBEC- China, suggested a strong involve-
aristolochic acid nephropathy and as-
rs1014971, SLC14A1-rs10775480, ment of aristolochic acid in bladder
sociated upper urinary tract urothelial
CCNE1-rs8102137, PSCA-rs2294008, cancer development in Asian coun-
carcinoma and bladder cancer may
UGT1A-rs1189203, and TP63 - tries, indicating an important public
become a major public health issue
rs35592567 confer increased risk health issue [22].
in the next few years [17].
[19].
Mutational signatures of
Genetics and genomics Mutational signature of tobacco smoking
aristolochic acid The mechanisms of tobacco car-
Genetic susceptibility After metabolic activation, aristolo- cinogenesis are very complex and
Some evidence supports a genetic chic acid reacts with DNA to form may vary between tumour sites.
predisposition to bladder cancer. aristolactam–DNA adducts. These Comparative studies of cancer ge-
Potential inheritable forms of blad- lesions concentrate in the renal cor- nome sequences from smokers and
der cancer, such as those that oc- tex, serving as a sensitive and spe- non-smokers found that smokers had
Fig. 5.15.1. (Left) Aristolochia clematitis blossoming in grasslands in Serbia, and (right) the corresponding fruits and seeds collected
during the harvest of the wheat crop from fields in the same area. A. clematitis is recognized as the causal agent of Balkan endemic
nephropathy and is associated with upper urinary tract urothelial carcinoma and bladder cancer.
CHAPTER 5.15
SECTION 5

Fig. 5.15.2. Metabolic activation and DNA adduct formation by aristolochic acid (AA). fangchi (known as Guang Fang Ji),
R = OCH3 in AAI, and R = H in AAII. COX, cyclooxygenase; CYP, cytochrome P450; dA- because they are morphologically
AAI, 7-(deoxyadenosin-N 6-yl)aristolactam I; dA-AAII, 7-(deoxyadenosin-N 6-yl)aristolac- similar (Fig. 5.15.3).
tam II; dG-AAI, 7-(deoxyguanosin-N 2-yl)aristolactam I; dG-AAII, 7-(deoxyguanosin-N 2-yl)
Originally, aristolochic acid ne-
aristolactam II; NQO1, NAD(P)H:quinone oxidoreductase; POR, NADPH:cytochrome
P450 oxidoreductase. phropathy was reported in Belgium
in more than 100 individuals who
O had ingested weight-loss capsules
O
O 4 C OOH
O
NH
O
containing powdered root extracts
5 O
O
N O2
Metabolic N
DNA adduct
O
HN
N
of Aristolochia fangchi. The causal
link with the intake of capsules
6
O
activation formation N
H
N N
7
dG-AAI
containing aristolochic acid was
8 DNA
NQO1 +DNA R
dG-AAII
9
POR +
R
COX R
O
dA-AAI (persistent)
demonstrated by the detection of
CYP1A1/2
Aristolochic acid I, II
AAI: R=OCH3
Aristolactam I, II
nitrenium ion
O dA-AAII aristolactam–DNA adducts in renal
NH
AAII: R=H O N N tissue samples. It is estimated that
N N
DNA
exposure to aristolochic acid af-
H
R
N fects 100 000 people in the Balkans
Mutation
induction
(where the total number of pa-
tients with kidney disease is about
A:T to T:A transversions
Urothelial
(in TP53 and other 25 000), 8 million people in Taiwan,
carcinoma
cancer-related genes) China, and more than 100 million
people in China [16].
In the initial cohorts for iatrogenic
aristolochic acid nephropathy, the
higher numbers of base substitutions process) in some susceptible tis- majority of patients were described
compared with non-smokers [23]. In sues, in particular in tissues directly as exhibiting a rapid and progressive
tumours of tissues directly exposed exposed to tobacco smoke [23,25]. evolution towards chronic kidney dis-
to tobacco smoke (the lung and the ease or end-stage renal disease [14].
larynx), COSMIC signature 4 was DNA methylation in urothelial In environmental aristolochic acid
prominent. This signature is similar carcinoma nephropathy, the progression rate
to that produced by benzo[a]pyrene Potential epigenetic signatures, is much slower, reaching end-stage
in cells in vitro and suggests a mis- mainly for DNA methylation alter- renal disease after 15–20 years [27].
replication of DNA damage (adducts) ations but also for mutations in chro- Activities such as mining, com-
formed by carcinogens present in matin regulators, have been linked bustion of fossil fuels, and the use
tobacco smoke. Other signatures, to specific carcinogens (see Chapter of arsenic-based pesticides are
such as signature 2 (which features 3.11). Their validation as potential known to potentiate the environ-
GC → AT mutations) and signature biomarkers in urine or tissue samples mental accumulation of arsenic.
13 (which features GC → CG muta- is still required [26]. This presents a major threat to hu-
tions), are considered to reflect an man health because exposure of
over-reactivity of the APOBEC family individuals through inhalation, inges-
of cytidine deaminases in DNA edit- Etiology tion, and skin contact can result in
ing [24]. A multiplatform analysis of numerous adverse health effects [9].
Risk factors
more than 400 patients with muscle- Consumption of drinking-water from
invasive bladder cancer confirmed In Asia, Aristolochia species are contaminated groundwater sources
a high mutational load driven by considered an integral part of the and ingestion of contaminated food
APOBEC-mediated mutagenesis. herbology used in traditional Chi (fish and grains) are the major routes
The detection of this signature cor- nese medicine, Japanese Kampō of human exposure. Biological fac-
responded to a 5-year survival rate medicine, and Ayur vedic medicine. tors (sex, race, and age) and lifestyle
of 75% [6]. Aristolochia is part of the same factors (nutrition and smoking sta-
Signature 5 is found in all tumour therapeutic family as the Akebia, tus) may influence the efficacy of the
types related to smoking and has Asarum, Cocculus, and Stephania pathways implicated in arsenic me-
a predominance of AT → GC and plants. These plants are referred tabolism and cytotoxic outcome, re-
GC → AT mutations. In smokers, to by common names such as sulting in inter-individual variations in
the frequency of mutations attribu- Mu Tong, Mokutsu, and Fang Ji, susceptibility to arsenic toxicity [9,10].
table to signature 5 has been found and they are used in a multitude
to increase with age at diagnosis; of herbal mixtures for therapeutic Evaluation and diagnosis
this has been suggested to reflect use. Stephania tetrandra (known as Patients suspected of having blad-
an acceleration of endogenous Han Fang Ji) is sometimes mistak- der cancer are usually evaluated by
mutagenic processes (a “clocklike” enly substituted with Aristolochia white-light cystoscopy, with adjunct
442
Fig. 5.15.3. (Left) Leaves of the Stephania tetrandra S. Moore plant (known as Han Fang Ji in traditional Chinese medicine). (Middle)
Leaves of the Aristolochia elegans plant; the shape is similar to that of Stephania tetrandra leaves. (Right) Transverse sections of
the roots of Aristolochia fangchi (known as Guang Fang Ji in traditional Chinese medicine) and weightloss capsules containing the
powdered root, ingested by a Belgian patient who developed end-stage renal disease. A. fangchi is recognized as the causal agent of
aristolochic acid-induced severe to end-stage renal failure and multifocal urothelial carcinoma in Belgium.
cytology performed to detect malig- The possibility of using circulat- prevention in populations at risk, in
nant cells. To date, no urinary-based ing tumour cells as a means, among the form of intensified screening.
tumour markers have demonstrated other things, to detect bladder cancer Recurrent prevention campaigns
sufficient sensitivity and specificity has been discussed [30]. Methylation can provide information about can-
to replace cystoscopy in the detec- markers in urine have been described cers related to tobacco smoking. In
tion of bladder cancer. for detection of bladder cancer, but contrast, measures to fight environ-
Cystoscopic detection may be the diagnostic accuracy is highly vari- mental arsenic contamination are
enhanced by optical imaging tech- able among reports [31]. difficult to implement. Specific equip-
nologies such as fluorescence cys- ment to remove arsenic from con-
toscopy or narrow-band imaging. taminated water is of poor efficiency
These technologies improve the dif-
Prevention (activated carbon-based filters) or
ferentiation of tumorous lesions from Reduced exposure to expensive (reverse osmosis). Other
normal tissue by taking advantage carcinogens approaches have been proposed on
of the increased metabolic activity the basis of animal studies: metal
(blue light) and vessel architecture With respect to urothelial malignan- chelators (partially successful), vita-
(narrow-band) that occur in cancer cies associated with aristolochic mins (vitamin C, vitamin B12, and fo-
cells, and they have higher specific- acid (Fig. 5.15.5), primary prevention lic acid) and trace elements for their
ity for bladder cancer than tradition- through regulation and education is antioxidant properties, glutathione
al cystoscopy does. Especially the possible. However, the general pop- as an antioxidant and an inhibitor of
detection rate of carcinoma in situ ulation considers traditional herbal reactive oxygen species, and plant-
could be significantly increased by remedies to be harmless because derived polyphenols with antioxidant
the use of these methods. they are of natural origin. Moreover, properties [10]. To date, only a few
CHAPTER 5.15
SECTION 5
Microscopic imaging techniques most patients who use these natural of these have been tested in a clini-
like confocal laser endomicroscopy products fail to inform their physicians cal setting. Because the proportions
and optical coherence tomography of their use. Therefore, these natural of possible responders vary among
permit a real-time high-resolution as- products, like all drugs, should be subgroups of the population, some
sessment of the bladder mucosa at submitted to rigorous pharmacologi- biomarker-based screening pro-
a cellular and subcellular level with cal and toxicological studies to deter- grammes are likely to be developed
spatial resolutions similar to those of mine their safety and efficacy. for individuals with high health risk
histology, but these techniques are In addition to opportunities for and arsenic exposure.
not yet approved for routine use in primary prevention, detection of
the diagnosis of bladder cancer [28]. exposure to aristolochic acid by Screening
Prognosis and management of blad- the use of molecular epidemiology No major organization recommends
der cancer depend on histopathology, studies (biomarkers and endog- screening asymptomatic adults for
the only reliable determining factor of enous mutagenic processes) would bladder cancer, and current evi-
tumour biology (Fig. 5.15.4) [29]. provide opportunities for secondary dence is insufficient to assess the

Fig. 5.15.4. Histological aspect of bladder carcinoma in situ observed adjacent to high-grade papillary urothelial carcinoma in a
Belgian patient who underwent a kidney transplant for end-stage aristolochic acid nephropathy. Urothelial carcinoma in situ is
characterized by flat, disordered proliferation of urothelial cells with marked cytological abnormalities. Haematoxylin and eosin
staining; magnification 100× (left) and 400× (right).
balance of benefits and harms of cinogens, such as aromatic amines. Immunotherapy with programmed
screening. However, non-random- UGT1A is involved in tumour pro- cell death 1 (PD-1) and programmed
ized trials have demonstrated the gression, and decreased levels of death-ligand 1 (PD-L1) checkpoint in-
ability to detect bladder cancer in UGT1A are associated with recur- hibitors has revolutionized the treat-
selected populations, such as those rence and progression of bladder ment paradigm of bladder cancer.
exposed to aristolochic acid [32,33]. cancer. UGT1A is differentially reg- Since 2016, five agents have been
ulated by estrogens, and androgen- approved to treat platinum-refractory
Improved methods of mediated signals promote bladder bladder cancer. The approved PD-1
detection and diagnosis and PD-L1 inhibitor agents have simi-
carcinogenesis by downregulating
Several urine biomarkers exist, but the expression of UGTs [36,37]. lar efficacy and safety profiles. There
until now these have had a limited is a lack of consensus on the utility
role for the detection of bladder Improved therapeutic of testing for PD-L1 as a predictive
cancer. Emerging studies have strategies biomarker, because patients with no
been published proposing panels of For nearly 30 years, the first-line expression also derive some clinical
protein biomarkers for the detection standard of care treatment for pa- benefit. Tumour mutation burden is
of bladder cancer, and the diagnos- another putative predictive biomark-
tients with locally advanced or me-
tic performance of multiplex urinary er, but further validation is needed
tastatic bladder cancer has been
protein profiling could be improved [38]. The improved tolerability of
cisplatin-containing combination
when it is combined with clinical in- immunotherapy over chemothera-
chemotherapy. The median survival
formation about the patient, such as py and radiation directly correlates
is now approximatively 15 months,
age, race, and smoking status [34]. with its targeted mechanism of ac-
compared with the estimated surviv-
tion. The current landscape is rapidly
New research paths al of 6 months for patients with me- evolving, and novel immunotherapy
Epidemiological studies have shown tastatic disease before the develop- combination trials are under way to
differences between the sexes in ment of modern chemotherapy. The further improve outcomes and define
the incidence and progression of 5-year survival rate with contempo- the ideal patients [39].
bladder cancer, suggesting an as- rary regimens remains poor, at 15%. With the increasing use of large-
sociation with steroid hormone About 21% of patients are treated scale genome-wide profiling studies,
pathways; therefore, the role of sex with cisplatin-based chemothera- the conventional two-pathway model
steroids is an emerging research py, and cisplatin ineligibility is com- of bladder cancer pathogenesis is
area in the development and pro- mon because of renal dysfunction, being superseded by a molecular
gression of bladder cancer [35]. an Eastern Cooperative Oncology description of disease pathogenesis
A member of the family of UDP- Group (ECOG) performance status and clinical behaviour. This approach
glucuronosyltransferases (UGTs), of 2, or both. Hearing loss, grade 2 should provide more adequate infor-
UGT1A, is an enzyme that is vital neuropathy, and heart failure may mation for personalized clinical and
for the detoxification of major car- also confer cisplatin ineligibility. therapeutic management.
444
Fig. 5.15.5. Nested variant of bladder carcinoma infiltrating the muscle wall in another Belgian kidney transplant recipient. Haematoxylin
and eosin staining; magnification 100× (left) and 400× (right).
References
1. Bray F, Ferlay J, Soerjomataram I, Siegel 7. Hurst CD, Platt FM, Taylor CF, Knowles 13. National Toxicology Program (2014).
RL, Torre LA, Jemal A (2018). Global MA (2012). Novel tumor subgroups of Aristolochic acids. Report on carcino-
cancer statistics 2018: GLOBOCAN es- urothelial carcinoma of the bladder de- gens. Research Triangle Park (NC),
timates of incidence and mortality world- fined by integrated genomic analysis. USA: Department of Health and Human
wide for 36 cancers in 185 countries. CA Clin Cancer Res. 18(21):5865–77. https:// Services; pp. 1–5. Available from: https://
Cancer J Clin. 68(6):394–424. https://doi. doi.org/10.1158/1078-0432.CCR-12-1807 ntp.niehs.nih.gov/ntp/roc/content/profiles/
org/10.3322/caac.21492 PMID:30207593 PMID:22932667 aristolochicacids.pdf.
2. Cumberbatch MGK, Jubber I, Black PC, 8. Miyazaki J, Nishiyama H (2017). 14. Gökmen MR, Cosyns JP, Arlt VM,
Esperto F, Figueroa JD, Kamat AM, et al. Epidemiology of urothelial carcinoma. Int J Stiborová M, Phillips DH, Schmeiser
(2018). Epidemiology of bladder cancer: Urol. 24(10):730–4. https://doi.org/10.1111/ HH, et al. (2013). The epidemiology,
a systematic review and contemporary iju.13376 PMID:28543959 diagnosis, and management of aristo-
update of risk factors in 2018. Eur Urol. lochic acid nephropathy: a narrative re-
74(6):784–95. https://doi.org/10.1016/j. 9. Minatel BC, Sage AP, Anderson C, Hubaux
view. Ann Intern Med. 158(6):469–77.
eururo.2018.09.001 PMID:30268659 R, Marshall EA, Lam WL, et al. (2018).
https://doi.org/10.7326/0003-4819-158-6-
Environmental arsenic exposure: from ge-
3. Willis D, Kamat AM (2015). Nonurothelial 201303190-00006 PMID:23552405
netic susceptibility to pathogenesis. Environ
bladder cancer and rare variant his-
Int. 112:183–97. https://doi.org/10.1016/j. 15. Chen CH, Dickman KG, Moriya M, Zavadil
tologies. Hematol Oncol Clin North Am.
envint.2017.12.017 PMID:29275244 J, Sidorenko VS, Edwards KL, et al. (2012).
29(2):237–52, viii. https://doi.org/10.1016/j.
hoc.2014.10.011 PMID:25836932 Aristolochic acid-associated urothelial
10. Rao CV, Pal S, Mohammed A, Farooqui
cancer in Taiwan. Proc Natl Acad Sci U S
M, Doescher MP, Asch AS, et al. (2017).
4. Cancer Genome Atlas Research Network A. 109(21):8241–6. https://doi.org/10.1073/
CHAPTER 5.15
Biological effects and epidemiological
SECTION 5
(2014). Comprehensive molecular charac- pnas.1119920109 PMID:22493262
consequences of arsenic exposure, and
terization of urothelial bladder carcinoma.
Nature. 507(7492):315–22. https://doi. reagents that can ameliorate arsenic dam- 16. Bunel V, Souard F, Antoine M-F, Stévigny C,
org/10.1038/nature12965 PMID:24476821 age in vivo. Oncotarget. 8(34):57605–21. Nortier JL (2018). Nephrotoxicity of natural
https://doi.org/10.18632/oncotarget.17745 products: aristolochic acid and fungal toxins.
5. Choi W, Czerniak B, Ochoa A, Su X, PMID:28915699 In: McQueen CA, editor. Comprehensive
Siefker-Radtke A, Dinney C, et al. (2014). toxicology. Volume 14. 3rd ed. Oxford,
Intrinsic basal and luminal subtypes of 11. Jadot I, Declèves A-E, Nortier J, Caron
UK: Elsevier Ltd; pp. 340–79. https://doi.
muscle-invasive bladder cancer. Nat Rev N (2017). An integrated view of aris-
org/10.1016/B978-0-12-801238-3.64093-X
Urol. 11(7):400–10. https://doi.org/10.1038/ tolochic acid nephropathy: update of
nrurol.2014.129 PMID:24960601 the literature. Int J Mol Sci. 18(2):1–23. 17. Grollman AP (2013). Aristolochic acid
https://doi.org/10.339 0/ijms18020297 nephropathy: harbinger of a global iat-
6. Robertson AG, Kim J, Al-Ahmadie H, PMID:28146082 rogenic disease. Environ Mol Mutagen.
Bellmunt J, Guo G, Cherniack AD, et
54(1):1–7. https://doi.org/10.1002/em.21756
al.; TCGA Research Network (2017). 12. IARC (2002). Some traditional herbal med-
PMID:23238808
Comprehensive molecular characteriza- icines, some mycotoxins, naphthalene and
tion of muscle-invasive bladder cancer. styrene. IARC Monogr Eval Carcinog Risks
Cell. 171(3):540–556.e25. https://doi.org/ Hum. 82:1–556. Available from: http://pub
10.1016/j.cell.2017.09.007 PMID:28988769 lications.iarc.fr/100 PMID:12687954

18. Phelan A, Lopez-Beltran A, Montironi 26. Schulz WA, Goering W (2016). DNA methyl- 33. Starke N, Singla N, Haddad A, Lotan Y
R, Zhang S, Raspollini MR, Cheng M, ation in urothelial carcinoma. Epigenomics. (2016). Long-term outcomes in a high-risk
et al. (2018). Inherited forms of bladder 8(10):1415–28. https://doi.org/10.2217/epi- bladder cancer screening cohort. BJU
cancer: a review of Lynch syndrome and 2016-0064 PMID:27624974 Int. 117(4):611–7. https://doi.org/10.1111/
other inherited conditions. Future Oncol. bju.13154 PMID:25891519
14(3):277–90. https://doi.org/10.2217/ 27. Jelaković B, Dika Ž, Arlt VM, Stiborova M,
fon-2017-0346 PMID:29345160 Pavlović NM, Nikolić J, et al. (2019). Balkan 34. Szarvas T, Nyirády P, Ogawa O, Furuya
endemic nephropathy and the causative H, Rosser CJ, Kobayashi T (2018).
19. de Maturana EL, Rava M, Anumudu C, role of aristolochic acid. Semin Nephrol. Urinary protein markers for the detection
Sáez O, Alonso D, Malats N (2018). Bladder 39(3):284–96. https://doi.org/10.1016/j. and prognostication of urothelial carci-
cancer genetic susceptibility. A system- semnephrol.2019.02.007 PMID:31054628 noma. Methods Mol Biol. 1655:251–273.
atic review. Bladder Cancer. 4(2):215–26. https://doi.org/10.1007/978-1-4939-7234-
ht t p s: //d o i .o r g /10 . 3 2 3 3 / B LC -17015 9 28. Schubert T, Rausch S, Fahmy O, Gakis 0_19 PMID:28889391
G, Stenzl A (2017). Optical improve-
PMID:29732392
ments in the diagnosis of bladder can- 35. Godoy G, Gakis G, Smith CL, Fahmy O
20. Stiborová M, Arlt VM, Schmeiser HH cer: implications for clinical practice. (2016). Effects of androgen and estrogen
(2017). DNA adducts formed by aristolochic Ther Adv Urol. 9(11):251–60. https:// receptor signaling pathways on blad-
acid are unique biomarkers of exposure d o i . o r g /10 .117 7/175 6 2 8 7 217 7 2 0 4 01 der cancer initiation and progression.
and explain the initiation phase of upper PMID:29662543 Bladder Cancer. 2(2):127–37. https://doi.
urothelial cancer. Int J Mol Sci. 18(10):2144. org/10.3233/BLC-160052 PMID:27376135
29. Soukup V, Čapoun O, Cohen D, Hernández
ht tps://doi.org/10.339 0/ijms1810214 4
V, Babjuk M, Burger M, et al. (2017). 36. Izumi K, Taguri M, Miyamoto H, Hara
PMID:29036902
Prognostic performance and reproduc- Y, Kishida T, Chiba K, et al. (2014).
21. Jelaković B, Karanović S, Vuković-Lela I, ibility of the 1973 and 2004/2016 World Androgen deprivation therapy prevents
Miller F, Edwards KL, Nikolić J, et al. (2012). Health Organization grading classification bladder cancer recurrence. Oncotarget.
Aristolactam-DNA adducts are a biomarker systems in non-muscle-invasive bladder 5(24):12665–74. https://doi.org/10.18632/
of environmental exposure to aristolochic cancer: a European Association of Urology oncotarget.2851 PMID:25557268
acid. Kidney Int. 81(6):559–67. https://doi. Non-Muscle Invasive Bladder Cancer
Guidelines Panel systematic review. Eur 37. Izumi K, Li Y, Ishiguro H, Zheng Y, Yao
org/10.1038/ki.2011.371 PMID:22071594
Urol. 72(5):801–13. https://doi.org/10.1016/j. JL, Netto GJ, et al. (2014). Expression of
22. Poon SL, Huang MN, Choo Y, McPherson eururo.2017.04.015 PMID:28457661 UDP-glucuronosyltransferase 1A in blad-
JR, Yu W, Heng HL, et al. (2015). Mutation der cancer: association with prognosis
signatures implicate aristolochic acid in 30. Busetto GM, Ferro M, Del Giudice F, Antonini and regulation by estrogen. Mol Carcinog.
bladder cancer development. Genome G, Chung BI, Sperduti I, et al. (2017). The 53(4):314–24. https://doi.org/10.1002/
Med. 7(1):38. https://doi.org/10.1186/ prognostic role of circulating tumor cells mc.21978 PMID:23143693
s13073-015-0161-3 PMID:26015808 (CTC) in high-risk non-muscle-invasive
bladder cancer. Clin Genitourin Cancer. 38. Stenehjem DD, Tran D, Nkrumah MA,
23. Alexandrov LB, Ju YS, Haase K, Van Loo P, 15(4):e661–6. https://doi.org/10.1016/j. Gupta S (2018). PD1/PDL1 inhibitors for
Martincorena I, Nik-Zainal S, et al. (2016). clgc.2017.01.011 PMID:28188046 the treatment of advanced urothelial blad-
Mutational signatures associated with to- der cancer. Onco Targets Ther. 11:5973–
bacco smoking in human cancer. Science. 31. Bosschieter J, Lutz C, Segerink LI, Vis 89. https://doi.org/10.2147/OTT.S135157
354(6312):618–22. https://doi.org/10.1126/ AN, Zwarthoff ECA, A van Moorselaar PMID:30275703
science.aag0299 PMID:27811275 RJ, et al. (2018). The diagnostic accuracy
of methylation markers in urine for the 39. Bellmunt J, Powles T, Vogelzang NJ
24. Phillips DH (2018). Mutational spectra detection of bladder cancer: a system- (2017). A review on the evolution of
and mutational signatures: insights into atic review. Epigenomics. 10(5):673–87. PD-1/PD-L1 immunotherapy for bladder
cancer aetiology and mechanisms of https://doi.org/10.2217/epi-2017- 0156 cancer: the future is now. Cancer Treat
DNA damage and repair. DNA Repair PMID:29692199 Rev. 54:58–67. https://doi.org/10.1016/j.
(Amst). 71:6–11. https://doi.org/10.1016/j. ctrv.2017.01.007 PMID:28214651
dnarep.2018.08.003 PMID:30236628 32. Zlotta AR, Roumeguere T, Kuk C,
Alkhateeb S, Rorive S, Lemy A, et al.
25. Fantini D, Seiler R, Meeks JJ (2019). (2011). Select screening in a specific
Molecular footprints of muscle-invasive high-risk population of patients suggests
bladder cancer in smoking and nonsmok- a stage migration toward detection of non-
ing patients. Urol Oncol. 37(11):818–25. muscle-invasive bladder cancer. Eur Urol.
https://doi.org/10.1016/j.urolonc.2018.09. 59(6):1026–31. https://doi.org/10.1016/j.
017 PMID:30446446 eururo.2011.03.027 PMID:21458152
446
5.16
Kidney cancer
Multiple risk factors, but currently
limited preventive strategies
Ghislaine Scelo Rosamonde E. Banks (reviewer)
Alexander Parker Carlo La Vecchia (reviewer)
carcinoma is diagnosed at an In this chapter, descriptive sta-

SUMMARY advanced stage in 25–30% of tistics are reported for the broad
patients. classification of kidney cancer; in
●● In 2018, there were an esti- discussions of features such as risk
mated 403 000 new cases of factors and prognosis, the focus is
kidney cancer worldwide, ac- “Kidney cancer” is a broad term reon the most common subtype (i.e.
counting for 2.4% of all new ferring to a histologically heteroge- renal cell carcinoma), with some
cancer cases. The predominant neous group of tumours that arise statements pertaining to other, less
tumour type is renal cell carci- in the renal parenchyma and the common subtypes of kidney cancer.
noma. Age-standardized inci- renal pelvis. Renal cell carcinoma,
dence rates in men are highest which denotes cancer originating Epidemiology
in Belarus, Estonia, Czechia, from the epithelial cells of the re-
Latvia, and Lithuania and low- nal parenchyma, accounts for more Incidence patterns
est in India, Thailand, and some than 90% of all cases of kidney can- In 2018, there were an estimated
countries in Africa. cer [1]. 403 000 new cases of kidney can-
The most common histological cer worldwide, accounting for 2.4%
●● Eight genetic syndromes have classification of renal cell carcino- of all new cancer cases [2].
been reported to increase the ma is clear cell renal cell carcinoma
risk of renal cell carcinoma. The (~80%), which is the most com- Geography and ethnicity
most common is von Hippel– monly diagnosed type of kidney There are large geographical vari-
Lindau syndrome. cancer in adults. Other histological ations in incidence rates of kidney
●● Genetic variants in 13 regions of subtypes of kidney cancer include cancer. Age-standardized incidence
the genome have been identi- papillary (10–15%), chromophobe rates in men vary from more than 20
(~5%), and collecting duct (< 2%) re- per 100 000 in five European coun-
fied as risk factors for renal cell
nal cell carcinomas. Oncocytomas tries (Belarus, Estonia, Czechia,
carcinoma through large-scale
are a benign histological subtype. A Latvia, and Lithuania) to less than
genome-wide association stud-
substantial proportion of renal cell 2 per 100 000 in low-risk countries
CHAPTER 5.16
ies. The implicated pathways in-
SECTION 5
carcinomas can be cured by surgi- such as India, Thailand, and some
clude the VHL-HIF pathway.
cal resection as the main treatment. countries in Africa (Fig. 5.16.1) [2].
●● The increase in risk of renal Kidney cancer that occurs in In the USA, age-standardized
cell carcinoma is about 30% in children (Wilms tumour, also known incidence rates of kidney cancer are
smokers compared with never- as nephroblastoma) is a different higher in Blacks (15.6 per 100 000
smokers. Excess body weight, entity, which is beyond the scope in males and 8.6 per 100  000
hypertension, chronic kidney of this chapter. Tumours that arise in females) than in Whites (14.0
disease, diabetes, and occupa- in the renal pelvis and the ureter per 100 000 in males and 7.6 per
tional exposure to trichloroethy- (urothelial carcinomas) are far less 100 000 in females) [3]. Incidence
lene are each associated with an common than renal cell carcino- rates in Hispanic Whites are simi-
mas and have different epidemio- lar to those in non-Hispanic Whites.
increased risk of kidney cancer.
logical features, which are similar Rates in American Indians and
●● Opportunities for early detec- to those of bladder cancer (see Alaska Natives are intermediate
tion are limited, and renal cell Chapter 5.15). (10.9 per 100 000 in males and 6.6
Chapter 5.16 • Kidney cancer 447

per 100 000 in females), and rates same pattern as for incidence rates.
in Asians and Pacific Islanders are Age-standardized mortality rates are
lower (6.4 per 100 000 in males and highest in Belarus (11 per 100 000 in
FUNDAMENTALS
2.9 per 100 000 in females) [3]. males) and the Baltic countries [2]. ■■ “Kidney cancer” refers to a
In Europe, large regional vari- Globally, mortality rates of kidney histologically heterogeneous
ations have been described within cancer have been stable since the group of tumours. There are
some countries, notably in Germany 1990s [5]. In recent years, mortality large geographical variations
(higher incidence rates in the east- rates have decreased in most coun- in incidence rates, which are
ern regions of the country) and in tries, with the notable exception poorly explained, as well as
Italy (higher incidence rates in the of Brazil, Croatia, Greece, Ireland,
increasing incidence rates in
northern part of the country) [4]. Portugal, and Slovenia, where rates
certain populations.
have increased.
Age and sex In general, mortality rates ap- ■■ Incidence rates of kidney
Incidence rates of kidney cancer pear to be decreasing faster in cancer have been increasing
increase steadily with age, with women than in men. In the USA, worldwide since the 1970s.
a peak of incidence at about age the decline in mortality rates is
75 years [3,5]. Worldwide, more ■■ The vast majority of cases
more pronounced in Blacks than
than half of all cases are diagnosed are sporadic, and known
in Whites, and mortality rates in
in people younger than 65 years [2]. Blacks have remained slightly low- risk factors, such as tobacco
The incidence of kidney can- er than those in Whites since the smoking, obesity, and
cer in men is about twice that in 1970s [5,8]. Competing mortality hypertension, confer only
women, across geographical re- may play a role, but ethnic differ- modest risk increases. This
gions and categories of race and ences in the biology and aggres- makes it difficult to identify
ethnicity [6]. The stability of the siveness of kidney cancer could high-risk groups and to
male-to-female incidence ratio over also explain this variation [9]. develop screening procedures.
time, across countries, and by age
groups substantiates that biologi- ■■ Enhanced early detection
cal differences between men and Genetics and genomics efforts that do not result in
women – rather than differences overdiagnosis are needed,
Genetic syndromes because these would reduce
in lifestyle factors, such as tobacco
smoking – are likely to account for Approximately 3–5% of renal cell car- mortality from kidney cancer.
much of this disparity in incidence. cinomas occur in a familial context For localized, early-stage, low-
[10]. Only a subset of the familial kid- grade tumours, the prognosis
Temporal trends ney cancer cases can be explained is very good after surgical
Incidence rates of kidney cancer by known genetic syndromes [10]. intervention.
have been increasing worldwide The most common syndrome
known to be associated with re- ■■ Kidney cancer is often
since the 1970s [5]. In the USA, inci-
nal cell carcinoma is von Hippel– diagnosed at an advanced
dence rates in males have increased
Lindau (VHL) syndrome. It affects stage; for such tumours,
steadily, from 8.0 per 100 000 in
1975 to 13.4 per 100 000 in 2008– an estimated 1 per 36 000 live prognosis is poor.
2012. In most countries, the aver- births in the United Kingdom and
age annual percentage increase is suggested to account for ap-
is about 2–3%. Only Austria and proximately 1% of patients with
Poland have reported significant renal cell carcinoma [11]. VHL syn- to increase the risk of renal cell car-
decreases in rates, since the early drome is caused by mutations in cinoma: familial clear cell renal car-
2000s. Because the effects of both the VHL tumour suppressor gene, cinoma with chromosome 3 trans-
birth cohort and calendar period which is located on the short arm location, hereditary papillary renal
contribute to the increases in inci- of chromosome 3. VHL syndrome carcinoma syndrome, Birt–Hogg–
dence rates, the observed temporal also increases the risk of a range Dubé syndrome, hereditary leio-
trends are likely to be due to a com- of other tumours: haemangioblasto- myomatosis and renal carcinoma
bination of changes in lifestyle and mas of the brain, spine, and retina; syndrome, PTEN hamartoma syn-
in exposures to risk factors, as well pheochromocytomas of the adrenal drome, succinate dehydrogenase
as changes in tumour detection and gland; and neuroendocrine tumours complex-associated renal carci-
in diagnostic practices over time [7]. of the pancreas. The risk of renal noma, and BAP1 mutant syndrome
cell carcinoma depends on the type [10]. These syndromes have been
Mortality patterns of mutation in the VHL gene. described in less detail than VHL
International variations in kidney Currently, there are seven other syndrome with respect to their as-
cancer mortality rates follow the syndromes that have been reported sociation with risk of kidney cancer,
448
Fig. 5.16.1. Global distribution of estimated age-standardized (World) incidence rates (ASR) per 100 000 person-years for
kidney cancer in men, 2018.
and their prevalence in the popula- it is present in up to 7% of eastern matin remodelling or chromatin
tion is mostly unknown. European populations. Finally, a vari- modifier genes, including PBRM1,
ant in MITF has also been reported ARID1A, SETD2, BAP1, KDM5C,
Genetic polymorphisms to increase the risk of developing and KDM6A [16,17]. Several of
Genetic variants in 13 regions of cutaneous melanoma, renal cell car- these genes are located on the X
the genome have been identified as cinoma, or both by about 5-fold [14]. chromosome, and this may play
risk factors for renal cell carcinoma a role in the difference in risk be-
through large-scale genome-wide Tumour molecular tween men and women.
association studies (GWAS) [12]. phenotypes An unusual tumour genomic pat-
The implicated pathways include Kidney cancers – even the most tern was reported in cases of clear
the VHL-HIF pathway – with vari- common subtype (i.e. renal cell car- cell renal cell carcinoma in Romania,
ants discovered in two regions: the cinomas) – are histologically hetero- marking the mutational signature of
EPAS1 gene, which encodes hypox- geneous clinical entities. A concerted exposure to aristolochic acid [17] (see
ia-inducible factor 2 alpha (HIF-2α), effort is being made to explore the Chapter 2.8). Although the exposure
and the 11q13.3 region, which im- molecular underpinnings of these tu- has been confirmed [18], the causal
pairs binding of HIF-2α and results mours (i.e. molecular phenotyping) to link between the exposure and the
in an allelic imbalance of cyclin D1 more accurately define the nature of occurrence of the tumour remains to
– as well as mediation of cholesterol these cancers. Most of the research be investigated.
transfer, obesity-related pathways, has focused on clear cell renal cell Moving beyond genomics, there
and pathways related to chromatin carcinomas. Sporadic and familial are several reports of the presence
CHAPTER 5.16
SECTION 5
remodelling. Much remains to be clear cell renal cell carcinomas are and clinical significance of other mo-
discovered; the risk loci identified so biologically similar; they almost al- lecular alterations at the RNA and
far for renal cell carcinoma are es- ways show a loss of the short arm of protein levels in renal cell carcinoma
timated to account for only 10% of chromosome 3, which carries VHL (see Chapter 3.8). For example,
the familial risk, leaving about 90% and other tumour suppressor genes. higher expression levels of survivin,
of the heritability unexplained. It was recently reported that some topoisomerase II alpha, and IMP3
Two rare genetic variants may genomic structural events, typically have all been reported in clear cell
also be implicated in the risk of renal through chromothripsis, can occur renal cell carcinoma and, more im-
cell carcinoma, with no evidence of during childhood or adolescence – portantly, linked to poor prognosis
familial syndromes. The I157T mis- decades before the development of after curative surgery [19,20]. These
sense variant in the cell-cycle control the renal cell carcinoma tumour [15]. biomarkers and others offer oppor-
gene CHEK2 increases the risk by For clear cell renal cell carcino- tunities to better manage post-oper-
about 50% [13]. Although the I157T mas, in addition to VHL, somatic ative follow-up for patients with clear
variant is very rare in most countries, mutations are recurrent in chro- cell renal cell carcinoma.

Non-clear cell renal cell carcino- horts [24]. In several studies the 5 years after blood pressure mea-
mas have different genomic profiles association was shown to be linear, surement in an attempt to minimize
[21]. For example, papillary renal with an increase in risk of about reverse causation [28]. In a study
cell carcinomas are typically char- 25% for each increase of 5 kg/m2 in with repeated measurements of
acterized by alterations of the MET body mass index (BMI). No data are blood pressure over time, the risk
pathway, and chromophobe renal available on the benefit of weight of renal cell cancer decreased with
cell carcinomas are characterized loss and/or long-term maintenance decreasing blood pressure [28].
by metabolic pathway alterations of a lower BMI in association with
with mitochondrial dysfunctions. risk of kidney cancer. High BMI is Alcohol consumption
estimated to be responsible for 26% Moderate consumption of alcohol
Etiology of incident cases of renal cell carci- reduces the risk of developing renal
noma worldwide [25]. cell carcinoma, and this protective ef-
Tobacco smoking Height has also been consis- fect may be stronger in women than
The effect size of tobacco smoking tently associated with risk of kidney in men. The identification of alcohol
on the risk of renal cell carcinoma cancer, independently of weight, consumption as a factor associated
is modest; the increase in risk is with an increase in risk of about with lower risk of renal cell carcinoma
36% in current smokers, 16% in for- 30% for each increase of 10 cm resulted from early observations in
mer smokers, and 31% in all smok- in height [26]. The mechanisms case–control studies and progressed
ers, compared with never-smokers involved could include levels of to much more robust and consistent
[22]. Epidemiological evidence for growth hormones, genetic back- evidence from large prospective co-
a causal role of tobacco smoking ground, and childhood exposures, horts, pooling projects, and meta-
includes a dose–response relation- rather than a direct link with renal analyses [29]. Investigators have
ship between risk and the quantity of cell carcinoma. begun to explore the possibility that
tobacco smoked per day, as well as the association between alcohol in-
decreased risks with a larger number Hypertension take and risk of renal cell carcinoma
of years after smoking cessation. In In the USA, a history of hyperten- may be modulated by variation in un-
high-income countries, an estimated sion has been estimated to double derlying genetics such as the genes
6% of deaths from kidney cancer are the risk of kidney cancer in Whites, coding for enzymes that metabolize
due to tobacco smoking [23]. and to triple the risk in Blacks [27]. alcohol [30].
Prospective cohort studies have
Anthropometric measures consistently reported dose–re-
Chronic kidney disease
The association between excess sponse associations between blood Chronic kidney disease increases
body weight and risk of renal cell pressure at baseline and risk of the risk of kidney cancer by 2–3-
carcinoma has been reported ex- kidney cancer, even when the risk fold [31]. Evidence suggests that
tensively in large prospective co- analysis is restricted to more than in the USA the increase in risk is
more pronounced in Blacks than in
Whites; this may contribute to the
Fig. 5.16.2. A patient’s blood pressure is monitored. Prospective cohort studies have higher observed incidence rates in
reported associations between blood pressure at baseline and risk of kidney cancer. Blacks, given that chronic kidney
disease is also more prevalent in
Blacks than in Whites [31,32].
Diabetes
The association between diabetes
and risk of kidney cancer has been
assessed in several prospective
cohort studies, but independence
from comorbidities of diabetes,
such as obesity, hypertension, and
chronic kidney disease, is still un-
clear [33]. A history of diabetes was
found to be associated with a 40%
excess risk of kidney cancer [33].
Trichloroethylene
The IARC Monographs classified
occupational exposure to trichloro-
ethylene as carcinogenic to humans
450
(Group 1), on the basis of a body of a range of medical conditions and the rise in incidence predates
sufficient evidence that this chemi- symptoms. Because renal cell car- widespread use of sensitive ab-
cal causes kidney cancer [34]. The cinoma usually remains clinically dominal imaging, and the incidence
most recent meta-analysis estimat- occult for most of its course, it is of late-stage tumours has also in-
ed that occupational exposure to tri- often diagnosed at an advanced creased [36].
chloroethylene confers a 30–40% stage, and 25–30% of patients Opportunities for primary pre-
excess risk of kidney cancer (see have metastases at diagnosis [37]. vention are limited, because the
Chapter 2.10) [35]. Because most kidney tumours factors that are responsible for the
develop outside the context of diag- geographical variations and time
Biology and early nosed genetic cancer syndromes, trends have not been identified. For
detection there is currently no recommended example, kidney cancer incidence
screening practice for primary renal rates have not benefited from the
Kidney cancer is characterized by general reduction in tobacco use.
cell carcinoma in people who are
the absence of early warning signs As discussed earlier, the factors
not known to carry genetic mutations
and by non-specific symptoms. that are known to be associated with
associated with increased risk of the
Patients who are diagnosed with lo- renal cell carcinoma confer modest
disease. Given that renal masses
calized renal cell carcinoma (stages risk increases (relative risks of about
can be detected with ultrasonogra-
I and II) are commonly cured after 1.2–2.5), resulting in population attri-
phy techniques, which are non-inva-
nephron-sparing nephrectomy as butable risks of less than 50% [40].
sive and harmless, the question of
the sole treatment, with limited This poses challenges for identify-
whether general screening for early
long-term side-effects. For tumours ing high-risk populations that could
detection of kidney cancer in the
that invade local tissues (stage III) benefit from enhanced screening
population is warranted has arisen
or have distant metastasis (stage
from patient associations as well as protocols. Nevertheless, the discov-
IV), prognosis is poor, with 5-year
clinicians. However, there has been ery of genetic polymorphisms as-
survival rates of about 50% and
no systematic evaluation of the con- sociated with development of renal
10%, respectively [36].
ditions for implementing a screening cell carcinoma and the identification
The majority of curable early-
programme (see Chapter 6.6). of refined molecular subtypes of the
stage tumours are detected inci-
In the absence of clear high-risk disease provide a clear opportunity
dentally through the wide use of
groups at the population level and to explore gene–environment inter-
ultrasonography examinations for
of non-invasive biomarkers for renal actions coupled with molecular sub-
cell carcinoma that could be mea- typing, which could reveal more indi-
Fig. 5.16.3. Although trichloroethylene sured in blood or urine, secondary vidualized risk estimates that would
has largely been replaced by tetrachloro- support the screening of certain
prevention for kidney cancer is still
ethylene as the main solvent used in dry populations. This approach is par-
cleaning, trichloroethylene is still used as a long way off. Research efforts are
a spot remover. A meta-analysis estimat- under way to identify such biomark- ticularly intriguing given the future
ed that occupational exposure to trichlo- ers. Plasma levels of KIM-1 were re- possibility of developing lower-cost
roethylene confers a 30–40% excess risk cently reported to predict the risk of and scalable screening tests based
of kidney cancer. being diagnosed with renal cell caron circulating biomarkers. However,
cinoma in the subsequent 5 years care must be taken to avoid the risk
[38]. However, the predictive ability of overdiagnosis that has occurred
would need to be improved for use with other cancer types (e.g. pros-
in a screening setting. tate cancer).
A systematic evaluation is war-
ranted of the conditions for imple-
Opportunities for
CHAPTER 5.16
SECTION 5
menting a screening programme.
prevention Kidney cancers are asymptomatic
Projections from Cancer Research and are usually detected inci-
UK indicate that over the next 20 dentally through routine imaging.
years kidney cancer will be one Therefore, most patients are treat-
of the cancer types with the most ed for a suspicious renal mass and
rapidly rising incidence [39]. These are only diagnosed with a cancer
estimates are based on increas- or a benign tumour after invasive
ing trends over the past decade surgery. The discovery of circulat-
and may be inflated as a result of ing biomarkers that could stratify
overdiagnosis during this period. renal masses into likely benign or
However, the increasing trends can- likely malignant would be extremely
not be explained solely by increased valuable to overcome the issues of
detection of asymptomatic tumours: overdiagnosis and overtreatment.

References
1. Hsieh JJ, Purdue MP, Signoretti S, 11. Maher ER, Neumann HP, Richard S 19. Parker AS, Eckel-Passow JE, Serie D,
Swanton C, Albiges L, Schmidinger M, (2011). von Hippel-Lindau disease: a clini- Hilton T, Parasramka M, Joseph RW, et
et al. (2017). Renal cell carcinoma. Nat cal and scientific review. Eur J Hum Genet. al. (2014). Higher expression of topoisom-
Rev Dis Primers. 3(1):17009. https://doi. 19(6):617–23. https://doi.org/10.1038/ejhg. erase II alpha is an independent marker
org/10.1038/nrdp.2017.9 PMID:28276433 2010.175 PMID:21386872 of increased risk of cancer-specific death
in patients with clear cell renal cell car-
2. Bray F, Ferlay J, Soerjomataram I, Siegel 12. Scelo G, Purdue MP, Brown KM, cinoma. Eur Urol. 66(5):929–35. https://
RL, Torre LA, Jemal A (2018). Global Johansson M, Wang Z, Eckel-Passow JE, doi.org /10.1016/ j.eur uro. 2013.12.017
cancer statistics 2018: GLOBOCAN es- et al. (2017). Genome-wide association PMID:24388441
timates of incidence and mortality world- study identifies multiple risk loci for renal
wide for 36 cancers in 185 countries. CA cell carcinoma. Nat Commun. 8(1):15724. 20. Pu Z, Wang Q, Xie H, Wang G, Hao H (2017).
Cancer J Clin. 68(6):394–424. https://doi. https://doi.org/10.1038/ncomms15724 Clinicalpathological and prognostic signifi-
org/10.3322/caac.21492 PMID:30207593 PMID:28598434 cance of survivin expression in renal cell
carcinoma: a meta-analysis. Oncotarget.
3. Bray F, Colombet M, Mery L, Piñeros M, 13. Brennan P, McKay J, Moore L, Zaridze 8(12):19825–33. https://doi.org/10.18632/
Znaor A, Zanetti R, et al., editors (2017). D, Mukeria A, Szeszenia-Dabrowska oncotarget.15082 PMID:28178644
Cancer incidence in five continents, Vol. N, et al. (2007). Uncommon CHEK2
XI (electronic version). Lyon, France: mis-sense variant and reduced risk of 21. Albiges L, Flippot R, Rioux-Leclercq N,
International Agency for Research on tobacco-related cancers: case control Choueiri TK (2018). Non-clear cell renal
Cancer. Available from: http://ci5.iarc.fr. study. Hum Mol Genet. 16(15):1794–801. cell carcinomas: from shadow to light. J
ht t p s: //d o i.o r g /10.10 9 3 / h m g /d dm127 Clin Oncol. [Epub ahead of print] https://
4. Li P, Znaor A, Holcatova I, Fabianova PMID:17517688 d o i . o r g / 10 .12 0 0 /J C O. 2 0 18 .7 9 . 2 5 31
E, Mates D, Wozniak MB, et al. (2015). PMID:30372389
Regional geographic variations in kid- 14. Bertolotto C, Lesueur F, Giuliano S,
ney cancer incidence rates in European Strub T, de Lichy M, Bille K, et al.; French 22. Cumberbatch MG, Rota M, Catto JW, La
countries. Eur Urol. 67(6):1134–41. https:// Familial Melanoma Study Group (2011). Vecchia C (2016). The role of tobacco
doi.org /10.1016/ j.eur uro. 2014.11.0 01 A SUMOylation-defective MITF germ- smoke in bladder and kidney carcino-
PMID:25465966 line mutation predisposes to melanoma genesis: a comparison of exposures and
and renal carcinoma. Nature. 480(7375): meta-analysis of incidence and mortali-
5. Znaor A, Lortet-Tieulent J, Laversanne M, 94–8. https://doi.org/10.1038/nature10539 ty risks. Eur Urol. 70(3):458–66. https://
Jemal A, Bray F (2015). International vari- PMID:22012259 doi.org/10.1016/j.eururo. 2015.0 6.0 42
ations and trends in renal cell carcinoma PMID:26149669
incidence and mortality. Eur Urol. 67(3): 15. Mitchell TJ, Turajlic S, Rowan A, Nicol D,
519–30. https://doi.org/10.1016/j.eururo. Farmery JHR, O’Brien T, et al.; TRACERx 23. Dy GW, Gore JL, Forouzanfar MH,
2014.10.002 PMID:25449206 Renal Consortium (2018). Timing the Naghavi M, Fitzmaurice C (2017). Global
landmark events in the evolution of burden of urologic cancers, 1990–
6. Scelo G, Li P, Chanudet E, Muller DC clear cell renal cell cancer: TRACERx 2013. Eur Urol. 71(3):437–46. https://
(2018). Variability of sex disparities in Renal. Cell. 173(3):611–623.e17. https:// doi.org/10.1016/j.eur uro. 2016.10.0 0 8
cancer incidence over 30 years: the strik- d o i . o r g / 1 0 .1 0 1 6 / j . c e l l . 2 0 18 . 0 2 . 0 2 0 PMID:28029399
ing case of kidney cancer. Eur Urol Focus. PMID:29656891
4(4):586–90. https://doi.org/10.1016/j.euf. 24. Wang F, Xu Y (2014). Body mass index and
2017.01.006 PMID:28753845 16. Cancer Genome Atlas Research Network risk of renal cell cancer: a dose-response
(2013). Comprehensive molecular char- meta-analysis of published cohort studies.
7. Znaor A, Laversanne M, Bray F (2017). acterization of clear cell renal cell Int J Cancer. 135(7):1673–86. https://doi.
Less overdiagnosis of kidney cancer? an carcinoma. Nature. 499(7456):43–9. org/10.1002/ijc.28813 PMID:24615287
age-period-cohort analysis of incidence ht t p s: //d o i.o r g /10.10 3 8 /n at u r e12 2 2 2
trends in 16 populations worldwide. Int PMID:23792563 25. Arnold M, Pandeya N, Byrnes G,
J Cancer. 141(5):925–32. https://doi.org/ Renehan PAG, Stevens GA, Ezzati PM, et
10.1002/ijc.30799 PMID:28543047 17. Scelo G, Riazalhosseini Y, Greger L, al. (2015). Global burden of cancer attri-
Letourneau L, Gonzàlez-Porta M, Wozniak butable to high body-mass index in 2012:
8. Lipworth L, Tarone RE, McLaughlin MB, et al. (2014). Variation in genomic land- a population-based study. Lancet Oncol.
JK (2011). Renal cell cancer among scape of clear cell renal cell carcinoma 16(1):36–46. https://doi.org/10.1016/S1470-
African Americans: an epidemiologic across Europe. Nat Commun. 5(1):5135. 2045(14)71123-4 PMID:25467404
review. BMC Cancer. 11(1):133. https:// ht tps: //doi.org /10.10 3 8 /nc omms 613 5
d o i . o r g / 1 0 .11 8 6 / 1 4 71 - 2 4 0 7-11 -13 3 PMID:25351205 26. Wirén S, Häggström C, Ulmer H, Manjer
PMID:21486465 J, Bjørge T, Nagel G, et al. (2014).
18. Turesky RJ, Yun BH, Brennan P, Mates Pooled cohort study on height and risk
9. Lipworth L, McLaughlin JK, Tarone RE, D, Jinga V, Harnden P, et al. (2016). of cancer and cancer death. Cancer
Blot WJ (2011). Renal cancer paradox: Aristolochic acid exposure in Romania Causes Control. 25(2):151–9. https://
higher incidence but not higher mortality and implications for renal cell car- d o i . o r g /10 .10 07/s10 5 5 2- 013 - 0 317-7
among African-Americans. Eur J Cancer cinoma. Br J Cancer. 114(1):76–80. PMID:24173535
Prev. 20(4):331–3. https://doi.org/10.1097/ ht t ps: //doi.o rg /10.10 3 8 / bjc . 2015. 4 02
CEJ.0b013e328345f9b3 PMID:21633203 PMID:26657656 27. Colt JS, Schwartz K, Graubard BI, Davis F,
Ruterbusch J, DiGaetano R, et al. (2011).
10. Haas NB, Nathanson KL (2014). Hypertension and risk of renal cell carci-
Hereditary kidney cancer syndromes. noma among white and black Americans.
Adv Chronic Kidney Dis. 21(1):81–90. Epidemiology. 22(6):797–804. https://doi.
https://doi.org/10.1053/j.ackd.2013.10.001 org/10.10 97/EDE.0b013e318230 0720
PMID:24359990 PMID:21881515
452
28. Chow WH, Gridley G, Fraumeni JF Jr, 33. Harding JL, Shaw JE, Peeters A, 38. Scelo G, Muller DC, Riboli E, Johansson
Järvholm B (2000). Obesity, hyperten- Cartensen B, Magliano DJ (2015). Cancer M, Cross AJ, Vineis P, et al. (2018). KIM-
sion, and the risk of kidney cancer in men. risk among people with type 1 and type 2 1 as a blood-based marker for early de-
N Engl J Med. 343(18):1305–11. https:// diabetes: disentangling true associations, tection of kidney cancer: a prospective
doi.org/10.1056/NEJM200011023431804 detection bias, and reverse causation. nested case-control study. Clin Cancer
PMID:11058675 Diabetes Care. 38(2):264–70. https://doi. Res. 24(22):5594–601. https://doi.org/
org/10.2337/dc14-1996 PMID:25488912 1 0 .11 5 8 / 1 0 7 8 - 0 4 3 2 . C C R - 1 8 - 1 4 9 6
29. Bellocco R, Pasquali E, Rota M, Bagnardi V,
PMID:30037816
Tramacere I, Scotti L, et al. (2012). Alcohol 34. IARC (2013). Trichloroethylene, tetrachlo-
drinking and risk of renal cell carcinoma: roethylene, and some other chlorinated 39. Smittenaar CR, Petersen KA, Stewart
results of a meta-analysis. Ann Oncol. agents. IARC Monogr Eval Carcinog Risks K, Moitt N (2016). Cancer incidence
23(9):2235–44. https://doi.org/10.1093/ Hum. 106:1–514. Available from: http:// and mortality projections in the UK un-
annonc/mds022 PMID:22398178 publications.iarc.fr/130 PMID:26214861 til 2035. Br J Cancer. 115(9):1147–55.
ht tps: //doi.org /10.10 3 8 / bjc. 2016. 3 0 4
30. Antwi SO, Eckel-Passow JE, Diehl
35. Karami S, Lan Q, Rothman N, Stewart PMID:27727232
ND, Serie DJ, Custer KM, Wu KJ, et al.
PA, Lee K-M, Vermeulen R, et al. (2012).
(2018). Alcohol consumption, variabil- 40. Scelo G, Larose TL (2018). Epidemiology
Occupational trichloroethylene exposure
ity in alcohol dehydrogenase genes and and risk factors for kidney cancer. J Clin
and kidney cancer risk: a meta-analysis.
risk of renal cell carcinoma. Int J Cancer. Oncol. [Epub ahead of print] https://
Occup Environ Med. 69(12):858–67. https://
142(4):747–56. https://doi.org/10.1002/ d o i . o r g / 10 .12 0 0 /J C O . 2 0 18 .7 9 .19 0 5
doi.org /10.113 6 /oemed -2012-10 0 932
ijc.31103 PMID:29023769 PMID:30372394
PMID:23000822
31. Hofmann JN, Corley DA, Zhao WK, Colt JS,
Shuch B, Chow W-H, et al. (2015). Chronic 36. Chow WH, Dong LM, Devesa SS (2010).
kidney disease and risk of renal cell carci- Epidemiology and risk factors for kid-
noma: differences by race. Epidemiology. ney cancer. Nat Rev Urol. 7(5):245–57.
26(1):59–67. https://doi.org/10.1097/EDE. https://doi.org/10.1038/nrurol.2010.46
0000000000000205 PMID:25393631 PMID:20448658
32. Lipworth L, Mumma MT, Cavanaugh KL, 37. Gupta K, Miller JD, Li JZ, Russell MW,
Edwards TL, Ikizler TA, Tarone RE, et al. Charbonneau C (2008). Epidemiologic
(2012). Incidence and predictors of end and socioeconomic burden of metastatic
stage renal disease among low-income renal cell carcinoma (mRCC): a literature
blacks and whites. PLoS One. 7(10): review. Cancer Treat Rev. 34(3):193–205.
e48407. https://doi.org/10.1371/journal. https://doi.org/10.1016/j.ctrv.2007.12.001
pone.0048407 PMID:23110237 PMID:18313224
CHAPTER 5.16
SECTION 5

5.17
Brain cancer
Increasing attention
on the immune response
Dominique S. Michaud John D. Mathews (reviewer)
Hiroko Ohgaki (reviewer)
Joachim Schüz (reviewer)
varicella zoster virus are each also highlighted the need for ad-
SUMMARY inversely associated with risk of ditional research on the causes of
glioma, and several markers of non-malignant brain tumours and
●● The revised WHO classifica- immune status are strongly as- childhood brain tumours [2].
tion of malignant tumours of the sociated with risk. About 68% of all brain and cen-
central nervous system includes
tral nervous system tumours are
molecular data, along with his-
non-malignant; about half of these
tology, in defining tumour types. In 2018, cancer of the brain and tumours are meningiomas, followed
●● The topic of mobile phones and central nervous system was the by pituitary tumours and nerve
brain tumours remains con- 17th most common cancer type, sheath tumours [3]. Meningiomas,
troversial despite decades of with an estimated 297 000 new even when they are non-malignant,
research and results from nu- cases worldwide. The study of the can have a devastating impact on
merous observational studies. etiology of brain tumours is particu- health by altering normal brain func-
Some studies have reported a larly challenging because of the rel- tion. Epidemiological studies that
higher relative risk for heavy atively low incidence rates of brain examine genetic and environmental
use of mobile phones, but in- and central nervous system can- determinants of brain tumours no
cidence rates of malignant tu- cers and the high heterogeneity of longer combine meningiomas with
mours have not increased over these tumours. As a result, most re-
other types of brain tumours, given
the past three decades. search in this field has been based
that they are etiologically (as well as
on case–control studies, which
●● Various genetic susceptibility clinically) distinct tumours. Among
have methodological limitations, or
loci have been identified for glio- the malignant tumours, heteroge-
cohort studies, which are often lim-
mas, and two distinct suscepti- neity is also substantial; almost half
ited by small numbers of cases.
bility loci have been associated of these are glioblastomas, followed
Because it is difficult to study
with meningiomas. Some sus- by other gliomas [3]. Most epide-
brain tumours within individual in-
ceptibility loci appear to be spe- stitutions, international brain cancer miological studies examine gliomas
cific to tumour grade, and risk consortia have been established to together, given that they originate
variants may also vary by sex. increase sample sizes, improve the from the same cell types (i.e. glial
classification of tumours, pool data cells), although often glioblasto-
●● Inherited variants or mutations mas – the most aggressive brain
for genetic and molecular analyses,
and acquired somatic mutations tumours – are examined separately.
and increase collaboration across
in or near telomerase genes are This chapter focuses on re-
different disciplines. These collabo-
associated with increased risk of search advances in the field of
rative efforts have been highly suc-
glioma. This suggests that long- epidemiology in the past 5 years.
cessful, resulting in advances in the
er telomere length may be a key
molecular classification of malig- It highlights findings from pooling
contributor to gliomagenesis.
nant brain tumours and the identifi- studies (consortium efforts) or co-
●● There is increasing evidence cation of new genetic susceptibility hort studies that have confirmed
that the immune response plays regions, and a consensus is being earlier findings, as well as new and
an important role in the etiology approached on the role of allergies promising results from studies ex-
of malignant glioma. Allergies [1] and other risk factors [2] in brain amining the role of the immune re-
and a history of infection with tumours. The collaborations have sponse in etiology.
454
There is increasing evidence and epidemiological studies may pro-

that the immune response plays vide new insights into etiological fac- FUNDAMENTALS
an important role in glioma devel- tors, because the differences in pat-
opment, and research that is under terns of acquired mutations across ■■ Brain and other primary central
way in this area should provide new different groups suggest that these nervous system tumours
opportunities for the identification tumours have distinct pathogenesis. comprise a group of very
of markers for early detection or heterogeneous tumours, both
prognosis prediction. In addition, Etiology malignant and non-malignant,
obtaining a better understanding of which exhibit a wide range of
underlying immune-related mecha- True etiological factors for brain
clinical signs and symptoms
nisms may provide new opportuni- tumours have been difficult to iden-
with varying prognosis.
ties for the development of immuno- tify, because findings for many
therapies to prolong survival. suspected risk factors have been ■■ Incidence rates of subtypes of
inconsistent or null. Many potential brain tumours vary substantially
risk factors have been studied, but by age; children develop differ-
Revised WHO most remain classified as “prob- ent types of brain tumours than
classification ably not risk factors”. These include adults. For example, embryonal
In 2016, the WHO classification of head injuries, occupational expo- tumours and pilocytic astrocy-
malignant tumours of the central sures, residential power-frequen- tomas are rarely observed in
nervous system was revised to in- cy electromagnetic fields, dental adults, whereas adult subtypes,
clude molecular data, along with X-rays, tobacco smoke, and alco- such as glioblastomas, are rare
histology, in defining tumour types hol consumption [2]. In two large in children.
[4]. The updated classification, prospective cohort studies, no as-
which includes molecular markers sociations were observed between ■■ Malignant brain cancers, pri-
(Fig. 5.17.1), demonstrates the het- meat intake, or carcinogens derived marily gliomas, are more com-
erogeneity of different malignant from meat, and risk of glioma [7,8]. mon in men than in women,
brain tumours and the difficulty of Although obesity has not been con- whereas non-malignant
classifying these tumours using his- sistently associated with risk of tumours are more common
tology alone. glioma, there is a consensus that in women. Incidence rates of
The importance of the revised obesity is associated with risk of malignant brain and other cen-
classification has been demonstrat- meningioma [9]. tral nervous system tumours
ed in large tumour data sets with are higher in Whites than in
clinical and demographic charac- Mobile phones Blacks, but incidence rates of
teristics. Tumours with certain mo- In 2011, an IARC Monographs Work non-malignant tumours are
lecular markers have been shown ing Group tasked with reviewing the higher in Blacks than in Whites.
to have distinct clinical behaviour. evidence on radiofrequency electro- Geographical variations in
In a data set that included both magnetic fields, including exposure incidence rates exist but could
high-grade and low-grade gliomas, from mobile phones, concluded that be attributable to differences in
tumours were classified into five there was limited evidence that these diagnostic, classification, and
groups on the basis of mutations exposures cause cancer in humans reporting practices.
in the TERT promoter, mutations and experimental animals, and clas-
in IDH, and co-deletion of chro- sified radiofrequency electromagnet- ■■ Genetic factors, including
mosome arms 1p and 19q (1p/19q ic fields as possibly carcinogenic to certain familial syndromes
co-deletion); the molecular groups humans (Group 2B) [10]. such as neurofibromatosis
CHAPTER 5.17
SECTION 5
were strongly associated with age at In the past 5 years, various com- and inherited genetic
diagnosis, survival, grade, and spe- mentaries, original studies, and susceptibilities, increase the
cific germline variants [5]. Similarly, meta-analyses have been published risk of brain tumours.
in a data set of lower-grade gliomas on this subject, which continues to ■■ Exposure to ionizing radiation,
from the Cancer Genome Atlas, receive substantial news coverage.
whether from atomic bombs
three groups of tumours, classi- However, causality remains question-
or therapeutic irradiation, has
fied on the basis of the presence able (see Chapter 2.5). Observational
been firmly established as a
or absence of mutations in IDH and study designs have limitations, some
cause of brain tumours.
1p/19q co-deletions, were strongly of which are particularly problematic
linked to clinical characteristics, in- when studying mobile phones and ■■ No excess risk of brain and
cluding histology, age at diagnosis, brain cancer. Limitations that contrib- other central nervous system
and survival (Fig. 5.17.2) [6]. ute to the complexity of determining tumours has been attributed to
Future widespread use of the re- causality include: difficulties in ac- use of tobacco products.
vised WHO classification in clinical curately measuring mobile phone
Chapter 5.17 • Brain cancer 455

Fig. 5.17.1. A simplified algorithm for classification of the diffuse gliomas on the basis of histological and genetic features. A caveat
to this diagram is that the diagnostic “flow” does not necessarily always proceed from histology first to molecular genetic features
next, because molecular signatures can sometimes outweigh histological characteristics in achieving an integrated diagnosis. A
similar algorithm can be followed for anaplastic-level diffuse gliomas. NOS, not otherwise specified. * Characteristic but not required
for diagnosis.
Histology Astrocytoma Oligoastrocytoma Oligodendroglioma Glioblastoma
IDH status
IDH mutant IDH wild-type IDH mutant IDH wild-type
1p/19q and Glioblastoma, IDH mutant

ATRX loss*
other genetic 1p/19q co-deletion
TP53 mutation*
parameters Glioblastoma, IDH wild-type
Diffuse astrocytoma, IDH mutant
Oligodendroglioma, IDH mutant and 1p/19q co-deleted Genetic testing not done
or inconclusive
After exclusion of other entities:

Diffuse astrocytoma, IDH wild-type
Oligodendroglioma, NOS
Diffuse astrocytoma, NOS

Oligodendroglioma, NOS
Oligoastrocytoma, NOS
Glioblastoma, NOS
use, with reference to both dose in the USA, age-standardized, of increase in incidence rates in
and duration; the potential for recall delay-adjusted incidence rates of countries with high use of mobile
bias, especially with respect to use malignant brain and other central phones, call causality into question.
of the phone on a particular side of nervous system cancers continued Nevertheless, this topic will continue
the head (i.e. laterality); the relative to decline in males (annual percent- to be highly controversial, because
recency of widespread use of mobile age change, −0.2%) and in females experts continue to disagree on the
phones, which is problematic for ex- (annual percentage change, −0.7%) interpretation of data that arise from
amining the possible impact of long in the most recent 5-year period different study designs. Results from
latency periods; and the heterogene- (2010–2014), even with adjustment prospective cohort studies that col-
ity of brain cancer subtypes. The re- for delays in reporting to cancer lect self-reported data on the use of
sults of experimental studies, wheth- registries [13]. Similarly, stable or mobile phones may shed light on the
er in vitro, in vivo, or animal studies, decreasing incidence rates of ma- associations, but a long waiting peri-
are similarly inconsistent [11]. lignant brain tumours (glioma and od is expected before these studies
Time trends in the incidence glioblastoma) were reported across provide results [14].
rates of brain cancer in countries all age groups in 2000–2014 in a
where mobile phones have been summary of the most recent and Genetic susceptibility
in widespread use for 25 years or comprehensive data on rates of Genome-wide association studies
more, including the USA, the Nordic malignant and non-malignant brain (GWAS) have identified several ge-
countries, the United Kingdom, and tumours for 99.9% of the population netic variants associated with risk
Australia [12], do not support the of the USA (Fig. 5.17.4) [3]. of different brain tumour subtypes.
strong positive relative risks report- It has been more than 25 years Single-nucleotide polymorphisms
ed in some case–control studies, since mobile phones were intro- (SNPs) in seven genes (TERT,
even after accounting for a 10-year duced, and they have been used TP53, CCDC26, EGFR, CDKN2B/
latency period. In the most recent by billions of people. These facts, CDKN2A, RTEL1, and PHLDB1)
(2018) report on cancer incidence combined with the consistent lack have been consistently linked to risk
456
Fig. 5.17.2. (A) Kaplan–Meier estimates of overall survival in patients with lower- of glioma in several large GWAS.
grade gliomas that are classified according to traditional histological type and grade. Recently, a large meta-analysis iden-
Glioblastoma samples (from previously published Cancer Genome Atlas data) tified 13 new susceptibility loci for
are also included for comparison. (B) Overall survival in patients with lower-grade
gliomas that are classified according to IDH mutation and 1p/19q co-deletion status.
glioma [15]. Although some suscep-
Glioblastoma samples classified according to IDH mutation status are also included. tibility loci are common to all glioma
GBM, glioblastoma; LGG, lower-grade glioma. subtypes (e.g. TP53), others appear
to be specific to glioblastoma (e.g.
EGFR) or non-glioblastoma glioma
A Gliomas Classified According to Histological Class and Grade (e.g. PHLDB1) [15] or to molecular
subgroups of glioma [5,16].
100 Astrocytoma II (N=30)
Two genetic variants located
Astrocytoma III (N=65)
Oligoastrocytoma II (N=41) near the telomerase genes TERC
Oligoastrocytoma III (N=33) and TERT are associated both with
80 Oligodendroglioma II (N=65) increased risk of high-grade glioma
Oligodendroglioma III (N=44) and with longer telomere length
Glioblastoma IV (N=397) [17]. The presence of frequent mu-
tations in telomerase genes in glio-
Overall Survival (%)
P<0.001 by log-rank test mas points to an important role for

60
telomere length in glioma develop-
ment [18].
New research suggests that
40 glioma risk variants are sex-spe-
cific [19]. These findings may pro-
vide insights into mechanisms that
may explain why incidence rates
20 of glioma are higher in men than
in women. For meningiomas, two
susceptibility loci (10p12.31 and
11p15.5) have been identified [20];
0 these are distinct from those identi-
0 5 10 15 fied for gliomas.
Years since Diagnosis It has been estimated that 25%
of the variation in the risk of develop-
ing all forms of glioma is associated
B Gliomas Classified According to Molecular Subtype with common genetic susceptibility
100 variants [21]. This estimate is de-
LGG with IDH mutation and
rived from a genome-wide complex
1p/19q co-deletion (N=84)
trait analysis that enables risk to be
LGG with IDH mutation and
no 1p/19q co-deletion (N=139) evaluated on the basis of the contri-
80 LGG with wild-type IDH (N=55) bution of all SNPs simultaneously
GBM with IDH mutation (N=24)
in GWAS (in contrast to an analysis
of the effects of single SNPs). The
GBM with wild-type IDH (N=373)
Overall Survival (%)
evidence to date strongly supports

60 a polygenic basis of genetic suscep-
CHAPTER 5.17
SECTION 5
tibility to glioma, and improved mo-
P<0.001 by log-rank test lecular classification of glioma sub-
types may result in the identification
40 of additional susceptibility loci.
Allergies, infections, and the

immune response
20 There is little or no evidence that
common cancer risk factors, includ-
ing tobacco smoke, obesity, and
diet, play a role in the etiology of
0 glioma. This suggests that the envi-
0 5 10 15
ronmental factors that influence car-
Years since Diagnosis cinogenesis in glial cells are unique.

Fig. 5.17.3. A man using a mobile phone. Observational studies have not yet established lergy status, have observed inverse
definitively whether use of mobile phones causes brain cancer. associations with risk of glioma [23–
25]. These findings provide support
for a causal relationship, because
cohort studies are not prone to re-
call bias or reverse causation.
Improved immunosurveillance
and protection against environ-
mental toxins in people with al-
lergies have been proposed as
mechanisms for how allergies may
confer protection against glioma [1].
However, the exact mechanisms for
these associations are not known,
and further research is required.
Varicella zoster virus

Unlike studies of polyomaviruses
(e.g. simian virus 40), which were
suspected to increase the risk of
brain tumours but were not subse-
quently confirmed as risk factors,
studies of varicella zoster virus (a
Allergies confirming inverse associations for herpesvirus that causes chicken-
It is becoming increasingly apparent allergies [1]. This large study report- pox and shingles) have reported
that the immune response plays a ed statistically significant reductions inverse associations between a his-
central role in the etiology of glioma. in risk of glioma of 30% for any res- tory of infection with the virus and
Numerous studies have reported piratory allergy, 23% for history of risk of glioma. Fewer studies have
inverse associations between aller- asthma, and 30% for history of ec- examined this association than
gies, including asthma and eczema, zema [1]. These associations were have investigated those for aller-
and risk of glioma [22]. Results from consistent in men and women and gies, but the inverse trend is simi-
the Glioma International Case-Con across most sites. larly consistent.
trol Study, which was conducted in In addition, several prospective The original study observed
2010–2013 and included 4533 cas- cohort studies with measurements inverse associations with risk of
es and 4171 controls, were consis- of pre-diagnostic plasma levels of glioma for self-reported history
tent with those of previous studies, immunoglobulin E, which reflect al- of chickenpox or shingles and for
Fig. 5.17.4. Annual age-adjusted incidence rates of primary brain and other central nervous system gliomas in the USA, and incidence
trends by age group for diagnosis years 2000–2014. APC, annual percentage change; CI, confidence interval.
458
elevated levels of immunoglobu- markers (differentially methylated on an improved understanding of

lin G antibodies to varicella zos- regions) for specific immune cell mechanisms.
ter virus [26]. These findings have types using peripheral blood DNA,
been reproduced in several studies and this has opened up the field of Prospects
[27–29]. In the Glioma International immunomethylomics [30].
Case-Control Study, a history of The development of high-dimen-
The identification of differentially
infection with varicella zoster virus sional technologies has opened up
methylated regions for immune cell
was associated with a 21% reduc- new doors to understanding brain
types, including neutrophils, lym-
tion in risk of glioma, and the as- cancer risk and survival. Large
phocytes, T cells, and regulatory T
sociation was slightly stronger for genomic studies have provided im-
cells, has provided new opportuni-
high-grade gliomas [28]. portant insights into key pathways
ties to study immune cells in relation
A cohort study with measure- that play a role in development of
to risk of glioma and survival [30,31]. brain cancer and have reinforced
ments of pre-diagnostic plasma lev- Lower levels of regulatory T cells
els of immunoglobulin G antibodies the importance of examining tu-
and lower levels of T cells were as- mour subtypes, because they are
to varicella zoster virus reported an
sociated with a higher risk of glioma likely to have different etiologies.
inverse association with risk of gli-
in a case–control setting [31], and Furthermore, improved classifica-
oma [27]. This result provides data
an elevated neutrophil-to-lympho- tion of brain tumours using molecu-
suggesting that the association ob-
cyte ratio, a marker of immunosup- lar markers can be used to better
served in case–control studies may
pression, was associated with poor predict prognosis and provide tar-
not be due solely to reverse causa-
survival in patients with glioma [30]. geted therapies.
tion. Although the biological mech-
Although case–control studies Epigenomic studies using high-
anism is not known, the immune re-
sponse clearly plays a central role are unable to examine pre-diagnos- dimensional arrays, as well as oth-
in this association. tic immune status, cohort studies er –omics analyses, will probably
examining other end-points have improve the understanding of the
Immunomethylomics suggested that these immune per- complex biological processes that
The difficulty of measuring immune turbations may exist years before lead to the development of brain
cells in archived blood samples – diagnosis [32]. Future studies using tumours. Given the lack of estab-
and thus in population studies – us- archived blood samples from pro- lished associations for modifiable
ing traditional methods (i.e. flow cy- spective cohorts will undoubtedly risk factors for brain tumours (with
tometry) has hindered progress in provide critical data in this field, the exception of exposure to ioniz-
studying altered immune states in which will offer new opportunities ing radiation), no recommendations
glioma etiology. Recently, research- for early detection and for the de- can be provided for primary or sec-
ers have identified DNA methylation velopment of therapeutics based ondary prevention.
References
CHAPTER 5.17
SECTION 5
1. Amirian ES, Zhou R, Wrensch MR, Olson 3. Ostrom QT, Gittleman H, Liao P, Vecchione- 5. Eckel-Passow JE, Lachance DH, Molinaro
SH, Scheurer ME, Il’yasova D, et al. Koval T, Wolinsky Y, Kruchko C, et al. AM, Walsh KM, Decker PA, Sicotte H, et al.
(2016). Approaching a scientific consen- (2017). CBTRUS statistical report: primary (2015). Glioma groups based on 1p/19q,
sus on the association between allergies brain and other central nervous system IDH, and TERT promoter mutations in tu-
and glioma risk: a report from the Glioma tumors diagnosed in the United States mors. N Engl J Med. 372(26):2499–508.
International Case-Control Study. Cancer in 2010–2014. Neuro Oncol. 19(Suppl https://doi.org/10.1056/NEJMoa1407279
Epidemiol Biomarkers Prev. 25(2):282–90. 5):v1–v88. https://doi.org/10.1093/neuonc/ PMID:26061753
https://doi.org/10.1158/1055-9965.EPI-15-
nox158 PMID:29117289 6. Brat DJ, Verhaak RG, Aldape KD, Yung
0847 PMID:26908595
4. Louis DN, Perry A, Reifenberger G, WK, Salama SR, Cooper LA, et al.;
2. Bondy ML, Scheurer ME, Malmer B, Cancer Genome Atlas Research Network
Barnholtz-Sloan JS, Davis FG, Il’yasova von Deimling A, Figarella-Branger D,
(2015). Comprehensive, integrative ge-
D, et al.; Brain Tumor Epidemiology Cavenee WK, et al. (2016). The 2016
nomic analysis of diffuse lower-grade
Consortium (2008). Brain tumor epidemi- World Health Organization classification
gliomas. N Engl J Med. 372(26):2481–98.
ology: consensus from the Brain Tumor of tumors of the central nervous system: a
Epidemiology Consortium. Cancer. summary. Acta Neuropathol. 131(6):803–
PMID:26061751
113(7 Suppl):1953–68. https://doi.org/10. 20. https://doi.org/10.1007/s00401-016-
1002/cncr.23741 PMID:18798534 1545-1 PMID:27157931

7. Michaud DS, Holick CN, Batchelor 16. Labreche K, Kinnersley B, Berzero G, 25. Schlehofer B, Siegmund B, Linseisen J,
TT, Giovannucci E, Hunter DJ (2009). Di Stefano AL, Rahimian A, Detrait I, Schüz J, Rohrmann S, Becker S, et al.
Prospective study of meat intake and diet al. (2018). Diffuse gliomas classified (2011). Primary brain tumours and specific
etary nitrates, nitrites, and nitrosamines by 1p/19q co-deletion, TERT promoter serum immunoglobulin E: a case-control
and risk of adult glioma. Am J Clin Nutr. and IDH mutation status are associ- study nested in the European Prospective
90(3):570–7. https://doi.org/10.3945/ajcn. ated with specific genetic risk loci. Acta Investigation into Cancer and Nutrition co-
2008.27199 PMID:19587083 Neuropathol. 135(5):743–55. https:// hort. Allergy. 66(11):1434–41. https://doi.
d o i .o r g /10 .10 07/s 0 0 4 01- 018 -18 2 5 -z org/10.1111/j.1398 -9995.2011.02670.x
8. Dubrow R, Darefsky AS, Park Y, Mayne PMID:21726235
ST, Moore SC, Kilfoy B, et al. (2010). PMID:29460007
Dietary components related to N-nitroso 26. Wrensch M, Weinberg A, Wiencke J,
17. Walsh KM, Codd V, Smirnov IV, Rice T,
compound formation: a prospective Masters H, Miike R, Barger G, et al.
Decker PA, Hansen HM, et al.; ENGAGE
study of adult glioma. Cancer Epidemiol (1997). Does prior infection with varicella-
Consortium Telomere Group (2014).
Biomarkers Prev. 19(7):1709–22. https:// zoster virus influence risk of adult glioma?
Variants near TERT and TERC influ- Am J Epidemiol. 145(7):594–7. https://doi.
doi.org/10.1158/1055-9965.EPI-10-0225
PMID:20570910 encing telomere length are associated org/10.1093/oxfordjournals.aje.a009155
with high-grade glioma risk. Nat Genet. PMID:9098175
9. Lauby-Secretan B, Scoccianti C, Loomis D, 46(7):731–5. https://doi.org/10.1038/
Grosse Y, Bianchini F, Straif K; International ng.3004 PMID:24908248 27. Sjöström S, Hjalmars U, Juto P, Wadell G,
Agency for Research on Cancer Handbook Hallmans G, Tjönneland A, et al. (2011).
Working Group (2016). Body fatness and 18. Walsh KM, Wiencke JK, Lachance Human immunoglobulin G levels of vi-
cancer – viewpoint of the IARC Working DH, Wiemels JL, Molinaro AM, Eckel- ruses and associated glioma risk. Cancer
Group. N Engl J Med. 375(8):794–8. Passow JE, et al. (2015). Telomere Causes Control. 22(9):1259–66. https://
https://doi.org/10.1056/NEJMsr1606602 maintenance and the etiology of adult d o i .o r g /10 .10 07/s10 5 5 2- 011- 97 9 9 - 3
PMID:27557308 glioma. Neuro Oncol. 17(11):1445–52. PMID:21717196
https://doi.org/10.1093/neuonc/nov082
10. Baan R, Grosse Y, Lauby-Secretan B, El 28. Amirian ES, Scheurer ME, Zhou R,
PMID:26014050
Ghissassi F, Bouvard V, Benbrahim-Tallaa Wrensch MR, Armstrong GN, Lachance
L, et al.; WHO International Agency for 19. Ostrom QT, Kinnersley B, Wrensch MR, D, et al. (2016). History of chickenpox
Research on Cancer Monograph Working Eckel-Passow JE, Armstrong G, Rice T, in glioma risk: a report from the Glioma
Group (2011). Carcinogenicity of radio- et al.; GliomaScan Consortium (2018). International Case-Control Study (GICC).
frequency electromagnetic fields. Lancet Sex-specific glioma genome-wide as- Cancer Med. 5(6):1352–8. https://doi.
Oncol. 12(7):624–6. https://doi.org/10.1016/ org/10.1002/cam4.682 PMID:26972449
sociation study identifies new risk lo-
S1470-2045(11)70147-4 PMID:21845765
cus at 3p21.31 in females, and finds 29. Wrensch M, Weinberg A, Wiencke J, Miike
11. Repacholi MH, Lerchl A, Röösli M, sex-differences in risk at 8q24.21. Sci R, Sison J, Wiemels J, et al. (2005). History
Sienkiewicz Z, Auvinen A, Breckenkamp Rep. 8(1):7352. https://doi.org/10.1038/ of chickenpox and shingles and prevalence
J, et al. (2012). Systematic review of wire- s41598-018-24580-z PMID:29743610 of antibodies to varicella-zoster virus and
less phone use and brain cancer and three other herpesviruses among adults
other head tumors. Bioelectromagnetics. 20. Claus EB, Cornish AJ, Broderick P, with glioma and controls. Am J Epidemiol.
33(3):187–206. https://doi.org/10.1002/ Schildkraut JM, Dobbins SE, Holroyd 161(10):929–38. https://doi.org/10.1093/
bem.20716 PMID:22021071 A, et al. (2018). Genome-wide asso- aje/kwi119 PMID:15870157
ciation analysis identifies a meningi-
12. Chapman S, Azizi L, Luo Q, Sitas F (2016). oma risk locus at 11p15.5. Neuro Oncol. 30. Wiencke JK, Koestler DC, Salas LA,
Has the incidence of brain cancer risen in 20(11):1485–93. https://doi.org/10.1093/ Wiemels JL, Roy RP, Hansen HM,
Australia since the introduction of mobile neuonc/noy077 PMID:29762745 et al. (2017). Immunomethylomic ap-
phones 29 years ago? Cancer Epidemiol. proach to explore the blood neutrophil
42:199–205. https://doi.org/10.1016/j. 21. Kinnersley B, Mitchell JS, Gousias K, lymphocyte ratio (NLR) in glioma sur-
canep.2016.04.010 PMID:27156022 Schramm J, Idbaih A, Labussière M, et vival. Clin Epigenetics. 9(1):10. https://
al. (2015). Quantifying the heritability of d o i . o r g /10 .118 6 /s1314 8 - 017- 0 316 - 8
13. Cronin KA, Lake AJ, Scott S, Sherman PMID:28184256
glioma using genome-wide complex trait
RL, Noone AM, Howlader N, et al. (2018).
analysis. Sci Rep. 5(1):17267. https://doi.
Annual report to the nation on the status 31. Wiencke JK, Accomando WP, Zheng S,
org/10.1038/srep17267 PMID:26625949 Patoka J, Dou X, Phillips JJ, et al. (2012).
of cancer, part I: national cancer statistics.
Cancer. 124(13):2785–800. https://doi. 22. Linos E, Raine T, Alonso A, Michaud D Epigenetic biomarkers of T-cells in hu-
org/10.1002/cncr.31551 PMID:29786848 (2007). Atopy and risk of brain tumors: man glioma. Epigenetics. 7(12):1391–
402. https://doi.org/10.4161/epi.22675
14. Toledano MB, Smith RB, Chang I, a meta-analysis. J Natl Cancer Inst.
PMID:23108258
Douglass M, Elliott P (2017). Cohort pro- 99(20):1544–50. https://doi.org/10.1093/
file: UK COSMOS – a UK cohort for study jnci/djm170 PMID:17925535 32. Barth SD, Schulze JJ, Kühn T, Raschke
of environment and health. Int J Epidemiol. E, Hüsing A, Johnson T, et al. (2015).
23. Davis FG, Al-Alem U (2011). Allergies and Treg-mediated immune tolerance and
46(3):775–87. https://doi.org/10.1093/ije/
adult gliomas: cohort results strengthen the risk of solid cancers: findings from
dyv203 PMID:26534947
evidence for a causal association. J Natl EPIC-Heidelberg. J Natl Cancer Inst.
15. Melin BS, Barnholtz-Sloan JS, Wrensch Cancer Inst. 103(21):1562–3. https://doi. 107(11):djv224. https://doi.org/10.1093/jnci/
MR, Johansen C, Il’yasova D, Kinnersley org/10.1093/jnci/djr397 PMID:22010179 djv224 PMID:26298011
B, et al.; GliomaScan Consortium (2017).
Genome-wide association study of glioma 24. Schwartzbaum J, Ding B, Johannesen
subtypes identifies specific differences in TB, Osnes LT, Karavodin L, Ahlbom A, et
genetic susceptibility to glioblastoma and al. (2012). Association between prediag-
non-glioblastoma tumors. Nat Genet. nostic IgE levels and risk of glioma. J Natl
49(5):789–94. https://doi.org/10.1038/ng. Cancer Inst. 104(16):1251–9. https://doi.
3823 PMID:28346443 org/10.1093/jnci/djs315 PMID:22855780
460
5.18
Thyroid cancer
The challenge of overdiagnosis
David O. Francis Luigino Dal Maso (reviewer)

Louise Davies Silvia Franceschi (reviewer)
Sabina Rinaldi (reviewer)
developing thyroid cancer, in- and most uniformly lethal subtype

SUMMARY cluding tumour predisposition is anaplastic thyroid cancer, a very
syndromes, multiple endocrine uncommon, poorly differentiated
●● Thyroid cancer consists of can- neoplasia type 2, or familial cancer, which originates mainly from
cers of several different histolo- medullary thyroid cancer. follicular cells [1].
gies, which differ in terms of Thyroid cancer is much more
●● Population-based screening for
cellular origin, incidence, and common in adults than in children.
thyroid cancer is not recom-
lethality. The most common In adults, survival rates for papillary
mended, because the harms
subtypes are papillary and fol- thyroid cancer are higher than 90%,
outweigh the benefits.
licular thyroid cancers. although there are more aggressive
subtypes of papillary thyroid carci-
●● In the past three decades, the noma, such as diffuse sclerosing,
incidence of thyroid cancer Thyroid cancer consists of cancers
tall cell, solid, trabecular, and onco-
(particularly of papillary thy- of several different histologies, cytic variants. In children younger
roid cancer) in adults has in- which differ in terms of cellular ori- than 10 years with a history of expo-
creased markedly, but thyroid gin, incidence, and lethality [1]. The sure to radiation, the solid variant is
cancer mortality rates have not most common subtypes are well- more common. Diffuse sclerosing
increased proportionally; this differentiated carcinomas (i.e. pa- papillary thyroid cancer is also more
pillary and follicular cancers), which common in children and in adults
suggests that overdiagnosis of
arise from the follicular cells within younger than 30 years, and these
thyroid cancer is occurring.
the thyroid gland. Papillary and fol- tumours generally do not show a
●● Although there are specific eti- licular thyroid cancers tend to have distinct nodule. Follicular thyroid
ologies that lead to the devel- a more benign course and a lower cancer is much less common and
opment of thyroid cancer, as mortality rate compared with other most often occurs in adults.
well as disparities in incidence subtypes, and together they com-
by sex and socioeconomic sta- prise more than 90% of thyroid
cancers; their proportion varies by Epidemiology
tus, most of the variation in in-
cidence trends is due to health- iodine sufficiency. Papillary thyroid In the past three decades, the in-
CHAPTER 5.18
SECTION 5
cancer spreads through the lym- cidence of thyroid cancer in adults
care system factors.
phatics, whereas follicular thyroid has doubled, tripled, or more in
●● Various risk factors associ- cancer spreads haematogenously several high-income countries [2].
ated with the development of and has a greater predilection for Dramatic increases in incidence
thyroid cancer have been in- distant metastases. have also been seen in middle-
vestigated. Robust evidence of Medullary thyroid cancer, which income countries, such as Brazil,
causal associations exists only arises from parafollicular calcitonin- China, and Turkey (Fig. 5.18.1) [3].
for radiation exposure during secreting C cells, has an interme- Studies from a few of the coun-
childhood. Emerging data indi- diate severity and mortality rate. tries with detailed registries show
cate an association with over- Medullary thyroid cancers comprise that almost the entire increase in in-
weight and obesity. fewer than 5% of thyroid cancers cidence has been due to increased
and can occur sporadically or as diagnosis of papillary thyroid can-
●● There are also genetic fac- part of the multiple endocrine neo- cer [4,5]. The size of the cancers
tors that increase the risk of plasia (MEN) syndromes. The rarest that are now being detected is
Chapter 5.18 • Thyroid cancer 461

also notable: most of the increase tain differentiated thyroid cancer,

in incidence has come from the and this rate has been stable over FUNDAMENTALS
detection of papillary thyroid can- time [10]. The high prevalence at au-
cers less than or equal to 2 cm in topsy explains the increasing identi- ■■ The most common subtypes
diameter. Given that cancers of this fication of these smaller tumours. of thyroid cancer are papillary
size are usually difficult to detect The mechanism is increas- thyroid cancer, which spreads
through physical examination (pal- ingly sensitive imaging studies. through the lymphatics, and
pation), the increased incidence of Asymptomatic thyroid nodules are follicular thyroid cancer, which
these small cancers is most likely to very common and are easily seen
spreads haematogenously.
be due to increased use of sensitive on medical imaging studies: up to
Papillary thyroid cancer is
imaging technologies. The implicat- 16% of computed tomography (CT)
ed technologies include ultrasonog- scans and magnetic resonance im- the subtype with the lowest
raphy and cross-sectional imaging aging (MRI) scans that include the mortality rate.
that includes the neck, which is thyroid gland show thyroid nodules, ■■ Less common subtypes
driven largely by practice patterns and with ultrasonography about of thyroid cancer include
of health-care providers [6]. two thirds of people will be found to
medullary thyroid cancer,
Recent studies have shown that have at least one nodule [11,12].
which has an intermediate
a large fraction of thyroid cancer di-
agnoses in high-income countries severity and mortality rate,
are likely to be due to the diagnosis
Factors affecting rates of and anaplastic thyroid cancer,
of lesions of no clinical significance
disease burden which is the rarest and most
[7]. In women, this fraction could The observed variation in thyroid uniformly lethal subtype.
be as high as 70–80% in Australia, cancer incidence rates by country
is driven by rates of well-differen- ■■ The incidence of thyroid cancer
France, Italy, and the USA and 90%
tiated thyroid cancer, in particular in women is 3 times that in men
in the Republic of Korea. In men, the
estimated fraction is 70% in France, papillary thyroid cancer. Although at all ages, but there are no
Italy, and the Republic of Korea and there are specific etiologies that differences by sex in mortality
45% in Australia and the USA. lead to the development of thyroid from thyroid cancer.
During the same period, thy- cancer, most of the variation in inci-
■■ Incidence rates of thyroid
roid cancer mortality rates have dence trends is due to health-care
system factors. cancer have increased around
not increased proportionally. This
the world. The main cause
pattern of dramatically increasing
incidence of thyroid cancer world- Sex is probably the detection of
wide, particularly of small papillary Worldwide, women are about 3 times subclinical disease that if left
thyroid cancers, with largely stable as likely as men to be diagnosed undetected would have been
mortality rates suggests that the with thyroid cancer. The reason for unlikely to cause harm to the
main cause is the diagnosis of le- this disparity is unclear. The differ- person.
sions that pose no significant risk ence may relate to the influence of
to the person [8]. For overdiag- menarche and pregnancies and cor- ■■ Detection of subclinical
nosis to occur, three factors must responding female hormonal varia- disease is largely attributable
be present: (i) subclinical disease tions, because the highest female-to- to health-care system factors
that is detectable by the screen- male ratio of thyroid cancer diagnosis and practice patterns of
ing test, (ii) a mechanism by which occurs during the reproductive peri- health-care providers.
the tumours can be identified, and od. Although hormonal factors may
(iii) health-care activities that lead play a role, the biological mechanism
to the detection [9]. The necessary of this association remains elusive
components for overdiagnosis of (see Chapter 3.6). dergo thyroid imaging because of
thyroid cancer are all present, as An argument against a biological referral bias, which results in higher
explained below. explanation for the higher incidence detection rates [13,14]. Reasons for
Thyroid cancer is a disease that rate of thyroid cancer in women is the striking disparity in thyroid can-
is readily detected subclinically. that multiple autopsy studies have cer incidence rates between men
Papillary thyroid cancer is common- shown nearly equivalent detected and women worldwide require fur-
ly found at autopsy in people who rates of thyroid cancer in men and ther elucidation.
died of other causes. Depending on women [4]. A more plausible ex-
the method of examination of the planation is the consideration that Health-care system model
thyroid, about 4% (partial examina- women have higher health-care Studies have shown that the in-
tion) to 11% (whole examination) of use during their reproductive period cidence of thyroid cancer is of-
thyroid glands can be shown to con- and therefore are more prone to un- ten higher in countries where the
462
Fig. 5.18.1. Global distribution of estimated age-standardized (World) incidence rates (ASR) per 100 000 person-years for
thyroid cancer in women, 2018.
health-care funding model includes cioeconomic status have lower rates Age at exposure is significantly re-
fee-for-service options [15]. In stud- of detection of thyroid cancer, more lated to risk. Among survivors of
ies of countries that have more than advanced stage at presentation, and the Hiroshima atomic bomb, those
one model of funding, patients treat- higher mortality from thyroid cancer who were younger than 19 years
ed at private hospitals that used a [23]. In the USA, the discrepancy in at the time of the bombing had an
fee-for-service payment model were mortality rates indicates that in some increased risk relative to the back-
found to be more likely to have thy- cases, patients with lower socioeco- ground risk, and that increased risk
roid cancer detected on unrelated nomic status may be undertreated persisted for at least five decades.
imaging compared with patients relative to those with higher socio- Those who were younger than
treated at public hospitals; this sug- economic status, although survival 5 years at the time of exposure had
gests that patients with private insur- is not always affected (see Chapter the highest risk, and those who were
ance were more likely to have thyroid 4.6) [24]. older than 19 years at the time of
cancer detected by imaging than by exposure did not have an increased
palpation [16,17]. This disparity may Etiology risk relative to the background risk
be explained by different factors, in- The vast majority of thyroid cancers [25].
cluding physician incentivization and are sporadic. However, there are Iodine deficiency can interact
the availability of advanced imaging specific risk factors. For medullary with the effects of radiation if the ra-
technology [18]. The larger the num- thyroid cancer in particular, heredi- diation is received from radioactive
bers of imaging tests ordered and tary syndromes contribute signifi- iodine. This affected the severity of
CHAPTER 5.18
SECTION 5
the more health-care providers in- cantly to the disease burden. the effects of the Chernobyl acci-
tervene for increasingly smaller find- dent, because iodine deficiency was
ings, the more thyroid cancers are Risk factors common in the populations of the
detected [19–21]. Various risk factors associated with affected areas. People who were
the development of thyroid cancer exposed thus absorbed more radio-
Socioeconomic status active iodine, and this increased the
have been investigated. Robust evi-
Recent detailed population-based dence of causal associations exists radiation dose received [26].
studies suggest that people with only for radiation exposure during After the Chernobyl accident,
higher socioeconomic status and childhood. early analyses suggested that ex-
those living in cities are more fre- posure to radiation led to more ag-
quently diagnosed with thyroid can- Radiation gressive thyroid cancer. Compared
cer, but that this does not correspond Exposure to radiation is the strong- with non-exposed children, many
with exposure to environmental pol- est known risk factor for papillary exposed children had disease that
lutants [22]. People with lower so- thyroid cancer (see Chapter 2.5). appeared to be more aggressive,

Fig. 5.18.2. The ruined reactor at the Chernobyl nuclear power plant in Ukraine. tors, additional research is required
Exposure to radiation is the strongest known risk factor for papillary thyroid cancer. to identify the mechanisms that
would lead to the development of
thyroid cancer [30].
In recent years, it has been sug-
gested that environmental and die-
tary exposure to nitrites may contrib-
ute to the development of papillary
thyroid cancer [31,32].
Hereditability
Several inherited conditions with
known genetic causes are associ-
ated with increased risk of thyroid
cancer of different cellular origins
[33]. Medullary thyroid cancer can
occur as a result of a germline ac-
tivating mutation in the RET onco-
gene. In children, medullary thyroid
cancer is most commonly associ-
ated with the MEN type 2 (MEN2)
syndrome. Increased risks of dif-
ferentiated thyroid cancer are seen
in people with PTEN hamartoma
tumour syndrome (Cowden syn-
with more extensive local invasion, In observational studies, over-
drome), DICER1 pleuropulmonary
lymph node involvement, and distant weight, obesity, and type 2 diabetes
blastoma syndrome, Carney com-
metastases. However, subsequent have all been found to be weakly plex type 1, and familial adenoma-
analysis suggested that this obser- associated with increased inci- tous polyposis syndrome.
vation was related to several varia- dence of papillary thyroid cancer. Familial differentiated thyroid
bles, including increased absorption These factors have been postulated cancer has also been noted, but no
of radioactive iodine by iodine-defi- to be associated with greater use of chromosomal abnormalities have
cient children and the initial lack of health care overall; as described yet been identified. For a patient to
a monitoring programme for chil- above, this is a known mechanism qualify as having familial non-med-
dren, who were the ones at risk of by which rates of thyroid cancer de- ullary thyroid cancer, there need to
developing thyroid cancer. When the tection may be higher in one region be three first-degree relatives with
clinical presentation and survival of than in another. For all of these fac- the disease.
exposed and non-exposed children
of the same age were compared, the
Fig. 5.18.3. A child undergoing a computed tomography (CT) scan. Exposure of
suspected difference in clinical ag- children to medical radiation may contribute to the development of thyroid cancer and
gressiveness was not observed [27]. should therefore be minimized.
Exposure to medical radiation
has increased in children, and this
may also contribute to the develop-
ment of thyroid cancer [28].
Other factors
In geographical areas where the
population has a low dietary intake
of stable iodine, there is a higher
incidence of goitre, follicular thyroid
cancer, and possibly anaplastic thy-
roid cancer. Iodine excess has been
proposed as a cause of increased
risk of papillary thyroid cancer, but
no plausible mechanism has been
identified [29].
464
Genetics and genomics tary variants (MEN2A, MEN2B, and with anaplastic carcinomas, thereby
familial medullary thyroid cancer) supporting the multistep de-differ-
Differentiated thyroid cancers have germline RET mutations, and entiation process [37,38]. The more
Abnormalities of the mitogen-acti- about 50% of sporadic cases have common somatic mutations are in
vated protein kinase (MAPK) path- somatic RET mutations. The so- the TP53 and β-catenin (CTNNB1)
way lead to both papillary and fol- matic RET codon M918T mutation genes. These mutations are rare
licular thyroid carcinoma [30]. in sporadic medullary thyroid cancer in differentiated thyroid cancers.
In adults, papillary thyroid can- has also been shown to portend a Other mutations, in BRAF and RAS,
cers commonly show point muta- more aggressive clinical course and are common in both differentiated
tions in BRAF and tend to have poorer prognosis. and anaplastic thyroid cancers and
relatively large numbers of genetic The genetics of medullary thy- are probably early events in thyroid
mutations overall. In children, rear- roid cancer are important for risk carcinogenesis that predispose to
rangements of the RET oncogene, stratification and treatment deci- tumour de-differentiation. Currently,
leading to activation of this area sion-making. Recent guidelines DNA or RNA analysis does not have
that is usually silent, are more com- designated risk categories of RET a role in the staging and manage-
mon than the BRAF mutations. In mutations as follows: “highest risk” ment of patients with anaplastic thy-
2014, the Cancer Genome Atlas includes patients with MEN2B and roid cancer [39,40].
Research Network showed a low the RET codon M918T mutation,
frequency of somatic alterations “high risk” includes patients with
RET codon C634 mutations and the
Prevention
(relative to other carcinomas for
which strong environmental risk RET codon A883F mutation, and The identification and treatment of
factors exist) and extended the “moderate risk” includes patients iodine deficiency is central to the
set of known papillary thyroid can- with RET codon mutations other prevention of thyroid cancer. In
cer driver alterations to include than M918T, C634, and A883F [36]. population-based studies, follicular
EIF1AX, PPM1D, and CHEK2 and thyroid cancer is more common in
Anaplastic thyroid cancer iodine-deficient areas in low- and
diverse gene fusions [34].
Papillary thyroid cancers in Anaplastic thyroid cancers are typi- middle-income countries, whereas
children tend to show more fusion cally aneuploid and have a complex papillary thyroid cancer is the pre-
events, rather than the pattern in karyotype with multiple chromoso- dominant subtype in countries with
adults of multiple point mutations. mal abnormalities. Loss of hetero- iodine sufficiency. Follicular thyroid
These genetic patterns may be why zygosity at multiple chromosomal cancers are more aggressive; they
thyroid cancers in children tend to regions is common. A progressive spread haematogenously, with a
be more iodine-avid and highly re- accumulation of chromosomal ab- predilection for lung metastases,
sponsive to treatment, whereas normalities is often seen when com- and have lower survival rates than
those in adults can have wider pat- paring differentiated carcinomas papillary thyroid cancers [41].
terns of spread and loss of differ-
entiation. Staging for thyroid cancer Fig. 5.18.4. Production of iodized salt on the outskirts of Vientiane, Lao People’s
reflects this: the American Joint Democratic Republic. The identification and treatment of iodine deficiency is central to
Committee on Cancer staging sys- the prevention of thyroid cancer.
tem defines all differentiated thyroid
cancers in children as stage I or II,
regardless of metastases [35].
Follicular thyroid cancers spe-
cifically are associated with point
CHAPTER 5.18
SECTION 5
mutations in other genes in the
MAPK pathway, such as RAS, or
with rearrangements of PPARγ.
Medullary thyroid cancer

The RET proto-oncogene, located
on chromosome 10q11.2, encodes a
single-pass transmembrane protein
of the receptor tyrosine kinase fam-
ily. RET is expressed in cells derived
from the neural crest, such as para-
follicular calcitonin-secreting C cells,
from which medullary thyroid cancer
arises. Most patients with heredi-

Avoiding unnecessary radiation accidents, provision of iodine thy- cause the harms outweigh the ben-
of the thyroid during childhood and roid blocking (i.e. saturating the thy- efits [42]. For the very small propor-
adolescence decreases the risk of roid gland with stable iodine) up to tion of the population with specific
papillary thyroid cancer. Even low- 24 hours before and up to 2 hours identified risks of thyroid cancer,
dose radiation of children from diag- after the exposure may be preven- such as associated tumour predis-
nostic X-rays, for example CT and tive, particularly for individuals liv- position syndromes as described
fluoroscopy, should be minimized. ing in iodine-deficient areas [33]. above, or with MEN2 or familial
Exposure to radiation increases the Population-based screening for medullary thyroid cancer, personal-
risk of thyroid cancer for decades thyroid cancer is not recommended ized screening with the appropriate
after the exposure. After nuclear by major task force bodies, be- testing method is appropriate.
References
1. Lloyd RV, Osamura RY, Klöppel G, Rosai 8. Davies L, Welch HG (2006). Increasing 15. Lee TJ, Kim S, Cho HJ, Lee JH (2012). The
J, editors (2017). WHO classification of tu- incidence of thyroid cancer in the incidence of thyroid cancer is affected by
mours of endocrine organs. Lyon, France: United States, 1973-2002. JAMA. the characteristics of a healthcare system.
International Agency for Research on 295(18):2164–7. https://doi.org/10.1001/ J Korean Med Sci. 27(12):1491–8. https://
Cancer (WHO Classification of Tumours jama.295.18.2164 PMID:16684987 doi.org/10.3346/jkms.2012.27.12.1491
series, 4th edition). Available from: http:// PMID:23255848
9. Welch HG, Black WC (2010). Overdiagnosis
publications.iarc.fr/554. in cancer. J Natl Cancer Inst. 102(9): 16. Zagzag J, Kenigsberg A, Patel KN,
605–13. https://doi.org/10.1093/jnci/djq099 Heller KS, Ogilvie JB (2017). Thyroid
2. La Vecchia C, Malvezzi M, Bosetti C,
PMID:20413742 cancer is more likely to be detected in-
Garavello W, Bertuccio P, Levi F, et al.
cidentally on imaging in private hospi-
(2015). Thyroid cancer mortality and in- 10. Furuya-Kanamori L, Bell KJL, Clark J, tal patients. J Surg Res. 215:239–44.
cidence: a global overview. Int J Cancer. Glasziou P, Doi SAR (2016). Prevalence https://doi.org/10.1016/j.jss.2017.03.059
136(9):2187–95. https://doi.org/10.1002/ of differentiated thyroid cancer in au- PMID:28688654
ijc.29251 PMID:25284703 topsy studies over six decades: a meta-
analysis. J Clin Oncol. 34(30):3672–9. 17. Altekruse S, Das A, Cho H, Petkov V,
3. Bray F, Ferlay J, Soerjomataram I, Siegel Yu M (2015). Do US thyroid cancer in-
https://doi.org/10.1200/JCO.2016.67.7419
RL, Torre LA, Jemal A (2018). Global PMID:27601555 cidence rates increase with socioeco-
cancer statistics 2018: GLOBOCAN es- nomic status among people with health
timates of incidence and mortality world- 11. Yoon DY, Chang SK, Choi CS, Yun EJ, Seo insurance? An observational study using
wide for 36 cancers in 185 countries. CA YL, Nam ES, et al. (2008). The prevalence SEER population-based data. BMJ Open.
Cancer J Clin. 68(6):394–424. https://doi. and significance of incidental thyroid nod- 5(12):e009843. https://doi.org/10.1136/
org/10.3322/caac.21492 PMID:30207593 ules identified on computed tomography. J bmjopen-2015-009843 PMID:26644126
Comput Assist Tomogr. 32(5):810–5. https://
4. Davies L, Welch HG (2014). Current thy- 18. Loehrer AP, Murthy SS, Song Z, Lubitz
doi.org/10.1097/RCT.0b013e318157fd38
roid cancer trends in the United States. CC, James BC (2017). Association of
PMID:18830117
insurance expansion with surgical man-
JAMA Otolaryngol Head Neck Surg.
12. Ezzat S, Sarti DA, Cain DR, Braunstein agement of thyroid cancer. JAMA Surg.
140(4):317–22. https://doi.org/10.1001/
GD (1994). Thyroid incidentalo- 152(8):734–40. https://doi.org/10.1001/
jamaoto.2014.1 PMID:24557566 jamasurg.2017.0461 PMID:28384780
mas. Prevalence by palpation and
5. Ahn HS, Kim HJ, Welch HG (2014). Korea’s ultrasonography. Arch Intern Med.
19. Smith-Bindman R, Miglioretti DL, Johnson
thyroid-cancer “epidemic” – screen- 154(16):1838–40. https://doi.org/10.1001/
E, Lee C, Feigelson HS, Flynn M, et al.
ing and overdiagnosis. N Engl J Med. a r c h i n t e .1 9 9 4 . 0 0 4 2 0 1 6 0 0 7 5 0 1 0 (2012). Use of diagnostic imaging stud-
371(19):1765–7. https://doi.org/10.1056/ PMID:8053752 ies and associated radiation exposure
NEJMp1409841 PMID:25372084 13. Germano A, Schmitt W, Almeida P, for patients enrolled in large integrated
Mateus-Marques R, Leite V (2018). health care systems, 1996-2010. JAMA.
6. Davies L, Morris LGT, Haymart M, Chen 307(22):2400–9. https://doi.org/10.1001/
AY, Goldenberg D, Morris J, et al.; AACE Ultrasound requested by general prac-
titioners or for symptoms unrelated to jama.2012.5960 PMID:22692172
Endocrine Surgery Scientific Committee
the thyroid gland may explain higher 20. Udelsman R, Zhang Y (2014). The epi-
(2015). American Association of Clinical
prevalence of thyroid nodules in fe- demic of thyroid cancer in the United
Endocrinologists and American College
males. Clin Imaging. 50:289–93. https:// States: the role of endocrinologists
of Endocrinology Disease State Clinical
doi.org/10.1016/j.clinimag.2018.05.003 and ultrasounds. Thyroid. 24(3):472–9.
Review: the increasing incidence of thy- PMID:29738997 ht tps://doi.org/10.1089/thy. 2013.0257
roid cancer. Endocr Pract. 21(6):686–96.
PMID:23937391
https://doi.org/10.4158/EP14466.DSCR 14. Singh Ospina N, Brito JP, Maraka S,
PMID:26135963 Espinosa de Ycaza AE, Rodriguez- 21. Zevallos JP, Hartman CM, Kramer JR,
Gutierrez R, Gionfriddo MR, et al. (2016). Sturgis EM, Chiao EY (2015). Increased
7. Vaccarella S, Franceschi S, Bray F, Wild CP, Diagnostic accuracy of ultrasound-guided thyroid cancer incidence corresponds
Plummer M, Dal Maso L (2016). Worldwide fine needle aspiration biopsy for thyroid to increased use of thyroid ultrasound
thyroid-cancer epidemic? The increasing malignancy: systematic review and meta- and fine-needle aspiration: a study of
impact of overdiagnosis. N Engl J Med. analysis. Endocrine. 53(3):651–61. https:// the Veterans Affairs health care sys-
375(7):614–7. https://doi.org/10.1056/ d o i .o r g /10 .10 07/s12 0 2 0 - 016 - 0 9 21-x tem. Cancer. 121(5):741–6. https://doi.
NEJMp1604412 PMID:27532827 PMID:27071659 org/10.1002/cncr.29122 PMID:25376872
466
22. Fei X, Lou Z, Christakos G, Liu Q, Ren 29. Zimmermann MB, Galetti V (2015). Iodine 36. Wells SA Jr, Asa SL, Dralle H, Elisei R,
Y, Wu J (2018). Contribution of industrial intake as a risk factor for thyroid cancer: a Evans DB, Gagel RF, et al.; American
density and socioeconomic status to the comprehensive review of animal and hu- Thyroid Association Guidelines Task Force
spatial distribution of thyroid cancer risk man studies. Thyroid Res. 8(1):8. https:// on Medullary Thyroid Carcinoma (2015).
in Hangzhou, China. Sci Total Environ. d o i.o r g /10.118 6 /s13 0 4 4 - 015 - 0 0 2 0 - 8 Revised American Thyroid Association
613–614:679–86. https://doi.org/10.1016/j. PMID:26146517 guidelines for the management of medul-
scitotenv.2017.08.270 PMID:28938210 lary thyroid carcinoma. Thyroid. 25(6):567–
30. Dunlap QD (2020). Differentiated thyroid 610. https://doi.org/10.1089/thy.2014.0335
23. Swegal WC, Singer M, Peterson E, cancer incidence. In: Randolph GW, edi- PMID:25810047
Feigelson HS, Kono SA, Snyder S, et al. tor. Surgery of the thyroid and parathyroid
(2016). Socioeconomic factors affect out- glands. 3rd ed. Philadelphia (PA), USA: 37. Wreesmann VB, Ghossein RA, Patel
comes in well-differentiated thyroid cancer. Elsevier. (forthcoming) SG, Harris CP, Schnaser EA, Shaha AR,
Otolaryngol Head Neck Surg. 154(3):440–5. et al. (2002). Genome-wide appraisal of
https://doi.org/10.1177/0194599815620778 31. Vigneri R, Malandrino P, Gianì F, Russo thyroid cancer progression. Am J Pathol.
PMID:26671905 M, Vigneri P (2017). Heavy metals in 161(5):1549–56. https://doi.org/10.1016/
the volcanic environment and thyroid S0002-9440(10)64433-1 PMID:12414503
24. Harari A, Li N, Yeh MW (2014). Racial and cancer. Mol Cell Endocrinol. 457:73–80.
socioeconomic disparities in presenta- https://doi.org/10.1016/j.mce.2016.10.027 38. Rodrigues RF, Roque L, Rosa-Santos J,
tion and outcomes of well-differentiated PMID:27794445 Cid O, Soares J (2004). Chromosomal
thyroid cancer. J Clin Endocrinol Metab. imbalances associated with anaplas-
99(1):133–41. https://doi.org/10.1210/jc. 32. Poulsen R, Cedergreen N, Hayes T, tic transformation of follicular thyroid
2013-2781 PMID:24243631 Hansen M (2018). Nitrate: an environmental carcinomas. Br J Cancer. 90(2):492–6.
endocrine disruptor? A review of evidence https://doi.org/10.1038/sj.bjc.6601530
25. Furukawa K, Preston D, Funamoto S, and research needs. Environ Sci Technol. PMID:14735198
Yonehara S, Ito M, Tokuoka S, et al. 52(7):3869–87. https://doi.org/10.1021/acs.
(2013). Long-term trend of thyroid can- est.7b06419 PMID:29494771 39. Smallridge RC, Ain KB, Asa SL, Bible KC,
cer risk among Japanese atomic-bomb Brierley JD, Burman KD, et al.; American
survivors: 60 years after exposure. Int 33. IARC Expert Group on Thyroid Health Thyroid Association Anaplastic Thyroid
J Cancer. 132(5):1222–6. https://doi. Monitoring after Nuclear Accidents (2018). Cancer Guidelines Taskforce (2012).
org/10.1002/ijc.27749 PMID:22847218 Thyroid health monitoring after nuclear American Thyroid Association guidelines
accidents. Lyon, France: International for management of patients with anaplas-
26. Iglesias ML, Schmidt A, Ghuzlan AA, Agency for Research on Cancer (IARC tic thyroid cancer. Thyroid. 22(11):1104–
Lacroix L, Vathaire F, Chevillard S, et al. Technical Publications, No. 46). Available 39. https://doi.org/10.1089/thy.2012.0302
(2017). Radiation exposure and thyroid from: http://publications.iarc.fr/571. PMID:23130564
cancer: a review. Arch Endocrinol Metab.
61(2):180–7. https://doi.org/10.1590/2359- 34. Agrawal N, Akbani R, Aksoy BA, Ally A, 40. Haddad RI, Lydiatt WM, Ball DW, Busaidy
3997000000257 PMID:28225863 Arachchi H, Asa SL, et al.; Cancer Genome NL, Byrd D, Callender G, et al. (2015).
Atlas Research Network (2014). Integrated Anaplastic thyroid carcinoma, ver-
27. Reiners C, Biko J, Haenscheid H, genomic characterization of papillary sion 2.2015. J Natl Compr Canc Netw.
Hebestreit H, Kirinjuk S, Baranowski thyroid carcinoma. Cell. 159(3):676–90. 13(9):1140–50. https://doi.org/10.6004/
O, et al. (2013). Twenty-five years after https://doi.org/10.1016/j.cell.2014.09.050 jnccn.2015.0139 PMID:26358798
Chernobyl: outcome of radioiodine treat- PMID:25417114
ment in children and adolescents with 41. Woodruff SL, Arowolo OA, Akute OO,
very high-risk radiation-induced differenti- 35. Amin MB, Greene FL, Compton CC, Afolabi AO, Nwariaku F (2010). Global var-
ated thyroid carcinoma. J Clin Endocrinol Gershenwald JE, Brookland RK, Meyer iation in the pattern of differentiated thyroid
Metab. 98(7):3039–48. https://doi.org/10. L, et al. (2017). The Eighth Edition AJCC cancer. Am J Surg. 200(4):462–6. https://
1210/jc.2013-1059 PMID:23616148 Cancer Staging Manual: continuing to doi.org/10.1016/j.amjsurg.2010.03.009
build a bridge from a population-based to PMID:20887838
28. Mathews JD, Forsythe AV, Brady Z, a more “personalized” approach to cancer
Butler MW, Goergen SK, Byrnes GB, et staging. CA Cancer J Clin. 67(2): 93–99. 42. Lin JS, Bowles EJA, Williams SB, Morrison
al. (2013). Cancer risk in 680 000 people h t t p s: //d o i . o r g /10 . 3 3 2 2 /c a a c . 213 8 8 CC (2017). Screening for thyroid cancer:
exposed to computed tomography scans PMID:28094848 updated evidence report and systematic
in childhood or adolescence: data linkage review for the US Preventive Services
study of 11 million Australians. BMJ. 346: Task Force. JAMA. 317(18):1888–903.
f2360. https://doi.org/10.1136/bmj.f2360 https://doi.org/10.1001/jama.2017.0562
PMID:23694687 PMID:28492904
CHAPTER 5.18
SECTION 5

5.19
Non-Hodgkin lymphoma
Complex etiology, including the role
of immune function
Sonja I. Berndt Franco Cavalli (reviewer)
Karin Ekström Smedby (reviewer)
●● To date, more than 120 genetic deaths (2.6% of all cancer deaths) in
SUMMARY susceptibility loci have been 2018 [2]. The incidence varies glob-
identified for lymphoid malig- ally (Fig. 5.19.1).
●● Non-Hodgkin lymphoma com- nancies, including variants in The age-standardized incidence
prises more than 50 different the human leukocyte antigen rate for both sexes combined is
neoplasms that arise from im- (HLA) region. 9.3 per 100 000 in more-developed
mature or mature B cells, T
regions, compared with 4.2 per
cells, or natural killer cells.
100 000 in less-developed regions
● ● The incidence varies glob- Non-Hodgkin lymphoma (NHL) com- (Fig. 5.19.2). Increased detection
ally. The age-standardized prises more than 50 different neo- (especially of more indolent lympho-
rate for both sexes combined plasms that arise from lymphocytes mas), solid organ transplantation,
is 9.3 per 100 000 in more- and can manifest in the lymph nodes, and immunosuppression are some
developed regions, compared lymphatic organs, and extranodal factors that are hypothesized to con-
with 4.2 per 100 000 in less- lymphatic tissue. The classification tribute to this difference, but environ-
developed regions. of these tumours has changed over mental, viral, or genetic factors may
●● Accurate diagnosis is impera- time with advances in molecular also play a role.
tive for disease management technology and the implementation Despite differences in incidence,
and treatment. Classification of of the WHO classification system. age-adjusted mortality rates are
lymphoid malignancies under- NHLs are classified broadly by line- similar in more-developed regions
pins diagnosis and is based on age as either B-cell neoplasms or (2.7 per 100 000) and less-devel-
a combination of morphologi- natural killer (NK)/T-cell neoplasms oped regions (2.3 per 100 000).
cal, phenotypic, genetic/molec- (Table 5.19.1). This may reflect better access to
ular, and clinical features. In the WHO classification sys- treatment and a higher proportion
tem, plasma cell tumours (e.g. mul- of indolent lymphomas in more-de-
●● The etiology of non-Hodgkin tiple myeloma) and lymphoid leu-
lymphoma is complex, with veloped regions. Less-developed
kaemias (see Chapter 5.20) are regions have a greater proportion of
multiple known or suspected
considered B-cell lymphoid malig- poor-prognosis NK/T-cell lympho-
risk factors. Evidence suggests
nancies. Lymphoid malignancies are
that some risk factors are com- mas (13.4%) and high-grade B-cell
then further subtyped within major
mon to multiple subtypes of lymphomas (59.6%) compared with
WHO categories on the basis of a
non-Hodgkin lymphoma, but more-developed regions (9.3% and
combination of morphological, phe-
others are subtype-specific. 39.2%, respectively) [3].
notypic, genetic/molecular, and clini-
The incidence trends and pat-
●● Established causes of non- cal features [1]. These subtype clas-
Hodgkin lymphoma include terns of NHL subtypes vary world-
sifications are used in determining
chronic infections (e.g. hepa- wide. In the USA, the incidence rates
disease management and treatment.
titis C virus), autoimmune dis- of most NHL types are high but ap-
eases (e.g. Sjögren syndrome), pear to be relatively stable or declin-
immune alterations (e.g. im- Epidemiology ing [4]. Although incidence rates in
munosuppression), exposure NHL is the 13th most common can- Asia are lower than those in the USA,
to lindane, and family history cer type worldwide, with an esti- the incidence rates of many lympho-
of non-Hodgkin lymphoma or mated 509 600 new cases (2.8% of ma subtypes are rising in Japan [5]
haematological malignancy. all new cancer cases) and 248 700 and other Asian countries, possibly
468
Table 5.19.1. Subtypes of non-Hodgkin lymphoma based on the 2016 WHO classification
B-cell neoplasms NK/T-cell neoplasms FUNDAMENTALS

Precursor acute lymphoblastic Precursor acute lymphoblastic leukaemia/
leukaemia/lymphoma, B-cell lymphoma, T-cell ■■ Lymphomas are clonal
Prolymphocytic leukaemia, B-cell Prolymphocytic leukaemia, T-cell tumours of lymphocytes and
Chronic lymphocytic leukaemia/small T-cell large granular lymphocytic leukaemia can manifest in the lymph
lymphocytic lymphoma
Hairy cell leukaemia Aggressive NK-cell leukaemia
nodes, lymphatic organs, and
Mantle cell lymphoma Adult T-cell leukaemia/lymphoma extranodal lymphatic tissue.
Marginal zone lymphoma Systemic EBV-positive T-cell lymphoma of
childhood ■■ Many lymphomas are charac-
Extranodal marginal zone lymphoma Extranodal NK/T-cell lymphoma, nasal type terized by recurrent chromoso-
of mucosa-associated lymphoid tissue mal translocations, such as the
(MALT lymphoma)
Nodal marginal zone lymphoma Peripheral T-cell lymphoma
t(11;14) translocation in mantle
Splenic marginal zone lymphoma Angioimmunoblastic T-cell lymphoma cell lymphoma.
Follicular lymphoma Hepatosplenic T-cell lymphoma
■■ These chromosomal transloca-
Paediatric-type follicular lymphoma Enteropathy-associated T-cell lymphoma
Primary cutaneous follicle centre Anaplastic large cell lymphoma, ALK-positive
tions may be generated during
lymphoma the extensive genetic remodel-
Diffuse large B-cell lymphoma Anaplastic large cell lymphoma, ALK-negative ling that occurs during normal
Diffuse large B-cell lymphoma, NOS Subcutaneous panniculitis-like T-cell lymphoma
lymphocyte maturation as part
Germinal centre B-cell subtype Primary cutaneous gamma delta T-cell lymphoma
of the adaptive immune system.
Activated B-cell subtype Monomorphic epitheliotropic intestinal T-cell
lymphoma
■■ Non-Hodgkin lymphomas are
T-cell/histiocyte-rich large B-cell Hydroa vacciniforme-like lymphoproliferative
lymphoma disorder classified broadly by lineage
Primary DLBCL of the central nervous Peripheral T-cell lymphoma, NOS as either B-cell neoplasms
system
or natural killer/T-cell neo-
Primary cutaneous DLBCL, leg type Primary cutaneous CD30-positive T-cell
lymphoproliferative disorders plasms. About 85–90% of
EBV-positive DLBCL, NOS Primary cutaneous anaplastic large cell lymphoma lymphomas are derived from
DLBCL associated with chronic Mycosis fungoides B lymphocytes, and natural
inflammation
killer/T-cell lymphomas are
Primary mediastinal (thymic) large Sézary syndrome
B-cell lymphoma much less common.
Intravascular large B-cell lymphoma
■■ In the WHO classification sys-
ALK-positive large B-cell lymphoma
Plasmablastic lymphoma
tem, non-Hodgkin lymphomas
Primary effusion lymphoma are further categorized into
HHV8-positive DLBCL, NOS specific types. The incidence
Burkitt lymphoma of non-Hodgkin lymphoma and
Burkitt-like lymphoma with 11q aberration the distribution of types vary
Lymphoplasmacytic lymphoma
worldwide.
Waldenström macroglobulinaemia
Multiple myeloma, plasma cell myeloma ■■ Chronic antigen stimulation,
Plasmacytoma
immunosuppression, immune
CHAPTER 5.19
SECTION 5
Heavy chain diseases, mu/gamma/alpha
dysfunction, and hereditary/
High-grade B-cell lymphoma with MYC
and BCL2 and/or BCL6 rearrangements genetic factors are thought to
High-grade B-cell lymphoma, NOS contribute to the risk of non-
DLBCL, diffuse large B-cell lymphoma; EBV, Epstein–Barr virus; HHV8, human herpesvirus type 8; Hodgkin lymphoma.
NK, natural killer; NOS, not otherwise specified.
as the result of changes in lifestyle or compared with rates in Whites in In the USA, the incidence of most
environmental exposures. the USA [5,6]. For Asians living in B-cell lymphomas is higher in non-
Unlike most NHL types, the in- the USA, the incidence appears to Hispanic Whites than in other racial
cidence rate of Epstein–Barr virus be intermediate [6], suggesting that or ethnic groups; however, NK/T-cell
(EBV)-related nasal NK/T-cell lym- both environmental and host fac- lymphomas, such as mycosis fun-
phoma is higher in Asian countries tors may contribute to risk. goides, peripheral T-cell lymphoma
Chapter 5.19 • Non-Hodgkin lymphoma 469

Fig. 5.19.1. Global distribution of estimated age-standardized (World) incidence rates (ASR) per 100 000 person-years for non-
Hodgkin lymphoma in both sexes, 2018.
Fig. 5.19.2. Estimated age-standardized (World) incidence and mortality rates (ASR) per 100 000 person-years for non-Hodgkin
lymphoma, by sex and region, 2018.
470
(PTCL), and adult T-cell leukaemia/ line genetic loci associated with the NK/T-cell lymphoma [10], and vari-
lymphoma, are more common in risk of different lymphoid malignan- ants in both HLA class I and class
non-Hispanic Blacks [4,7]. The per- cies (Fig. 5.19.3). The majority of II are associated with follicular lym-
centage of PTCL cases is also higher discovered loci confer only a small phoma and MZL [11,12].
in southern Africa [3]; this suggests a increase in susceptibility and appear Susceptibility loci (see Chapter
possible genetic component. to be subtype-specific. However, a 3.2) have been discovered for both
Overall, incidence rates of NHL few loci overlap among subtypes, DLBCL and follicular lymphoma at
are higher in males than in females and some chromosomal regions are chromosome 8q24 near MYC [9,11],
(Fig. 5.19.2), but this difference var- important for multiple subtypes even a region known to be associated
ies substantially by subtype; the if the variants are subtype-specific. with multiple different cancer types.
greatest excess risk is seen for Genetic variants in the human Genetic variation near LPP at chro-
mantle cell lymphoma, Burkitt lym- leukocyte antigen (HLA) region, mosome 3q27.3–3q28, a region
phoma, and hairy cell leukaemia [4]. a gene encoding the major histo- also associated with immune-relat-
Little difference between the sexes compatibility complex proteins re- ed diseases, is associated with both
is observed for marginal zone lym- sponsible for immune function, are follicular lymphoma in subjects of
phoma (MZL); this may reflect the associated with multiple lymphoma European descent [11] and DLBCL
higher prevalence of autoimmune
subtypes, including follicular lym- in the Chinese population [13].
diseases in women and the strong
phoma, diffuse large B-cell lympho- Many B-cell lymphomas are
association between autoimmune
ma (DLBCL), and MZL in European characterized by chromosomal
disease and risk of MZL [8].
populations and extranodal NK/T- translocations, often involving the
cell lymphoma in Asian popula- immunoglobulin heavy chain locus,
Genetics and genomics tions. Variants in HLA class I are although copy number alterations
Genome-wide association studies associated with DLBCL [9], variants and mutations may also be pres-
have identified more than 120 germ- in HLA class II are associated with ent. These somatic alterations may
Fig. 5.19.3. Established genetic loci for specific lymphoid malignancies. To date, most loci have been discovered in populations of
European ancestry. Two loci have been identified in populations of East Asian ancestry: one locus for B-cell lymphoma, particularly
diffuse large B-cell lymphoma (DLBCL), and one locus for natural killer/T-cell lymphoma (NKTCL). ALL, acute lymphoblastic
leukaemia; CLL, chronic lymphocytic leukaemia; FL, follicular lymphoma; HL, Hodgkin lymphoma; MM, multiple myeloma; MZL,
marginal zone lymphoma; NSHL, nodular sclerosing Hodgkin lymphoma; WM, Waldenström macroglobulinaemia.
CHAPTER 5.19
SECTION 5

be the result of a deviation in the Fig. 5.19.4. Lymphocyte development and cell lineages. Lymphoid malignancies
normal lymphocyte maturation pro- arise from immature or mature B cells, T cells, or natural killer cells. MHC, major
cess as part of the adaptive immune histocompatibility complex.
system, which generates broad
antibody diversity and specificity
through V(D)J gene recombination
(which involves DNA double-strand
breaks), germinal centre reaction,
clonal expansion, somatic hypermu-
tation of immunoglobulin G genes,
class-switch recombination, selec-
tion, and differentiation/apoptosis.
Etiology and biological

characteristics
Lymphomas arise from clonal tu-
mours of B cells, T cells, or NK
cells (Fig. 5.19.4) that have ar-
rested during different stages of
differentiation. Emerging evidence
indicates that the etiology of NHL
is complex, with subtype-specific
patterns of risk [14].
Known or suspected risk factors
include immune alterations (e.g. im-
munosuppression), viral infections
(e.g. hepatitis C virus [HCV] and hu-
man T-cell lymphotropic virus type 1
[HTLV-1]), autoimmune diseases
(e.g. Sjögren syndrome), environ-
mental or occupational exposures
(e.g. benzene and pentachlorophe- germinal centre B-cell subtype, among sites of origin. Chronic
nol), and lifestyle factors. Some risk activated B-cell subtype, or other. infections (e.g. HCV) and autoim-
factors are shared across multiple About 1–12% of DLBCL tumours mune diseases, particularly B-cell
subtypes and may be generally as- are high-grade B-cell lymphomas activating diseases (e.g. Sjögren
sociated with risk of NHL, but others with MYC and BCL2 and/or BCL6 syndrome) are associated with an
are likely to be specific to individual rearrangements; these double-hit increased risk of DLBCL, implicat-
subtypes. or triple-hit lymphomas, which are ing chronic immune stimulation in
now classified separately [1], have the pathogenesis of DLBCL. Solid
Diffuse large B-cell lymphoma a worse prognosis. More recently, organ transplantation is a risk fac-
DLBCL is an aggressive B-cell lym- next-generation sequencing tech- tor, possibly as a result of chronic
phoma that accounts for 25–45% of nologies have been used to further immune activation in response to
NHL cases. It is the most common classify DLBCLs on the basis of the donor organ, immunosuppres-
adult lymphoma worldwide. DLBCL specific mutations and chromoso- sion therapy, or both, resulting in
can originate in lymph nodes or mal rearrangements into four or five immune dysfunction [18]. HIV infec-
extranodal sites, such as the gas- additional categories with poten- tion is also a risk factor for DLBCL,
trointestinal tract, the testis, and tial prognostic significance [15,16]. particularly primary central nervous
the central nervous system. It is a Recurrent mutations in MYD88 and system lymphoma, possibly due to
heterogeneous group of lympho- CD79B, frequently found in primary immunosuppression. Family his-
mas on the basis of histology, im- central nervous system lymphoma, tory of NHL is associated with an
munophenotype, and clinical pre- may lead to activation of the nuclear increased risk, implicating genetic
sentation, and some DLBCLs, such factor kappa-light-chain-enhancer factors. Other suggestive risk fac-
as primary mediastinal large B-cell of activated B (NF-κB) signalling tors include higher body mass in-
lymphoma, are classified separate- pathway [17]. dex, lower socioeconomic status,
ly by WHO [1]. The etiology of DLBCL is com- working as a farmer or field crop
DLBCL tumours can be catego- plex, with multiple known or sus- worker, and occupation as a hair-
rized according to cell of origin as pected risk factors and differences dresser [19].
472
Fig. 5.19.5. A crop duster spraying a cornfield. Nodal MZL, which accounts for
10–25% of MZL cases, occurs in
the lymph nodes and has a hetero-
geneous morphology and cytology.
Transformation to DLBCL occurs
in about 15% of patients with nodal
MZL. Splenic MZL occurs in the
spleen, blood, and bone marrow; de-
letion of 7q and NOTCH2 mutations
are characteristic of the malignancy.
Chronic infection, autoimmune
disorders, inflammation, and anti-
gen stimulation are thought to be
strong contributors to the etiology
of MZL. Infection with Helicobacter
pylori is observed in most cases
of gastric MALT lymphoma, and
eradication of H. pylori with antibi-
otic treatment leads to regression
of MALT lymphoma in 75–80% of
cases. H. pylori infection is thought
Follicular lymphoma protective against follicular lym- to trigger inflammation and immu-
Follicular lymphoma is a slow- phoma, possibly because of an in- nological responses, leading to
growing B-cell malignancy that ac- creased response against cancer- the positive selection of malignant
counts for 12–20% of NHL cases. It specific or cancer-related antigens B cells. HCV infection is associ-
is the second most common adult and early eradication of tumour ated with an increased risk of MZL,
lymphoma in Europe and the USA. cells. Exposure to the chlorinated particularly splenic MZL and nodal
insecticide lindane, which is classi- MZL. Chronic antigen stimula-
The 5-year survival rates tend to be
fied as carcinogenic to humans on tion leading to B-cell stimulation is
higher than 80%, with 2–3% of cas-
the basis of epidemiological studies thought to underlie the association,
es transforming to DLBCL per year.
showing an increased risk of NHL, and interferon-based antiviral treat-
Follicular lymphoma arises from
may be a stronger risk factor for fol- ment leads to disease regression
the transformation of germinal cen-
licular lymphoma [22]. Exposures in more than 70% of cases [25].
tre B cells with varying proportions
to trichloroethylene and other chlo- B-cell activating autoimmune con-
of centroblasts and centrocytes,
rinated solvents are suspected risk ditions, such as Sjögren syndrome
which determine the pathological
factors [23,24]. and systemic lupus erythematosus,
grade of the lymphoma. Grades 1
are strongly associated with an in-
and 2 are considered low-grade creased risk of MZL [8].
disease, whereas grade 3B is more
Marginal zone lymphoma
aggressive. About 80–90% of folli- MZL is a slow-growing B-cell ma- Mantle cell lymphoma
cular lymphomas display a t(14;18) lignancy that accounts for 7–11%
Mantle cell lymphoma is a rare,
translocation, in which the BCL2 of NHL cases. MZL arises from the
aggressive B-cell lymphoma that
gene is joined to an immunoglobu- marginal zone or edge of lymphoid
makes up about 3–6% of NHL cas-
lin heavy (IGH) gene, and a subset tissue. There are three distinct
es. It occurs more often in men than
CHAPTER 5.19
SECTION 5
of cases have BCL6 translocations. types of MZL: extranodal, nodal, in women, and more often in Whites
Pesticide exposure has been asso- and splenic. Extranodal MZL of than in Blacks or Asians [4]. It often
ciated with t(14;18) translocations mucosa-associated lymphoid tis- involves the bone marrow, spleen,
[20], suggesting a possible etiologi- sue (MALT lymphoma) is the most peripheral blood, and gastrointestinal
cal link. common type of MZL, accounting tract. Mantle cell lymphoma is char-
Evidence from epidemiological for about two thirds of MZL cases. It acterized by the chromosomal trans-
studies points to several risk factors occurs outside the lymph nodes at location t(11;14) and overexpression
for follicular lymphoma, including a variety of anatomical sites, includ- of cyclin D1, which is observed in
family history of NHL [21]. Unlike ing the stomach, salivary glands, most cases. Overexpression of the
DLBCL, solid organ transplanta- thyroid, and lung. Several chromo- transcription factor SOX11 is of-
tion does not appear to increase somal translocations, some of which ten present, but absence of SOX11
risk [18], and autoimmune diseases involve genes encoding NF-κB expression is associated with a
appear to play a smaller role [21]. regulators, have been reported for more favourable prognosis. Overall,
Allergy and hay fever appear to be MALT lymphoma. mantle cell lymphoma has a poor

prognosis, with 5-year survival rates Fig. 5.19.6. A girl aged 9 years sits with her mother and baby sister before undergoing
of less than 50%. treatment for Burkitt lymphoma at Bugando Medical Centre in Mwanza, United
The etiology of mantle cell Republic of Tanzania.
lymphoma is not well understood.
Unlike many other NHL subtypes,
solid organ transplantation and
most autoimmune diseases do not
appear to be associated with risk
of mantle cell lymphoma. Hay fever
and allergy appear to be protective
against mantle cell lymphoma, and
having a first-degree relative with a
haematological malignancy is asso-
ciated with an increased risk [26].
Living on a farm may also be asso-
ciated with increased risk.
Burkitt lymphoma
Burkitt lymphoma is an aggressive,
rapidly growing B-cell NHL involving
the jaw, central nervous system, colo-
rectum, kidney, or other organs. The
hallmark of Burkitt lymphoma is the
presence of translocations involving
MYC and an immunoglobulin gene
(e.g. IGH). Although they are histo- The risk factors for sporadic tations in RHOA. Anaplastic large
logically indistinguishable, there are Burkitt lymphoma are not well un- cell lymphomas are characterized
three etiological subtypes of Burkitt derstood. In developed countries, by the chromosomal translocation
lymphoma: endemic, immunodefi- the incidence of Burkitt lymphoma t(2;5)(p23;q35) involving ALK.
ciency-associated, and sporadic. peaks in childhood and then again Except for a history of coeliac
Endemic Burkitt lymphoma oc- in late adulthood, and the incidence disease, which is associated pri-
curs primarily in equatorial Africa rate in males is substantially higher marily with enteropathy-associated
and Papua New Guinea, where than that in females. For younger T-cell lymphoma, there are few
Plasmodium falciparum malaria is cases, a history of allergy is asso- established risk factors for PTCL.
holoendemic. It is the most common ciated with a reduced risk of Burkitt HIV infection has been linked to an
childhood cancer in those coun- lymphoma, suggesting that immuno- increased risk of PTCL [28], impli-
tries, and nearly 100% of tumours logical hypersensitivity may be im- cating immune dysregulation in the
are positive for EBV. Although re- portant [27]. For older cases, HCV pathogenesis. Although this is rare,
cent malaria infections are hypoth- infection may be a risk factor. textured breast implants appear
esized to contribute to endemic to increase the risk of anaplastic
Burkitt lymphoma, the mechanism Peripheral T-cell lymphoma large cell lymphoma [29], possibly
and interaction with EBV are not PTCL is the most common T-cell through chronic immune stimula-
well understood. lymphoma and accounts for 4–7% tion. Recent evidence suggests that
Immunodeficiency-associated of NHL cases in Europe and the psoriasis and eczema may be asso-
Burkitt lymphoma occurs primar- USA. It is a heterogeneous group ciated with increased risk of PTCL,
ily in individuals with HIV infection of lymphomas with diverse mor- whereas allergy may be protective
and less commonly after organ phological and clinical features. [30]. Family history of any haema-
transplantation. Sporadic Burkitt Predominantly nodal PTCLs include tological malignancy is associated
lymphoma makes up about 30% of anaplastic large cell lymphoma, an- with an increased risk.
lymphoid malignancies in children gioimmunoblastic T-cell lymphoma,
and about 1–5% in adults in devel- and PTCL not otherwise specified.
oped countries and often occurs in A subset of PTCLs, including an- Socioeconomic
the abdomen. In contrast to endem- gioimmunoblastic T-cell lymphoma differences
ic Burkitt lymphoma, EBV is identi- and some PTCLs not otherwise The diagnosis and classification
fied in only 30–60% of immunode- specified, have features of folli- of lymphomas remain challenging
ficiency-related Burkitt lymphoma cular helper T cells. Up to 75% of in low- and middle-income coun-
tumours and only 15–30% of spo- these lymphomas have mutations tries, where immunohistochemis-
radic Burkitt lymphoma tumours. in TET2, and about 60% have mu- try and other technologies needed
474
to make an accurate diagnosis are areas have a substantially lower per- and immunosuppression increase
often unavailable. Less-developed centage of cases in women [3], sug- the risk of NHL. Prevention or early
countries tend to have a higher per- gesting that sex disparities in medi- treatment of these infections and
centage of unclassifiable cases and cal care may exist in some regions. diseases can decrease the inci-
more misclassified cases compared dence of some subtypes of NHL.
with more-developed countries [3]. Reduced exposure to lindane and
As accurate and more refined clas-
Prevention other suspected lymphomagens
sification becomes more critical to Much progress has been made in (such as benzene) may also be ben-
disease management and treat- identifying risk factors associated eficial. Further research on the eti-
ment, these disparities could result with specific NHL types. There is ology of specific NHL subtypes and
in greater mortality differences in convincing evidence that some in- the identification of early biomark-
the future. Although NHL is more fections (e.g. HCV), autoimmune ers may offer insights into pathways
common in men, some geographical diseases (e.g. Sjögren syndrome), of prevention.
References
1. Swerdlow SH, Campo E, Pileri SA, 7. Adams SV, Newcomb PA, Shustov AR 13. Tan DE, Foo JN, Bei JX, Chang J, Peng
Harris NL, Stein H, Siebert R, et (2016). Racial patterns of peripheral R, Zheng X, et al. (2013). Genome-wide
al. (2016). The 2016 revision of the T-cell lymphoma incidence and surviv- association study of B cell non-Hodgkin
World Health Organization classifica- al in the United States. J Clin Oncol. lymphoma identifies 3q27 as a suscep-
tion of lymphoid neoplasms. Blood. 34(9):963–71. https://doi.org/10.1200/JCO. tibility locus in the Chinese population.
127(20):2375–90. https://doi.org/10.1182/ 2015.63.5540 PMID:26962200 Nat Genet. 45(7):804–7. https://doi.
blood-2016-01-643569 PMID:26980727 org/10.1038/ng.2666 PMID:23749188
8. Bracci PM, Benavente Y, Turner JJ,
2. Bray F, Ferlay J, Soerjomataram I, Siegel Paltiel O, Slager SL, Vajdic CM, et al. 14. Morton LM, Slager SL, Cerhan JR, Wang
RL, Torre LA, Jemal A (2018). Global (2014). Medical history, lifestyle, family SS, Vajdic CM, Skibola CF, et al. (2014).
cancer statistics 2018: GLOBOCAN es- history, and occupational risk factors for Etiologic heterogeneity among non-Hodg-
timates of incidence and mortality world- marginal zone lymphoma: the InterLymph kin lymphoma subtypes: the InterLymph
wide for 36 cancers in 185 countries. CA Non-Hodgkin Lymphoma Subtypes Non-Hodgkin Lymphoma Subtypes Project.
Cancer J Clin. 68(6):394–424. https://doi. Project. J Natl Cancer Inst Monogr. J Natl Cancer Inst Monogr. 2014(48):130–
org/10.3322/caac.21492 PMID:30207593 2014(48):52–65. https://doi.org/10.1093/ 44. https://doi.org/10.1093/jncimonographs/
3. Perry AM, Diebold J, Nathwani BN, jncimonographs/lgu011 PMID:25174026 lgu013 PMID:25174034
MacLennan KA, Müller-Hermelink HK, 9. Cerhan JR, Berndt SI, Vijai J, 15. Schmitz R, Wright GW, Huang DW,
Bast M, et al. (2016). Non-Hodgkin lym- Ghesquières H, McKay J, Wang SS, et al. Johnson CA, Phelan JD, Wang JQ, et al.
phoma in the developing world: review of (2014). Genome-wide association study (2018). Genetics and pathogenesis of dif-
4539 cases from the International Non-
identifies multiple susceptibility loci for fuse large B-cell lymphoma. N Engl J Med.
Hodgkin Lymphoma Classification Project.
diffuse large B cell lymphoma. Nat Genet. 378(15):1396–407. https://doi.org/10.1056/
Haematologica. 101(10):1244–50. https://
46(11):1233–8. https://doi.org/10.1038/ NEJMoa1801445 PMID:29641966
doi.org/10.3324/haematol.2016.148809
ng.3105 PMID:25261932
PMID:27354024 16. Chapuy B, Stewart C, Dunford AJ, Kim
10. Li Z, Xia Y, Feng LN, Chen J-R, Li H-M, J, Kamburov A, Redd RA, et al. (2018).
4. Teras LR, DeSantis CE, Cerhan JR,
Cui J, et al. (2016). Genetic risk of ex- Molecular subtypes of diffuse large B
Morton LM, Jemal A, Flowers CR (2016).
tranodal natural killer T-cell lymphoma: cell lymphoma are associated with dis-
2016 US lymphoid malignancy statistics by
World Health Organization subtypes. CA a genome-wide association study. tinct pathogenic mechanisms and out-
CHAPTER 5.19
Lancet Oncol. 17(9):1240–7. https://doi. comes. Nat Med. 24(5):679–90. https://
SECTION 5
Cancer J Clin. 66(6):443–59. https://doi.
org/10.3322/caac.21357 PMID:27618563 org /10.1016/S1470 -20 4 5(16) 3 014 8 - 6 d o i.o r g /10.10 3 8 /s 415 91- 018 - 0 016 - 8
PMID:27470079 PMID:29713087
5. Chihara D, Ito H, Matsuda T, Shibata A,
Katsumi A, Nakamura S, et al. (2014). 11. Skibola CF, Berndt SI, Vijai J, Conde L, 17. Phelan JD, Young RM, Webster DE,
Differences in incidence and trends of Wang Z, Yeager M, et al. (2014). Genome- Roulland S, Wright GW, Kasbekar M, et
haematological malignancies in Japan wide association study identifies five sus- al. (2018). A multiprotein supercomplex
and the United States. Br J Haematol. ceptibility loci for follicular lymphoma out- controlling oncogenic signalling in lym-
164(4):536–45. https://doi.org/10.1111/bjh. side the HLA region. Am J Hum Genet. phoma. Nature. 560(7718):387–91. https://
12659 PMID:24245986 95(4):462–71. https://doi.org/10.1016/j. d o i.o r g /10.10 3 8 /s 415 8 6 - 018 - 0 2 9 0 - 0
ajhg.2014.09.004 PMID:25279986 PMID:29925955
6. Bassig BA, Au WY, Mang O, Ngan
R, Morton LM, Ip DKM, et al. (2016). 12. Vijai J, Wang Z, Berndt SI, Skibola CF, 18. Clarke CA, Morton LM, Lynch C, Pfeiffer
Subtype-specific incidence rates of lym- Slager SL, de Sanjose S, et al. (2015). A RM, Hall EC, Gibson TM, et al. (2013).
phoid malignancies in Hong Kong com- genome-wide association study of mar- Risk of lymphoma subtypes after solid
pared to the United States, 2001–2010. ginal zone lymphoma shows association to organ transplantation in the United
Cancer Epidemiol. 42:15–23. https:// the HLA region. Nat Commun. 6(1):5751. States. Br J Cancer. 109(1):280–8.
doi.org /10.1016/ j.c anep. 2016.02.0 07 ht t ps: //doi.org /10.10 3 8 /nc omms 6751 ht t ps: //doi.org /10.10 3 8 / bjc. 2013. 2 9 4
PMID:26991956 PMID:25569183 PMID:23756857

19. Cerhan JR, Kricker A, Paltiel O, Flowers 23. Cocco P, Vermeulen R, Flore V, Nonne 27. Mbulaiteye SM, Morton LM, Sampson JN,
CR, Wang SS, Monnereau A, et al. (2014). T, Campagna M, Purdue M, et al. (2013). Chang ET, Costas L, de Sanjosé S, et
Medical history, lifestyle, family history, Occupational exposure to trichloro- al. (2014). Medical history, lifestyle, fam-
and occupational risk factors for diffuse ethylene and risk of non-Hodgkin lym- ily history, and occupational risk factors
large B-cell lymphoma: the InterLymph phoma and its major subtypes: a pooled for sporadic Burkitt lymphoma/leukemia:
Non-Hodgkin Lymphoma Subtypes InterLymph analysis. Occup Environ Med. the Interlymph Non-Hodgkin Lymphoma
Project. J Natl Cancer Inst Monogr. 70(11):795–802. https://doi.org/10.1136/ Subtypes Project. J Natl Cancer Inst
2014(48):15–25. https://doi.org/10.1093/ oemed-2013-101551 PMID:23881218 Monogr. 2014(48):106–14. https://doi.
jncimonographs/lgu010 PMID:25174023 o r g /10.10 9 3 / j nc i m o n o g r ap hs / l g u 0 0 3
24. Callahan CL, Stewart PA, Friesen MC, PMID:25174031
20. Chiu BC, Dave BJ, Blair A, Gapstur SM, Locke S, De Roos AJ, Cerhan JR, et al.
Zahm SH, Weisenburger DD (2006). (2018). Case-control investigation of oc- 28. Gibson TM, Morton LM, Shiels MS, Clarke
Agricultural pesticide use and risk of cupational exposure to chlorinated sol- CA, Engels EA (2014). Risk of non-Hodgkin
t(14;18)-defined subtypes of non-Hodgkin vents and non-Hodgkin’s lymphoma. lymphoma subtypes in HIV-infected people
during the HAART era: a population-based
lymphoma. Blood. 108(4):1363–9. https:// Occup Environ Med. 75(6):415–20. https://
study. AIDS. 28(15):2313–8. https://doi.
doi.org/10.1182/blood-2005 -12- 008755 doi.org /10.113 6 /oemed -2017-10 4 8 9 0
org/10.1097/Q AD.0000000000000428
PMID:16621961 PMID:29588333
PMID:25111081
21. Linet MS, Vajdic CM, Morton LM, de Roos 25. Arcaini L, Besson C, Frigeni M, Fontaine
29. Leberfinger AN, Behar BJ, Williams NC,
AJ, Skibola CF, Boffetta P, et al. (2014). H, Goldaniga M, Casato M, et al. (2016).
Rakszawski KL, Potochny JD, Mackay
Medical history, lifestyle, family history, Interferon-free antiviral treatment in B-cell
DR, et al. (2017). Breast implant-as-
and occupational risk factors for follicular lymphoproliferative disorders associated sociated anaplastic large cell lympho-
lymphoma: the InterLymph Non-Hodgkin with hepatitis C virus infection. Blood. ma: a systematic review. JAMA Surg.
Lymphoma Subtypes Project. J Natl 128(21):2527–32. https://doi.org/10.1182/ 152(12):1161–8. https://doi.org/10.1001/
Cancer Inst Monogr. 2014(48):26–40. blood-2016-05-714667 PMID:27605512 jamasurg.2017.4026 PMID:29049466
https://doi.org/10.1093/jncimonographs/
lgu006 PMID:25174024 26. Smedby KE, Sampson JN, Turner JJ, 30. Wang SS, Flowers CR, Kadin ME, Chang
Slager SL, Maynadié M, Roman E, et al. ET, Hughes AM, Ansell SM, et al. (2014).
22. Alavanja MC, Hofmann JN, Lynch CF, (2014). Medical history, lifestyle, family Medical history, lifestyle, family history,
Hines CJ, Barry KH, Barker J, et al. history, and occupational risk factors for and occupational risk factors for periph-
(2014). Non-Hodgkin lymphoma risk and mantle cell lymphoma: the InterLymph eral T-cell lymphomas: the InterLymph
insecticide, fungicide and fumigant use in Non-Hodgkin Lymphoma Subtypes Non-Hodgkin Lymphoma Subtypes
the Agricultural Health Study. PLoS One. Project. J Natl Cancer Inst Monogr. Project. J Natl Cancer Inst Monogr.
9(10):e109332. https://doi.org/10.1371/ 2014(48):76–86. https://doi.org/10.1093/ 2014(48):66–75. https://doi.org/10.1093/
journal.pone.0109332 PMID:25337994 jncimonographs/lgu007 PMID:25174028 jncimonographs/lgu012 PMID:25174027
476
5.20
Leukaemias
Understanding pathogenesis through
similarities and differences
Eve Roman Martha S. Linet (reviewer)
Alexandra G. Smith Joachim Schüz (reviewer)
The leukaemias (literally “white Critically, most of the neoplasms

SUMMARY blood”) comprise a heterogeneous listed in the ICD-O-3 categories of
group of more than 30 lymphoid and myelodysplastic syndromes and my-
●● Globally, there is a lack of pop- myeloid malignancies with diverse eloproliferative neoplasms still ap-
ulation-based descriptive data etiologies, treatment pathways, and pear with a code beginning with “D”
for many leukaemia subtypes,
outcomes [1]. They are classified by (neoplasms of unknown or uncer-
of which there are more than
cell of origin (Fig. 5.20.1). tain behaviour) in the International
30. This information is required
Leukaemias were first recog- Statistical Classification of Diseases
to inform etiological hypothe-
ses, plan health-care services, nized as a distinct entity in the and Related Health Problems, 10th
and monitor the impact of ther- 1850s [2]. The taxonomy of leukae- revision (ICD-10).
apeutic change. mias has changed markedly over Such radical changes in classifi-
time, as biological understanding cation, together with the breadth of
●● Different subtypes of leukaemia of the similarities and differences investigations required to implement
dominate at different ages. For between the various haematologi- the classification system (histol-
example, B-cell acute lympho- cal malignancies – leukaemias, ogy, cytology, immunophenotyping,
blastic leukaemia is most com-
lymphomas, and myelomas – and cytogenetics, flow cytometry, and
mon in children younger than
their relationship to the normal clinical data), continue to pose sig-
15 years, and chronic lympho-
bone marrow and immune system nificant challenges for population-
cytic leukaemia, myeloprolif-
has increased. However, contem- based cancer registries; many strug-
erative neoplasms, and acute
myeloid leukaemia are far more porary population-based informa- gle to capture all diagnoses, and
common at older ages. tion about the occurrence and out- often continue to report using the
come for many leukaemia subtypes traditional leukaemia grouping [4,5].
●● For reasons that are unknown, is sparse, and for some of the rarer In 2018, there were an estimated
almost every leukaemia sub- entities is mostly non-existent. 437 000 new cases of leukaemia
type has a male predominance. This absence of data largely worldwide, and leukaemia was the
●● In high-income countries, sur- reflects the paradigm-changing 15th most common cancer type, ac-
vival rates vary widely from one nature of the WHO classification counting for 2.4% of all new cancer
CHAPTER 5.20
SECTION 5
subtype to another. The 5-year implemented in 2001 (the basis for cases [6]. However, because many
relative survival is more than the International Classification of countries still do not have high-qual-
80% for chronic lymphocytic Diseases for Oncology, third edition ity and representative cancer reg-
leukaemia and chronic myeloid [ICD-O-3]), which, for the first time, istration systems, examining global
leukaemia but less than 20% incorporated genetic data with infor- variation and trends over time is
for other subtypes, such as mation on immunology, morphology, challenging for any cancer type; for
acute myeloid leukaemia. and clinical parameters [3]. This re- leukaemias, the situation is exacer-
●● Increased understanding of sulted not only in significant refine- bated by the diagnostic challenges
pathogenesis has resulted in ments to previously defined catego- associated with identifying the vari-
marked improvements in sur- ries but also in the addition of several ous leukaemia subtypes, coupled
vival for some leukaemia sub- new entities, including the myelodys- with the inconsistent implementation
types, including chronic my- plastic syndromes and myeloprolif- of the WHO classification [1,7,8].
eloid leukaemia. erative neoplasms, which form part Furthermore, even in countries with
of the myeloid leukaemia spectrum. good cancer registration systems,
Chapter 5.20 • Leukaemias 477

there is a lack of consistency in the as well as the development of sev-

policies applied to progressions and eral specialist registries [12,13]. One FUNDAMENTALS
transformations (e.g. from myelo- such source is the United Kingdom
dysplastic syndromes to acute my- Haematological Malignancy Re ■■ Originating in blood-forming
eloid leukaemia [AML]); for exam- search Network (HMRN; https:// tissues, usually the bone mar-
ple, the United States Surveillance, www.hmrn.org), which since 2004 row, the leukaemias comprise
a heterogeneous group of lym-
Epidemiology, and End Results has collated detailed information on phoid and myeloid malignan-
(SEER) programme has different all newly diagnosed haematologi- cies. This simple topographic
rules to the European Network of cal malignancies arising in a popu- categorization – cancer in the
Cancer Registries [9,10]. lation of about 4 million [14]. The blood – reflects the pattern of
In low-income countries, where HMRN data for the 12 years from spread rather than the origin.
mortality and morbidity from infec- September 2004 to August 2016 ■■ The 2001 WHO classification
tions and nutritional conditions are (n = 29 329) for the major subtypes of haematological malignan-
often high, diagnosing leukaemia (Fig. 5.20.2) illustrate where the leu- cies, which groups cancers
presents additional challenges. kaemias sit within the broad WHO according to their cell of origin,
The symptoms of many types of ICD-O-3 cell-of-origin haematologi- was adopted into worldwide
leukaemia are broadly similar to cal malignancy spectrum. clinical practice but did not
The leukaemias account for have an immediate effect on
those of infectious and/or para-
population-based epidemio-
sitic illnesses, and the diagnostic about 40% of all haematological ma-
logical research. Increasing
expertise and/or technologies re- lignancies. They comprise all my- recognition that the lack of
quired to enable leukaemia to be eloid subtypes and several lymphoid data on clinically meaningful
distinguished from background in- subtypes. The main leukaemia groups was impeding scientific
fections are often lacking. subtypes are shown in Fig. 5.20.3. progress has led to recent im-
Mature B-cell chronic lymphocytic provements in national cancer
leukaemia (CLL) is the largest cate- registration procedures as well
Descriptive epidemiology gory, followed by the myeloprolifera- as the development of special-
Good-quality population-based de- ist registries.
tive neoplasms, the AMLs, and the
scriptive data are required not myelodysplastic syndromes. ■■ Although the majority of
only to inform etiological hypoth- Historically, when CLL cells leukaemia subtypes do not
eses and plan health-care services were found in lymph nodes rather appear to have major environ-
but also to monitor the impact of mental determinants, a few
than in peripheral blood, the dis-
therapeutic change in the general well-established risk factors
ease was termed small lymphocytic continue to produce strong
population. This need is particu- lymphoma. The different names associations, for example
larly pertinent in fast-moving ar- reflected differences in disease cytotoxic chemotherapy and/
eas like haemato-oncology, where spread rather than in origin. For or radiotherapy and acute
treatment protocols are subject to research purposes, CLL is increas- myeloid leukaemia/myelodys-
rapid change, and “gold standard” ingly grouped with other mature plastic syndromes.
randomized controlled trials, which B-cell malignancies, both lympho- ■■ Knowledge relating to ge-
tend to be conducted almost exclu- mas and myelomas, and/or with netic determinants has
sively in higher-income countries, the non-Hodgkin lymphomas, both increased markedly over the
are frequently restricted to specific T-cell and B-cell; the same is true past 5 years. The number of
patient subgroups, often compris- for hairy cell leukaemia, which also predisposition syndromes
ing younger people with fewer co- recognized to be associ-
has a mature B-cell origin [15,16].
ated with certain leukaemia
morbidities. Furthermore, in some However, most population-based subtypes is increasing, and a
countries, particularly low-income registries still include CLL and hairy chapter on myeloid neoplasms
countries and/or those where uni- cell leukaemia in their “all leukae- with germline predisposition
versal health coverage is lacking, mia” category [4,11]. is included in the most recent
the likelihood of both treatment and For information and complete- WHO classification.
trial entry often varies with socio- ness, data on monoclonal B-cell ■■ The leukaemias have led
economic status, sex, and ethnicity. lymphocytosis, which has an the field of cancer geno-
In recent years, there has been ICD-O-3 behaviour code of 1 (and mics. Since the advent of the
an increasing recognition that sci- is not listed in ICD-10), are also in- first targeted cancer therapy
entific progress is being impeded by cluded in Fig. 5.20.3. Monoclonal (tyrosine kinase inhibitors),
the lack of reliable population-based B-cell lymphocytosis is defined by advances in molecular biology
incidence and survival data on the a monoclonal B-cell count of less and therapy have continued
to transform the landscape for
various leukaemia subtypes [11]. than 5 × 10 9/L in peripheral blood several – but by no means all –
This has led to improvements in na- [1]. Because about 75% of cases leukaemia subtypes.
tional cancer registration procedures have a CLL phenotype, monoclonal
478
Fig. 5.20.1. Overview of haematopoiesis. Leukaemias are classified by cell of origin.
B-cell lymphocytosis is increasingly contrast, at older ages, CLL, the and the relative contribution of each
being studied with a view to increas- myeloproliferative neoplasms, and varies from one subtype to another.
ing the understanding of pathogen- the AMLs are far more common With respect to environmental ex-
esis of CLL (defined by a monoclo- (Fig. 5.20.4). Variations with sex posures, relatively little has changed
nal B-cell count of ≥ 5 × 10 9/L with are also marked; the overall male in the past 5 years; well-established
CLL morphology and phenotype). predominance is evident across the risk factors continue to produce
The overall incidence of leukae- full age spectrum and the main di- strong associations but explain only
mia, like that of many other types of agnostic subtypes. a small proportion of the total burden
cancer, increases with increasing Additional differences are evident of disease. Examples of such asso-
age, and the incidence rate is high- within subtypes [17], as illustrated ciations include those with cytotoxic
er in men than in women. However, in Fig. 5.20.5, which presents sex chemotherapy, benzene, ionizing
in contrast to many other cancer rate ratios for myelodysplastic syn- radiation, and viral infections such
types, leukaemias can occur at any dromes and AML. Myelodysplastic as human T-cell lymphotropic virus
syndrome with deletion of chromo-
CHAPTER 5.20
SECTION 5
age, and different subtypes domi- type 1 (HTLV-1), which is a neces-
nate at different ages. The hetero- some 5q has a strong female pre- sary but not sufficient cause of the
geneity of the various leukaemia dominance, in contrast to the other comparatively rare adult T-cell leu-
subtypes (excluding monoclonal subtypes of myelodysplastic syn- kaemia/lymphoma (see Chapter 2.2).
drome. AML with myelodysplasia-
B-cell lymphocytosis) is illustrated HTLV-1 causes leukaemia in about
related changes has a strong male
in Fig. 5.20.4, which distributes the 5% of people infected with the virus.
predominance, whereas AML with
data by age at diagnosis and sex. Although HTLV-1 is endemic in parts
MLL rearrangement is more com-
Acute lymphoblastic leukaemias of Japan, South America, Papua New
mon in females.
(ALL), notably B-cell ALL, which Guinea, Africa, and the Middle East,
accounts for less than 4% of the it is hardly ever found elsewhere.
total, predominate in children youn- Risk factors With respect to broader envi-
ger than 15 years, an age group in Like all diseases, the leukaemias ronmental associations, systematic
which some leukaemia subtypes have both genetic and environmen- trends with frequently used proxies
are often rare or non-existent. In tal determinants to their etiology, of exposure are rarely observed

Fig. 5.20.2. Diagnostic distribution of haematological malignancies classified by the International Classification of Diseases for
Oncology, third edition (ICD-O-3). Data from the Haematological Malignancy Research Network (HMRN) for 2004–2016 (n = 29 329).
DLBCL, diffuse large B-cell lymphoma; HL, Hodgkin lymphoma; NOS, not otherwise specified.
for leukaemias, in contrast to many potential etiological role of specific tion of Down syndrome with AML
other cancer types. For example, risk factors, such as exposure to and ALL and of Fanconi anaemia
in high-income countries the inci- antibiotics, non-ionizing radiation, and other bone marrow failure syn-
dence of several common cancer or hair dyes, often produce results dromes with myelodysplastic syn-
types tends to vary with regularly that are weak and inconsistent. An dromes and AML.
used markers of socioeconomic extensive up-to-date review of all In contrast to knowledge about
status or lifestyle, including educa- the evidence relating to the envi- environmental determinants, knowl-
tion level, income, and deprivation ronmental determinants of leukae- edge relating to the genetic deter-
level, for reasons that are related mia in children and adults can be minants of several leukaemia sub-
either to etiology – exemplified by found in the latest edition of Cancer types has increased markedly over
lung cancer and smoking, or cervi- Epidemiology and Prevention [18]. the past 5 years. This increase is,
cal cancer and human papillomavi- As with environmental deter- at least in part, due to the advent of
rus (HPV) infection – or to detec- minants, certain genetic features new genomic technologies and their
tion, as illustrated by colon cancer that predispose towards leukaemia growing accessibility to the wider
and screening. The consistency of have long been known. Perhaps scientific community. As a result, the
such observations often helps to the most notable is male sex, number of predisposition syndromes
target public health interventions which is generally associated with recognized to be associated with
and policies that aim either to pre- an increased risk across the age certain leukaemia subtypes, particu-
vent the development of disease spectrum (Fig. 5.20.4). In addition, larly (but not exclusively) those of the
(see Chapter 6.1) or to detect it at certain congenital disorders are myeloid lineage, has increased con-
an early stage (see Chapter 6.6). strongly associated with the subse- siderably. Knowledge in this area is
However, for the leukaemias, quent development of the acute leu- advancing rapidly. A chapter on my-
coherent patterns of this type are kaemias, usually those occurring eloid neoplasms with germline pre-
rarely observed. Findings from epi- in children, adolescents, or young disposition (inherited and de novo)
demiological studies examining the adults. Examples are the associa- is, for the first time, included in the
480
Fig. 5.20.3. Diagnostic distribution of leukaemias (including monoclonal B-cell lymphocytosis) classified by the International
Classification of Diseases for Oncology, third edition (ICD-O-3). Data from the Haematological Malignancy Research Network
(HMRN) for 2004–2016 (n = 11 231).
most recent WHO classification, and part of routine clinical practice of which – unlike the single chromo-
associations between genetic condi- for many decades [1]. However, somal translocation and resulting
tions and lymphoid leukaemias, no- these methods have limitations. aberrant fusion protein in CML –
tably B-cell ALL, are also discussed Conventional cytogenetics are lim- have complex pathogenic pathways.
in the relevant chapters [1]. ited to detecting structural changes Although next-generation sequenc-
at a chromosome level, whereas ing and other techniques are rapidly
smaller abnormalities such as point becoming part of routine diagnostic
Genomics, survival, and
mutations are not detectable, and practice in some settings, the incor-
treatment molecular cytogenetics can only be poration of this information in other
CHAPTER 5.20
SECTION 5
The leukaemias have led the field of targeted at known abnormalities. settings, particularly in low-income
cancer genomics. In the 1960s, the Accordingly, new techniques that countries, remains challenging.
Philadelphia translocation was dis- have been developed in the past The scientific advances that
covered in chronic myeloid leukae- 15 years are increasingly being used have led to improvements in sur-
mia (CML), a subtype of myelopro- for the diagnosis, classification, and vival for some leukaemia subtypes
liferative neoplasms. This discovery prognostication of the leukaemias. are a major success story. In high-
eventually resulted in the develop- These include DNA sequencing and income countries, survival rates for
ment of the first targeted therapy in array-based platforms with next- paediatric B-cell ALL now exceed
cancer, a BCR-ABL tyrosine kinase generation sequencing, which cur- 90%, and survival rates for acute
inhibitor, which has transformed rently provides the greatest genomic promyelocytic leukaemia, a subtype
outcomes in CML [19]. resolution (see Chapter 3.2). Recent of AML, are about 80%. Tyrosine
Chromosomal analysis, under- studies using these techniques are kinase inhibitors have transformed
taken through either classical or revealing the complexity of many CML from a comparatively rare fa-
molecular techniques, has been leukaemia subtypes [20–23], many tal cancer to a long-term condition

Fig. 5.20.4. Incidence proportions of leukaemias (excluding monoclonal B-cell lymphocytosis) distributed by subtype within age
strata, age-standardized (world, 2000–2025) rates per 100 000, and 5-year overall survival (OS) and relative survival (RS). Sex
rate ratio is male rate divided by female rate. Data from the Haematological Malignancy Research Network (HMRN) for 2004–2016,
followed up September 2018.
with a survival rate that approaches population-based data is 13.2%. survival, 84.1%) and the myelopro-
that of the general population. Such For AML, the median age at diag- liferative neoplasms (5-year relative
progress has redirected the re- nosis is about 70 years. Although survival, 92.1%), is relatively good,
search efforts to other types of leu- the frequency of curative therapy is despite the fact that these cancers
kaemia, and to other cancer types. relatively high in younger patients, are currently incurable. The path-
However, despite these improve- who often comprise the focus of ways of patients with these more
ments, the outlook for older people clinical trials involving ALLs as well chronic cancers often follow a remit-
and those with aggressive subtypes as AMLs, the inability of some pa- ting–relapsing course, with patients
remains poor. Contemporary esti- tients, notably older patients, to tol- being monitored until chemothera-
mates of 5-year overall survival and erate intensive chemotherapy regi- py treatment is required, and some
relative survival from the HMRN mens remains problematic. never receiving treatment at all.
population-based patient cohort The increased application of
are shown in Fig. 5.20.4, both by genomic technologies is leading to Prevention and early
age strata for all subtypes com- the development of new targeted detection
bined and by major subtype by all agents, including monoclonal anti- In recent decades, advances in mo-
ages combined. The corresponding bodies. However, at present, most lecular biology and therapy have
relative survival curves are shown of these agents still need to be used transformed the landscape for sev-
in Fig. 5.20.6. in conjunction with intensive che- eral leukaemia subtypes. However,
Although some subtypes of motherapy, so little progress has in general this progress has not
AML are potentially curable with been made to date for the treatment been matched by similar insights
intensive chemotherapy, over the of patients who cannot tolerate into the etiological determinants of
past three decades there has been such regimens [24]. the majority of leukaemias. In such
little improvement for the majority In contrast, the outlook for pa- circumstances, the development of
of AML patients. The 5-year rela- tients with more indolent leukae- preventive strategies that will af-
tive survival for AML in the HMRN mias, including CLL (5-year relative fect the total burden of leukaemia
482
is challenging. However, it is clear Fig. 5.20.5. Sex rate ratios (male rate divided by female rate) for subtypes of myelo-
that reduction in population expo- dysplastic syndromes and acute myeloid leukaemia (AML).
sures to well-known leukaemogenic
agents such as polycyclic aromatic
hydrocarbons should be pursued.
In addition, radiological diagnostic
and therapeutic procedures involv-
ing ionizing radiation should be
used only when clinically required,
and at the lowest possible doses.
With respect to the potential
impact on high-risk groups, more
careful monitoring of individuals
with recognized leukaemia predis-
position syndromes or other genetic
susceptibilities is one area where
improvements could be made. For
example, the onset of bone mar-
row failure, a prelude to AML, could
perhaps be detected at an earlier
stage, enabling pre-emptive hae-
matopoietic stem cell transplanta-
tion to be undertaken.
However, in situations where
primary prevention is not possible,
early detection and improved treat-
ments tend to be the major focus.
In this respect, the landscape for
the leukaemias is changing rapidly,
with new diagnostic technologies
and less toxic targeted novel agents
emerging, providing considerable
promise for the future.
Fig. 5.20.6. Relative survival curves for leukaemias classified by the International Classification of Diseases for Oncology, third edition
(ICD-O-3). Data from the Haematological Malignancy Research Network (HMRN) for 2004–2016, followed up September 2018.
CHAPTER 5.20
SECTION 5

References
1. Swerdlow SH, Campo E, Harris NL, Jaffe 9. Ruhl J, Adamo M, Dickie L (2015). 17. Roman E, Smith A, Appleton S, Crouch
ES, Pileri SA, Stein H, et al., editors (2017). Hematopoietic and lymphoid neoplasm S, Kelly R, Kinsey S, et al. (2016). Myeloid
WHO classification of tumours of hae- coding manual. Bethesda (MD), USA: malignancies in the real-world: occur-
matopoietic and lymphoid tissues. Lyon, National Cancer Institute. rence, progression and survival in the
France: International Agency for Research UK’s population-based Haematological
on Cancer (WHO Classification of Tumours 10. HAEMACARE Working Group (2010). Malignancy Research Network 2004–15.
series, revised 4th edition). Available from: Manual for coding and reporting haema- Cancer Epidemiol. 42:186–98. https://
http://publications.iarc.fr/556. tological malignancies. Tumori. 96(4):i– doi.org /10.1016 / j.c anep. 2016.03.011
A32. PMID:20968151 PMID:27090942
2. Piller G (2001). Leukaemia – a brief his-
torical review from ancient times to 1950. 11. Miranda-Filho A, Piñeros M, Ferlay J, 18. Thun M, Linet MS, Cerhan JR, Haiman
Br J Haematol. 112(2):282–92. https://doi. Soerjomataram I, Monnereau A, Bray CA, Schottenfeld D, editors (2018).
org/10.1046/j.1365 -2141.2001.02411.x F (2018). Epidemiological patterns of Cancer epidemiology and prevention.
PMID:11167820 leukaemia in 184 countries: a popu- 4th ed. New York (NY), USA: Oxford
lation-based study. Lancet Haematol. University Press.
3. Jaffe ES, Harris NL, Stein H, Vardiman JW, 5(1):e14–24. https://doi.org/10.1016/S2352-
editors (2001). Pathology and genetics of tu- 3026(17)30232-6 PMID:29304322 19. Apperley JF (2015). Chronic myeloid leukae-
mours of haematopoietic and lymphoid tis- mia. Lancet. 385(9976):1447–59. https://
sues. Lyon, France: International Agency for 12. Juliusson G, Lazarevic V, Hörstedt A-S, doi.org/10.1016/S0140-6736(13)62120-0
Research on Cancer (WHO Classification Hagberg O, Höglund M; Swedish Acute PMID:25484026
of Tumours series, 3rd edition). Leukemia Registry Group (2012). Acute
myeloid leukemia in the real world: 20. Papaemmanuil E, Gerstung M, Bullinger
4. Cronin KA, Lake AJ, Scott S, Sherman why population-based registries are L, Gaidzik VI, Paschka P, Roberts ND,
RL, Noone A-M, Howlader N, et al. (2018). needed. Blood. 119(17):3890–9. https:// et al. (2016). Genomic classification
Annual report to the nation on the status of doi.org/10.1182/blood-2011-12-379008 and prognosis in acute myeloid leuke-
cancer, part I: national cancer statistics. PMID:22383796 mia. N Engl J Med. 374(23):2209–21.
Cancer. 124(13):2785–800. https://doi. https://doi.org/10.1056/NEJMoa1516192
org/10.1002/cncr.31551 PMID:29786848 13. Østgård LSG, Nørgaard JM, Raaschou-
PMID:27276561
Jensen KK, Pedersen RS, Rønnov-
5. Allemani C, Matsuda T, Di Carlo V, Jessen D, Pedersen PT, et al. (2016). 21. Papaemmanuil E, Gerstung M, Malcovati
Harewood R, Matz M, Nikšić M, et al.; The Danish National Acute Leukemia L, Tauro S, Gundem G, Van Loo P, et al.;
CONCORD Working Group (2018). Global Registry. Clin Epidemiol. 8:553–60. Chronic Myeloid Disorders Working Group
surveillance of trends in cancer survival ht tps://doi.org/10. 2147/CLEP.S9 9 46 0 of the International Cancer Genome
2000-14 (CONCORD-3): analysis of indi- PMID:27822099 Consortium (2013). Clinical and biological
vidual records for 37 513 025 patients di- implications of driver mutations in mye-
agnosed with one of 18 cancers from 322 14. Smith A, Howell D, Crouch S, Painter D,
lodysplastic syndromes. Blood. 122(22):
population-based registries in 71 coun- Blase J, Wang HI, et al. (2018). Cohort
3616–27, quiz 3699. https://doi.org/10.1182/
tries. Lancet. 391(10125):1023–75. https:// profile: the Haematological Malignancy
blood-2013-08-518886 PMID:24030381
doi.org/10.1016/S0140-6736(17)33326-3 Research Network (HMRN); a UK popula-
PMID:29395269 tion-based patient cohort. Int J Epidemiol. 22. Strefford JC (2015). The genomic land-
47(3):700–700g. https://doi.org/10.1093/ije/ scape of chronic lymphocytic leukaemia:
6. Bray F, Ferlay J, Soerjomataram I, Siegel dyy044 PMID:29618056 biological and clinical implications. Br
RL, Torre LA, Jemal A (2018). Global J Haematol. 169(1):14–31. https://doi.
cancer statistics 2018: GLOBOCAN es- 15. Teras LR, DeSantis CE, Cerhan JR,
org/10.1111/bjh.13254 PMID:25496136
timates of incidence and mortality world- Morton LM, Jemal A, Flowers CR (2016).
wide for 36 cancers in 185 countries. CA 2016 US lymphoid malignancy statis- 23. Vainchenker W, Kralovics R (2017).
Cancer J Clin. 68(6):394–424. https://doi. tics by World Health Organization sub- Genetic basis and molecular pathophysi-
org/10.3322/caac.21492 PMID:30207593 types. CA Cancer J Clin. 66(6):443–59. ology of classical myeloproliferative neo-
h t t p s: //d o i . o r g /10 . 3 3 2 2 /c a a c . 213 5 7 plasms. Blood. 129(6):667–79. https://
7. Arber DA, Orazi A, Hasserjian R, Thiele PMID:27618563 doi.org/10.1182/blood-2016-10- 695940
J, Borowitz MJ, Le Beau MM, et al. PMID:28028029
(2016). The 2016 revision to the World 16. Morton LM, Slager SL, Cerhan JR,
Health Organization classification of Wang SS, Vajdic CM, Skibola CF, et al. 24. Dombret H, Gardin C (2016). An update of
myeloid neoplasms and acute leuke- (2014). Etiologic heterogeneity among current treatments for adult acute myeloid
mia. Blood. 127(20):2391–405. https:// non-Hodgkin lymphoma subtypes: the leukemia. Blood. 127(1):53–61. https://
doi.org/10.1182/blood-2016-03-643544 InterLymph Non-Hodgkin Lymphoma doi.org/10.1182/blood-2015-08-604520
PMID:27069254 Subtypes Project. J Natl Cancer Inst PMID:26660429
Monogr. 2014(48):130–44. https://doi.
8. Swerdlow SH, Campo E, Pileri SA, o r g /10.10 9 3 / j n c i m o n o g r a p h s / l g u 013
Harris NL, Stein H, Siebert R, et PMID:25174034
al. (2016). The 2016 revision of the
World Health Organization classifica-
tion of lymphoid neoplasms. Blood.
127(20):2375–90. https://doi.org/10.1182/
blood-2016-01-643569 PMID:26980727
484
WHO Report on Cancer: Setting priorities,

investing wisely and providing care for all
André M. Ilbawi
Background and rationale intended to shape the global agenda registries and other information
Over the past two decades, there and highlight priority actions. systems; and
has been rapid progress in the un- • draw attention to cancer research
Aim and scope to better understand the causes
derstanding of cancer prevention
and treatment. Cancer now features The WHO Report on Cancer: Setting of cancer, to evaluate interven-
in global development targets, in- priorities, investing wisely and pro- tions, and to formulate a research
cluding the United Nations 2030 viding care for all provides evidence- agenda to develop new policies
based public health- and policy- and programmes.
Sustainable Development Goals,
oriented guidance on cancer, based The primary target audience for
and is a critical element of universal
on the latest available evidence and the WHO Report on Cancer is policy-
health coverage.
international experience. The report makers. The report is also intended
However, the reality for cancer
catalyses global collaboration and for a broad multisectoral audience,
patients suggests that progress
provides guidance on next steps to including nongovernmental organi-
has been inadequate and inequita-
improve cancer control in countries. zations, philanthropic foundations,
ble, particularly in low- and middle-
The aim of the WHO Report on academic institutions, and private
income countries. At the current
Cancer is to set the agenda for ac- sector entities. It is global in its reach,
rate, global targets to reduce prema-
celerated action on evidence-based, providing clear guidance for policy-
ture mortality will not be achieved.
comprehensive cancer control pro- makers in all settings.
Exacerbating the problem, the num-
ber of new cancer cases is projected grammes and to raise awareness Link to IARC World Cancer
to double over the next two or three about cancer as a preventable and Report
decades. The greatest impact of controllable public health priority
The WHO Report on Cancer is a
cancer and the fastest increase in globally.
complement to the existing IARC
the cancer burden will be in low- and The scope of the WHO Report
World Cancer Reports, which pro-
middle-income countries, many of on Cancer is to:
vide extensive details and scientific
which are ill-equipped to cope with • present the cancer burden and
background on cancer patterns and
the current burden. trends, and the social and eco-
causes and tested preventive in-
The time is now to set the can- nomic impact of the disease;
terventions. This new IARC World
cer policy agenda promoting health • inform policy-makers about the Cancer Report comprehensively
for all, consistent with universal need to prioritize investment in presents the most up-to-date sci-
health coverage and the Sustainable cancer, and provide recommenda- ence in cancer prevention.
Development Goals. Implementation tions on the way forward; The WHO Report on Cancer
of evidence-based cancer policies • describe effective public health translates this structured evidence
will shift the trajectory and save mil- strategies to mitigate common risk and other scientific findings into ac-
lions of lives each year. Governments factors for cancer; tionable policies and programmes.
expressed their commitment to ac- • provide the most up-to-date evi- This has been achieved by promot-
celerate action through the 2017 dence on effective cancer control ing clear linkages between the two
World Health Assembly resolution programmes for all resource levels, documents and integration of con-
on cancer prevention and control with a focus on access and equity; tent. The WHO Report on Cancer
(WHA70.12). As part of resolution • facilitate evidence-based decision- summarizes the current state of the
WHA70.12, governments specifically making by policy-makers in select- science to advance understanding
requested WHO, in collaboration ing a basic cancer control package of how the science of cancer informs
with IARC, to produce a global report relevant for their national context. policy. In effect, the new IARC World
on cancer, a landmark document • highlight the importance of cancer Cancer Report and the WHO Report
WHO Report on Cancer: Setting priorities, investing wisely and providing care for all 485
on Cancer have complementary underwent regional consultations of the WHO Report on Cancer will
roles, respectively summarizing the to ensure that it presents the per- be measured by its impact in shift-
evidence and promoting evidence- spectives of stakeholders around ing the global dialogue, supporting
based policies, based on the highest the world and summarizes the best the formulation and implementation
quality science. global understanding of cancer pol- of effective cancer policies, and
icies. The work does not stop with changing the trajectory of cancer
Next steps the release of the report; there will for communities around the world.
Before global release of the WHO be spin-off products and broad dis-
Report on Cancer, the report semination strategies. The success
486
6 The basis for,

and outcomes from,
prevention strategies
The burden of death from the multiple different Vaccination is effective for some cancers
cancer types can be decreased in all commu- caused by infectious agents. Deaths from
nities and countries. Cancer incidence can be sporadic cancer may be decreased through
reduced by decreasing or eliminating exposure chemoprevention and diagnosis of early-stage
to carcinogens in multiple contexts. Success disease by screening and emerging molecular
in reducing the incidence of smoking-related methods of early diagnosis. An increased risk
cancers in some countries indicates a range of cancer may be indicated by family history
of measures that may be researched for their and can be addressed by monitoring the affect-
efficacy in other situations. Interventions to ed individuals. The extent to which the options
change behaviour related to nutrition, exercise, summarized here are realized across national
and weight gain are being actively researched. boundaries warrants continuing research.
Tobacco cessation: the WHO perspective

Cessation support can more than double
the chance of successfully quitting
Tobacco use, particularly cigarette than 60% of smokers indicated that

smoking, remains the leading pre- they intend to quit, and more than FUNDAMENTALS
ventable cause of death from can- 40% had attempted to quit in the
cer and other conditions worldwide. 12 months preceding the survey ■■ The success of tobacco
In 2017, about 8 million people died (Fig. P1.1). Tobacco cessation sup- control policies has increased
from a tobacco-related disease [1,2]. port services complement coun- the demand for support to quit
The global costs of smoking are tries’ tobacco control measures tobacco use.
equivalent to 18% of what countries and can contribute to reducing the
■■ Without cessation assistance,
spend on health care [3]. prevalence of tobacco use.
only 4% of attempts to quit
Globally, there are 1.1 billion Nicotine, a pharmacologically
tobacco succeed.
adult smokers and at least 303 mil- active drug that occurs naturally
lion users of smokeless tobacco in the tobacco plant, is highly ad- ■■ Proven cessation medications
[4], many of whom say they want, dictive and is delivered rapidly to and professional support can
or intend, to quit [5,6]. Although this the brain after the inhalation or in- double a tobacco user’s chance
is encouraging, the availability of to- gestion of tobacco products or the of successfully quitting.
bacco cessation support worldwide use of non-tobacco products that
contain nicotine [7]. Nicotine is so ■■ Several different approaches
remains low, and many people do
addictive that the autonomy over have been developed to help
not have adequate cessation sup-
smoking of one quarter of adoles- people stop using tobacco.
port available to them. Currently,
cents starts to diminish after smok- These vary in terms of intensi-
only about 30% of the world’s popu-
ing just three or four cigarettes, and ty, cost, and effectiveness, and
lation have access to appropriate
after smoking five packs (i.e. 100 can broadly be categorized as
tobacco cessation services [6].
cigarettes), nearly 60% are depen- behavioural or pharmacologi-
Over the past decade, coun-
dent [8]. Most people who use to- cal interventions.
tries have made substantial prog-
bacco regularly do so because they ■■ Tobacco cessation support
ress in establishing evidence-based
are addicted to nicotine, and they should be made readily
and cost-effective tobacco control
can therefore benefit greatly from
measures. In numerous countries, accessible in order to have a
a range of effective tobacco ces-
many indoor public spaces are now greater impact on reducing the
sation interventions. It is estimated
smoke-free, warnings about the dan- prevalence of tobacco use.
that the highest-level cessation pol-
gers of tobacco use appear on pack-
icies, adopted in 14 countries from
aging and in mass media messages,
2007 to 2014, will result in about
higher tobacco product prices and
1.5 million fewer future tobacco-
taxes have reduced the affordability stroke returns to that of a never-
related deaths up to 2030 [9].
of tobacco products, and tobacco smoker within 5–15 years. The risk
product advertising, promotion, and of death due to lung cancer is re-
sponsorship have been prohibited. The health benefits of duced by 30–50% within 10 years
All of these efforts have con- quitting tobacco of quitting smoking [10].
tributed to reduced demand for to- The risk of death due to tobacco People who quit tobacco can
bacco products and have increased use begins to decrease soon after live longer, healthier, and more pro-
existing tobacco users’ intention to quitting. Current evidence suggests ductive lives. Quitting smoking at
quit. On average, across countries that the risk of death due to ischae- any time in life is likely to extend
where the Global Adult Tobacco mic heart disease is halved within life expectancy; for example, quit-
Survey has been conducted, more 5 years of quitting, and the risk of ting at age 30 years can add up to
488
Fig. P1.1. Proportion of current smokers who intend to quit (countries with Global Adult Tobacco Survey data [4], various years).
Proportions include those who indicated they were thinking of quitting in the next month, within the next 12 months, or sometime in
the future.
Botswana 2017
Senegal 2015
Mexico 2015
Kenya 2014
Argentina 2012
Costa Rica 2015
Uruguay 2017
Malaysia 2011
Ukraine 2017
Ethiopia 2016
Qatar 2013
Nigeria 2012
Cameroon 2013
Bangladesh 2017
Panama 2013
Kazakhstan 2014
Uganda 2013
Romania 2011
Philippines 2015
Russian Federation 2016
Turkey 2012
Thailand 2011
Viet Nam 2015
Greece 2013
Brazil 2008
Poland 2010
India 2017
Indonesia 2011
Egypt 2009
China 2010
Pakistan 2014
0 10 20 30 40 50 60 70 80 90
10 years of life expectancy. Even at increasing the years of economically lar diseases, cancer, chronic lung
age 50 years, quitting results in an and socially productive life. diseases, and diabetes – and ac-
average of 6 years of life expectancy celerate action against the leading
gained [11]. Hence, it is never too risk factors for NCDs. The agreed
Policy actions
late to gain the health benefits of target for tobacco control is a 30%
quitting tobacco use. The life years
recommended by WHO relative reduction in the prevalence
gained can also be expected to be Following the Political Declaration of current (daily and occasional) to-
lived in better health, because the on the prevention and control of bacco use in people aged 15 years
diseases caused by tobacco use noncommunicable diseases (NCDs) and older between 2010 and 2025,
are commonly chronic and debilitat- adopted by the United Nations which was endorsed by the World
ing and lead to years of diminished General Assembly in 2011, WHO Health Assembly in May 2013. To
PERSPECTIVE
SECTION 6
quality of life. Therefore, quitting can developed nine voluntary global tar- achieve this target, it is essential
reduce the health-care costs associ- gets to reduce global mortality from not only to prevent the uptake of to-
ated with long-term illness while also the four main NCDs – cardiovascu- bacco use but also to ensure that
Tobacco cessation: the WHO perspective 489

more tobacco users quit. Several Despite these commitments, approaches have been developed
highly effective and inexpensive in- progress towards best-practice ces- to help people stop using tobacco
terventions exist to help make this sation support in countries is slow (Table P1.1). These vary in terms of
happen, as summarized below. compared with progress on other intensity, cost, and effectiveness,
The importance of helping cur- WHO-recommended policy mea- and can broadly be categorized
rent tobacco users quit is reflected sures, such as smoke-free places as behavioural or pharmacological
in the WHO Global Action Plan for and bans on tobacco advertising, interventions.
the Prevention and Control of NCDs promotion, and sponsorship.
2013–2020 [12]. The Global Action Behavioural interventions
Plan lists a menu of “best buys” Although behavioural interventions
and cost-effective policy options for
Effective cessation
for tobacco cessation are generally
countries to address the NCD bur-
interventions are low-cost, they can be very effective.
den. These include the recommen- available Brief advice from health profession-
dation that countries should provide A wide choice of behavioural and als as part of their routine consulta-
cost-covered, effective, and popula- pharmacological tobacco cessation tions or interactions is an approach
tion-wide cessation support, includ- interventions is available. Without that makes use of existing health-
ing brief advice, national toll-free cessation assistance, only 4% of at- care systems. When a tobacco user
quitline services, and mCessation tempts to quit tobacco succeed [13]. visits a primary or specialized care
(a mobile phone-based intervention Proven cessation medications and service, this presents an opportuni-
providing text messages supporting professional support can double a ty for the health-care worker to offer
individual efforts to stop smoking), to tobacco user’s chance of success- and provide them with personalized
all those who want to quit [12]. fully quitting [14]. Several different counselling. Brief advice is a key
Table P1.1. Types of tobacco cessation interventions
Behavioural Population-level Brief advice Advice to stop using tobacco, usually taking only a few minutes,
interventions approaches is given to all tobacco users during the course of a routine
consultation and/or interaction with a physician or health-care
worker.
Quitlines A national toll-free quitline is a telephone counselling service that

can provide both proactive and reactive counselling. A reactive
quitline provides an immediate response to a call initiated by the
tobacco user, but only responds to incoming calls. A proactive
quitline involves setting up a schedule of follow-up calls to tobacco
users to provide ongoing support.
mCessation Tobacco cessation interventions are delivered via mobile phone

text messaging. Mobile technologies provide the opportunity to
expand access to a wider population, and text messaging can
provide personalized tobacco cessation support in an efficient and
cost-effective manner.
Individual specialist Intensive behavioural Behaviour support refers to multiple sessions of individual or
approaches support group counselling aimed at helping people stop their tobacco use.
It includes all cessation assistance that imparts knowledge about
tobacco use and quitting, and provides support and resources to
develop skills and strategies for changing behaviour.
Cessation clinics In many countries, clinics specializing in tobacco cessation

services are available. These clinics offer intensive behavioural
support and, where appropriate, medications or advice on
the provision of medications, delivered by specially trained
practitioners.
Pharmacological Nicotine replacement therapies (NRTs) NRTs are available in several forms, including gum, lozenges,
interventions patches, inhalers, and nasal spray. These cessation tools reduce
cravings and withdrawal symptoms by providing a low, controlled
dose of nicotine without the toxins found in cigarettes. The doses
of NRT are gradually reduced over time to help the tobacco user
wean off nicotine by getting used to less and less stimulation.
Non-nicotine pharmacotherapies These include medications such as bupropion, varenicline,

and cytisine. These pharmacotherapies reduce cravings and
withdrawal symptoms and decrease the pleasurable effects of
cigarettes and other tobacco products.
490
means of motivating people who more effective and can double the Therefore, every country can use its
might not otherwise seek tobacco chances of successfully quitting [16]. existing systems and resources to
cessation support and encouraging ensure that tobacco users at least
them to quit, and thus it is an essen- Mechanisms for receive brief advice (Fig. P1.3).
tial component of tobacco cessation Incorporating brief advice into
services. Countries can easily train
developing tobacco existing health-care programmes
physicians and health-care workers
cessation support has the potential to reach more than
to provide brief advice effectively to Implementing tobacco 80% of all tobacco users in a coun-
the population they serve. cessation measures try each year if delivered routinely
Toll-free quitlines are a conve- alongside other tobacco and widely across a health-care
nient way for tobacco users who are control policies maximizes system [19]. Tobacco cessation in-
ready to quit to access brief and po- their impact terventions should be integrated into
tentially intensive behavioural coun- any existing health programmes in
selling. People who use quitlines Tobacco cessation support has primary care where feasible, as well
increase their absolute quit rate by 4 optimal effect when implemented as disease- and population-specific
percentage points, which represents in conjunction with other demand- programmes such as national tuber-
a doubling of success in quitting reduction tobacco control policies, culosis programmes [20], NCD pro-
compared with those who attempt such as raising tobacco taxes, grammes, oral health programmes
to quit without assistance [14]. This establishing smoke-free environ- [21], HIV/AIDS programmes, men-
rate can be further increased if the ments, banning tobacco advertis- tal health programmes, and pro-
quitline is proactive and counsellors ing, promotion, and sponsorship, grammes addressing the needs of
make follow-up calls to potential to- printing large pictorial health warn- women’s, children’s, and adoles-
bacco quitters. ing labels on tobacco packages, and cents’ health. In particular, there
With the advent and spread of delivering anti-tobacco mass media has been a major drive globally to
mobile phone technologies, people campaigns. In turn, these tobacco integrate cessation services into
who want to quit can now be ac- control measures promote tobacco tuberculosis programmes and into
cessed not only through telephone cessation by encouraging quitting sexual and reproductive health
calls but also via text messages. Text and creating a supportive environ- programmes. Both of these pro-
message interventions can increase ment. A good example of synergiz- grammes reach populations at
the absolute quit rate by 4% [15]. ing efforts is to include the local particular risk from the harms of to-
mCessation register portal/number bacco and present an opportunity to
Pharmacological or quitline number on cigarette and address tobacco dependence when
interventions tobacco packs and in mass media people make (potentially rare) con-
Pharmacotherapy cessation inter- anti-tobacco campaigns; this can tact with the health system.
ventions include nicotine replace- significantly increase the demand Countries should also consider
ment therapies (NRTs) as well as for tobacco cessation services [17]. leveraging existing infrastructure
medications that do not contain to provide wide-reaching intensive
nicotine but act to alleviate tobacco
Using existing infrastructure behavioural support for tobacco us-
withdrawal symptoms. Both forms
to develop cessation support ers. Many countries have existing
of therapy are effective aids to help is feasible and affordable call centres and substance abuse
people to quit tobacco use. The Integrating brief advice into exist- or other health-related hotlines that
efficacy of pharmacotherapies is ing primary health-care systems is can be expanded to provide tobac-
generally high, and compared with one of the first actions that countries co quitline services.
people who do not use an interven- can take to develop tobacco cessa-
tion, increases in the absolute quit tion support. Guidelines for imple- Provide comprehensive
rate can range from 6% for a sin- mentation of Article 14 of the WHO tobacco cessation support
gle type of NRT to almost 15% for Framework Convention on Tobacco and treatment when
varenicline [16]. Combining more Control recommend that countries resources allow
than one type of NRT (patches and adopt a stepwise approach to de- The cost and effectiveness of differ-
a faster-acting form) can also in- velop and strengthen national to- ent cessation approaches vary, and
crease the effectiveness of NRTs bacco cessation systems as rapidly therefore the affordability of the dif-
(see “Combined NRT” in Fig. P1.2). and cost-effectively as possible [18]. ferent approaches varies across low-,
Both behavioural cessation sup- Much of the needed infrastructure middle-, and high-income countries.
port and pharmacotherapies are for promoting tobacco cessation Overall, almost all population-level
effective in helping people to quit measures, such as a primary health- behavioural interventions are global-
PERSPECTIVE
SECTION 6
tobacco use (Fig. P1.2). However, care system, already exists in most ly affordable, whereas intensive face-
combining both behavioural and countries, making such promotion to-face therapy is affordable for mid-
pharmacological interventions is not only feasible but also affordable. dle- and high-income countries [16].

Fig. P1.2. Increased proportion of people who abstain from smoking for 6 months or more due to a specific intervention. Each
bar represents the findings of a meta-analysis, and the strength of evidence associated with each study will vary. The vertical axis
represents the projected percentage point increase in 6–12-month abstinence compared with no intervention. The authors adjusted
the published percentage point increase in 6–12-month abstinence to allow for direct comparison between each intervention where
the meta-analyses did not use a comparator equivalent to “no intervention”. Assessments were based on the published effectiveness
of the comparison intervention through a consensus [16]. NRT, nicotine replacement therapy.
Behavioural interventions
20
18
Absolute increase in proportion of people
who quit smoking (percentage points)
16
14
12
10
Print-based Brief advice from Automated Proactive Face-to-face

self-help physician text messaging telephone behavioural
support
Pharmacological interventions
20
Absolute increase in proportion of people
who quit smoking (percentage points)
18
16
14
12
10
Single NRT Cytisine Buproprion Notriptyline Combined Varenicline

NRT
If resources allow, countries should effective way to quit, but uptake of (individual, group, or telephone),
provide tobacco users with the high- interventions also relies on people’s pharmacological therapy, and other
est level of support to facilitate a suc- preferences, which is likely to vary cessation services via text messag-
cessful quit attempt. Countries may across different social and cultural ing or online tools [22,23]. Providing
follow a stepwise approach to devel- contexts. Tobacco users may pre- a diverse range of tobacco cessa-
op their tobacco cessation support fer using multiple tobacco cessation support options, as often as
systems (Fig. P1.3). tion interventions, including health possible, is also important to ensure
Combining behavioural and phar- education materials, advice from maximal uptake and effectiveness
macological interventions is the most health professionals, counselling (Table P1.2).
492
Fig. P1.3. Stepwise approach to developing and strengthening national tobacco cessation systems.
Increase the likelihood of quit

attempts succeeding
STEP 3
Medications
Specialized treatment
STEP 2
Establish free, proactive quitline
and/or mCessation service
STEP 1
Prompt quit
Establish system components attempts
Address any issues related to health-care workers
Integrate brief advice into existing health systems
E-cigarettes and other create an aerosol that is inhaled These aerosols are inhaled by
products promoted as by the user. The liquid contains users during a process of suck-
nicotine (but not tobacco) and ing or smoking involving a device.
“cessation aids”
other chemicals that may be toxic They contain nicotine and non-
In recent years the tobacco industry to people’s health. tobacco additives, and are often
(and other non-tobacco commercial • Electronic non-nicotine delivery flavoured. The tobacco may be
actors, such as those manufactur- systems (ENNDS) are similar to in the form of specially designed
ing e-cigarettes) has introduced ENDS, but the heated solution cigarettes (e.g. “heat sticks”, “Neo
a wide array of products, the ma- delivered as an aerosol through sticks”) or pods or plugs.
jority of which simulate the act of the device does not generally These products are aggressive-
smoking while typically delivering contain nicotine. ly marketed or promoted as “clean-
nicotine. There are currently three • Heated tobacco products (HTPs) er” alternatives to conventional cig-
broad categories of these products: are tobacco products that pro- arettes, as smoking cessation aids,
• Electronic nicotine delivery sys- duce aerosols containing nicotine or as “reduced risk” products (see
tems (ENDS), which are some- and toxic chemicals upon heating Chapter 2.1). They have proliferated
times referred to as e-cigarettes, of the tobacco or activation of a in several markets around the globe
are devices that heat a liquid to device containing the tobacco. and present a unique challenge to
Table P1.2. Examples of minimal, expanded, and advanced cessation interventionsa
Minimal Expanded Advanced

Brief advice integrated into primary care Brief advice integrated into primary care Brief advice integrated into primary care,
services and hospital services hospital, and specialized services
Quitline: toll-free quitline provided Quitline: toll-free quitline provided
mCessation: text messaging mCessation: text messaging
Specialized tobacco dependence
treatment services: behavioural
PERSPECTIVE
SECTION 6
counselling and/or medication

a
All countries should implement, at a minimum, brief advice. Once this is well established, countries can apply expanded and advanced measures, subject
to the availability of resources.

regulators. Although some of these harmful relative to conventional to- conclusions could be drawn from the
products have lower emissions bacco products [24,25]. available studies [28,30]. This is con-
than conventional cigarettes, they There remains a great deal of sistent with the conclusion of the 2018
are not risk-free, and the long-term uncertainty surrounding the risks review by the National Academies of
impact on health and mortality is associated with ENDS (Table P1.4). Sciences, Engineering, and Medicine
still unknown. There is insufficient Although some have been shown of the evidence on ENDS (referred to
independent evidence to support to help smokers quit conventional as e-cigarettes in this and the subse-
the use of these products as a pop- smoking under certain conditions quent reports): “Overall, there is lim-
[26,27], the evidence is inconclusive ited evidence that e-cigarettes may
ulation-level tobacco cessation in-
[28–30]. There have been only a be effective aids to promote smoking
tervention to help people quit use of
limited number of randomized con- cessation” [31].
conventional tobacco (Table P1.3).
trolled trials and longitudinal studies In contrast, a randomized con-
HTPs contain tobacco, and the use investigating the role of ENDS as trolled trial of e-cigarettes versus
of these products constitutes to- potential cessation aids offered to NRT concluded that e-cigarettes
bacco use, thereby contributing to a population, and their conclusions were more effective for smoking
the burden of tobacco in countries are equivocal [28,30]. cessation than NRT when both
where they are sold. In addition, Two systematic reviews – which products were accompanied by
some studies do not support the were published in 2016 and 2017, behavioural support, based on a
claims that these products are less respectively – established that no 1-year abstinence rate of 18.0%
Table P1.3. Questions and summaries of the evidence for heated tobacco products (HTPs)
Question Summary of the evidence
Do HTPs contain harmful chemicals? From available evidence, we know that many of the harmful chemicals that are generated by
HTPs are similar to those generated by conventional cigarettes, but generally at lower levels
[46,47]. However, there is also some evidence that there are new chemicals in HTPs that are not
present in the emissions of conventional cigarettes, and that could have some degree of toxicity
and associated harm [24].
Are HTPs less harmful than cigarettes? To date, the available evidence demonstrates that exposure to harmful and potentially harmful
chemicals from these products may be lower relative to cigarettes [48] (but higher compared
with electronic nicotine delivery systems [ENDS]). However, the evidence does not show that
these products will reduce tobacco-related diseases, or that they are exclusively used as
substitutes for cigarettes. If they attract users who were not previously tobacco users, their
overall impact on health would be negative.
Are HTPs useful as a cessation aid? HTPs are tobacco products and, therefore, even if a tobacco user converts from the use of
conventional cigarettes to HTPs, this would not constitute cessation. Claims that smokers
switch from conventional cigarettes to exclusive use of HTPs are unsubstantiated [49]. Further
independent studies are needed to gather more information and inform policy options.
Table P1.4. Questions and summaries of the evidence for electronic nicotine delivery systems (ENDS)
Question Summary of the evidence
What are the consequences of taking Recent surveys in the USA and some European countries have shown marked increases in
up ENDS use at a younger age? ENDS use among young people [50]. Between 2011 and 2018 in the USA, rates of e-cigarette
use in young people increased from 1.5% to a staggering 20.8% [44]. Young people who use
ENDS are exposed to nicotine, which can have long-term effects on the developing brain, and
there is a risk of nicotine addiction, given that tobacco product use is primarily established in
adolescence [37]. Furthermore, there is a growing body of evidence in some settings that never-
smoker minors who use ENDS at least double their chance of starting to smoke cigarettes later
in life [51,52].
What is the harm of ENDS relative to ENDS’ aerosols are likely to be less toxic than cigarettes, but there is insufficient evidence
conventional cigarettes? to quantify the precise level of risk associated with them [39]. Also, many factors will have an
impact on the relative risk associated with their use, for example the amount of nicotine and
other toxicants in the heated liquid.
What are the health effects associated ENDS pose risks to users and non-users [39]. There is insufficient evidence to quantify this risk,
with ENDS? and the long-term effects of exposure to ENDS’ toxic emissions are unknown [39,50]. In addition
to risks associated with emissions of ENDS, there are also risks of physical injury brought about
by fires or explosions related to ENDS devices [53].
Do ENDS help smokers quit tobacco? The effectiveness of ENDS as a smoking cessation aid is still being debated. To date, in part
due to the diversity of ENDS products and the low certainty surrounding many studies, the
potential for ENDS to play a role as a population-level tobacco cessation intervention is unclear
[28–30].
494
in the e-cigarette group compared WHO does not endorse ENDS tobacco users to quit, but currently
with 9.9% in the NRT group [32]. as cessation aids only about 30% of the world’s popu-
However, the study has several The scientific evidence on e-ciga- lation have access to comprehen-
limitations. For example, although rettes as cessation aids is inconclusive tobacco cessation services.
people who were assigned to the e- sive, and there is a lack of clarity as Countries – in particular low- and
cigarette group were more likely to to whether these products have any middle-income countries, where
abstain from using traditional ciga- role to play in smoking cessation. the majority of tobacco users in
rettes compared with those who There are also real concerns about the world live – should implement
were assigned to the NRT group, the risk they pose to nonsmokers these proven tobacco cessation
80% of people in the e-cigarette who start to use them, especially measures, alongside other tobacco
group continued to use e-cigarettes young people. Unlike for the tried control policies, to maximize their
1 year after the study started, and tested nicotine and non-nic- impact on reducing the prevalence
whereas only 9% of those in the otine pharmacotherapies that are of tobacco use and the risk of death
NRT group continued to use NRTs known to help people quit tobacco from all tobacco-related diseases,
at 1 year. In most countries where use, WHO does not endorse e-cig- including cancer.
e-cigarettes are available, the ma- arettes as cessation aids. Resources are finite. In order
jority of users of e-cigarettes con- As ENDS are increasingly intro- for tobacco cessation interventions
tinue to use e-cigarettes and con- duced to the market, careful moni- to reach as many tobacco users
ventional cigarettes concurrently, toring of cessation rates is vital. The as possible at the lowest achiev-
which has little or no beneficial im- possibility of tobacco industry inter- able cost and have the most im-
pact on health risk and effects [33]. ference in tobacco cessation efforts pact, governments should prioritize
Some reviews have also sug- through misinformation about the population-wide tobacco cessation
gested that use of e-cigarettes potential benefits of these products approaches as recommended by
could in fact hinder smoking ces- – which are presented as alterna- the WHO Global Action Plan for the
sation [34]. Furthermore, beyond tives but in most cases are comple- Prevention and Control of NCDs
the scope of cessation, novel and mentary to the use of conventional 2013–2020: integrating brief advice
emerging tobacco and nicotine tobacco products – is a present and into primary care, providing national
products are increasingly being real threat (Box P1.1 and Box P1.2). toll-free quitline services, and mak-
taken up by never-users of tobacco ing mCessation support available.
[35]. These products therefore play If resources allow, countries should
an important role in expanding the Conclusions also provide tobacco users with
market of nicotine users, with a high A wide range of proven behavioural combined behavioural and pharma-
associated risk of addiction, particu- and pharmacological cessation in- cological interventions to facilitate a
larly in children and adolescents. terventions can be used to support successful quit attempt.
Box P1.1. Excerpt on electronic nicotine delivery systems (ENDS) from the WHO Director-General’s Commentary in The Lancet [54].
Much has been written and said about the potential of electronic nicotine delivery systems (ENDS) such as
e-cigarettes to help tobacco users quit [31,36–38]. Although tobacco and related industries promote these
products as tools for quitting, the evidence does not support their use as part of population-based cessation
strategies. The aerosols of ENDS contain toxic chemicals that are harmful to both users and non-users and
are, therefore, products that come with health risks of their own [31,39]. And in combination with smoking,
which is the practice with the majority of ENDS users, the health effects of two or more products are combined
[35]. ENDS on their own are associated with increased risk of cardiovascular diseases [40] and lung disorders
[41] and adverse effects on the development of the fetus during pregnancy [37]. For adolescents, the addictive
nature of nicotine can lead to dependence and may harm adolescent brain development, including reduced ac-
tivity in the prefrontal cortex [42,43]. Use of ENDS could also lead to a new generation of nicotine and tobacco
users, as seen in some countries [44], especially given how these products are marketed to young people [37].
Although the specific level of risk associated with ENDS has not yet been conclusively estimated, ENDS are
undoubtedly harmful, should be strictly regulated, and, most importantly, must be kept away from children. It is
also incorrect to think that heated tobacco products are the answer, as they simply move tobacco users from
one harmful tobacco product to another.
To truly help tobacco users quit and to strengthen global tobacco control, governments need to scale up
PERSPECTIVE
SECTION 6
policies and interventions that we know work. Tried and tested interventions, such as nicotine and non-nicotine
pharmacotherapies, should be promoted for cessation.

Box P1.2. Summary on electronic nicotine delivery systems (ENDS) and electronic non-nicotine delivery systems (ENNDS) and
cessation, from the seventh report of the WHO Study Group on Tobacco Product Regulation [45].
ENDS are a heterogeneous class of products, with various profiles of nicotine and non-nicotine toxicants,
which depend on factors including their construction, power, liquid constituents, nicotine concentration, and
user behaviour. The amount of nicotine delivered can range from none to doses that exceed those delivered
by tobacco cigarettes in the same number of puffs. Nicotine from ENDS reaches users’ blood faster than from
most types of nicotine replacement therapy (NRT), and, at least with some ENDS, at higher concentrations.
ENDS could be effective in cessation for some smokers under some circumstances, while, for other smokers,
in different circumstances, it might have the opposite effect. Whether an ENDS has beneficial or detrimental
effects on smoking cessation appears to depend on the technology, the motivation and consumer behaviour
of the ENDS user, the type of smoker who seeks ENDS use, and the regulatory environment for ENDS and
tobacco use.
Translating the evidence into a potential role of ENDS and ENNDS in smoking cessation is difficult. The
evidence does not allow a blanket policy recommendation for or against general use of ENDS and ENNDS as
cessation aids.
References
1. IHME (2018). Findings from the Global 7. CDC (1994). Preventing tobacco use among 13. Cohen S, Lichtenstein E, Prochaska JO,
Burden of Disease Study 2017. Seattle young people: a report of the Surgeon Rossi JS, Gritz ER, Carr CR, et al. (1989).
(WA), USA: Institute for Health Metrics and General. Atlanta (GA), USA: Centers for Debunking myths about self-quitting.
Evaluation. Available from: http://www. Disease Control and Prevention. Available Evidence from 10 prospective studies of
healthdata.org/policy-repor t /findings- from: https://www.cdc.gov/mmwr/PDF/rr/ persons who attempt to quit smoking by
global-burden-disease-study-2017. themselves. Am Psychol. 44(11):1355–65.
rr4304.pdf.
https://doi.org/10.1037/0003 - 066X.44.
2. WHO (2019). WHO report on the global 11.1355 PMID:2589730
8. Ursprung WW, DiFranza JR (2010).
tobacco epidemic, 2019: offer help to quit
tobacco use. Geneva, Switzerland: World The loss of autonomy over smoking in 14. Stead LF, Hartmann-Boyce J, Perera R,
Health Organization. Licence: CC BY- relation to lifetime cigarette consump- Lancaster T (2013). Telephone counsel-
NC-SA 3.0 IGO. Available from: https:// tion. Addict Behav. 35(1):14–8. https:// ling for smoking cessation. Cochrane
www.who.int/tobacco/global_report/en/. doi.org/10.1016/j.addbeh.2009.08.001 Database Syst Rev. (8):CD002850. https://
PMID:19717241 doi.org/10.1002/14651858.CD002850.
3. WHO (2018). WHO global report on trends in pub3 PMID:23934971
prevalence of tobacco smoking 2000–2025. 9. Levy DT, Yuan Z, Luo Y, Mays D (2018).
2nd ed. Geneva, Switzerland: World Health Seven years of progress in tobacco con- 15. Whittaker R, McRobbie H, Bullen C, Rodgers
Organization. Available from: https://www. trol: an evaluation of the effect of nations A, Gu Y (2016). Mobile phone-based inter-
who.int/tobacco/publications/surveillance/ meeting the highest level MPOWER mea- ventions for smoking cessation. Cochrane
trends-tobacco-smoking-second-edition/ sures between 2007 and 2014. Tob Control. Database Syst Rev. (4):CD006611. https://
en/. doi.org/10.1002/14651858.CD006611.pub4
27(1):50–7. https://doi.org/10.1136/tobacco
PMID:27060875
4. Asma S, Mackay J, Song SY, Zhao L, control-2016-053381 PMID:27956650
Morton J, Palipudi KM, et al. (2015). The 16. West R, Raw M, McNeill A, Stead L,
GATS atlas: Global Adult Tobacco Survey. 10. DHHS (1990). The health benefits of Aveyard P, Bitton J, et al. (2015). Health-
Atlanta (GA), USA: CDC Foundation. smoking cessation: a report of the care interventions to promote and as-
Available from: http://gatsatlas.org/. Surgeon General. Rockville (MD), sist tobacco cessation: a review of ef-
USA: Department of Health and Human ficacy, effectiveness and affordability for
5. GBD 2016 Risk Factors Collaborators Services. Available from: https://profiles. use in national guideline development.
(2017). Global, regional, and national nlm.nih.gov/ps/access/NNBBCT.pdf. Addiction. 110(9):1388–403. https://doi.
comparative risk assessment of 84 be- org/10.1111/add.12998 PMID:26031929
havioural, environmental and occupation- 11. Doll R, Peto R, Boreham J, Sutherland I
al, and metabolic risks or clusters of risks, (2004). Mortality in relation to smoking: 17. Park J, Minh LN, Shin SH, Oh JK, Yun EH,
1990–2016: a systematic analysis for the 50 years’ observations on male British Lee DH, et al. (2019). Influence of new
Global Burden of Disease Study 2016. doctors. BMJ. 328(7455):1519–27. https:// tobacco control policies and campaigns
Lancet. 390(10100):1345–422. https:// on Quitline call volume in Korea. Tob
doi.org/10.1136/bmj.38142.554479.AE
doi.org/10.1016/S0140-6736(17)32366-8 Induc Dis. 17:21. https://doi.org/10.18332/
PMID:15213107
PMID:28919119 tid/104674 PMID:31582932
6. WHO (2017). WHO report on the global to- 12. WHO (2013). Global action plan for the 18. WHO (2010). Guidelines for implementa-
bacco epidemic, 2017: monitoring tobac- prevention and control of noncommu- tion of Article 14 of the WHO Framework
co use and prevention policies. Geneva, nicable diseases 2013–2020. Geneva, Convention on Tobacco Control. Geneva,
Switzerland: World Health Organization. Switzerland: World Health Organization. Switzerland: World Health Organization.
Available from: https://www.who.int/tobac Available from: https://www.who.int/nmh/ Available from: http://www.who.int/fctc/
co/global_report/2017/en/. events/ncd_action_plan/en. guidelines/adopted/article_14/en/.
496
19. WHO (2008). WHO report on the global to- 30. El Dib R, Suzumura EA, Akl EA, Gomaa 39. WHO (2014). Electronic nicotine delivery
bacco epidemic, 2008: the MPOWER pack- H, Agarwal A, Chang Y, et al. (2017). systems. Report by WHO. Conference of the
age. Geneva, Switzerland: World Health Electronic nicotine delivery systems and/ Parties to the WHO Framework Convention
Organization. Available from: https://www. or electronic non-nicotine delivery sys- on Tobacco Control, sixth session, FCTC/
who.int/tobacco/mpower/2008/en/. tems for tobacco smoking cessation or COP/6/10 Rev.1. Available from: https://
reduction: a systematic review and meta- apps.who.int/iris/handle/10665/147110.
20. WHO (2007). A WHO/the Union mono-
graph on TB and tobacco control: join- analysis. BMJ Open. 7(2):e012680. https://
40. Alzahrani T, Pena I, Temesgen N, Glantz
ing efforts to control two related global doi.org/10.1136/bmjopen-2016 - 012680
SA (2018). Association between electron-
epidemics. Geneva, Switzerland: World PMID:28235965 ic cigarette use and myocardial infarction.
Health Organization. Available from:
31. National Academies of Sciences, En gi Am J Prev Med. 55(4):455–61. https://
https://www.who.int/tobacco/resources/
neering, and Medicine (2018). Public health doi.org/10.1016/j.amepre.2018.05.004
publications/tb_tobac_monograph.pdf.
consequences of e-cigarettes. Washing PMID:30166079
21. WHO (2017). WHO monograph on to- ton (DC), USA: National Academies Press.
bacco cessation and oral health inte- 41. Wills TA, Pagano I, Williams RJ, Tam
https://doi.org/10.17226/24952 EK (2019). E-cigarette use and respira-
gration. Geneva, Switzerland: World
Health Organization. Available from: tory disorder in an adult sample. Drug
32. Hajek P, Phillips-Waller A, Przulj D,
https://apps.who.int /iris/bitstream/han Alcohol Depend. 194:363–70. https://doi.
Pesola F, Myers Smith K, Bisal N, et
dle/10665/255692/9789241512671-eng. org/10.1016/j.drugalcdep.2018.10.004
al. (2019). A randomized trial of e-cig-
pdf?sequence=1. PMID:30472577
arettes versus nicotine-replacement
22. Cox LS, Cupertino AP, Tercyak KP (2011). therapy. N Engl J Med. 380(7):629–37. 42. England LJ, Aagaard K, Bloch M, Conway
Interest in participating in smoking cessa- https://doi.org/10.1056/NEJMoa1808779 K, Cosgrove K, Grana R, et al. (2017).
tion treatment among Latino primary care PMID:30699054 Developmental toxicity of nicotine: a trans-
patients. J Clin Psychol Med Settings. disciplinary synthesis and implications for
18(4):392–9. https://doi.org/10.1007/s108 33. Robertson L, Hoek J, Blank ML, Richards emerging tobacco products. Neurosci
80-011-9259-y PMID:21984387 R, Ling P, Popova L (2019). Dual use of Biobehav Rev. 72:176–89. https://doi.
electronic nicotine delivery systems (ENDS) o r g /10.1016 / j.neubi o rev. 2 016 .11.013
23. Ybarra ML, Bağci Bosi AT, Bilir N, Holtrop and smoked tobacco: a qualitative analy- PMID:27890689
JS, Korchmaros J, Emri S (2011). Interest sis. Tob Control. 28(1):13–9. https://doi.
in technology-based and traditional smok- 43. Morean ME, Krishnan-Sarin S, O’Malley
ing cessation programs among adult SS (2018). Assessing nicotine dependence
PMID:29419488
smokers in Ankara, Turkey. Tob Induc in adolescent E-cigarette users: the 4-item
Dis. 9(10):10. https://doi.org/10.1186/1617- 34. Kalkhoran S, Glantz SA (2016). Patient-Reported Outcomes Measurement
9625-9-10 PMID:21806793 E-cigarettes and smoking cessation in re- Information System (PROMIS) Nicotine
24. Glantz SA (2018). Heated tobacco al-world and clinical settings: a systematic Dependence Item Bank for electron-
products: the example of IQOS. Tob review and meta-analysis. Lancet Respir ic cigarettes. Drug Alcohol Depend.
Control. 27(Suppl 1):s1–6. https://doi.org/ Med. 4(2):116–28. https://doi.org/10.1016/ 188:60–3. https://doi.org/10.1016/j.drugalc
10.1136/tobaccocontrol-2018 - 05 4601 S2213-2600(15)00521-4 PMID:26776875 dep.2018.03.029 PMID:29753155
PMID:30352841
35. Wang JB, Olgin JE, Nah G, Vittinghoff 44. Cullen KA, Ambrose BK, Gentzke AS,
25. TPSAC (2018). Tobacco Products Scientific E, Cataldo JK, Pletcher MJ, et al. (2018). Apelberg BJ, Jamal A, King BA (2018).
Advisory Committee. Washington (DC), Cigarette and e-cigarette dual use and Notes from the field: use of electronic
USA: Food and Drug Administration. risk of cardiopulmonary symptoms in cigarettes and any tobacco product
Available from: https://www.fda.gov/advi the Health eHeart Study. PLoS One. among middle and high school students –
sory-committees/committees-and-meet United States, 2011–2018. MMWR Morb
13(7):e0198681. https://doi.org/10.1371/
ing-materials/tobacco-products-scientific- Mortal Wkly Rep. 67(45):1276–7. https://
journal.pone.0198681 PMID:30044773
advisory-committee. d o i . o r g / 10 .15 5 8 5 / m m w r. m m 6 74 5 a 5
26. Rahman MA, Hann N, Wilson A, 36. WHO (2016). Electronic nicotine delivery PMID:30439875
Mnatzaganian G, Worrall-Carter L systems and/or electronic non nicotine
delivery systems for tobacco smoking ces- 45. WHO (2019). WHO Study Group on
(2015). E-cigarettes and smoking ces-
sation or reduction: a systematic review Tobacco Product Regulation. Report
sation: evidence from a systematic re-
view and meta-analysis. PLoS One. and meta-analysis. Geneva, Switzerland: on the scientific basis of tobacco prod-
10(3):e0122544. https://doi.org/10.1371/ World Health Organization. Available from: uct regulation: seventh report of a WHO
journal.pone.0122544 PMID:25822251 study group. Geneva, Switzerland: World
https://w w w.who.int /tobacco/industr y/
Health Organization (WHO Technical
product_regulation/BackgroundPapers
27. Hartmann-Boyce J, McRobbie H, Bullen C, Report Series, No. 1015). Licence: CC
Begh R, Stead LF, Hajek P (2016). Electronic ENDS4_4November.pdf?ua=1.
BY-NC-SA 3.0 IGO. Available from:
cigarettes for smoking cessation. Cochrane 37. WHO (2016). Electronic nicotine de- https://www.who.int/publications-detail/
Database Syst Rev. (9):CD010216. https:// who-study-group-on-tobacco-product-
livery systems and electronic non-
doi.org/10.1002/14651858.CD010216.pub3 regulation-report-on-the-scientific-basis-
nicotine delivery systems (ENDS/
PMID:27622384 of-tobacco-product-regulation-seventh-
ENNDS). Report by WHO. Conference
28. Malas M, van der Temple J, Schwartz R, of the Parties to the WHO Framework report-of-a-who-study-group.
Minichiello A, Lightfoot C, Noormohamed Convention on Tobacco Control, sev- 46. Simonavicius E, McNeill A, Shahab L,
A, et al. (2016). Electronic cigarettes for enth session, FCTC/COP/7/11. Available Brose LS (2019). Heat-not-burn tobacco
smoking cessation: a systematic review. from: https://www.who.int/fctc/cop/cop7/ products: a systematic literature review.
Nicotine Tob Res. 18(10):1926–36. https:// FCTC_COP_7_11_EN.pdf.
doi.org/10.1093/ntr/ntw119 PMID:27113014 Tob Control. 28(5):582–94. https://doi.
38. McNeill A, Brose LS, Calder R, Bauld L,
29. Khoudigian S, Devji T, Lytvyn L, Campbell PMID:30181382
Robson D (2019). Vaping in England, an
K, Hopkins R, O’Reilly D (2016). The ef-
ficacy and short-term effects of electronic evidence update, February 2019. A report 47. Auer R, Concha-Lozano N, Jacot-
cigarettes as a method for smoking ces- commissioned by Public Health England. Sadowski I, Cornuz J, Berthet A (2017).
PERSPECTIVE
London, UK: Public Health England. Heat-not-burn tobacco cigarettes: smoke

SECTION 6
sation: a systematic review and a meta-

analysis. Int J Public Health. 61(2):257–67. Available from: https://www.gov.uk/govern by any other name. JAMA Intern Med.
https://doi.org/10.1007/s00038-016-0786-z ment/publications/vaping-in-england-an- 177(7):1050–2. https://doi.org/10.1001/
PMID:26825455 evidence-update-february-2019. jamainternmed.2017.1419 PMID:28531246

48. Committee on Toxicity, Carcinogenicity 50. Perikleous EP, Steiropoulos P, Paraskakis 52. Chaffee BW, Watkins SL, Glantz SA
and Mutagenicity of Chemicals in Food, E, Constantinidis TC, Nena E (2018). (2018). Electronic cigarette use and
Consumer Products and the Environment E-cigarette use among adolescents: an progression from experimentation to es-
(2017). Statement on the toxicological overview of the literature and future per- tablished smoking. Pediatrics. 141(4):4.
spectives. Front Public Health. 6:86. https://doi.org/10.1542/peds.2017-3594
evaluation of novel heat-not-burn tobac-
https://doi.org/10.3389/fpubh.2018.00086 PMID:29507167
co products. Assessment made for UK
Department of Health and Public Health PMID:29632856 53. Rossheim ME, Livingston MD, Soule EK,
England. Available from: https://cot.food. Zeraye HA, Thombs DL (2019). Electronic
51. Berry KM, Fetterman JL, Benjamin EJ,
gov.uk/sites/default/files/heat_not_burn_ cigarette explosion and burn injuries, US
Bhatnagar A, Barrington-Trimis JL,
tobacco_statement.pdf. Emergency Departments 2015–2017.
Leventhal AM, et al. (2019). Association Tob Control. 28(4):472–4. https://doi.
of electronic cigarette use with subse- org/10.1136/tobaccocontrol-2018-054518
49. WHO (2017). Heated tobacco prod-
quent initiation of tobacco cigarettes in US PMID:30219795
ucts (HTPs) information sheet. Geneva,
youths. JAMA Netw Open. 2(2):e187794.
Switzerland: World Health Organization. 54. Ghebreyesus TA (2019). Progress in
ht t p s: //d o i.o r g /10.10 01/ j a m a n et wo r k
Available from: https://www.who.int/ beating the tobacco epidemic. Lancet.
open.2018.7794 PMID:30707232
tobacco/publications/prod_regulation/ 394(10198):548–9. https://doi.org/10.1016/
heated-tobacco-products/en/. S0140-6736(19)31730-1 PMID:31371094
498
6.1
Changing behaviour
The need for sustainable
implementation
Graham A. Colditz Steinar Tretli (reviewer)
Sydney E. Philpott-Streiff Giske Ursin (reviewer)
balance could achieve comparable report in Canada in 1974 [3] and the
SUMMARY benefit. Vaccination is effective in Healthy People report in the USA in
preventing some cancers caused 1979 [4], the focus on health equity
●● Prevention strategies over the by infectious agents. Variations in and reducing disparities in disease
past 5 years have made strides the implementation of prevention burden has taken centre stage in
in cancer prevention through the strategies across countries and the the past decade. In 2009, Australia
modification of various causal benefits that extend beyond individ- established a National Preventative
pathways. ual countries deserve further study. Health Taskforce with the intention
●● Two of the most notable suc- of Australia becoming the world’s
cesses in prevention have been Scope of the preventive healthiest country by 2020 [5].
through tobacco control and vac- approach Similarly, the USA expanded the
cination policies. Healthy People goals with targets
There has been a renewed focus on
to reduce disparities by 2020 [6],
●● Despite advances in evidence- the increasing global cancer burden,
and some progress has been re-
based interventions, widespread which rose to an estimated 18.1 mil-
ported [7].
implementation of these preven- lion new cases and 9.6 million deaths
Globally, the most notable suc-
tion strategies varies between in 2018 [1]. Currently, a growing
cesses in prevention have been in
countries. emphasis is on how to increase the
two contrasting domains. The first
availability of evidence-based pre-
●● For effective prevention prac- vention and treatment strategies [2]. is in tobacco control. In 2008, WHO
tices, the cultural context, mea- With respect to primary preven- identified the MPOWER measures
surement strategies, and sus- tion, the nine principles of preven- (Fig. 6.1.1), a set of six cost-effective
tainability for implementation tion associated with effective pro- and high-impact changes that help
must be considered. grammes are still relevant to ensure countries reduce demand for to-
that the approach will be effective. bacco. More than half of the world’s
Interventions must include the fol- countries have implemented at least
The burden of death from the mul- lowing characteristics: they should one MPOWER measure at the high-
tiple different cancer types can be (i) be comprehensive, (ii) be ap- est level of achievement [8].
reduced in all communities and propriately timed, (iii) use varied In addition, the WHO Framework
countries by implementing evi- teaching methods, (iv) have suf- Convention on Tobacco Control,
dence-based prevention and treat- ficient dosage, (v) be administered which is implemented variably by
ment strategies. The incidence of by well-trained staff, (vi) provide country, has led to increases in regu-
cancer can be reduced by decreas- opportunities for positive relation- latory approaches to reduce ciga-
ing or eliminating exposure to car- ships, (vii) be socioculturally rel- rette smoking, resulting in a decline
cinogens in multiple contexts and evant, (viii) be theory-driven, and in lung cancer mortality [9]. However,
by maximizing adherence to a life- (ix) include outcome evaluation. this decline is restricted to high-
style that lowers risk. Success in Population health and preven- income countries, and the preva-
reducing the incidence of smoking- tion strategies have evolved over lence of smoking and the rates of
related cancers is well established the past 50 years, with an increas- lung cancer remain high in low- and
SECTION 6
CHAPTER 6.1
but varies by country. Interventions ing awareness that the social con- middle-income countries [9]. In many
to change behaviour related to nu- text drives exposures and health countries, a broad spectrum of pre-
trition, physical activity, and energy habits. Evolving from the Lalonde vention research is occurring, with a
Chapter 6.1 • Changing behaviour 499

Fig. 6.1.1. The MPOWER measures, es ductions in incidence of liver disease

tablished by WHO to help reduce demand
for tobacco.
and deaths from liver cancer [12]. FUNDAMENTALS
In many high- and middle-income
countries, rates of delivering the ■■ During the past 40 years, rates
HBV vaccine on time are fairly high have decreased for some
(> 80%); in contrast, in many low-in- cancer types, but they have
come countries, rates of adherence
increased for other cancer
to vaccination schedules are lower
types. Cancer remains a lead-
[13]. Variations in childhood HBV
vaccination schedule and population ing cause of death worldwide.
coverage are shown in Fig. 6.1.2. ■■ Behaviour is central to the
These inequalities emphasize
etiology and management
the continued need to strengthen
of cancer prevention and
the infrastructure for immunization
systems, especially in low-income outcomes. This allows for
countries. Such approaches have several avenues for targeted
been successful. For example, the and sustained interventions.
Maldives has sustained an immuni-
■■ Successful preventive
zation programme that trains health
workers across the country on vari- interventions have focused
ous aspects of immunization and on tobacco use, vaccinations,
surveillance. These workers com- nutrition, and physical activity.
bine the work of health profession-
■■ However, successful preven-
als with community engagement
and strong public awareness [14]. tive interventions are not
There is overwhelming evidence effective without sustainable
that HPVs are responsible for di- implementation strategies that
verse preventable cancer types (see are widespread and scalable.
Chapter 2.2), accounting for an esti-
mated 4.5% of all new cancer cases ■■ Sustainable implementa-
worldwide [15]. HPV vaccines have tion requires organizing and
been determined to provide safe maximizing community assets
and durable protection against these and resources, institutionalizing
tumorigenic viruses [16]. Strong evi- policies and practices within
dence of reduced incidence of early communities and organizations,
cervical lesions [17] and follow-up ev- considering the context and
idence of population benefits [18] with
infrastructure of the community,
HPV vaccination led to changes in
cervical cancer screening guidelines and involving a multiplicity of
focus remaining on tobacco control.
In Australia, cigarette taxes have in- [18]. As a result, Australia has moved stakeholders who can develop
creased by 12.5% each year since to vaginal HPV testing every 5 years, long-term buy-in and support.
2016, and there is a plan to continue which consequently saves lives and
the increase for another 2 years [10]. reduces the patient burden and costs
Tobacco taxes are a proven strategy of prevention programmes [19]. in achieving high vaccination rates.
to reduce the prevalence of smok- However, despite compelling ob- Discussing HPV vaccination at ev-
ing, particularly in adolescents and jective evidence of the benefit of HPV ery well-child checkup for children
groups with low socioeconomic sta- vaccines, vaccination rates and poli- starting at about age 9 years as well
tus. In addition, in Canada graphic cies differ markedly by country (see as implementing school-based vac-
warning labels about the harms of Chapter 6.3). Personal reasons for cination programmes could help to
smoking have had a significant im- low vaccination rates include: need- increase acceptance of the vaccine
pact on the prevalence of smoking ing more information, no recom- and increase vaccination rates.
and on quit attempts [11]. mendation by physician, confusion In addition, there are established
The second notable success about the age requirement, and the effective approaches to screening for
in prevention is in vaccination pro- perception that the vaccine will en- the prevention and early detection of
grammes, particularly those for hep- courage sexual promiscuity. It has colorectal cancer, which is the third
atitis B virus (HBV) and for human been shown that physician recom- most commonly diagnosed cancer
papillomavirus (HPV). Uptake of mendation [20] and widespread vac- worldwide. In some countries, there
HBV vaccination has resulted in re- cine availability [21] are major factors have been promising improvements
500
Fig. 6.1.2. Coverage and timing of birth dose of hepatitis B virus (HBV) vaccine for children aged 12–60 months in 13 countries with
national vaccination schedules that include a vaccine dose at birth, 2005–2014. Coverage is the percentage of children receiving
the birth dose of HBV vaccine. Timing of vaccination is the percentage of children receiving the birth dose within 7 days of birth and
within 24 hours of birth.
100
80
Percentage of children
60
40
20
Overall
Albania
Armenia
Azerbaijan
Cambodia
Colombia
Dominican Republic
Kyrgyzstan
Maldives
Nigeria
Peru
Sierra Leone
Tajikistan
Republic of Moldova
Coverage Timing: within 7 days Timing: within 24 hours
in population screening rates after data suggest that the prevalence was successfully implemented on
the establishment of countrywide of childhood obesity has plateaued 1 January 2014, resulted in a 5.5%
colorectal cancer screening pro- in some countries, groups with low reduction in purchases of taxed bev-
grammes. Mathematical modelling socioeconomic status face a dis- erages in 2014 and a 9.7% reduction
studies have shown that screening proportionate impact, including in in 2015 [28].
by colonoscopy is potentially highly populations in South Asia [25]. Broader application of effective
cost-effective at combatting colo- No countrywide programmes prevention strategies to address
rectal cancer in countries in sub- against childhood obesity are cur- these top public health initiatives
Saharan Africa [22]. rently being implemented, but some must move beyond tobacco control
In addition to these notable pre- interventions are showing promise. and singular interventions. Applying
vention strategies, other countries One example is a school-based pro- the principles of implementation
have implemented successful programme evaluated in urban Pakistan science to evidence-based inter-
grammes that are showing progress that demonstrated favourable trends ventions will speed up the transla-
in improving various public health ini- in blood pressure and body mass in- tion of research into practice and
the achievement of the global ben-
tiatives. For example, in Brazil con- dex at follow-up [26]. Cost-effective
efit of a reduced disease burden.
ditional cash transfer programmes school-based programmes have
Implementation science provides a
(which provide low-income families been implemented successfully in
framework to study and identify the
with cash conditional on invest- Australia and in groups with low so-
effective strategies to move from
ments in health and education) have cioeconomic status [27]. Although
research to practice [29].
been shown to increase the odds of these programmes show promise in
children’s visits for preventive ser- the field of physical activity, proper
vices and vaccinations [23]. implementation requires scaling up Background information
Childhood obesity is a global through a transdisciplinary approach. required before
public health problem with conse- An established driver of child- implementation
quences such as premature car- hood and adolescent obesity is con- Defining evidence-based interven-
diovascular disease and premature sumption of sugar-sweetened bever- tions is a necessary first step for
SECTION 6
CHAPTER 6.1
mortality [24]. Adolescent obesity ages (see Chapter 2.6). In Mexico, the implementation of effective pre-
increases the risk of several can- an excise tax of 1 peso per litre on vention strategies. Evidence-based
cer types. Although trends from sugar-sweetened beverages, which health care can provide access to

more and higher-quality information Fig. 6.1.3. A poster about human papillomavirus (HPV) vaccination in the Sinhala
on what works, resulting in a higher language, in Sri Lanka.
likelihood of successful programmes
and policies being implemented,
greater workforce productivity, and
more efficient use of public and pri-
vate resources [30].
Despite the gold standard pro-
vided by timely implementation of
these evidence-based interven-
tions, much less attention has been
focused on how to effectively imple-
ment these practices [30]. A useful
framework ties the evidence-based
strategies to implementation sci-
ence for effective uptake, dissemi-
nation, and scale-up. The context
for the preventive interventions (e.g.
public health or clinical systems,
regulatory strategies, or community-
or group-based interventions, such
as in the workplace, at schools, and
at childcare centres) must be con-
sidered when identifying strategies
for implementation [31].
In addition to the characteristics
of the intervention, the capacity of the
public health infrastructure and the
health delivery system to implement
and sustain a prevention strategy is
fundamental to the success of the in-
tervention. In the setting of tobacco
control, partners of WHO assess
the commitment and organizational
structure for implementing evidence-
based tobacco control programmes.
For other interventions, including
vaccination programmes, the un-
derlying structure of health systems
and the goals of access to universal
health coverage are integral to the
programme’s success. Universal ac- egies. For example, national cam- Australia has led the world with
cess to health care is important for paigns engage public awareness to simple messages about sun protec-
the delivery of the preventive inter- support the changes in culture that tion [34], which have been comple-
vention and also for cancer care and have removed the acceptability of mented by professional education,
outcomes of care [32]. indoor smoke exposure (and indoor mass media messaging, and envi-
smoking). These campaigns are ronmental modifications, resulting in
most effective when the messages population-wide changes in beliefs
Considerations for are reinforced by health-care provid- about sun exposure and prevention
national campaigns ers and by other structural changes, (see Chapter 5.8). As a result, the in-
A common tension of implementing including restricting access to ciga- cidence of and mortality from mela-
prevention strategies is the trade- rettes or putting in place workplace noma have fallen [35].
off of population-wide coverage policies, facilities, and practices.
versus targeting prevention to the However, for national, popu-
groups at highest risk. lation-wide campaigns, the com-
Considerations that limit
Vaccination programmes and ponents of health literacy and cul- wider applicability
taxation on cigarettes demonstrate tural context within a country must Sustainability has been defined as
the value of population-wide strat- be considered [33]. For example, the continued use of components
502
of a programme to achieve its goals Fig. 6.1.4. A poster about healthy nutrition from the Cancer Prevention 4 Africa
and the desired population health campaign, which is designed to improve people’s understanding about the early
outcomes. Often, interventions are signs of cancer and how simple lifestyle changes can greatly reduce the likelihood of
developing many cancer types.
adapted to fit in a new applied con-
text, and the study of this process
shows promise to inform broader
prevention goals. Specifically, inter-
ventions that address children’s nu-
trition, physical activity, and energy
balance can be adapted to diverse
school settings and student popula-
tions to align with local relevance
and account for the norms and cul-
ture within schools [36]. Similarly,
the components of programmes
that lead to sustainability are now
considered within frameworks that
may help to bring prevention to
broader populations [37].
Public health capacity is a key
variable that underpins the suc-
cessful implementation of pro-
grammes. When evaluating and
implementing a programme, cues
can be taken from other strategies
that have proven effective in build-
ing capacity [38].
Finally, approaches to measur-
ing and evaluating the success of
interventions (and their component
parts) must be defined and as-
sessed within the constraints of re-
al-world delivery. Using appropriate
measures in the context of imple-
mentation models bring a sharper
focus to quantification of the impact
of programmes [30].
Health behavioural interventions,
such as prevention, should follow the
dimensions of the RE-AIM framework
(i.e. reach, efficacy, adoption, imple-
mentation, and maintenance). To
ensure that preventive interventions
are effective, a focus must be placed
on the maintenance and sustainabil-
ity of the intervention. Furthermore,
given the clear role that policy and the integration of research findings strategies for overcoming barriers
environmental approaches play in and evidence into health-care policy to the adoption, adaptation, integra-
ensuring population-level access to and practice [39]. Implementation tion, scale-up, and sustainability of
prevention, increased research illus- science seeks to understand the evidence-based interventions, tools,
trating a more systematic increase in behaviour of health-care profes- policies, and guidelines. Expanding
implementation of these approaches sionals and other stakeholders as a the focus of implementation science
is critical, although such research is key variable in the sustainable up- to include policy research could be
rarely funded. take, adoption, and implementation very fruitful.
In the past 5 years there has been of evidence-based interventions. Brownson et al. [40] summarized
SECTION 6
CHAPTER 6.1
an increasing emphasis on imple- The field of implementation sci- lessons learned related to population-
mentation science research, which ence offers innovative approaches level prevention of chronic disease,
is the study of methods to promote to identify, understand, and develop including several that are relevant

to implementation science in cancer department) and individuals. Key ment of effective programming, the
prevention specifically: (i) start with questions include the following: cultural context, measurement strat-
environmental and policy interven- • How does the intervention align egies, and sustainability for imple-
tions as the key to initiating and sus- with local needs and provide avail- mentation must be considered.
taining systematic change, (ii) think able resources for feasible moni- Future priorities in the area of
across multiple levels of influence, toring strategies? changing behaviour include:
(iii) make better use of existing tools • Will additional technical assis- • identifying the components of
tance be needed for broader interventions that are key to sus-
for implementation, (iv) understand
implementation? tained change, and those that are
the local context and politics, (v) build
• How is this developed, delivered, most readily adapted to fit a popu-
new and non-traditional partner-
and sustained? lation group;
ships, (vi) address health disparities, • How flexible can and must the in- • a clearer understanding of when
and (vii) conduct more and better pol- tervention be? prevention strategies are not ad-
icy research. These lessons deserve • What are the measures of organi- equate and should be abandoned
particular attention in terms of identi- zational success and of overall or replaced;
fying untapped levers for increasing outcome? • a greater use of implementation
implementation of the evidence base
science to move from research to
for cancer prevention. Conclusions broader application of prevention
When planning to scale up inter- strategies;
Numerous effective prevention strat-
ventions for wider population cover- • maintaining programmes for the
egies have been evaluated over the
age, questions arise, such as the sustained achievement of desir-
past 5 years. Vaccination and to-
strength of the evidence base, the bacco control strategies have been able goals and population out-
ability to deliver the intervention at shown to be scalable and effec- comes; and
low cost, the approaches to monitor- tive in widespread implementation. • a better understanding of the
ing the consistency or integrity of the However, there is continuing devel- benefits of prevention through
delivery of the intervention, and out- opment in the areas of nutrition and the leading modifiable risk factors
comes across levels of health sys- physical activity, among other pre- and the benefits that extend be-
tem (health-care provider or health vention strategies. For the develop- yond individual countries.
References
1. Bray F, Ferlay J, Soerjomataram I, Siegel 5. National Preventative Health Taskforce 9. Islami F, Torre LA, Jemal A (2015).
RL, Torre LA, Jemal A (2018). Global (2009). Australia: the healthiest country Global trends of lung cancer mortali-
cancer statistics 2018: GLOBOCAN es- by 2020. National Preventative Health ty and smoking prevalence. Transl Lung
timates of incidence and mortality world- Strategy – the roadmap for action. Cancer Res. 4(4):327–38. https://doi.
wide for 36 cancers in 185 countries. CA Canberra: Commonwealth of Australia. org/10.3978/j.issn.2218-6751.2015.08.04
Cancer J Clin. 68(6):394–424. https://doi. PMID:26380174
org/10.3322/caac.21492 PMID:30207593 6. Koh HK, Graham G, Glied SA (2011).
Reducing racial and ethnic disparities: 10. Kirby T (2016). Australia tax increases
2. Sullivan R, Peppercorn J, Sikora K, the action plan from the Department to price cigarettes out of reach. Lancet
Zalcberg J, Meropol NJ, Amir E, et al. of Health and Human Services. Health Oncol. 17(6):e228. https://doi.org/10.1016/
(2011). Delivering affordable cancer care Aff (Millwood). 30(10):1822–9. https://
S1470-2045(16)30136-X PMID:27183848
in high-income countries. Lancet Oncol. d o i . o r g / 1 0 .13 7 7/ h l t h a f f . 2 0 11 . 0 6 7 3
12(10):933–80. https://doi.org/10.1016/ PMID:21976322 11. Azagba S, Sharaf MF (2013). The ef-
S1470-2045(11)70141-3 PMID:21958503 fect of graphic cigarette warning labels
7. NCHS (2016). Healthy People 2020 mid-
on smoking behavior: evidence from the
3. Lalonde M (1974). Social values and course review. Hyattsville (MD), USA:
public health. Can J Public Health. National Center for Health Statistics. Canadian experience. Nicotine Tob Res.
65(4):260–8. PMID:4849733 Available from: https://www.cdc.gov/nchs/ 15(3):708–17. https://doi.org/10.1093/ntr/
healthy _ people/ hp2020/ hp2020 _ mid nts194 PMID:22990228
4. DHEW (1979). Healthy people: the Sur course_review.htm.
geon General’s report on health promo- 12. Chiang CJ, Yang YW, You SL, Lai MS,
tion and disease prevention. Washington 8. WHO (2015). WHO report on the global to- Chen CJ (2013). Thirty-year outcomes
(DC), USA: Department of Health, bacco epidemic, 2015: raising taxes on to- of the national hepatitis B immunization
Education, and Welfare. Available from: bacco. Geneva, Switzerland: World Health program in Taiwan. JAMA. 310(9):974–6.
https://profiles.nlm.nih.gov/ps/access/ Organization. Available from: http://www. https://doi.org/10.1001/jama.2013.276701
NNBBGK.pdf. who.int/tobacco/global_report/2015/en/. PMID:24002285
504
13. Schweitzer A, Akmatov MK, Krause G 22. Ginsberg GM, Lauer JA, Zelle S, Baeten S, 32. Colditz GA, Emmons KM (2016). The
(2017). Hepatitis B vaccination timing: Baltussen R (2012). Cost effectiveness of role of universal health coverage in re-
results from demographic health surveys strategies to combat breast, cervical, and ducing cancer deaths and disparities.
in 47 countries. Bull World Health Organ. colorectal cancer in sub-Saharan Africa Lancet. 388(10045):638–40. https://doi.
95(3):199–209G. https://doi.org/10.2471/ and South East Asia: mathematical mod- org /10.1016/S0140 - 673 6(16) 3 0376 -2
BLT.16.178822 PMID:28250533 elling study. BMJ. 344:e614. https://doi. PMID:27236343
org/10.1136/bmj.e614 PMID:22389347
14. WHO (2017). Expanded Programme on 33. Kaufman J, Ames H, Bosch-Capblanch X,
Immunization (EPI) fact sheet: Maldives 23. Shei A, Costa F, Reis MG, Ko AI (2014). Cartier Y, Cliff J, Glenton C, et al. (2017).
2017. Geneva, Switzerland: World Health The impact of Brazil’s Bolsa Família condi- The comprehensive ‘Communicate to
Organization. Available from: http://origin. tional cash transfer program on children’s Vaccinate’ taxonomy of communication in-
searo.who.int /immunization/data /mal health care utilization and health outcomes. terventions for childhood vaccination in rou-
dives_2017.pdf. BMC Int Health Hum Rights. 14(1):10. tine and campaign contexts. BMC Public
https://doi.org/10.1186/1472-698X-14-10 Health. 17(1):423. https://doi.org/10.1186/
15. de Martel C, Plummer M, Vignat J, s12889-017-4320-x PMID:28486956
PMID:24690131
Franceschi S (2017). Worldwide burden
of cancer attributable to HPV by site, 24. Franks PW, Hanson RL, Knowler WC, 34. Montague M, Borland R, Sinclair C (2001).
country and HPV type. Int J Cancer. Sievers ML, Bennett PH, Looker HC Slip! Slop! Slap! and SunSmart, 1980-
141(4):664–70. https://doi.org/10.1002/ (2010). Childhood obesity, other car- 2000: skin cancer control and 20 years
ijc.30716 PMID:28369882 diovascular risk factors, and premature of population-based campaigning. Health
death. N Engl J Med. 362(6):485–93. Educ Behav. 28(3):290–305. https://
16. Huh WK, Joura EA, Giuliano AR, Iversen doi.org/10.1177/109019810102800304
OE, de Andrade RP, Ault KA, et al. (2017). PMID:11380050
PMID:20147714
Final efficacy, immunogenicity, and safety
analyses of a nine-valent human papil- 25. Jafar TH, Qadri Z, Islam M, Hatcher J, 35. Hill D, Marks R (2008). Health promotion
lomavirus vaccine in women aged 16-26 Bhutta ZA, Chaturvedi N (2008). Rise in programs for melanoma prevention: screw
years: a randomised, double-blind trial. childhood obesity with persistently high or spring? Arch Dermatol. 144(4):538–40.
Lancet. 390(10108):2143–59. https://doi. rates of undernutrition among urban https://doi.org/10.1001/archderm.144.4.538
org /10.1016/S 014 0 - 673 6 (17) 31821- 4 school-aged Indo-Asian children. Arch Dis PMID:18427051
PMID:28886907 Child. 93(5):373–8. https://doi.org/10.1136/ 36. Lee RM, Gortmaker SL (2018). Health
adc.2007.125641 PMID:17942586 dissemination and implementation within
17. Brotherton JM, Gertig DM, May C, Chappell
G, Saville M (2016). HPV vaccine impact 26. Almas A, Islam M, Jafar TH (2013). School- schools. In: Brownson RC, Colditz GA,
in Australian women: ready for an HPV- based physical activity programme in pre- Proctor EK, editors. Dissemination and
based screening program. Med J Aust. adolescent girls (9-11 years): a feasibility implementation research in health: trans-
204(5):184, e1. https://doi.org/10.5694/ trial in Karachi, Pakistan. Arch Dis Child. lating science to practice. 2nd ed. New
mja15.01038 PMID:26985843 98(7):515–9. https://doi.org/10.1136/arch York (NY), USA: Oxford University Press;
dischild-2012-303242 PMID:23661575 pp. 401–16.
18. Medical Services Advisory Committee
(2013). National Cervical Screening 27. Craike M, Wiesner G, Hilland TA, 37. Shelton RC, Cooper BR, Stirman SW (2018).
Program Renewal: evidence review. Bengoechea EG (2018). Interventions The sustainability of evidence-based in-
Canberra: Commonwealth of Australia. to improve physical activity among so- terventions and practices in public health
Available from: http://www.cancerscreening. cioeconomically disadvantaged groups: and health care. Annu Rev Public Health.
gov.au/internet/screening/publishing.nsf/ an umbrella review. Int J Behav Nutr 39(1):55–76. https://doi.org/10.1146/
Content /E6A 211A6FFC29E2CCA 257C Phys Act. 15(1):43. https://doi.org/10.1186/ a n n u r ev - p u b l h e a l t h - 0 4 0 617- 0147 31
ED007FB678/$File/Review%20of%20 PMID:29328872
s12966-018-0676-2 PMID:29764488
Evidence%20notated%2013.06.14.pdf.
38. Leeman J, Calancie L, Hartman MA,
28. Colchero MA, Rivera-Dommarco J, Popkin
19. Simms KT, Laprise JF, Smith MA, Lew Escoffery CT, Herrmann AK, Tague
BM, Ng SW (2017). In Mexico, evidence of
JB, Caruana M, Brisson M, et al. (2016). LE, et al. (2015). What strategies are
sustained consumer response two years
Cost-effectiveness of the next generation used to build practitioners’ capacity to
after implementing a sugar-sweetened
nonavalent human papillomavirus vaccine implement community-based interven-
beverage tax. Health Aff (Millwood). 36(3):
in the context of primary human papilloma- tions and are they effective?: a system-
564–71. https://doi.org/10.1377/hlthaff.2016.
virus screening in Australia: a comparative atic review. Implement Sci. 10(1):80.
1231 PMID:28228484
modelling analysis. Lancet Public Health. h t t p s: //d o i .o r g /10 .118 6 /s13 012- 015 -
1(2):e66–75. https://doi.org/10.1016/S2468- 29. Emmons KM, Colditz GA (2017). 0272-7 PMID:26018220
2667(16)30019-6 PMID:29253419 Realizing the potential of cancer pre-
39. DHHS (2016). Dissemination and imple-
vention – the role of implementation sci-
20. Holman DM, Benard V, Roland KB, Watson mentation research in health, PAR-16-
ence. N Engl J Med. 376(10):986–90.
M, Liddon N, Stokley S (2014). Barriers to 238. Washington (DC), USA; Department
https://doi.org/10.1056/NEJMsb1609101
human papillomavirus vaccination among of Health and Human Services. Available
PMID:28273020
US adolescents: a systematic review of from: https://grants.nih.gov/grants/guide/
the literature. JAMA Pediatr. 168(1):76– 30. Brownson RC, Colditz GA, Proctor EK, pa-files/par-16-238.html.
82. https://doi.org/10.1001/jamapediatrics. editors (2018). Dissemination and imple-
40. Brownson RC, Haire-Joshu D, Luke DA
2013.2752 PMID:24276343 mentation research in health: translating
(2006). Shaping the context of health: a
science to practice. 2nd ed. New York
21. Walling EB, Benzoni N, Dornfeld J, review of environmental and policy ap-
(NY), USA: Oxford University Press.
Bhandari R, Sisk BA, Garbutt J, et al. proaches in the prevention of chronic
(2016). Interventions to improve HPV 31. Lobb R, Colditz GA (2013). Implementation diseases. Annu Rev Public Health.
vaccine uptake: a systematic review. science and its application to population 27(1):341–70. https://doi.org/10.1146/
Pediatrics. 138(1):e20153863. https:// health. Annu Rev Public Health. 34(1): annurev.publhealth.27.021405.102137
d o i . o r g / 1 0 .1 5 4 2 / p e d s . 2 0 1 5 - 3 8 6 3 235–51. https://doi.org/10.1146/annurev- PMID:16533121
PMID:27296865 publhealth-031912-114444 PMID:23297655
SECTION 6
CHAPTER 6.1

6.2
Improving diet and nutrition, physical

activity, and body weight
From evidence to practice
Annie S. Anderson Christine M. Friedenreich (reviewer)
Martin J. Wiseman (reviewer)
political, environmental, and cul- the importance of using compre-

SUMMARY tural factors; global efforts to reduce hensive approaches that take into
the burden of cancer need to take consideration health-enhancing en-
●● There is now clear evidence that
account of these social determi- vironments, behaviour change com-
the greatest change in diet and
physical activity across a popula- nants in order to produce equitable munications, and systems change.
tion can be achieved when pop- changes in health and well-being [1]. Health services, including cancer
ulation-wide approaches, such The NOURISHING framework and screening programmes, can contrib-
as policy specification, are com- the new Driving Action framework ute to national efforts [3].
bined with individually targeted from World Cancer Research Fund Single strategies, such as those
approaches. International [2] (Fig. 6.2.1) highlight focusing on communications and
●● Approaches to changing diet
and physical activity should Fig. 6.2.1. The Driving Action framework from World Cancer Research Fund International.
take into consideration health-
enhancing environments, be-
haviour change communica-
tions, and systems change.
●● Government regulatory mea-
sures, such as product nutrient
specification, and fiscal interven-
tions can be used to successfully
affect dietary patterns, but indus-
try opposition can influence the
design of optimal programmes.
●● Educational approaches and
awareness-raising strategies can
motivate and support people
to change their behaviour, but
their impact on dietary intake
alone is small and may be low-
est in vulnerable groups.
●● No single intervention can ad-
dress the challenge of achieving
healthy dietary patterns.
Behavioural risk factors for cancer,

such as diet and physical activity,
are influenced by underlying social
determinants, including economic,
506
education, have limited effects and tion for effective change, the imple-
can be associated with increases mentation of effective intervention FUNDAMENTALS
in health inequalities. A compre- policies is dependent on govern-
hensive community approach to ment knowledge, the capacity and ■■ Although evidence from trials,
changing health behaviours has will to act, and the governance struc- modelling (i.e. theoretical anal-
been demonstrated historically in tures to translate evidence into prac- ysis estimated from existing
the North Karelia Project in Finland, tice. In addition, actions have to take data), and practical experience
which showed significant reduc- account of the local context and the can guide action for effective
tions in cardiovascular outcomes, specific needs of the population (see change, the implementation
followed by reductions in cancer Chapter 6.1). of effective and equitable
mortality, arising from “the correct intervention policies, such as
theory base, comprehensive work Diet and nutrition a sugar tax, is dependent on
with the population, and much hard government action.
work in the community” [4]. Food ■■ Natural experiments can
There is a growing evidence base For cancer prevention, both dietary provide useful evidence for
on the impact of behaviour change quantities (i.e. appropriate energy intervention planning and
communications and programmes, intake) and diet quality are impor- policy development.
which include individual-level coun- tant. Plant-based dietary patterns –
selling by health professionals, edu- with an emphasis on whole grains, ■■ Evidence from comprehensive
cation, and social support, such as vegetables, fruits, and beans, and community programmes sug-
demonstrated by the diabetes pre- limited intake of red meat, processed gests that a combination of be-
vention programmes. However, these meat, sugar, ultra-processed foods, havioural theory, commitment,
approaches tend to be intensive and sugar-sweetened beverages, and and national and local action
may have low generalizability, espe- alcoholic beverages – are desirable are key factors in the design of
cially in the most vulnerable commu- (see Chapter 2.6). programmes and policies.
nities [5]. There is now clear evidence It is clear that multiple factors, ■■ When implementing pro-
that the greatest change in diet and beyond personal decision-making, grammes that were successful
physical activity across a population influence food choice and dietary in other regions, care needs to
can be achieved when population- patterns, including sociocultural back- be taken to consider the local
wide approaches, such as policy ground, lifestyle patterns, and eco- context and the specific needs
specification, are combined with indi- nomic and commercial pressures. of the population.
vidually targeted approaches. Therefore, to achieve equitable, se-
Although evidence from trials, cure, sustainable, and optimal die- ■■ Most research evidence
modelling (i.e. theoretical analysis tary intake, wider environmental fac- has short- to medium-term
estimated from existing data), and tors need to be embraced in addition outcomes, and more research
practical experience can guide ac- to individually focused approaches. is needed on programme
sustainability, reach, and long-
term outcomes to assess the
Fig. 6.2.2. Vegetables and fruits at a market in France. impact of programmes and
policies on cancer outcomes
across all population groups.
No single intervention can address

the challenge of achieving healthy
dietary patterns.
Hawkes et al. [6] described four
mechanisms through which food
policies can have an impact on diet
throughout the life-course: (i) pro-
viding an enabling environment for
the learning of healthy preferences
in childhood (because preferences
are often persistent and resis-
tant to change); (ii) identifying and
CHAPTER 6.2
SECTION 6
overcoming barriers to the expres-

sion of healthy preferences, such
as strategies related to physical
Chapter 6.2 • Improving diet and nutrition, physical activity, and body weight 507
resources, information, and skills; sessed. Implementing regulations vending machines, nutrition edu-
(iii) approaches that encourage for food composition (e.g. maximum cation, and provision of access to
people to reassess existing un- limits) and standards for product safe drinking-water – have been
healthy preferences at the point availability (e.g. trans fatty acids and associated with modest reductions
of purchase through changes in salt) for use in food marketing and in consumption in many countries.
price, availability, and presentation procurement (such as in local and However, intakes remain high, nota-
(sometimes referred to as choice national government catering set- bly in children (compared with adults)
architecture); and (iv) the ability to tings, worksites, nurseries, schools, and in groups with lower socioeco-
stimulate food-systems response and food assistance programmes), nomic status. Sales of sugar-sweet-
so that changes made by one ac- accompanied by mandatory label- ened beverages are continuing to
tion, such as mandatory nutrition ling, can have a significant effect on increase in low- and middle-income
labelling, have an impact elsewhere population dietary patterns [8]. The countries; this is most likely to be re-
in the food environment, for exam- impact is likely to be greatest when lated to the low cost, large unit size,
ple product reformulation. regulatory rather than voluntary ap- and marketing.
For decades, nutrition pro- proaches are used [9]. Recent efforts have focused
grammes have focused primarily on on the additional, population-wide
behaviour change communications Beverages strategy of introducing taxes on
such as education programmes, Caloric beverages can make a sig- sugar-sweetened beverages, with
food labelling information (e.g. traf- nificant contribution to excess energy the aims of decreasing consump-
fic-light labelling), and skills (e.g. intake and the development of weight tion, encouraging beverage compa-
food preparation). These are con- gain, or may decrease appetite for nies to reformulate their products,
sidered to be important strategies more nutrient-dense foods, thus de- and generating income to support
to support people to practically im- creasing dietary quality. In addition, public health. Taxes are commonly
plement advice, to help frame pub- alcoholic beverages are of concern identified as the single most impor-
lic understanding, and to generate because of the established associa- tant policy approach for reducing
support for healthy public policy, but tion between alcohol consumption intakes of sugar-sweetened bever-
their impact on dietary intake alone and the incidence of cancer at sev- ages. Although taxes are financially
is small and may be lowest in vulner- eral sites (see Chapter 2.3). regressive for low-income groups,
able groups. More recently, many this financial impact can be bal-
countries have developed voluntary Sugar-sweetened beverages anced by using tax revenues to
codes of practice in conjunction with Consumption of sugar-sweetened reduce the prices of healthier food
the food industry, for example reduc- beverages is associated with weight options [10]. It is estimated that in
tion in sugar intake, but these have gain, overweight, and obesity, which 2018 at least 26 countries had intro-
not been demonstrated to achieve increase the risk of cancer. Health duced a sugar tax, with a significant
desirable levels of change. promotion efforts – including nutri- impact on purchases. For exam-
Increasingly, it is recognized that ent regulations in schools, bans on ple, in 2014 Mexico introduced an
government regulatory measures,
such as product nutrient specifica-
Fig. 6.2.3. A young man in South Africa drinking water. Reducing consumption of sugar-
tion, and fiscal interventions can be
sweetened beverages is a positive step towards a healthier diet.
used to successfully affect dietary
patterns, but industry opposition
can influence the design of optimal
programmes. Actions by govern-
ments should be monitored, and ac-
countability mechanisms should be
in place at the local, national, and
international levels [7]. Fiscal incen-
tives and disincentives, such as food
prices, subsidies, and financial re-
wards and penalties, are considered
to be positive approaches in chang-
ing dietary behaviours, notably when
implemented as part of an integrated
package of mutually reinforcing ac-
tivities, such as education and mar-
keting. However, the level of financial
impact needed to improve health
outcomes needs to be carefully as-
508
excise tax of 10% on sugar-sweet- The approaches considered to and health-care organizations [17].
ened beverages, accompanied by be the least effective in decreasing Positive effects were also reported
campaigns to raise awareness of alcohol consumption are educa- from environmental and policy ap-
the association between consumption in schools, public service an- proaches that include the creation
tion of sugar-sweetened beverages nouncements, and voluntary regu- or enhancement of access to places
and diabetes; this was followed by lation by the alcohol industry [16]. to be active, through infrastructural
an average decrease of 7.6% in initiatives such as community-scale
purchases of taxed beverages in and street-scale urban design and
Physical activity
2014 and 2015 [11]. land use, an active transport policy
Consistent with the new Driving and practices, and community-wide
Alcohol Action policy framework from World policies and planning [17].
Reviews of approaches to reduce Cancer Research Fund Interna The same review recommended
alcohol consumption indicate that tional [2], evidence suggests that the informational approaches of
the most cost-effective strategies health-enhancing environments and community-wide and mass media
include taxes that increase prices, behaviour change communications campaigns, as well as short mes-
restrictions on the physical avail- are key components for increasing sages about physical activity tar-
ability of alcohol, drink–driving physical activity. In addition, a sys- geting key community sites. Given
laws, brief interventions with at-risk tems approach is needed to provide the importance of social support,
drinkers, and the treatment of drink- a structural framework for national behavioural and social approaches
ers with alcohol dependence [12]. and local action. Examples include are effective for increasing physical
Data from natural experiments government policies that ensure ad- activity within communities, neigh-
suggest that the level of price restric- equate and affordable access to and bourhoods, and worksites. For chil-
tion is important and that similar in- use of natural environments for ac- dren, school-based strategies that
terventions can have different effects tivity, recreation, and play. encompass physical education,
depending on context and culture A 2012 review of physical activ- classroom activities, after-school
[13]. The effects are influenced by ity interventions around the world sports, and active transport can
availability and licensing, acceptabil- reported that initiatives to pro- produce positive impacts. A key
ity of alcohol use within society, mar- mote physical activity can have in- message from the review is that
keting (including sponsorship), and creased effectiveness when health although individuals need to be
labelling information (i.e. alcohol con- agencies form partnerships and co- informed and motivated to adopt
tent, calories, serving size). Changing ordinate efforts with several stake- physical activity, the public health
consumer attitudes and norms about holders: schools; businesses; poli- priority should be to ensure that
alcohol consumption and garnering cy, advocacy, nutrition, recreation, environments are safe and sup-
support for comprehensive policy ap- planning, and transport agencies; portive of health and well-being. In
proaches may be challenging in con-
texts where knowledge levels about
Fig. 6.2.4. Young men playing football on the beach in Rio de Janeiro, Brazil.
the association between alcohol
consumption and cancer risk are low
[14]. Opportunities to provide warning
labels related to cancer are consid-
ered to be a useful avenue to raise
awareness of cancer risk, although
such approaches are not supported
by the alcohol industry.
At the individual level, opportunis-
tic screening (assessment of alcohol
consumption) in primary care and
other health-care settings, followed
by brief interventions, is an effective
approach, which has been demon-
strated to have a moderate effect
on reducing alcohol consumption
and increasing the number of people
drinking alcohol below levels associ-
ated with increased risk. Brief inter-
CHAPTER 6.2
SECTION 6
ventions with multiple contacts or

follow-up sessions appear to be the
most effective [15].
Fig. 6.2.5. Women and girls participating in a free public yoga course held every De tivity will compete with other health-
cember morning in Yangon, Myanmar. care demands.
Sedentary behaviour
Research on changing sedentary
behaviour (i.e. time spent sitting) in
the workplace, during leisure time,
commuting, and in the household is
relatively recent (see Chapter 2.7).
Several reviews have highlighted
that interventions that target both
physical activity and sedentary be-
haviour are generally ineffective
in changing time spent sitting [19].
This finding underlines the impor-
tance of an intervention having a
primary aim of reducing sedentary
behaviour; otherwise, effects on
this outcome tend to be small.
Current evidence from behav-
iour change studies indicates that
environmental restructuring, per-
addition, the authors noted that to physical activity in built environ- suasion, education, and training
properly support initiatives for the ments, such as limiting street access generally show promise in reduc-
promotion of physical activity, work- to cars, increasing access to cyclists ing sedentary behaviour. A recent
forces need to be trained in physi- and pedestrians, and improving walk- systematic review evaluated the
cal activity and health, core public ability, especially when combined evidence from randomized con-
health disciplines, and methods of with promotional efforts. In addition, trolled trials on the effectiveness of
intersectoral collaboration [17]. although most countries have adopt- workplace interventions to reduce
More recently, a review of inter- ed national physical activity policies time spent sitting at work [20]. The
vention studies in low- and middle- and plans, major challenges with im- review concluded that sit–stand
income countries highlighted that plementation are evident. In low- and desks are effective in reducing sit-
although the number of interven- middle-income countries, resources ting time at work, total sitting time,
tions is increasing, the challenge to scale up effective interventions and duration of sitting bouts. In ad-
is greater because the prevalence and train workforces in physical ac- dition, short breaks (1–2 minutes
of physical inactivity is higher in ur-
ban versus rural communities at a
time when there is a rising global Fig. 6.2.6. Using a standing workstation effectively reduces sitting time at work.
trend towards urbanization [18].
The review of intervention studies
in low- and middle-income coun-
tries, including examples from the
Islamic Republic of Iran, China,
India, South Africa, and Vanuatu,
reported an increasing number
of promising approaches, includ-
ing community-wide campaigns
(e.g. using multiple communication
media to raise programme aware-
ness), strategies that include social
support (e.g. walking groups), and
school-based programmes, al-
though not all of these approaches
were found to be effective.
There is increasing evidence of
the effectiveness of community-wide
policies and planning to enhance
510
every 30 minutes) were more effec- ity is considered particularly helpful needed for sustainable behaviour
tive than long breaks (two 15-min- in maintenance of weight loss. change. It is clear that relevant policy
ute breaks per workday) in the short The global burden of obesity actions for addressing obesity need
term. Computer prompting resulted highlights an urgent need to iden- to be identified in a systematic man-
in decreases in the average number tify and implement policies that will ner. The Food Environment Policy
and duration of sitting bouts lasting have an impact on prevention and Index [24], which offers a useful tool
30 minutes or more [20]. management. To date, no country for developing consensus for action,
In randomized controlled tri- has reversed the obesity epidemic has been used in Thailand, New
als, interventions to reduce non- in its population, and evidence on Zealand, Australia, and England.
occupational sedentary behaviour effective national programmes is For example, in England the top-
have been shown to be effective lacking. Much of the work in this priority policy actions identified for
in adults. The current evidence arena has been focused on tackling government were those that affect
suggests that use of technology to childhood obesity, given the burden both children and adults: (i) control
reduce sedentary time (e.g. alert- of noncommunicable diseases that the advertising of unhealthy foods
ing the user to accumulated time are now presenting in adolescence. to children; (ii) implement the levy
spent sedentary), use of specific However, many children who are on sugary beverages; (iii) reduce
behaviour change techniques (e.g. overweight also have parents who the sugar, fat, and salt content in
self-monitoring), or a combination are overweight, and the adult world processed foods; (iv) monitor school
of both are characteristics of ef- shapes what children see and re- and nursery food standards; (v) pri-
fective programmes [21]. Reduced spond to. Societal actions that have oritize health and the environment in
television viewing, computer use, favourable impacts on vulnerable the 25-year Food and Farming Plan;
and total transport-related sitting groups of all ages and backgrounds (vi) adopt a national food action
time and the use of smart technolo- offer the greatest potential for equi- plan; (vii) monitor the food environ-
gies need further investigation, and table effects. ment; (viii) apply buying standards to
these are promising areas for fur- Tackling obesity is more com- all public institutions; (ix) strengthen
ther investigation. plex than addressing either en- planning laws to discourage less-
ergy intake or energy expenditure, healthy food offers; and (x) evaluate
and there are no simple solutions. food-related programmes and poli-
Obesity Approaches that tackle both envi- cies [25].
Excess body fat results from an im- ronmental factors, which support or The combined forces of regula-
balance between energy consumed undermine the ability of people to tory actions from governments and
from food and beverages and en- participate in healthful behaviours, increased efforts from industry and
ergy expenditure, notably through and individual action are desirable. civil society will be necessary to
physical activity. Data from weight- Roberto et al. [23] highlighted how address obesity (see Chapter 6.9).
loss studies clearly show that ener- food environments exploit people’s Public advocacy efforts [26] (includ-
gy intake is the most important driv- biological, psychological, social, ing those from cancer organizations)
er for achieving changes in energy and economic vulnerabilities, mak- are considered to be a key compo-
balance, although physical activity ing it easier for them to eat pro- nent in creating demand and sup-
is also important. Review-level evi- cessed foods and follow unhealthy port for effective obesity policies and
dence demonstrates that combined dietary patterns (see Chapter 2.6). in mitigating reaction against their
diet plus physical activity interven- This situation reinforces preferenc- implementation. Important issues
tions can result in a loss of 8–11% es and demands for foods of poor for obesity coalitions to address
of body weight within 6 months, nutritional quality, thus maintaining include challenges from the food
whereas moderate- to high–inten- unhealthy food environments. and beverage industry and ways to
sity interventions without reduction Approaches by governments to avoid stigmatization by insensitive
in energy intake achieve a loss of address obesity have tended to fo- programmes and campaigns, and
about 2–3% of body weight within cus on one or two target areas and thus lose support for obesity pro-
the same period [22]. Physical activ- lack the comprehensive approach grammes by civil society. CHAPTER 6.2
SECTION 6
References
1. Marmot M (2018). Diet, cancer, and NCD 11. Álvarez-Sánchez C, Contento I, Jiménez- 20. Shrestha N, Kukkonen-Harjula KT,
prevention. Lancet Oncol. 19(7):863–4. Aguilar A, Koch P, Gray HL, Guerra LA, Verbeek JH, Ijaz S, Hermans V, Pedisic Z
https://doi.org/10.1016/S1470-2045(18) et al. (2018). Does the Mexican sugar- (2018). Workplace interventions for reduc-
30382-6 PMID:29803702 sweetened beverage tax have a signaling sitting at work. Cochrane Database
ing effect? ENSANUT 2016. PLoS One. Syst Rev. (6):CD010912. https://doi.
2. World Cancer Research Fund International 13(8):e0199337. https://doi.org/10.1371/ org/10.1002/14651858.CD010912.pub4
(2018). Driving action to prevent cancer journal.pone.0199337 PMID:30133438 PMID:29926475
and other non-communicable diseases:
a new policy framework for promoting 12. Alcohol and Public Policy Group 21. Thraen-Borowski KM, Ellingson LD,
healthy diets, physical activity, breastfeed- (2010). Alcohol: no ordinary commod- Meyer JD, Cadmus-Bertram L (2017).
ing and reducing alcohol consumption. ity – a summary of the second edition. Nonworksite interventions to reduce sed-
Available from: https://www.wcrf.org/sites/ Addiction. 105(5):769–79. https://doi. entary behavior among adults: a system-
default/files/driving-action.pdf. org/10.1111/j.1360 - 0443.2010.02945.x atic review. Transl J Am Coll Sports Med.
PMID:20331569 2(12):68–78. PMID:28993817
3. Anderson AS, Mackison D, Boath C,
Steele R (2013). Promoting changes in diet 13. Nelson JP, McNall AD (2017). What hap- 22. Chin SH, Kahathuduwa CN, Binks M
and physical activity in breast and colorec- pens to drinking when alcohol policy
(2016). Physical activity and obesity:
tal cancer screening settings: an unex- changes? A review of five natural experi-
what we know and what we need to know.
plored opportunity for endorsing healthy ments for alcohol taxes, prices, and avail-
Obes Rev. 17(12):1226–44. https://doi.
behaviors. Cancer Prev Res (Phila). ability. Eur J Health Econ. 18(4):417–34.
org/10.1111/obr.12460 PMID:27743411
6(3):165–72. https://doi.org/10.1158/1940- ht t p s: //d o i.o r g /10 .10 07/s1019 8 - 016 -
6207.CAPR-12-0385 PMID:23324132 0795-0 PMID:27055901 23. Roberto CA, Swinburn B, Hawkes C,
14. Buykx P, Li J, Gavens L, Hooper L, Lovatt Huang TT, Costa SA, Ashe M, et al.
4. Puska P, Vartiainen E, Nissinen A,
M, Gomes de Matos E, et al. (2016). Public (2015). Patchy progress on obesity pre-
Laatikainen T, Jousilahti P (2016).
awareness of the link between alcohol vention: emerging examples, entrenched
Background, principles, implementation,
and cancer in England in 2015: a popula- barriers, and new thinking. Lancet.
and general experiences of the North
Karelia Project. Glob Heart. 11(2):173–8. tion-based survey. BMC Public Health. 385(9985):2400–9. https://doi.org/10.1016/
https://doi.org/10.1016/j.gheart.2016.04. 16(1):1194. https://doi.org/10.1186/s12889- S0140-6736(14)61744-X PMID:25703111
010 PMID:27242083 016-3855-6 PMID:27899099
24. Swinburn B, Sacks G, Vandevijvere
5. Aziz Z, Absetz P, Oldroyd J, Pronk NP, 15. Álvarez-Bueno C, Rodríguez-Martín S, Kumanyika S, Lobstein T, Neal B,
Oldenburg B (2015). A systematic re- B, García-Ortiz L, Gómez-Marcos MÁ, et al.; INFORMAS (2013). INFORMAS
view of real-world diabetes prevention Martínez-Vizcaíno V (2015). Effectiveness (International Network for Food and
programs: learnings from the last 15 of brief interventions in primary health care Obesity/non-communicable diseases
years. Implement Sci. 10(1):172. https:// settings to decrease alcohol consumption Research, Monitoring and Action Support):
d o i.o r g /10 .118 6 /s13 012- 015 - 0 3 5 4 - 6 by adult non-dependent drinkers: a system- overview and key principles. Obes Rev.
PMID:26670418 atic review of systematic reviews. Prev Med. 14(Suppl 1):1–12. https://doi.org/10.1111/
76(Suppl):S33–8. https://doi.org/10.1016/j. obr.12087 PMID:24074206
6. Hawkes C, Smith TG, Jewell J, Wardle ypmed.2014.12.010 PMID:25514547
J, Hammond RA, Friel S, et al. (2015). 25. Watson F, Taylor A, Rayner M, Lobstein
Smart food policies for obesity preven- 16. WHO (2011). Reducing risks and prevent- T, Hinks R (2018). Priority actions for ad-
tion. Lancet. 385(9985):2410–21. https:// ing disease: population-wide interventions. dressing the obesity epidemic in England.
doi.org/10.1016/S0140-6736(14)61745-1 In: Global status report on noncommunica- Public Health Nutr. 21(5):1002–10. https://
PMID:25703109 ble diseases 2010. Geneva, Switzerland: doi.org/10.1017/S13689 80 0170 0350 0
World Health Organization; pp. 47–60. PMID:29233230
7. Swinburn B, Kraak V, Rutter H,
Vandevijvere S, Lobstein T, Sacks G, 17. Heath GW, Parra DC, Sarmiento OL, 26. Huang TT, Cawley JH, Ashe M, Costa
et al. (2015). Strengthening of account- Andersen LB, Owen N, Goenka S, et al.; SA, Frerichs LM, Zwicker L, et al.
ability systems to create healthy food Lancet Physical Activity Series Working (2015). Mobilisation of public support
environments and reduce global obe- Group (2012). Evidence-based intervention for policy actions to prevent obesity.
sity. Lancet. 385(9986):2534–45. https:// in physical activity: lessons from around the Lancet. 385(9985):2422–31. https://doi.
doi.org/10.1016/S0140-6736(14)61747-5 world. Lancet. 380(9838):272–81. https:// org /10.1016/S 014 0 - 673 6 (14) 6174 3 - 8
PMID:25703108 doi.org/10.1016/S0140-6736(12)60816-2
PMID:25703113
PMID:22818939
8. Mytton OT, Clarke D, Rayner M (2012).
Taxing unhealthy food and drinks to im- 18. Sallis JF, Bull F, Guthold R, Heath
prove health. BMJ. 344:e2931. https://doi. GW, Inoue S, Kelly P, et al.; Lancet
org/10.1136/bmj.e2931 PMID:22589522 Physical Activity Series 2 Executive
Committee (2016). Progress in physical
9. Mozaffarian D, Angell SY, Lang T, Rivera activity over the Olympic quadrennium.
JA (2018). Role of government policy in Lancet. 388(10051):1325–36. https://doi.
nutrition – barriers to and opportunities org /10.1016/S0140 - 673 6(16) 3 0 5 81- 5
for healthier eating. BMJ. 361:k2426. PMID:27475270
h t t p s : / / d o i . o r g / 1 0 .11 3 6 / b m j . k 2 4 2 6
PMID:29898890 19. Howlett N, Trivedi D, Troop NA, Chater
AM (2019). Are physical activity interven-
10. Knai C, James L, Petticrew M, Eastmure tions for healthy inactive adults effective
E, Durand MA, Mays N (2017). An evalu- in promoting behavior change and main-
ation of a public–private partnership to re- tenance, and which behavior change
duce artificial trans fatty acids in England, techniques are effective? A systematic
2011–16. Eur J Public Health. 27(4):605– review and meta-analysis. Transl Behav
8. https://doi.org/10.1093/eurpub/ckx002 Med. 9(1):147–57. https://doi.org/10.1093/
PMID:28339665 tbm/iby010 PMID:29506209
512
6.3
Vaccination
The prospect of eliminating some
cancer types
Silvia Franceschi Julia Brotherton (reviewer)
Iacopo Baussano Laia Bruni (reviewer)
85 countries had established Africa, the Amazon basin, China, the

SUMMARY HPV vaccination programmes. Republic of Korea, and countries in
South-East Asia [2]. In 2018, there
●● Hepatitis B virus (HBV) infection ●● Comprehensive data document
were an estimated 841  000 new
is very common in some areas the safety and high efficacy of
cases of liver cancer and 781 000
of the world. In 2016, an esti- HPV vaccines, especially in ad-
deaths from liver cancer worldwide
mated 292 million people were olescent girls, who are the prior-
[3]. Vaccines against HBV have been
living with chronic HBV infection ity target for HPV vaccination.
available for several decades, and
worldwide. HBV infection is also ●● Anti-vaccination campaigns and their efficacy in preventing chronic
responsible for approximately the relatively high cost, coupled HBV infection and liver cancer has
1 million deaths per year. with the necessarily protracted been clearly demonstrated in chil-
time frame to cancer prevention, dren and adolescents. It is expected
●● Highly effective vaccines against
hamper adequate coverage and that HBV vaccination will nearly elimi-
HBV infection have been avail-
universal implementation of HPV nate HBV-associated liver cancer in
able since 1982. By 2016, vaccination. many areas when the vaccinated
185 countries had introduced populations reach adulthood [4].
HBV vaccination, and vaccina- HPV is the most common sexu-
tion coverage in children had A notable fraction of cancer cases ally transmitted virus. Infection
reached 87% globally. in humans (~15%) are caused by typically resolves asymptomatically
●● HBV vaccination of babies infections [1], and these are largely within 1–2 years, but certain types
at birth is necessary to pre- amenable to effective preventive of HPV (called oncogenic types) can
vent mother-to-child transmis- interventions. Among the most im- cause cancers of the cervix, anus,
portant infections associated with vulva, vagina, penis, and orophar-
sion, but more than half of the
cancers are human papillomavirus ynx over extended time periods in
world’s children fail to receive a
(HPV), Helicobacter pylori (see individuals in whom HPV infection
birth dose.
Chapters 2.2 and 5.4), hepatitis B is not cleared by the immune sys-
●● Thirteen high-risk human papil- virus (HBV), and hepatitis C virus tem. Highly effective vaccines have
lomavirus (HPV) types, particu- (HCV). To date, only cancers relat- been available since 2006 to pre-
larly HPV type 16, cause cervi- ed to HPV and HBV can be prevent- vent infection by HPV16 and HPV18,
cal cancer (about 570 000 new ed through vaccination. Because of which are the most oncogenic types
cases per year in 2018) and anal the long latency between the occur- and are responsible for most HPV-
cancer, and substantial fractions rence of infection and the diagnosis related cancers. Recently, a vaccine
of cancers of the vulva, vagina, of cancer, data on efficacy against has become available that also tar-
penis, and oropharynx. invasive cancers remain limited, but gets oncogenic types HPV31, 33,
findings on precancerous lesions 45, 52, and 58.
●● Three prophylactic vaccines, and viral end-points are extremely The efficacy and cost–effective-
consisting of empty viral cap- favourable and robust. ness of the HPV vaccine are great-
sids of HPV types 16 and 18, Chronic infection with HBV is est in previously unexposed wom-
CHAPTER 6.3
SECTION 6
alone or with an additional one of the most important causes en. Therefore, HPV vaccination
two or seven types, have been of liver cancer, particularly in highly is preferentially recommended for
available since 2006. By 2018, endemic areas such as sub-Saharan pre-adolescent girls. By 2018, 85
Chapter 6.3 • Vaccination 513

countries had established HPV vac- if transmission occurs during birth

cination programmes [5]. However, and early childhood. Overall, up to FUNDAMENTALS
most girls in low- and middle-in- 40% of men and 15% of women
come countries, who are at highest with a perinatally acquired HBV in- ■■ In some low-income countries,
risk of cervical cancer, are not yet fection will die of liver cirrhosis or up to one third of all cases of
immunized [6]. HPV vaccines are hepatocellular carcinoma [10]. cancer are directly associated
efficacious at preventing infections In high-risk areas, HBV is re- with various infections. This of-
and lesions not only in the cervix sponsible for 50–80% of cases of fers the prospect of prevention
but also at other anatomical sites liver cancer [11].The attributable through vaccination.
where they have been investigated, fractions for liver cancers due to
■■ The hepatitis B virus (HBV)
but only global high-coverage mass HBV and HCV vary substantially by
vaccine was the first vaccine
vaccination programmes are ex- country (Fig. 6.3.1) [2]. HBV caus-
designed to prevent a major
pected to reduce the incidence of es about two thirds of liver cancer
human cancer type. The vac-
and mortality from cancers asso- cases in less-developed countries
cine can safely and effectively
ciated with vaccine-targeted HPV but only about one quarter of cases
be administered simultane-
types in the next few decades [7]. in more-developed countries. For
ously with many other routine
This chapter summarizes the HCV-attributable cases, the pattern
childhood immunizations.
epidemiological features of HBV is nearly opposite.
and HPV infections and the perfor- ■■ One of the first nationwide HBV
mance of vaccines against these
Hepatitis B virus vaccine vaccination programmes was
infections and the associated can- The HBV vaccine was the first vac- implemented in Taiwan, China,
cers, with a focus on the large cine designed to prevent a major and has resulted in a marked
amounts of data that have accumu- human cancer type [12]. A highly decrease in the incidence of
lated in the past 5 years. effective vaccine has been avail- hepatocellular carcinoma.
able since 1982, but worldwide
■■ Prevention of chronic HBV
vaccination only ramped up after
Hepatitis B virus GAVI, the Vaccine Alliance, started
infection through vaccination
is anticipated to result in
Hepatitis B virus and liver supporting HBV vaccine in 2001 [8].
decreases in the rates of liver
The current vaccine is a recombi-
cancer cancer, but several decades
nant HBV surface antigen (HBsAg)
HBV is a highly contagious DNA vi- will be required to confirm
produced in yeast or mammalian
rus that is transmitted by exposure this outcome.
cells into which the HBsAg gene is
to HBV-contaminated blood and inserted using plasmids. The vac- ■■ Prophylactic human papillo-
other body fluids, including semen cine, administered as a three-dose mavirus (HPV) vaccines were
and vaginal fluids [8]. The virus is series, is highly safe and 95% ef- initially developed to prevent
transmitted from mother to infant fective in preventing HBV infection infection with a small number
and from child to child, as well as and its chronic consequences. In of oncogenic HPV types. The
by unsafe injections, sexual contact, settings with a high prevalence of scope and effectiveness of
and blood transfusions. Perinatal HBV infection, the first dose should such vaccines has improved,
transmission from infected moth- be given to newborn babies as soon by expanding the range of
ers to their newborn babies or from as possible after birth, to prevent types covered and because of
one child to another is very common mother-to-child transmission. unforeseen cross-protection
in highly endemic areas, and HBV The introduction of HBV vacci- against related types.
can also be transmitted by fomites nation programmes has resulted in
■■ Nationwide HPV vaccination
[8]. HBV infection is a major global a decrease in the incidence of HBV of adolescent girls (in some
health problem. In 2016, an estimat- infection and hepatocellular carcino- cases, together with boys) in
ed 292 million people were chroni- ma [10] (see Chapter 5.6). In Taiwan, some countries is now recog-
cally infected with HBV, i.e. about China, where a nationwide HBV nized as offering, in combina-
3.9% (uncertainty interval, 3.4– vaccination programme for new- tion with cervical screening,
4.6%) of the world’s population [9]. born babies was started in 1983, the prospect of the elimination
Chronic HBV infection, through the proportion of children who were of cervical cancer as a public
persistent inflammation, liver necro- seropositive for HBsAg decreased health problem.
sis, and regenerative proliferation, from 10% before the vaccination
may eventually lead to cirrhosis and programme started to 0.5% in 2009 ■■ HPV vaccination can also pre-
hepatocellular carcinoma. About [13]. The reduction in prevalence vent a fraction of cases of can-
80% of hepatocellular carcinomas was accompanied by a 70% reduc- cer of the anus, vulva, vagina,
develop in cirrhotic livers. The risk tion in the incidence of liver cancer penis, and oropharynx.
of chronic HBV infection is greatest in children and adolescents [14].
514
Fig. 6.3.1. Attributable fraction (AF) for liver cancers due to (A) hepatitis B virus (HBV) and (B) hepatitis C virus (HCV).
In the USA, the incidence of acute trends in HBV vaccine coverage in Mali, Nigeria, Haiti, Guatemala,
HBV infection decreased by 81% be- all WHO regions (Fig. 6.3.3), with a Iraq, the Syrian Arab Republic, and
tween 1990 and 2006 [15]. major increase in coverage at the Papua New Guinea [9].
By 2016, 185 countries had in- beginning of the 21st century [4]. In The recommended introduction
troduced HBV vaccination, and 2016, vaccine coverage was still low of universal HBV vaccination of ba-
three-dose vaccination coverage (≤ 80%) in some high-risk popula- bies at birth has been successful in
CHAPTER 6.3
SECTION 6
in children had reached 87% glob- tions, such as in Kenya, the Central far fewer countries. In 2016, cover-
ally [9]. There have been favourable African Republic, Chad, Gabon, age of birth dose of HBV vaccine

Fig. 6.3.2. An eight-week-old baby is vaccinated against eight antigens, including hep- occurred in less-developed countries
atitis B virus (HBV), at the Madarounfa Health Centre in Niger. [3]. Multiple epidemiological studies
over the past three decades have
confirmed the carcinogenicity of 13
oncogenic types in cervical cancer
(HPV16, 18, 31, 33, 35, 39, 45, 51, 52,
56, 58, 59, and probably 68). Types
HPV16 and HPV18 are detectable in
about 70% of cervical cancers, with
little variation around the world [17].
Substantial fractions of other
cancer types, including cancers of
the anus (88%), vulva and vagina
(41%), penis (50%), and oropharynx
(30%), are also attributable to HPV,
nearly always due to type HPV16
[17]. The relative importance of HPV
in oropharyngeal cancer is much
greater in more-developed countries
in which the prevalence of tobacco
use has been declining. Non-cervical
cancers account for about 100 000
HPV-related cases per year globally.
The incidence of HPV-associated
was estimated to be 46% globally Human papillomaviruses cancers is especially high in immu-
and only 10% in sub-Saharan Africa
nodeficient individuals, especially
[9,16]. The United Nations includ- Human papillomaviruses anal cancer in HIV-positive men who
ed combating viral hepatitis in the and cancer have sex with men.
Sustainable Development Goals, HPV is a sexually transmitted infec- The natural history and molecu-
with the target of achieving 90% tion that is acquired by most wom- lar mechanisms involved in HPV
global coverage of birth dose by en and men shortly after the onset carcinogenesis are best understood
2030. Because of the increasing ef- of sexual activity. HPV infection is in the cervix [7]. The most common
ficacy and the decreasing cost of an- considered a necessary cause of morphological manifestation of HPV
tiviral treatments for HBV and HCV cervical cancer (see Chapter 5.10). infection consists of minor epithe-
infection, WHO also has an aim of In 2018, there were an estimated lial abnormalities (equivocal and low-
identifying and treating at least 80% 570 000 new cases of cervical can- grade cellular changes). In a minority
of chronic carriers of HBV and HCV cer and 311 000 deaths from cervi- of women (~10%) in whom the infec-
infections by 2030 [9]. cal cancer worldwide, 95% of which tion is not cleared by the immune
Fig. 6.3.3. Three-dose hepatitis B virus (HBV) vaccine coverage, by WHO region, 1990–2015.
100
90 African Region
80
Region of the Americas
70
60 Eastern Mediterranean Region
Coverage (%)
50
European Region
40
30 South-East Asia Region
20
Western Pacific Region
10
0 Global
1990 1995 20 0 0 2005 2010 2015
Year
516
system, precancerous lesions (ad- from 26 randomized controlled tri- (CIN3+) were consistent with those
vanced intraepithelial neoplasia) can als that included a placebo or other for CIN2+ [18].
develop. If these lesions are not treat- vaccine control arm and involved a Vaccines also prevented CIN2+
ed, they can lead to cervical cancer total of 73 428 women, mainly aged in HPV-negative women vaccinat-
after many years, usually decades. 15–26 years, with a follow-up of ed at age 24–45 years. However,
1.3–8 years. the protection against HPV16/18-
Human papillomavirus Vaccine efficacy and the corre- associated CIN2+ of all women,
vaccines sponding 95% confidence intervals regardless of baseline HPV status,
Three subunit vaccines against HPV (CIs) against cervical intraepithe- was weaker than that in younger
are currently available. All are com- lial neoplasia grade 2 and above women and was not statistically
posed of virus-like particles and are (CIN2+), and adenocarcinoma in significant (vaccine efficacy, 26%;
produced by expression of the HPV situ were evaluated by computing 95% CI, −5% to 48%); the lower fre-
L1 gene in insect cells or yeast. The risks in the vaccination group ver- quency of CIN2+ in this age group
bivalent vaccine is against HPV16 sus the control group separately was noted [18].
and HPV18. The quadrivalent vac- by women’s HPV DNA status, i.e. The risk of serious adverse
cine also includes HPV6 and HPV11, the presence of oncogenic HPV events, including autoimmune dis-
which are the cause of most genital infection at vaccination. In HPV- eases, was similar in the vaccinated
warts, and the more recent nonava- negative women aged 15–26 years, and control groups (relative risk,
lent vaccine also targets HPV31, 33, vaccines reduced the risk of CIN2+ 0.94; 95% CI, 0.72–1.06) [18]. Total
45, 52, and 58. associated with HPV16/18 from 164 death rates were similar (11 per
HPV vaccines also differ by the to 2 per 10 000 (vaccine efficacy, 10 000 in the control group and 14
adjuvant. An alum adjuvant is used 99%; 95% CI, 95–100%). Vaccine per 10 000 in the HPV vaccinated
in the quadrivalent and nonavalent efficacy was about 90% also for group), and no pattern in the cause
vaccines, and a complex adju- relatively rare adenocarcinoma in or timing of death was detected. In
vant system (ASO4) consisting of situ (Table 6.3.1). Among all young addition, HPV vaccines did not sig-
monophosphoryl lipid A and alum is women, regardless of baseline HPV nificantly increase the risk of miscar-
used in the bivalent vaccine. status, the risk of CIN2+ associated riage, pregnancy termination, con-
with HPV16/18 fell from 341 to 157 genital abnormality, or stillbirth [18].
Vaccine efficacy and safety per 10 000 (vaccine efficacy, 54%; The effectiveness [19] and safety
A systematic review [18] combined 95% CI, 43–63%). Reductions in [20] of HPV vaccines continue to
published and unpublished findings risk for the most severe precancer be monitored in many countries
Table 6.3.1. Efficacy of human papillomavirus (HPV) vaccines in women aged 15–26 years who were negative for oncogenic HPV
infection at vaccination
Outcome Anticipated absolute effectsa (95% CI) Vaccine efficacy (%) Number of
(95% CI) c participants
Risk with placebo Risk with HPV (number of studies)
(per 10 000) vaccination b
(per 10 000)
CIN2+ associated with HPV16/18 164 2 (0 to 8) 99 (95 to 100) 23 676
Follow-up: 3–5 years (3 RCTs)
CIN3+ associated with HPV16/18 70 0 (0 to 7) 99 (90 to 100) 20 214
AIS associated with HPV16/18 9 0 (0 to 7) 90 (18 to 99) 20 214
Any CIN2+ irrespective of HPV type, 287 106 (72 to 158) 63 (45 to 75) 25 180
bivalent or quadrivalent vaccine (5 RCTs)
Follow-up: 2–6 years
Any CIN3+ irrespective of HPV type, 109 23 (4 to 120) 79 (−10 to 96) 20 719
bivalent or quadrivalent vaccine (3 RCTs)
Follow-up: 3.5–4 years
Any AIS irrespective of HPV type 10 0 (0 to 8) 90 (24 to 99) 20 214
AIS, adenocarcinoma in situ; CI, confidence interval; CIN2+, cervical intraepithelial neoplasia grade 2 and above; HPV, human papillomavirus, RCTs,
randomized controlled trials.
a
The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its
CHAPTER 6.3
SECTION 6
95% CI). When risk in the vaccinated group is zero, the 95% CI is computed using an exact binomial method.
b
Assumed risk calculated from the sum of control group event rates.
c
Vaccine efficacy (%) = (1 − relative risk)*100.

through population-based surveil- document the long-term non-inferior- and boys who had missed a dose in
lance systems, ad hoc studies, and ity of one dose are under way [25]. school [28].
follow-up of trial participants. National programmes exist in
The most recently licensed no- Immunization rates many low-income countries in Latin
navalent HPV vaccine (not includ- Between 2006 and 2014, 64 coun- America, but not yet in India, China,
ed in the systematic review [18]) tries implemented national HPV and most countries in Africa [6].
was compared with the quadriva- vaccination programmes, but vac- Bhutan, Malaysia, and Rwanda
lent HPV vaccine in a randomized cine uptake varies widely across pioneered the implementation of
trial involving 14 215 women aged and within countries [6]. HPV vaccination [24] before GAVI
15–26 years. The nonavalent HPV Nearly all European countries started supporting HPV vaccine
vaccine prevented infection and offer HPV vaccination [26]. The in 2012. Since then, more than 30
precancers related to HPV31, 33, average time between first vac- GAVI-eligible countries have start-
45, 52, and 58 and generated an cine authorization and universal ed implementing vaccination [29]
antibody response to HPV6, 11, 16, mass vaccination was 36 months, and have achieved good levels of
and 18 that was non-inferior to that ranging from 5 months in Spain participation (> 70%) in the targeted
generated by the quadrivalent HPV to 117 months in Croatia. The tar- girls [30]. However, the GAVI target
vaccine [21]. In HPV-uninfected get age is generally 12–13 years, of vaccinating 40 million girls in the
women, the efficacy of the nonava- but some countries recommended lowest-income countries by 2020 is
lent vaccine against CIN2+ associ- starting at older ages or including considered to be at risk, because of
ated with the nine targeted onco- several birth cohorts in the first a slow ramp-up from demonstration
genic HPV types was 100% (95% rounds. Immunization rates ranged projects to national programmes
CI, 70.4–100%). from 14.1% in Bulgaria to 85.9% in and because of challenges with the
Both trial data [18] and popula- the United Kingdom. Coverage of supply of vaccines [5].
tion-based studies [22] demonstrate
less than 30% was reported in east-
that the bivalent vaccine induces
substantial and significant cross-
ern European countries, Greece, Conclusions
and France, but the accuracy of Despite the effectiveness and safe-
protection against HPV31, 33, and
vaccination monitoring also varies ty of HPV vaccines, anti-vaccina-
45 at least. Preliminary findings also
greatly in Europe [26]. tion campaigns and the relatively
suggest that the vaccines can pre-
In the USA, the HPV vaccines high cost, coupled with the delayed
vent HPV infection in the entire ano-
were recommended for girls in 2006 benefits of anti-cancer vaccines,
genital tract and in the mouth [23].
and for boys in 2011, but uptake hamper the universal implementa-
Doses has been slow compared with that tion of HPV vaccination. There are
for other adolescent vaccines [27]. projected to be 770 000 new cases
The initial recommendation for HPV
According to a nationwide database of cervical cancer per year by 2040
vaccination was a three-dose sched-
of medical billings, in 2014 cumula- [3]. To seize a unique opportunity to
ule for everybody. A significant de-
velopment to improve population tive vaccination coverage of one or tackle a major disease in women,
coverage was the endorsement by more doses by age 18 years was in 2018 the WHO Director-General,
WHO in 2014 of two-dose instead 53.3% in girls and 30.3% in boys. Dr Tedros Adhanom Ghebreyesus,
of three-dose schedules up to age Although coverage is still lower in made a call for coordinated global
15 years, supported by stronger boys, the ramp-up in vaccination in action against cervical cancer.
immune responses in children and boys was quicker than that in girls, Modelling studies are being
adolescents than in young women which indicates good acceptability. done to identify the best vaccina-
[24]. One-dose-only vaccination Immunization rates were found to be tion and screening strategy to elimi-
could greatly further augment the substantially affected by area of resi- nate cervical cancer as a public
feasibility and affordability of mass dence and type of health insurance. health problem [31]. The higher the
vaccination (see Chapter 4.4). The Vaccination at later than age 12 years pre-vaccination prevalence of HPV
earliest non-randomized evidence was frequent among girls in the USA, infection, the more difficult HPV
that one dose of vaccine could pro- and vaccination is administered by a elimination will be [32]. Fortunately,
vide durable protection against HPV variety of providers: paediatricians, if coverage is equal, herd protection
infection came from the Costa Rica family doctors, and gynaecologists. is predictably stronger against a
Vaccine Trial [25]. The antibody lev- In Australia, 80.1% of girls and sexually transmitted virus like HPV
els after one dose, although lower 74.1% of boys aged 15 years had than it is against airborne and food-
than the levels elicited by three dos- been fully vaccinated in 2015– borne infections [33].
es, were 9 times as high as the levels 2016, thanks to an especially strong In the absence of vaccination,
elicited by natural infection. A formal societal advocacy and a close inter- the prevalence of HPV16 infec-
randomized controlled trial and other action between school-based vac- tion may increase in populations
complementary studies to further cination and active recall of girls in less-developed countries, as a
518
result of the transition from tradi- and programme factors responsible from a population more rapidly, but it
tional to gender-similar age-related for vaccine delivery and completion. is less cost-effective than increasing
sexual behaviour, i.e. the sexual The implementation of additional the coverage or the number of birth
pattern that is most conducive to strategies to increase population- cohorts in girls [33].
rapid spread of the infection in level protection, such as vaccinating Because there are currently
young people. A prompt introduc- older women or men, would be de- only two manufacturers of vac-
tion of HPV vaccination before the pendent on greatly reduced vaccine cines, a shortage of vaccines is
transition of sexual behaviour would prices. Offering vaccination to multi- threatening the global action of
decrease the prevalence of HPV16 ple cohorts of girls, for example up WHO [5]. Therefore, more abun-
infection, whereas introduction to age 15 years or 18 years, is very dant availability of fair-priced HPV
of vaccination after the transition cost-effective, even at current vac- vaccines greatly depends on the
would mean that vaccination will cine prices, and accelerates cervical advent of new manufacturers in
take longer to decrease the preva- cancer prevention. Beyond a certain low- and middle-income countries.
lence of HPV16 infection by the age, vaccination remains attractive Multivalent vaccines are ideal, but
same amount (Fig. 6.3.4) [32]. but has limited return, because of bivalent vaccines would be wel-
Key factors to improve HPV vac- the age-related accumulation of per- come, because of the preponder-
cination coverage include educating sistent HPV infections whose fate ant role of HPV16/18 in the onset of
communities – including adoles- is not ameliorated by the vaccines HPV-associated cancer in the cer-
cents, families, and health workers – [34]. Gender-neutral vaccination is vix and at other sites [17].
and better addressing organizational highly desirable to eliminate HPV
Fig. 6.3.4. Expected variations of vaccination effectiveness according to pre-vaccination HPV prevalence in women and changes
in sexual behaviour. Changes in the prevalence of HPV16 among women aged 20–34 years in relation to the number of years
since the beginning of a population’s transition from traditional to gender-similar age-related sexual behaviour and the introduction
of vaccination among girls aged 11 years (with an assumption of 70% coverage) before and after the transition. The shaded area
shows an assumption of a 15-year transition period. The arrows show the approximate timing of the introduction of vaccination,
before or after the transition. Traditional sexual behaviour indicates a population in which genders have different age-specific sexual
activity rates and a wide gap in ages (e.g. an average of 5.6 years, as observed in India) of spouses or cohabitating sexual partners.
Gender-similar sexual behaviour indicates a population in which genders have similar age-specific sexual activity rates and a
narrow gap in ages (e.g. an average of 2.1 years, as observed in the USA) of spouses or cohabitating sexual partners.
CHAPTER 6.3
SECTION 6

References
1. Plummer M, de Martel C, Vignat J, Ferlay 11. de Martel C, Maucort-Boulch D, Plummer 21. Joura EA, Giuliano AR, Iversen OE,
J, Bray F, Franceschi S (2016). Global M, Franceschi S (2015). World-wide Bouchard C, Mao C, Mehlsen J, et al.;
burden of cancers attributable to infec- relative contribution of hepatitis B and Broad Spectrum HPV Vaccine Study
tions in 2012: a synthetic analysis. Lancet C viruses in hepatocellular carcinoma. (2015). A 9-valent HPV vaccine against
Glob Health. 4(9):e609–16. https://doi. Hepatology. 62(4):1190–200. https://doi. infection and intraepithelial neoplasia in
org /10.1016/S2 214 -10 9X(16) 3 014 3 -7 org/10.1002/hep.27969 PMID:26146815 women. N Engl J Med. 372(8):711–23.
PMID:27470177 https://doi.org/10.1056/NEJMoa1405044
12. WHO (2009). Hepatitis B vaccines: WHO PMID:25693011
2. Maucort-Boulch D, de Martel C, Franceschi position paper. Wkly Epidemiol Rec.
S, Plummer M (2018). Fraction and in- 84(40):405–19. PMID:19817017 22. Kavanagh K, Pollock KG, Cuschieri K,
cidence of liver cancer attributable to Palmer T, Cameron RL, Watt C, et al.
hepatitis B and C viruses worldwide. Int 13. Ni YH, Chang MH, Wu JF, Hsu HY, Chen (2017). Changes in the prevalence of hu-
J Cancer. 142(12):2471–7. https://doi. HL, Chen DS (2012). Minimization of hepa- man papillomavirus following a national
org/10.1002/ijc.31280 PMID:29388206 titis B infection by a 25-year universal vac- bivalent human papillomavirus vaccina-
cination program. J Hepatol. 57(4):730–5. tion programme in Scotland: a 7-year
3. Ferlay J, Ervik M, Lam F, Colombet M, Mery
https://doi.org/10.1016/j.jhep.2012.05.021 cross-sectional study. Lancet Infect Dis.
L, Piñeros M, et al. (2018). Global Cancer
Observatory. Lyon, France: International
Agency for Research on Cancer. Available 14. Chang MH, You SL, Chen CJ, Liu CJ, S1473-3099(17)30468-1 PMID:28965955
from: https://gco.iarc.fr/. Lee CM, Lin SM, et al.; Taiwan Hepatoma 23. Beachler DC, Jenkins G, Safaeian M,
4. WHO (2017). Global hepatitis report Study Group (2009). Decreased incidence Kreimer AR, Wentzensen N (2016). Natural
2017. Geneva, Switzerland: World Health of hepatocellular carcinoma in hepatitis B acquired immunity against subsequent
Organization. Available from: https:// vaccinees: a 20-year follow-up study. J Natl genital human papillomavirus infection:
w w w.w h o. i nt / h e p a t i t i s /p u b l i c a t i o n s / Cancer Inst. 101(19):1348–55. https://doi. a systematic review and meta-analysis.
global-hepatitis-report2017/en/. org/10.1093/jnci/djp288 PMID:19759364 J Infect Dis. 213(9):1444–54. https://doi.
org/10.1093/infdis/jiv753 PMID:26690341
5. WHO (2018). Meeting of the Strategic 15. Wasley A, Grytdal S, Gallagher K; Centers
Advisory Group of Experts on Immuniza for Disease Control and Prevention (CDC) 24. Brotherton JM, Jit M, Gravitt PE, Brisson
tion, October 2018 – conclusions and (2008). Surveillance for acute viral hepati- M, Kreimer AR, Pai SI, et al. (2016).
recommendations. Wkly Epidemiol Rec. tis – United States, 2006. MMWR Surveill Eurogin Roadmap 2015: how has HPV
93(49):661–79. Available from: https:// Summ. 57(2):1–24. PMID:18354374 knowledge changed our practice: vac-
apps.who.int/iris/handle/10665/276545. cines. Int J Cancer. 139(3):510–7. https://
16. Li X, Dumolard L, Patel M, Gacic-Dobo
doi.org/10.1002/ijc.30063 PMID:26916230
6. Bruni L, Diaz M, Barrionuevo-Rosas L, M, Hennessey K (2018). Implementation
Herrero R, Bray F, Bosch FX, et al. (2016). of hepatitis B birth dose vaccination – 25. Kreimer AR, Herrero R, Sampson JN,
Global estimates of human papillomavi- worldwide, 2016. Wkly Epidemiol Rec. Porras C, Lowy DR, Schiller JT, et al.;
rus vaccination coverage by region and 93(7):61–72. PMID:29450989 Costa Rica HPV Vaccine Trial (CVT)
income level: a pooled analysis. Lancet Group (2018). Evidence for single-dose
Glob Health. 4(7):e453–63. https://doi. 17. de Martel C, Plummer M, Vignat J,
protection by the bivalent HPV vaccine –
org /10.1016/S2 214 -10 9X(16) 3 0 0 9 9 -7 Franceschi S (2017). Worldwide burden
review of the Costa Rica HPV vac-
PMID:27340003 of cancer attributable to HPV by site,
cine trial and future research studies.
country and HPV type. Int J Cancer.
7. Schiffman M, Doorbar J, Wentzensen N, Vaccine. 36(32 Pt A):4774–82. https://
141(4):664–70. https://doi.org/10.1002/ijc.
de Sanjosé S, Fakhry C, Monk BJ, et al. doi.org/10.1016/j.vaccine.2017.12.078
30716 PMID:28369882
(2016). Carcinogenic human papillomavi- PMID:29366703
rus infection. Nat Rev Dis Primers. 2:16086. 18. Arbyn M, Xu L, Simoens C, Martin-Hirsch
26. Sheikh S, Biundo E, Courcier S, Damm O,
ht tps: //doi.org /10.103 8 /nrdp. 2016. 8 6 PP (2018). Prophylactic vaccination against
Launay O, Maes E, et al. (2018). A report
PMID:27905473 human papillomaviruses to prevent cervi-
on the status of vaccination in Europe.
cal cancer and its precursors. Cochrane
8. IARC (2012). Biological agents. IARC Vaccine. 36(33):4979–92. https://doi.
Database Syst Rev. (5):CD009069. https://
Monogr Eval Carcinog Risks Hum. 100B:1– o r g / 10 .10 16 / j .v a c c i n e . 2 0 18 . 0 6 . 0 4 4
doi.org/10.1002/14651858.CD009069.
441. Available from: http://publications.iarc. PMID:30037416
pub3 PMID:29740819
fr/119 PMID:23189750
27. Gargano JW, Zhou F, Stokley S, Markowitz
9. Razavi-Shearer D, Gamkrelidze I, Nguyen 19. Drolet M, Bénard É, Boily MC, Ali H, LE (2018). Human papillomavirus vacci-
MH, Chen D-S, Van Damme P, Abbas Z, Baandrup L, Bauer H, et al. (2015). nation in commercially-insured vaccine-
et al.; Polaris Observatory Collaborators Population-level impact and herd ef- eligible males and females, United States,
(2018). Global prevalence, treatment, fects following human papillomavirus 2007–2014. Vaccine. 36(23):3381–6. https://
and prevention of hepatitis B virus infec- vaccination programmes: a system- doi.org/10.1016/j.vaccine.2018.03.045
tion in 2016: a modelling study. Lancet atic review and meta-analysis. Lancet PMID:29735321
Gastroenterol Hepatol. 3(6):383–403. Infect Dis. 15(5):565–80. https://doi.
https://doi.org/10.1016/S2468-1253(18) org /10.1016 /S1473 - 3 0 9 9 (14)71073 - 4 28. Brotherton JM, Winch KL, Bicknell L,
30056-6 PMID:29599078 PMID:25744474 Chappell G, Saville M (2017). HPV vaccine
coverage is increasing in Australia. Med J
10. Trépo C, Chan HL, Lok A (2014). Hepatitis 20. WHO (2017). Human papillomavirus vac- Aust. 206(6):262. https://doi.org/10.5694/
B virus infection. Lancet. 384(9959):2053– cines: WHO position paper, May 2017. mja16.00958 PMID:28359009
63. https://doi.org/10.1016/S0140-6736(14) Wkly Epidemiol Rec. 92(19):241–68.
60220-8 PMID:24954675 PMID:28530369
520
29. GAVI (2018). Gavi welcomes call for 31. Brisson M, Bénard É, Drolet M, Bogaards 33. Baussano I, Lazzarato F, Ronco G,
coordinated global action against cer- JA, Baussano I, Vänskä S, et al. (2016). Franceschi S (2018). Impacts of human
vical cancer. 19 May 2018. Available Population-level impact, herd immunity, papillomavirus vaccination for differ-
from: https://www.gavi.org/library/news/ and elimination after human papillomavi- ent populations: a modeling study. Int
statements/2018/gavi-welcomes- call- rus vaccination: a systematic review and J Cancer. 143(5):1086–92. https://doi.
for-coordinated-global-action-against- meta-analysis of predictions from trans- org/10.1002/ijc.31409 PMID:29603224
cervical-cancer/. mission-dynamic models. Lancet Public 34. Hildesheim A, Gonzalez P, Kreimer AR,
Health. 1(1):e8–17. https://doi.org/10.1016/ Wacholder S, Schussler J, Rodriguez AC,
30. Gallagher KE, Howard N, Kabakama S, S2468-2667(16)30001-9 PMID:29253379 et al.; Costa Rica HPV Vaccine Trial (CVT)
Mounier-Jack S, Griffiths UK, Feletto
Group (2016). Impact of human papil-
M, et al. (2017). Lessons learnt from hu- 32. Baussano I, Lazzarato F, Brisson M, lomavirus (HPV) 16 and 18 vaccination
man papillomavirus (HPV) vaccination Franceschi S (2016). Human papillomavi- on prevalent infections and rates of cer-
in 45 low- and middle-income countries. rus vaccination at a time of changing sexu- vical lesions after excisional treatment.
PLoS One. 12(6):e0177773. https:// al behavior. Emerg Infect Dis. 22(1):18–23. Am J Obstet Gynecol. 215(2):212.e1–15.
doi.org/10.1371/journal.pone.0177773 https://doi.org/10.3201/eid2201.150791 https://doi.org/10.1016/j.ajog.2016.02.021
PMID:28575074 PMID:26691673 PMID:26892991
CHAPTER 6.3
SECTION 6

6.4
Preventive therapy
Certain interventions clearly
established
Jack Cuzick Karen Brown (reviewer)
Diona L. Damian (reviewer)
prevent cancer is still in its infancy. [3]. Four prevention trials have sub-
SUMMARY This partly reflects the fact that sequently confirmed this finding
cancers are more heterogeneous in high-risk women without breast
●● For women at high risk of breast and biologically complex than car- cancer (Table 6.4.1).
cancer, reductions of 30–70% diovascular disease, and the causal Overall, these trials show a 38%
in the incidence of breast can- pathways are less well understood. reduction in breast cancer incidence
cer can be achieved with use of Good biomarkers for identifying in- [4], as a result of a 50% reduction
anti-estrogenic agents. dividuals at increased risk of spe- for estrogen receptor (ER)-positive
●● Widespread use of low-dose as- cific cancer types are also missing, breast cancers but no effect for ER-
pirin for 10 years between ages and even less is known about fac- negative tumours. Two of these tri-
50 years and 65 years could tors that are predictive of response als with long-term follow-up have
have a major impact on cancer to specific treatments. shown that the protection persists
incidence and mortality. Interventions have been divided long after stopping use of the medi-
into four groups: those for which cation [5,6]. In the International
●● Many other agents, including there is good evidence of efficacy, Breast Cancer Intervention Study I
some medicines used for other those with findings that are promis- (IBIS-I) trial, 5 years of treatment
purposes and some food coming but not fully convincing, those with tamoxifen resulted in a greater
ponents, seem promising for for which there is a substantial reduction in the incidence of breast
cancer prevention but have not amount of evidence of no benefit, cancer in the 10–20-year follow-up
been fully evaluated in humans. and those for which there is good period than in the first 10 years of
evidence of harm. follow-up (cumulative risk, 4.6% vs
●● Good short-term biomarkers
for response to treatment are 6.3% in years 0–10, 3.3% vs 6.3%
needed to efficiently evaluate Breast cancer in years 10–20) (Fig. 6.4.1).
new agents. Breast cancer prevention has been The two major side-effects of
facilitated by the fact that studies of tamoxifen are endometrial can-
the treatment of existing cancers cer and venous thromboembolism
Cancer prevention is a large field can also provide reliable evidence (Table 6.4.2). Endometrial cancer
comprising lifestyle changes to re- for a preventive effect on new tu- was increased in postmenopaus-
duce risk, screening interventions mours in the contralateral breast. al women by about 2.5-fold above
to detect early lesions, and preven- the baseline rate of about 60 per
tive interventions aimed at more ac- Tamoxifen 100 000 per year at age 60 years,
tively interrupting the carcinogenic The Cancer Research Campaign II whereas venous thromboembolism
pathway. Although tobacco use is (CRC-II) trial provided the first occurred about twice as often in
clearly the strongest known avoid- evidence of a preventive effect for the tamoxifen arm compared with
able cause of cancer [1], only thera- tamoxifen [2]. A subsequent meta- placebo. Less serious but more
peutic interventions to reduce risk analysis of 20 randomized clinical common side-effects of tamoxifen
are considered in this chapter. trials of 5 years of treatment with include vasomotor symptoms such
Compared with cardiovascu- tamoxifen as adjuvant therapy in as hot flushes and night sweats,
lar disease, for which preventive about 15 000 women overall docu- and gynaecological symptoms such
treatments are firmly established, mented a reduction of about one as bleeding and uterine polyps.
the development of therapies to third in contralateral breast tumours Topical formulations of tamoxifen
522
Table 6.4.1. Established therapeutic agents for cancer prevention

Agent Type of study Relevant Number Key findings FUNDAMENTALS
references evaluated
Aspirin RCTs and [17,18,21] 69 224 in RCTs 7–10% reduction in all- ■■ “Preventive therapy” has
observational 52 926 in case– cancer incidence and been widely adopted as
studies control studies 9–12% in mortality for
10-year use. Mostly for
an appropriately focused
colorectal, stomach, and term to address many
oesophageal cancer (30% interventions, anticipated or
each), with smaller and
less certain reductions for
established, once referred to
breast, prostate, and lung as “chemoprevention”.
cancer (5–15%)
■■ Over recent decades, evi-
Oral Meta-analysis of [15] > 20 000 cases 27% reduction for ovarian
contraceptives 17 case–control cancer for any use; > 50% dence of preventive benefit
studies and 7 reduction for > 10-year from clinical trials has largely
cohort studies use
involved drugs rather than mi-
Anti-estrogenic compounds for breast cancer prevention cronutrients or supplements.
Selective estrogen-receptor modulators
■■ Confidence to introduce
Tamoxifen 4 RCTs in women [5,6] 28 193 33% reduction for all
at high risk breast cancer, based on preventive therapy is markedly
44% reduction for ER+ increased if relevant mecha-
invasive and no effect for
nistic data are available.
ER− breast cancer
Raloxifene 3 RCTs (1 vs [7,8] 37 296 34% reduction overall, ■■ Some reduction in risk of can-
tamoxifen in with 56% reduction for cer associated with consump-
women at ER+ invasive breast
high risk) cancer; 25% less effective tion of fruits and vegetables is
than tamoxifen in direct evident, but the relative impact
comparison in women at of particular food items has not
high risk
become clear.
Lasofoxifene 1 RCT in women [4,9] 8856 79% reduction for all
with osteoporosis breast cancer for higher ■■ Particular micronutrients
dose; 18% reduction for
lower dose and supplements for cancer
prevention have been subject
Arzoxifene 1 RCT in women [4] 9354 58% reduction for all
with osteoporosis breast cancer; 70% to large clinical trials, but few,
reduction for ER+ breast if any, benefits have emerged,
cancer
and evidence of increased risk
Aromatase inhibitors has accrued.
Anastrozole 1 RCT in women [13] 3864 53% reduction for all
at high risk; breast cancer; 58% ■■ When available, integration of
contralateral reduction for ER+ invasive preventive therapy involving
tumours in RCTs breast cancer
in adjuvant
particular drugs with other
setting initiatives including screening
Exemestane 1 RCT in women [12] 4560 53% reduction for all represents the broadest basis
at high risk; RCT breast cancer; 73% for reducing cancer incidence.
of contralateral reduction for ER+ invasive
tumours in breast cancer
adjuvant setting
ER, estrogen receptor; RCTs, randomized controlled trials. oxifene than for tamoxifen [7], but
a direct comparison in the Study of
metabolites applied directly to the ated. Raloxifene is a second-gener- Tamoxifen and Raloxifene (STAR)
breast are now under study, with ation SERM originally developed to trial found that the reduction in
the hope that the local dose will be prevent osteoporosis in postmeno- breast cancer risk for raloxifene
high enough to maintain its preven- pausal women. It has estrogenic was about 25% less than that for ta-
tive effects but the reduction in sys- effects on bone and lipid metabo- moxifen [8]. However, no excess of
temic dose will limit its side-effects. lism, and anti-estrogenic effects on endometrial cancer or other gynae-
the endometrium and breast tissue. cological problems were observed,
Other selective estrogen- Because of this tissue selectivity, and raloxifene may be more ac-
receptor modulators raloxifene has fewer side-effects ceptable than tamoxifen for post-
CHAPTER 6.4
SECTION 6
The effects of three other selective than tamoxifen. menopausal women, because it is
ER modulators (SERMs) on breast Trials in women with osteoporo- already widely used to treat and
cancer risk have now been evalu- sis suggested larger effects for ral- prevent osteoporosis.
Chapter 6.4 • Preventive therapy 523

Fig. 6.4.1. Long-term effect of tamoxifen treatment on cumulative incidence of breast cancer over time in the International Breast
Cancer Intervention Study I (IBIS-I) trial. Solid lines indicate all breast cancers, and dashed lines indicate invasive estrogen receptor
(ER)-positive breast cancers.
Table 6.4.2. Potential common or major side-effects of pharmacological agents considered for cancer prevention
Agent Side-effect Findings
Tamoxifen/SERMs Endometrial cancer 2–3-fold increase, except with raloxifene
Venous thromboembolic 73% increase overall; smaller increase with raloxifene

events
Vasomotor symptoms 20% increase during treatment; no effect subsequently
Aromatase inhibitors Bone fractures 50% increase in adjuvant trials without baseline bone density scan; non-
significant 11% increase in prevention studies with baseline identification and
treatment of women with low bone density
Musculoskeletal symptoms, Increase from 58% in placebo to 64% with anastrozole (10% relative increase)
arthralgia
Carpal tunnel syndrome 3.6-fold increase in adjuvant setting vs tamoxifen (3% vs 1%); 58% increase in
prevention setting (3% vs 2%)
Vasomotor symptoms 15% increase overall; 20% increase in severe symptoms
Vaginal dryness, dyspareunia, 20% increase in prevention setting vs placebo (19% vs 16%); 3-fold increase
loss of libido in adjuvant setting vs tamoxifen (1% vs 0.3%)
LHRH agonist Bone loss, menopausal 7% loss in bone mineral density at lumbar spine; substantial increase in
symptoms vasomotor symptoms
Aspirin/NSAIDs Gastrointestinal bleeding Increase of ~50%, mostly in initial period after starting treatment
Haemorrhagic stroke 35% increase, but larger reduction in occlusive stokes; net decrease in
incidence, but increase in fatal events
LHRH, luteinizing hormone-releasing hormone; NSAIDs, non-steroidal anti-inflammatory drugs; SERMs, selective estrogen-receptor modulators.
524
Two other SERMs – lasofox- bined (primary end-point) was seen and no data are available on reduc-
ifene [4,9] and arzoxifene [4] – have (hazard ratio, 0.47; 95% confidence tion in cancer risk.
been investigated again in post- interval, 0.32–0.68; P < 0.0001) None of these agents have had
menopausal women with osteo- [13], which was similar to the results an effect on ER-negative breast can-
porosis, with reduction in the inci- reported in the MAP3 trial. For ER- cer; this remains an unmet need.
dence of fractures as the primary positive invasive cancer, the reduc-
end-point. For both agents, a re- tion was 58%, but – as in the MAP3 Ovarian cancer
duction in the incidence of breast trial – no effect was found for ER-
cancer was found that was larger Although no randomized trials have
negative breast cancer. Vasomotor
than has been seen for tamoxifen been conducted, there is clear
and musculoskeletal side-effects
(Table 6.4.1). Lasofoxifene was also evidence from case–control and
were increased with both agents,
associated with reductions in the in- cohort studies that use of oral con-
but only by 10–15%, and these ad-
cidence of fractures, heart disease, traceptives has a large protective
verse events were also reported
and strokes [9], suggesting that it effect for ovarian cancer. In an over-
by many women who received pla-
could be an ideal preventive agent, view of 24 such studies, Havrilesky
cebo (64% for anastrozole vs 58%
but the manufacturer is not pursu- et al. demonstrated a 27% reduc-
for placebo in the IBIS-II trial). This
ing the licensing of this drug for any tion for any use and a 57% reduc-
illustrates the need to have a place-
of these indications. tion for more than 10 years of use
bo arm to accurately assess subjec-
[15]. Oral contraceptives have im-
tive side-effects. Blinded long-term
Aromatase inhibitors pacts on other cancer types, includ-
follow-up is continuing in IBIS-II,
The third-generation aromatase ining a small increase in the risk of
so that the long-term efficacy and
hibitors anastrozole, letrozole, and breast cancer and cervical cancer
side-effects of anastrozole can be
exemestane have all been found to but larger reductions in the risk of
evaluated.
be more effective than tamoxifen endometrial cancer and colorectal
Overall, the reported reductions
for the treatment of ER-positive cancer [16].
in breast cancer incidence for both
breast cancer in postmenopausal
exemestane and anastrozole were
women [10] and are now routinely
larger than those seen for tamox- Aspirin and other
used for this indication. In these tri- non-steroidal anti-
ifen or raloxifene, and indicate that
als, the incidence of contralateral
these two drugs are attractive op- inflammatory drugs
breast tumours, a good surrogate
tions for breast cancer prevention There is now overwhelming evi-
for new cancers, was also reduced
in postmenopausal women at high dence for a reduction of about one
by a further 50% compared with ta-
risk. Although both SERMs and third in colorectal cancer incidence
moxifen [11].
aromatase inhibitors increase men- and mortality from long-term regu-
Two large breast cancer preven-
opausal symptoms, SERMs also lar aspirin use [17]. Beneficial ef-
tion trials have reported on the use
increase the incidence of endome- fects of a similar size have been
of aromatase inhibitors in high-risk
women without breast cancer. The trial cancer and thromboembolic seen for oesophageal cancer and
Mammary Prevention 3 (MAP3) trial events, whereas aromatase inhibi- stomach cancer, and smaller, less
randomized 4560 postmenopaus- tors increase the incidence of frac- convincing reductions of 5–15%
al women to either exemestane or tures and musculoskeletal symp- have recently also been found for
placebo for 5 years and found a 65% toms (Table 6.4.2). The fracture risk lung cancer, breast cancer, and
reduction in invasive breast cancers seems to be largely controlled by prostate cancer (Table 6.4.3) [18],
[12]. No reduction was observed for a baseline dual-energy X-ray ab- but there appears to be little or no
ER-negative disease, but the effect sorptiometry (DEXA) scan and use effect on other major cancer sites.
on ER-positive disease was even of bisphosphonates in women with Long-term use of about 10 years
greater than the overall effect (haz- low bone density [13]. was estimated to reduce overall
ard ratio, 0.27; 95% confidence inter- For premenopausal women, the cancer incidence by about 9% in
val, 0.12–0.60; P < 0.001). However, only well-studied option remains men and 7% in women, and over-
these conclusions are limited by the tamoxifen, although a small ran- all cancer mortality by 13% in men
short median follow-up period of domized trial in 75 women has ex- and 9% in women (Table 6.4.3) [18].
35 months. amined a combination of the lutein- The relative impact appears to be
The IBIS-II trial compared anas- izing hormone-releasing hormone similar between the sexes, but the
trozole with placebo in 3864 post- (LHRH) agonist goserelin with ral- overall effects are greater for men
menopausal women at increased oxifene in women at high risk [14]. because these cancer types are
risk of breast cancer. After a me- A 4.7% absolute reduction in breast relatively more common in men.
CHAPTER 6.4
SECTION 6
dian follow-up of 5 years, a 53% density was seen after 2 years of Gastrointestinal and cerebral
reduction in invasive breast cancer treatment, but this was not main- bleeding are the most important
and ductal carcinoma in situ com- tained after treatment completion, harms associated with aspirin use,

Table 6.4.3. Estimated impact of 10 years of aspirin use for individuals aged 50– is also found in other fruits and veg-
70 years on percentage reduction in 15-year incidence and 20-year mortality from six etables. Both sulforaphane and ly-
cancer sites affected by aspirin use, and for all cancers separately for men and women copene have been linked to reduced
risk of prostate cancer [26,27].
Site Reduction in incidence (%) Reduction in mortality (%)
Several spices have also been
Colorectum 35 40 put forward as having protective ef-
fects. Curcumin, which comes from
Oesophagus 30 50
turmeric, has been the most studied,
Stomach 30 35 but there is still very limited evidence
in humans for cancer prevention
Breast 10 5
[28]. Of the many hundreds of other
Prostate 10 15 compounds that have been studied
Lung 5 15
[29], those that have received the
most attention are resveratrol (which
Overall – men 9 13 is found mostly in red wine and ber-
Overall – women 7 9 ries) [30], green tea polyphenols
[31], and pomegranate juice [32],
but again convincing evidence of ef-
and their risk and fatality rate in- The impact of aspirin use on ficacy in humans is lacking.
crease with age [19]. Use of prophy- cancer mortality appears to be larg- Reports on vitamin D with or
lactic aspirin in the general popula- er than that for incidence, suggest- without calcium are very mixed,
tion aged 50–65 years is likely to ing an anti-metastatic effect as well with no compelling evidence for
be beneficial when the reduction as a separate effect on incidence benefit at any cancer site [33].
in risk of cancer and cardiovascu- [22,23]. The mechanisms that me-
lar disease and the risk of excess diate these effects are currently not Agents that have not
bleeding are all considered. The established, and trials are under worked
benefit–risk ratio is highly favour- way to examine aspirin as an adju-
Epidemiological and laboratory evi-
able for the general population for vant treatment for individuals with
dence suggested a potential anti-can-
both men and women and is about colorectal, stomach, oesophageal,
cer effect of vitamin A, β-carotene,
5:1 for serious events and at least breast, and prostate cancer [24].
and their analogues. Despite ran-
7:1 for deaths [18]. Markers that Data on other non-steroidal an-
domized evidence of a benefit of
identify individuals most likely to ti-inflammatory drugs, such as ibu-
β-carotene, vitamin E, and sele-
benefit would enable treatment to profen, sulindac, or celecoxib, are
nium in a severely deficient popula-
be more focused, and this is a cur- less extensive, and there are no tri-
tion in Linxian, China [34], subse-
rent research priority. The United als with long-term follow-up, except
quent studies in Europe and North
States Preventive Services Task for studies of colorectal adenomas.
America have been negative. Two
Force currently supports the use of However, observational studies have
large studies of β-carotene in heavy
aspirin for those at increased risk of found similar overall effects on can-
smokers and in workers exposed to
cardiovascular disease or colorec- cer incidence [22].
asbestos found that it actually led to
tal cancer [20]. increases in the incidence of lung
The effects of daily use of as- Other agents cancer [35,36], and one found an
pirin on cancer incidence are not Vaccination against human papil- increase in all-cause mortality [35].
apparent until at least 3 years after lomavirus (HPV) has proven to be An overview of all randomized trials
the start of use (Fig. 6.4.2), with a highly effective in reducing precur- of β-carotene confirmed an increase
relative reduction in incidence after sor lesions for cervical cancer and in the incidence of lung cancer and
that time for all cancers of about is very likely to prevent cervical also found an increase in the inci-
24% [21]. Some benefits appear to cancer and other HPV-related can- dence of stomach cancer but no sig-
be sustained for several years after cers (see Chapter 6.3). nificant effect on other cancer types,
treatment cessation in long-term us- Many studies have suggested a either individually or overall [37].
ers. Relative reductions in cancer in- protective effect of consumption of Vitamin E and selenium were
cidence appear to be similar in men fruits and vegetables, with a strong- thought to have a beneficial effect on
and women [21], although data are er effect for vegetables [25]. Specific prostate cancer, on the basis of labo-
less extensive for women and men potentially active components in- ratory and epidemiological studies
have a higher incidence of the can- clude sulforaphane, which is found [38], but randomized trials have been
cer types for which the incidence is in cruciferous vegetables, and lyco- negative. In particular, neither sele-
reduced by aspirin use, leading to pene, which is found at particularly nium nor vitamin E supplementation
greater absolute reductions. high levels in cooked tomatoes but reduced the incidence of prostate
526
Fig. 6.4.2. Impact of aspirin use on cancer mortality by scheduled duration of treatment.
cancer in the Selenium and Vitamin both the Prostate Cancer Prevention stands out as having the largest po-
E Cancer Prevention Trial (SELECT), Trial (PCPT) [40], which investigat- tential impact on the population at
in which prostate cancer was the pri- ed finasteride, and the Reduction large. This is because it has a major
mary end-point, and Klein et al. [39] by Dutasteride of Prostate Cancer effect on three common gastrointes-
reported that the incidence of pros- Events (REDUCE) trial [41], which tinal cancer types – colorectal, stom-
tate cancer increased by 17% with assessed dutasteride. ach, and oesophageal cancer – and
vitamin E supplementation. Other There has also been much inter- potentially provides small reductions
studies have not shown any effects est in the role of statins for cancer in three other major cancer types:
of supplementation on the incidence prevention, but the overall evidence lung, breast, and prostate cancer.
of prostate cancer, colorectal cancer, is largely negative [42]. However, questions still remain about
or cause-specific mortality. aspirin’s optimal dose, duration, effi-
The use of 5α-reductase inhibi- cacy, safety, and impact on different
tors either for prevention or for man-
Conclusions and subtypes of specific cancers, and
agement of early prostate cancer has challenges more research is needed.
produced complex outcomes, with Despite its early stage of develop- In terms of relative overall impor-
CHAPTER 6.4
SECTION 6
substantial reductions in disease of ment, important discoveries have tance for cancer prevention, tobacco
low Gleason grade but an apparent already been made for preventive cessation remains the most impor-
increase in high-grade cancers in therapy. Of these, low-dose aspirin tant factor. Parkin et al. estimated that

tobacco use is responsible for 19% of ifen in women at high risk of breast only been reported more recently,
all new cancer cases but calculated cancer is only 10–20% [43], and largely because they were not ap-
that no other activity is responsible much of this low uptake is due to parent until after 3–5 years of as-
for more than 10% of cancers [1]. The a lack of knowledge and interest in pirin use. Education of both health
estimate that 7–10% of cancers could prevention from health profession- professionals and the general pub-
be avoided with daily use of low-dose als. Aspirin has had earlier recom- lic about the benefits of therapeutic
aspirin for 10 years between ages mendations from professional bod- prevention needs to be a major goal
50 years and 65 years, with a larger ies against using it in the general and activity.
reduction of 9–13% for mortality [18], population [44]. However, those Also, activities to promote pre-
makes this a key element of any can- recommendations were based on ventive therapy need to be inte-
cer prevention strategy. comparing cardiovascular benefits grated with those to encourage a
However, several major chal- with risks of bleeding, and now healthy lifestyle. Neither of these
lenges remain. Key among these need to be updated in view of the alone will eliminate cancer, and
is to find ways to encourage more much larger benefits seen for can- adoption of one does not preclude
widespread use of agents with es- cer prevention than for cardiovas- the need for the other.
tablished utility. Uptake of tamox- cular disease. These benefits have
References
1. Parkin DM, Boyd L, Walker LC (2011). 16. 7. Barrett-Connor E, Mosca L, Collins P, 13. Cuzick J, Sestak I, Forbes JF, Dowsett M,
The fraction of cancer attributable to life- Geiger MJ, Grady D, Kornitzer M, et al.; Knox J, Cawthorn S, et al.; IBIS-II investi-
style and environmental factors in the UK Raloxifene Use for The Heart (RUTH) Trial gators (2014). Anastrozole for prevention of
in 2010. Br J Cancer. 105(Suppl 2):S77– Investigators (2006). Effects of raloxifene breast cancer in high-risk postmenopausal
81. https://doi.org/10.1038/bjc.2011.489 on cardiovascular events and breast cancer women (IBIS-II): an international, double-
PMID:22158327 in postmenopausal women. N Engl J Med. blind, randomised placebo-controlled
355(2):125–37. https://doi.org/10.1056/ trial. Lancet. 383(9922):1041–8. https://
2. Cuzick J, Baum M (1985). Tamoxifen NEJMoa062462 PMID:16837676 doi.org/10.1016/S0140-6736(13)62292-8
and contralateral breast cancer. Lancet. PMID:24333009
8. Vogel VG, Costantino JP, Wickerham DL,
2(8449):282. https://doi.org/10.1016/S0140-
Cronin WM, Cecchini RS, Atkins JN, et 14. Howell A, Ashcroft L, Fallowfield L,
6736(85)90338-1 PMID:2862460
al.; National Surgical Adjuvant Breast and Eccles DM, Eeles RA, Ward A, et al.
3. Early Breast Cancer Trialists’ Collaborative Bowel Project (2010). Update of the National
(2018). RAZOR: a phase II open ran-
Group (EBCTCG) (2005). Effects of che- Surgical Adjuvant Breast and Bowel Project
domized trial of screening plus goserelin
motherapy and hormonal therapy for early Study of Tamoxifen and Raloxifene (STAR)
and raloxifene versus screening alone
breast cancer on recurrence and 15-year P-2 trial: preventing breast cancer. Cancer
in premenopausal women at increased
survival: an overview of the randomised Prev Res (Phila). 3(6):696–706. https://doi.
org/10.1158/1940 - 6207.CAPR-10 - 0076 risk of breast cancer. Cancer Epidemiol
trials. Lancet. 365(9472):1687–717. https:// Biomarkers Prev. 27(1):58–66. https://
PMID:20404000
doi.org/10.1016/S0140-6736(05)66544-0 doi.org/10.1158/1055-9965.EPI-17-0158
PMID:15894097 9. Cummings SR, Ensrud K, Delmas PD, PMID:29097444
LaCroix AZ, Vukicevic S, Reid DM, et
4. Cuzick J, Sestak I, Bonanni B, Costantino 15. Havrilesky LJ, Moorman PG, Lowery WJ,
al.; PEARL Study Investigators (2010).
JP, Cummings S, DeCensi A, et al.; Lasofoxifene in postmenopausal wom- Gierisch JM, Coeytaux RR, Urrutia RP,
SERM Chemoprevention of Breast en with osteoporosis. N Engl J Med. et al. (2013). Oral contraceptive pills as
Cancer Overview Group (2013). Selective 362(8):686–96. https://doi.org/10.1056/ primary prevention for ovarian cancer:
oestrogen receptor modulators in preven- NEJMoa0808692 PMID:20181970 a systematic review and meta-analysis.
tion of breast cancer: an updated meta- Obstet Gynecol. 122(1):139–47. https://
analysis of individual participant data. 10. Dowsett M, Cuzick J, Ingle J, Coates doi.org/10.1097/AOG.0b013e318291c235
Lancet. 381(9880):1827–34. https://doi. A, Forbes J, Bliss J, et al. (2010). Meta- PMID:23743450
org /10.1016/S 0140 - 673 6(13) 6 0140 - 3 analysis of breast cancer outcomes in ad-
PMID:23639488 juvant trials of aromatase inhibitors versus 16. Gierisch JM, Coeytaux RR, Urrutia RP,
tamoxifen. J Clin Oncol. 28(3):509–18. Havrilesky LJ, Moorman PG, Lowery WJ, et
5. Powles TJ, Ashley S, Tidy A, Smith IE, https://doi.org/10.1200/JCO.2009.23.1274 al. (2013). Oral contraceptive use and risk
Dowsett M (2007). Twenty-year follow- PMID:19949017 of breast, cervical, colorectal, and endome-
up of the Royal Marsden randomized, trial cancers: a systematic review. Cancer
11. Cuzick J (2005). Aromatase inhibitors for
double-blinded tamoxifen breast cancer Epidemiol Biomarkers Prev. 22(11):1931–
breast cancer prevention. J Clin Oncol.
prevention trial. J Natl Cancer Inst. 99(4): 43. https://doi.org/10.1158/1055-9965.EPI-
23(8):1636–43. https://doi.org/10.1200/
283–90. https://doi.org/10.1093/jnci/djk050 JCO.2005.11.027 PMID:15755971 13-0298 PMID:24014598
PMID:17312305
12. Goss PE, Ingle JN, Alés-Martínez JE, 17. Rothwell PM, Fowkes FG, Belch JF,
6. Cuzick J, Sestak I, Cawthorn S, Hamed H, Cheung AM, Chlebowski RT, Wactawski- Ogawa H, Warlow CP, Meade TW (2011).
Holli K, Howell A, et al.; IBIS-I Investigators Wende J, et al.; NCIC CTG MAP.3 Effect of daily aspirin on long-term risk
(2015). Tamoxifen for prevention of breast Study Investigators (2011). Exemestane of death due to cancer: analysis of in-
cancer: extended long-term follow-up of the for breast-cancer prevention in post- dividual patient data from randomised
IBIS-I breast cancer prevention trial. Lancet menopausal women. N Engl J Med. trials. Lancet. 377(9759):31–41. https://
Oncol. 16(1):67–75. https://doi.org/10.1016/ 364(25):2381–91. https://doi.org/10.1056/ doi.org/10.1016/S0140-6736(10)62110-1
S1470-2045(14)71171-4 PMID:25497694 NEJMoa1103507 PMID:21639806 PMID:21144578
528
18. Cuzick J, Thorat MA, Bosetti C, Brown 28. Devassy JG, Nwachukwu ID, Jones PJ 37. Druesne-Pecollo N, Latino-Martel P, Norat
PH, Burn J, Cook NR, et al. (2015). (2015). Curcumin and cancer: barriers T, Barrandon E, Bertrais S, Galan P, et al.
Estimates of benefits and harms of to obtaining a health claim. Nutr Rev. 73 (2010). Beta-carotene supplementation
prophylactic use of aspirin in the gen- (3):155–65. https://doi.org/10.1093/nutrit/ and cancer risk: a systematic review and
eral population. Ann Oncol. 26(1):47–57. nuu064 PMID:26024538 metaanalysis of randomized controlled tri-
https://doi.org/10.1093/annonc/mdu225 als. Int J Cancer. 127(1):172–84. https://
PMID:25096604 29. Hackshaw-McGeagh LE, Perry RE, doi.org/10.1002/ijc.25008 PMID:19876916
Leach VA, Qandil S, Jeffreys M, Martin
19. Thorat MA, Cuzick J (2015). Prophylactic RM, et al. (2015). A systematic review of 38. Pak RW, Lanteri VJ, Scheuch JR,
use of aspirin: systematic review of harms dietary, nutritional, and physical activity Sawczuk IS (2002). Review of vitamin E
and approaches to mitigation in the gen- interventions for the prevention of pros- and selenium in the prevention of pros-
eral population. Eur J Epidemiol. 30(1):5– tate cancer progression and mortality. tate cancer: implications of the Selenium
18. https://doi.org/10.1007/s10654-014- and Vitamin E Chemoprevention Trial.
Cancer Causes Control. 26(11):1521–50.
9971-7 PMID:25421783 Integr Cancer Ther. 1(4):338–44. https://
ht t p s: //d o i.o r g /10 .10 07/s10 5 52- 015 -
20. Chubak J, Whitlock EP, Williams SB, 0659-4 PMID:26354897 d o i .o r g /10 .117 7/15 3 47 3 5 4 0 2 2 3 818 6
Kamineni A, Burda BU, Buist DSM, et al. PMID:14664728
30. Gescher A, Steward WP, Brown K (2013).
(2016). Aspirin for the prevention of can-
Resveratrol in the management of hu- 39. Klein EA, Thompson IM Jr, Tangen CM,
cer incidence and mortality: systematic
man cancer: how strong is the clinical Crowley JJ, Lucia MS, Goodman PJ, et
evidence reviews for the U.S. Preventive
Services Task Force. Ann Intern Med. evidence? Ann N Y Acad Sci. 1290(1): al. (2011). Vitamin E and the risk of pros-
164(12):814–25. https://doi.org/10.7326/ 12–20. https://doi.org/10.1111/nyas.12205 tate cancer: the Selenium and Vitamin E
M15-2117 PMID:27064482 PMID:23855461 Cancer Prevention Trial (SELECT). JAMA.
306(14):1549–56. https://doi.org/10.1001/
21. Rothwell PM, Price JF, Fowkes FG, 31. Li MJ, Yin YC, Wang J, Jiang YF (2014). jama.2011.1437 PMID:21990298
Zanchetti A, Roncaglioni MC, Tognoni Green tea compounds in breast cancer pre-
G, et al. (2012). Short-term effects of vention and treatment. World J Clin Oncol. 40. Thompson IM Jr, Goodman PJ, Tangen
daily aspirin on cancer incidence, mor- 5(3):520–8. https://doi.org/10.5306/wjco. CM, Parnes HL, Minasian LM, Godley PA,
tality, and non-vascular death: analy- v5.i3.520 PMID:25114865 et al. (2013). Long-term survival of partici-
sis of the time course of risks and ben- pants in the Prostate Cancer Prevention
efits in 51 randomised controlled trials. 32. Malik A, Afaq F, Sarfaraz S, Adhami VM, Trial. N Engl J Med. 369(7):603–10.
Lancet. 379(9826):1602–12. https://doi. Syed DN, Mukhtar H (2005). Pomegranate https://doi.org/10.1056/NEJMoa1215932
org /10.1016/S 014 0 - 673 6 (11) 61720 - 0 fruit juice for chemoprevention and che- PMID:23944298
PMID:22440946 motherapy of prostate cancer. Proc
Natl Acad Sci U S A. 102(41):14813–8. 41. Andriole GL, Bostwick DG, Brawley OW,
22. Chan AT, Ogino S, Fuchs CS (2009). https://doi.org/10.1073/pnas.0505870102 Gomella LG, Marberger M, Montorsi F, et
Aspirin use and survival after diagnosis of PMID:16192356 al.; REDUCE Study Group (2010). Effect
colorectal cancer. JAMA. 302(6):649–58. of dutasteride on the risk of prostate can-
https://doi.org/10.1001/jama.2009.1112 33. Bjelakovic G, Gluud LL, Nikolova D, cer. N Engl J Med. 362(13):1192–202.
PMID:19671906 Whitfield K, Krstic G, Wetterslev J, et al. https://doi.org/10.1056/NEJMoa0908127
(2014). Vitamin D supplementation for PMID:20357281
23. Holmes MD, Chen WY, Li L, Hertzmark prevention of cancer in adults. Cochrane
E, Spiegelman D, Hankinson SE (2010). Database Syst Rev. (6):CD007469. 42. Emberson JR, Kearney PM, Blackwell
Aspirin intake and survival after breast L, Newman C, Reith C, Bhala N, et al.;
h t t p s : / / d o i . o r g / 1 0 .1 0 0 2 / 1 4 6 51 8 5 8 .
cancer. J Clin Oncol. 28(9):1467–72. Cholesterol Treatment Trialists’ (CTT)
CD007469.pub2 PMID:24953955
https://doi.org/10.1200/JCO.2009.22.7918 Collaboration (2012). Lack of effect of
PMID:20159825 34. Blot WJ, Li JY, Taylor PR, Guo W, Dawsey lowering LDL cholesterol on cancer:
24. Coyle C, Cafferty FH, Rowley S, S, Wang GQ, et al. (1993). Nutrition in- meta-analysis of individual data from
MacKenzie M, Berkman L, Gupta S, et tervention trials in Linxian, China: sup- 175,000 people in 27 randomised trials
al.; Add-Aspirin investigators (2016). plementation with specific vitamin/min- of statin therapy. PLoS One. 7(1):e29849.
ADD-ASPIRIN: a phase III, double-blind, eral combinations, cancer incidence, PMID:22276132
placebo controlled, randomised trial as- and disease-specific mortality in the
general population. J Natl Cancer Inst. 43. Hackett J, Thorneloe R, Side L, Wolf M,
sessing the effects of aspirin on disease
recurrence and survival after primary 85(18):1483–92. https://doi.org/10.1093/ Horne R, Cuzick J, et al. (2018). Uptake
therapy in common non-metastatic solid jnci/85.18.1483 PMID:8360931 of breast cancer preventive therapy in the
tumours. Contemp Clin Trials. 51:56–64. UK: results from a multicentre prospec-
35. Omenn GS, Goodman GE, Thornquist tive survey and qualitative interviews.
https://doi.org/10.1016/j.cct.2016.10.004
PMID:27777129 MD, Balmes J, Cullen MR, Glass A, et Breast Cancer Res Treat. 170(3):633–40.
al. (1996). Effects of a combination of ht t p s: //d o i.o r g /10.10 07/s10 5 4 9 - 018 -
25. Negri E, La Vecchia C, Franceschi beta carotene and vitamin A on lung 4775-1 PMID:29687178
S, D’Avanzo B, Parazzini F (1991). cancer and cardiovascular disease. N
Vegetable and fruit consumption and Engl J Med. 334(18):1150–5. https://doi. 44. Thun MJ, Jacobs EJ, Patrono C (2012).
cancer risk. Int J Cancer. 48(3):350–4. org/10.105 6/ NEJM19 9 6 050233 418 02 The role of aspirin in cancer preven-
https://doi.org/10.1002/ijc.2910480307 PMID:8602180 tion. Nat Rev Clin Oncol. 9(5):259–67.
PMID:2040528 https://doi.org/10.1038/nrclinonc.2011.199
36. Albanes D, Heinonen OP, Taylor PR, PMID:22473097
26. Juge N, Mithen RF, Traka M (2007). Virtamo J, Edwards BK, Rautalahti M, et
Molecular basis for chemoprevention by al. (1996). Alpha-tocopherol and beta-
sulforaphane: a comprehensive review. carotene supplements and lung cancer
Cell Mol Life Sci. 64(9):1105–27. https:// incidence in the Alpha-Tocopherol, Beta-
d o i.o r g /10.10 07/s 0 0 018 - 0 07- 6 4 8 4 - 5 Carotene Cancer Prevention Study: ef-
PMID:17396224
fects of base-line characteristics and
27. Chen P, Zhang W, Wang X, Zhao K, Negi study compliance. J Natl Cancer Inst.
DS, Zhuo L, et al. (2015). Lycopene and 88(21):1560–70. https://doi.org/10.1093/
jnci/88.21.1560 PMID:8901854
CHAPTER 6.4
SECTION 6
risk of prostate cancer: a systematic review

and meta-analysis. Medicine (Baltimore).
94(33):e1260. https://doi.org/10.1097/MD.
0000000000001260 PMID:26287411

6.5
Managing people with high and

moderate genetic risk
Genomic tools to promote effective
cancer risk reduction
Patricia Ashton-Prolla Paul Brennan (reviewer) Mieke Van Hemelrijck (reviewer)
Jeffrey N. Weitzel Ian S. Fentiman (reviewer)
Hereditary cancer is caused pre- phylactic surgery [4]), there are still
SUMMARY dominantly by one (or, rarely, more health system, ethnic, and socioeco-
than one) moderately or highly pen- nomic disparities in access to risk
●● The identification of individu- etrant pathogenic or likely pathogenic assessment, especially in low- and
als and families with hereditary (P/LP) germline variant in a cancer middle-income countries [5].
cancer is an important opportu- predisposition gene (see Chapter The phenotypic effect is hetero-
nity for cancer prevention. 3.2). The identification of individuals geneous for most variants associ-
●● Cancer risk-reducing interven- and families with hereditary cancer is ated with hereditary cancer. Genetic
tions (e.g. lifestyle changes, an important opportunity for cancer variants can be classified according
enhanced surveillance, che- prevention [1]. to their frequency and the associ-
moprevention, and prophylac- About 5–10% of all solid tumours ated risk of cancer (Fig. 6.5.1). In ad-
tic surgery) are available, and and haematological malignancies dition, there are geographical, popu-
are associated with P/LP germline lation-derived differences in variant
identification of the causative
variants in cancer predisposition type and frequency among different
germline genetic variant is key
genes [2,3]. Carriers of such vari- regions of the world, and founder
to the development of rational
ants have significantly higher risks mutations account for a substantial
management guidelines ac-
of developing multiple cancer types, fraction of the cancer burden in cer-
cording to specific cancer risks.
often at an early age, compared with tain regions (Table 6.5.2). Therefore,
●● Recent advances in diagnos- the general population. Therefore, characterization of the mutational
tic tools using multigene panel these cases contribute a significant landscape of cancer predisposition
testing have enabled the simul- proportion of the cancer burden genes and variant penetrance is of
taneous and more affordable worldwide, given that lifetime cancer great importance.
analysis of multiple cancer pre- risks in these individuals may reach Although specific cancer predis-
disposition genes. up to 80% (e.g. for hereditary breast position syndromes are clearly iden-
and ovarian cancer syndrome) or tified (Table 6.5.1), phenotypic over-
●● Health system, ethnic, and so- even close to 100% (e.g. for Li– lap exists among several of them
cioeconomic disparities in ac- Fraumeni syndrome). (Fig. 6.5.2) [5]. Most P/LP germline
cess to risk assessment still Genetic/genomic cancer risk as- variants are inherited, but in some
exist, especially in low- and sessment (GCRA) is standard-of- cancer predisposition syndromes,
middle-income countries, and care medical practice that uses such as Li–Fraumeni syndrome and
add to the complexity of en- genetic and genomic tools to iden- familial adenomatous polyposis, de
abling universal access to this tify individuals and families with in- novo mutations in TP53 and APC,
important strategy to reduce creased risk of cancer. This enables respectively, have been described in
the global cancer burden. early and frequent screening to de- 5–10% of affected patients [6]. As tu-
These disparities must be ad- tect smaller, more curable cancers, mour genetic testing becomes more
dressed to ensure that all ben- and to propose cancer prevention common, previously unrecognized
efits of incorporating genetic measures (Box 6.5.1). Despite such germline mutations will be detected,
or genomic information into an high risks and the availability of can- and the percentage of cancers ac-
individual’s clinical care are at- cer risk-reducing interventions (e.g. counted for by high or moderate
tained at a global level. lifestyle changes, enhanced surveil- genetic risk as a result of de novo
lance, chemoprevention, and pro- mutations will be better understood.
530
Box 6.5.1. Components of genetic/genomic cancer risk assessment (GCRA), indi-

cating the most common features of pre- and post-test counselling.
FUNDAMENTALS
Pre-test counselling ■■ About 5–10% of all solid tumours
and haematological malignan-
• Initial assessment and engagement with patient. cies are associated with inherited
• Document patient and family history of cancer; perform physical ex- predisposition from moderately
amination whenever necessary. or highly penetrant pathogenic or
likely pathogenic germline vari-
• Assess psychosocial and interpersonal dynamics (communication ants. An estimated 1.7 million new
within family) and support system; discuss cultural beliefs. cases of hereditary cancer were
diagnosed worldwide in 2018.
• Discuss basic principles of cancer genetics (including medical, genetic,
and technical information); describe features of hereditary cancer syn- ■■ There is significant heterogeneity
dromes, and consider factors that limit interpretation and assessment. in cancer risks associated with
pathogenic or likely pathogenic
• Assess mutation probabilities and empirical cancer risks. variants in cancer predisposition
• Discuss genetic testing process, potential test outcomes, cost, turn- genes, and management differs
according to the penetrance of
around time, and insurance coverage. each variant. Phenotypic overlap
• Develop genetic testing strategies and facilitate informed consent, is common and is observed for
and assess and address psychological, cultural, communication, and multiple genes.
ethical issues. Anticipate potential cancer risk management options ■■ Advances in sequencing technol-
according to test result. ogy have enhanced knowledge
about the genes and pathogenic
• Ensure protection of anonymity, privacy, and confidentiality, and fa- or likely pathogenic variants as-
cilitate communicating genetic information to at-risk family members sociated with hereditary cancer
and/or medical caregivers. and have increased access to
• Discuss alternatives to genetic testing. more affordable and comprehen-
sive genetic testing.
• Anticipate increasing referrals from patients participating in direct-
■■ Options are available for cancer
to-consumer testing schemes and those with a possible germline risk reduction in carriers, includ-
mutation detected on tumour testing. ing lifestyle changes, enhanced
surveillance, chemoprevention,
Post-test counselling and risk-reduction surgery.
• Disclose, interpret, and communicate test results. However, evidence on the ef-
ficacy and cost–effectiveness
• Address psychological and ethical concerns. of these interventions has been
• Identify at-risk family members who would also benefit from genetic generated only for high-pene-
testing and/or increased screening or preventive care. trance genes such as BRCA1 and
BRCA2, and most guidelines cite
• Discuss communication of results to at-risk family members (strate- inadequate evidence for some
gies, resources, and barriers). or all aspects of management for
moderate-risk gene variants, for
• Arrange contacts and resources for patient and at-risk family members. which more work is needed to
• Develop personalized risk management plan by applying evidence- calibrate risks and interventions.
based guidelines. ■■ Genetic/genomic cancer risk as-
• Propose empirical risk screening and prevention recommendations sessment is a standard-of-care
in setting of uninformative genetic test results. multidisciplinary process that
ideally involves genetic counsel-
• Identify research options or clinical trials appropriate to patients and ling, experienced cancer risk
at-risk family members when applicable. consultants, and medical/surgi-
cal risk management teams.
• Summarize and disseminate personalized risk management plan
with patient and patient-authorized care providers. ■■ Despite major advances in the
field, the remaining challenges
include difficulties in the breadth
of variants and their curation,
Recent advances in diagnostic ing turnaround times and a higher
limited accuracy of the associated
tools using multigene panel testing yield in the identification of disease- risk estimation and establishment
have enabled the identification of causing variants. Despite these ad- of clinical utility, limited access
previously unidentified genotype– vances, important challenges arise to professional genetic counsel-
phenotype relationships and the from multigene panel testing. These ling and testing, and the need
for professional education about
CHAPTER 6.5
SECTION 6
simultaneous and more affordable include uncertain clinical action-

genetics and genomics and train-
analysis of multiple cancer predis- ability of P/LP variants identified in ing of multidisciplinary teams.
position genes, with shorter test- genes with moderate penetrance,
Chapter 6.5 • Managing people with high and moderate genetic risk 531
Fig. 6.5.1. Genetic risk categories in hereditary cancer and their characterization according to penetrance, actionability, screening
and management recommendations, and implications for family members. Clinical utility increases with higher cancer risk
predisposition; the gradient in the arrow denotes the potential significant overlap between the categories. Odds ratios are presented
as estimates of the generalized odds over the baseline population for organ-specific cancer risk. More studies, especially on genes
in the low- and moderate-risk categories, are needed to better clarify the associated cancer risks and penetrance. It is important to
note that penetrance and expressivity can widely vary with specific mutations within the same gene. HBOC syndrome, hereditary
breast and ovarian cancer syndrome; LFS, Li–Fraumeni syndrome; OR, odds ratio.
Penetrance: high; causes a well-known cancer syndrome with Examples:

well-defined cancer risks. OR: ≥ 5.0. Allele frequency: very low BRCA1/BRCA2
to low. Actionability: high; evidence-based risk-reducing and HBOC
High risk guidelines exist for at least one organ system (i.e. salpingo- syndrome;
oophorectomy for BRCA1/BRCA2 carriers, colectomy for APC TP53 and LFS.
Clinical utility
carriers). Implications for other family members: well defined.
Penetrance: moderate; organ-specific cancer risks are defined Examples:

Moderate for at least one cancer site. OR: < 5.0 and ≥ 2.0. Allele frequency: ATM and breast
risk intermediate. Actionability: moderate; limited evidence for cancer; CHEK2
enhanced surveillance for carriers. Implications for other family and colon
members: challenging, may be difficult to determine. cancer.
Penetrance: low or uncertain; vague organ-specific cancer risks. Examples:

OR: < 2.0 and ≥ 1.0. Allele frequency: high to very high. MRE11A and
Low risk Actionability: low, due to lack of established evidence-based NBN and breast
guidelines. Screening and management recommendations: based cancer;
on empirical risk estimates and case-by-case literature and GALNT12 and
laboratory data review. Implications for other family members: colon cancer.
poorly defined.
or in newly identified or very rare cancer incidence statistics for solid Considering the prevalence of
genes for which validation stud- tumours and assuming that 10% are hereditary cancer and its effects in
ies are required, and the pervasive hereditary, an estimated 1.7 million terms of cancer risks and preven-
conundrum of variants of uncertain new cases of hereditary cancer were tion, it is noteworthy that only a few
significance. Another complication diagnosed worldwide in 2018 (http:// countries address this issue in their
is that causality of observed moder- gco.iarc.fr). The identification of can- strategic plans for cancer control.
ate- or low-risk variants is difficult cer patients with genetic predisposi- In the World Cancer Declaration
to infer, and many are simply an tion can influence oncological man-
Progress Report 2016, only
incidental finding with respect to a agement and can direct screening
Bermuda, France, and Greece for-
given patient’s history of cancer [7]. and prevention strategies to ame-
mally included GCRA among their
liorate the risk of second primary tu-
strategies to reduce cancer risks
mours. Critically, cascade testing of
Scope of the preventive relatives for a familial mutation has (https://www.uicc.org/wcd-report).
approach in cancer the greatest potential to enhance Nationwide guidelines and/or gov-
genomics cancer prevention and improve the ernment programmes of GCRA
Moderately or highly penetrant P/LP cost–benefit ratio for society, as well and genetic testing have been es-
germline variants have been de- as enable the avoidance of cancer tablished in some countries, includ-
scribed in individuals with most, if mortality and treatment-related mor- ing Canada, France, the United
not all, tumour types. On the basis of bidity for individuals [8,9]. Kingdom, and the USA [10].
532
Table 6.5.1. Common cancer predisposition syndromes, associated genes, and phenotype
Hereditary cancer Associated Most commonly associated tumours Additional/distinctive findings

syndrome gene(s)
Ataxia telangiectasia ATM Leukaemia, breast cancer, pancreatic Ataxia, telangiectasias, recessive
cancer inheritance; female carriers at increased
risk of breast cancer
Cowden syndrome PTEN Breast cancer, thyroid cancer, colorectal Macrocephaly, Lhermitte–Duclos
cancer, endometrial cancer disease, acral keratosis, trichilemmomas,
papillomatous papules; developmental
delay and/or autism spectrum disorders
Familial adenomatous APC Colorectal cancer, pancreatic cancer, Osteomas, dental abnormalities, congenital
polyposis, attenuated gastric cancer, thyroid cancer, desmoid hypertrophy of the retinal pigment
and classic forms tumours, tumours of the central nervous epithelium, benign cutaneous lesions
system, hepatoblastoma
Gorlin syndrome PTCH Basal cell carcinoma, medulloblastoma, Pits in palms and soles, macrocephaly
ovarian tumours and prominent forehead keratocystic
odontogenic tumours, cardiac and ovarian
fibromas, calcified falx cerebri
Li–Fraumeni TP53 Breast cancer (ER+ and HER2+),

syndrome sarcomas, tumours of the central nervous
system, adrenocortical carcinoma
Hereditary breast BRCA1/BRCA2 Breast cancer, ovarian cancer, prostate Biallelic mutations in BRCA2 cause
and ovarian cancer cancer, melanoma, pancreatic cancer Fanconi syndrome
syndrome
BRIP1, RAD51C, Ovarian cancer
RAD51D
BRCA1, BRCA2, Triple-negative breast cancer Includes both high-risk and moderate-risk
PALB2, RAD51D? genes
ATM, BRCA1, Male breast cancer

BRCA2, CHEK2,
PALB2
Hereditary diffuse CDH1 Diffuse gastric cancer, lobular breast

gastric cancer cancer, colorectal cancer
Juvenile polyposis BMPR1A, SMAD4 Colorectal and small intestine cancer, Hamartomatous polyps
syndrome pancreatic cancer, gastric cancer
Lynch syndrome MLH1, MSH2, Colon and small intestine, gastric,

MSH6, PMS2, hepatobiliary, endometrial, ovarian,
EPCAM pancreatic, and ureteral tumours
Melanoma– CDKN2, CDK4 Pancreatic cancer, melanoma

pancreatic cancer
syndrome
Multiple endocrine MEN1 Pancreatic cancer, pituitary and parathyroid Familial isolated hyperparathyroidism,
neoplasia type 1 tumours, well-differentiated endocrine facial angiofibromas, collagenomas,
(MEN1) tumours of the gastroenteropancreatic lipomas, meningiomas, ependymomas,
tract, carcinoid and adrenal tumours leiomyomas
Multiple endocrine RET Medullary thyroid carcinoma, Mucocutaneous neuromas, gastrointestinal

neoplasia type 2 pheochromocytoma, benign thyroid symptoms, muscular hypotonia, Marfanoid
(MEN2) tumours habitus
MUTYH-associated MUTYH Colorectal (polyps) and small intestine Recessive inheritance; carriers may be at
polyposis (MAP) cancers increased risk of colon cancer
Peutz–Jeghers STK11 Colon cancer (polyps), breast cancer, Hyperpigmented lesions, Peutz–Jeghers
syndrome ovarian cancer, pancreatic cancer, gastric polyps
cancer, endometrial and cervical carcinoma
Von Hippel–Lindau VHL Haemangioblastoma, clear cell renal Retinal angiomas; renal, pancreatic, and
syndrome cell carcinoma, pheochromocytoma, genital cysts
CHAPTER 6.5
SECTION 6
endolymphatic sac tumours
ER, estrogen receptor; HER2, human epidermal growth factor receptor 2.
Differences between disease, and access to periodic need for a skilled workforce include
countries in effective cancer screening methods, predic- programmes such as the American
implementation of GCRA tive genetic testing, and appropriate Society of Clinical Oncology (ASCO)
therapeutic options remains limited University curricula and the Cancer
In high-income countries, prog- (see Chapter 1.3). Therefore, the Genomics Education Program at City
ress in the detection and diagno- proportion of potentially curable tu- of Hope (https://www.cityofhope.org/
sis of hereditary cancer in the past mours at diagnosis is decreased in education/health-professional-edu
20 years has been enormous and lower-income countries, especially cation/cancer-genomics-education-
well documented [5]. Important disin patients who rely solely on public program).
coveries about the biology of he- health-care systems. The incidence Hereditary cancer syndromes
reditary cancers have resulted in and mortality rates of several cancer also have distinct clinical patterns
efforts to increase the awareness types (e.g. breast cancer and endo- and distribution globally. One of the
and education of both the general metrial cancer) have been correlated main factors to explain such dif-
public and health-care profession- with the Human Development Index ferences is the occurrence of spe-
als, as well as the discovery of (HDI) level of a country; decreases cific founder mutations at higher
targeted treatments for hereditary in mortality-to-incidence ratios have frequency in certain geographical
cancers, and even the development been observed with increments in regions or populations, leading to
of public policies. These, in turn, HDI level, probably resulting from large clusters of specific inherited
have enabled early – and often pre- better access to cancer screening, cancers (Table 6.5.2), in addition
symptomatic – detection of carriers, diagnosis, and treatment [11,12]. to the geographical differences in
prompt and effective intervention, In addition to regional economic cancer detection rate, registration,
and thus effective reduction of the and social constraints, an important diagnosis, prevention initiatives,
cancer burden in families with he- barrier is lack of awareness of he- and management [13–16]. Only a
reditary cancer. reditary cancer and of the cancer few national cancer institutions in
However, this progress has prevention opportunities that result low- and middle-income countries
reached only part of the world’s pop- from proper GCRA. An additional have formulated coordinated pro-
ulation. In many countries, diagno- challenge is the limited availabili- grammes towards the identification
ses still rely on overt clinical signs, ty of GCRA practitioners. GCRA and management of patients with
which appear late in the course of training resources to address the inherited cancers.
Table 6.5.2. Hereditary cancer in different populations: examples of founder mutations identified in selected cancer predisposition
genes
Continent Country Gene Mutation(s) Details Reference
Africa Algeria, Morocco, BRCA1 c.798_799delTT 22% of BRCA1 mutations in North [35]
and Tunisia African families
South Africa BRCA1 c.1374delC, c.2641G>T 77.8% of mutation carriers had 1 of [36]
BRCA2 c.7934delG the 3 Afrikaner founder mutations
North America Canada MSH2 c.942+3A>T 27% of Lynch syndrome mutations in [37]
families from Newfoundland
Mexico BRCA1 Exon 9–12 del 35% and 29% of the BRCA- [38]
associated cases of ovarian cancer
and breast cancer, respectively
South America Brazil TP53 c.1010G>A 70–80% of TP53 mutations in patients [13]
with breast cancer in Brazil
Asia China MSH2 c.1452_1455del 21% of Lynch syndrome mutations in [37]
Guangdong
Palestine TP53 c.541G>A 2% of women with early-onset breast [39]
cancer in one study
Europe Denmark MLH1 c.1667+2_1667+8del7ins4 25% of Lynch syndrome mutations in [37]
Amsterdam
Portugal BRCA2 c.156_157insAlu 37.9% of BRCA2 pathogenic variants [40]
in Portuguese families
Germany VHL c.T292C, c.T334C Observed in families from the Black [41]
Forest and east central regions
Oceania Australia BRCA1 c.3331_3334del Also recurrent in Hispanic [15]
populations, Europe, the United
Kingdom, and the USA
Australia BRCA2 c.6275_6276del Also recurrent in the United Kingdom, [15]
Belgium, Spain, the Netherlands, and
North America
534
Models for implementa- In integrated GCRA, consulta- tary cancer remain undetected; this
tion of GCRA tions and/or referrals are embedded is the main argument in favour of
in pathology services or oncology proposing population-based genet-
The high cost of diagnostic cancer
practices. For example, a reflexive ic testing for high-risk variants, re-
gene sequencing has historically
statement may be included in the pa- gardless of clinical or pathological
been a barrier to access, although
thology report (e.g. BRCA mutation features. However, robust evidence
costs associated with newer meth-
testing is suggested in all high-grade for the cost–effectiveness and fea-
odologies (high-throughput mas-
serous ovarian cancers) and/or elec- sibility of such practices is still lack-
sively parallel or next-generation
tronic medical record. Genetic coun- ing outside populations with a high
sequencing) are decreasing signifi-
sellors may be embedded in oncol- frequency of founder mutations in
cantly. Intrinsic to the identification
ogy clinics to identify eligible patients hereditary cancer. Finally, there are
and management of patients with
and provide point-of-service GCRA numerous initiatives to increase the
hereditary cancer is the complexity
of interpreting and communicating during oncology visits [19,20]. low rate of cascade testing to iden-
the genetic information as well as These established models are tify at-risk relatives [21–23].
the desirability for a multidisciplinary being challenged by (i) the practice
shift from single-gene testing to mul-
approach in patient care. Different
tigene panel testing, thus increasing
Management of patients
models have been proposed to pro-
the complexity of risk interpretation
with hereditary cancer
vide comprehensive GCRA, includ-
and communication; (ii) the expan- In the past two decades, major ad-
ing the following three.
sion of direct-to-consumer genetic vances have occurred in the diagno-
Integrated nationwide reference
testing platforms, which offer, but sis and management of patients with
centres provide GCRA, genetic test-
ing, and long-term management of do not require, genetic counselling; hereditary cancer [5,7]. The advent
patients and families with heredi- and (iii) the increasing identifica- and decreasing costs of next-gen-
tary cancer in specialized networks. tion of germline hereditary cancer eration sequencing have resulted in
There are examples of such net- predisposition through genomic tu- the development of multigene panel
works in the United Kingdom (within mour analysis, which provides un- testing and a significant expansion
specialist genetic services of the solicited and previously unsuspect- of knowledge about the genes in-
National Health System), France (in ed diagnoses of hereditary cancer. volved and the degree of phenotypic
the national health-care system), Furthermore, the availability of tar- overlap among them (Fig. 6.5.2). In
and Canada (http://ocp.cancercare. geted drug therapies has created addition, incidental (unrelated to the
on.ca/cms/One.aspx?portalId=7751 a medical necessity and increased respective phenotype) but clinically
5&pageId=10051) [17]. the frequency of germline genetic important findings have become in-
The community of practice ap- testing in a treatment-focused con- creasingly common, such as those
proach relies on the collaboration text; however, inclusion of personal identified through tumour genetic
of academic centres with commu- and family counselling has not been testing and those that do not match
nity-based providers in practice emphasized. the clinical picture (e.g. the pres-
networks, leveraging their practice Whatever the approach, several ence of a P/LP germline variant in a
with the experience and the multi- studies indicate that genetic testing mismatch repair gene in a proband
disciplinary nature of academic pro- rates are still low and a significant with no personal and/or family his-
grammes [5,18,19]. proportion of patients with heredi- tory of colon cancer) [24–26].
Fig. 6.5.2. Phenotypic overlap of solid tumours observed in association with germline pathogenic or likely pathogenic variants in
cancer predisposition genes. SNPs, single-nucleotide polymorphisms.
Ovarian
Pancreatic Prostate RAD51D
PRSS1 HOXB13 BRIP1
MMR PRS (SNPs) BRCA1 TP53
CDKN2A MRE11A BRCA2 ATM
Melanoma Colon BARD1 PALB2
BRCA2 STK11
RAD51C Breast
PALB2 BRCA1
ATM BRCA2 CHEK2
MMR MMR CDH1
CHEK2 EPCAM PTEN
CHAPTER 6.5
SECTION 6
Breast Breast
Uterine
Compared with sequential sin- sive Cancer Network change every first cancer diagnosis, and less often
gle-gene testing, multigene panel year in response to this dynamic. met established testing criteria for
testing is more efficient in identify- Furthermore, the identification of a Li–Fraumeni syndrome, compared
ing a P/LP variant, less expensive, P/LP variant in a moderate-pene- with those identified by single-gene
and faster, and it also identifies trance gene may not necessarily testing [27]. These findings are likely
variants in intermediate-penetrance be associated with causality of the to result in a revision of the pheno-
(moderate- or low-risk) genes. For phenotype that motivated testing. type and genetic testing criteria for
these reasons, next-generation se- In these situations, a critical review Li–Fraumeni syndrome.
quencing is often regarded as an of results of genetic testing in light The definition of the penetrance
increasingly economical diagnostic of the family history of cancer and of variants has also been shown
tool, with the potential to democra- segregation analyses may add to to be of great importance in cas-
tize access to effective risk assess- the interpretation of the significance cade testing of a patient’s relatives.
ment and cancer prevention. of the results. As demonstrated recently, residual
However, for many of the genes Multigene panel testing has also cancer risk for relatives who test
included in multigene panel test- resulted more often in the identifica- negative for a familial P/LP variant
ing, there are still limited data tion of unexpected, phenotype-unre- is inversely proportional to variant
on cancer-specific penetrance, and lated P/LP variants. Apart from the penetrance and is influenced by
therefore screening or risk-reducing question of causality, one has to con- family history of cancer. Therefore,
interventions are less established sider the incomplete knowledge of negative results of familial testing
(Box 6.5.2). As a result, clinical man- allelic heterogeneity and genotype– for high-penetrance variants have
agement of patients harbouring a phenotype associations in these a higher negative predictive value
P/LP variant in a moderate-pene- situations. A recent study showed than those for low-penetrance vari-
trance gene or a newly identified that carriers of germline TP53 mu- ants, and counselling of a relative
gene with little associated information tations identified by multigene panel who is unaffected by cancer and
can be very challenging. The guide- testing had fewer tumours in child- has a P/LP germline variant in a low-
lines of the National Comprehen hood, had an older median age at penetrance gene should take into
account family history of cancer [28].
Another challenge that arises
Box 6.5.2. Different levels of information and clinical utility of results of multigene panel
as a consequence of the improved
testing, according to the genes harbouring pathogenic or likely pathogenic variants.
diagnostic capacity is the frequent
identification of variants of uncertain
Genes associated with a well-known cancer predisposition syndrome significance; this is an especially fre-
• Highest cancer risks quent occurrence in populations that
• High-penetrance, low-frequency alleles are less well represented. Despite
multiple efforts to standardize the
• Risk well defined for most associated cancers process of variant calling, discor-
• Screening and management guidelines well defined dant classifications among different
laboratories are still fairly common.
• Clear implications for family members
In a study of 518 patients (603 ge-
Genes not associated with a well-known syndrome but well netic variants) tested in more than
known/researched one laboratory, the interpretation dif-
fered among the laboratories for 155
• Moderate to high cancer risks
(26%) of the variants [29].
• Moderate-penetrance and high- or moderate-frequency alleles In addition to variant calling, dis-
• Risk fairly well defined for some but not all cancers closure of a result of variants of un-
certain significance is a challenge
• Screening and management guidelines dependent on test results both for the patient and for the cli-
and family history nician, because the result does not
• Implications for family members less well defined have an associated clinical utility.
To overcome this challenge, efforts
Recently described genes should be directed towards reclas-
• Cancer risks not well defined (usually moderate or low) sification of variants of uncertain
• Management guidelines not well defined significance, but that process usu-
ally takes years, and when it is
• Implications for family members not clear available, patient contact and coun-
• Frequent variants of uncertain significance selling may be difficult.
Taken together, the benefits
• May not change medical management
and challenges of multigene panel
536
testing underscore the importance those not undergoing surveillance. cancer and to qualify them to pro-
of genetic counselling and taking a This result shows that long-term vide genetic services. Inform and
family history of cancer – an afford- compliance with surveillance for empower patients with hereditary
able tool that can still drive patient early tumour detection in patients cancer, to enhance the adherence
care and data-sharing initiatives in with Li–Fraumeni syndrome is ef- to and effectiveness of interven-
providing clinically useful genetic fective [32]. Similar strategies have tions to reduce cancer risk. These
information [8,26]. been applied in other countries, and actions aim at a reduction of the
Finally, although the benefit of although results from screening are harms that have been reported
risk-reducing mastectomy and sal- positive overall, there are still lim- as a result of lack of access to ad-
pingo-oophorectomy for BRCA mu- ited data on the effect on mortality equate genetic testing, inaccurate
tation carriers is well established, and a lack of consensus on the best interpretation of results, or failure
there have also been important long-term follow-up protocol [33]. to tailor risk-reducing interventions
advances in evidence to support A very recent advance is the de- appropriately.
cancer risk-reducing strategies in velopment of polygenic risk scores • Address the challenge of a lim-
other scenarios. The most emblem- that combine information on multi- ited workforce in GCRA through
atic example is that of Li–Fraumeni ple single-nucleotide polymorphisms, the development of tailored initia-
syndrome (OMIM no. 151623), one which are associated with very mod- tives aimed at increasing access
of the most aggressive cancer pre- est risk individually. As these tools to service provision for at-risk
disposition syndromes, which is are clinically validated, they will refine individuals.
described and characterized by a the capacity to predict risk and apply • Improve the quality of care (i.e.
high and early-onset risk of cancer. tailored interventions [34]. accuracy of genetic testing and
The disease is caused by germline interpretation of results) through
TP53 mutations, and carriers have research efforts, data shar-
an estimated lifetime risk of 80%
Current challenges,
ing, training of multidisciplinary
(males) to 100% (females) of devel-
and interventions to teams, and regulatory actions.
oping at least one malignancy. In a overcome barriers This includes (i) in-depth study
recent study of 214 families with Li– The identification and manage- of the clinical utility (i.e. associ-
Fraumeni syndrome, 4% of carriers ment of patients with hereditary ated cancer risks and appropri-
developed a malignancy in the first cancer should be regarded as a ate screening and risk-reducing
year of life, 41% were diagnosed public health concern, because options) of P/LP variants in
with cancer by age 18 years, and public health plays an important moderate-penetrance genes and
40% developed second neoplasms role in ensuring access to interven- in newly identified genes; (ii) in-
[30]. In another study of 286 carri- tions that can prevent disease. The depth study of the clinical sig-
ers from 107 families, the cumula- timely identification of patients with nificance of unexpected, pheno-
tive cancer incidence was 50% by hereditary cancer and their at-risk type-unrelated findings obtained
age 31 years in females and 50% relatives can drastically change the by multigene panel testing; and
by age 46 years in males, and near- management of individuals who (iii) characterization of the mu-
ly 100% by age 70 years for the en- have already been diagnosed with tational landscape and associ-
tire cohort [31]. cancer, and enables the implemen- ated phenotypes in populations
Recently, survival benefits from tation of cancer prevention strat- or countries with reduced ac-
intensive cancer screening in pa- egies or early detection options cess to genetic risk assessment
tients with Li–Fraumeni syndrome among at-risk relatives who are un- and, thus, very limited available
have been reported, and this has affected by cancer. information.
completely changed the approach However, to ensure that all bene- • Invest in population-specific re-
towards managing affected fami- fits of incorporating genetic or geno- search to better define the land-
lies. In 2004, a clinical surveillance mic information into an individual’s scape of genetic variants (indi-
protocol using physical examina- clinical care are attained at a global vidually or in combination) that
tion and frequent biochemical and level, several barriers must still be significantly influence cancer risk.
imaging studies was introduced in overcome. Actions suggested to re- • Develop public policies aimed at
three tertiary care centres in North duce such barriers include, but are increasing access to GCRA and
America. An 11-year follow-up not restricted to, the following: management, including genetic
showed that 5-year overall survival • Invest in the education of health- counselling, testing, risk-reducing
was significantly higher in individu- care providers, to reduce variabil- interventions, and targeted cancer
als undergoing surveillance than in ity in knowledge about hereditary therapies whenever applicable.
CHAPTER 6.5
SECTION 6
References
1. Rodriguez JL, Thomas CC, Massetti GM, 10. Green RF, Ari M, Kolor K, Dotson WD, 18. Blazer KR, Christie C, Uman G, Weitzel
Duquette D, Avner L, Iskander J, et al. Bowen S, Habarta N, et al. (2019). JN (2012). Impact of web-based case
(2016). CDC grand rounds: family his- Evaluating the role of public health in im- conferencing on cancer genetics train-
tory and genomics as tools for cancer plementation of genomics-related recom- ing outcomes for community-based cli-
prevention and control. MMWR Morb mendations: a case study of hereditary nicians. J Cancer Educ. 27(2):217–25.
Mortal Wkly Rep. 65(46):1291–4. https:// cancers using the CDC Science Impact https://doi.org/10.1007/s13187-012-0313-8
d o i . o r g /10 .15 5 8 5 / m m w r. m m 6 5 4 6 a 3 Framework. Genet Med. 21(1):28–37. PMID:22328115
PMID:27880748 ht t p s: //d o i.o r g /10.10 3 8 /s 414 3 6 - 018 -
0028-2 PMID:29907802 19. McCuaig JM, Stockley TL, Shaw P, Fung-
2. Euhus DM, Robinson L (2013). Genetic Kee-Fung M, Altman AD, Bentley J, et
predisposition syndromes and their 11. Liedke PE, Finkelstein DM, Szymonifka al.; BRCA TtoT Community of Practice
management. Surg Clin North Am. J, Barrios CH, Chavarri-Guerra Y, Bines (2018). Evolution of genetic assessment
93(2):341–62. https://doi.org/10.1016/j. J, et al. (2014). Outcomes of breast can- for BRCA-associated gynaecologic ma-
suc.2013.01.005 PMID:23464690 cer in Brazil related to health care cover- lignancies: a Canadian multisociety road-
3. Desai AV, Perpich M, Godley LA (2017). age: a retrospective cohort study. Cancer map. J Med Genet. 55(9):571–7. https://
Clinical assessment and diagnosis of Epidemiol Biomarkers Prev. 23(1):126–33. doi.org/10.1136/jmedgenet-2018-105472
germline predisposition to hematopoi- https://doi.org/10.1158/1055-9965.EPI-13- PMID:30042185
etic malignancies: the University of 0693 PMID:24165578
20. Kentwell M, Dow E, Antill Y, Wrede
Chicago experience. Front Pediatr. 5:252.
12. Martínez-Mesa J, Werutsky G, Michiels CD, McNally O, Higgs E, et al. (2017).
https://doi.org/10.3389/fped.2017.00252
S, Pereira Filho CAS, Dueñas-González Mainstreaming cancer genetics: a
PMID:29270394
A, Zarba JJ, et al. (2017). Exploring dis- model integrating germline BRCA test-
4. Daly MB, Pilarski R, Berry M, Buys SS, parities in incidence and mortality rates ing into routine ovarian cancer clinics.
Farmer M, Friedman S, et al. (2017). of breast and gynecologic cancers ac- Gynecol Oncol. 145(1):130–6. https://
NCCN guidelines insights: genetic/famil- cording to the Human Development d o i.o r g /10.1016 / j.yg y n o. 2 017.01.0 3 0
ial high-risk assessment: breast and ovar- Index in the Pan-American region. Public PMID:28162234
ian, version 2.2017. J Natl Compr Canc Health. 149:81–8. https://doi.org/10.1016/j.
Netw. 15(1):9–20. https://doi.org/10.6004/ puhe.2017.04.017 PMID:28577441 21. Slavin TP, Banks KC, Chudova D,
jnccn.2017.0003 PMID:28040716 Oxnard GR, Odegaard JI, Nagy RJ,
13. Achatz MI, Zambetti GP (2016). The in- et al. (2018). Identification of inciden-
5. Weitzel JN, Blazer KR, MacDonald DJ, herited p53 mutation in the Brazilian pop- tal germline mutations in patients with
Culver JO, Offit K (2011). Genetics, ge- ulation. Cold Spring Harb Perspect Med. advanced solid tumors who underwent
nomics, and cancer risk assessment: 6(12):a026195. https://doi.org/10.1101/ cell-free circulating tumor DNA sequenc-
state of the art and future directions in cshperspect.a026195 PMID:27663983 ing. J Clin Oncol. [Epub ahead of print]
the era of personalized medicine. CA ht tps://doi.org/10.120 0/JCO.18.0 0328
Cancer J Clin. 61(5):327–59. https://doi. 14. Foulkes WD, Knoppers BM, Turnbull C PMID:30339520
org/10.3322/caac.20128 PMID:21858794 (2016). Population genetic testing for can-
cer susceptibility: founder mutations to ge- 22. Manchanda R, Legood R (2018).
6. Acuna-Hidalgo R, Veltman JA, Hoischen nomes. Nat Rev Clin Oncol. 13(1):41–54. Population based germline testing for
A (2016). New insights into the generation https://doi.org/10.1038/nrclinonc.2015.173 primary cancer prevention. Oncotarget.
and role of de novo mutations in health PMID:26483301 9(69):33062–3. https://doi.org/10.18632/
and disease. Genome Biol. 17(1):241. oncotarget.25995 PMID:30237851
https://doi.org/10.1186/s13059-016-1110-1 15. Rebbeck TR, Friebel TM, Friedman
PMID:27894357 E, Hamann U, Huo D, Kwong A, et al.; 23. Caswell-Jin JL, Zimmer AD, Stedden
EMBRACE; GEMO Study Collaborators; W, Kingham KE, Zhou AY, Kurian AW
7. Slavin TP, Niell-Swiller M, Solomon I,
HEBON (2018). Mutational spectrum in a (2019). Cascade genetic testing of rela-
Nehoray B, Rybak C, Blazer KR, et al.
worldwide study of 29,700 families with tives for hereditary cancer risk: results
(2015). Clinical application of multigene
BRCA1 or BRCA2 mutations. Hum Mutat. of an online initiative. J Natl Cancer Inst.
panels: challenges of next-generation
39(5):593–620. https://doi.org/10.1002/ 111(1):95–8. https://doi.org/10.1093/jnci/
counseling and cancer risk management.
humu.23406 PMID:29446198 djy147 PMID:30239769
Front Oncol. 5:208. https://doi.org/10.3389/
fonc.2015.00208 PMID:26484312 16. Bray F, Jemal A, Grey N, Ferlay J, Forman 24. Yurgelun MB, Allen B, Kaldate RR,
8. George R, Kovak K, Cox SL (2015). D (2012). Global cancer transitions ac- Bowles KR, Judkins T, Kaushik P, et al.
Aligning policy to promote cascade cording to the Human Development Index (2015). Identification of a variety of muta-
genetic screening for prevention and (2008–2030): a population-based study. tions in cancer predisposition genes in
early diagnosis of heritable diseases. Lancet Oncol. 13(8):790–801. https:// patients with suspected Lynch syndrome.
J Genet Couns. 24(3):388–99. https:// doi.org/10.1016/S1470-2045(12)70211-5 Gastroenterology. 149(3):604–13.e20.
d o i.o r g /10.10 07/s10 8 97- 014 - 9 8 0 5 - 5 PMID:22658655 ht tps://doi.org/10.1053/j.gastro. 2015.
PMID:25577298 05.006 PMID:25980754
17. Gadzicki D, Evans DG, Harris H, Julian-
9. Tuffaha HW, Mitchell A, Ward RL, Connelly Reynier C, Nippert I, Schmidtke J, et al. 25. Roberts ME, Jackson SA, Susswein
L, Butler JRG, Norris S, et al. (2018). Cost- (2011). Genetic testing for familial/heredi- LR, Zeinomar N, Ma X, Marshall ML, et
effectiveness analysis of germ-line BRCA tary breast cancer – comparison of guide- al. (2018). MSH6 and PMS2 germ-line
testing in women with breast cancer and lines and recommendations from the UK, pathogenic variants implicated in Lynch
cascade testing in family members of mu- France, the Netherlands and Germany. syndrome are associated with breast
tation carriers. Genet Med. 20(9):985–94. J Community Genet. 2(2):53–69. https:// cancer. Genet Med. 20(10):1167–74.
ht tps: //doi.org /10.103 8 /gim. 2017. 231 d o i .o r g /10 .10 07/s12 6 87- 011- 0 0 42- 4 ht tps://doi.org /10.103 8 /gim. 2017. 25 4
PMID:29300376 PMID:22109790 PMID:29345684
538
26. Graffeo R, Livraghi L, Pagani O, Goldhirsch 32. Villani A, Shore A, Wasserman JD, 38. Villarreal-Garza C, Alvarez-Gómez RM,
A, Partridge AH, Garber JE (2016). Time Stephens D, Kim RH, Druker H, et al. Pérez-Plasencia C, Herrera LA, Herzog
to incorporate germline multigene panel (2016). Biochemical and imaging surveil- J, Castillo D, et al. (2015). Significant
testing into breast and ovarian cancer lance in germline TP53 mutation carriers clinical impact of recurrent BRCA1 and
patient care. Breast Cancer Res Treat. with Li-Fraumeni syndrome: 11 year follow- BRCA2 mutations in Mexico. Cancer.
160(3):393–410. https://doi.org/10.1007/ up of a prospective observational study. 121(3):372–8. https://doi.org/10.1002/cncr.
s10549-016-4003-9 PMID:27734215 Lancet Oncol. 17(9):1295–305. https:// 29058 PMID:25236687
doi.org/10.1016/S1470-2045(16)30249-2
27. Rana HQ, Gelman R, LaDuca H, PMID:27501770 39. Lolas Hamameh S, Renbaum P, Kamal L,
McFarland R, Dalton E, Thompson J, Dweik D, Salahat M, Jaraysa T, et al. (2017).
et al. (2018). Differences in TP53 muta- 33. Asdahl PH, Ojha RP, Hasle H (2017). Genomic analysis of inherited breast can-
tion carrier phenotypes emerge from Cancer screening in Li-Fraumeni syn- cer among Palestinian women: genetic
panel-based testing. J Natl Cancer Inst. drome. JAMA Oncol. 3(12):1645–6. heterogeneity and a founder mutation in
110(8):863–70. https://doi.org/10.1093/ https://doi.org/10.1001/jamaoncol.2017. TP53. Int J Cancer. 141(4):750–6. https://
jnci/djy001 PMID:29529297 2459 PMID:28772307 doi.org/10.1002/ijc.30771 PMID:28486781
28. Lee AJ, Cunningham AP, Tischkowitz 34. Mavaddat N, Michailidou K, Dennis 40. Peixoto A, Santos C, Pinheiro M, Pinto
M, Simard J, Pharoah PD, Easton DF, et J, Lush M, Fachal L, Lee A, et al.; P, Soares MJ, Rocha P, et al. (2011).
al. (2016). Incorporating truncating vari- ABCTB Investigators; kConFab/AOCS International distribution and age es-
ants in PALB2, CHEK2, and ATM into the Investigators; NBCS Collaborators (2019). timation of the Portuguese BRCA2
BOADICEA breast cancer risk model. Polygenic risk scores for prediction of c.156_157insAlu founder mutation. Breast
Genet Med. 18(12):1190–8. https://doi. breast cancer and breast cancer sub- Cancer Res Treat. 127(3):671–9. https://
org/10.1038/gim.2016.31 PMID:27464310 types. Am J Hum Genet. 104(1):21–34. d o i.o r g /10.10 07/s10 5 4 9 - 010 -10 3 6 - 3
https://doi.org/10.1016/j.ajhg.2018.11.002 PMID:20652400
29. Balmaña J, Digiovanni L, Gaddam P, PMID:30554720
Walsh MF, Joseph V, Stadler ZK, et al. 41. Nielsen SM, Rhodes L, Blanco I, Chung
(2016). Conflicting interpretation of ge- 35. Laraqui A, Uhrhammer N, Rhaffouli HE, WK, Eng C, Maher ER, et al. (2016).
netic variants and cancer risk by com- Sekhsokh Y, Lahlou-Amine I, Bajjou T, Von Hippel-Lindau disease: genetics
mercial laboratories as assessed by the et al. (2015). BRCA genetic screening in and role of genetic counseling in a mul-
Prospective Registry of Multiplex Testing. Middle Eastern and North African: muta- tiple neoplasia syndrome. J Clin Oncol.
J Clin Oncol. 34(34):4071–8. https:// tional spectrum and founder BRCA1 mu- 34(18):2172–81. https://doi.org/10.1200/
d o i . o r g / 10 .12 0 0 /J C O. 2 016 . 6 8 . 4 316 tation (c.798_799delTT) in North African. JCO.2015.65.6140 PMID:27114602
PMID:27621404 Dis Markers. 2015:194293. https://doi.
org/10.1155/2015/194293 PMID:25814778
30. Bougeard G, Renaux-Petel M, Flaman JM,
Charbonnier C, Fermey P, Belotti M, et al. 36. Seymour HJ, Wainstein T, Macaulay S,
(2015). Revisiting Li-Fraumeni syndrome Haw T, Krause A (2016). Breast cancer in
from TP53 mutation carriers. J Clin Oncol. high-risk Afrikaner families: is BRCA found-
33(21):2345–52. https://doi.org/10.1200/ er mutation testing sufficient? S Afr Med
JCO.2014.59.5728 PMID:26014290 J. 106(3):264–7. https://doi.org/10.7196/
SAMJ.2016.v106i3.10285 PMID:26915939
31. Mai PL, Khincha PP, Loud JT, DeCastro
RM, Bremer RC, Peters JA, et al. (2017). 37. Ponti G, Castellsagué E, Ruini C,
Prevalence of cancer at baseline screen- Percesepe A, Tomasi A (2015). Mismatch
ing in the National Cancer Institute Li- repair genes founder mutations and can-
Fraumeni syndrome cohort. JAMA Oncol. cer susceptibility in Lynch syndrome.
3(12):1640–5. https://doi.org/10.1001/jama Clin Genet. 87(6):507–16. https://doi.org/
oncol.2017.1350 PMID:28772286 10.1111/cge.12529 PMID:25345868
CHAPTER 6.5
SECTION 6
6.6
Screening
From biology to public health
Raúl Murillo Partha Basu (reviewer) Robert A. Smith (reviewer)

Ophira Ginsburg (reviewer)
Julietta Patnick (reviewer)
findings and by errors in sam- a suitable screening programme [2].

SUMMARY pling techniques for microscopic These proposed principles highlight
analysis. the need to detect the disease at a
●● Early detection of cancer is a preclinical stage and provide timely
critical component of cancer ●● The hallmarks of cancer may
treatment to reduce the associated
control. In addition to reduc- offer a new approach to cancer
mortality, the need for screening
tion of mortality from a specific screening by combining onco-
tests with good accuracy, and the
cancer type, a proper approach proteins, cell damage markers,
need for population-based screen-
to cancer screening should en- and epigenetic markers.
ing programmes with quality assur-
sure that the harms do not out- ●● Cost–effectiveness analyses on ance and access to confirmatory
weigh the benefits. organization of cervical cancer diagnosis and treatment, among
and breast cancer screening other characteristics (Box 6.6.1).
●● A linear evolution has been the
report variable results depend- Cancer screening programmes
underlying concept of carcino-
ing on the assumptions in the aim to comply with these princi-
genesis. However, a better un-
models. ples. However, recent research has
derstanding of tumour biology
would help to broaden cancer revealed more clearly that cancer
screening coverage while re- The available evidence consistently screening is a complex scenario in
ducing the overdiagnosis of in- shows that survival rates are sig- which there are both benefits and
dolent tumours and the under- nificantly higher for cancers that harms, and that in some instances
diagnosis of interval cancers. are detected at early stages and the harms may outweigh the bene-
properly treated than for advanced fits or the determination of whether
●● Alternative screening algorithms cancers [1]. Early detection of can- the benefits outweigh the harms can
should not only overcome the cer is achievable either by earlier be made only by the individual pa-
challenges of morphology-based diagnosis in symptomatic patients tient [4]. After decades of research
diagnosis but also help to im- or by systematic screening of and development, only screening
prove adherence in the context asymptomatic individuals. Although for cervical cancer, breast cancer,
of population-based screening, prolonged survival is a desired out- and colorectal cancer has been
to reduce the gap in mortality come for the evaluation of treat- successfully implemented, gener-
reduction between high-income ment, reduction of mortality from a ally in high-income countries [6–8],
countries and low- and middle- specific cancer is the primary ob- whereas screening for other can-
income countries. jective for cancer screening [2]. cer types, such as prostate cancer,
The principles of screening for lung cancer, and stomach cancer,
●● After decades of research and
disease proposed by Wilson and continues to be debated [4]. In
development, only screening for
Jungner [3] have been regularly low- and middle-income countries,
cervical cancer, breast cancer,
used to analyse the progress of im- where the burden of cancer mor-
and colorectal cancer has been
plementation of organized cancer tality is growing, there has been no
successfully implemented, gen-
screening [4,5]. More recently, dos significant progress in the imple-
erally in high-income countries.
Santos Silva summarized the es- mentation of cancer screening [9].
●● Observer-dependent techniques sential components of successful Contradictory results from both
are limited by inter-observer cancer screening as a suitable dis- clinical research and effectiveness
variability in the interpretation of ease, a suitable screening test, and research have promoted an intense
540
scientific debate about the valid meth- to intervene in the natural history
ods for the assessment and evalua- of the disease [4,15]. However, the FUNDAMENTALS
tion of cancer screening [10], as well use of morphological features for
as about alternative approaches for the diagnosis of pre-neoplastic le- ■■ The essential components of
programme organization to pursue sions poses the inherent challenge successful cancer screening
a better balance between diagnostic of accessing the target organ [15]. are a suitable disease, a
accuracy and treatment rates [11,12]. In addition, breast cancer, prostate suitable screening test,
The uncertainty derived from this cancer, and lung cancer have re- and a suitable screening
controversy can be reduced only by vealed great heterogeneity of dis- programme. Although there
progressively understanding tumour ease, with controversial results in is relative consensus about
biology, the factors associated with mortality reduction by screening [5]. these principles of screening,
successful screening, and technolo- The existence for the same can- improved knowledge about the
gy development as a binding element cer type of indolent, less aggressive critical components is required.
between cancer biology and public (slow-progressing), and aggressive ■■ The natural history of the
health programmes. This chapter re- tumours is currently one of the big- disease does not enable
views the contribution and potential gest challenges for cancer screen- an understanding of the
use of knowledge about these ele- ing, given the possibility of overdi- differences between
ments as a means to improve early agnosis of tumours without clinical indolent, less aggressive,
detection of cancer. significance and, at the same time, and aggressive tumours.
the difficulty of detecting lethal tu- This challenge can lead
mours in early phases (interval can- to both overtreatment and
Biological bases of cers). Furthermore, the identification undertreatment of cancers
screening of only a limited number of driver detected by screening.
genes, the discouraging results of
Natural history of the disease ■■ Improved knowledge of
mutation-targeted therapies on over-
A linear model with consecutive tumour biology warrants new
all survival, and the variable progres-
steps explains carcinogenesis from approaches in developing
sion of precancerous lesions, most
initiation to invasion [13]. The clonal screening tests or in combining
of which return spontaneously, chal-
evolution theory states that a first screening tests in alternative
lenge the theory of successive linear
mutation in a driver gene induces algorithms to improve the
somatic mutations as the only route
abnormal cell proliferation; a sec- accuracy and reliability.
of carcinogenesis [5,14,16].
ond mutation contributes to abnor- Next-generation sequencing has ■■ The experience in both high-
mal cell division and the alteration shown for a single tumour thousands income countries and low- and
of cellular architecture, resulting in of genetic alterations not contained middle-income countries on
benign tumours or identifiable pre- in germlines, and has enabled a implementation of cancer
cancerous conditions; and subse- better understanding of the roles of screening has furthered
quent mutations produce the final these alterations not only by differen- innovative programmatic
transformation to a cell with inva- tiating driver genes from passenger approaches that are suited to
sive capacity [13]. genes but also by elucidating the role different levels of resources
With this approach, actionable of epigenetic alterations involved in and contexts.
models of carcinogenesis are best malignant cellular transformation.
expressed by the progress of cervi- Moreover, recent publications have
cal intraepithelial neoplasia to inva- highlighted the role of the tissue and states that many anomalies involv-
sive cervical cancer and the devel- tumour microenvironment [16] and ing genomic instability could rapidly
opment of adenomatous polyps that have proposed new approaches to occur, reshaping the entire genome
progress to invasive cancer of the better explain tumour heterogeneity from one or two dominant clones [18]
colon [5]. However, the approach is and the onset of aggressive tumours (Fig. 6.6.1).
also proposed in the development over a short period, such as the con- The new theories enable a better
of cutaneous naevi to melanoma, cept of the field effect, which sug- understanding of tumour diversity.
the progression of Barrett oesopha- gests multiple initiating cells with in- Srivastava et al. have argued that
gus to oesophageal adenocarcino- dependent evolution [17]. In addition, the difference between indolent and
ma, and the progression of ductal alternative models of clonal evolution aggressive tumours may not rely ex-
adenocarcinomas in the pancreas suggest branched and punctuated clusively on the characteristics of tu-
and the breast [5,13,14]. evolutions; branched evolution en- mour cells, but is instead determined
In this context, dysplasia is the tails multiple clonal lineages evolv- by interactions among the host, en-
CHAPTER 6.6
SECTION 6
ideal surrogate marker for cancer, ing in parallel and cellular coopera- vironmental exposures, and neopla-
and its detection in asymptomatic tion via paracrine interactions, and sia [14]. Therefore, understanding
individuals is seen as the best way the model of punctuated evolution these interactions could determine
Chapter 6.6 • Screening 541

Box 6.6.1. Principles of cancer screening.
Principles of early disease detection, from Wilson Essential components of successful cancer
and Jungner (1968) [3]: screening, from dos Santos Silva (1999) [2]:
1. The condition sought should be an important 1. Suitable disease:

health problem. • Detectable preclinical phase
2. There should be a recognizable latent or early • Early treatment
symptomatic stage. • Relative burden of disease.
3. The natural history of the condition, including

development from latent to declared disease,
should be adequately understood. 2. Suitable screening test:
4. There should be an accepted treatment for • Validity (sensitivity and specificity)
patients with recognized disease, and treatment • Acceptability and costs.
should be better at an earlier stage.
5. There should be an agreed policy on whom to
treat, for patient care as a whole. 3. Suitable screening programme:
• There is a clear definition of the target population.
• The individuals to be screened are identifiable.
• Measures are available to ensure high coverage
6. There should be a suitable test or examination.
and attendance.
7. The test should be acceptable to the population. • There are adequate field facilities for collecting
the screening material and adequate laboratory
facilities to examine it.
• There is an organized quality control programme to
8. Facilities for diagnosis and treatment should be
assess the screening material and its interpretation.
available.
• Adequate facilities exist for diagnosis and
9. Case-finding should be a continuing process and appropriate treatment of confirmed neoplastic
not a “once-and-for-all” project. lesions and for the follow-up of treated individuals.
10. The cost of case-finding (including diagnosis • There is a carefully designed referral system for
and treatment of patients diagnosed) should be management of any abnormality found.
economically balanced in relation to possible • Evaluation and monitoring of the total programme
expenditure on medical care as a whole. is organized.
the ideal time to effectively use a unrelated to a specific evolution- the molecular profile of premalignant
screening test and significantly re- ary model. From this perspective, lesions remains challenging, because
duce the chance of overdiagnosis. therapies targeted to precise sig- individual mutations do not follow a
nalling pathways have been devel- consistent pattern between premalig-
Hallmarks of cancer oped irrespective of clinical stage nant and malignant states, suggest-
Hanahan and Weinberg proposed at diagnosis, with the idea that each ing a variable order and timing in the
a set of characteristics of malignant tumour expresses its hallmark ca- process of carcinogenesis [21]. In ad-
cells as the basis of molecular mech- pabilities within a certain clinical dition to cellular properties, changes
anisms that enable tumour growth and molecular course, which might in the surrounding tissue and in the
and metastatic invasion [19]. They differ from patient to patient [14]. cell microenvironment have been
proposed acquired capabilities as Although the described hallmarks proposed as early indicators of malig-
the hallmarks of cancer cells, includ- are distinctive of malignant cells, nant transformation, including pro-in-
ing sustaining proliferative signalling, many of them must be expressed flammatory and immune responses,
evading growth suppressors, resist- early in the process of carcinogen- changes in energy metabolism, and
ing cell death, enabling replicative esis. Accordingly, alterations in cell increased angiogenesis.
immortality, sustaining angiogenesis, proliferation and differentiation, an-
evading immune destruction, repro- ti-growth signalling, and apoptosis
gramming energy metabolism, and have been reported for different pre- Screening tests
activating invasion and metastasis. cancerous conditions [20]. Therefore, Cancer diagnosis continues to be
In addition to an improved un- early detection of anomalies in the morphology-based. Therefore, tis-
derstanding of cancer biology, the cell circuits involved has prompted sue or cell samples are needed to
hallmarks of cancer offer an alter- enthusiastic research into cancer verify the malignant transformation,
native approach to carcinogenesis screening. However, understanding and this condition may influence the
542
Fig. 6.6.1. Models of clonal evolution and tumour progression. The relationship between models of carcinogenesis (clonal evolution
with tumours shown in black), tumour progression through clinical stages, and progression time with regard to early detection by
screening. A linear evolution of carcinogenesis (A) is more plausible in tumours that have a long sojourn time, progressively transit
through clinical stages, and are detectable by screening. However, some tumours that are due to this pattern may have a slow
growth rate (even regression) and would not be detected by screening (indolent tumours, shown by the dotted lines). In tumours
with branched evolution (B), clones derive from a common ancestor but evolve in parallel. Such tumours may have more rapid pro-
gression, but the sojourn time is still long enough to enable their detection by screening. Some tumours have punctuated evolution
(C), with a large number of mutations in short periods and one or two clones progressing rapidly. Therefore, they are more difficult
to detect by screening (interval cancers).
Clonal evolution Tumour progression Progression time

A Interval cancer
Normal Dysplasia/ Cancer Cancer Cancer Screen-detectable

cell Ca in situ Stage I Stage II/III Stage IV
Indolent tumour
B Interval cancer
Normal Dysplasia/ Cancer Cancer Cancer
cell Ca in situ Stage I Stage II/III Stage IV
Screen-detectable
Indolent tumour
C Interval cancer
Normal Cancer Cancer
cell Stage I–III Stage IV Screen-detectable
Indolent tumour
development of technologies for the events associated with invasive An additional characteristic of
early detection of cancer. procedures highlight the need to morphology-based diagnosis is ob-
To date, the epidemiological ax- confer higher value to the specific- server dependency. Most screening
iom favours the combination of a ity of screening tests, in addition to tests in use today (Table 6.6.1) seek
highly sensitive screening test with reassessing their capability to avoid macroscopic or microscopic visuali-
a highly specific diagnostic test [4]. the detection of indolent tumours. zation of changes related to malignant
Despite the low sensitivity of cer-
vical cytology and faecal occult Fig. 6.6.2. Safe specimen sampling is possible if direct anatomical access is available,
blood tests, the achievements of as is the case for skin cancer screening.
screening for cervical cancer and
colorectal cancer reinforce this ap-
proach. Highly frequent screening
(i.e. with a short interval) usually
rectifies the low sensitivity; how-
ever, this is possible only if the dis-
ease has a long sojourn time and
if it is not difficult to obtain tissue
or cell samples, thus resulting in a
positive balance between the ben-
efits and the risk [15].
Safe specimen sampling is pos-
sible if direct anatomical access is
available, as to the skin or the oral
cavity, and even for organs that are
accessible by endoscopy, such as
CHAPTER 6.6
SECTION 6
the stomach. In contrast, the inac-

cessibility of visceral organs and
the potential severity of adverse

Table 6.6.1. Cancer screening practices

Screening interval Main age range Mortality
Cancer site Screening test
(years) (years) reduction?
Image-based screening
Breast Mammography 1–3 50–69 Yes
Lung Low-dose computed tomography (CT) 1–2 55–74 a

Yes
Stomach Upper gastrointestinal X-ray series 1–2 ≥ 40 Uncertain
Direct or endoscopic visual screening
Cervix Visual inspection with acetic acid (VIA) 1–3 30–49 Yesb
Oral Direct visual inspection 1–3 ≥ 35 a

Yesb
Colon Colonoscopy 5–10 50–69 Yes
Colon Flexible sigmoidoscopy 3–5 55–69 Yes
Stomach Upper gastrointestinal endoscopy 1–2 40–64 Uncertain
Clinical examination screening
Breast Clinical breast examination 1 40–69 Unknown
Breast Breast self-examination – – No
Cell sampling screening
Cervix Cervical cytology 1–3 25–69 Yesb
Biomarker-based screening
Cervix Human papillomavirus (HPV) testing 3–5 30–65 Yes
Colon Faecal occult blood test (FOBT) 1–2 50–69 Yes
Stomach Pepsinogen I/II 40–64 Unknown
Prostate Prostate-specific antigen (PSA) 1–5 50–74 Uncertain
Liver α-Fetoprotein (AFP) c

Every 6 months High risk a
Uncertain
Ovary CA125 – – No
a
Restricted to individuals at high risk: tobacco use for lung cancer and oral cancer; chronic hepatitis or cirrhosis for liver cancer.
b
Limited evidence: VIA, one trial using “screen and treat” in one visit; direct visual inspection, one trial without adjustment by cluster design; cervical
cytology, based on observational studies (effectiveness).
c
Regularly combined with ultrasound.
transformation. Observer-dependent nario would be to combine different ficity for early identification of le-
techniques share some limitations, tests in diagnostic algorithms to guar- sions with high malignant potential.
such as variability in the character- antee adequate sensitivity and speci- To date, only human papillomavirus
istics of premalignant and malignant ficity. However, gains in diagnostic (HPV) tests and faecal occult blood
lesions, inter-observer variability in accuracy could be counterbalanced tests have solid evidence for reduc-
the interpretation of findings, and er- by the effects of such algorithms on tion of cancer mortality when used
rors in sampling techniques for mi- the number of visits and patient ad- as screening tests (Table 6.6.1).
croscopic analysis [15,22]. Although herence to clinical protocol [23]. The search for new diagnostic
these techniques are complemented Biomarkers have several ad- biomarkers requires prolonged pro-
by histological verification, the limi- vantages for cancer screening, in- cesses and faces several challeng-
tations noted must be compensated cluding the possibility of measuring es, which increase if asymptomatic
for with short screening intervals and them in body fluids, measuring in individuals with low prevalence of
high reassessment rates. quantitative terms, lowering pro- disease are envisioned as the target
Currently, research on alternative vider dependency, using automated population. The accessibility of body
approaches to cancer screening fo- platforms with high throughput, re- fluids is countered by the lack of
cuses on functional images and bio- ducing costs by large-scale produc- specificity to the site of tumour origin
markers of early disease. To date, no tion, and reducing the number of and by the low concentration of mark-
single technology has overcome the visits through simultaneous testing ers released in the early stages of
limitations of anatomical accessibility in a single specimen [24]. However, tumour development [24]. Moreover,
or diagnostic accuracy and reliability. most existing biomarkers suffer mortality reduction as the main re-
Therefore, the most likely future sce- from limited sensitivity or low speci- search outcome and avoidance of
544
detection of indolent tumours contin- Although mortality rates from tified as major factors that influ-
ue to be major challenges in translat- cervical cancer are low in Europe, ence cost–effectiveness ratios [36].
ing basic research into clinical prac- data from programme evaluation in Hence, quality assurance plays a
tice. These limitations are common to European countries show that popu- central role in minimizing false-neg-
tests based on cells, DNA, proteins, lation-based screening programmes ative and false-positive results, and
and circulating metabolites, which do not cover most of the region [31] observer-dependent tests present a
are currently the most widespread (see Chapter 4.5). Moreover, case– challenge in this respect. However,
research approaches to early detec- control studies reveal greater effec- the broad concept should be tailored
tion of cancer. tiveness for organized screening ver- according to the level of resources,
Some novel approaches to sus opportunistic screening [32,33], because certain quality assurance
overcome these limitations include but cohort analyses have shown vari- guidelines from high-income coun-
combining oncoproteins, cell dam- able results over time, with a greater tries propose more than 40 indica-
age markers, and epigenetic mark- effect of organized screening on cer- tors per programme [6–8], a stan-
vical cancer incidence revealed in dard that is difficult to meet in most
ers to increase specificity to lesions
earlier studies [34,35]. low- and middle-income countries.
with high malignant potential [25],
Similarly, cost–effectiveness anal- In addition to deficient participa-
combining circulating markers with
yses on organization of cervical can- tion and quality, deficient follow-up
tumour antigens to improve specific-
cer (see Chapter 5.10) and breast of positive screening results and
ity to the site of tumour origin [26],
cancer (see Chapter 5.9) screening reassessment of equivocal results
searching for markers in fluids spe- contribute to the lack of mortali-
report variable results depending
cific to the site of tumour origin [27], on the assumptions in the models. ty reduction in low- and middle-
and combining functional tests with In general, organized screening is income countries [30], as well as
anatomical images [28]. Thus, new more cost-effective than opportu- to the higher mortality in socially
technologies in genomics, proteo- nistic screening. However, analy- disadvantaged populations in high-
mics, and metabolomics, as well as ses of the effectiveness of screen- income countries. Cervical cytology
the growing number of high-quality ing reveal no significant differences screening has reduced mortality
biorepositories and a greater capac- when data from real scenarios are from cervical cancer in high-income
ity for data analysis, are opening up fed into the models [36,37], as op- countries, but short screening inter-
new avenues to search for cancer posed to models with hypothetical vals and high reassessment rates
screening biomarkers. scenarios that assume substantially hinder adherence in women with
Furthermore, the use of big da- lower participation rates for oppor- limited access to health care [30].
tabases and machine learning offer tunistic screening [38]. The gap in mortality reduction
new opportunities to improve the Although the definition of orga- between high-income countries and
accuracy of screening tests (par- nized and opportunistic screening low- and middle-income countries
ticularly for image-based screen- is not consistent across studies, has invigorated the search for alter-
ing) and to improve individual risk screening accuracy and excessive native programmatic approaches,
stratification in order to better guide use of diagnostics have been iden- accompanied by the introduction of
screening protocols.
Fig. 6.6.3. Women waiting at a mobile clinic for free breast cancer screening in Moscow,
Screening programmes Russian Federation.
Population-based programmes are

considered to be essential for suc-
cessful cancer screening (Box 6.6.1).
The main effects expected from such
programmes are increased cover-
age, improved cost–effectiveness,
and improved equity. Early studies
in Europe showed an inverse rela-
tionship between screening cover-
age and cervical cancer incidence
and mortality [29]. However, this
relationship is less clear in regions
without population-based screen-
ing, such as Latin America, where
CHAPTER 6.6
SECTION 6
screening coverage has increased

but recall attendance after positive
screening results remains low [30].

technologies that conform to these the requirements for facilities and Recently, stratified screening ac-
approaches. A “screen and treat” ap- professional skills are challenges cording to individual risk has been
proach in one or two visits has been for patient access in low-income proposed for early detection of breast
proven to result in a significant reduc- settings [43]. Moreover, controver- cancer [44]. This is the underlying
tion in mortality from cervical cancer sial data on effectiveness, cost– concept of HPV testing in cervical
in low-income settings, either with effectiveness, and overdiagnosis cancer screening, and similar ap-
visual inspection with acetic acid or have impaired the implementation proaches have been developed for
with HPV testing [39,40]. HPV testing of mammography screening pro- screening of colorectal cancer (by fa-
has also enabled self-sampling and grammes in low- and middle-income milial and genetic risk) and lung can-
the identification of women at higher countries. A stepwise approach ac- cer (by smoking history). Although
risk. Self-sampling favours participa- cording to level of resources and preliminary data on effectiveness
tion in reluctant populations [41], and health system capacity seems more and cost–effectiveness are positive,
the identification of women at higher suited to these scenarios, moving the ultimate success of the strategy
risk has led to a greater reduction from breast awareness (based on will depend on the predictive capacity
in the incidence of cervical cancer breast self-examination) to a shift of risk assessment and the final im-
[42]. However, the lower specificity of the stage distribution of detect- pact on mortality reduction.
must be corrected for by additional ed disease towards a lower stage
visits to triage HPV-positive women (based on clinical breast examina-
(Fig. 6.6.4). tion) and progressive implementa- Conclusions
Mammography screening has re- tion of mammography screening The connections among disease,
duced mortality from breast cancer (from hospital-based to population- screening tests, and screening pro-
in high-income countries. However, based) [11]. grammes remain valid. However, in
Fig. 6.6.4. Alternative approaches for cervical cancer screening according to natural history of the disease. Available technologies for
self-collection and physician collection. * Visual inspection with acetic acid (VIA) has not reliably demonstrated high specificity.
Screening tests Natural history Programmatic approaches
Positive screening results

Normal
Sensitivity
HPV DNA and mRNA
High-risk HPV
HPV E6/E7 markers
Specificity
Persistent HPV
Cytology, VIA*
Precancer
Triage
Cancer
Programmatic approaches Colposcopy

Screen, triage, diagnose, and treat (3 or more visits) (biopsy)
Screen, diagnose, and treat (2–3 visits)
Screen and treat (1–2 visits)
Treatment
546
Fig. 6.6.5. Alternative approaches to cancer screening.
Programme Modifiable Critical elements Alternatives

component factor for analysis under research
Tumour biology
Target population Age range Burden of disease
Cost–effectiveness
Group risk stratification

Sensitivity
Screening test
Screening test Specificity
Screening interval
Access/Adherence
Threshold
Individual risk
stratification
Combined algorithms
Specificity
Diagnostics Diagnostic test Overdiagnosis
Number of visits Access/Adherence
Triage testing
Tumour biology
Intervention
Treatment Access/Adherence
Number of visits
Active surveillance
understanding tumour biology, the capacity of new technologies, must (Fig. 6.6.5). Knowledge accumulated
prevailing linear approach to iden- be reviewed. from years of experience, not only
tifying tumours with aggressive be- Currently, the implementation of in high-income countries but also in
haviour that merit early detection cancer screening might be improved low- and middle-income countries,
must be overcome. Overdiagnosis by variations in programmatic ap- reveals the need to rethink screen-
of indolent tumours and the mor- proaches, including, as necessary, ing programmes on the basis of
phological basis of cancer diagnosis decreased screening intensity, a re- the level of resources available and
are the most relevant challenges in duced number of visits for the clini- the specific conditions of each sce-
searching for alternative approach- cal protocol, increased cut-off points nario. Combining programmatic ap-
es to cancer screening. These con- for referrals on diagnostic confirma- proaches with suitable technologies
cepts elicit a change in the traditional tion, stratified screening accord- ensures broader participation and
epidemiological approach, in which ing to population risk, and expect- increased treatment rates.
the balance between sensitivity and ing behaviour against lesions that
specificity, as well as the predictive are suspected to be indolent [45]
CHAPTER 6.6
SECTION 6

References
1. Allemani C, Matsuda T, Di Carlo V, 9. Sullivan T, Sullivan R, Ginsburg OM 19. Hanahan D, Weinberg RA (2011). Hallmarks
Harewood R, Matz M, Nikšić M, et al.; (2015). Screening for cancer: con- of cancer: the next generation. Cell.
CONCORD Working Group (2018). Global siderations for low- and middle-in- 144(5):646–74. https://doi.org/10.1016/j.
surveillance of trends in cancer survival come countries. In: Gelband H, Jha P, cell.2011.02.013 PMID:21376230
2000–14 (CONCORD-3): analysis of indi- Sankaranarayanan R, Horton S, editors.
vidual records for 37 513 025 patients di- Disease control priorities. 3rd ed. Vol. 3, 20. Srivastava S, Grizzle WE (2010).
agnosed with one of 18 cancers from 322 Cancer. Washington (DC), USA: World Biomarkers and the genetics of early
population-based registries in 71 coun- neoplastic lesions. Cancer Biomark. 9(1–
Bank; pp. 211–222.
tries. Lancet. 391(10125):1023–75. https:// 6):41–64. https://doi.org/10.3233/CBM-
doi.org/10.1016/S0140-6736(17)33326-3 10. Jacklyn G, Bell K, Hayen A (2017). 2011-0204 PMID:22112469
PMID:29395269 Assessing the efficacy of cancer screening.
21. Ryan BM, Faupel-Badger JM (2016). The
Public Health Res Pract. 27(3):e2731727.
2. dos Santos Silva I (1999). Cancer preven- hallmarks of premalignant conditions: a
ht tps://doi.org/10.170 61/phrp2731727
tion. In: Cancer epidemiology: principles molecular basis for cancer prevention.
PMID:28765860
and methods. Lyon, France: International Semin Oncol. 43(1):22–35. https://doi.
Agency for Research on Cancer; pp. 355– 11. WHO (2014). WHO position paper on mam- org/10.1053/j.seminoncol.2015.09.007
380. Available from: http://publications.iarc. mography screening. Geneva, Switzerland: PMID:26970122
fr/421. World Health Organization. Available from:
22. Baron JA (2012). Screening for cancer
https://www.who.int/cancer/publications/
3. Wilson JMG, Jungner G (1968). Principles with molecular markers: progress comes
mammography_screening/en/.
and practice of screening for disease. with potential problems. Nat Rev Cancer.
Public Health Papers No. 34. Geneva, 12. WHO (2013). WHO guidelines for screen- 12(5):368–71. https://doi.org/10.1038/
Switzerland: World Health Organization. ing and treatment of precancerous le- nrc3260 PMID:22495319
Available from: http://whqlibdoc.who.int/ sions for cervical cancer prevention. 23. Herrero R, Murillo R (2018). Cervical
php/WHO_PHP_34.pdf. Geneva, Switzerland: World Health cancer. In: Thun M, Linet MS, Cerhan
Organization. Available from: https://www. JR, Haiman CA, Schottenfeld D, edi-
4. Croswell JM, Ransohoff DF, Kramer BS who.int/reproductivehealth/publications/
(2010). Principles of cancer screening: les- tors. Cancer epidemiology and preven-
c a n c e r s /s c r e e n i n g _ a n d _ t r e a t m e nt _ tion, 4th ed. New York (NY), USA: Oxford
sons from history and study design issues. of_precancerous_lesions/en/.
Semin Oncol. 37(3):202–15. https://doi. University Press; pp. 925–46.
org/10.1053/j.seminoncol.2010.05.006 13. Vogelstein B, Kinzler KW (2015). The 24. Duffy MJ (2015). Use of biomarkers in
PMID:20709205 path to cancer – three strikes and you’re screening for cancer. Adv Exp Med Biol.
out. N Engl J Med. 373(20):1895–8. 867:27–39. https://doi.org/10.1007/978-94-
5. Shieh Y, Eklund M, Sawaya GF, Black https://doi.org/10.1056/NEJMp1508811
WC, Kramer BS, Esserman LJ (2016). 017-7215-0_3 PMID:26530358
PMID:26559569
Population-based screening for cancer: 25. Wentzensen N, Schiffman M, Palmer
hope and hype. Nat Rev Clin Oncol. 13(9): 14. Srivastava S, Reid BJ, Ghosh S, Kramer T, Arbyn M (2016). Triage of HPV posi-
550–65. https://doi.org/10.1038/nrclinonc. BS (2016). Research needs for under- tive women in cervical cancer screen-
2016.50 PMID:27071351 standing the biology of overdiagnosis in ing. J Clin Virol. 76(Suppl 1):S49–55.
cancer screening. J Cell Physiol. 231(9): https://doi.org/10.1016/j.jcv.2015.11.015
6. Dimitrova N, Parkinson ZS, Bramesfeld 1870–5. https://doi.org/10.1002/jcp.25227
A, Ulutürk A, Bocchi G, López-Alcalde PMID:26643050
PMID:26505642
J, et al. (2016). European guidelines for 26. Cohen JD, Li L, Wang Y, Thoburn C,
breast cancer screening and diagno- 15. Khunger M, Kumar U, Roy HK, Tiwari AK Afsari B, Danilova L, et al. (2018).
sis – the European breast guidelines. (2014). Dysplasia and cancer screen- Detection and localization of surgically
Luxembourg: Publications Office of the ing in 21st century. APMIS. 122(8):674– resectable cancers with a multi-analyte
European Union. Available from: https://doi. 82. https://doi.org/10.1111/apm.12283 blood test. Science. 359(6378):926–30.
org/10.2788/503032. PMID:24910362 https://doi.org/10.1126/science.aar3247
PMID:29348365
7. Arbyn M, Anttila A, Jordan J, Ronco G, 16. Brücher BL, Jamall IS (2016). Somatic mu-
Schenck U, Segnan N, et al. (2010). tation theory – why it’s wrong for most can- 27. Martin KJ, Fournier MV, Reddy GP,
European guidelines for quality assurance cers. Cell Physiol Biochem. 38(5):1663– Pardee AB (2010). A need for basic re-
in cervical cancer screening. Second edi- 80. https://doi.org/10.1159/000443106 search on fluid-based early detection
tion – summary document. Ann Oncol. PMID:27160408 biomarkers. Cancer Res. 70(13):5203–6.
21(3):448–58. https://doi.org/10.1093/ https://doi.org/10.1158/0008-5472.CAN-
annonc/mdp471 PMID:20176693 17. Lochhead P, Chan AT, Nishihara R, Fuchs 10-0987 PMID:20587531
CS, Beck AH, Giovannucci E, et al. (2015).
8. von Karsa L, Patnick J, Segnan N, Atkin Etiologic field effect: reappraisal of the field 28. Minamimoto R, Senda M, Jinnouchi S,
W, Halloran S, Lansdorp-Vogelaar I, effect concept in cancer predisposition Terauchi T, Yoshida T, Inoue T (2015).
et al.; European Colorectal Cancer and progression. Mod Pathol. 28(1):14–29. Detection of breast cancer in an FDG-
Screening Guidelines Working Group https://doi.org/10.1038/modpathol.2014.81 PET cancer screening program: results
(2013). European guidelines for qual- PMID:24925058 of a nationwide Japanese survey. Clin
ity assurance in colorectal cancer Breast Cancer. 15(2):e139–46. https://
screening and diagnosis: overview 18. Davis A, Gao R, Navin N (2017). Tumor evo- d o i . o r g / 10 .1016 / j . c l b c . 2 014 . 0 9 . 0 0 8
and introduction to the full supplement lution: linear, branching, neutral or punctu- PMID:25454690
publication. Endoscopy. 45(1):51–9. ated? Biochim Biophys Acta Rev Cancer.
https://doi.org/10.1055/s-0032-1325997 1867(2):151–61. https://doi.org/10.1016/j.
PMID:23212726 bbcan.2017.01.003 PMID:28110020
548
29. Quinn M, Babb P, Jones J, Allen E (1999). 35. Ronco G, Pilutti S, Patriarca S, Montanari 41. Madzima TR, Vahabi M, Lofters A (2017).
Effect of screening on incidence of and G, Ghiringhello B, Volante R, et al.; Turin Emerging role of HPV self-sampling in
mortality from cancer of cervix in England: Cervical Screening Working Group (2005). cervical cancer screening for hard-to-
evaluation based on routinely collect- Impact of the introduction of organised reach women: focused literature review.
ed statistics. BMJ. 318(7188):904–8. screening for cervical cancer in Turin, Can Fam Physician. 63(8):597–601.
https://doi.org/10.1136/bmj.318.7188.904 Italy: cancer incidence by screening his- PMID:28807952
PMID:10102852 tory 1992–98. Br J Cancer. 93(3):376–8.
https://doi.org/10.1038/sj.bjc.6602705 42. Ronco G, Dillner J, Elfström KM, Tunesi S,
30. Murillo R, Almonte M, Pereira A, Ferrer PMID:16012518 Snijders PJ, Arbyn M, et al.; International
E, Gamboa OA, Jerónimo J, et al. (2008). HPV screening working group (2014).
Cervical cancer screening programs 36. de Gelder R, Bulliard JL, de Wolf C, Efficacy of HPV-based screening for pre-
in Latin America and the Caribbean. Fracheboud J, Draisma G, Schopper D, vention of invasive cervical cancer: follow-
Vaccine. 26(Suppl 11):L37–48. https:// et al. (2009). Cost-effectiveness of oppor- up of four European randomised controlled
doi.org/10.1016/j.vaccine.2008.06.013 tunistic versus organised mammography trials. Lancet. 383(9916):524–32. https://
PMID:18945401 screening in Switzerland. Eur J Cancer. doi.org/10.1016/S0140-6736(13)62218-7
45(1):127–38. https://doi.org/10.1016/j. PMID:24192252
31. Basu P, Ponti A, Anttila A, Ronco G, ejca.2008.09.015 PMID:19038540
Senore C, Vale DB, et al. (2018). Status 43. Panieri E (2012). Breast cancer screen-
of implementation and organization of 37. Adab P, McGhee SM, Yanova J, Wong ing in developing countries. Best Pract
cancer screening in the European Union CM, Hedley AJ (2004). Effectiveness Res Clin Obstet Gynaecol. 26(2):283–90.
and efficiency of opportunistic cervi- https://doi.org/10.1016/j.bpobgyn.2011.
Member States – summary results from
cal cancer screening: comparison
the second European screening report. 11.007 PMID:22222136
with organized screening. Med Care.
Int J Cancer. 142(1):44–56. https://doi.
42(6):600–9. https://doi.org/10.1097/01.mlr. 44. Gray E, Donten A, Karssemeijer N, van
org/10.1002/ijc.31043 PMID:28940326
0000128007.04494.29 PMID:15167328 Gils C, Evans DG, Astley S, et al. (2017).
32. Makkonen P, Heinävaara S, Sarkeala Evaluation of a stratified national breast
38. Kim JJ, Leung GM, Woo PP, Goldie SJ
T, Anttila A (2017). Impact of organized screening program in the United Kingdom:
(2004). Cost-effectiveness of organized
and opportunistic Pap testing on the an early model-based cost-effectiveness
versus opportunistic cervical cytology
risk of cervical cancer in young women analysis. Value Health. 20(8):1100–9.
screening in Hong Kong. J Public Health
– a case-control study from Finland. https://doi.org/10.1016/j.jval.2017.04.012
(Oxf). 26(2):130–7. https://doi.org/10.1093/
Gynecol Oncol. 147(3):601–6. https:// pubmed/fdh138 PMID:15284314 PMID:28964442
d o i.o r g /10.1016 / j.yg y n o. 2 017.0 9.010
39. Sankaranarayanan R, Nene BM, Shastri 45. Esserman LJ, Thompson IM, Reid B, Nelson
PMID:28942994
SS, Jayant K, Muwonge R, Budukh AM, P, Ransohoff DF, Welch HG, et al. (2014).
33. Nieminen P, Kallio M, Anttila A, Hakama et al. (2009). HPV screening for cervi- Addressing overdiagnosis and overtreat-
M (1999). Organised vs. spontaneous cal cancer in rural India. N Engl J Med. ment in cancer: a prescription for change.
Pap-smear screening for cervical can- 360(14):1385–94. https://doi.org/10.1056/ Lancet Oncol. 15(6):e234–42. https://
cer: a case-control study. Int J Cancer. NEJMoa0808516 PMID:19339719 doi.org/10.1016/S1470-2045(13)70598-9
83(1):55–8. https://doi.org/10.1002/(SICI) PMID:24807866
10 97- 0 215 (19 9 9 0 9 24) 8 3:1< 5 5:: A I D - 40. Sankaranarayanan R, Esmy PO, Rajkumar
IJC11>3.0.CO;2-U PMID:10449608 R, Muwonge R, Swaminathan R,
Shanthakumari S, et al. (2007). Effect of vi-
34. Lynge E, Madsen M, Engholm G (1989). sual screening on cervical cancer incidence
Effect of organized screening on inci- and mortality in Tamil Nadu, India: a clus-
dence and mortality of cervical cancer in ter-randomised trial. Lancet. 370(9585):
Denmark. Cancer Res. 49(8):2157–60. 398–406. https://doi.org/10.1016/S0140-
PMID:2702657 6736(07)61195-7 PMID:17679017
CHAPTER 6.6
SECTION 6

6.7
Circulating DNA and other biomarkers

for early diagnosis
Great potential, but challenges recognized
Anna Babayan Shaoqing Ju (reviewer)
Natalie Reimers James McKay (reviewer)
Klaus Pantel
(see Chapter 5.12). The family of to become an essential element of

SUMMARY liquid biopsy analytes includes cir- personalized medicine (Fig. 6.7.1).
culating tumour cells (CTCs), circu- This chapter discusses some of
●● The analysis of tumour-derived the recent highlights on the use of
lating cell-free tumour DNA (ctDNA),
products, including circulating ctDNA and CTCs for early detection
circulating non-coding nucleic acids
cell-free tumour DNA (ctDNA) and monitoring of cancer.
and related biomarkers, in body such as microRNAs (miRNAs) and
fluids is increasingly recognized long non-coding RNAs (lncRNAs),
as an aid in the early diagnosis extracellular vesicles, and tumour- ctDNA for early detection
of malignant disease. educated platelets [2–6]. of cancer
Over the past 10 years, many Early detection of cancer in the
●● For application in screening or
liquid biopsy tests have been estab- context of a screening programme
early diagnosis, ctDNA analy-
lished and validated [3]. Clinical ap- for healthy individuals at high risk
sis and related techniques re-
quire well-validated tests with plications of liquid biopsy in patients is a much sought-after goal in can-
exceptionally high sensitivity with early-stage cancer include early cer research. Current therapeutic
and specificity. detection of small tumours, improved strategies, in particular surgery, en-
risk assessment (tumour staging), able many patients with cancer to be
●● Recent approaches have com- and monitoring of minimal residual cured, provided the disease is de-
bined the evaluation of soluble tected early in its anticipated clinical
disease [7]. Thus, liquid biopsy is set
tumour biomarkers with ctDNA
analysis of cancer-related mu-
tations in multiple genes. Fig. 6.7.1. Liquid biopsy in cancer. Schematic representation of the liquid biopsy
concept as the analysis of circulating tumour cells (CTCs), circulating cell-free tumour
●● These technologies face chal- DNA (ctDNA), non-coding RNAs (ncRNAs), exosomes, and tumour-educated platelets
lenges, including low concen- in the blood of patients with cancer. Key applications of liquid biopsy are listed.
trations of ctDNA and other liq-
uid biomarker analytes.
●● Progress in technology (e.g.
next-generation sequencing) is
paving the way for the develop-
ment of diagnostic tests for early
detection of cancer and the in-
troduction of precision medicine
into clinical practice.
The analysis of tumour cells and

tumour-derived products detectable
in blood and other body fluids, which
was introduced by Pantel and Alix-
Panabieres and has been referred to
as a liquid biopsy [1], has garnered
substantial interest in recent years
550
course. However, metastatic disease cation of ctDNA analysis for early di-
remains largely incurable, with very agnosis of small cancerous lesions. FUNDAMENTALS
few exceptions, which specifically in- The proper choice of markers is
clude testicular cancer or small liver also very important. Markers detect- ■■ The analysis of circulating
metastases in colon cancer. ed and validated in patients with ad- tumour cells, circulating cell-
Liquid biopsy, as a minimally in- vanced disease, such as CEA, lack free tumour DNA, and other
vasive and easily repeatable meth- specificity and sensitivity for early tumour-derived products
od, seems to offer an attractive al- detection. Concentrations of the detectable in the blood and
ternative to invasive tissue biopsies marker are lower at early stages of other body fluids has been
as the current definitive methodol- disease than at late stages. In addi- referred to as liquid biopsy.
ogy in tumour diagnostics. However, tion, the biology of these two disease
states varies; therefore, a late-stage ■■ Most research has been
programmes for early detection or
marker may not be suitable to detect focused on prognosis and
screening require well-validated tests
small tumours at early stages. Also, therapy, including real-time
with exceptionally high sensitivity
blood markers of early lesions may assessment of the stage
and specificity.
In the context of the TRACERx be masked by comorbidities, such of malignant disease in
study, Abbosh et al. investigated as chronic inflammatory diseases individual patients.
the potential of ctDNA analysis for [9], as well as by the accumulation
■■ Liquid biopsy tests have the
early diagnosis and monitoring in of cancer-related mutations with age
potential to aid in the detection
patients with non-small cell lung in healthy individuals [10,11].
of minimal residual disease.
carcinoma (NSCLC). The sequenc- These limitations may be illus-
ing of single-nucleotide variants in trated by the recently published work ■■ The presence of circulating
resected tumour tissue was used to of Cohen et al., who introduced the tumour cells as potential
create a patient-specific panel for CancerSeek panel for the detection seeds of distant metastases is
next-generation sequencing-based of the eight most common cancer highly predictive of metastatic
ctDNA analysis of plasma collected types [12]. This complex approach outgrowth and worse outcome
before surgery. With the detection combined the evaluation of eight in patients with both early-
soluble tumour biomarkers, including stage and late-stage disease.
threshold of at least two tumour-
standard tumour markers such as
specific single-nucleotide variants, ■■ Analysis of therapy-relevant
CEA, with ctDNA analysis of cancer-
the sensitivity of personalized tests genomic aberrations in circu-
related mutations in 16 genes. The
in pre-surgery plasma samples was
panel reached an overall median lating tumour cells and circulat-
97% for lung squamous cell carci-
sensitivity of 70%, with specificity of ing cell-free tumour DNA ena-
nomas but only 19% for lung adeno-
99% or more, but significant differ- bles the guidance of precision
carcinomas [8]. The authors calcu-
ences in sensitivity were observed therapy and the prediction of
lated that a tumour with a diameter
among the tumour types analysed, resistance to therapy.
of about 2.7 cm (volume, 10 cm3)
including 98% in ovarian cancer,
would result in a mean ctDNA plas-
60% in lung cancer, and 33% in
ma variant allele frequency of 0.1%. cancer face similar challenges to
breast cancer [12]. These findings
Modern low-dose computed tomog- require validation, ideally in an in- tests for early detection, including
raphy lung screening enables the dependent prospectively sampled, low concentrations of ctDNA and
detection of tumours of diameter pre-diagnostic cohort. Moreover, other liquid biomarker analytes [7].
0.4 cm (volume, 0.34 cm3), which the study analysed only healthy con- Tie et al. evaluated the ability of
would correspond to a plasma vari- trols; therefore, the high specificity of ctDNA analysis to detect minimal
ant allele frequency of 1.8 × 10 −4%, the CancerSeek approach requires residual disease in blood samples
below the detection limit of most further validation with non-cancer obtained from patients with stage II
current ctDNA technologies [8]. controls with comorbidities such as colon cancer after surgical removal
Another aspect is the cost of inflammatory diseases, which are of the primary tumour. The method
the patient-tailored next-generation common in ageing individuals. was able to predict recurrence at
sequencing-based ctDNA approach 36 months with a sensitivity of 48%
for the detection of single-nucleotide and a specificity of 100% [13]. In the
variants. Abbosh at al. estimated
ctDNA for monitoring of above-mentioned study of Abbosh
the current cost of targeted ctDNA minimal residual disease et al. in patients with lung cancer,
profiling to be US$ 1750 per pain patients with early- detection of ctDNA mutations that
tient, which is likely to be too high stage cancer were also present in the respective
SECTION 6
CHAPTER 6.7
for routine implementation as a test Liquid biopsy tests for the detection primary tumour was predictive of
for population cancer screening [8]. and monitoring of minimal residual relapse in 93% of cases, with a me-
These findings challenge the appli- disease in patients with early-stage dian of 70 days before radiological
Chapter 6.7 • Circulating DNA and other biomarkers for early diagnosis 551
confirmation [8]. Both of these stud- whether chromosome 1q23.1 ampli- provides a snapshot of the actual
ies demonstrate the feasibility and fication was enriched in tumour-initi- disease status. It has been shown
potential clinical value of ctDNA ating cells from patients with breast that regular enumeration of CTCs
analysis for monitoring of minimal cancer. Amplification of the region can be used for disease prognosis,
residual disease. However, ctDNA was detected in ctDNA as the aver- diagnosis of minimal residual dis-
detection required knowledge of age copy-number ratio of three genes ease, and monitoring of effective-
primary tumour-specific mutations, (TUFT1, S100A7, and S100A8) rela- ness of therapy [17–19].
and the mutational spectrum may tive to a reference gene by droplet Although reliable information
change during progression from digital polymerase chain reaction can easily be obtained in patients
minimal residual disease to overt (PCR). Detection of the amplifica- with advanced disease, patients with
metastatic disease. tion in ctDNA samples already at first early-stage cancer usually pres-
ctDNA analysis without prior diagnosis was predictive of relapse ent with very low concentrations of
knowledge of the genetics of the within 5 years in 67% of patients with CTCs. Nonetheless, a pooled analy-
primary tumour was applied in a re- early-stage breast cancer (stage I or sis including data from 3173 patients
cent study of patients with stage I–III II) and within 3 years in 40% of pa- with non-metastatic breast cancer
lung cancer. Chaudhuri et al. used tients with locally advanced breast (stage I–III) provided strong evi-
the highly sensitive cancer person- cancer (stage II or III), with 100% dence for CTCs as an independent
specificity in both cohorts [15]. prognostic factor with regard to poor
alized profiling by deep sequencing
Taken together, these results overall, breast cancer-specific, and
(CAPP-Seq) approach targeting 128
demonstrate the power of ctDNA disease-free survival [20]. Detection
genes that are recurrently mutated in
analysis to predict minimal residual of CTCs in patients with breast can-
lung cancer. Detection of ctDNA af-
disease in patients with cancer. cer receiving neoadjuvant therapy
ter the initial treatment of the primary
tumour was predictive of progres- is a significant predictor of outcome
sion in 72% of patients, with a me- CTCs for early detection independent of the response of the
and monitoring of primary tumour to therapy [21,22].
dian of 5.2 months before radiologi-
minimal residual disease This suggests that the presence
cal evidence of disease recurrence.
of CTCs signals the occurrence of
Remarkably, ctDNA was detectable Over the past decade, in addition to clinically relevant minimal residual
in 94% of patients experiencing re- the measurement of ctDNA, various disease at distant sites.
currence at the “minimal residual methods have been developed to Currently, most CTC assays rely
disease landmark” time point, which detect CTCs in the peripheral blood on epithelial markers such as EpCAM,
was defined as the first post-treat- of patients with cancer [16]. As for and most of the CTCs detected are
ment blood draw within 4 months of any other liquid biopsy analyte, single isolated cells. Despite the rel-
treatment completion [14]. quantification and characterization evance of epithelial–mesenchymal
Goh et al. used in vitro and pa- of CTCs in the blood of patients transition to cancer, the presence of
tient-derived xenograft assays to test with cancer at any particular time these “epithelial” CTCs is associated
with an unfavourable prognosis in
Fig. 6.7.2. A patient receiving chemotherapy in the context of cancer management. The cancer of the breast, prostate, colon,
currently available data suggest improved clinical management based on the power of and lung [23]. The clinical relevance
circulating cell-free tumour DNA (ctDNA) analysis to detect and monitor minimal residual of “mesenchymal” CTCs lacking any
disease in patients with cancer. epithelial markers as well as of CTC
clusters is still under investigation,
but the additional detection of these
subsets of CTCs may improve the
early detection of cancer and mini-
mal residual disease. The sensitivity
of current CTC assays seems to be
too low to enable them to be used
for early cancer detection. Only one
report has shown that detection of
CTCs in the blood of patients with
chronic obstructive pulmonary dis-
ease was able to predict the occur-
rence of lung cancer [24].
It has been shown that the pres-
ence of CTCs after completion of
adjuvant therapy is a predictor of
metastatic relapse and poor survival
552
[18]. Moreover, information provided could identify NSCLC with high ac- particular during the early stages
by CTCs may extend to the prote- curacy (AUC, 0.942) [28]. of cancer that are relevant to early
omic, transcriptomic, and genomic Recently, tumour-educated plate- detection programmes.
levels. Although single-cell analysis lets have emerged as new members Researchers have used various
is challenging, investigations of proof the family of liquid biopsy analytes. targeted DNA sequencing tech-
tein expression and genome-wide External stimuli, such as activation niques, such as digital PCR (quan-
studies on single cells are becoming of platelet surface receptors and titative PCR), BEAMing (beads,
the state of the art [25,26]. Molecular lipopolysaccharide-mediated platelet emulsion, amplification, magnetics)
characterization of CTCs provides a activation, induce specific splicing of technology, the safe-sequencing
powerful tool to assess intrapatient precursor messenger RNAs (mR- system, CAPP-Seq, and tagged-
heterogeneity and to obtain informa- NAs) in circulating tumour-educated amplicon deep sequencing [33].
tion about the clonal origin of CTCs platelets. The combination of specific These methods can reach ctDNA
and clonal selection under therapy. splice events in response to external detection limits of less than 0.01%.
The identification of clones that are signals and the capacity of platelets A disadvantage of these technolo-
sensitive and resistant to therapy to directly ingest (spliced) circulating gies is the requirement for detailed
may provide new insights and po- mRNA can provide tumour-educated prior information on the mutational
tential targets for cancer treatment. platelets with a highly dynamic mRNA spectrum of the tumour in the indi-
repertoire, with potential applicability vidual patient. This may be a limita-
to cancer diagnostics [6,29]. tion if these techniques are used for
Liquid biopsy beyond The first results on the use of cancer screening. Such information
ctDNA and CTC analyses tumour-derived exosomes and other is not required if non-targeted next-
In addition, the analysis of circulat- extracellular vesicles [30] are prom- generation sequencing is applied
ing non-coding nucleic acids such as ising, and their potential as cancer to investigate ctDNA, enabling the
miRNAs and lncRNAs (see Chapter biomarkers has been explored in genome-wide analysis of mutations
3.8) is a highly promising liquid biopsy multiple studies. However, the lack by whole-genome sequencing or
approach [4]. miRNAs and lncRNAs of standardization of protocols for whole-exome sequencing. However,
were found to provide additional lev- pre-analytical handling and ana- the drawbacks of genome-wide ctD-
els of transcriptional and translational lytical workflows limits interstudy NA analyses compared with target-
regulation and to be strongly involved comparisons and large international ed approaches include the need for
in cancer development. multicentre studies [31]. Moreover, higher concentrations of ctDNA and
Although levels of upregulation the investigation of extracellular ves- the lower overall assay sensitivity.
icles and their content in combina- In addition to next-generation
and downregulation of individual
tion with other liquid biopsy analytes sequencing-based mutation analy-
miRNAs or lncRNAs are probably in-
(e.g. CTCs, ctDNA) may provide sis (see Chapter 3.2), which is the
sufficient for a reliable test to detect
new opportunities for the develop- most prominent approach in ctDNA
cancer, signatures of 3–6 non-coding
ment of diagnostic tests [32]. analysis, copy number alteration
RNAs may be powerful and sensitive
In addition to ctDNA and CTCs, (CNA) and methylation analyses are
tools for early detection of cancer
the biomarkers discussed in this garnering substantial interest [34].
(reviewed in [4]). For example, a sig-
chapter provide information not Shallow whole-genome sequenc-
nature of serum miR-21 and miR-155
only on tumour cells but also on the ing of ctDNA, which enables the
was reported as a sensitive and spe- cost-effective global assessment of
tumour microenvironment – such as
cific biomarker for diagnosis of breast CNAs [35], has introduced the glob-
stromal and immune cells. This ad-
cancer; for miR-21 the receiver oper- al CNA score as a reliable biomarker
ditional information may be helpful
ating characteristic (ROC) area un- associated with active disease and
to detect the body’s response to the
der the curve (AUC) value was 0.788, survival in patients with melanoma.
development of small cancerous le-
the sensitivity was 66.67%, and the Similarly, genome-wide CNA as-
sions, which in turn could be used
specificity was 88.89%, and for miR- sessment has been used to screen
for early cancer detection.
155 the ROC AUC value was 0.749, cfDNA for the detection of incipient
the sensitivity was 100%, and the haematological malignancies in ap-
specificity was 51.02% [27]. Technologies for parently healthy individuals [36].
lncRNAs can also be suc- detection of ctDNA and Epigenomic tumour-specific al-
cessfully used in diagnostic tests. CTCs terations can be detected in ctDNA
Tang et al. found that three lncR- Circulating cell-free DNA (cfDNA) in and have the potential to serve as
NAs (LINC01627, LINC01628, and blood plasma is highly fragmented biomarkers. Shen et al. demon-
ERICH1-AS1) were upregulated in DNA derived mainly from apoptotic strated the ability to identify large-
SECTION 6
CHAPTER 6.7
the plasma of patients with NSCLC cells. The concentration of ctDNA scale tumour-specific ctDNA meth-
compared with healthy individuals. in blood may be less than 0.01% of ylation patterns [37]. The method
The suggested diagnostic signature the total cfDNA concentration, in they established was successfully
applied for cancer detection and Fig. 6.7.3. A women having blood drawn. Liquid biopsy is recognized as a means of
classification in a patient cohort indicating prognosis for patients with cancer, but its potential is also being explored for
across several tumour types [37]. the purpose of early diagnosis.
Besides large-scale methylation as-
sessment, smaller panels have the
benefit of being less expensive and
easier to interpret. Thus, methylation
of 12 genes investigated by droplet
digital methylation-specific PCR in
ctDNA was successfully applied to
accurately distinguish between pa-
tients with breast cancer and healthy
volunteers [38].
Furthermore, the physicochemi-
cal properties of methylated DNA
assessed as the methylation land-
scape of cfDNA could be used to
accurately discriminate between
healthy individuals and patients
with cancer (accuracy > 70%) [39].
These recent findings demonstrate
the high potential as biomarkers of Although some antigens are applica- biopsy should be performed to de-
cfDNA CNA and methylome analy- ble to various different cancer types termine whether the NSCLC mu-
ses. However, these findings require (e.g. keratins are suitable for cancers tations are present. Insofar as the
further validation in larger cohorts of the breast, colon, and prostate test provides positive results, it may
and groups of patients with early- and other epithelial tumours), tissue- benefit patients who may be too ill
stage cancer or benign disease. specific antigens such as prostate- or are otherwise unable to provide a
Efficient enrichment of CTCs specific antigen or breast-specific tumour specimen for EGFR testing.
can be achieved by approaches mammaglobin are also suitable. Blood is a rich source of infor-
that exploit the differences between mation through which solid can-
tumour cells and blood cells, includ- cers can be detected, classified,
From discovery to
ing the differential expression of cell and matched to a specific thera-
membrane proteins (e.g. EpCAM,
clinical validation and py. Different approaches such as
the most widely used marker for the utility ctDNA, non-coding nucleic acids,
enrichment of CTCs in blood from Currently, only two liquid biopsy extracellular vesicles, tumour-ed-
patients with carcinoma) as well as tests are approved in the USA by ucated platelets, or CTC analyses
different sizes, densities, electric the FDA, but not for the early detec- will provide complementary infor-
charges, and deformabilities [5,16]. tion of cancer. The FDA approved mation, depending on the tumour
After enrichment, the CTCs are still the above-mentioned system for type and the intended clinical use.
surrounded by hundreds to thou- detecting CTCs in metastatic can- Despite the potential of individual
sands of leukocytes, and therefore cer in 2018 and an EGFR mutation techniques, each has its own limi-
reliable methods are required to iden- test for ctDNA analysis in 2016 [3]. tations; this leads to the idea of
tify a CTC at the single-cell level. The EGFR mutation test can detect combining different analytes for the
CTCs can be detected by anti- EGFR gene mutations in patients early detection of cancer. Technical
bodies against membrane and cyto- with NSCLC. Such mutations are and clinical validation of assays is
plasmic antigens, including epithelial, present in about 10–20% of pa- very important and can be achieved
mesenchymal, tissue-specific, and tients with NSCLC. The EGFR mu- in independent, international con-
tumour-associated markers. Most tation test identifies the presence sortia such as the European IMI
current CTC assays use the same of 42 specific NSCLC mutations in Cancer-ID network (https://www.
identification step as the system ap- exons 18–21, including the L858R cancer-id.eu). Similar to the devel-
proved by the United States Food mutation, exon 19 deletions, and opment of new drugs, the pipeline
and Drug Administration (FDA) for the T790M mutation. On the basis for the development of new diag-
detecting CTCs in patients with meta- of these data, patients who may nostic tools needs more standardi-
static cancer: cells are fluorescently benefit from treatment with erlotin- zation to bridge the gap between
stained for epithelial keratins as a ib or osimertinib may be selected. the plethora of published biomark-
marker of CTCs, and CD45 is used However, if such mutations are not er studies and the paucity of new
as a leukocyte exclusion marker. detected in the blood, then a tumour markers entering clinical practice.
554
References
1. Pantel K, Alix-Panabières C (2010). 11. Fernandez-Cuesta L, Perdomo S, Avogbe 20. Janni WJ, Rack B, Terstappen LWMM,
Circulating tumour cells in cancer pa- PH, Leblay N, Delhomme TM, Gaborieau Pierga J-Y, Taran F-A, Fehm T, et al. (2016).
tients: challenges and perspectives. V, et al. (2016). Identification of circulat- Pooled analysis of the prognostic relevance
Trends Mol Med. 16(9):398–406. https:// ing tumor DNA for the early detection of of circulating tumor cells in primary breast
doi.org/10.1016/j.molmed.2010.07.001 small-cell lung cancer. EBioMedicine. cancer. Clin Cancer Res. 22(10):2583–93.
PMID:20667783 10:117–23. https://doi.org/10.1016/j.ebiom. https://doi.org/10.1158/1078-0432.CCR-15-
2016.06.032 PMID:27377626 1603 PMID:26733614
2. Babayan A, Pantel K (2018). Advances in
liquid biopsy approaches for early detec- 12. Cohen JD, Li L, Wang Y, Thoburn C, 21. Riethdorf S, Müller V, Loibl S, Nekljudova
tion and monitoring of cancer. Genome Afsari B, Danilova L, et al. (2018). V, Weber K, Huober J, et al. (2017).
Med. 10(1):21. https://doi.org/10.1186/ Prognostic impact of circulating tumor
Detection and localization of surgically
s13073-018-0533-6 PMID:29558971 cells for breast cancer patients treated
resectable cancers with a multi-analyte
in the neoadjuvant “Geparquattro” trial.
3. Kwapisz D (2017). The first liquid biopsy test blood test. Science. 359(6378):926–30.
Clin Cancer Res. 23(18):5384–93. https://
approved. Is it a new era of mutation testing https://doi.org/10.1126/science.aar3247
doi.org/10.1158/1078-0432.CCR-17-0255
for non-small cell lung cancer? Ann Transl PMID:29348365
PMID:28679772
Med. 5(3):46. https://doi.org/10.21037/atm.
2017.01.32 PMID:28251125 13. Tie J, Wang Y, Tomasetti C, Li L, Springer
22. Bidard FC, Michiels S, Riethdorf S,
S, Kinde I, et al. (2016). Circulating tu-
Mueller V, Esserman LJ, Lucci A, et al.
4. Anfossi S, Babayan A, Pantel K, Calin mor DNA analysis detects minimal re-
(2018). Circulating tumor cells in breast
GA (2018). Clinical utility of circulat- sidual disease and predicts recurrence cancer patients treated by neoadjuvant
ing non-coding RNAs – an update. Nat in patients with stage II colon cancer. chemotherapy: a meta-analysis. J Natl
Rev Clin Oncol. 15(9):541–63. https:// Sci Transl Med. 8(346):346ra92. https:// Cancer Inst. 110(6):560–7. https://doi.
d o i .o r g /10 .10 3 8 /s 415 71- 018 - 0 0 3 5 -x doi.org/10.1126/scitranslmed.aaf 6219 org/10.1093/jnci/djy018 PMID:29659933
PMID:29784926 PMID:27384348
23. Alix-Panabières C, Mader S, Pantel K
5. Poudineh M, Sargent EH, Pantel K, Kelley 14. Chaudhuri AA, Chabon JJ, Lovejoy AF, (2017). Epithelial-mesenchymal plasticity
SO (2018). Profiling circulating tumour
Newman AM, Stehr H, Azad TD, et al. in circulating tumor cells. J Mol Med (Berl).
cells and other biomarkers of invasive
(2017). Early detection of molecular 95(2):133–42. https://doi.org/10.1007/
cancers. Nat Biomed Eng. 2(2):72–84.
residual disease in localized lung can- s00109-016-1500-6 PMID:28013389
https://doi.org/10.1038/s41551-018-0190-5
cer by circulating tumor DNA profiling.
PMID:31015625 24. Ilie M, Hofman V, Long-Mira E, Selva E,
Cancer Discov. 7(12):1394–403. https://
6. Best MG, Sol N, In ’t Veld SGJG, Vancura doi.org/10.1158/2159-8290.CD-17-0716 Vignaud J-M, Padovani B, et al. (2014).
A, Muller M, Niemeijer AN, et al. (2017). PMID:28899864 “Sentinel” circulating tumor cells allow
Swarm intelligence-enhanced detection of early diagnosis of lung cancer in patients
non-small-cell lung cancer using tumor-ed- 15. Goh JY, Feng M, Wang W, Oguz G, with chronic obstructive pulmonary dis-
ucated platelets. Cancer Cell. 32(2):238– Yatim SMJM, Lee PL, et al. (2017). ease. PLoS One. 9(10):e111597. https://
252.e9. https://doi.org/10.1016/j.ccell.2017. Chromosome 1q21.3 amplification is a doi.org/10.1371/journal.pone.0111597
07.004 PMID:28810146 trackable biomarker and actionable target PMID:25360587
for breast cancer recurrence. Nat Med.
7. Bardelli A, Pantel K (2017). Liquid biop- 25. Wang D, Bodovitz S (2010). Single cell
23(11):1319–30. https://doi.org/10.1038/
sies, what we do not know (yet). Cancer analysis: the new frontier in ‘omics’.
nm.4405 PMID:28967919 Trends Biotechnol. 28(6):281–90. https://
Cell. 31(2):172–9. https://doi.org/10.1016/
j.ccell.2017.01.002 PMID:28196593 16. Alix-Panabières C, Pantel K (2016). doi.org/10.1016/j.tibtech. 2010.03.0 02
Clinical applications of circulating tumor PMID:20434785
8. Abbosh C, Birkbak NJ, Wilson GA, Jamal-
cells and circulating tumor DNA as liq- 26. Babayan A, Alawi M, Gormley M, Müller
Hanjani M, Constantin T, Salari R, et al.;
uid biopsy. Cancer Discov. 6(5):479–91. V, Wikman H, McMullin RP, et al. (2016).
TRACERx consortium; PEACE consor-
https://doi.org/10.1158/2159-8290.CD-15- Comparative study of whole genome
tium (2017). Phylogenetic ctDNA analysis
depicts early-stage lung cancer evolution. 1483 PMID:26969689 amplification and next generation se-
Nature. 545(7655):446–51. https://doi. quencing performance of single can-
17. Fehm T, Müller V, Alix-Panabières C,
org/10.1038/nature22364 PMID:28445469 cer cells. Oncotarget. 8(34):56066–80.
Pantel K (2008). Micrometastatic spread in
ht t p s: //d o i.o r g /10.18 6 32 /o nc ot a r g et .
9. Pantel K, Denève E, Nocca D, Coffy A, breast cancer: detection, molecular char-
10701 PMID:28915574
Vendrell JP, Maudelonde T, et al. (2012). acterization and clinical relevance. Breast
Circulating epithelial cells in patients Cancer Res. 10(Suppl 1):S1. https://doi. 27. Han JG, Jiang YD, Zhang CH, Yang YM,
with benign colon diseases. Clin Chem. org/10.1186/bcr1869 PMID:19091005 Pang D, Song YN, et al. (2017). A novel
58(5):936–40. https://doi.org/10.1373/ panel of serum miR-21/miR-155/miR-
18. Pantel K, Alix-Panabières C, Riethdorf 365 as a potential diagnostic biomarker
clinchem.2011.175570 PMID:22205690
S (2009). Cancer micrometastases. Nat for breast cancer. Ann Surg Treat Res.
10. Krimmel JD, Schmitt MW, Harrell MI, Rev Clin Oncol. 6(6):339–51. https:// 92(2):55–66. https://doi.org/10.4174/astr.
Agnew KJ, Kennedy SR, Emond MJ, et d o i . o r g / 1 0 .1 0 3 8 / n r c l i n o n c . 2 0 0 9 . 4 4 2017.92.2.55 PMID:28203552
al. (2016). Ultra-deep sequencing de- PMID:19399023
tects ovarian cancer cells in peritoneal 28. Tang Q, Ni Z, Cheng Z, Xu J, Yu H,
fluid and reveals somatic TP53 muta- 19. Pantel K, Alix-Panabières C (2013). Real- Yin P (2015). Three circulating long
tions in noncancerous tissues. Proc time liquid biopsy in cancer patients: fact non-coding RNAs act as biomark-
Natl Acad Sci U S A. 113(21):6005–10. or fiction? Cancer Res. 73(21):6384–8. ers for predicting NSCLC. Cell Physiol
https://doi.org/10.1073/pnas.1601311113 https://doi.org/10.1158/0008-5472.CAN- Biochem. 37(3):1002–9. https://doi.org/
PMID:27152024 13-2030 PMID:24145355 10.1159/000430226 PMID:26393913
SECTION 6
CHAPTER 6.7
29. Nilsson RJ, Balaj L, Hulleman E, van 33. Heitzer E, Haque IS, Roberts CES, 37. Shen SY, Singhania R, Fehringer G,
Rijn S, Pegtel DM, Walraven M, et Speicher MR (2019). Current and future Chakravarthy A, Roehrl MHA, Chadwick D,
al. (2011). Blood platelets contain tu- perspectives of liquid biopsies in geno- et al. (2018). Sensitive tumour detection and
mor-derived RNA biomarkers. Blood. mics-driven oncology. Nat Rev Genet. classification using plasma cell-free DNA
118(13):3680–3. https://doi.org/10.1182/ 20(2):71–88. https://doi.org/10.1038/s41576- methylomes. Nature. 563(7732):579–83.
blood-2011-03-344408 PMID:21832279 018-0071-5 PMID:30410101 https://doi.org/10.1038/s41586-018-0703-0
PMID:30429608
30. Hoshino A, Costa-Silva B, Shen TL, 34. Heitzer E, Ulz P, Geigl JB, Speicher
Rodrigues G, Hashimoto A, Tesic Mark MR (2016). Non-invasive detection of 38. Uehiro N, Sato F, Pu F, Tanaka S,
M, et al. (2015). Tumour exosome in- genome-wide somatic copy number al- Kawashima M, Kawaguchi K, et al.
tegrins determine organotropic me- terations by liquid biopsies. Mol Oncol. (2016). Circulating cell-free DNA-based
tastasis. Nature. 527(7578):329–35. 10(3):494–502. https://doi.org/10.1016/j. epigenetic assay can detect early breast
ht t p s: //d o i.o r g /10 .10 3 8 /n a t u r e15 75 6 cancer. Breast Cancer Res. 18(1):129.
molonc.2015.12.004 PMID:26778171
PMID:26524530 https://doi.org/10.1186/s13058-016-0788-z
35. Silva S, Danson S, Teare D, Taylor F, PMID:27993161
31. Lane RE, Korbie D, Hill MM, Trau M
Bradford J, McDonagh AJG, et al. (2018).
(2018). Extracellular vesicles as circu- 39. Sina AA, Carrascosa LG, Liang Z, Grewal
Genome-wide analysis of circulating cell-
lating cancer biomarkers: opportunities YS, Wardiana A, Shiddiky MJA, et al.
free DNA copy number detects active mel-
and challenges. Clin Transl Med. 7(1):14. (2018). Epigenetically reprogrammed
https://doi.org/10.1186/s40169-018-0192-7 anoma and predicts survival. Clin Chem. methylation landscape drives the DNA self-
PMID:29855735 64(9):1338–46. https://doi.org/10.1373/ assembly and serves as a universal cancer
clinchem.2018.290023 PMID:29941468 biomarker. Nat Commun. 9(1):4915. https://
32. Bracht JWP, Mayo-de-Las-Casas C, doi.org/10.1038/s 41467- 018 - 07214 -w
Berenguer J, Karachaliou N, Rosell R 36. Lenaerts L, Vandenberghe P, Brison N,
PMID:30514834
(2018). The present and future of liquid Che H, Neofytou M, Verheecke M, et al.
biopsies in non-small cell lung cancer: (2019). Genomewide copy number al-
combining four biosources for diagno- teration screening of circulating plasma
sis, prognosis, prediction, and disease DNA: potential for the detection of in-
monitoring. Curr Oncol Rep. 20(9):70. cipient tumors. Ann Oncol. 30(1):85–95.
https://doi.org/10.1007/s11912-018-0720-z https://doi.org/10.1093/annonc/mdy476
PMID:30030656 PMID:30371735
556
6.8
Governmental action to control

carcinogen exposure
Multiple options covering diverse scenarios
Vincent J. Cogliano Dorota Jarosińska (reviewer)
Sakari Karjalainen (reviewer)
Kurt Straif (reviewer)
an agent that is known or suspected lations, voluntary (non-enforceable)

SUMMARY to be a carcinogen through individu- guidelines, incentives, and educa-
al choice. Often, however, individu- tion campaigns that help individuals
●● Governmental action has been als cannot control – or sometimes make informed choices.
effective in reducing exposure do not even know about – their ex- Nongovernmental organizations
to known and suspected car- posure to carcinogens in the air they also develop guidelines, incen-
cinogens. These actions can breathe, the food and water they tives, and education campaigns.
involve legislation, regulation (to eat and drink, the places they work, Examples include the guideline for
eliminate or restrict exposure), or the products they can afford to primary prevention of cervical can-
enforcement of legislation and buy and use. This opens up a role cer from the American Society of
regulations, voluntary (non-en- for national governments and inter- Clinical Oncology, reduced insur-
forceable) guidelines, incentives, governmental organizations to act ance premiums for nonsmokers from
and education campaigns. in ways that complement individual various insurance organizations, the
choices to avoid exposure to car- SunSmart campaign from Cancer
●● Depending on the legal author-
cinogens. These actions can take Council Australia, and the European
ity, the basis of regulation can
several forms: legislation, regulation Code Against Cancer, which was
be hazard, exposure, or risk.
(to eliminate or restrict exposure), updated in 2014 (https://cancer-code-
●● Hazard-based regulation can enforcement of legislation and regu- europe.iarc.fr/index.php/en/).
be effective. Notable examples
include reduction of tobacco
Fig. 6.8.1. The European Parliament. After a specific cause of cancer in humans is
use and international action identified, national governments and intergovernmental organizations can act to prevent
to eliminate persistent organic or control exposure to the carcinogen.
pollutants.
●● New methods of toxicity testing
are emerging and transforming
the science of carcinogen iden-
tification. The goal is to identify
and evaluate cancer hazards on
the basis of data on precancer-
ous effects.
●● National and international health
agencies are still identifying ad-
ditional carcinogens.
After research identifies a cause of

cancer in humans, primary preven-
CHAPTER 6.8
SECTION 6
tion efforts can be directed towards

reducing human exposure. In some
cases, people can avoid exposure to
Chapter 6.8 • Governmental action to control carcinogen exposure 557

To provide authoritative, impar- protect vulnerable populations with-

tial scientific information on agents out attempting to quantify accept-
able levels of exposure or risk. For ex-
FUNDAMENTALS
that are known or suspected to be
carcinogens, several national and ample, governments worldwide have ■■ Historically, legislative action
international health agencies de- acted for several decades to prevent to reduce carcinogen exposure
velop evaluations of epidemiologi- smoking and other exposures to to- focused on measures directed
cal and experimental evidence on bacco and tobacco smoke, espe- towards the prevention of
carcinogenicity (Table 6.8.1). cially for young people. This chapter occupational cancer. Over
This chapter discusses exam- describes similar actions in response decades, legislation in many
ples of governmental action to con- to the recent identification of other countries has been based
trol carcinogen exposure, with a carcinogenic hazards to which there on recognition of specific
focus on developments during the is widespread exposure. chemicals as carcinogens in
past 5 years. Given the breadth of this context.
the subject, this chapter is not a Obesity and overweight
■■ Measures to prevent occu-
comprehensive global assessment; Worldwide, an estimated 640 mil- pational cancer indicate the
rather, it aims to provide up-to-date lion adults were obese (body mass scope of relevant initiatives,
examples of new developments, index ≥ 30 kg/m2) in 2014, a 6-fold which include prohibition of
relevant country experiences, and increase since 1975. An estimated certain chemicals, restriction
novel approaches. For a discussion 110 million children and adolescents of manufacturing processes to
of actions to control cervical cancer were obese in 2013, a doubling since reduce emissions, and manda-
and exposure to carcinogenic hu- 1980. If the people who are over- tory requirements for the use of
man papillomavirus (HPV) types, weight (body mass index ≥ 25 kg/m2 personal protective equipment.
see Chapters 5.10 and 6.4. and < 30 kg/m2) are also considered,
the totals are about triple. ■■ Involuntary exposure to
carcinogenic pollutants in the
In 2016, being obese or over-
Restrictions qualitatively weight was established as a risk fac-
air, water, and soil may be
based on hazard tor for cancers of the gastric cardia,
limited by the specification of
maximum levels of known or
In some cases, the identification of gall bladder, pancreas, ovary, and
suspected carcinogens.
an agent as a known or suspected thyroid, and for multiple myeloma
carcinogen can be sufficient basis and meningioma [1,2]. This added to ■■ The extent of possible
for action. Depending on the legal previous findings for cancers of the human exposure to known
authority, preventive measures can colorectum, oesophagus, kidney, carcinogens as a result of
using consumer products and
prescription drugs is usually
Table 6.8.1. Some sources of authoritative evaluations of carcinogenicity from gov- subject to regulation.
ernment agencies and intergovernmental organizations
■■ Governments may intervene
Authority Agency or programme in relation to carcinogen
National authorities exposure that occurs through
Australia National Industrial Chemicals Notification and individual choices.
Assessment Scheme, Department of Health
■■ In most countries, the sale of
Canada Health Canada tobacco products is limited by
USA Environmental Protection Agency regulations on advertising, pur-
Food and Drug Administration chase by children, packaging,
National Institute for Occupational Safety and Health and product identification.
National Toxicology Program Report on Carcinogens
■■ Governments may play a role
Occupational Safety and Health Administration
in education campaigns about,
Several state health or environmental agencies
for example, sun exposure and
International authorities tobacco use.
European Union European Chemicals Agency
European Food Safety Authority
Several national health agencies postmenopausal breast, and endo-
World Health Organization IARC (IARC Monographs programme) metrium [3]. When these newly es-
International Programme on Chemical Safety tablished cancer sites are included,
Joint Expert Committee on Food Additives (joint with the as much as 9% of the cancer burden
Food and Agriculture Organization of the United Nations) in women in North America, Europe,
Joint Meeting on Pesticide Residues (joint with the Food and the Middle East may be attribu-
and Agriculture Organization of the United Nations)
table to obesity.
558
In the past, obesity was viewed “Effective modern approaches for the income countries. Most of the users
as a matter of personal responsibility control of obesity”) [4]. Many of these are young women. Use of ultravio-
that could be controlled through in approaches could also be applicable let-emitting tanning devices is clas-
dividual choice. Governmental inter to the reduction of exposure to other sified by the IARC Monographs as
vention focused on education cam- known or suspected carcinogens. carcinogenic to humans (Group 1);
paigns and on taxation of unhealthy Public health interventions to reduce such devices cause malignant mela-
foods and beverages to urge individ- the prevalence of obesity are likely noma of the skin and eye. The risks
uals to adopt healthy lifestyles (see to accelerate with the recognition are higher for exposure at younger
Chapter 6.2). More recently, a wider that the cancer burden attributable ages (see Chapter 2.4). Risks of
variety of governmental interven- to obesity and overweight is greater cutaneous melanoma are higher
tions have been recognized as hav- than was previously believed. for people who first used tanning
ing value in reducing the prevalence devices before about age 30 years
of obesity. A recent survey described Ultraviolet-emitting tanning (overall relative risk, 1.75). Risks of
worldwide trends towards strength- devices ocular melanoma are higher for peo-
ening existing interventions and Indoor tanning using ultraviolet-emit- ple whose first use was before age
introducing novel approaches to re- ting devices, such as sunlamps and 20 years. There is also a positive as-
duce the prevalence of obesity (see sunbeds, is common in many high- sociation with risk of squamous cell
Effective modern approaches for the control of obesity
These approaches for the control have included trans fat content salted foods to include a govern-
of obesity [1] are also applicable to since 2008, there has been a ment health message.
other health concerns. documented decrease in levels of
Restrictions, standards, and
trans fatty acids in the population.
Stronger taxes bans on specific ingredients
For example, in 2014, the Navajo Built environment For example, many European
Nation in the USA imposed higher Obstacles to obtaining healthy countries have adopted legislation
taxes on sugar-sweetened bever- food include lack of supermarkets, that restricts the trans fat content
ages and foods high in salt, fat, lack of public transportation, and of foods. Also, Ghana has a law
and/or sugar, and eliminated tax- unsafe neighbourhoods. For ex- to restrict the fat content of meats,
es on fresh fruits, vegetables, and ample, since 2011, Canada has and several Pacific island coun-
nuts. Mexico placed an 8% tax on worked with supermarkets to pro- tries have banned the sale or im-
high-calorie foods in 2013 and a vide nutritious perishable foods to port of certain fatty animal parts.
10% tax on sugar-sweetened bev- more than 70 000 people living in
Screening to target high-risk
erages in 2014. isolated northern communities. individuals
Stronger educational messages School-based interventions For example, Japan has a law that
For example, in 2012, Western Many countries promote healthy requires adults to have their waist
Australia launched a campaign meals in schools or restrict the circumference measured annually
featuring graphic images of obese provision of unhealthy foods. Since and compared to population stan-
people coupled with messages 2013, at least 19 states in the USA dards. There are fines for employ-
about “toxic fat”. Evidence from re- have required schools to provide ers and local governments that do
search on anti-tobacco campaigns parents with body mass index as- not meet population health goals,
shows that advertising featuring sessments of their children. but no penalties for individuals.
powerful images and health warn- Sustainable agriculture,
Restrictions on advertising and
ings can affect public opinion. environment, and healthy food
marketing
Labelling Many countries have long re- These are integrated programmes
Labelling provides better informa- stricted advertising of unhealthy that engage government, multi-
tion on more food products. For foods directed at children. In ad- ple private-sector industries, and
example, since 2012, Cameroon dition, some countries are mov- stakeholders.
has mandated nutritional label- ing towards restricting advertis-
Reference
ling, and Chile, Ecuador, and the ing of certain products aimed at
1. Taylor AL, Parento EW, Schmidt LA
United Kingdom have introduced the broader public. For example,
(2015). The increasing weight of regula-
CHAPTER 6.8
SECTION 6
front-of-package, traffic-light la- France requires advertisements tion: countries combat the global obesity
belling. In the USA, where labels for processed, sweetened, or epidemic. Indiana Law J. 90(1):7.

Fig. 6.8.2. A woman using a sunbed. Brazil was the first country to ban indoor tanning 7 years, and sellers are required to
for people of all ages. disclose a phone’s specific absorp-
tion rate of energy [9]. In 2017, the
state of California in the USA is-
sued guidance to reduce exposure
to energy from mobile phones [10].
Occupational exposures
to chemical, physical, and
biological agents
Worldwide, an estimated 740 000
people per year die from exposure
to carcinogens in the workplace [11]
(see Chapter 2.10). Many such can-
cers occur in high-income countries,
because of longer life expectancies.
However, exposures can be higher
in low- and middle-income countries
if there is low compliance with safety
norms, if there is weak enforcement
of hazard control in workplaces, if
worker organizations are not strong
enough to ensure compliance with
standards, and/or if there is a large
informal economy that is not subject
to regular inspection [12].
The Globally Harmonized Sys
tem of Classification and Labelling
of Chemicals [13] is becoming an
international standard for the com-
munication of chemical hazards. The
Globally Harmonized System defines
two categories of carcinogenic haz-
carcinoma of the skin, especially for Mobile phones ards: known or presumed human
use before age 20 years [5]. carcinogens (Categories 1A and
Concern about children’s health is 1B) and suspected human carcino-
Soon after the announcement evident in some actions to reduce
of the IARC Monographs conclu- gens (Category 2). The European
exposure from mobile phones. Chil Chemicals Agency aligned its clas-
sions, Brazil became the first coun- dren hold phones closer to their
try to ban indoor tanning for people sification and labelling practices with
brains than adults do, and the bone the Globally Harmonized System in
of all ages, and Australia followed and marrow in children’s skulls have 2011, the United States Occupational
in 2015. In view of the higher sus- higher conductivity. Radiofrequency Safety and Health Administration in
ceptibility of younger users, age electromagnetic fields from mobile 2012, and the Scientific Committee
restriction has been a more com- phones have been classified by the on Occupational Exposure Limits
mon type of action. In Europe, 11 IARC Monographs as possibly car- for the European Union in 2017 [14].
countries ban indoor tanning under cinogenic to humans (Group 2B), Labelling provides workers with in-
age 18 years, as do New Zealand with positive associations for glio- formation on the potential hazards of
and each province in Canada. In ma and acoustic neuroma [8]. chemicals in the workplace.
the USA, 17 states ban commercial Although regulation of mobile
indoor tanning under age 18 years. phone use mostly aims to reduce
Most other states have restrictions distractions while driving or in the Restrictions quantita-
for minors, such as bans under classroom, some health agencies tively based on levels of
age 14–17 years or requirements have acted in response to the pos- exposure or risk
for parental consent or accompani- sible risk of cancer, citing the IARC Some laws require a quantitative
ment [6]. Research shows that laws Monographs findings. Since 2014, evaluation of exposure or risk be-
with age restrictions are effective Belgium has banned the sale and fore acting to reduce risks to ac-
in reducing rates of indoor tanning advertising of mobile phones de- ceptable levels. Determining what
among female students [7]. signed for children younger than level of risk is acceptable can entail
560
intense debate, especially when the derive an exposure–response rela- bacco products. Guidance proposed
scientific evidence is inconclusive tionship, often linear, to estimate risk in 2012 describes the need to dem-
or when the benefits and costs of as a function of exposure. Final regu- onstrate whether a new or modified
exposure reduction accrue to dif- latory limits consider these health- tobacco product will reduce levels of
ferent segments of the population. based estimates along with political, exposure to hazardous substances
Government agencies often dis- socioeconomic, technical, and other or will reduce the risk of tobacco-
tinguish between the underlying considerations, depending on the related disease.
health science (known as risk as- governing legislation (Fig. 6.8.3).
sessment) and the legal, political, In 2016, a new law amended the Incorporation of
social, economic, and technical as- United States Toxic Substances Con
pects of a decision (known as risk trol Act [16]. This law directs the En
increased understanding
management) [15]. vironmental Protection Agency to
and new types of
Risk assessment of carcinogens develop risk-based evaluations that information
generally proceeds in distinct steps consider individuals who may be at Although observational epidemiol-
(Fig. 6.8.3): (i) hazard identification greater risk than the general popu- ogy has led to the identification of
determines whether an agent can lation because of biological suscep- about 100 known human carcino-
cause cancer under some condi- tibility or higher exposure. The new gens, animal bioassays are the pri-
tions; (ii) dose–response assess- law also prescribes timelines to mary support for the identification of
ment describes cancer risk as a accelerate the pace of risk evalua- most suspected carcinogens. In the
function of exposure to the agent; tions. Most of the first 10 substances past decades, the pace of animal
(iii) exposure assessment identifies selected to undergo risk evaluation bioassays has slowed. The United
human exposure pathways and es- are classified by the Environmental States National Toxicology Program
timates the levels of human expo- Protection Agency as known or sus- published its first 200 technical re-
sure; and (iv) risk characterization pected carcinogens. ports during 1976–1982 (a period of
integrates these steps for a conclu- 6 years), the next 200 during 1982–
sion about cancer risk. Exposure and risk 1993 (11 years), and the most recent
assessment of tobacco 200 during 1993–2018 (25 years)
Occupational and products (https://ntp.niehs.nih.gov/results/
environmental exposures In 2009, a new law authorized the pubs/index.html). At the same time,
Many government agencies use two United States Food and Drug Admin data on cancer mechanisms have
approaches to set regulatory limits istration to regulate tobacco products. become increasingly pivotal, and in
for known or suspected carcinogens, The law mandates several preventive the IARC Monographs programme
although specific procedures and ter- measures that have been success- mechanistic data have led to the
minology differ. Threshold approach- fully implemented in other countries. classification of more than a dozen
es estimate an exposure level below It also provides a unique risk-based agents as known human carcino-
which carcinogenic effects should approach for evaluating claims of re- gens (https://monographs.iarc.fr/list-
not occur. Non-threshold approaches duced harm from new or modified to- of-classifications-volumes/).
Fig. 6.8.3. The steps involved in risk assessment and risk management.
Hazard Dose–Response
Identification Assessment
Can the agent How is exposure
cause cancer related to human
under some cancer risk?
Risk
conditions?
Management
Risk
Characterization Identify options
Integrate Hazard, Evaluate risks
Dose–Response, Legal and other
and Exposure Political considerations
Exposure Social Take action
Assessment
Economic
How are people Technical considerations
exposed?
CHAPTER 6.8
SECTION 6
To what levels are

they exposed?
Risk Assessment Risk Management

Fig. 6.8.4. The United States Capitol Building. In 2016, the 114th Congress passed a erties of chemicals, and read-across
new law on chemical safety to amend and update the Toxic Substances Control Act, approaches for filling data gaps.
which went into force in 1976.
In the USA, Section 4 of the
2016 law that amended the Toxic
Substances Control Act directs the
Environmental Protection Agency
to develop and implement alterna-
tive testing methods to reduce ver-
tebrate animal testing. Examples
include computational toxicology
and bioinformatics, high-throughput
screening methods, testing of cate-
gories of substances, tiered testing
methods, in vitro studies, systems
biology, and new methods identified
by authoritative validation bodies.
The role of international

agreements
International agreements are a means
for addressing global health and en-
Radically new methods of toxicity The overall goal is to identify and vironmental concerns when govern-
testing are emerging to transform or evaluate apical hazards (i.e. observ- ments acting alone cannot achieve
contribute to the science of carcino- able disease in vivo, such as can- the results they seek. International
gen identification. Rather than time- cer) on the basis of non-apical data.
agreements support and guide ac-
consuming tests of single chemicals The research question will shift from
tions at the national level by articulat-
in experimental animals, in vitro tests whether an agent causes cancer
ing general principles and areas of
on human cells or cell components when tested alone as a single agent
consensus. Details of implementation
are investigating the ability to perturb to whether an agent can contribute
are a matter for each country.
disease pathways. High-throughput to an increased incidence of cancer
assays can test thousands of chemi- that can involve multiple risk factors.
cals over a wide range of concentra- Full implementation will require a Fig. 6.8.5. A warning sign about con-
tions. Modelling will estimate human better understanding of human dis- tamination by polychlorinated biphe-
intake rates that yield target-tissue ease pathways, the development nyls (PCBs), which are listed under the
concentrations analogous to those of methods to incorporate the new Stockholm Convention on Persistent Or
ganic Pollutants.
that perturb disease pathways in data, characterization of the uncer-
vitro [17]. Pathways can involve mul- tainties associated with using the
tiple agents, some genetic and some new data, and the development of
environmental [18]. case studies to promote discussion
In the realm of exposure assess- and acceptance among scientists
ment, advances in environmental and stakeholders [21]. Acceptance
sampling technology, biomarkers, is critical if data on precancerous ef-
genomics, and informatics are ex- fects are to support the type of regu-
panding the ability to measure the lation that now requires extensive
exposome, which is the totality of en- animal testing or the demonstration
vironmental exposures received dur- of cancer in humans.
ing a lifetime. This will provide data In the European Union, the Regis
for evaluating interactions between a tration, Evaluation, Authorisation and
chemical agent and other chemical Restriction of Chemicals (REACH)
and non-chemical stressors, includ- regulation has encouraged the re-
ing gene–environment interactions placement of animal testing. The
(see Chapter 3.3). These approaches European Union is actively promot-
promise to facilitate specific linkages ing research into the development
of exposures to biological effects and and validation of alternatives to ani-
to indicate molecular pathways in- mal testing. Examples include quan-
volved in carcinogenesis [19,20] (see titative structure–activity relationship
Chapter 3.11). (QSAR) models for predicting prop-
562
Stockholm Convention on Table 6.8.2. Persistent organic pollutants listed under the Stockholm Convention
Persistent Organic Pollutants Persistent organic pollutant IARC Monographs classification a
The Stockholm Convention on Per Annex A: Elimination
sistent Organic Pollutants is a legally Aldrinb Group 2A
binding treaty initiated by the United Chlordaneb Group 2B
Nations Environment Programme Dieldrinb Group 2A
and adopted in 2001 (http://chm. Endrinb Group 3
pops.int/). Countries undertake to Heptachlor b Group 2B
Hexachlorobenzeneb Group 2B
eliminate or restrict the production,
Mirexb Group 2B
use, import, and export of persistent
Polychlorinated biphenyls (PCBs) b Group 1
organic pollutants, which can cross
Toxapheneb Group 2B
national boundaries, persist in the
Chlordecone Group 2B
environment, bioaccumulate, and
Short-chain chlorinated paraffins Group 2B
harm human health and the environ-
Decabromodiphenyl ether (commercial mixture) –
ment (see Chapter 2.9). To date, 181
Technical endosulfan and its related isomers –
countries plus the European Union Hexabromobiphenyl –
have ratified the treaty. There are 28 Hexabromocyclododecane –
listed pollutants, most of which are Hexabromodiphenyl ether and heptabromodiphenyl –
known or suspected human carcino- ether (commercial octabromodiphenyl ether)
gens (Table 6.8.2). Hexachlorobutadiene Group 3
An example of research translat- Alpha hexachlorocyclohexane Group 2B
ed into governmental action involves Beta hexachlorocyclohexane Group 2B
perfluorooctanoic acid (PFOA). In Lindane Group 1
2014, the IARC Monographs classi- Pentachlorobenzene –
fied PFOA as possibly carcinogenic Pentachlorophenol and its salts and esters Group 1
to humans (Group 2B), based in Polychlorinated naphthalenes –
part on evidence of testicular can- Tetrabromodiphenyl ether and pentabromodiphenyl –
ether (commercial pentabromodiphenyl ether)
cer and kidney cancer in humans
Annex B: Restriction
[22,23]. Subsequently, PFOA, its 4,4′-Dichlorodiphenyltrichloroethane (DDT) b Group 2A
salts, and PFOA-related compounds Perfluorooctane sulfonic acid (PFOS), its salts, and –
were proposed for listing under the perfluorooctane sulfonyl fluoride
Stockholm Convention. In addition Annex C: Unintentional production
to testicular cancer and kidney can- Polychlorinated dibenzo-para-dioxinsb Group 3
cer, the proposal cites thyroid dis- 2,3,7,8-Tetrachlorodibenzo-para-dioxin Group 1
ease, pregnancy-induced hyperten- Polychlorinated dibenzofuransb Group 3
sion, and high cholesterol as health 2,3,4,7,8-Pentachlorodibenzofuran Group 1
issues linked to PFOA. Chemicals proposed for listing
A recent example of national Dicofol Group 3
legislative action on persistent or- Pentadecafluorooctanoic acid (PFOA), its salts, and Group 2B
PFOA-related compounds
ganic pollutants is Section 6 of the
Perfluorohexane sulfonic acid, its salts, and related –
2016 law that amended the Toxic compounds
Substances Control Act. The law a
Group 1, carcinogenic to humans; Group 2A, probably carcinogenic to humans; Group 2B, possibly
specifies that exposure shall be re- carcinogenic to humans; Group 3, not classifiable as to its carcinogenicity to humans; –, not evaluated
duced “to the extent practicable” for (https://monographs.iarc.fr/agents-classified-by-the-iarc/).
certain persistent, bioaccumulative, The 12 initial persistent organic pollutants.

b
and toxic substances. The law does

not require risk evaluation for these WHO Framework Convention use is the leading cause of cancer
substances, only a reasonable ba- on Tobacco Control worldwide. To date, 180 countries
sis to conclude that there is a toxic, plus the European Union have rati-
The WHO Framework Convention
persistent, bioaccumulative hazard. fied the treaty. The 2018 global prog-
on Tobacco Control is a legally
This is similar to the treatment of
binding treaty initiated by WHO ress report on the implementation of
these substances in the European
and adopted in 2003 (http://www. the WHO Framework Convention on
Union, where the aim of REACH is
the substitution of persistent, bio- who.int/fctc/en/). Concerted inter- Tobacco Control documents many
national action was undertaken to national examples of effective action
CHAPTER 6.8
SECTION 6
accumulative, and toxic chemicals,

and the minimization of exposures in address the globalization of the to- to reduce the prevalence of tobacco
the interim [24]. bacco epidemic, given that tobacco use in adults and children [25].

In 2018, a legally binding sup- system, and through law enforce- Disclaimer
plement, the Protocol to Eliminate ment and other measures to enable
The views expressed in this chap-
Illicit Trade in Tobacco Products, international cooperation.
was ratified (http://www.who.int/fctc/ ter are those of the author and do
protocol/en/). The protocol provides not necessarily represent the views
tools to prevent illicit trade by secur- or the policies of the United States
ing the supply chain, by establishing Environmental Protection Agency.
an international tracking and tracing
References
1. Lauby-Secretan B, Scoccianti C, Loomis D, 10. California Department of Public Health 18. Goodson WH 3rd, Lowe L, Carpenter
Grosse Y, Bianchini F, Straif K; International (2017). How to reduce exposure to ra- DO, Gilbertson M, Manaf Ali A, Lopez
Agency for Research on Cancer Handbook diofrequency energy from cell phones. de Cerain Salsamendi A, et al. (2015).
Working Group (2016). Body fatness and Available from: https://www.cdph.ca.gov/ Assessing the carcinogenic potential of
cancer – viewpoint of the IARC Working Pro gr ams /C C D PH P/ D EO D C / EH I B / low-dose exposures to chemical mixtures
Group. N Engl J Med. 375(8):794–8. CDPH%20Document%20Librar y/Cell- in the environment: the challenge ahead.
https://doi.org/10.1056/NEJMsr1606602 Phone-Guidance.pdf. Carcinogenesis. 36(Suppl 1):S254–96.
PMID:27557308 https://doi.org/10.1093/carcin/bgv039
11. Hämäläinen P, Takala J, Kiat TB (2017). PMID:26106142
2. IARC (2018). Absence of excess body Global estimates of occupational ac-
fatness. IARC Handb Cancer Prev. 16:1– cidents and work-related illnesses 19. National Research Council (2012).
646. Available from: http://publications. 2017. Singapore: Workplace Safety and Exposure science in the 21st century: a
iarc.fr/570. Health Institute. Available from: http:// vision and a strategy. Washington (DC),
www.icohweb.org/site/images/news/pdf/ USA: National Academies Press. https://
3. IARC (2002). Weight control and physical Rep o r t %2 0 G l o bal%2 0 Est imate s%2 0 doi.org/10.17226/13507
activity. IARC Handb Cancer Prev. 6:1– of%20Occupational%20Accidents%20
315. Available from: http://publications. 20. Wild CP, Scalbert A, Herceg Z (2013).
and%20Work-related%20Illnesses%20 Measuring the exposome: a powerful ba-
iarc.fr/376. 2017%20rev1.pdf. sis for evaluating environmental exposures
4. Taylor AL, Parento EW, Schmidt LA 12. Santana VS, Ribeiro FSN (2011). and cancer risk. Environ Mol Mutagen.
(2015). The increasing weight of regula- Occupational cancer burden in develop- 54(7):480–99. https://doi.org/10.1002/em.
tion: countries combat the global obesity ing countries and the problem of informal 21777 PMID:23681765
epidemic. Indiana Law J. 90(1):7. workers. Environ Health. 10(Suppl 1):S10. 21. National Academies of Sciences,
https://doi.org/10.1186/1476-069X-10-S1- Engineering, and Medicine (2017). Using
5. IARC (2012). Radiation. IARC Monogr
S10 PMID:21489206 21st century science to improve risk-relat-
Eval Carcinog Risks Hum. 100D:1–437.
Available from: http://publications.iarc. ed evaluations. Washington (DC), USA:
13. United Nations (2017). Globally Harmonized
fr/121 PMID:23189752 National Academies Press. https://doi.
System of Classification and Labelling of
org/10.17226/24635
Chemicals (GHS). Seventh revised edition.
6. CDC (2018). Indoor tanning is not safe.
Available from: http://www.unece.org/trans/ 22. Benbrahim-Tallaa L, Lauby-Secretan
Atlanta (GA), USA: Centers for Disease
danger/publi/ghs/ghs_rev07/07files_e.html. B, Loomis D, Guyton KZ, Grosse Y, El
Control and Prevention. Available from:
https://www.cdc.gov/cancer/skin/basic_ Ghissassi F, et al.; International Agency for
14. Scientific Committee on Occupational
Research on Cancer Monograph Working
info/indoor_tanning.htm. Exposure Limits (SCOEL) (2017). Group (2014). Carcinogenicity of per-
Methodology for derivation of occupa- fluorooctanoic acid, tetrafluoroethylene,
7. Guy GP Jr, Berkowitz Z, Jones SE, Olsen
tional exposure limits of chemical agents. dichloromethane, 1,2-dichloropropane,
EO, Miyamoto JN, Michael SL, et al.
Luxembourg: Publications Office of the and 1,3-propane sultone. Lancet Oncol.
(2014). State indoor tanning laws and ado-
European Union. Available from: https:// 15(9):924–5. https://doi.org/10.1016/S1470-
lescent indoor tanning. Am J Public Health.
op.europa.eu/en/public ation - detail/- / 2045(14)70316-X PMID:25225686
104(4):e69–74. https://doi.org/10.2105/
publication/3c8ef3e0 - 48fc-11e8 -be1d-
AJPH.2013.301850 PMID:24524515
01aa75ed71a1. 23. IARC (2017). Some chemicals used as sol-
8. IARC (2013). Non-ionizing radiation, Part vents and in polymer manufacture. IARC
15. National Research Council (1983). Risk Monogr Eval Carcinog Risks Hum. 110:1–
2: Radiofrequency electromagnetic fields.
assessment in the federal government: 276. Available from: http://publications.
IARC Monogr Eval Carcinog Risks Hum.
managing the process. Washington (DC), iarc.fr/547
102:1–460. Available from: http://publica
USA: National Academies Press. https://
tions.iarc.fr/126 PMID:24772662 24. ECHA (2018). Management of PBT/vPvB
doi.org/10.17226/366
substances under REACH. Helsinki,
9. Belgian Federal Public Service Health,
16. United States Congress (2016). Frank R. Finland: European Chemicals Agency.
Food Chain Safety and Environment
Lautenberg Chemical Safety for the 21st Available from: https://echa.europa.eu/
(2014). New rules for selling mobile phones.
Century Act. Available from: https://www. management-of-pbt-vpvb-substances.
Available from: https://www.health.belgium.
congress.gov/114/plaws/publ182/PLAW-
b e /site s /default / f ile s /up loads / f ie lds / 25. WHO (2018). Global progress report on
114publ182.pdf.
fpshealth_theme_file/19099488/FAQ_ implementation of the WHO Framework
KBs_mobiele_telefoon_version5_EN.pdf. 17. National Research Council (2007). Convention on Tobacco Control. Geneva,
Toxicity testing in the 21st century: a vision Switzerland: World Health Organization.
and a strategy. Washington (DC), USA: Available from: https://www.who.int/fctc/
National Academies Press. https://doi. rep o r t ing / W H O - FCTC -2 018 _ gl o bal _
org/10.17226/11970 progress_report.pdf.
564
6.9
Prevention strategies common

to noncommunicable diseases
A focus on tobacco, alcohol, obesity,
and physical inactivity
David J. Hunter Kunjan Kunjan (reviewer) Christopher P. Wild (reviewer)
K. Srinath Reddy Neil Pearce (reviewer)
noncommunicable diseases. relevant to four disease clusters:

SUMMARY National cancer control pro- cancers, cardiovascular diseases,
grammes should seek potential chronic respiratory diseases, and
●● In 2016 there were 40.5 million synergies with programmes for diabetes, which together account
deaths from noncommunicable the prevention of other non- for about 78% of global deaths from
diseases worldwide, accounting communicable diseases in re- NCDs [3]. According to WHO esti-
for 72% of all deaths globally in lation to alcohol consumption, mates, in 2016 there were 40.5 mil-
that year. diet, and physical exercise. lion deaths from NCDs worldwide,
●● Tobacco use is estimated to accounting for 72% of all deaths
cause 22% of cancers world- globally in that year [3]. About 78%
wide and contributes to multiple With increases in life expectancy and of the NCD-related deaths oc-
other diseases. the growth of populations, more peo- curred in low- and middle-income
ple worldwide are living into the age countries, which also had a high
●● A range of dietary factors, in- groups of peak cancer incidence. proportion of deaths in middle age.
cluding alcohol consumption, Many cancer prevention strategies This staggering toll of NCDs and
that are implicated in cancer eti- are specific to cancer, such as hu- premature mortality in low- and
ology are also relevant to risk of man papillomavirus (HPV) vaccina- middle-income countries reflects
cardiovascular disease, result- tion. Some strategies for cancer pre- the transition in the main causes
ing in similar dietary recommen- vention also reduce the risk of other of death – from maternal and child
dations for both disease types. noncommunicable diseases (NCDs). deaths and infectious and parasitic
●● An estimated 24% of disability- The United Nations and WHO diseases to NCDs.
called for a 25% reduction in prema- Cardiovascular diseases are the
adjusted life years lost due to
ture deaths (i.e. at ages 30–69 years) biggest contributor to NCD-related
tracheal, bronchial, and lung
from NCDs by 2025, compared with deaths, followed by cancer, chronic
cancers worldwide are attribu-
2010, with a slogan of “25 by 25” respiratory diseases, and diabetes
table to air pollution (both in-
[1]. This was later modified within (Fig. 6.9.1) [3]. High-income coun-
door and outdoor), which also
the Sustainable Development Goals tries have a lower burden of mater-
contributes to the burden of
agenda to an overarching target
cardiovascular disease, stroke, nal and child deaths and infectious
(Target 3.4) of reducing the total pre-
and chronic obstructive pulmo- diseases, and therefore a higher
mature mortality from NCDs by one
nary disease. proportional mortality due to NCDs.
third by 2030, relative to 2015 [2].
However, because low- and mid-
●● In some high-income coun- Win–win strategies that have
dle-income countries have larger
tries, the mortality rates of non- benefits across several NCDs are
population sizes, they have a larger
communicable diseases have attractive in attempting to reach this
absolute number of deaths due to
peaked – particularly with re- goal and are reviewed in this chapter.
NCDs.
spect to cardiovascular diseas-
Surprisingly, age-standardized
es and, possibly, cancer. Burden of disease death rates due to NCDs are also
●● Reducing the prevalence of Four common behavioural risk fac- higher in low- and middle-income
CHAPTER 6.9
SECTION 6
tobacco smoking is key to re- tors – tobacco use, excess alco- countries than in high-income coun-
ducing the risk of many can- hol consumption, unhealthy diet, tries. For example, rates of car-
cer types as well as other and lack of physical activity – are diovascular disease and death are
Chapter 6.9 • Prevention strategies common to noncommunicable diseases 565

Fig. 6.9.1. Global distribution of number of deaths (thousands) within noncommunica-

ble diseases, 1990–2016. COPD, chronic obstructive pulmonary disease; CVD, car-
GLOBALdiseases;
diovascular DISTRIBUTION OF Deaths
DM, diabetes mellitus.(thousands) within NCD’s, 1990-2016
FUNDAMENTALS
■■ The current toll of noncom-
20 000 municable diseases and
premature mortality in low-
18 000 and middle-income countries
16 000 reflects the transition in the
main causes of death – from
14 000 maternal and child deaths
and infectious and parasitic
12 000
diseases to noncommunica-
1990
10 000 ble diseases.
2010 ■■ Some risk factors for cancer,
8 000
2016
such as tobacco use, alcohol
6 000 consumption, unhealthy diet,
4 000
and lack of physical activity,
also contribute to the burden
2 000 of other noncommunicable dis-
eases, notably cardiovascular
0
diseases, chronic respiratory
CVD Cancer COPD DM Others diseases, and diabetes.
■■ Tobacco smoking causes
multiple tumour types,
substantially higher in low- and mid- Prevalence of risk factors respiratory disease, and
dle-income countries even though cardiovascular disease.
the prevalence of risk factors is low- Behavioural risk factors ■■ Excess alcohol consumption
er [4]. This may reflect the relatively The four common behavioural risk causes several cancer types
unprepared state of health systems factors that contribute to the etiol- and also cardiovascular
in low- and middle-income countries ogy of NCDs can all be subject to disease and stroke.
in responding to this fresh challenge intervention: tobacco use, alco-
posed by the rapid health transition ■■ Policy interventions to
hol consumption, unhealthy diet,
by providing pharmacological thera- reduce the prevalence of
and lack of physical activity. All
these risk factors are likely
pies and revascularization to those four of these factors contribute to
to be win–win strategies
at risk [4]. increased cancer incidence and
with benefits across several
The four main disease clusters mortality, although for most cancer noncommunicable diseases.
(cancers, cardiovascular diseases, types there are no readily measur-
chronic respiratory diseases, and able intermediate non-malignant in- ■■ Individual behaviour change
diabetes) account for a lower pro- dicators other than obesity, where- should lower personal risk of
portion of disability-adjusted life as the prevention of cardiovascular multiple noncommunicable
years (DALYs) lost compared with disease and stroke benefits from diseases, including cancer.
NCDs in the neuropsychiatric, mus- measurable intermediate indica- Primary care workers can be
culoskeletal, renal, hearing, and vi- tors, such as blood pressure and trained to offer advice about
sion clusters. hypercholesterolaemia. the reduction of risk factors as
WHO has estimated that the Tobacco use is estimated to well as about understanding
absolute number and proportion of the signs and symptoms of
cause 22% of cancers worldwide,
deaths due to NCDs will increase the major noncommunicable
and alcohol consumption 7% [6].
diseases, to promote earlier
worldwide, rising to about 70% of The role of diet and physical activity
diagnosis and referral for
all deaths in 2030 [5]. This trend is in cancer may be mediated mainly
treatment.
mainly as a result of increases in through obesity or may be at least
the size and age of the world’s pop- partially independent. It has been
ulation, as well as continuing reduc- noted that in Asian populations, in-
tions in child mortality and deaths creased body fat and visceral adi- therefore, the associations of diet
from infectious diseases. Projected posity pose risks for NCDs at body and physical inactivity with NCDs
increases in the prevalence of risk mass index thresholds that are low- may have been underestimated in
factors for NCDs also contribute. er than the conventional criteria [7]; these populations.
566
It has been particularly difficult carcinogenic to humans (Group 1) dynamic diameter less than 2.5 µm
to specifically characterize diet as a and indoor emissions from house- (PM2.5), are thought to be mainly
risk factor for cancer, with the ex- hold combustion of biomass fuel responsible for the excess in lung
ceptions of contaminants such as (primarily wood) as probably carci- cancer, because these particles can
aflatoxin contamination of mouldy nogenic to humans (Group 2A). In penetrate deeply into the lungs. In
foods as a cause of liver cancer, 2013, the IARC Monographs classi- addition to particulate matter, air-
and arsenic in drinking-water as a fied outdoor air pollution as carcino- borne gases such as ozone, carbon
cause of bladder cancer and skin genic to humans (Group 1) and esti- monoxide, nitrogen dioxide, and sul-
cancer [8]. In the World Cancer mated that 223 000 deaths per year fur dioxide may be associated with
Research Fund/American Institute worldwide (about 15% of all deaths risk of diseases such as asthma,
for Cancer Research (WCRF/AICR) from lung cancer) are attributable to cardiovascular disease, and stroke.
2018 Expert Report, the only other outdoor air pollution.
diet–cancer relationship for which On the basis of estimates from
Experience in high-
the evidence was categorized as the Global Burden of Disease Study
convincing was between consump- 2015, The Lancet Commission on
income countries
tion of processed meat and risk of Pollution and Health estimated In some high-income countries,
colorectal cancer [8]. However, the that 24% of DALYs lost due to tra- the mortality rates of NCDs have
WCRF/AICR committee catego- cheal, bronchial, and lung cancers peaked – particularly with respect
rized a large number of associations worldwide are attributable to air to cardiovascular diseases and,
as probable, and these directions of pollution, both indoor and outdoor, possibly, cancer. The 60-year
association are mostly considered with a higher burden in low- and trends in the USA (Fig. 6.9.2) show
to be the same for risk of cardiovas- middle-income countries [10]. The that age-adjusted mortality rates for
cular disease. Hence, dietary rec- association of air pollution with cardiovascular diseases have de-
ommendations for cancer and car- cardiovascular disease, stroke, creased by about 75% from a peak
diovascular disease largely overlap; and chronic obstructive pulmonary in the 1960s, those for cerebrovas-
they both emphasize consuming a disease means that much larger cular disease have decreased by
largely plant-based diet and eating numbers of DALYs lost and deaths 78%, and those for cancer have
whole foods rather than processed due to these other NCDs are attrib- decreased by 17% since 1980
foods [8,9]. uted to air pollution than for cancer. (https://w w w.cdc.gov/nchs/data/
However, reduction in exposure to hus/2011/024.pdf).
Air pollution air pollution would be predicted to Similar reductions in the inci-
In recent years, the contribution reduce the incidence of all four of dence of cardiovascular diseases
of air pollution to NCDs has be- these NCDs. and of lung cancer in men have been
come far more widely appreciated. The precise components of air seen in many high-income countries
In 2006, the IARC Monographs pollution that are causal are not fully [11]. However, despite these declines
classified indoor emissions from identified. Particulates, notably par- in age-adjusted risks, reductions in
household combustion of coal as ticulate matter with particles of aero- the absolute number of deaths per
Fig. 6.9.2. Decline in age-adjusted death rates per 100 000 people for major noncommunicable diseases in the USA, 1950–2010.
1200
1000
800
Per 100,000
Diabetes
Chronic lower respiratory diseases
600
Malignant neoplasms
400 Cerebrovascular diseases
Diseases of the heart
200
CHAPTER 6.9
SECTION 6
0
1960 1970 1980 1990 2000 2010
Year

year are smaller, because of popu- ing leading to early detection and ternational cooperation in tobacco
lation growth and ageing. The es- treatment [14]. Therefore, the ex- and agricultural policies; (ii) promo-
timates from the Global Burden of perience in high-income countries tion of health literacy, to increase
Disease Study 2017 of age- and sex- suggests that the size of the NCD self-efficacy in avoiding risks and
specific rates for deaths and DALYs epidemic is not predetermined, and maintaining health; and (iii) health
lost due to NCDs globally demon- the challenge for low- and middle- services that combine timely and
strate a substantial reduction in inci- income countries is whether they cost-effective management of NCD
dence rates, but because of increas- can intervene sufficiently early to risk factors and clinically manifest
ing population sizes and population mitigate the epidemic. NCDs (Table 6.9.1).
ageing the absolute number of cases However, forecasting of future The major priority for cancer
continues to increase [12]. cancer rates suggests an increas- prevention is tobacco control (see
Modelling suggests that reducing global burden in the absence of Chapter 2.1), which could prevent
tions in the prevalence of risk fac- major interventions. The projections about 29% of all cancer deaths in
tors explain about 44–76% of the from the Global Burden of Disease the USA and also greatly reduce the
decline in mortality from coronary Study 2017 of the leading causes of number of deaths from cardiovascu-
heart disease in the USA and other years of life lost predict that between lar disease and chronic respiratory
high-income countries, and im- 2016 and 2040, cancers of the lung, disease [16]. Tobacco smoking is
proved treatments and access to liver, colorectum, and breast will the second largest cause of deaths
treatments explain about 23–47% move up the rankings, as a result worldwide [10]. Recent experience
of the decline [13]. The causes of a combination of changes in the documents that when the preva-
of the decline in cancer mortali- prevalence of risk factors, popula- lence of smoking declines, there are
ty rates in the USA are less well tion growth and ageing, and declin- almost immediate reductions in the
quantified, although a reduction in ing mortality from NCDs [15]. incidence of myocardial infarction
lung cancer mortality in men as a and hospital admissions for asthma,
result of a decrease in the preva- and the incidence of lung cancer de-
lence of smoking is clearly a major Prevention and control of creases within a decade [17].
contributor. Over the 20th century, NCDs The most effective way to re-
mortality rates from cervical cancer A comprehensive NCD control produce tobacco use is by increasing
decreased dramatically in high-in- gramme should include: (i) policy the price of tobacco products (see
come countries, mainly because of interventions that assist people to Chapter 6.1), and the most effec-
organized cervical cancer screen- avoid risky behaviours, including in- tive way to do this is by increasing
Table 6.9.1. Opportunities for the prevention, detection, and treatment of noncommunicable diseases in low- and middle-income
countries
Category Prevention Detection Treatment
Government policy Anti-tobacco policy Promotion of awareness of Ensure access to affordable

Regulation and labelling of processed foods NCDs, signs and symptoms, essential medicines
and high-sugar beverages and need for early detection
Planning for safe, healthy environments that

promote physical activity and limit transition
to sedentary lifestyle
Mitigation of harmful effects of alcoholic
beverages
Reduction in outdoor air pollution; provision
of cleaner fuels where indoor air pollution
due to burning of coal or biomass occurs
Health system Intersectoral planning for health promotion Surveillance for risk factor and Facilities and equipment for
Training of health personnel, including task- NCD prevalence affordable treatments
shifting for cancer detection and treatment Facilities and equipment for Recognition of need for both
low-cost detection of patients acute and chronic treatment
who should be referred for of NCDs
cancer workup
Clinicians Counselling of patients in risk factor Evaluation of intermediate Evidence-based treatment with
reduction risk factors; lifestyle and drug affordable essential medicines
Treatment for tobacco addiction interventions to lower risk Procedural or surgical
factor profiles interventions if appropriate
Appropriate screening (e.g.
HPV detection)
HPV, human papillomavirus; NCDs, noncommunicable diseases.
568
Fig. 6.9.3. This display advertisement income for those who are financially The evidence shows that sugar-
from Nepal illustrates one of the many dependent on growing tobacco. sweetened beverages are important
aspects of tobacco control plans. The increase in the prevalence of causes of childhood obesity, and
obesity is predicted to result in an in- substitution of lower-calorie options
crease in cancer incidence and mor- is associated with weight loss in ran-
tality (see Chapter 2.7). This implies domized trials [20,21]. Taxes on sug-
that prevention of obesity is a prior- ar-sweetened beverages have been
ity for the prevention of cancer, as successful in lowering consumption,
well as cardiovascular diseases and particularly in Central and South
diabetes. The environment in much America, and preliminary evidence
of the world has been described as suggests some reductions in the
obesogenic because of the increas- prevalence of obesity (see Chapter
ing availability of lower-cost pro- 6.2). However, much larger societal
cessed foods, combined with lower changes will be needed in intersec-
levels of daily physical activity. toral management of agriculture and
Fig. 6.9.4. This poster is part of the “no fast food” campaign in Azerbaijan.
excise taxes. In both France and

South Africa, tripling the price of cig-
arettes halved cigarette consumption
in less than 15 years and doubled
tobacco revenues to the state, which
could be used to fund other smoking-
reduction activities, such as advertis-
ing and nicotine replacement therapy
[18].
Worldwide, the age-standard-
ized prevalence of daily smok-
ing in 2016 was estimated to be
25% in men and 5.4% in women
[19]. Between 1990 and 2015, the
global age-standardized preva-
lence decreased by 28% in men
and by 34% in women; there was
substantial heterogeneity across
countries both in smoking preva-
lence and in change in prevalence
[19]. Implementation of the WHO
Framework Convention on Tobacco
Control, called for in Target 3a of
the United Nations Sustainable
Development Goals, is still patchy,
and a greatly decreased prevalence
of smoking will be required to coun-
ter the demographic effects of an
increase in the younger age groups
that the tobacco industry targets to
become new smokers. Increasing
CHAPTER 6.9
SECTION 6
the quit rates among current smok-

ers is also critical. Another part of
the solution is alternative sources of

the food supply, as well as urban de- tion has recently become an issue vascular diseases, chronic obstruc-
sign to promote healthier transport in several cities in sub-Saharan tive pulmonary disease, stroke, and
options. In many countries, the im- Africa. Policies that reduce levels of diabetes. HPV vaccination and cervi-
pact of rapid urbanization has meant air pollution are urgently needed, to cal cancer screening are also a “best
that these considerations are given a reduce the burden of air pollution- buy” but would not be predicted to
low priority. related morbidity and mortality. directly alter the risk of other NCDs.
Exposure to indoor air pollution Early-life vaccination against
has become less prevalent glob- hepatitis B virus has sharply reduced Health system
ally but is still highly prevalent in the prevalence of chronic hepatitis
low- and middle-income countries; B virus infection (see Chapter 5.6),
challenges
about 3 billion people worldwide are and thus the incidence of liver can- The global NCD epidemic challeng-
exposed to household air pollution, cer [25]. The recent development es all health systems, although the
which accounts for an estimated of direct-acting antiviral agents that challenges vary according to the
3.5–4 million deaths per year [22]. can cure hepatitis C virus infection level of development. In low- and
Exposure to outdoor air pollution in more than 95% of people who middle-income countries, limited fi-
has increased substantially in recent take a 12-week course offers the nancial protection from the costs of
decades (see Chapter 2.9). In 2016, potential to remove hepatitis C virus cancer treatment drives many peo-
an estimated 95% of the world’s infection as a cause of liver cancer. ple into bankruptcy. The health-care
population lived in areas with ambi- Reductions in the prevalence of in- infrastructure is inadequate to meet
ent PM2.5 levels that exceeded the fections with hepatitis B virus and the needs, with limited facilities for
WHO air quality guideline of 10 μg/ hepatitis C virus will also reduce the advanced care and shortages of
m3 for outdoor PM2.5 (annual aver- incidence of non-cancer liver dis- trained health workers. In general,
age), and 58% lived in areas with eases, such as cirrhosis. the health systems are configured
levels that exceeded 35 μg/m3 [23]. Many of these interventions have to provide acute episodic care and
The sources of PM2.5 vary sub- been identified by WHO as being need to be adapted to provide chron-
stantially geographically. Sand is cost-effective. A set of 16 “best buys” ic continuous care across multiple
a major component in North Africa out of 88 interventions have been disciplines. Although investment in
and the Middle East. In India, burn- selected on the basis of cost–effec- secondary and tertiary hospitals
ing of crop wastes, construction tiveness and feasibility of implemen- may provide the physical facilities
dust, and vehicular emissions com- tation [26]. Five of these are policies for cancer care, the specialization
bine to create high levels of outdoor designed to reduce the prevalence involved means that economies of
air pollution in the growing metropo- of tobacco use, three are to reduce scale or clinical experience may not
lis of Delhi, which is surrounded by alcohol consumption, and one aims be readily achievable between the
agricultural states. In China, coal- to increase physical activity. These treatment of cancer and treatment
fired power plants, automobiles, interventions involve legislative ac- of other NCDs. This contrasts with
and industrial facilities are thought tions, public awareness campaigns, the win–win component of a joint
to be the dominant contributors to and public health interventions. approach to reducing the preva-
air pollution in the Beijing–Shanghai These steps would be expected to lence of risk factors.
corridor [24]. Outdoor air pollu- decrease the risk of cancers, cardio- In many countries, access to es-
sential drugs is not assured. As a re-
sult, some countries, such as India
Fig. 6.9.5. Part of the “This Girl Can” campaign from Sport England, which encourages and Thailand, are resorting to com-
women to get active. pulsory licensing to domestically
produce the more expensive cardio-
vascular or anticancer drugs. Many
patients with cancer are deprived
of low-cost drugs such as morphine
that can provide pain relief; this is
due to both national policies and in-
ternational regulations that restrict
the trade in opioid drugs [27].
There is a need to train and
deploy non-physician health-care
workers in primary care to provide
appropriate referral for potential
cancer cases in an attempt to detect
cancers at an earlier stage. This is
a complex endeavour because of a
570
lack of knowledge in many popula- at three levels (state, capital, and health-care providers. The global
tions about the signs and symptoms district), is being pioneered by the movement for universal health cov-
of cancer and its potential to be treat- Tata Trust in India to offer afford- erage means that many countries
ed. Necessary diagnostic facilities able clinical care closer to patients’ are adopting policies that provide
include imaging, biopsy, and histo- homes [28]. Telemedicine and mo- greater financial protection to peo-
pathology. Treatment facilities range bile phone consultations may be ple for health care, including the
from outpatient oncology treatments helpful in initial assessment before more treatable cancer types [29].
to surgery and/or radiotherapy. Early referral as well as in continuing clini- It is essential to ensure that na-
presentation by patients, along with cal management. tional cancer control programmes
rapid referral, diagnosis, and treat- The costs of acute interventions actively explore potential synergies
ment initiation, requires substantial and chronic care for NCDs, includ- with the prevention and acute and
specialist staffing. The development ing cancer, are a formidable bar- chronic care of other NCDs.
of local and regional cancer centres, rier for patients, governments, and
References
1. United Nations (2011). Political declaration 8. WCRF/AICR (2018). Diet, nutrition, physi- 14. Mathew A, George PS (2009). Trends in
of the high-level meeting of the General cal activity and cancer: a global perspec- incidence and mortality rates of squamous
Assembly on the prevention and control tive. Continuous Update Project Expert cell carcinoma and adenocarcinoma of
Report 2018. World Cancer Research cervix – worldwide. Asian Pac J Cancer
of non-communicable diseases. Available
Fund/American Institute for Cancer Prev. 10(4):645–50. PMID:19827887
from: http://www.un.org/ga/search/view_
Research. Available from: https://www.
doc.asp?symbol=A/66/L.1. wcrf.org/dietandcancer. 15. Foreman KJ, Marquez N, Dolgert A,
Fukutaki K, Fullman N, McGaughey M,
2. United Nations (2015). Transforming our 9. American Heart Association (2019). The et al. (2018). Forecasting life expectancy,
world: the 2030 Agenda for Sustainable American Heart Association diet and life- years of life lost, and all-cause and cause-
Development. Available from: https://sus style recommendations. Available from:
specific mortality for 250 causes of death:
tainabledevelopment.un.org/post2015/ https://www.heart.org/en/healthy-living/
reference and alternative scenarios for
transformingourworld. healthy-eating/eat-smart/nutrition-basics/
2016-40 for 195 countries and territories.
aha-diet-and-lifestyle-recommendations.
3. WHO (2018). Global health estimates Lancet. 392(10159):2052–90. https://doi.
10. Landrigan PJ, Fuller R, Acosta NJR, Adeyi org /10.1016/S0140 - 673 6(18) 316 9 4 - 5
2016: deaths by cause, age, sex, by coun-
O, Arnold R, Basu NN, et al. (2018). The PMID:30340847
try and by region, 2000–2016. Geneva,
Lancet Commission on Pollution and
Switzerland: World Health Organization. Health. Lancet. 391(10119):462–512. 16. Islami F, Goding Sauer A, Miller KD,
Available from: https://www.who.int/health https://doi.org/10.1016/S0140- 6736(17) Siegel RL, Fedewa SA, Jacobs EJ, et al.
info/global_burden_disease/estimates/en/. 32345-0 PMID:29056410 (2018). Proportion and number of cancer
cases and deaths attributable to poten-
4. Yusuf S, Rangarajan S, Teo K, Islam S, 11. Ezzati M, Riboli E (2012). Can noncom- tially modifiable risk factors in the United
Li W, Liu L, et al.; PURE Investigators municable diseases be prevented? States. CA Cancer J Clin. 68(1):31–54.
(2014). Cardiovascular risk and events Lessons from studies of populations and h t t p s: //d o i . o r g /10 . 3 3 2 2 /c a a c . 214 4 0
in 17 low-, middle-, and high-income individuals. Science. 337(6101):1482–7. PMID:29160902
https://doi.org/10.1126/science.1227001
countries. N Engl J Med. 371(9):818–27.
PMID:22997325 17. Glantz S, Gonzalez M (2012). Effective
tobacco control is key to rapid progress in
PMID:25162888 12. Kyu HH, Abate D, Abate KH, Abay SM,
reduction of non-communicable diseas-
Abbafati C, Abbasi N, et al.; GBD 2017
5. Mathers CD, Loncar D (2006). Projections DALYs and HALE Collaborators (2018). es. Lancet. 379(9822):1269–71. https://
of global mortality and burden of dis- Global, regional, and national disability- doi.org/10.1016/S0140-6736(11)60615-6
ease from 2002 to 2030. PLoS Med. adjusted life-years (DALYs) for 359 diseas- PMID:21963004
3(11):e442. https://doi.org/10.1371/journal. es and injuries and healthy life expectancy
18. Jha P, Peto R (2014). Global effects
pmed.0030442 PMID:17132052 (HALE) for 195 countries and territories,
of smoking, of quitting, and of taxing
1990-2017: a systematic analysis for the
tobacco. N Engl J Med. 370(1):60–8.
6. Institute for Health Metrics and Evaluation Global Burden of Disease Study 2017.
https://doi.org/10.1056/NEJMra1308383
(2019). Global Burden of Disease Study Lancet. 392(10159):1859–922. https://
doi.org/10.1016/S0140-6736(18)32335-3 PMID:24382066
2017 (GBD 2017) data resources: GBD
results tool. Available from: http://ghdx. PMID:30415748 19. GBD 2015 Tobacco Collaborators (2017).
healthdata.org/gbd-results-tool. 13. Ford ES, Capewell S (2011). Proportion Smoking prevalence and attributable dis-
of the decline in cardiovascular mortality ease burden in 195 countries and territo-
7. Krishnaveni GV, Yajnik CS (2017). De ries, 1990-2015: a systematic analysis from
disease due to prevention versus treat-
velopmental origins of diabetes – an In the Global Burden of Disease Study 2015.
CHAPTER 6.9
SECTION 6
ment: public health versus clinical care.

dian perspective. Eur J Clin Nutr. 71(7): Annu Rev Public Health. 32(1):5–22. Lancet. 389(10082):1885–906. https://
865–9. https://doi.org/10.1038/ejcn.2017.87 ht t ps: //doi.o rg /10.114 6 /annurev- publ doi.org/10.1016/S0140-6736(17)30819-X
PMID:28537579 health-031210-101211 PMID:21417752 PMID:28390697

20. de Ruyter JC, Olthof MR, Seidell JC, Katan 24. Rohde RA, Muller RA (2015). Air pollu- 28. Sirohi B, Chalkidou K, Pramesh CS,
MB (2012). A trial of sugar-free or sugar- tion in China: mapping of concentrations Anderson BO, Loeher P, El Dewachi O,
sweetened beverages and body weight in and sources. PLoS One. 10(8):e0135749. et al. (2018). Developing institutions for
children. N Engl J Med. 367(15):1397–406. ht t ps: //doi.org /10.1371/ jour nal.p one. cancer care in low-income and middle-in-
https://doi.org/10.1056/NEJMoa1203034 0135749 PMID:26291610 come countries: from cancer units to com-
PMID:22998340 prehensive cancer centres. Lancet Oncol.
25. Plymoth A, Viviani S, Hainaut P (2009).
Control of hepatocellular carcinoma 19(8):e395–406. https://doi.org/10.1016/
21. Ebbeling CB, Feldman HA, Chomitz VR,
through hepatitis B vaccination in areas S1470-2045(18)30342-5 PMID:30102234
Antonelli TA, Gortmaker SL, Osganian
SK, et al. (2012). A randomized trial of of high endemicity: perspectives for glob-
al liver cancer prevention. Cancer Lett. 29. Frenk J, de Ferranti D (2012). Universal
sugar-sweetened beverages and ado- health coverage: good health, good eco-
lescent body weight. N Engl J Med. 286(1):15–21. https://doi.org/10.1016/j.
canlet.2009.08.024 PMID:19836128 nomics. Lancet. 380(9845):862–4. https://
367(15):1407–16. https://doi.org/10.1056/ doi.org/10.1016/S0140-6736(12)61341-5
NEJMoa1203388 PMID:22998339 26. WHO (2017). Tackling NCDs: ‘best PMID:22959372
buys’ and other recommended interven-
22. Gordon SB, Bruce NG, Grigg J, Hibberd
tions for the prevention and control of
PL, Kurmi OP, Lam KB, et al. (2014).
noncommunicable diseases. Geneva,
Respiratory risks from household air pol-
lution in low and middle income countries. Available from: https://apps.who.int/iris/
Lancet Respir Med. 2(10):823–60. https:// handle/10665/259232.
doi.org/10.1016/S2213-2600(14)70168-7
PMID:25193349 27. Knaul FM, Farmer PE, Krakauer EL, De
Lima L, Bhadelia A, Jiang Kwete X, et al.;
23. Shaddick G, Thomas ML, Amini H, Broday Lancet Commission on Palliative Care and
D, Cohen A, Frostad J, et al. (2018). Data Pain Relief Study Group (2018). Alleviating
integration for the assessment of popula- the access abyss in palliative care and pain
tion exposure to ambient air pollution for relief – an imperative of universal health
Global Burden of Disease assessment. coverage: the Lancet Commission report.
Environ Sci Technol. 52(16):9069–78. Lancet. 391(10128):1391–454. https://
https://doi.org/10.1021/acs.est.8b02864 doi.org/10.1016/S0140-6736(17)32513-8
PMID:29957991 PMID:29032993
572
Contributors
Christian C. Abnet Bruce K. Armstrong Rosamonde E. Banks
National Cancer Institute The University of Western Australia University of Leeds
Rockville, Maryland Perth Leeds
abnetc@mail.nih.gov and R.Banks@leeds.ac.uk
The University of Sydney
Clement A. Adebamowo Sydney Partha Basu
University of Maryland School of bruce@brucekarmstrong.org International Agency for Research
Medicine on Cancer
Baltimore, Maryland Patricia Ashton-Prolla Lyon
cadebamowo@ihv.umaryland.edu Universidade Federal do Rio basup@iarc.fr
Grande do Sul
Demetrius Albanes and Linda Bauld
National Cancer Institute Hospital de Clínicas de Porto The University of Edinburgh
Bethesda, Maryland Alegre Edinburgh
daa@nih.gov Porto Alegre and
pprolla@hcpa.edu.br Cancer Research UK
Laia Alemany Vilches London
Institut Català d’Oncologia (ICO) – Dagfinn Aune Linda.Bauld@ed.ac.uk
Institut d’Investigació Biomèdica de Imperial College London
Bellvitge (IDIBELL) London Iacopo Baussano
and and International Agency for Research
CIBER en Epidemiología y Salud on Cancer
Bjørknes University College
Pública (CIBERESP) Lyon
Oslo
Barcelona baussanoi@iarc.fr
d.aune@imperial.ac.uk
lalemany@iconcologia.net
Laura E. Beane Freeman
Anssi Auvinen
Maribel Almonte National Cancer Institute
Tampere University
International Agency for Research Bethesda, Maryland
Tampere
on Cancer freemala@mail.nih.gov
Lyon anssi.auvinen@tuni.fi
almontem@iarc.fr Sonja I. Berndt
Anna Babayan National Cancer Institute
Devasena Anantharaman University Medical Center Bethesda, Maryland
Rajiv Gandhi Centre for Hamburg-Eppendorf berndts@mail.nih.gov
Biotechnology Hamburg
Thiruvananthapuram Anna.Babayan@qiagen.com Margherita Bignami
devasena@rgcb.res.in Istituto Superiore di Sanità
Chunxue Bai Rome
Annie S. Anderson Zhongshan Hospital, Fudan margherita.bignami@gmail.com
University of Dundee University
Dundee and Maria Blettner
a.s.anderson@dundee.ac.uk Chinese Alliance Against Lung Universitätsmedizin der Johannes
Cancer Gutenberg Universität Mainz
Benjamin O. Anderson and Mainz
University of Washington International Society for blettner@uni-mainz.de
and Respiratory Diseases
Fred Hutchinson Cancer Research and Ron Borland
Center Chinese Society of e-Health Cancer Council Victoria
Seattle, Washington Shanghai Melbourne, Victoria
banderso@fredhutch.org bai.chunxue@zs-hospital.sh.cn ron.borland@cancervic.org.au
World Cancer Report 573

Freddie Bray Karen Canfell Aaron J. Cohen

International Agency for Research Cancer Council NSW Health Effects Institute
on Cancer and Boston, Massachusetts
Lyon The University of Sydney and
brayf@iarc.fr Sydney, New South Wales Institute for Health Metrics
karen.canfell@nswcc.org.au and Evaluation, University of
Paul Brennan Washington
International Agency for Research Bochen Cao
Seattle, Washington
on Cancer acohen@healtheffects.org
Geneva
Lyon
caob@who.int
brennanp@iarc.fr Graham A. Colditz
Franco Cavalli Washington University in St. Louis
Louise A. Brinton Foundation for the Institute of School of Medicine
National Cancer Institute Oncology Research St. Louis, Missouri
Bethesda, Maryland Ospedale San Giovanni colditzg@wustl.edu
brinton6610@gmail.com Bellinzona
Franco.Cavalli@eoc.ch Pietro Comba
Jennifer D. Brooks Istituto Superiore di Sanità
University of Toronto Stephen J. Chanock Rome
Toronto, Ontario National Cancer Institute
and
jennifer.brooks@utoronto.ca Bethesda, Maryland
WHO Collaborating Centre
chanocks@mail.nih.gov
for Environmental Health in
Julia Brotherton
Isabelle Chemin Contaminated Sites
VCS Foundation
Inserm U1052 – Centre de Rome
Melbourne, Victoria
Recherches en Cancérologie pietro.comba@iss.it
jbrother@vcs.org.au
de Lyon
Lyon David I. Conway
Karen Brown
isabelle.chemin@inserm.fr University of Glasgow
University of Leicester
Glasgow
Leicester Chien-Jen Chen David.Conway@glasgow.ac.uk
kb20@leicester.ac.uk Academia Sinica
Taipei Ian A. Cree
Laia Bruni chencj@gate.sinica.edu.tw International Agency for Research
Institut Català d’Oncologia (ICO)
on Cancer
Barcelona Wanqing Chen
lbruni@iconcologia.net Lyon
National Cancer Center, Chinese
Academy of Medical Sciences creei@iarc.fr
Nele Brusselaers and
Karolinska Institutet Peking Union Medical College Jack Cuzick
Stockholm Beijing Queen Mary University of London
nele.brusselaers@ki.se chenwq@cicams.ac.cn London
j.cuzick@qmul.ac.uk
Christopher Bullen Zhengming Chen
The University of Auckland University of Oxford Luigino Dal Maso
Oxford Centro di Riferimento Oncologico
Auckland
zhengming.chen@ndph.ox.ac.uk (CRO), IRCCS
c.bullen@auckland.ac.nz
Aviano
Zvavahera Mike Chirenje
Gloria M. Calaf dalmaso@cro.it
University of Zimbabwe
Universidad de Tarapacá Harare
Arica je@uzchs-ctrc.org Diona L. Damian
and The University of Sydney
Columbia University Medical Vincent J. Cogliano and
Science Environmental Protection Agency Royal Prince Alfred Hospital
New York, New York Washington, DC Sydney, New South Wales
gmc24@cumc.columbia.edu cogliano.vincent@gmail.com Diona.Damian@health.nsw.gov.au
574
Robert D. Daniels Eugenia Dogliotti Jessica N. Everett

Centers for Disease Control and Istituto Superiore di Sanità NYU Langone Health
Prevention Rome New York, New York
and eugenia.dogliotti@iss.it Jessica.Everett@nyulangone.org
National Institute for Occupational
Safety and Health Veronika Fedirko
Laure Dossus
Cincinnati, Ohio Emory University
International Agency for Research
rtd2@cdc.gov Atlanta, Georgia
on Cancer veronika.fedirko@emory.edu
Lyon
George Davey Smith
dossusl@iarc.fr Ian S. Fentiman
University of Bristol
Bristol The Harley Street Breast Clinic
KZ.Davey-Smith@bristol.ac.uk Ronny Drapkin London
University of Pennsylvania isf@ianfentiman.co.uk
Louise Davies Perelman School of Medicine
and Jacques Ferlay
Geisel School of Medicine
Basser Center for BRCA International Agency for Research
and
on Cancer
The Dartmouth Institute for and
Lyon
Health Policy & Clinical Practice, Penn Ovarian Cancer Research
ferlayj@iarc.fr
Dartmouth College Center
Lebanon, New Hampshire Philadelphia, Pennsylvania Pietro Ferrari
Louise.Davies@dartmouth.edu rdrapkin@pennmedicine.upenn. International Agency for Research
edu on Cancer
Sanford M. Dawsey Lyon
National Cancer Institute Eric J. Duell ferrarip@iarc.fr
Rockville, Maryland Institut Català d’Oncologia (ICO)
dawseys@mail.nih.gov Barcelona Miranda M. Fidler-Benaoudia
eduell@idibell.cat CancerControl Alberta, Alberta
Harry J. de Koning Health Services
Erasmus University Medical Center Calgary, Alberta
Karin Ekström Smedby
Rotterdam Miranda.Fidler-Benaoudia@ahs.ca
Karolinska Institutet
h.dekoning@erasmusmc.nl
Stockholm James Flanagan
Catherine de Martel Karin.Ekstrom.Smedby@ki.se Imperial College London
International Agency for Research London
on Cancer A. Heather Eliassen j.flanagan@imperial.ac.uk
Lyon Brigham and Women’s Hospital
demartelc@iarc.fr and Silvia Franceschi
Harvard Medical School Centro di Riferimento Oncologico
Lynette Denny Boston, Massachusetts (CRO), IRCCS
South African Medical Research nhahe@channing.harvard.edu Aviano
Council Gynaecological Cancer silvia.franceschi@cro.it
Research Centre Steffen Emmert David O. Francis
and University Medical Center Rostock University of Wisconsin–Madison
University of Cape Town Rostock Madison, Wisconsin
Cape Town steffen.emmert@med.uni-rostock. dofrancis@wisc.edu
lynette.denny@uct.ac.za de
Neal D. Freedman
Carol E. DeSantis National Cancer Institute
Karen M. Emmons
American Cancer Society Bethesda, Maryland
Harvard T.H. Chan School of
Atlanta, Georgia freedmanne@mail.nih.gov
Public Health
carol.desantis@cancer.org
Boston, Massachusetts
Christine M. Friedenreich
Joanna Didkowska kemmons@hsph.harvard.edu
CancerControl Alberta, Alberta
Polish National Cancer Registry Health Services
and Carolina Espina and
Maria Skłodowska-Curie Institute – International Agency for Research University of Calgary
Oncology Center on Cancer Calgary, Alberta
Warsaw Lyon Christine.Friedenreich@
joanna.didkowska@coi.pl espinac@iarc.fr albertahealthservices.ca

Peter P. Fu Rüdiger Greinert Rolando Herrero

Food and Drug Administration Elbe Clinics International Agency for Research
Jefferson, Arkansas Buxtehude on Cancer
peter.fu@fda.hhs.gov ruediger.greinert@elbekliniken.de Lyon
herreror@iarc.fr
Koraljka Gall Trošelj John D. Groopman
Ruđer Bošković Institute Johns Hopkins Bloomberg School Rayjean J. Hung
Zagreb of Public Health Lunenfeld-Tanenbaum Research
troselj@irb.hr and Institute, Sinai Health System
Johns Hopkins School of Medicine Toronto, Ontario
Judy E. Garber Baltimore, Maryland and
Dana-Farber Cancer Institute jgroopm1@jhu.edu Dalla Lana School of Public Health,
and University of Toronto
Harvard Medical School Giuseppe Grosso Toronto, Ontario
Boston, Massachusetts University of Catania rayjean.hung@lunenfeld.ca
Judy_Garber@dfci.harvard.edu Catania
giuseppe.grosso@studium.unict.it David J. Hunter
Gail Garvey University of Oxford
Menzies School of Health Marc Gunter Oxford
Research, Charles Darwin International Agency for Research david.hunter@ndph.ox.ac.uk
University on Cancer
Casuarina, Northern Territory Lyon Ivano Iavarone
gail.garvey@menzies.edu.au gunterm@iarc.fr Istituto Superiore di Sanità
and
Gemma Gatta Jason Gurney WHO Collaborating Centre
Fondazione IRCCS Istituto University of Otago for Environmental Health in
Nazionale dei Tumori Contaminated Sites
Wellington
Rome
Milan jason.gurney@otago.ac.nz
ivano.iavarone@iss.it
Gemma.Gatta@istitutotumori.mi.it
Kathryn Z. Guyton
André M. Ilbawi
Cindy L. Gauvreau International Agency for Research
International Agency for Research on Cancer
Geneva
on Cancer Lyon
ilbawia@who.int
Lyon guytonk@iarc.fr
cindy.gauvreau@utoronto.ca
Lisa Iversen
Bothwell Takaingofa Guzha University of Aberdeen
Adi F. Gazdar (deceased) University of Zimbabwe Aberdeen
Hamon Center for Therapeutic Harare l.iversen@abdn.ac.uk
Oncology Research bothwellguzha@gmail.com
and Charles W. Jameson
University of Texas Southwestern Janet Hall CWJ Consulting, LLC
Medical Center Inserm U1052 – Centre de Cape Coral, Florida
Dallas, Texas Recherches en Cancérologie drjameson@embarqmail.com
de Lyon Dorota Jarosińska
Ophira Ginsburg Lyon WHO European Centre for
NYU Langone Health janet.hall@inserm.fr Environment and Health
and Bonn
NYU School of Medicine Susan E. Hankinson jarosinskad@who.int
New York, New York University of Massachusetts
Ophira.Ginsburg@nyulangone.org Amherst, Massachusetts Mazda Jenab
and International Agency for Research
Edward L. Giovannucci Brigham and Women’s Hospital on Cancer
Harvard T.H. Chan School of Boston, Massachusetts Lyon
Public Health shankinson@schoolph.umass.edu jenabm@iarc.fr
and
Brigham and Women’s Hospital Zdenko Herceg Mattias Johansson
and International Agency for Research International Agency for Research
Harvard Medical School on Cancer on Cancer
Boston, Massachusetts Lyon Lyon
egiovann@hsph.harvard.edu hercegz@iarc.fr johanssonm@iarc.fr
576
Michael E. Jones Alan Prem Kumar Sarah Lewis

The Institute of Cancer Research National University of Singapore University of Bristol
and Singapore Bristol
Royal Cancer Hospital and S.J.Lewis@bristol.ac.uk
London Cancer Science Institute of
Michael.Jones@icr.ac.uk Singapore Donghui Li
Singapore The University of Texas
Shaoqing Ju and MD Anderson Cancer Center
Affiliated Hospital of Nantong Curtin University Houston, Texas
University Perth, Western Australia dli@mdanderson.org
Nantong csiapk@nus.edu.sg
jrjr2020@163.com He Li
Kunjan Kunjan National Cancer Center, Chinese
Rudolf Kaaks Academy of Medical Sciences
Postgraduate Institute of Medical
German Cancer Research Center and
Education and Research
(DKFZ) Peking Union Medical College
Chandigarh
Heidelberg Beijing
dr.kunjan2human@gmail.com
r.kaaks@Dkfz-Heidelberg.de Lihe_2017@163.com
Carlo La Vecchia Terry Lichtor
Sakari Karjalainen Università degli Studi di Milano Rush University Medical Center
Cancer Society of Finland Milan Chicago, Illinois
Helsinki carlo.lavecchia@unimi.it Terry_Lichtor@rush.edu
Sakari.Karjalainen@cancer.fi
Dirk W. Lachenmeier Martha S. Linet
Ausrele Kesminiene Chemical and Veterinary National Cancer Institute
International Agency for Research Investigation Agency Karlsruhe Bethesda, Maryland
on Cancer Karlsruhe linetm@mail.nih.gov
Lyon Dirk.Lachenmeier@CVUAKA.
kesminienea@visitors.iarc.fr BWL.DE Johan P. Mackenbach
Erasmus University Medical Center
Timothy J. Key Marc Ladanyi Rotterdam
University of Oxford Memorial Sloan Kettering Cancer j.mackenbach@erasmusmc.nl
Oxford Center
tim.key@ndph.ox.ac.uk New York, New York Núria Malats
ladanyim@mskcc.org Spanish National Cancer Research
Malcolm King Centre (CNIO)
Mississaugas of the New Credit Béatrice Lauby-Secretan Madrid
First Nation International Agency for Research nmalats@cnio.es
and on Cancer
University of Saskatchewan Lyon Reza Malekzadeh
Saskatoon, Saskatchewan secretanb@iarc.fr Tehran University of Medical
malcolm.king@usask.ca Sciences
Tehran
Dominique Laurier
Manolis Kogevinas malek@tums.ac.ir
Institute for Radiological Protection
Barcelona Institute for Global and Nuclear Safety
Health Mohandas K. Mallath
Fontenay-aux-Roses
Barcelona Tata Medical Center
dominique.laurier@irsn.fr Kolkata
manolis.kogevinas@isglobal.org
mohandas.mallath@tmckolkata.
C. René Leemans com
Anita Koushik Amsterdam University Medical
University of Montreal School of Center Alberto Mantovani
Public Health Amsterdam Humanitas Clinical and Research
Montreal, Quebec cr.leemans@vumc.nl Center
anita.koushik@umontreal.ca Milan
Michael Leitzmann and
James R. Krycer University of Regensburg Queen Mary University of London
The University of Sydney Regensburg London
Sydney, New South Wales Michael.Leitzmann@klinik.uni- alberto.mantovani@
james.krycer@sydney.edu.au regensburg.de humanitasresearch.it

Richard M. Martin J. David Miller Alexander Parker

University of Bristol Carleton University University of Florida
Bristol Ottawa, Ontario Jacksonville, Florida
richard.martin@bristol.ac.uk david.miller@carleton.ca Alexander.Parker@jax.ufl.edu
John D. Mathews Steven C. Moore Electra D. Paskett

The University of Melbourne National Institutes of Health The Ohio State University
Melbourne, Victoria Bethesda, Maryland Columbus, Ohio
Electra.Paskett@osumc.edu
mathewsj@unimelb.edu.au steve.moore@nih.gov
Julietta Patnick
Valerie McCormack Colin R. Muirhead
University of Oxford
International Agency for Research Newcastle upon Tyne Oxford
on Cancer colin.muirhead6@virginmedia.com julietta.patnick@ndph.ox.ac.uk
Lyon
mccormackv@iarc.fr Raúl Murillo Graham Pawelec
University Hospital San Ignacio University of Tübingen
Marjorie L. McCullough Bogotá Tübingen
American Cancer Society and graham.pawelec@uni-tuebingen.
Atlanta, Georgia Pontificia Universidad Javeriana de
marji.mccullough@cancer.org Bogotá
and Neil Pearce
James McKay International Agency for Research London School of Hygiene &
International Agency for Research on Cancer Tropical Medicine
on Cancer Lyon London
Lyon raulhmurillo@yahoo.com Neil.Pearce@LSHTM.ac.uk
mckayj@iarc.fr
David H. Phillips
Robert Newton
King’s College London
Francis Mégraud University of York
London
Inserm U1053 – BaRITOn York
david.phillips@kcl.ac.uk
Bordeaux Research in robert.newton@york.ac.uk
Translational Oncology – Sydney E. Philpott-Streiff
University of Bordeaux Chikako Nishigori Washington University in St. Louis
Bordeaux Kobe University School of Medicine
francis.megraud@chu-bordeaux.fr Kobe St. Louis, Missouri
chikako@med.kobe-u.ac.jp sphilpott@wustl.edu
Ronald L. Melnick
National Institutes of Health Joëlle L. Nortier Martyn Plummer
Bethesda, Maryland Université libre de Bruxelles University of Warwick
ron.melnick@gmail.com Brussels Coventry
Joelle.Nortier@erasme.ulb.ac.be Martyn.Plummer@warwick.ac.uk
Filip Meheus
International Agency for Research Josiah Ochieng Igor Pogribny
on Cancer Meharry Medical College Food and Drug Administration
Lyon Nashville, Tennessee Jefferson, Arkansas
jochieng@mmc.edu Igor.Pogribny@fda.hhs.gov
meheusf@iarc.fr
Kornelia Polyak
Wenbo Meng Hiroko Ohgaki
Dana-Farber Cancer Institute
The First Hospital of Lanzhou Charité Medical University
and
University Berlin Harvard Medical School
Lanzhou hiroko.ohgaki@charite.de Boston, Massachusetts
mengwb@lzu.edu.cn Kornelia_Polyak@dfci.harvard.edu
Klaus Pantel
Dominique S. Michaud University Medical Center Nagarajan Rajendra Prasad
Tufts University School of Medicine Hamburg-Eppendorf Annamalai University
Boston, Massachusetts Hamburg Annamalai Nagar
Dominique.Michaud@tufts.edu pantel@uke.de drprasadnr@gmail.com
578
Liang Qiao Natalie Reimers Lesley Rushton

The Westmead Institute for University Medical Center Imperial College London
Medical Research Hamburg-Eppendorf London
and Hamburg l.rushton@imperial.ac.uk
The University of Sydney and n.reimers@uke.de
Westmead Hospital Aoife Ryan
Westmead, New South Wales Leandro Fórnias Machado University College Cork
liang.qiao@sydney.edu.au de Rezende Cork
Universidade de São Paulo a.ryan@ucc.ie
You-Lin Qiao São Paulo
Cancer Hospital, Chinese lerezende@usp.br Rengaswamy Sankaranarayanan
RTI International India
Academy of Medical Sciences
Sabina Rinaldi New Delhi
Beijing
International Agency for Research sankardr@hotmail.com
and
Peking Union Medical College on Cancer
Lyon Diana Sarfati
Beijing
rinaldis@iarc.fr University of Otago
and Wellington
Affiliated Cancer Hospital of diana.sarfati@otago.ac.nz
Zhengzhou University Bridget H. Robson
Zhengzhou University of Otago
Wellington Catherine Sauvaget
qiaoy@cicams.ac.cn International Agency for Research
bridget.robson@otago.ac.nz
on Cancer
Ewa Rajpert-De Meyts Lyon
Eve Roman
Copenhagen University Hospital, sauvagetc@iarc.fr
University of York
Rigshospitalet
York
Copenhagen Augustin Scalbert
eve.roman@york.ac.uk
erm@rh.dk International Agency for Research
on Cancer
Martin Röösli
Kunnambath Ramadas Lyon
Swiss Tropical and Public Health
Regional Cancer Centre scalberta@iarc.fr
Institute
Thiruvananthapuram
and
ramdasrcc@gmail.com Ghislaine Scelo
University of Basel
Timothy R. Rebbeck International Agency for Research
Basel on Cancer
Dana-Farber Cancer Institute martin.roosli@swisstph.ch
and Lyon
Harvard T.H. Chan School of ghislaine.scelo@gmail.com
Thierry Roumeguère
Public Health Université libre de Bruxelles
Boston, Massachusetts David Schottenfeld
Brussels University of Michigan
Timothy_Rebbeck@dfci.harvard.edu Thierry.Roumeguere@erasme.ulb. Ann Arbor, Michigan
ac.be daschott@umich.edu
K. Srinath Reddy
Public Health Foundation of India Esther Roura Fornells Mary K. Schubauer-Berigan
New Delhi Institut Català d’Oncologia (ICO) – International Agency for Research
ksrinath.reddy@phfi.org Institut d’Investigació Biomèdica de on Cancer
Bellvitge (IDIBELL) Lyon
Jürgen Rehm Barcelona beriganm@iarc.fr
Centre for Addiction and Mental and
Health CIBER en Epidemiología y Salud Wolfgang A. Schulz
Toronto, Ontario Pública (CIBERESP) Heinrich Heine University
and Madrid Düsseldorf
Dalla Lana School of Public Health, eroura@iconcologia.net wolfgang.schulz@hhu.de
University of Toronto
Toronto, Ontario Anja Rudolph Joachim Schüz
and IQVIA Commercial GmbH & Co. International Agency for Research
Technische Universität Dresden OHG on Cancer
Dresden Frankfurt Lyon
jtrehm@gmail.com Anja.Rudolph@web.de schuzj@iarc.fr

Nereo Segnan Colinda Simons Herbert Tilg

Center for Epidemiology and Maastricht University Medical University of Innsbruck
Prevention in Oncology (CPO Maastricht Innsbruck
Piedmont) colinda.simons@maastricht Herbert.Tilg@i-med.ac.at
and university.nl
University Hospital Città della Massimo Tommasino
Salute e della Scienza Niels E. Skakkebæk International Agency for Research
Turin Copenhagen University Hospital, on Cancer
nereo.segnan@cpo.it Rigshospitalet Lyon
Copenhagen tommasinom@iarc.fr
Carlo Senore nes@rh.dk
Center for Epidemiology and Steinar Tretli
Prevention in Oncology (CPO Alexandra G. Smith
Cancer Registry of Norway
Piedmont) University of York
Oslo
and York
Steinar.Tretli@kreftregisteret.no
University Hospital Città della alexandra.smith@york.ac.uk
Salute e della Scienza Ioannis P. Trougakos
Turin Martyn T. Smith
University of California, Berkeley National and Kapodistrian
carlo.senore@cpo.it University of Athens
Berkeley, California
martynts@berkeley.edu Athens
Gautam Sethi itrougakos@biol.uoa.gr
National University of Singapore
Robert A. Smith
Singapore Michelle C. Turner
American Cancer Society
phcgs@nus.edu.sg Barcelona Institute for Global
Atlanta, Georgia
robert.smith@cancer.org Health (ISGlobal)
Muthu K. Shanmugam
Barcelona
National University of Singapore
Isabelle Soerjomataram michelle.turner@isglobal.org
Singapore
International Agency for Research
phcsmk@nus.edu.sg
on Cancer Renée Turzanski Fortner
Lyon German Cancer Research Center
Tatsuhiro Shibata
soerjomatarami@iarc.fr (DKFZ)
The University of Tokyo
Heidelberg
and
Aswathy Sreedevi r.fortner@dkfz.de
National Cancer Center Research Amrita Institute of Medical
Institute Sciences Giske Ursin
Tokyo Kochi
tashibat@ncc.go.jp Cancer Registry of Norway
aswathys@aims.amrita.edu Oslo
Kevin D. Shield Giske.Ursin@kreftregisteret.no
Bernard W. Stewart
Centre for Addiction and Mental University of New South Wales
Health Toshikazu Ushijima
and
Toronto, Ontario National Cancer Center Research
South Eastern Sydney Local
and Institute
Health District
Dalla Lana School of Public Health, Tokyo
Sydney, New South Wales
University of Toronto tushijim@ncc.go.jp
Bernard.Stewart@health.nsw.gov.
Toronto, Ontario au
kevin.david.shield@gmail.com Salvatore Vaccarella
Kurt Straif International Agency for Research
Jack Siemiatycki International Agency for Research on Cancer
University of Montreal on Cancer Lyon
Montreal, Quebec Lyon vaccarellas@iarc.fr
j.siemiatycki@umontreal.ca straif.kurt@gmail.com
Piet van den Brandt
Diane M. Simeone Michael J. Thun Maastricht University
NYU Langone Health American Cancer Society Maastricht
New York Atlanta, Georgia pa.vandenbrandt@
Diane.Simeone@nyulangone.org michael.thun62@gmail.com maastrichtuniversity.nl
580
Mieke Van Hemelrijck Elisabete Weiderpass Diana R. Withrow

King’s College London International Agency for Research National Cancer Institute
London on Cancer Rockville, Maryland
mieke.vanhemelrijck@kcl.ac.uk Lyon diana.withrow@nih.gov
director@iarc.fr
Katherine Van Loon Zhixun Yang
University of California, San Jeffrey N. Weitzel National Cancer Center, Chinese
Francisco City of Hope Cancer Center Academy of Medical Sciences
San Francisco, California Duarte, California and
Katherine.VanLoon@ucsf.edu jweitzel@coh.org
Peking Union Medical College
Beijing
Elizabeth A. Whelan
Christine Varon mailyangzx@126.com
National Institute for Occupational
Inserm U1053 – BaRITOn
Safety and Health
Bordeaux Research in Cincinnati, Ohio Jiri Zavadil
Translational Oncology – ewhelan@cdc.gov International Agency for Research
University of Bordeaux on Cancer
Bordeaux David Whiteman Lyon
christine.varon@u-bordeaux.fr QIMR Berghofer Medical Research zavadilj@iarc.fr
Institute
Paolo Vineis Brisbane, Queensland Georg Zeller
Imperial College London David.Whiteman@qimrberghofer. European Molecular Biology
London edu.au Laboratory
and Heidelberg
Italian Institute for Genomic Christopher P. Wild zeller@embl.de
Medicine International Agency for Research
Turin on Cancer Ariana Znaor
p.vineis@imperial.ac.uk Lyon International Agency for Research
wildshouse@gmail.com on Cancer
Elizabeth Ward Lyon
American Cancer Society Walter C. Willett znaora@iarc.fr
Atlanta, Georgia Harvard T.H. Chan School of
Public Health
and
and
North American Association of
Brigham and Women’s Hospital
Central Cancer Registries
and
eward04@gmail.com Harvard Medical School
Boston, Massachusetts
Penelope M. Webb wwillett@hsph.harvard.edu
QIMR Berghofer Medical Research
Institute Martin J. Wiseman
and World Cancer Research Fund
University of Queensland International
Brisbane, Queensland London
Penny.Webb@qimrberghofer.edu.au m.wiseman@wcrf.org

Disclosures of interests
Laia Alemany Vilches reports that her unit at the Harry J. de Koning reports that his unit, the Department
Catalan Institute of Oncology benefited from research of Public Health at Erasmus University Medical Center,
funding from GSK, Merck, and Seegene. benefits from research funding from SCOR Global Life.
Bruce K. Armstrong reports having received personal Joanna Didkowska reports that her unit at the Maria
consultancy fees from Maurice Blackburn Lawyers. Skłodowska-Curie Institute benefits from research
funding from Roche Polska, AstraZeneca, and Logistic
Patricia Ashton-Prolla reports having received per- Speed.
sonal consultancy fees from AstraZeneca.
Ronny Drapkin reports having received personal
Anssi Auvinen reports having received personal
consultancy fees from Repare Therapeutics and from
consultancy fees from Epid Research Inc. and Merck
Mersana Therapeutics.
Sharp & Dohme.
Jessica N. Everett reports that her unit at NYU
Linda Bauld reports providing expert opinion on to-
bacco control to the parliaments of the European Union Langone Health benefits from research funding from
and Iceland, members of the German Bundestag, and Immunovia.
several governments.
Judy E. Garber reports receiving personal consul-
Julia Brotherton reports that her unit at the Victorian tancy fees from Helix Pharma and benefiting from re-
Cytology Service benefited from an unrestricted research funding from Ambry Genetics.
search grant from Seqirus.
Gemma Gatta reports that her unit at Istituto Nazionale
Karen Brown reports that her unit at the University of dei Tumori benefited from research funding from
Leicester benefited from research funding from Indena Amgen Dompé.
SpA.
Adi F. Gazdar (deceased) reported receiving personal
Laia Bruni reports that her unit at the Catalan Institute consultancy fees from Genentech and Bristol-Myers
of Oncology benefits from research funding from Merck Squibb.
and GlaxoSmithKline.
Charles W. Jameson reports providing expert tes-
Franco Cavalli reports that his unit at the Oncology timony for plaintiffs in litigation related to glyphosate
Institute of Southern Switzerland benefited from re- products and talc products.
search funding from Roche.
Dominique Laurier reports that his unit at the Institute
Aaron J. Cohen reports that his unit at the Health
for Radiological Protection and Nuclear Safety benefits
Effects Institute benefits from funding from companies
from research funding from Areva and EDF.
that manufacture or market motor vehicles for sale in
the USA.
Alberto Mantovani reports receiving honoraria from
Jack Cuzick reports that his unit at the Wolfson Biovelocità, Novartis, Merck, Compugen, Roche,
Institute of Preventive Medicine benefits from research AstraZeneca, and Chiesi.
funding from AstraZeneca, and reports having received
personal consultancy fees in his capacity as a member Francis Mégraud reports that his unit at the University
of an advisory board at Merck. of Bordeaux benefits from research funding from
Allergan, reports having benefited from research fund-
Diona L. Damian reports receiving non-monetary sup- ing from Biocodex, and reports having received hono-
port from Blackmores Ltd. raria from Mayoly Spindler.
582
Klaus Pantel reports that his unit at the Institute of Martyn T. Smith reports having received personal
Tumour Biology at University Medical Center Hamburg- consultancy fees from several law firms in connection
Eppendorf benefits from research funding from Janssen with providing expert testimony for plaintiffs in litigation
and from EU/IMI CANCER-ID EFPIA partners, and re- related to pharmaceutical, occupational, and environ-
ports holding EPO patents No. 2016128125 A1 and ap- mental exposure cases.
plication No. 17157020.3 – 1405.
Ioannis P. Trougakos reports that his unit at the
Electra D. Paskett reports that her unit at The Ohio National and Kapodistrian University of Athens ben-
State University benefits from research funding from efited from research funding from Amgen.
the Merck Foundation.
Michelle C. Turner reports having received personal
David H. Phillips reports receiving personal consul- consultancy fees from ICF Incorporated LLC.
tancy fees from the law firm Shook, Hardy & Bacon.
Giske Ursin reports that her institution, the Cancer
You-Lin Qiao reports having benefited from person- Registry of Norway, benefits from research funding
al consultancy fees and support for travel, and re- from Merck/Merck Sharp & Dohme.
ports that his unit at the Chinese Academy of Medical
Sciences benefited from research support from Merck Toshikazu Ushijima reports that his unit at the
Sharp & Dohme. National Cancer Center Research Institute benefits
from research funding from Ohara Pharmaceutical Inc.
Martin Röösli reports having been an unpaid mem-
ber of the foundation board of the Swiss Research Elizabeth Ward reports that her unit at the American
Foundation for Electricity and Mobile Communication, Cancer Society Intramural Research Program benefit-
a non-profit research foundation at ETH Zurich. ed from research funding from Merck-Serrano.
Wolfgang A. Schulz reports that his unit at Heinrich Penelope M. Webb reports that her unit at QIMR
Heine University benefited from research support from Berghofer Medical Research Institute receives funding
4SC. for a research project from AstraZeneca.
Nereo Segnan reports that his unit at the Center Georg Zeller reports holding shares on the patent
for Epidemiology and Prevention in Oncology (CPO EP2955232A1 on “Method for diagnosing colorectal
Piedmont) benefited from equipment support from cancer based on analyzing the gut microbiome”.
Medtronic and EndoChoice.
Carlo Senore reports that the University Hospital Città

della Salute e della Scienza, where his unit is based,
benefited from equipment support from Medtronic and
EndoChoice.

Sources
Boxes
2.1.1 Reproduced from Thun M, Linet MS, Cerhan JR, Haiman 6.5.2 Patricia Ashton-Prolla.
CA, Schottenfeld D, editors (2018). Cancer epidemiology and
6.6.1 Raúl Murillo.
prevention. 4th ed. New York (NY), USA: Oxford University Press,
Table 11.1, p. 186, by permission of Oxford University Press, USA P1.1 Reprinted from Ghebreyesus TA (2019). Progress in beat-
(www.oup.com). ing the tobacco epidemic. Lancet. 394(10198):548–9. https://doi.
3.2.1 Thomas Hudson. org/10.1016/S0140-6736(19)31730-1 PMID:31371094, Copyright
2019, with permission from Elsevier.
4.4.1 Rengaswamy Sankaranarayanan and Kunnambath Ramadas.
P1.2 Reprinted from WHO (2019). WHO Study Group on
5.10.1 Adapted/translated from Lax SF, Horn LC, Löning T (2016). Tobacco Product Regulation. Report on the scientific basis
Categorization of uterine cervix tumors: what’s new in the 2014
of tobacco product regulation: seventh report of a WHO study
WHO classification [in German]. Pathologe. 37(6):573–84. https://
group. Geneva, Switzerland: World Health Organization (WHO
doi.org/10.1007/s00292-016-0247-8 PMID:27770187, by permis-
Technical Report Series, No. 1015). Licence: CC BY-NC-SA 3.0
sion from Springer Nature © 2016.
IGO. Available from: https://www.who.int/publications-detail/
5.14.1 Adapted from Ulbright TM, Amin MB, Balzer B, Berney DM, who-study-group-on-tobacco-product-regulation-report-on-the-
Epstein JI, Guo C, et al. (2016). Germ cell tumours. In: Moch H, scientific-basis-of-tobacco-product-regulation-seventh-report-
Humphrey PA, Ulbright TM, Reuter VE, editors. WHO classifica- of-a-who-study-group.
tion of tumours of the urinary system and male genital organs.
Lyon, France: International Agency for Research on Cancer
(WHO Classification of Tumours series, 4th edition); pp. 189–226.
6.5.1 Adapted from Weitzel JN, Blazer KR, MacDonald DJ, Culver
JO, Offit K (2011). Genetics, genomics, and cancer risk assess-
ment: state of the art and future directions in the era of person-
alized medicine. CA Cancer J Clin. 61(5):327–59. https://doi.
org/10.3322/caac.20128 PMID:21858794, with permission from
John Wiley & Sons.
Figures
The photographs in this material are used for illustrative purposes 1.1.4 Courtesy of Bochen Cao, Isabelle Soerjomataram,
only; they do not imply any particular health status, attitudes, be- and Freddie Bray. Compiled from Cao B, Bray F, Ilbawi A,
haviours, or actions on the part of any person who appears in the Soerjomataram I (2018). Effect on longevity of one-third reduc-
photographs. tion in premature mortality from non-communicable diseases by
2030: a global analysis of the Sustainable Development Goal
1.1.1 Reproduced from Cao B, Bray F, Beltrán-Sánchez H, health target. Lancet Glob Health. 6(12):e1288–e1296. https://
Ginsburg O, Soneji S, Soerjomataram I (2017). Benchmarking doi.org/10.1016/S2214-109X(18)30411-X PMID:30420032
life expectancy and cancer mortality: global comparison with
cardiovascular disease 1981–2010. BMJ. 357:j2765. https://doi. 1.1.5 vicmillon. Courtesy of Pixabay.
org/10.1136/bmj.j2765 PMID:28637656 1.1.6 iStockphoto.com/PavelSinitcyn.
1.1.2 Courtesy of Bochen Cao, Isabelle Soerjomataram, and
1.2.1A Compiled from Ferlay J, Colombet M, Bray F (2018).
Freddie Bray. Compiled from WHO Global Health Estimates (https://
www.who.int/healthinfo/global_burden_disease/en/). Cancer Incidence in Five Continents, CI5plus: IARC CancerBase
No. 9 [Internet]. Lyon, France: International Agency for Research
1.1.3 Reproduced from Ferlay J, Ervik M, Lam F, Colombet M, on Cancer. Available from: http://ci5.iarc.fr.
Mery L, Piñeros M, et al. (2018). Global Cancer Observatory:
Cancer Today. Lyon, France: International Agency for Research 1.2.1B WHO Mortality Database. Available from: http://www.who.
on Cancer. Available from: https://gco.iarc.fr/today. int/healthinfo/statistics/mortality_rawdata/en/index.html.
584
1.2.2A Compiled from Ferlay J, Colombet M, Bray F (2018). 1.3.8 Adapted from Soerjomataram I, Lortet-Tieulent J, Parkin
Cancer Incidence in Five Continents, CI5plus: IARC CancerBase DM, Ferlay J, Mathers C, Forman D, et al. (2012). Global bur-
No. 9 [Internet]. Lyon, France: International Agency for Research den of cancer in 2008: a systematic analysis of disability-adjust-
on Cancer. Available from: http://ci5.iarc.fr. ed life-years in 12 world regions. Lancet. 380(9856):1840–50.
https://doi.org/10.1016/S0140-6736(12)60919-2 PMID:23079588,
1.2.2B WHO Mortality Database. Available from: http://www.who. Copyright 2012, with permission from Elsevier.
int/healthinfo/statistics/mortality_rawdata/en/index.html.
1.3.9 Compiled from Ferlay J, Colombet M, Bray F (2018). Cancer
1.2.3A Compiled from Ferlay J, Colombet M, Bray F (2018). Incidence in Five Continents, CI5plus: IARC CancerBase No. 9
Cancer Incidence in Five Continents, CI5plus: IARC CancerBase [Internet]. Lyon, France: International Agency for Research on
No. 9 [Internet]. Lyon, France: International Agency for Research Cancer. Available from: http://ci5.iarc.fr.
on Cancer. Available from: http://ci5.iarc.fr.
2.1.1 © Vincent Poulinet. Courtesy of Flickr.
1.2.3B WHO Mortality Database. Available from: http://www.who.
int/healthinfo/statistics/mortality_rawdata/en/index.html. 2.1.2 Courtesy of the U.S. National Cancer Institute; courtesy of
Clifford Watson.
1.2.4 SoleneC1. Courtesy of Pixabay.
2.1.3 Reproduced from Huang J, Duan Z, Kwok J, Binns S, Vera LE,
1.2.5A Compiled from Ferlay J, Colombet M, Bray F (2018). Kim Y, et al. (2019). Vaping versus JUULing: how the extraordinary
Cancer Incidence in Five Continents, CI5plus: IARC CancerBase growth and marketing of JUUL transformed the US retail e-cig-
No. 9 [Internet]. Lyon, France: International Agency for Research arette market. Tob Control. 28(2):146–51. https://doi.org/10.1136/
on Cancer. Available from: http://ci5.iarc.fr. tobaccocontrol-2018-054382 PMID:29853561 © Huang J, Duan Z,
1.2.5B WHO Mortality Database. Available from: http://www.who. Kwok J, et al. 2018. All rights reserved.
int/healthinfo/statistics/mortality_rawdata/en/index.html 2.1.4 Reproduced from GBD 2015 Tobacco Collaborators
1.2.6A Compiled from Ferlay J, Colombet M, Bray F (2018). (2017). Smoking prevalence and attributable disease bur-
Cancer Incidence in Five Continents, CI5plus: IARC CancerBase den in 195 countries and territories, 1990-2015: a system-
No. 9 [Internet]. Lyon, France: International Agency for Research atic analysis from the Global Burden of Disease Study 2015.
on Cancer. Available from: http://ci5.iarc.fr. Lancet. 389(10082):1885–906. https://doi.org/10.1016/S0140-
6736(17)30819-X PMID:28390697 © 2017, GBD 2015 Tobacco
1.2.6B WHO Mortality Database. Available from: http://www.who. Collaborators. Published by Elsevier Ltd.
2.1.5 Reproduced from Wang TW, Gentzke A, Sharapova S,
1.2.7A Compiled from Ferlay J, Colombet M, Bray F (2018). Cullen KA, Ambrose BK, Jamal A (2018). Tobacco product use
Cancer Incidence in Five Continents, CI5plus: IARC CancerBase among middle and high school students - United States, 2011-
No. 9 [Internet]. Lyon, France: International Agency for Research 2017. MMWR Morb Mortal Wkly Rep. 67(22):629–33. https://doi.
on Cancer. Available from: http://ci5.iarc.fr. org/10.15585/mmwr.mm6722a3 PMID:29879097
1.2.7B WHO Mortality Database. Available from: http://www.who. 2.1.6 hsyncoban/Getty Images.
2.2.1 © 2012 Grace Wilentz, Courtesy of Photoshare.
1.2.8A Compiled from Ferlay J, Colombet M, Bray F (2018).
2.2.2 Courtesy of Churches Health Association of Zambia (CHAZ).
Cancer Incidence in Five Continents, CI5plus: IARC CancerBase
No. 9 [Internet]. Lyon, France: International Agency for Research 2.2.3 © Ted Alcorn. www.tedalcorn.com.
on Cancer. Available from: http://ci5.iarc.fr.
2.3.1 © 2015 Daniel Waistell, Courtesy of Photoshare.
1.2.8B WHO Mortality Database. Available from: http://www.who. 2.3.2 iStockphoto.com/west.
2.3.3–2.3.7 Jürgen Rehm, Kevin D. Shield, and Elisabete
1.2.9 cegoh. Courtesy of Pixabay. Weiderpass. Compiled from data in WHO (2018). Global sta-
1.2.10A Compiled from Ferlay J, Colombet M, Bray F (2018). tus report on alcohol and health 2018. Geneva, Switzerland:
Cancer Incidence in Five Continents, CI5plus: IARC CancerBase World Health Organization. Available from: https://www.who.int/
No. 9 [Internet]. Lyon, France: International Agency for Research substance_abuse/publications/global_alcohol_report/en/.
on Cancer. Available from: http://ci5.iarc.fr. 2.4.1 Chikako Nishigori.
1.2.10B WHO Mortality Database. Available from: http://www. 2.4.2 Adapted from Hiraku Y, Kawanishi S, Ohshima H, editors
who.int/healthinfo/statistics/mortality_rawdata/en/index.html. (2014). Cancer and inflammation mechanisms: chemical, biologi-
1.3.1 Miranda M. Fidler-Benaoudia and Freddie Bray. Compiled cal, and clinical aspects. 1st ed. Hoboken (NJ), USA: John Wiley
from UNDP Human Development Report Office (2015). What & Sons. © 2014 John Wiley & Sons, Inc. Published 2014 by John
is human development? New York (NY), USA: United Nations Wiley & Sons, Inc.
Development Programme; pp. 1–10. Available from: http://hdr. 2.4.3 Chikako Nishigori. Compiled from the Japanese Ministry of
undp.org/en/content/what-human-development. Health, Labour and Welfare.
1.3.2–1.3.6 Reproduced from Ferlay J, Ervik M, Lam F, Colombet 2.4.4 © Wayan Susila.
M, Mery L, Piñeros M, et al. (2018). Global Cancer Observatory:
Cancer Today. Lyon, France: International Agency for Research 2.4.5 taniadimas. Courtesy of Pixabay.
on Cancer. Available from: https://gco.iarc.fr/today. 2.5.1 Dominique Laurier. Compiled from UNSCEAR (2010).
1.3.7 Reprinted from Fidler M, Bray F, Soerjomataram I Sources and effects of ionizing radiation. UNSCEAR 2008 Report
(2018). The global cancer burden and human development: to the General Assembly, with scientific annexes. Volume I: Sources.
a review. Scand J Public Health. 46(1):27–36. https://doi. New York (NY), USA: United Nations Scientific Committee on the
org/10.1177/1403494817715400 PMID:28669281 Copyright 2018, Effects of Atomic Radiation. Available from: http://www.unscear.
SAGE Publishing. Compiled from Soerjomataram I, Lortet-Tieulent org/unscear/publications/2008_1.html.
J, Parkin DM, Ferlay J, Mathers C, Forman D, et al. (2012). Global 2.5.2 © Phillip Jeffrey fadetoplay.com. License CC BY-NC-SA 2.0.
burden of cancer in 2008: a systematic analysis of disability-adjusted
2.5.3 Gill Tudor/IAEA. Available on Flickr. License CC BY-SA 2.0.
life-years in 12 world regions. Lancet. 380(9856):1840–50. https://
doi.org/10.1016/S0140-6736(12)60919-2 PMID:23079588; Arnold 2.5.4 Martin Röosli. Compiled from Foerster M, Thielens A, Joseph
M, Pandeya N, Byrnes G, Renehan PAG, Stevens GA, Ezzati PM, W, Eeftens M, Röösli M (2018). A prospective cohort study of adoles-
et al. (2015). Global burden of cancer attributable to high body-mass cents’ memory performance and individual brain dose of microwave
index in 2012: a population-based study. Lancet Oncol. 16(1):36–46. radiation from wireless communication. Environ Health Perspect.
https://doi.org/10.1016/S1470-2045(14)71123-4 PMID:25467404 126(7):077007. https://doi.org/10.1289/EHP2427 PMID:30044230

2.5.5 © 2018 Avishek Das, Courtesy of Photoshare. 3.1.4 GregMontani. Courtesy of Pixabay.
2.6.1 © UNICEF India/2014/Dhiraj Singh. 3.2.1 Reprinted from Manolio TA, Collins FS, Cox NJ, Goldstein
DB, Hindorff LA, Hunter DJ, et al. (2009). Finding the missing
2.6.2 (Left) Elina Sazanova/Pexels.com; (right) iStockphoto.com/ heritability of complex diseases. Nature. 461(7265):747–53.
frederique wacquier. https://doi.org/10.1038/nature08494 PMID:19812666, by per-
B2.6.1 Adapted from Farvid MS, Cho E, Chen WY, Eliassen AH, mission from Springer Nature © 2009; adapted from McCarthy
Willett WC (2015). Adolescent meat intake and breast cancer risk. MI, Abecasis GR, Cardon LR, Goldstein DB, Little J, Ioannidis
Int J Cancer. 136(8):1909–20. https://doi.org/10.1002/ijc.29218 JP, et al. (2008). Genome-wide association studies for complex
PMID:25220168, with permission from John Wiley & Sons. traits: consensus, uncertainty and challenges. Nat Rev Genet.
9(5):356–69. https://doi.org/10.1038/nrg2344 PMID:18398418,
B2.6.2 Walter C. Willett and Hilary Farmer. Compiled from by permission from Springer Nature © 2008.
Richman EL, Kenfield SA, Chavarro JE, Stampfer MJ,
Giovannucci EL, Willett WC, et al. (2013). Fat intake after diag- 3.2.2 Reproduced from Garcia-Closas M, Chanock S (2008). Genetic
nosis and risk of lethal prostate cancer and all-cause mortality. susceptibility loci for breast cancer by estrogen receptor status. Clin
JAMA Intern Med. 173(14):1318–26. https://doi.org/10.1001/jama Cancer Res. 14(24):8000–9. https://doi.org/10.1158/1078-0432.
CCR-08-0975 PMID:19088016, by permission from the American
internmed.2013.6536 PMID:23752662
Association for Cancer Research.
B2.6.3 Reproduced from Wang DD, Li Y, Afshin A, Springmann
3.2.3 Reprinted from Lawrence MS, Stojanov P, Polak P, Kryukov
M, Mozaffarian D, Stampfer MJ, et al. (2019). Global improvement
GV, Cibulskis K, Sivachenko A, et al. (2013). Mutational heteroge-
in dietary quality could lead to substantial reduction in premature
neity in cancer and the search for new cancer-associated genes.
death. J Nutr. 149(6):1065–74. https://doi.org/10.1093/jn/nxz010 Nature. 499(7457):214–8. https://doi.org/10.1038/nature12213
PMID:31049577, by permission of Oxford University Press. PMID:23770567, by permission from Springer Nature © 2013.
2.7.1 Christine M. Friedenreich and Michael Leitzmann. 3.2.4 Reprinted from Stratton MR, Campbell PJ, Futreal PA
2.7.2 iStockphoto.com/simonkr. (2009). The cancer genome. Nature. 458(7239):719–24. https://
doi.org/10.1038/nature07943 PMID:19360079, by permission
2.7.3 baona/Getty Images. from Springer Nature © 2009.
2.7.4 © 2011 SHER & GUL, Courtesy of Photoshare. 3.2.5 Stephen J. Chanock.
2.8.1 © International Institute of Tropical Agriculture. Available on 3.3.1 Courtesy of Martyn T. Smith.
Flickr. License CC BY-NC-SA 2.0.
3.3.2 klimkin. Courtesy of Pixabay.
2.8.2 Thomas Lumpkin/CIMMYT. Available on Flickr. License CC
BY-NC-SA 2.0. 3.3.3 © 2012 Micah Albert, Courtesy of Photoshare.
2.8.3 © Kew Royal Botanical Gardens. Available on Flickr. 3.3.4 Reproduced from Simonds NI, Ghazarian AA, Pimentel CB,
Schully SD, Ellison GL, Gillanders EM, et al. (2016). Review of
License CC BY-NC-SA 2.0.
the gene-environment interaction literature in cancer: what do we
2.8.4 congerdesign. Courtesy of Pixabay. know? Genet Epidemiol. 40(5):356–65. https://doi.org/10.1002/
gepi.21967 PMID:27061572, with permission from John Wiley &
2.9.1–2.9.3 Reproduced from Health Effects Institute (2018).
Sons.
State of global air 2018. Data source: Global Burden of Disease
Study 2016. Institute for Health Metrics and Evaluation, 2017. 3.3.5 U.S. Air Force photo/Staff Sgt Eric T. Sheler. Released un-
Available from: www.stateofglobalair.org. der US public domain.
2.9.4 Courtesy of Marco De Santis. 3.4.1 Adapted from Tubbs A, Nussenzweig A (2017). Endogenous
DNA damage as a source of genomic instability in cancer.
2.10.1 Photo by Tim Wilson (timwilson@mackandtim.net). Cell. 168(4):644–56. https://doi.org/10.1016/j.cell.2017.01.002
2.10.2 © 2016 Sudipta Dutta Chowdhury, Courtesy of Photoshare. PMID:28187286, Copyright 2017, with permission from Elsevier.
2.10.3 Chevanon Photography. Courtesy of Pexels. 3.4.2 Eugenia Dogliotti and Margherita Bignami.
2.10.4 skeeze. Courtesy of Pixabay. 3.4.3 © MICHEL GANGNE/AFP.
2.10.5 Sebastian Cem Kreuzer-Erenay. www.sebastiancem.com. B3.4.1 Eugenia Dogliotti and Margherita Bignami.
2.11.1 & 2.11.2 Kaylan Veera. B3.4.2 Eugenia Dogliotti and Margherita Bignami. Compiled
from Ng AWT, Poon SL, Huang MN, Lim JQ, Boot A, Yu W, et al.
3.1.1 Reproduced from Hanahan D, Weinberg RA (2014). Hallmarks (2017). Aristolochic acids and their derivatives are widely impli-
of cancer: an organizing principle for cancer medicine. In: Devita cated in liver cancers in Taiwan and throughout Asia. Sci Transl
VT Jr., Lawrence TS, Rosenberg SA, editors. Cancer: principles Med. 9(412):eaan6446. https://doi.org/10.1126/scitranslmed.
and practice of oncology. Philadelphia (PA), USA: Wolters Kluwer aan6446 PMID:29046434
Health; adapted from Hanahan D, Weinberg RA (2000). The hall-
3.5.1–3.5.3 Muthu K. Shanmugam.
marks of cancer. Cell. 100(1):57–70. https://doi.org/10.1016/s0092-
8674(00)81683-9 PMID:10647931, Copyright 2000, with permis- 3.5.4 Digital Vision/Getty Images.
sion from Elsevier; Hanahan D, Weinberg RA (2011). Hallmarks 3.5.5 Ed Uthman. Available on Flickr. License CC BY 2.0.
of cancer: the next generation. Cell. 144(5):646–74. https://doi.
org/10.1016/j.cell.2011.02.013 PMID:21376230, Copyright 2011, 3.6.1 Reproduced from Collaborative Group on Hormonal Factors
with permission from Elsevier. in Breast Cancer (2012). Menarche, menopause, and breast cancer
risk: individual participant meta-analysis, including 118 964 women
3.1.2 Reproduced from Bolden JE, Lowe SW (2015). Cellular with breast cancer from 117 epidemiological studies. Lancet Oncol.
senescence. In: Mendelsohn J, Gray JW, Howley PM, Israel 13(11):1141–51. https://doi.org/10.1016/S1470-2045(12)70425-4
MA, Thompson CB, editors. The molecular basis of cancer. PMID:23084519 © 2012 Elsevier Ltd. Published by Elsevier Ltd.
Philadelphia (PA), USA: Elsevier; p. 235. Copyright 2015, with
permission from Elsevier. 3.6.2 Reproduced from Figueroa JD, Pfeiffer RM, Patel DA,
Linville L, Brinton LA, Gierach GL, et al. (2014). Terminal duct
3.1.3 Reproduced from Kierszenbaum AL, Tres LL (2016). lobular unit involution of the normal breast: implications for breast
Histology and cell biology: an introduction to pathology. cancer etiology. J Natl Cancer Inst. 106(10):dju286. https://doi.
Philadelphia (PA), USA: Elsevier; p. 102. Copyright 2016, with org/10.1093/jnci/dju286 PMID:25274491, by permission of Oxford
permission from Elsevier. University Press.
586
3.6.3 Reproduced from Sampson JN, Falk RT, Schairer C, Moore 3.10.3 Adapted from Gagnaire A, Nadel B, Raoult D, Neefjes J,
SC, Fuhrman BJ, Dallal CM, et al. (2017). Association of estro- Gorvel JP Gagnaire A, Nadel B, Raoult D, Neefjes J, Gorvel JP
gen metabolism with breast cancer risk in different cohorts of (2017). Collateral damage: insights into bacterial mechanisms that
postmenopausal women. Cancer Res. 77(4):918–25. https://doi. predispose host cells to cancer. Nat Rev Microbiol. 15(2):109–28.
org/10.1158/0008-5472.CAN-16-1717 PMID:28011624, Copyright https://doi.org/10.1038/nrmicro.2016.171 PMID:28045107, by per-
2017, American Association for Cancer Research. mission from Springer Nature © 2017.
3.6.4 Reproduced from Chlebowski RT, Anderson GL, Sarto 3.10.4 Photo by Eric Erbe, digital colorization by Christopher
GE, Haque R, Runowicz CD, Aragaki AK, et al. (2015). Pooley. United States Department of Agriculture. Released under
Continuous combined estrogen plus progestin and endometrial US public domain.
cancer: the Women’s Health Initiative randomized trial. J Natl
3.10.5 Georg Zeller.
Cancer Inst. 108(3):djv350. https://doi.org/10.1093/jnci/djv350
PMID:26668177, by permission of Oxford University Press. 3.11.1 Thomas Ried/NCI Center for Cancer Research. Released
under US public domain.
3.6.5 Reproduced from Trabert B, Wentzensen N, Yang HP,
Sherman ME, Hollenbeck AR, Park Y, et al. (2013). Is estrogen 3.11.2 Reproduced from Guyton KZ, Rusyn I, Chiu WA, Corpet
plus progestin menopausal hormone therapy safe with respect DE, van den Berg M, Ross MK, et al. (2018). Application of the
to endometrial cancer risk? Int J Cancer. 132(2):417–26. https:// key characteristics of carcinogens in cancer hazard identifica-
doi.org/10.1002/ijc.27623 PMID:22553145, with permission from tion. Carcinogenesis. 39(4):614–22. https://doi.org/10.1093/car-
John Wiley & Sons. cin/bgy031 PMID:29562322 © Guyton et al. 2018. Published by
Oxford University Press.
3.6.6 Reproduced from Beral V, Gaitskell K, Hermon C, Moser
K, Reeves G, Peto R; Collaborative Group on Epidemiological 4.1.1 Adapted from Ferlay J, Ervik M, Lam F, Colombet M, Mery
Studies of Ovarian Cancer (2015). Menopausal hormone use and L, Piñeros M, et al. (2018). Global Cancer Observatory: Cancer
ovarian cancer risk: individual participant meta-analysis of 52 Today. Lyon, France: International Agency for Research on
epidemiological studies. Lancet. 385(9980):1835–42. https://doi. Cancer. Available from: https://gco.iarc.fr/today.
org/10.1016/S0140-6736(14)61687-1 PMID:25684585, © 2015
4.1.2 Reproduced from Vaccarella S, Lortet-Tieulent J, Saracci
Collaborative Group on Epidemiological Studies of Ovarian
R, Conway DI, Straif K, Wild CP, editors (2019). Reducing so-
Cancer. Open Access article distributed under the terms of CC
cial inequalities in cancer: evidence and priorities for research
BY. Published by Elsevier Ltd.
(IARC Scientific Publication No. 168). Lyon, France: International
3.6.7 Reproduced from International Collaboration of Agency for Research on Cancer. Available from http://publications.
Epidemiological Studies of Cervical Cancer (2006). Cervical car- iarc.fr/580.
cinoma and reproductive factors: collaborative reanalysis of indi-
4.1.3 Johnny Miller/Unequal Scenes.
vidual data on 16,563 women with cervical carcinoma and 33,542
women without cervical carcinoma from 25 epidemiological stud- 4.1.4 U.S. Navy photo by Chief Warrant Officer 4 Seth Rossman.
ies. Int J Cancer. 119(5):1108–24. https://doi.org/10.1002/ijc.21953 Released under US public domain.
PMID:16570271, with permission from John Wiley & Sons.
4.2.1 & 4.2.2 Compiled from WHO (2015). World health statistics
3.6.8 Reproduced from Roddam AW, Allen NE, Appleby P, Key 2015. Geneva, Switzerland: World Health Organization. Available
TJ; Endogenous Hormones and Prostate Cancer Collaborative from: https://www.who.int/gho/publications/world_health_statistics/
Group (2008). Endogenous sex hormones and prostate can- 2015/en/.
cer: a collaborative analysis of 18 prospective studies. J Natl
4.2.3 Sofia Christensen/Voice of America.
Cancer Inst. 100(3):170–83. https://doi.org/10.1093/jnci/djm323
PMID:18230794, by permission of Oxford University Press. 4.2.4 UNICEF Zimbabwe/2018/Moetsabi.
3.7.1 Courtesy of Augustin Scalbert and Isabelle Romieu. 4.3.1 Rob Smith at www.photorobsmith.com.
3.7.2 Reprinted with permission from Carayol M, Leitzmann MF, 4.3.2 © 2009 Xiaoyun Zheng, Courtesy of Photoshare.
Ferrari P, Zamara-Ros R, Achaintre D, Stepien M, et al. (2017).
4.3.3 Wanqing Chen.
Blood metabolic signatures of body mass index: a targeted metab-
olomics study in the EPIC cohort. J Proteome Res. 16(9):3137–46. 4.3.4 Courtesy of Dr Volker Brinkmann/Max Planck Institute for
https://doi.org/10.1021/acs.jproteome.6b01062 PMID:28758405 Infection Biology.
© 2017, American Chemical Society.
4.4.1 & 4.4.2 Courtesy of Rengaswamy Sankaranarayanan and
3.8.1 Reproduced from Herceg Z, Lambert MP, van Veldhoven Kunnambath Ramadas. Compiled from annual reports of Mumbai,
K, Demetriou C, Vineis P, Smith MT, et al. (2013). Towards in- Chennai, Bangalore, Delhi, and Barshi population-based cancer
corporating epigenetic mechanisms into carcinogen identifica- registries and from the reports of the National Cancer Registry
tion and evaluation. Carcinogenesis. 34(9):1955–67. https://doi. Programme of India.
org/10.1093/carcin/bgt212 PMID:23749751, by permission of
4.4.3 iStockphoto.com/THEPALMER.
Oxford University Press.
4.4.4 Courtesy of Zachary Weber.
3.8.2 Reproduced from Herceg Z, Ghantous A, Wild CP, Sklias
A, Casati L, Duthie SJ, et al. (2018). Roadmap for investigating 4.4.5 © 2011 Biocon Foundation, Courtesy of Photoshare.
epigenome deregulation and environmental origins of cancer.
4.5.1 WHO/Sergey Volkov.
Int J Cancer. 142(5):874–82. https://doi.org/10.1002/ijc.31014
PMID:28836271, © 2017 International Agency for Research on 4.5.2 Photo by Denise Bradley. © Archant Norfolk.
Cancer (IARC/WHO); licensed by UICC. Open Access. 4.5.3 Lili Sohn.
3.8.3 & 3.8.4 Toshikazu Ushijima. 4.6.1 Reprinted from Morris AM, Rhoads KF, Stain SC, Birkmeyer
3.9.1 & 3.9.2 Alberto Mantovani. JD (2010). Understanding racial disparities in cancer treatment and
outcomes. J Am Coll Surg. 211(1):105–13. https://doi.org/10.1016/j.
3.10.1 bryan. Available on Flickr. License CC BY 2.0.
jamcollsurg.2010.02.051 PMID:20610256, Copyright 2010, with
3.10.2 Georg Zeller. permission from Elsevier.

4.6.2 Reprinted from Henley SJ, Jemal A (2018). Rural can- 5.1.6 Reprinted from Vargas AJ, Harris CC (2016). Biomarker de-
cer control: bridging the chasm in geographic health inequity. velopment in the precision medicine era: lung cancer as a case
Cancer Epidemiol Biomarkers Prev. 27(11):1248–51. https://doi. study. Nat Rev Cancer. 16(8):525–37. https://doi.org/10.1038/
org/10.1158/1055-9965.EPI-18-0807 PMID:30385497, by per- nrc.2016.56 PMID:27388699, by permission from Springer Nature
mission from the American Association for Cancer Research. © 2016; adapted from National Research Council (2011). Toward
precision medicine: building a knowledge network for biomedi-
4.6.3 Reproduced from Siegel RL, Jemal A, Wender RC, Gansler
cal research and a new taxonomy of disease. Washington (DC),
T, Ma J, Brawley OW (2018). An assessment of progress in cancer
control. CA Cancer J Clin. 68(5):329–39. https://doi.org/10.3322/ USA: National Academies Press. https://doi.org/10.17226/13284,
caac.21460 PMID:30191964, with permission from John Wiley & with permission of National Academies Press; permission con-
Sons. veyed through Copyright Clearance Center, Inc.
4.6.4 Reprinted from Siegel RL, Sahar L, Robbins A, Jemal A 5.2.1 © 2012 Terese Winslow LLC, U.S. Govt.
(2015). Where can colorectal cancer screening interventions have 5.2.2 & 5.2.3 Reproduced from Ferlay J, Ervik M, Lam F, Colombet
the most impact? Cancer Epidemiol Biomarkers Prev. 24(8):1151–6. M, Mery L, Piñeros M, et al. (2018). Global Cancer Observatory:
https://doi.org/10.1158/1055-9965.EPI-15-0082 PMID:26156973, by Cancer Today. Lyon, France: International Agency for Research
permission from the American Association for Cancer Research. on Cancer. Available from: https://gco.iarc.fr/today.
4.6.5 Reproduced from The New Hampshire Colorectal Cancer 5.2.4 Reproduced from de Martel C, Plummer M, Vignat J,
Screening Program (NHCRCSP) Patient Navigation (PN) Model. Franceschi S (2017). Worldwide burden of cancer attributable to
Centers for Disease Control and Prevention and New Hampshire HPV by site, country and HPV type. Int J Cancer. 141(4):664–
Colorectal Cancer Screening Program (2016). New Hampshire 70. https://doi.org/10.1002/ijc.30716 PMID:28369882, © 2017
Colorectal Cancer Screening Program Patient Navigation Model International Agency for Research on Cancer (IARC/WHO); li-
for increasing colonoscopy quality and completion: a replication censed by UICC. Open Access.
manual. p. 8. Available from: https://www.cdc.gov/cancer/crccp/
pn-replication-manual.htm. 5.2.5 Reprinted from Leemans CR, Snijders PJF, Brakenhoff RH
(2018). The molecular landscape of head and neck cancer. Nat
B4.6.1 Reprinted from McKenney KM, Martinez NG, Yee LM (2018). Rev Cancer. 18(5):269–82. https://doi.org/10.1038/nrc.2018.11
Patient navigation across the spectrum of women’s health care in PMID:29497144, by permission from Springer Nature © 2018.
the United States. Am J Obstet Gynecol. 218(3):280–86. https://doi.
org/10.1016/j.ajog.2017.08.009 PMID:28844825, Copyright 2018, 5.3.1 Reproduced from Brown IS, Fujii S, Kawachi H, Lam AK,
with permission from Elsevier. Saito T (2019). Oesophageal squamous cell carcinoma NOS. In:
WHO Classification of Tumours Editorial Board. Digestive sys-
B4.6.2 Carol E. DeSantis. Compiled from 2012–2016 data from tem tumours. Lyon: International Agency for Research on Cancer
the National Center for Health Statistics, Centers for Disease (WHO Classification of Tumours series, 5th edition); pp. 48–53.
Control and Prevention, 2018. Available from: http://publications.iarc.fr/579.
4.7.1 Diana Sarfati, Bridget H. Robson, and Gail Garvey. 5.3.2 Reproduced from Ferlay J, Ervik M, Lam F, Colombet M,
4.7.2 © iStockphoto/filipefrazao. Mery L, Piñeros M, et al. (2018). Global Cancer Observatory:
Cancer Today. Lyon, France: International Agency for Research
4.7.3 © 2012 Victor Casillas Romo, Courtesy of Photoshare. on Cancer. Available from: https://gco.iarc.fr/today.
4.7.4 Courtesy of Nunukul Yuggera Aboriginal Dance Company; 5.3.3 © 2012 Meagan Harrison, Courtesy of Photoshare.
Reon Enoch (left) and Aaron Ruska (right). Photo by Jen Dainer.
B5.3.1 (Map) Reproduced with permission from World Drug
4.7.5 Diana Sarfati, Bridget H. Robson, and Gail Garvey. Report 2006, United Nations Publication, Sales No. E.06.XI.10;
5.1.1 Reproduced from Ferlay J, Ervik M, Lam F, Colombet M, ISBN 92-1-148214-3, Volume 1: Analysis. (Inset photo) iStock-
Mery L, Piñeros M, et al. (2018). Global Cancer Observatory: photo.com/sadikgulec.
Cancer Today. Lyon, France: International Agency for Research B5.3.2 Courtesy of Digestive Disease Research Institute, Tehran
on Cancer. Available from: https://gco.iarc.fr/today. University of Medical Sciences, Shariati Hospital.
5.1.2 Compiled from Ferlay J, Colombet M, Bray F (2018). Cancer 5.4.1 AJ Cann. Courtesy of Flickr.
Incidence in Five Continents, CI5plus: IARC CancerBase No. 9
[Internet]. Lyon, France: International Agency for Research on 5.4.2 Reproduced from Zamani M, Ebrahimtabar F, Zamani V,
Cancer. Available from: http://ci5.iarc.fr. Miller WH, Alizadeh-Navaei R, Shokri-Shirvani J, et al. (2018).
Systematic review with meta-analysis: the worldwide preva-
5.1.3 Adi F. Gazdar. A Compiled from Cancer Genome Atlas lence of Helicobacter pylori infection. Aliment Pharmacol Ther.
Research Network (2014). Comprehensive molecular profiling 47(7):868–76. https://doi.org/10.1111/apt.14561 PMID:29430669,
of lung adenocarcinoma. Nature. 511(7511):543–50. https://doi. with permission from Wiley.
org/10.1038/nature13385 PMID:25079552. B & C Compiled from
Cancer Genome Atlas Research Network (2012). Comprehensive 5.4.3 Reproduced from Anderson WF, Rabkin CS, Turner N,
genomic characterization of squamous cell lung cancers. Fraumeni JF Jr, Rosenberg PS, Camargo MC (2018). The chang-
Nature. 489(7417):519–25. https://doi.org/10.1038/nature11404 ing face of noncardia gastric cancer incidence among US non-
PMID:22960745. D Compiled from George J, Lim JS, Jang SJ, Hispanic whites. J Natl Cancer Inst. 110(6):608–15. https://doi.
Cun Y, Ozretić L, Kong G, et al. (2015). Comprehensive geno- org/10.1093/jnci/djx262 PMID:29361173, by permission of Oxford
mic profiles of small cell lung cancer. Nature. 524(7563):47–53. University Press.
https://doi.org/10.1038/nature14664 PMID:26168399 5.4.4 Adapted from Carrasco-Garcia E, García-Puga M, Arevalo
5.1.4 Adi F. Gazdar. S, Matheu A (2018). Towards precision medicine: linking genetic
and cellular heterogeneity in gastric cancer. Ther Adv Med Oncol.
5.1.5 Tina Encarnacion/UConn Health Photo.
10:1758835918794628. https://doi.org/10.1177/1758835918794628
PMID:30181784, © Carrasco-Garcia et al., 2018.
588
5.4.5 Reproduced from Ferreira RM, Pereira-Marques J, Pinto- 5.9.1 & 5.9.2 Reproduced from Ferlay J, Ervik M, Lam F, Colombet
Ribeiro I, Costa JL, Carneiro F, Machado JC, et al. (2018). Gastric M, Mery L, Piñeros M, et al. (2018). Global Cancer Observatory:
microbial community profiling reveals a dysbiotic cancer-asso- Cancer Today. Lyon, France: International Agency for Research
ciated microbiota. Gut. 67(2):226–36. https://doi.org/10.1136/ on Cancer. Available from: https://gco.iarc.fr/today.
gutjnl-2017-314205 PMID:29102920, © Ferreira et al. 2018. All
5.9.3 Courtesy of Catherine Askew. Compiled from Liu Y, Zhang J,
rights reserved. No commercial use is permitted unless otherwise
Huang R, Feng W, Kong Y, Xu F, et al. (2017). Influence of occupa-
expressly granted.
tion and education level on breast cancer stage at diagnosis, and
5.4.6 chrisfxwolf0. Courtesy of Pixabay. treatment options in China: a nationwide, multicenter 10-year epi-
demiological study. Medicine (Baltimore). 96(15):e6641. https://doi.
5.4.7 iStockphoto.com/kot63.
org/10.1097/MD.0000000000006641 PMID:28403116; Lousdal ML,
5.5.1 Reproduced from Fitzmaurice C, Akinyemiju TF, Al Lami Kristiansen IS, Møller B, Støvring H (2014). Trends in breast cancer
FH, Alam T, Alizadeh-Navaei R, Allen C, et al.; Global Burden of stage distribution before, during and after introduction of a screening
Disease Cancer Collaboration (2018). Global, regional, and nation- programme in Norway. Eur J Public Health. 24(6):1017–22. https://
al cancer incidence, mortality, years of life lost, years lived with dis- doi.org/10.1093/eurpub/cku015 PMID:24596400; Castillo CSM,
ability, and disability-adjusted life-years for 29 cancer groups, 1990 Cabrera MEC, Derio PL, Gaete VF, Cavada CG (2017). Impact of the
to 2016: a systematic analysis for the Global Burden of Disease Chilean Explicit Guaranties Health System (GES) on breast cancer
Study. JAMA Oncol. 4(11):1553–68. https://doi.org/10.1001/ treatment. Rev Med Chile. 145(12):1507–13. https://doi.org/10.4067/
jamaoncol.2018.2706 PMID:29860482, eFigure 7, Supplement, p. s0034-98872017001201507 PMID:29652946; Jedy-Agba E,
184, © 2018 Global Burden of Disease Cancer Collaboration. McCormack V, Adebamowo C, Dos-Santos-Silva I (2016). Stage
at diagnosis of breast cancer in sub-Saharan Africa: a systematic
5.5.2 & 5.3.3 Kaboompics.com. Courtesy of Pexels.
review and meta-analysis. Lancet Glob Health. 4(12):e923–e935.
5.5.4 Joshua Paul Shefman. https://doi.org/10.1016/S2214-109X(16)30259-5 PMID:27855871;
Iqbal J, Ginsburg O, Rochon PA, Sun P, Narod SA (2015). Differences
5.6.1 Geoff Whiteway.
in breast cancer stage at diagnosis and cancer-specific survival
5.6.2 Adapted from Chen CJ, Yang HI (2011). Natural history by race and ethnicity in the United States. JAMA. 313(2):165–73.
of chronic hepatitis B REVEALed. J Gastroenterol Hepatol. https://doi.org/10.1001/jama.2014.17322 PMID:25585328
26(4):628–38. https://doi.org/10.1111/j.1440-1746.2011.06695.x
5.9.4 Reprinted from Michailidou K, Lindström S, Dennis J,
PMID:21323729, with permission from John Wiley & Sons.
Beesley J, Hui S, Kar S, et al. (2017). Association analysis iden-
5.6.3 Adapted from Yang HI, Lee MH, Liu J, Chen CJ (2014). Risk tifies 65 new breast cancer risk loci. Nature. 551(7678):92–4.
calculators for hepatocellular carcinoma in patients affected with https://doi.org/10.1038/nature24284 PMID:29059683, by permis-
chronic hepatitis B in Asia. World J Gastroenterol. 20(20):6244– sion from Springer Nature, Copyright © 2017.
51. https://doi.org/10.3748/wjg.v20.i20.6244 PMID:24876745
5.9.5 Reproduced from Palmer JR, Viscidi E, Troester MA, Hong
5.6.4 Courtesy of Dr Christian Wittekind, Institute of Pathology, CC, Schedin P, Bethea TN, et al. (2014). Parity, lactation, and breast
University of Leipzig. Reproduced with permission from Wiegand J, cancer subtypes in African American women: results from the
Berg T (2013). The etiology, diagnosis and prevention of liver cirrho- AMBER Consortium. J Natl Cancer Inst. 106(10):dju237. https://doi.
sis: part 1 of a series on liver cirrhosis. Dtsch Arztebl Int. 110(6):85– org/10.1093/jnci/dju237 PMID:25224496, by permission of Oxford
91. https://doi.org/10.3238/arztebl.2013.0085 PMID:23451000 University Press.
5.6.5 Reprinted from Yang HI, Sherman M, Su J, Chen PJ, Liaw 5.9.6 Courtesy of Catherine Askew. Compiled from Key TJ,
YF, Iloeje UH, et al. Nomograms for risk of hepatocellular carci- Appleby PN, Reeves GK, Travis RC, Alberg AJ, Barricarte A,
noma in patients with chronic hepatitis B virus infection. J Clin et al.; Endogenous Hormones and Breast Cancer Collaborative
Oncol. 28(14):2437–44. https://doi.org/10.1200/JCO.2009.27.4456 Group (2013). Sex hormones and risk of breast cancer in pre-
PMID:20368541, with permission. © 2010, American Society of menopausal women: a collaborative reanalysis of individual
Clinical Oncology. All rights reserved. participant data from seven prospective studies. Lancet Oncol.
14(10):1009–19. https://doi.org/10.1016/S1470-2045(13)70301-2
5.7.1 & 5.7.2 Jessica N. Everett and Diane M. Simeone.
PMID:23890780; Ge W, Clendenen TV, Afanasyeva Y, Koenig
5.7.3 Reproduced from The Cancer Genome Atlas Research KL, Agnoli C, Brinton LA, et al. (2018). Circulating anti-Müllerian
Network (2017). Integrated genomic characterization of pan- hormone and breast cancer risk: a study in ten prospective co-
creatic ductal adenocarcinoma. Cancer Cell. 32(2):185–203. horts. Int J Cancer. 142(11):2215–26. https://doi.org/10.1002/
e13. https://doi.org/10.1016/j.ccell.2017.07.007 PMID:28810144, ijc.31249 PMID:29315564.
© 2017 The Cancer Genome Atlas Research Network. Published
5.9.7 Reproduced from Cuzick J, Sestak I, Cawthorn S, Hamed H,
by Elsevier Inc.
Holli K, Howell A, et al.; IBIS-I Investigators (2015). Tamoxifen for
5.7.4 Reprinted from Sharma A, Smyrk TC, Levy MJ, Topazian MA, prevention of breast cancer: extended long-term follow-up of the
Chari ST (2018). Fasting blood glucose levels provide estimate of IBIS-I breast cancer prevention trial. Lancet Oncol. 16(1):67–75.
duration and progression of pancreatic cancer before diagnosis. https://doi.org/10.1016/S1470-2045(14)71171-4 PMID:25497694,
Gastroenterology. 155(2):490–500.e2. https://doi.org/10.1053/j. © 2015 Cuzick et al. Open Access article distributed under the
gastro.2018.04.025 PMID:29723506, Copyright 2018, with permis- terms of CC BY. Published by Elsevier Ltd.
sion from Elsevier.
5.10.1 & 5.10.2 Reproduced from Ferlay J, Ervik M, Lam F,
5.8.1 Courtesy of National Cancer Institute. Released under US Colombet M, Mery L, Piñeros M, et al. (2018). Global Cancer
public domain. Observatory: Cancer Today. Lyon, France: International Agency
for Research on Cancer. Available from: https://gco.iarc.fr/today.
5.8.2 Daniel Sone. Courtesy of National Cancer Institute.
Released under US public domain. 5.10.3 & 5.10.4 Courtesy of Dr Hue-Tsi Wu.
5.8.3 Ben Kerckx. Courtesy of Pixabay. 5.10.5 Reprinted from Litjens RJ, Hopman AH, van de Vijver KK,
Ramaekers FC, Kruitwagen RF, Kruse AJ (2013). Molecular bio-
5.8.4 David Whiteman.
markers in cervical cancer diagnosis: a critical appraisal. Expert
5.8.5 Hans Braxmeier. Courtesy of Pixabay. Opin Med Diagn. 7(4):365–77. https://doi.org/10.1517/17530059.
2013.808621 PMID:23777477, © 2013 Taylor & Francis, reprinted
5.8.6 Courtesy of Tom Sales.
by permission of the publisher Taylor & Francis Ltd, http://www.
5.8.7 ambermb. Courtesy of Pixabay. tandfonline.com.

5.11.1 Compiled from Ferlay J, Colombet M, Bray F (2018). 5.14.3 Compiled from Ferlay J, Colombet M, Bray F (2018).
Cancer Incidence in Five Continents, CI5plus: IARC CancerBase Cancer Incidence in Five Continents, CI5plus: IARC CancerBase
No. 9 [Internet]. Lyon, France: International Agency for Research No. 9 [Internet]. Lyon, France: International Agency for Research
on Cancer. Available from: http://ci5.iarc.fr. on Cancer. Available from: http://ci5.iarc.fr.
5.11.2 Penelope M. Webb. Compiled from Setiawan VW, Yang HP, 5.14.4 Reproduced from Ferlay J, Ervik M, Lam F, Colombet M,
Pike MC, McCann SE, Yu H, Xiang YB, et al.; Australian National Mery L, Piñeros M, et al. (2018). Global Cancer Observatory:
Endometrial Cancer Study Group (2013). Type I and II endometrial Cancer Today. Lyon, France: International Agency for Research
cancers: have they different risk factors? J Clin Oncol. 31(20):2607– on Cancer. Available from: https://gco.iarc.fr/today.
18. https://doi.org/10.1200/JCO.2012.48.2596 PMID:23733771
5.14.5 Rex Medlen. Courtesy of Pixabay.
5.11.3 Penelope M. Webb. Compiled from Whiteman DC, Webb
PM, Green AC, Neale RE, Fritschi L, Bain CJ, et al. (2015). 5.14.6 Ewa Rajpert-De Meyts and Niels E. Skakkebæk.
Cancers in Australia in 2010 attributable to modifiable factors: 5.15.1 Courtesy of Nikola M. Pavlovic.
summary and conclusions. Aust N Z J Public Health. 39(5):477–84.
https://doi.org/10.1111/1753-6405.12471 PMID:26437735; Parkin 5.15.2 Courtesy of Volker Arlt. Adapted from Stiborová M, Arlt
DM, Boyd L, Walker LC (2011). 16. The fraction of cancer attribu- VM, Schmeiser HH (2017). DNA adducts formed by aristolochic
table to lifestyle and environmental factors in the UK in 2010. Br J acid are unique biomarkers of exposure and explain the initia-
Cancer. 105(Suppl 2):S77–81. https://doi.org/10.1038/bjc.2011.489 tion phase of upper urothelial cancer. Int J Mol Sci. 18(10):2144.
PMID:22158327 https://doi.org/10.3390/ijms18102144 PMID:29036902
5.11.4 Penelope M. Webb. Compiled from Luo J, Chlebowski 5.15.3A Courtesy of Jean-Louis Vanherweghem. Reproduced
RT, Hendryx M, Rohan T, Wactawski-Wende J, Thomson from Nortier J, Pozdzik A, Roumeguere T, Vanherweghem J-L
CA, et al. (2017). Intentional weight loss and endometrial can- (2013). Néphropathie aux acides aristolochiques (« néphropa-
cer risk. J Clin Oncol. 35(11):1189–93. https://doi.org/10.1200/ thie aux herbes chinoises »). Encyclopédie Médico-Chirurgicale
JCO.2016.70.5822 PMID:28165909 (EMC). Néphrologie. 10(2):1–14 [Article 18-040-J-10]. Copyright
2013 Elsevier Masson SAS. All rights reserved.
Colombet M, Mery L, Piñeros M, et al. (2018). Global Cancer 5.15.3B Courtesy of Jessica Maufort.
Observatory: Cancer Today. Lyon, France: International Agency 5.15.3C Courtesy of 天問小窩. Released under CC BY 2.0.
5.15.4 & 5.15.5 Courtesy of Sandrine Rorive.
5.12.3 Reproduced from Coburn SB, Bray F, Sherman ME, Trabert
B (2017). International patterns and trends in ovarian cancer inci- 5.16.1 Reproduced from Ferlay J, Ervik M, Lam F, Colombet M,
dence, overall and by histologic subtype. Int J Cancer. 140(11):2451– Mery L, Piñeros M, et al. (2018). Global Cancer Observatory:
60. https://doi.org/10.1002/ijc.30676 PMID:28257597, with permis- Cancer Today. Lyon, France: International Agency for Research
sion from John Wiley & Sons. on Cancer. Available from: https://gco.iarc.fr/today.
5.12.4 Austin Kirk. Available on Flickr. License CC BY 2.0. 5.16.2 hamiltonpaviana. Courtesy of Pixabay.
5.12.5 © 2017 Magali Rochat/VectorWorks, Courtesy of Photoshare. 5.16.3 iStockphoto.com/LuckyBusiness.
5.13.1 & 5.13.2 Reproduced from Ferlay J, Ervik M, Lam F, 5.17.1 Courtesy of David N. Louis.
Colombet M, Mery L, Piñeros M, et al. (2018). Global Cancer
5.17.2 Reproduced from Brat DJ, Verhaak RG, Aldape KD, Yung
Observatory: Cancer Today. Lyon, France: International Agency
WK, Salama SR, Cooper LA, et al.; Cancer Genome Atlas Research
Network (2015). Comprehensive, integrative genomic analysis of dif-
5.13.3 Compiled from NORDCAN (https://www.ancr.nu/); Cancer fuse lower-grade gliomas. N Engl J Med. 372(26):2481–98. https://
Incidence in Five Continents (http://ci5.iarc.fr); Australian Institute doi.org/10.1056/NEJMoa1402121 PMID:26061751, Copyright 2015,
of Health and Welfare (https://www.aihw.gov.au); National Cancer Massachusetts Medical Society. Reprinted with permission from
Institute Surveillance, Epidemiology, and End Results Program, Massachusetts Medical Society.
USA (https://seer.cancer.gov)
5.17.3 fizkes/Getty Images.
5.13.4 Compiled from WHO Mortality Database. Available from:
http://www.who.int/healthinfo/statistics/mortality _rawdata/en/ 5.17.4 Reproduced from Ostrom QT, Gittleman H, Liao P,
index.html. Vecchione-Koval T, Wolinsky Y, Kruchko C, et al. (2017). CBTRUS
statistical report: primary brain and other central nervous sys-
5.13.5 nappy. Courtesy of Pexels. tem tumors diagnosed in the United States in 2010–2014. Neuro
5.13.6 Darryl Leja, NHGRI. Available on Flickr. License CC BY 2.0. Oncol. 19(Suppl 5):v1–v88. https://doi.org/10.1093/neuonc/nox158
PMID:29117289, by permission of Oxford University Press.
5.13.7 iStockphoto.com/Jolkesky.
5.18.1 Reproduced from Ferlay J, Ervik M, Lam F, Colombet M,
5.14.1 Reproduced from Skakkebaek NE, Rajpert-De Meyts E, Mery L, Piñeros M, et al. (2018). Global Cancer Observatory:
Buck Louis GM, Toppari J, Andersson AM, Eisenberg ML, et al.
(2016). Male reproductive disorders and fertility trends: influences
on Cancer. Available from: https://gco.iarc.fr/today.
of environment and genetic susceptibility. Physiol Rev. 96(1):55–
97. https://doi.org/10.1152/physrev.00017.2015 PMID:26582516, 5.18.2 Stefan Krasowski. Courtesy of Flickr. License CC BY 2.0.
© 2016 the American Physiological Society. Adapted from Rajpert-
5.18.3 iStockphoto.com/jjneff.
De Meyts E (2006). Developmental model for the pathogenesis of
testicular carcinoma in situ: genetic and environmental aspects. 5.18.4 iStockphoto.com/FabVietnam_Photography.
Hum Reprod Update. 106(12):303–23. https://doi.org/10.1093/
humupd/dmk006 PMID:16540528, by permission of Oxford
Colombet M, Mery L, Piñeros M, et al. (2018). Global Cancer
University Press.
Observatory: Cancer Today. Lyon, France: International Agency
5.14.2 Reproduced from Ferlay J, Ervik M, Lam F, Colombet M, for Research on Cancer. Available from: https://gco.iarc.fr/today.
Mery L, Piñeros M, et al. (2018). Global Cancer Observatory:
5.19.3 Sonja I. Berndt.
on Cancer. Available from: https://gco.iarc.fr/today.
590
5.19.4 Courtesy of Sonja I. Berndt. Compiled from Todar K. Immune 6.5.1 Adapted from Slavin TP, Niell-Swiller M, Solomon I,
defense against bacterial pathogens: adaptive or acquired immu- Nehoray B, Rybak C, Blazer KR, et al. (2015). Clinical application
nity. In: Todar’s online textbook of bacteriology. http://textbookof of multigene panels: challenges of next-generation counseling
bacteriology.net/adaptive_2.html; Lunardi A, Guarnerio J, Wang and cancer risk management. Front Oncol. 5:208. https://doi.
G, Maeda T, Pandolfi PP (2013). Role of LRF/Pokemon in lineage org/10.3389/fonc.2015.00208 PMID:26484312, © 2015 Slavin,
fate decisions. Blood. 121(15):2845–53. https://doi.org/10.1182/ Niell-Swiller, Solomon, Nehoray, Rybak, Blazer, and Weitzel.
blood-2012-11-292037 PMID:23396304; Halin C, Mora JR, 6.5.2 Patricia Ashton-Prolla.
Sumen C, von Andrian UH (2005). In vivo imaging of lympho-
cyte trafficking. Annu Rev Cell Dev Biol. 21:581–603. https://doi. 6.6.1 Raúl Murillo.
org/10.1146/annurev.cellbio.21.122303.133159 PMID:16212508; 6.6.2 35007/Getty Images.
He Y, Jiang X, Chen J (2014). The role of miR-150 in normal and
malignant hematopoiesis. Oncogene. 33(30):3887–93. https:// 6.6.3 WHO/Sergey Volkov.
doi.org/10.1038/onc.2013.346 PMID:23955084 6.6.4 Reproduced from Herrero R, Murillo R (2018). Cervical can-
cer. In: Thun M, Linet MS, Cerhan JR, Haiman CA, Schottenfeld
5.19.5 dlewis33/Getty Images.
D, editors. Cancer epidemiology and prevention. 4th ed. New York
5.19.6 Christopher Glass/IMA World Health. (NY), USA: Oxford University Press; pp. 925–46. Reproduced
with permission of the Licensor through PLSclear.
5.20.1–5.20.6 Eve Roman and Alexandra G. Smith.
6.6.5 Raúl Murillo.
6.1.1 Reprinted from WHO (2018). Tobacco Free Initiative (TFI).
MPOWER brochures and other resources. Available from: http:// 6.7.1 Anna Babayan, Natalie Reimers, and Klaus Pantel.
www.who.int/tobacco/mpower/publications/en/. 6.7.2 iStockphoto.com/Willowpix.
6.1.2 Reprinted from Schweitzer A, Akmatov MK, Krause G (2017). 6.7.3 Amornthep Srina. Courtesy of Pexels.
Hepatitis B vaccination timing: results from demographic health
surveys in 47 countries. Bull World Health Organ. 95(3):199– 6.8.1 © European Union 2018 – European Parliament. Released
209G. https://doi.org/10.2471/BLT.16.178822 PMID:28250533, under License CC BY-NC-ND 2.0.
Copyright 2016. 6.8.2 Stockbyte/Getty Images.
6.1.3 Reprinted with permission from Dr Deepa Gamage, Consultant 6.8.3 Republished with permission from National Research Council
Epidemiologist, Epidemiology Unit, Ministry of Health, Sri Lanka. (1983). Risk assessment in the federal government: managing
6.1.4 AfrOx (Africa Oxford Cancer Foundation). the process. Washington (DC), USA: National Academies Press.
https://doi.org/10.17226/366, © 1983 by the National Academy of
6.2.1 Reproduced from World Cancer Research Fund International Sciences, Courtesy of the National Academies Press.
Driving Action policy framework (www.wcrf.org/drivingaction).
6.8.4 iStockphoto.com/flySnow.
6.2.2 Stokpic. Courtesy of Pixabay.
6.8.5 USEPA. Courtesy of Wikimedia Commons. Released under
6.2.3 Thomas Chauke. Courtesy of Pexels. US public domain.
6.2.4 ASSY. Courtesy of Pixabay. 6.9.1 David J. Hunter and K. Srinath Reddy. Compiled from WHO
(2018). Global health estimates 2016: deaths by cause, age, sex,
6.2.5 © 2013 Min Zaw, Courtesy of Photoshare.
by country and by region, 2000–2016. Geneva, Switzerland:
6.2.6 iStockphoto.com/mokee81. World Health Organization. Available from: https://www.who.int/
healthinfo/global_burden_disease/estimates/en/.
6.3.1 Reproduced from Maucort-Boulch D, de Martel C, Franceschi
S, Plummer M (2018). Fraction and incidence of liver cancer at- 6.9.2 David J. Hunter and K. Srinath Reddy. Compiled from
tributable to hepatitis B and C viruses worldwide. Int J Cancer. NCHS (2012). Health, United States, 2011: with special feature on
142(12):2471–7. https://doi.org/10.1002/ijc.31280 PMID:29388206. socioeconomic status and health. Hyattsville (MD), USA: National
John Wiley & Sons, Inc. © 2018 International Agency for Research Center for Health Statistics. PMID:22812021
on Cancer (IARC/WHO); licensed by UICC. Open Access. 6.9.3 Courtesy of National Health Education, Information and
6.3.2 © Séverine Bonnet/Médecins Sans Frontières. Communication Centre, Ministry of Health and Population, Nepal.
6.3.3 Reproduced from WHO (2017). Global hepatitis report 6.9.4 Courtesy of Public Health & Reforms Center of Ministry of
2017. Geneva, Switzerland: World Health Organization. Available Health of Azerbaijan Republic, 2018.
from: https://www.who.int/hepatitis/publications/global-hepatitis- 6.9.5 This Girl Can. Courtesy of Sport England.
report2017/en/; p. 24. Copyright 2017.
P1.1 From Asma S, Mackay J, Song SY, Zhao L, Morton J,
6.3.4 Reproduced from Baussano I, Lazzarato F, Brisson M, Palipudi KM, et al. (2015). The GATS atlas: Global Adult Tobacco
Franceschi S (2016). Human papillomavirus vaccination at a time of Survey. Atlanta (GA), USA: CDC Foundation. Available from:
changing sexual behavior. Emerg Infect Dis. 22(1):18–23. https://doi. http://gatsatlas.org/.
org/10.3201/eid2201.150791 PMID:26691673
P1.2 Compiled from West R, Raw M, McNeill A, Stead L, Aveyard
6.4.1 Reproduced from Cuzick J, Sestak I, Cawthorn S, Hamed H, P, Bitton J, et al. (2015). Health-care interventions to promote and
Holli K, Howell A, et al.; IBIS-I Investigators (2015). Tamoxifen for assist tobacco cessation: a review of efficacy, effectiveness and
prevention of breast cancer: extended long-term follow-up of the affordability for use in national guideline development. Addiction.
IBIS-I breast cancer prevention trial. Lancet Oncol. 16(1):67–75. 110(9):1388–403. https://doi.org/10.1111/add.12998 PMID:26031929
https://doi.org/10.1016/S1470-2045(14)71171-4 PMID:25497694, P1.3 WHO.
© 2015 Cuzick et al. Open Access article distributed under the
terms of CC BY. Published by Elsevier Ltd. page 15 Pogonici/Getty Images.
6.4.2 Reprinted from Rothwell PM, Fowkes FG, Belch JF, Ogawa page 49 rawpixel.com. Courtesy of Pexels.
H, Warlow CP, Meade TW (2011). Effect of daily aspirin on long- page 145 Reproduced from WCRF/AICR (2018). Diet, nutrition,
term risk of death due to cancer: analysis of individual patient physical activity and cancer: a global perspective. Continuous
data from randomised trials. Lancet. 377(9759):31–41. https://doi. Update Project Expert Report 2018. World Cancer Research
org/10.1016/S0140-6736(10)62110-1 PMID:21144578, Copyright Fund/American Institute for Cancer Research. Available from:
2011 with permission from Elsevier. http://dietandcancerreport.org.

page 147 Pogonici/Getty Images. page 297 Courtesy of Dmitry V. Kazakov, MD, Charles University
Medical Faculty Hospital, Pilsen, Czechia.
page 237 Johnny Miller/Unequal Scenes.
page 487 © 2017 Afshan Najafi, Courtesy of Photoshare.
page 296 © IARC.
Tables
2.1.1 Reproduced from Thun M, Linet MS, Cerhan JR, Haiman 3.4.1 Eugenia Dogliotti and Margherita Bignami.
CA, Schottenfeld D, editors (2018). Cancer epidemiology and
3.6.1 Reproduced from Pearce CL, Templeman C, Rossing MA,
prevention. 4th ed. New York (NY), USA: Oxford University
Lee A, Near AM, Webb PM, et al.; Ovarian Cancer Association
Press, Table 11.3, p. 198, by permission of Oxford University
Consortium (2012). Association between endometriosis and risk
Press, USA (www.oup.com); adapted from U.S. Department of
of histological subtypes of ovarian cancer: a pooled analysis of
Health and Human Services (2014). The health consequences of
case-control studies. Lancet Oncol. 13(4):385–94. https://doi.
smoking – 50 years of progress: a report of the Surgeon General.
Atlanta (GA), USA: U.S. Department of Health and Human org/10.1016/S1470-2045(11)70404-1 PMID:22361336, © 2012
Services, Centers for Disease Control and Prevention, National Elsevier Ltd. Published by Elsevier Ltd
Center for Chronic Disease Prevention and Health Promotion, 3.6.2 Reproduced from Brinton LA, Cook MB, McCormack V,
Office on Smoking and Health. Available from: https://www. Johnson KC, Olsson H, Casagrande JT, et al.; European Rare
ncbi.nlm.nih.gov/books/NBK179276/; from Carter BD, Abnet CC, Cancer Study Group (2014). Anthropometric and hormonal risk
Feskanich D, Freedman ND, Hartge P, Lewis CE, et al. (2015). factors for male breast cancer: Male Breast Cancer Pooling
Smoking and mortality – beyond established causes. N Engl J Project results. J Natl Cancer Inst. 106(3):djt465. https://doi.
Med. 372(7):631–40. https://doi.org/10.1056/NEJMsa1407211 org/10.1093/jnci/djt465 PMID:24552677, by permission of Oxford
PMID:25671255, Copyright 2015, Massachusetts Medical University Press.
Society. Reprinted with permission from Massachusetts Medical
Society; from Siegel RL, Jacobs EJ, Newton CC, Feskanich D, 3.6.3 Reproduced from Black A, Pinsky PF, Grubb RL 3rd,
Freedman ND, Prentice RL, et al. (2015). Deaths due to cigarette Falk RT, Hsing AW, Chu L, et al. (2014). Sex steroid hormone
smoking for 12 smoking-related cancers in the United States. metabolism in relation to risk of aggressive prostate cancer.
JAMA Intern Med. 175(9):1574–6. https:/doi.org/10.1001/jama Cancer Epidemiol Biomarkers Prev. 23(11):2374–82. https://doi.
internmed.2015.2398 PMID:26076120, Copyright 2015, American org/10.1158/1055-9965.EPI-14-0700 PMID:25178985, Copyright
Medical Association. All rights reserved. 2014, American Association for Cancer Research.
2.2.1 Robert Newton. Compiled from IARC (2012). Biological 3.11.1 Reproduced from Smith MT, Guyton KZ, Gibbons CF, Fritz
agents. IARC Monogr Eval Carcinog Risks Hum. 100B:1–441. JM, Portier CJ, Rusyn I, et al. (2016). Key characteristics of carcino-
Available from: http://publications.iarc.fr/119. PMID:23189750; gens as a basis for organizing data on mechanisms of carcinogene-
Plummer M, de Martel C, Vignat J, Ferlay J, Bray F, Franceschi sis. Environ Health Perspect. 124(6):713–21. https://doi.org/10.1289/
S (2016). Global burden of cancers attributable to infections in ehp.1509912 PMID:26600562, Reproduced from Environmental
2012: a synthetic analysis. Lancet Glob Health. 4(9):e609–16. Health Perspectives, https://ehp.niehs.nih.gov/15-09912/
https://doi.org/10.1016/S2214-109X(16)30143-7 PMID:27470177; 4.2.1 Reprinted from Allemani C, Matsuda T, Di Carlo V, Harewood
de Martel C, Georges D, Bray F, Ferlay J, Clifford G (2020). Global R, Matz M, Nikšić M, et al.; CONCORD Working Group (2018).
burden of cancers attributable to infections in 2018: a worldwide
Global surveillance of trends in cancer survival 2000–2014
incidence analysis. Lancet Glob Health. 8(2):e180–90. https://doi.
(CONCORD-3): analysis of individual records for 37 513 025 pa-
org/10.1016/S2214-109X(19)30488-7 PMID:31862245
tients diagnosed with one of 18 cancers from 322 population-based
2.3.1 & 2.3.2 Jürgen Rehm, Kevin D. Shield, and Elisabete registries in 71 countries. Lancet. 391(10125):1023–75. https://doi.
Weiderpass. Compiled from data in WHO (2018). Global sta- org/10.1016/S0140-6736(17)33326-3 PMID:29395269, Copyright
tus report on alcohol and health 2018. Geneva, Switzerland: 2018, with permission from Elsevier.
World Health Organization. Available from: https://www.who.int/
4.2.2 Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin
substance_abuse/publications/global_alcohol_report/en/.
DM (2010). GLOBOCAN 2008, Cancer Incidence and Mortality
2.7.1 Christine M. Friedenreich and Michael Leitzmann. Worldwide: IARC CancerBase No. 10 [Internet]. Lyon, France:
2.9.1 Reproduced from IARC Monographs on the Identification International Agency for Research on Cancer.
of Carcinogenic Hazards to Humans, Agents Classified by the 4.5.1–4.5.3 Harry J. de Koning. Compiled from EU-topia (2018).
IARC Monographs: https://monographs.iarc.fr/agents-classified- EU-topia: towards improved cancer screening in all of Europe.
by-the-iarc/. Available from: www.eu-topia.org; Ponti A, Anttila A, Ronco
2.10.1 & 2.10.2 Reproduced from IARC Monographs on the G, Senore C, Basu P, Segnan N, et al. (2017). Against Cancer.
Identification of Carcinogenic Hazards to Humans, Agents Cancer screening in the European Union. Report on the imple-
Classified by the IARC Monographs: https://monographs.iarc.fr/ mentation of the Council Recommendation on cancer screening.
agents-classified-by-the-iarc/. Brussels, Belgium: European Commission. Available from: https://
ec.europa.eu/health/sites/health/files/major_chronic_diseases/
2.10.3 Jack Siemiatycki and Lesley Rushton. docs/2017_cancerscreening_2ndreportimplementation_en.pdf.
2.11.1 Lisa Iversen. 4.6.1 Reproduced from Smith RA, Andrews KS, Brooks D,
3.2.1 Stephen J. Chanock. Fedewa SA, Manassaram-Baptiste D, Saslow D, et al. (2017).
Cancer screening in the United States, 2017: a review of cur-
3.3.1 Adapted from Smith MT, Guyton KZ, Gibbons CF, Fritz JM,
rent American Cancer Society guidelines and current issues in
Portier CJ, Rusyn I, et al. (2016). Key characteristics of carcinogens
cancer screening. CA Cancer J Clin. 67(2):100–21. https://doi.
as a basis for organizing data on mechanisms of carcinogenesis.
org/10.3322/caac.21392 PMID:28170086, with permission from
Environ Health Perspect. 124(6):713–21. https://doi.org/10.1289/
John Wiley & Sons.
ehp.1509912 PMID:26600562, Reproduced from Environmental
Health Perspectives, https://ehp.niehs.nih.gov/15-09912/ 5.1.1 Adi F. Gazdar.
592
5.1.2 Adi F. Gazdar. Compiled from Langevin SM, Kelsey KT B5.6.2 Compiled from Chiang CJ, Yang YW, You SL, Lai MS,
(2017). Clinical epigenetics of lung cancer. In: Laurence J, Van Chen CJ (2013). Thirty-year outcomes of the national hepatitis B
Beusekom M, editors. Translating epigenetics to the clinic. immunization program in Taiwan. JAMA. 310(9):974–6. https://doi.
London, UK: Elsevier; pp. 97–133. https://doi.org/10.1016/B978- org/10.1001/jama.2013.276701 PMID:24002285; Chiang CJ, Yang
0-12-800802-7.00005-8; Duruisseaux M, Esteller M (2018). YW, Chen JD, You SL, Yang HI, Lee MH, et al. (2015). Significant
Lung cancer epigenetics: from knowledge to applications. reduction in end-stage liver diseases burden through the national
Semin Cancer Biol. 51:116–28. https://doi.org/10.1016/j.semcan viral hepatitis therapy program in Taiwan. Hepatology. 61(4):1154–
cer.2017.09.005 PMID:28919484; Gazdar AF, Bunn PA, Minna 62. https://doi.org/10.1002/hep.27630 PMID:25476749
JD (2017). Small-cell lung cancer: what we know, what we need
5.7.1 Jessica N. Everett and Diane M. Simeone.
to know and the path forward. Nat Rev Cancer. 17(12):725–37.
https://doi.org/10.1038/nrc.2017.87 PMID:29077690 5.8.1 Adapted from Gershenwald JE, Scolyer RA, Hess KR, Sondak
VK, Long GV, Ross MI, et al. (2017). Melanoma staging: evidence-
5.2.1 & 5.2.2 Reproduced from de Martel C, Plummer M, Vignat
J, Franceschi S (2017). Worldwide burden of cancer attributable based changes in the American Joint Committee on Cancer eighth
to HPV by site, country and HPV type. Int J Cancer. 141(4):664– edition cancer staging manual. CA Cancer J Clin. 67(6):472–92.
70. https://doi.org/10.1002/ijc.30716 PMID:28369882, © 2017 https://doi.org/10.3322/caac.21409 PMID:29028110, with permis-
International Agency for Research on Cancer (IARC/WHO); lision from John Wiley & Sons; used with permission of the American
censed by UICC. Open Access. College of Surgeons, Chicago, Illinois. The original source for this
information is the AJCC Cancer Staging Manual, 8th edition (2017),
5.2.3 Reprinted from Leemans CR, Snijders PJF, Brakenhoff RH published by Springer International Publishing.
(2018). The molecular landscape of head and neck cancer. Nat
Rev Cancer. 18(5):269–82. https://doi.org/10.1038/nrc.2018.11 5.9.1 Adapted from Colditz GA, Bohlke K (2014). Priorities for the
PMID:29497144, by permission from Springer Nature © 2018. primary prevention of breast cancer. CA Cancer J Clin. 64(3):186–
94. https://doi.org/10.3322/caac.21225 PMID:24647877, with per-
5.3.1 Reza Malekzadeh, Christian C. Abnet, and Sanford M. Dawsey. mission from John Wiley & Sons.
5.4.1 Compiled from Laurén P (1965). The two histological main 5.10.1 Adapted/translated from Lax SF, Horn LC, Löning T (2016).
types of gastric carcinoma: diffuse and so-called intestinal-type Categorization of uterine cervix tumors: what’s new in the 2014
carcinoma. An attempt at a histo-clinical classification. Acta Pathol WHO classification [in German]. Pathologe. 37(6):573–84. https://
Microbiol Scand. 64:31–49. https://doi.org/10.1111/apm.1965.64.1.31 doi.org/10.1007/s00292-016-0247-8 PMID:27770187, by permis-
PMID:14320675; Lauwers GY, Carneiro F, Graham DY, Curado MP, sion from Springer Nature © 2016.
Franceschi S, Montgomery E, et al. (2010). Gastric carcinoma. In:
Bosman FT, Carneiro F, Hruban RH, Theise ND, editors. WHO 5.10.2 Adapted from Kurman RJ, Carcangiu ML, Herrington S,
classification of tumours of the digestive system. Lyon, France: Young RH, editors (2014). WHO classification of tumours of fe-
International Agency for Research on Cancer (WHO Classification male reproductive organs. Lyon, France: International Agency for
of Tumours series, 4th edition); pp. 48–58. Research on Cancer (WHO Classification of Tumours series, 4th
edition). Available from: https://publications.iarc.fr/16.
5.4.2 Reproduced from Matsuoka T, Yashiro M (2018). Biomarkers
of gastric cancer: current topics and future perspective. World J 5.10.3 & 5.10.4 Adapted from Kurman RJ, Carcangiu ML, Herrington
Gastroenterol. 24(26):2818–32. https://doi.org/10.3748/wjg.v24.i26. S, Young RH, editors (2014). WHO classification of tumours of
2818 PMID:30018477, © Matsuoka T and Yashiro M, 2018. female reproductive organs. Lyon, France: International Agency
Published by Baishideng Publishing Group Inc. All rights reserved. for Research on Cancer (WHO Classification of Tumours series,
4th edition); Lax SF, Horn LC, Löning T (2016). Categorization of
5.5.1 Carlo Senore, Nereo Segnan, and Marc Gunter. Compiled uterine cervix tumors: what’s new in the 2014 WHO classification
from WCRF/AICR (2018). Diet, nutrition, physical activity and [in German]. Pathologe. 37(6):573–84. https://doi.org/10.1007/
colorectal cancer. Continuous Update Project Expert Report s00292-016-0247-8 PMID:27770187
2018. World Cancer Research Fund/American Institute for
Cancer Research. Available from: https://www.aicr.org/continuous- 5.10.5 Reproduced from Bhatla N, Berek J, Cuello M, Denny L,
update-project/reports/colorectal-cancer-2017-report.pdf; IARC Grenman S, Karunaratne K, et al. (2018). New revised FIGO stag-
(2012). Personal habits and indoor combustions. IARC Monogr ing of cervical cancer (2018). Abstract S020.2. Presented at the
Eval Carcinog Risks Hum. 100E:1–575. Available from: http:// XXII FIGO World Congress of Gynecology and Obstetrics, Rio
publications.iarc.fr/122 PMID:23193840 de Janeiro, Brazil, 14–19 October 2018. Int J Gynecol Obstet.
143(S3):43–157. https://doi.org/10.1002/ijgo.12584, with permis-
5.5.2 Carlo Senore, Nereo Segnan, and Marc Gunter. Compiled sion from John Wiley & Sons.
from Lauby-Secretan B, Vilahur N, Bianchini F, Guha N, Straif K;
International Agency for Research on Cancer Handbook Working 5.11.1 Penelope M. Webb. Compiled from Getz G, Gabriel SB,
Group (2018). The IARC perspective on colorectal cancer screen- Cibulskis K, Lander E, Sivachenko A, Sougnez C, et al.; Cancer
ing. N Engl J Med. 378(18):1734–40. https://doi.org/10.1056/ Genome Atlas Research Network (2013). Integrated genomic char-
NEJMsr1714643 PMID:29580179; Armaroli P, Villain P, Suonio acterization of endometrial carcinoma. Nature. 497(7447):67–73.
E, Almonte M, Anttila A, Atkin WS, et al. (2015). European Code https://doi.org/10.1038/nature12113 PMID:23636398; McAlpine J,
Against Cancer, 4th edition: cancer screening. Cancer Epidemiol. Leon-Castillo A, Bosse T (2018). The rise of a novel classification
39(Suppl 1):S139–52. https://doi.org/10.1016/j.canep.2015.10.021 system for endometrial carcinoma; integration of molecular sub-
PMID:26596722 classes. J Pathol. 244(5):538–49. https://doi.org/10.1002/path.5034
PMID:29344951
B5.5.1 Reproduced from IARC (2019). Colorectal cancer screen-
ing. IARC Handb Cancer Prev. 17:1–300. Available from: http:// 5.11.2 Penelope M. Webb.
publications.iarc.fr/573.
5.12.1 Adapted from Prat J (2012). Ovarian carcinomas: five dis-
5.6.1–5.6.3 Chien-Jen Chen. tinct diseases with different origins, genetic alterations, and clin-
icopathological features. Virchows Arch. 460(3):237–49. https://
B5.6.1 Adapted from Yang HI, Tseng TC, Liu J, Lee MH, Liu CJ, Su
doi.org/10.1007/s00428-012-1203-5 PMID:22322322, by permis-
TH, et al. (2016). Incorporating serum level of hepatitis B surface
sion from Springer Nature © 2012.
antigen or omitting level of hepatitis B virus DNA does not affect
calculation of risk for hepatocellular carcinoma in patients without 5.12.2 Renée Turzanski Fortner and Rudolf Kaaks.
cirrhosis. Clin Gastroenterol Hepatol. 14(3):461–468.e2. https://doi.
org/10.1016/j.cgh.2015.10.033 PMID:26598229, Copyright 2016,
with permission from Elsevier.

5.19.1 Reproduced from Swerdlow SH, Campo E, Harris NL, Villarreal-Garza C, Alvarez-Gómez RM, Pérez-Plasencia C,
Jaffe ES, Pileri SA, Stein H, et al., editors (2017). WHO classifi- Herrera LA, Herzog J, Castillo D, et al. (2015). Significant clinical im-
cation of tumours of haematopoietic and lymphoid tissues. Lyon, pact of recurrent BRCA1 and BRCA2 mutations in Mexico. Cancer.
France: International Agency for Research on Cancer (WHO 121(3):372–8. https://doi.org/10.1002/cncr.29058 PMID:25236687;
Classification of Tumours series, revised 4th edition). Available Lolas Hamameh S, Renbaum P, Kamal L, Dweik D, Salahat M,
from: http://publications.iarc.fr/556. Jaraysa T, et al. (2017). Genomic analysis of inherited breast cancer
among Palestinian women: genetic heterogeneity and a founder mu-
6.3.1 Adapted with permission from Arbyn M, Xu L, Simoens C,
tation in TP53. Int J Cancer. 141(4):750–6. https://doi.org/10.1002/
Martin-Hirsch PP (2018). Prophylactic vaccination against hu-
ijc.30771 PMID:28486781; Peixoto A, Santos C, Pinheiro M, Pinto
man papillomaviruses to prevent cervical cancer and its precur-
P, Soares MJ, Rocha P, et al. (2011). International distribution and
sors. Cochrane Database Syst Rev. (5):CD009069. https://doi.
age estimation of the Portuguese BRCA2 c.156_157insAlu found-
org/10.1002/14651858.CD009069.pub3 PMID:29740819, John
er mutation. Breast Cancer Res Treat. 127(3):671–9. https://doi.
Wiley & Sons, Inc.
org/10.1007/s10549-010-1036-3 PMID:20652400; Nielsen SM,
6.4.1 & 6.4.2 Adapted from Cuzick J (2017). Preventive therapy for Rhodes L, Blanco I, Chung WK, Eng C, Maher ER, et al. (2016). Von
cancer. Lancet Oncol. 18(8):e472–e482. https://doi.org/10.1016/ Hippel-Lindau disease: genetics and role of genetic counseling in a
S1470-2045(17)30536-3 PMID:28759386, Copyright 2017, with multiple neoplasia syndrome. J Clin Oncol. 34(18):2172–81. https://
permission from Elsevier. doi.org/10.1200/JCO.2015.65.6140 PMID:27114602
6.4.3 Reproduced from Cuzick J, Thorat MA, Bosetti C, Brown 6.6.1 Raúl Murillo.
PH, Burn J, Cook NR, et al. (2015). Estimates of benefits and 6.8.1 Vincent J. Cogliano.
harms of prophylactic use of aspirin in the general population.
Ann Oncol. 26(1):47–57. https://doi.org/10.1093/annonc/mdu225 6.8.2 Vincent J. Cogliano. Compiled from Stockholm Convention:
PMID:25096604, by permission of Oxford University Press. http://chm.pops.int/TheConvention/ThePOPs/AllPOPs/tabid/
2509/Default.aspx; http://chm.pops.int/TheConvention/ThePOPs/
6.5.1 Patricia Ashton-Prolla. ChemicalsProposedforListing/tabid/2510/Default.aspx.
6.5.2 Compiled from Achatz MI, Zambetti GP (2016). The inherited 6.9.1 Reproduced from Hunter DJ, Reddy KS (2013).
p53 mutation in the Brazilian population. Cold Spring Harb Perspect Noncommunicable diseases. N Engl J Med. 369(14):1336–43.
Med. 6(12):a026195. https://doi.org/10.1101/cshperspect.a026195 https://doi.org/10.1056/NEJMra1109345 PMID:24088093, Copyright
PMID:27663983; Rebbeck TR, Friebel TM, Friedman E, Hamann 2013, Massachusetts Medical Society. Reprinted with permission
U, Huo D, Kwong A, et al.; EMBRACE; GEMO Study Collaborators; from Massachusetts Medical Society.
HEBON (2018). Mutational spectrum in a worldwide study of 29,700
families with BRCA1 or BRCA2 mutations. Hum Mutat. 39(5):593– P1.1–P1.4 WHO.
620. https://doi.org/10.1002/humu.23406 PMID:29446198; Laraqui pages 45–48 Reproduced from IARC Monographs on the
A, Uhrhammer N, Rhaffouli HE, Sekhsokh Y, Lahlou-Amine I, Identification of Carcinogenic Hazards to Humans, Agents
Bajjou T, et al. (2015). BRCA genetic screening in Middle Eastern Classified by the IARC Monographs: https://monographs.iarc.fr/
and North African: mutational spectrum and founder BRCA1 muta- agents-classified-by-the-iarc/.
tion (c.798_799delTT) in North African. Dis Markers. 2015:194293.
https://doi.org/10.1155/2015/194293 PMID:25814778; Seymour HJ,
Wainstein T, Macaulay S, Haw T, Krause A (2016). Breast cancer in
high-risk Afrikaner families: is BRCA founder mutation testing suffi-
cient? S Afr Med J. 106(3):264–7. https://doi.org/10.7196/SAMJ.2016.
v106i3.10285 PMID:26915939; Ponti G, Castellsagué E, Ruini C,
Percesepe A, Tomasi A (2015). Mismatch repair genes founder
mutations and cancer susceptibility in Lynch syndrome. Clin Genet.
87(6):507–16. https://doi.org/10.1111/cge.12529 PMID:25345868;
Text
page 250 Reproduced with permission from CSDH (2008).
Closing the gap in a generation: health equity through action on
the social determinants of health. Final Report of the Commission
on Social Determinants of Health. Geneva, Switzerland: World
Health Organization. Available from: https://www.who.int/social_
determinants/thecommission/finalreport/en/.
594
Subject index
A ALK 302, 307, 474
alkaline phosphatase 436
ACD 376
allergy 458, 473, 474
acetaldehyde 68, 69, 166, 338, 358
allogeneic haematopoietic stem cell transplantation
acinar adenocarcinoma 421 226
acoustic neuroma 88, 89, 560 Alternate Healthy Eating Index 97, 98
acrylamide 113 aluminium production 128
actinic keratoses 379, 380 AMACR 425
acute lymphoblastic leukaemia (ALL) 479–482 American Cancer Society 95, 99
acute myeloid leukaemia (AML) 53, 212, 478–483 American Institute for Cancer Research 68, 92–94,
acute promyelocytic leukaemia 481 101, 144, 145, 567
ADAM29 324 anal cancer 66, 513, 516
adefovir 364 anaplastic large cell lymphoma 474
adenomatous polyposis syndromes 349 anaplastic thyroid cancer 461, 464, 465
ADH 69, 166 anastrozole 391, 525
adiposity 102, 103, 106, 144, 351, 406, 566 androgen deprivation therapy 426
adoptive cell therapy 219 androgen insensitivity syndrome 434
adult T-cell leukaemia/lymphoma 66, 471, 479 androgen receptor 356, 425
aerodigestive cancers 69, 70, 72, 166 androgens 404, 406, 414, 425
aflatoxins 109–111, 160, 175, 209, 356, 358, 359, angiogenesis 216, 542
363, 567
angioimmunoblastic T-cell lymphoma 474
age at first birth 25, 31, 189, 193, 262, 382, 387, 389
anogenital cancers 63, 518
age at last birth 405
anogenital warts 64
age at menarche 25, 189, 193
antibiotics 185, 224–227, 480
age at menopause 193, 389
anti-Müllerian hormone 389
ageing 16, 149, 151
α-1-antitrypsin deficiency 356
Agency for Toxic Substances and Disease Registry
(ATSDR) 121 antiviral drugs 362, 364
AHRR 303 AP2-gamma 436
air pollution 20, 19, 40, 116–120, 299, 326, 330, 567, APC 177, 336, 346, 530
570 APOBEC 318, 413, 442
alcohol consumption 19, 20, 26, 40, 68–75, 166, APOBEC3 160
186, 238, 242, 254, 256, 261, 290, 313, 317, 320,
apoptosis 232, 346, 351
325, 330, 341, 355, 356, 358, 363, 368, 372, 407,
450, 509, 565, 566, 570 areca nut 52, 260–262, 317
alcohol-related cancers 68–75 ARID1A 324, 405, 413, 449
ALDH 69, 166 aristolactam–DNA adducts 441, 442

Aristolochia 111, 178, 440–442 biomarkers 103, 110–113, 201–204, 209, 210, 305,
aristolochic acid 109–111, 160, 175, 178, 179, 306, 318, 320, 321, 328, 329, 338, 358, 361, 365,
440–444, 449 370, 387, 395, 398, 400, 416, 425, 426, 441–444,
449, 451, 544, 545, 550–554
aromatase inhibitors 391, 525
biomass combustion 116, 119, 326, 330, 567
arsenic 116, 120, 122, 209, 291, 379, 440, 442, 443,
567 Birt–Hogg–Dubé syndrome 448
aryl hydrocarbon receptor (AhR) 229, 232 bisphosphonates 408, 525
arzoxifene 525 blackfoot disease 120, 440
asbestos 116, 119–121, 127, 129, 134, 135, 231, 242, bladder cancer 53, 66, 101, 116, 120, 121, 123, 129,
299 134, 202, 254, 439, 440–444, 567
ASIP 376, 379 bladder urothelial carcinoma 439, 441, 442
aspirin 27, 94, 167, 186, 218, 330, 352, 372, 407–409, BNC2 379
415, 418, 525–528 body mass index (BMI) 41, 106, 107, 204, 291, 387,
asthma 458, 567, 568 407, 415, 418, 450, 558, 559, 566
asymmetric division 150 body weight 106, 107, 144, 145, 383, 418, 450, 508,
511
ataxia–telangiectasia 174
Bowen disease 379
ATM 318, 335, 369, 370, 376
brachytherapy 249
atrophic gastritis 341
BRAF 346–348, 376, 377, 413, 465
atypical proliferative serous tumours 412
brain cancer 454–459
autoimmune diseases 471–475
BRCA1 154–156, 174, 177, 369, 370, 383, 387, 391,
autophagy 232 412, 413, 417, 418
azacitidine 211 BRCA2 156, 174, 177, 300, 335, 369, 370, 383, 387,
412, 413, 417, 418, 422
breast cancer 17, 25–27, 36, 40, 41, 69, 70, 72, 75,
B 92–97, 101, 103–106, 119, 123, 134, 137, 138,
140–142, 154, 157, 158, 164, 167, 174, 177, 204,
bacteria 61–63, 185–187, 221–227, 333, 337, 338, 217, 225, 240, 246, 248–250, 254, 255, 259, 260,
351 262, 263, 266, 267, 271, 273, 274, 279–281, 283,
285, 290, 291, 382–391, 412, 522, 523, 525–528,
Bacteroides fragilis 223 552–554, 558
Balkan endemic nephropathy 111, 178, 440 breast cancer screening 25, 26, 262, 266, 267, 271,
BAP1 376, 448, 449 273, 274, 279, 281, 285, 286, 292, 388, 389, 540,
541, 545, 546
Barrett oesophagus 328, 330, 541
breast density 384, 387–389
basal cell carcinoma (BCC) 378–380
breastfeeding 66, 122, 189, 385, 387, 405, 408
base excision repair 171, 173, 175, 176
breast implants 474
BCL2 472, 473
1-bromopropane 233
BCL6 472, 473
Burkitt lymphoma 65, 471, 474
BCR-ABL1 159
1,3-butadiene 116, 129
behavioural support 490–495
behaviour change 259, 408, 490–492, 499–511, 568
benzene 53, 116, 123, 127–129, 299, 472, 475, 479
C
benzo[a]pyrene 116, 129, 175
betel quid 317 CA125 416
bidis 51, 52, 56, 57, 261 CA19-9 338
bile ducts 66 CADM1 400
596
cadmium 231, 291 cervical cancer vaccines 517

calcitonin 461, 465 cervical intraepithelial neoplasia (CIN) 395, 400,
calcium 28, 93, 94, 99, 352, 526 517, 518, 541
canakinumab 216 cervical squamous cell carcinoma 395, 397
Cancer Genome Atlas 158, 301, 334, 405, 439, 455, checkpoint blockade 218, 219
465 CHEK2 449, 465
cancer hazard identification 230, 232–234 chemopreventive agents 352, 408, 418, 427
cancer stem cells 150, 152 chemotherapeutic agents 225
cannabis smoking 300 chewing tobacco 317, 320
cannabis use 435 chickenpox 458
captafol 124 childhood cancer 122
carbon monoxide 53, 567 chimeric antigen receptor T cells 218
carcinogen identification 558, 561, 562 Chinese herbs nephropathy 440
carcinogen mechanisms 229–234, 561 Chlamydia trachomatis 415, 418
carcinogens in alcoholic beverages 68
chlordane 435
carcinogens in natural products 109–111
chlorination by-products 120
cardia stomach cancer 30, 40
cholangiocarcinoma 66, 355–359
cardiovascular disease 16, 17, 19, 166, 437, 489,
choriocarcinoma 431
526, 528, 565–570
CHRNA5 301
Carney complex 464
α-carotene 92 chromium 116
β-carotene 92, 526 chromophobe renal cell carcinoma 447, 450
carotenoids 92, 384 chromosomal instability 208, 336, 346, 398
cascade testing 532, 535, 536 chromothripsis 160, 177, 449
CASP8 319 chronic gastritis 338
β-catenin 356, 405, 413, 465 chronic kidney disease 440–442, 450
CCDC26 456 chronic liver infections 355–358, 360–364
CCND1 356 chronic lymphocytic leukaemia (CLL) 478, 479, 482
CCNE1 413 chronic myeloid leukaemia (CML) 159, 481
CD79B 472 chronic obstructive pulmonary disease (COPD) 300,
552, 567, 570
CDH1 335, 336, 338, 341, 383
chronic pancreatitis 368, 369
CDK4 376
CDKN2A 318, 324, 356, 369–371, 376, 377, 380, ciclosporin 232
456 cigarettes 50–53, 56–58, 255, 261, 299, 300, 488,
celecoxib 526 491, 493, 495, 499, 500, 502
cell death 232 cigar smoking 51, 52, 54, 56
cellular senescence 149, 150 circulating cell-free DNA (cfDNA) 553, 554
cervical adenocarcinoma 394, 395, 397 circulating cell-free tumour DNA (ctDNA) 550–554
cervical cancer 19, 30, 31, 36, 41–43, 53, 63, 64, circulating tumour cells (CTCs) 550, 552, 554
66, 106, 137, 138, 240–242, 246–250, 254, 259, cirrhosis 186, 202, 357, 358, 360, 362, 363, 365,
260, 263, 266, 267, 270, 273, 291, 394–400, 513, 514, 570
514, 516–519, 525, 526, 568
cisplatin 225, 434, 444
cervical cancer screening 31, 63, 64, 263, 266, 267,
270, 271, 274, 279, 281, 292, 398–400, 500, 540, clear cell renal cell carcinoma 447, 449
543, 545, 546, 568, 570 clinical breast examination 262

clomiphene citrate 142, 406 Crohn disease 351

clonal evolution 541, 553 cryptorchidism 196, 430, 434, 436
Clonorchis sinensis 61, 66, 359, 360 CTNNA1 335
Clostridium 337 CTNNB1 356, 405, 413, 465
coal combustion 119, 326, 567 curcumin 187, 526
coal tar 129 cutaneous melanoma 374–377, 559
coffee 97, 204, 407 cyclin D1 449, 473
colitis 186 cyclin E1 413
collecting duct renal cell carcinoma 447 cyclobutane pyrimidine dimers 80, 81
colon cancer 93, 95, 101–105, 141, 167, 173–175, 177, cyclooxygenase 2 (COX-2) 81, 181, 184, 186, 352,
181, 347, 348, 351, 352, 551 387
colonization 288–290, 292 cyclophosphamide 225
colonoscopy 348, 501 CYFRA 21.1 398
colorectal cancer 17, 27, 28, 36, 40, 41, 53, 68–70, CYP19A1 404
72, 73, 75, 93–97, 99, 106, 112, 121, 137, 138, cystoscopy 442, 443
141, 142, 167, 173–175, 177, 179, 186, 202–204,
cytokines 181, 183, 184, 186, 216, 218, 368, 387
218, 223, 225, 227, 240, 250, 254, 255, 259, 260,
263, 266, 267, 270, 280, 281, 282, 290, 344–352, cytology screening 266, 270, 399, 400, 543, 545
525–527, 558, 567 cytotoxic T-lymphocyte-associated protein 4 (CTLA-4)
colorectal cancer screening 27, 266, 267, 270, 273, 218, 226
279, 281, 282, 292, 347, 349, 351, 352, 501, 540,
543, 546
colposcopy 400
D
combined estrogen–progestogen contraceptives
137, 138, 140 dairy products 28, 93, 94, 350
combined estrogen–progestogen menopausal DAZL 435
therapy 141, 406, 414
decitabine 211
combustion emissions 116, 119, 567
demographic transition 16, 17, 346
Commission on Global Governance for Health 250 denosumab 391
Commission on Social Determinants of Health 250 deoxycholic acid 224
complement-dependent cytotoxicity 216 diabetes 16, 95, 105, 110, 112, 187, 203, 285, 351,
computed tomography (CT) 25, 84–87, 304, 305, 360, 368–372, 407, 450, 464, 489, 507, 509, 565,
462, 466 566, 569, 570
conjunctival cancer 66 diazinon 124
construction industry 129, 134 DICER1 464
contemporary hormonal contraceptives 140, 406 4,4′-dichlorodiphenyltrichloroethane (DDT) 122, 124,
435
Continuous Update Project 68, 69, 144, 145
dieldrin 124
cortical inclusion cysts 412
diesel engine exhaust 116, 127, 129, 134, 299
costs of cancer 34, 248
diet 17, 19, 20, 25, 27, 40, 92–99, 144, 145, 167,
Cowden syndrome 404, 464
185, 186, 221, 223, 225, 227, 290, 326, 328, 330,
CpG island methylator phenotype (CIMP) 211, 347, 333–335, 340, 368, 372, 384, 385, 389, 407, 464,
348 506–508, 511, 565–567
CPS1 356 dietary carcinogens 109–113
C-reactive protein 186 dietary exposures 92–113, 204, 389
creosotes 129 dietary fibre 92, 93, 97, 99, 227
598
dietary patterns 92–98, 254, 262, 348, 350, 507, electronic nicotine delivery systems (ENDS) 51–53,
508, 511 57, 58, 300, 493–495
diet quality 93, 98, 507 electronic non-nicotine delivery systems (ENNDS)
493, 495
diffuse large B-cell lymphoma (DLBCL) 471–473
electronic waste (e-waste) 122, 123
dioxins 115, 122–124
electrophilicity 231
direct-acting antiviral agents 362, 365, 570
ellagic acid 186
disability-adjusted life years (DALYs) 41, 70, 72, 73,
75, 348, 566–568 embryonal carcinoma 431, 436
discrimination 276, 278, 289 emphysema 300
disinfection by-products 120 endocrine disrupters 115, 121–123, 209, 384, 434,
435
disorders of sexual development 431, 434, 436
endogenous hormones 102, 103, 106, 191, 193,
disparities 17, 20, 238–243, 246–250, 253, 254, 196, 387, 389
258, 263, 276–279, 281–283, 285, 291, 352, 398,
399, 425, 426, 463, 499, 504, 530 endometrial cancer 40, 97, 101, 104–106, 137, 138,
140–142, 174, 191–193, 263, 291, 403–409, 522,
d -limonene 229 523, 525, 558
DMRT1 431, 435 endometriosis 193, 407, 412, 413, 415
DNA damage 78, 80–82, 88, 170–179, 440, 442 endosalpingiosis 412
DNA glycosylases 173, 175–177 endoscopy 328, 330, 338, 341, 348, 351, 371
DNA methylation 207–213, 436, 459 entecavir 362, 364
DNA photolesions 78–82 environmental degradation 291, 292
DNA repair 160, 165, 170–179, 207, 231, 347, 370, environmental pollution 115–124, 563
383, 441
Environmental Protection Agency (EPA) 120, 121,
dose–response assessment 561 229, 561, 562, 564
Down syndrome 480 EPAS1 449
drinking-water contaminants 120, 121, 440, 442, EPCAM 404
443, 567 epidemiological transition 16, 17, 19, 20, 34, 40, 43,
ductal carcinoma in situ (DCIS) 391 344
dutasteride 527 epidermal growth factor receptor (EGFR) 348
dysbiosis 221, 227, 335, 338, 351 epigenetic alterations 184, 206–213, 440, 442, 541
epigenetics 150, 151, 206–213, 302, 303
epigenetic therapy 211, 212
E epigenomics 206–209, 306
epithelial–mesenchymal transition (EMT) 149, 336,
early detection 540–547, 550–554 552
early-life exposures 209 Epstein–Barr virus (EBV) 61, 65, 317, 318, 336–338,
E-cadherin 94, 335 340, 469, 474
e-cigarettes 51–53, 57, 58, 262, 300, 493–495 ERBB2 336, 380, 413, 440
eczema 458, 474 ERBB3 336
education campaigns 557, 559 ERG 425
education level 239, 241, 253, 255, 258, 260, 262, erionite 120
271, 279, 283 erlotinib 554
EGFR 301–304, 307, 336, 397, 456, 457, 554 Escherichia coli 223
EIF1AX 465 estradiol 194, 196, 387, 388
electromagnetic fields 88, 89, 455 estrogen-only menopausal therapy 141, 403, 406, 414

estrogen–progestogen contraceptives 137, 138, 140 fast foods 96

estrogen–progestogen menopausal therapy 141, fat intake 92, 93, 95, 186
406, 414 fatty acids 202
estrogen receptor (ER) 382–385, 387, 389, 391 fermentation 93
estrogens 191, 193, 404–407, 414 fertility drugs 142, 406
ethanol 68, 69, 166, 358 FGFR3 431, 439–441
ethylene dibromide 124 fibres 120, 121
ethylene oxide 229 fibroids 407
etoposide 229 finasteride 196, 427, 527
European Chemicals Agency 560 fine particles 116, 118, 567
European Code Against Cancer 295, 557 5-fluorouracil 348
European Council recommendations on screening 266 fluoro-edenite 120, 121
European Prospective Investigation into Cancer and fluxomics 202
Nutrition (EPIC) 202, 203
folate 94, 96
European Study of Cohorts for Air Pollution Effects
folic acid 96
(ESCAPE) 116
follicular lymphoma 471, 473
evaluation of prevention strategies 499, 503
follicular thyroid cancer 461, 464, 465
evidence-based preventive interventions 499,
501–503, 559 food 92–98, 109–113, 144, 145, 186, 290, 507, 508,
511, 559
exemestane 391, 525
Food and Drug Administration (FDA) 554, 561
exercise 102, 103, 408, 509
food contaminants 109–111, 115, 121, 122, 124, 440,
expenditure on health 247, 248 442, 567
Expert Reports 144, 145 formaldehyde 119
exposome 115, 119, 120, 124, 204, 562 founder mutations 530, 534, 535
Exposome-Explorer 204 FOXP1 425
extracellular vesicles 550, 553, 554 Framework Convention on Tobacco Control 54, 56,
extranodal marginal zone lymphoma of mucosa- 242, 491, 499, 563, 564, 569
associated lymphoid tissue 473 fruits 92, 93, 95, 97–99, 326, 350, 507, 526, 559
EZH2 207, 212, 303 fumonisins 110–112
Fusobacterium nucleatum 223, 338
F
faecal immunochemical test (FIT) 270, 348, 351, 352
G
faecal occult blood test (FOBT) 267, 270, 348, 351, gall bladder cancer 40, 106, 558
352, 543, 544
gastric adenocarcinoma and proximal polyposis of
fallopian tube 412 the stomach 335
FAM135B 324 gastric atrophy 335, 338, 341
familial adenomatous polyposis 173, 336, 349, 464, gastric cancer 61–63, 65, 101, 106, 177, 221, 227,
530 333–341, 525–527, 558
familial medullary thyroid cancer 465, 466 gastric cancer screening 341
familial non-medullary thyroid cancer 464 gastric lymphoma 61
familial pancreatic cancer 369 gastric microbiome 335, 337, 338
Fanconi anaemia 174, 324, 383, 480 gastrin-17 341
600
gastritis 185, 335, 337, 338 glyphosate 124

gastrointestinal tract cancers 222–224 GNAQ 377
gastro-oesophageal reflux disease 328 Golestan Cohort Study 327, 329
gemcitabine 225 GOLM1 425
Gender Inequality Index 247 gonadal dysgenesis 434
gene–environment interactions 164–169, 209, 562 gonadoblastoma 431
gene expression 206–212 gonadotropin-releasing hormone agonists and
genetic counselling 535, 537 antagonists 142
genetic/genomic cancer risk assessment (GCRA) gonadotropins 142
530, 532, 534, 535, 537 gonocytes 434, 436
genetic mosaicism 161 Gorlin syndrome 379
genetic risk score 349 goserelin 525
genetic susceptibility 154–160, 164–169, 300, 301, GSTM1 441
318, 324, 328, 335, 336, 349, 355, 369, 376, 379,
382, 383, 391, 396, 397, 404, 412, 422, 425, 430, GSTP1 425
435, 441, 448, 449, 456, 457, 464, 471, 480, 483, guaiac faecal occult blood test (gFOBT) 270, 348,
530, 531, 532, 535, 537 351, 352
genetic testing 530, 532, 534–537 gut microbiota 186, 187, 221–227, 351
genital malformations 430, 434 gynaecomastia 196
genital warts 64, 517
genome-wide association studies (GWAS) 156–158,
167, 168, 300, 301, 349, 376, 379, 383, 388, 396,
404, 423, 435, 441, 449, 456, 457, 471
H
genomic instability 102, 103, 149, 151, 170, 171, 174, haematological malignancies 474, 477, 478, 553
175, 177, 370, 372, 440, 541
Haematological Malignancy Research Network
genomics 155–160, 530, 532, 534, 535, 537 (HMRN) 478, 482
genotoxicity 229, 231 haematopoietic stem cell transplantation 483
geographical disparities 247, 249, 253–255, 258, haemochromatosis 355
259, 276, 279
hair dyes 472, 480
germ cell neoplasia in situ (GCNIS) 430, 431,
434–436 hairy cell leukaemia 471, 478
germ cell tumours 411, 430–432, 434–436 hallmarks of cancer 148, 542
germline mutations 154, 156, 171, 173–175, 177, hazard identification 558, 561
300, 349, 376, 379, 412, 413, 464, 465, 530, 532, hazardous waste 121, 123
535–537
head and neck cancers 310, 311, 313, 316–321
glioblastoma 454, 456, 457
health disparities 17, 20, 238–243, 246–250, 253–255,
glioma 88, 89, 175, 454–459, 560 258, 263, 276–283, 285, 291, 352, 426, 499, 504
global burden of cancer 16, 17, 19–21, 23–32, 34, health inequalities 17, 20, 238–243, 246–250, 253,
35, 40, 43 254, 258, 263, 271, 273, 274, 289–291, 500, 507
Global Burden of Disease Study 55–58, 72, 74, 116, health insurance coverage 276, 277, 279, 281, 282,
118, 119, 355, 567, 568 305, 463
Globally Harmonized System of Classification and health literacy 259, 276, 279, 283, 502, 568
Labelling of Chemicals 560
heated tobacco products (HTPs) 52, 53, 493, 494
glycaemic load 407
height 368, 372, 425, 450
glycine 202
Helicobacter pylori 29, 30, 41, 61–63, 185, 221, 222,
glycine decarboxylase 202 227, 231, 254, 255, 291, 328, 333–335, 337, 338,
glycolysis 202 341, 473

hepatitis B virus (HBV) 61, 64, 65, 110, 111, 175, human papillomavirus (HPV) 30, 31, 41, 43, 61, 63,
254, 255, 291, 355–365, 500, 513–516, 570 64, 70, 194, 249, 263, 270, 291, 313, 316–321,
hepatitis B virus (HBV) vaccination 255, 291, 362, 379, 394–398, 400, 500, 513–519, 526
364, 500, 513–515, 570 human papillomavirus (HPV) screening 242, 263,
hepatitis C virus (HCV) 61, 64, 65, 255, 355–358, 270, 394, 399, 400, 500, 544, 546
361–365, 472–475, 513, 514, 516, 570 human papillomavirus (HPV) vaccination 31, 242,
hepatocellular carcinoma 64, 65, 97, 175, 179, 202, 263, 279, 320, 394, 399, 500, 513, 517–519, 526,
203, 224, 355–365, 514 570
hepatocellular carcinoma screening 363 Human Poverty Index 249

HER2 300, 302 human T-cell lymphotropic virus type 1 (HTLV-1) 61,
66, 472, 479
herbal products 111, 178, 179, 440, 442, 443
hydrochlorothiazide 80, 82
hereditary cancer syndromes 530, 531, 534–537
8-hydroxydeoxyguanosine (8-OHdG) 80
hereditary diffuse gastric cancer 335, 336, 341
8-hydroxyguanine (8-oxoG) 80, 81, 173, 175–177
hereditary non-polyposis colorectal cancer (HNPCC)
173, 175, 177, 349, 412 hypertension 407, 450
hereditary pancreatitis 369 hypogonadism 437
hereditary papillary renal carcinoma 448 hypopharyngeal cancer 53, 54, 310, 311, 317
hereditary prostate cancer 422 hypoxia-inducible factor 2 alpha (HIF-2α) 449
heterocyclic amines (HCAs) 327 hysterectomy 193, 403, 408, 418
heterocyclic aromatic amines (HAAs) 113, 340
β-hexachlorocyclohexane 123
HHIP 356 I
histone deacetylases (HDACs) 207, 209, 212
IARC Handbooks of Cancer Prevention 106, 236
histone modifications 207, 208, 211, 303, 436
IARC Monographs 45, 53, 54, 61, 68, 73, 79, 112,
HNF1B 404 116, 119–121, 124, 128, 129, 233, 455, 559–561,
Hodgkin lymphoma 65, 66, 218 563, 567
homologous recombination 171, 174, 177, 179 ibuprofen 186, 526
hormonal contraceptives 137, 138, 140, 141, 189, IDH 455
192–194 IDH1 356, 376
hormonal factors 25, 387, 388, 405–408, 414 IDH2 356
hormone therapy 141, 142, 191, 194, 232, 385, 389, IFNL3 357, 362
403, 406–408, 414
IGH 473, 474
hot beverages 317, 326, 329
imiquimod 82
HOXB13 422, 423
immigrants 274
HPGDS 435
immortalization 232
HRAS 319, 380, 431, 441
immune checkpoint inhibitors 179, 218, 219, 226,
human chorionic gonadotropin 142
319, 340, 444
Human Development Index (HDI) 16–19, 24–27, 31,
immune function 81, 82, 181, 182, 186, 215–219,
32, 35, 36, 40–43, 73, 239, 247, 249, 250, 258, 263,
396, 471, 472, 474
344–346, 398, 399, 534
immune response 455, 458, 459, 542
human epidermal growth factor receptor 2 (HER2)
324, 336, 340, 382, 391 immunomethylomics 459
human immunodeficiency virus (HIV) 61, 63, 65, 66, immunosuppression 78, 81, 82, 216–219, 231, 376,
70, 231, 248, 249, 291, 379, 398, 472, 474 379, 398, 459, 468, 472, 475
human leukocyte antigen (HLA) 158, 356, 397, 471 immunosurveillance 81, 82, 227, 458
602
immunotherapy 177, 179, 217–219, 225, 226, 307, J

348, 356, 444
IMP3 449 JAK2 336
implementation science 501–504 Janus kinases (JAKs) 183
incidental findings 532, 535
Indigenous populations 122, 288–292
indoor air pollution 119, 299, 326, 330, 567, 570 K
industrial hygiene 131, 133, 135
Kaposi sarcoma 40, 65, 66, 246, 249, 291
inequalities 17, 20, 31, 42, 238–243, 246–250, 253,
Kaposi sarcoma-associated herpesvirus (KSHV) 61, 65
254, 258, 260–263, 271, 273, 274, 276–279, 281,
289–291, 500, 507 kataegis 160
inequities 20, 246, 249, 250, 253, 263, 266, 270, KDM5C 449
276, 278, 281, 282, 352, 394 KDM6A 449
infections 16, 17, 19, 29–31, 36, 40, 41, 43, 61–66, keratinocyte cancers 378–380
181, 185, 240–242, 246, 248, 249, 254, 255, 291,
333–335, 337, 338, 355–357, 359, 360, 362–364, Keshan disease 328
458, 459, 472–475, 479, 513–519, 570 key characteristics of carcinogens 165, 167,
infectious agents 61–66 229–233
infiltrating pancreatic ductal adenocarcinoma 367 Ki-67 387, 390, 400, 425
inflammation 62, 66, 78, 80–82, 102, 103, 105, 106, kidney cancer 40, 53, 69, 101, 102, 106, 121, 254,
181–187, 215, 216, 219, 221–225, 231, 300, 329, 447–451, 558, 563
334, 338, 351, 368, 387, 473, 514, 542 KIM-1 451
inflammatory bowel disease 186, 187 KIT 376, 377, 435
insecticides 124, 473 KITLG 157, 435
insulin 103, 105, 106, 203 Klinefelter syndrome 196
insulin-like growth factor 1 (IGF-1) 387 koningic acid 202
interferon 215, 216, 356, 362, 364, 365
KRAS 301, 302, 304, 346, 348, 370, 397, 405, 413,
interleukins 78, 82, 181, 183, 184, 186, 215–218, 387 417, 435
International Atomic Energy Agency (IAEA) 249 kreteks 51
International Cancer Genome Consortium (ICGC) 158
International Childhood Cancer Cohort Consortium
(I4C) 122
L
International Federation of Gynecology and
Obstetrics (FIGO) 398, 400 Lactobacillus 338
International Labour Organization 134 lamivudine 362, 364
International Lung Cancer Consortium 300 laryngeal cancer 53, 54, 58, 69, 70, 72, 166, 241,
International Nuclear Workers Study (INWORKS) 86 310, 311, 313, 316, 317
intraductal papillary mucinous neoplasms 370, 371 lasofoxifene 525
intrahepatic bile duct carcinoma 178 lead compounds 129
intrahepatic cholangiocarcinoma 355–357, 359 legislative initiatives 557–564
involuntary smoking 57 letrozole 525
iodine 461, 463–466 leukaemia 85–87, 174, 477–483
ionizing radiation 84–87, 122, 129, 170, 174, 379, levonorgestrel-releasing intrauterine system (LNG-IUS)
459, 463, 464, 466, 479, 483 140, 406, 408
IRF4 379 Leydig cell tumours 430
irinotecan 225 Li–Fraumeni syndrome 300, 336, 530, 536, 537

LIN28 436 mammography 266, 267, 271, 388, 546

lindane 124, 473, 475 mantle cell lymphoma 471, 473, 474
linkage disequilibrium 155, 157 manufacturing 134
lipopolysaccharide 183 marginal zone lymphoma (MZL) 471, 473
liquid biopsy 328, 416, 417, 550–554 mass spectrometry (MS) 200–204
liver cancer 19, 40, 41, 53, 64, 68, 70, 72, 73, 97, mastectomy 391, 537
106, 110–112, 137, 175, 178, 181, 186, 203, 223,
MC1R 376, 379
227, 240, 241, 246, 250, 252, 254, 255, 290, 291,
355–365, 500, 513, 514, 567, 570 meat 93, 97–99, 112, 113, 122, 340, 350, 455, 507, 559
liver cancer screening 363 mechanisms of carcinogenesis 229–234, 561
liver cancer vaccine 362, 514 mediastinal large B-cell lymphoma 472
liver flukes 66, 356, 359, 360, 363 Medicaid 279, 281
lobular involution 191 medical radiation 84, 85, 87, 464, 466
low-dose computed tomography (LDCT) 304–307 Medicare 305
LPP 471 Mediterranean diet 97, 98, 384
LSD1 207, 212 medullary thyroid cancer 461, 463–465
lung adenocarcinoma 24, 299, 301, 302, 304, 551 melanoma 79–81, 160, 240, 374–378, 502, 541, 559
lung cancer 24, 25, 40, 53, 54, 58, 69, 87, 101, Mendelian randomization 69, 166, 383
102, 105, 116, 118–120, 129, 134, 160, 167, 216, meningioma 40, 88, 106, 454, 455, 457, 558
225, 241, 248–250, 252, 255, 259, 260, 279, 289,
299–307, 488, 499, 525–527, 551–554, 567, 568 menopausal hormone therapy 25, 141, 191, 194,
385, 403, 406–408, 414
lung cancer screening 25, 304–306, 540, 541, 546
menstrual factors 189, 191
lung squamous cell carcinoma 24, 299, 301, 302, 551
mesothelioma 119, 120, 121, 134
lycopene 526
messenger RNA (mRNA) 207, 553
lymphoid malignancies 468, 469, 471, 473, 474
metabolic change 200–204
lymphoma 468, 471–474
metabolic syndrome 407, 437
Lynch syndrome 173, 175, 336, 349, 404, 408, 412,
422, 441 metabolism 200–204
metabolomics 200–204, 389
metals 115, 120–123, 299
M metal working 134
metastasis 215–219
macrophages 181, 184, 215–219, 387
metformin 369, 372, 390, 407–409
magnetic resonance cholangiopancreatography
(MRCP) 371 methylation 207–212, 302–304, 442, 443
magnetic resonance imaging (MRI) 371, 462 MGMT 347
maize 110–112 microarrays 156, 157
major histocompatibility complex (MHC) 471 microbe-associated molecular patterns (MAMPs)

221, 223, 225
MAL 400
microbiome 99, 145, 221–227, 337, 338, 351, 370
malaria 65, 474
microbiome-wide association studies (MWAS)
malathion 124 223–225, 227
male breast cancer 196 microRNA (miRNA) 184, 207, 306, 338, 436, 550, 553
male infertility 430, 434, 436 microsatellite instability 173, 179, 336, 340, 347,
MALT lymphoma 473 348, 405, 413
mammographic density 383, 388, 389 mineral oils 121, 129, 134
604
minimal residual disease 550–552 naswar 52

mining 87, 134 NAT2 441
minor allele frequency (MAF) 155, 157, 164 National Breast and Cervical Cancer Early Detection
mismatch repair 171, 173, 175, 177, 179, 336, 340, Program (NBCCEDP) 281
347, 404, 405, 413, 422, 423 national cancer control programmes 571
MITF 376, 449 National Comprehensive Cancer Network 536
mitogen-activated protein kinase (MAPK) 186, 346, National Lung Screening Trial 304–306
348, 465
National Toxicology Program (NTP) 229, 233, 234, 561
MLH1 173, 175, 336, 347, 348, 369, 404, 413
natural killer cells 184, 215, 218, 219, 468, 472
mobile phones 88, 89, 455, 456, 560
naturally occurring chemical carcinogens 110–112
monoclonal antibodies 482
natural products 109–112, 187
monoclonal B-cell lymphocytosis (MBL) 478, 479
necrosis 232
mononucleosis 65
MPOWER 54, 58, 499 nervous system tumours 454–456
MSH2 173, 175, 336, 369, 404, 413 neuroblastoma 157, 177, 211
MSH6 336, 369, 404, 413 next-generation sequencing 154–158, 161, 329,
370, 383, 472, 481, 535, 536, 541, 551, 553
MTHFR 69
NF1 376, 377
mucinous cystic neoplasms 355, 370
NFE2L2 319, 324
multigene panel testing 531, 535–537
NF-κB 81, 182–185, 187, 387, 472, 473
multiple endocrine neoplasia (MEN) 461, 464–466
multiple myeloma 40, 106, 212, 468, 558 nickel compounds 128
muscle-invasive bladder cancer 439, 442 nicotine 51–53, 58, 488, 491, 493–495
mutational signatures 80, 174–179, 348, 375, 397, nicotine replacement therapy (NRT) 491, 494, 495, 569
413, 441, 442, 449 Nijmegen breakage syndrome 174
mutator phenotype 175, 177, 179 nitrates 120, 121
MUTYH 173, 175–177 nitrative stress 231
MUTYH-associated polyposis (MAP) 177, 349 nitrites 464
MYC 216, 356, 404, 425, 471, 472, 474
nitrosamines 327, 328, 330
MYCN 211, 380
NKX3-1 425
mycosis fungoides 469
N-nitroso compounds (NOCs) 112, 338, 340
mycotoxins 110–112
non-adenomatous polyposis syndromes 349
MYD88 472
non-alcoholic fatty liver disease 360
myelodysplastic syndrome (MDS) 211, 477–480
non-alcoholic steatohepatitis 360
myeloid-derived suppressor cells (MDSCs) 216, 218
non-cardia stomach cancer 29, 30
myeloproliferative neoplasms (MPNs) 477–479, 481,
482 non-coding RNA (ncRNA) 207, 209, 306, 338, 550, 553
noncommunicable diseases (NCDs) 16–21, 34,
489–491, 495, 565–571
N non-Hodgkin lymphoma (NHL) 61, 65, 66, 123, 249,

468, 469, 471–475, 478
naevi 376, 377, 541 non-homologous end joining 171
NANOG 436 non-ionizing radiation 88, 89, 170, 480
nasopharyngeal cancer 53, 65, 310, 311, 317, 318 non-melanoma skin cancer 79, 80, 82, 378–380

non-muscle-invasive bladder cancer 439 oesophageal cancer screening 328

non-piped water 326, 328, 330 oesophageal squamous cell carcinoma 323–330
non-seminomatous tumours 430, 431, 435, 436 OGG1 175–177
non-small cell lung carcinoma (NSCLC) 302, 303, oil drilling 291
307, 551, 553, 554
oncogenes 158, 159, 177, 182, 211, 212, 216, 318,
non-steroidal anti-inflammatory drugs (NSAIDs) 167, 336, 346, 348, 356, 370, 398, 400, 413, 441, 464,
186, 187, 330, 407–409, 525, 526 465
NOTCH1 160, 324, 380 oophorectomy 391, 418
NOTCH2 473 opioid analgesics 570
NRAS 376, 377, 431 Opisthorchis viverrini 61, 66, 359, 360
NTHL1 173, 175 opium 326, 327
nuclear accidents 87, 463, 466 oral cancer 53, 54, 57, 68, 70, 72–74, 102, 106, 241,
nuclear magnetic resonance (NMR) spectroscopy 259–262, 290, 310, 311, 316, 317
200, 203 oral cancer screening 262, 319, 320
nuclear testing 292 oral contraceptives 31, 137, 138, 140, 141, 192–194,
nucleoside/nucleotide analogues 362 254, 398, 405, 406, 408, 411, 413, 414, 417, 418, 525
nucleosome positioning 206 orchidectomy 436, 437
nucleotide excision repair 170, 171, 174, 175 organ transplantation 379, 468, 472–474
nutrient patterns 350, 351 oropharyngeal cancer 53, 54, 63, 68, 310, 311, 313,
316–318, 320, 513, 516
nutrient supply 232
osimertinib 554
nutrition 92–99, 144, 145, 227, 290, 499, 503, 504,
508, 509, 511, 559 outdoor air pollution 116, 118, 119, 299, 567, 570
nutrition transition 92, 97, 98 ovarian cancer 36, 40, 102, 106, 137, 138, 140–142,
154, 174, 177, 193, 194, 202, 260, 263, 391,
411–418, 525, 558
ovarian cancer screening 416
O
ovary-stimulating drugs 142
O -methylguanine-DNA methyltransferase (MGMT)
6 overdiagnosis 240, 243, 416, 451, 462, 541, 542,
171, 175, 211 546, 547
obesity 19, 26–28, 30, 40, 41, 93, 97, 106, 107, 181, overweight 19, 27, 30, 93, 97, 106, 107, 258, 260, 262,
185, 187, 191, 254, 256, 258, 260, 262, 263, 279, 263, 291, 351, 408, 450, 464, 508, 511, 558, 559
280, 285, 291, 328, 351, 356, 358, 360, 363, 368, oxaliplatin 225
372, 382, 385, 387, 403, 405, 407–409, 415, 425,
450, 455, 464, 501, 508, 511, 558, 559, 566, 569 oxidative stress 78, 80, 102, 103, 231, 440
OCA2 379
occupational carcinogens 119, 124, 127–136, 560, 561
occupational exposures 78, 79, 86, 87, 120, 124,
P
127–136, 239, 242, 440, 450, 560, 561
p16 347
Occupational Safety and Health Administration 560
p27 387
OCT4 436
paan masala 260–262
ocular melanoma 559
PALB2 335, 369, 370
oesophageal adenocarcinoma 40, 101, 323, 324,
328, 330, 541 pancreatic cancer 40, 53, 54, 69, 102, 106, 112, 177,
203, 217, 224, 225, 367–372, 558
oesophageal cancer 40, 53, 54, 68, 70, 72, 74,
101, 106, 111, 166, 241, 250, 252, 254, 255, 263, pancreatic intraepithelial neoplasia 370
323–330, 525–527, 558 pancreatitis 368, 369
606
papillary renal cell carcinoma 447, 450 pickled vegetables 97

papillary thyroid cancer 461–466 pigmentation 375, 376, 379
Pap smear 266, 267, 417 PIK3C 346
parity 25, 31, 191, 193, 262, 385, 387, 389, 398, PIK3CA 318, 319, 324, 336, 380, 397, 405, 413, 440
405, 411, 413
pipe smoking 51, 53, 54, 56
PARP1 376
PLA2G6 376
α-particles 84
plant foods 92, 93, 96, 98, 507
β-particles 84
plants 111, 440–442
particulate matter (PM) 115, 116, 118, 299, 360, 567,
570 plasma cell myeloma 468
patient navigation 281, 283, 284 Plasmodium falciparum 474
Patient Protection and Affordable Care Act 249, pleuropulmonary blastoma syndrome 464
279, 281, 282 PMS2 336, 369, 404, 413
pattern recognition receptors (PRRs) 221, 223–225 pneumonia 300
PBRM1 449 POLE 405
pembrolizumab 179, 340 pollutants 115–123
penile cancer 513, 516 poly(ADP-ribose) polymerase (PARP) inhibitors 370
pentachlorophenol 124, 472 polybrominated diphenyl ethers (PBDEs) 122
pepsinogen 338, 341 polychlorinated biphenyls (PCBs) 122, 231
peptic ulcer disease 333, 335 polycyclic aromatic hydrocarbons (PAHs) 52, 113, 129,
perfluorinated alkylated substances 120, 121 170, 299, 300, 325–328, 340, 379, 483
perfluorooctanoic acid (PFOA) 121, 563 polycystic ovary syndrome 407
perineal powder 415, 418 polygenic risk score 158, 537
peripheral T-cell lymphoma (PTCL) 471, 474 polymerase chain reaction (PCR) 210, 552–554
persistent organic pollutants 121, 122, 563 polyps 346
pesticides 122, 124, 442, 473 population attributable fraction (PAF) 99, 152
Peutz–Jeghers syndrome 336 population-based screening programmes 267, 271,
phagocytosis 216 274, 540, 545, 546
pharmaceutical drugs 137–142 population-wide prevention strategies 502, 503
pharmacotherapies for tobacco cessation 490–492, POT1 376
495 POU5F1 436
pharyngeal cancer 70, 72, 241, 310, 311, 317 PPARγ 465
phenylketonuria 168 PPM1D 465
Philadelphia translocation 159, 481 PRDM14 435
PHLDB1 456, 457 prebiotics 226
photocarcinogenesis 78–82 precision medicine 168, 177, 306, 307, 361, 550, 551
photosensitivity 82, 171
precision prevention 168
phototherapy 375
pregnancy 189, 194, 383, 387
phthalates 435
premature mortality 16–20, 34, 41
physical activity 95, 101–105, 144, 145, 254, 258–260,
prevention strategies 499–504, 518, 522, 523, 525–
262, 263, 291, 407, 408, 499, 501, 503, 504, 506,
528, 542–546, 557–559, 561, 564–566, 568–570
507, 509–511, 565, 566, 569, 570
primary effusion lymphoma 65
physical inactivity 19, 20, 26, 27, 103, 185, 254, 408,
510, 565, 566 probiotics 226
phytochemicals 92, 93 processed foods 93, 96, 97, 507, 511, 567, 569

processed meat 93, 97, 99, 112, 186, 340, 350, 507, radiation 77–82, 84–89, 231, 375–380, 459, 463,
567 464, 466
progenitor cells 150, 151, 387 radiofrequency electromagnetic fields 88, 89, 455, 560
progesterone receptor (PR) 382 radiomics 307
progestin-containing intrauterine devices 406, 408, 409 radionuclides 84, 87
progestins 191–193, 406 radiotherapy 84, 249
progestogen-only contraceptives 137, 140, 141, 406 radon 84, 87, 127, 299
prognostic markers 210, 211, 398 raloxifene 391, 523, 525
programmed cell death 1 (PD-1) 179, 217, 218, 226, ramucirumab 340
307, 444 RANK ligand 391
programmed death-ligand 1 (PD-L1) 340, 444 RB1 301, 324, 356, 376, 377, 425, 440
prolactin 388 reactive nitrogen species 231
prostaglandins 181, 218, 408 reactive oxygen species 68, 80, 81, 176, 231, 368
prostate cancer 17, 28, 29, 40, 93, 95, 102, 106, receptor-mediated effects 232–234
112, 157, 158, 196–198, 203, 218, 240, 246,
248–250, 254, 260, 263, 421–427, 525–527 receptor tyrosine kinases 336
prostate cancer screening 28, 29, 422, 426, 427, rectal cancer 141, 344, 346, 347, 350, 352
540, 541 red meat 93, 97–99, 112, 186, 340, 350, 368, 507
prostate-specific antigen (PSA) 28, 29, 40, 421, 422, 5α-reductase inhibitors 427, 527
426, 554 Registration, Evaluation, Authorisation and
prostatic intraepithelial neoplasia 426, 427 Restriction of Chemicals (REACH) 562, 563
Protocol to Eliminate Illicit Trade in Tobacco regulatory initiatives 508, 511, 557, 558, 560–564
Products 564 regulatory T cells (Treg cells) 217, 219
proton-pump inhibitors 330 renal cell carcinoma 40, 111, 178, 447–451
PRSS1 369 reproductive factors 25, 189–198, 254, 385, 389, 405
psoriasis 474 respiratory disease 16, 19, 565, 566, 568
PTCH1 379, 380 resveratrol 69, 187, 526
PTEN 346, 376, 377, 404, 405, 413, 440, 448, 464 RET 302, 464, 465
PTX3 216 retinoblastoma 165, 300
pulmonary nodules 306, 307 RHOA 336, 474
pyrimidine dimers 78, 81 ribavirin 362, 365
risk assessment 561
risk prediction models 191, 193, 349, 360–362, 378,
Q 383, 388, 389, 536, 537
quality assurance in screening 540, 545 ROS1 301, 307
quantitative structure–activity relationship (QSAR) RTEL1 456

models 562 rural–urban disparities 253–255, 259, 260, 276,
quitlines 490, 491, 495 279, 510
R S
racial and ethnic disparities 276–279, 282, 283, 285, salivary glands 310
286, 289, 291, 389, 426 Salmonella 222
racism 278, 289 salpingectomy 418
γ-radiation 84 salpingo-oophorectomy 537
608
salt-preserved food 318 small-molecule inhibitors 227

Schistosoma haematobium 61, 66, 440 SMO 380
Scientific Committee on Occupational Exposure smokeless tobacco 51–54, 57, 58, 260, 261, 317, 368
Limits 560 smoking 24, 25, 50–54, 56–58, 241, 254–256, 299,
screening 240, 242, 255, 259, 262, 263, 266–274, 300, 303–305, 313, 317, 320, 325–327, 340, 341,
277–283, 285, 286, 292, 304–306, 319, 320, 328, 368, 372, 383, 384, 398, 407, 440–443, 450, 488,
347, 349, 351, 352, 363, 371, 399, 400, 416, 417, 490, 494, 495, 499, 500, 502, 558, 568, 569
422, 426, 451, 466, 500, 501, 506, 509, 530, 532, smoking cessation 299, 304, 330, 368, 372, 384,
534, 536, 537, 540–547, 550, 551, 553, 559 493, 494
SDHB 335 social inequalities 238–243, 246–250, 253–255,
secondary bile acids 186, 223, 224, 227 258–263, 271, 273, 274, 276, 278, 279, 281, 289,
291, 292
second-hand smoke 50, 54, 57, 119
Sociodemographic Index (SDI) 344, 345
sedentary behaviour 104, 105, 254, 258, 262, 291,
407, 510, 511 socioeconomic status 29, 238–243, 246, 247, 249,
250, 253–255, 258–263, 271, 273, 274, 276–279,
selective estrogen-receptor modulators (SERMs) 285, 326, 330, 345, 352, 389, 399, 426, 463, 480,
142, 391, 523, 525 500, 501, 508, 530
selenium 328, 330, 526 soil contamination 121, 122, 291
self-renewal 150 solar radiation 40, 77–82, 127, 129, 133, 134, 375,
seminoma 430, 431, 435, 436 377, 378
serous borderline tumours 412 solid fuel combustion 116, 119, 299, 567
serous tubal intraepithelial carcinomas 412, 417 somatic mutations 156, 158–160, 411, 413
serrated pathway 346 somatic stem cells 150, 152
sessile serrated adenoma 346–348 soot 129
SETD2 449 SOX11 473
sex cord stromal tumours 411, 430 spermatocytic tumour 431
sex hormone-binding globulin 194, 196 spermatogenesis 434, 437
sexual and gender minorities 276 SPINK1 425
shift work 129, 133, 134 sporadic cancer 148–153
shingles 458 SPRY4 435
sigmoidoscopy 270, 273, 352 squamous cell carcinoma antigen (SCC-Ag) 398
signet ring cell carcinoma 333, 341 squamous intraepithelial lesions 395, 398–400
silica 116, 134, 135, 129, 231, 299 STAG2 440
simian virus 40 458 STAT3 183, 187

statins 330, 372, 407, 408, 527
single-nucleotide polymorphisms (SNPs) 155–157,
160, 164–166, 349, 369, 383, 389, 456, 457, 537 stem cells 150–153, 387
Sjögren syndrome 472, 473, 475 STK11 335, 336, 369
skin cancer 69, 77–82, 120, 134, 374–380, 567 Stockholm Convention on Persistent Organic
Pollutants 563
skin cancer screening 378
stomach cancer 19, 29, 30, 53, 61–63, 65, 69, 121,
skin squamous cell carcinoma (SCC) 79, 81, 82,
185, 240, 241, 250, 252, 254, 255, 259, 290,
378–380, 560
333–341, 525–527, 558
SLC45A2 376, 379
stomach cancer screening 341, 540
SMAD4 324, 346, 370
stroke 488, 566, 567, 570
small cell lung carcinoma (SCLC) 299, 301–303, 307
sugar-sweetened beverages 96, 98, 350, 501,
small lymphocytic lymphoma (SLL) 478 507–509, 559, 569

sulforaphane 186, 187, 526 TFAP2C 436

sulfur mustard 231 thyroid cancer 40, 86, 87, 106, 159, 240, 243, 292,
sulindac 526 461–466, 558
sunburn 78, 80, 81, 376 thyroid cancer screening 466
sunlamps and sunbeds 79, 375, 559 tissue-specific stem cells 150
TMPRSS2-ERG 425
sunlight 77–82, 377, 379, 380
tobacco cessation 488–495
sun protection 79, 82, 377, 502
tobacco control 54–56, 58, 242, 280, 304, 488, 489,
sun protection campaigns 377, 502, 557
491, 495, 499, 500, 502, 504, 563, 564, 568, 569
sunscreen 377, 380
tobacco products 50–54, 56–58, 261, 262, 317, 488,
survivin 449 493–495, 561, 564
survivorship 95 tobacco smoking 20, 24, 25, 50–58, 160, 209, 238,
susceptibility alleles 154–158, 160, 166, 167, 383 242, 255, 260, 261, 289, 299, 300, 304, 305, 313,
317, 320, 325, 326, 328, 340, 341, 368, 372, 407,
sustainability of implementation strategies 502–504
440–443, 450, 488, 494, 499, 500, 558, 568
Sustainable Development Goals 17–20, 362, 516,
tobacco-specific nitrosamines 52, 53, 299
565, 569
topoisomerase II alpha 449
synthetic lethality 177
toxicity testing 562
Toxic Substances Control Act 561–563
TP53 mutations 80, 111, 155, 156, 160, 175, 301,
T 304, 318, 324, 325, 336, 346, 356, 359, 369, 370,
376, 377, 380, 383, 405, 413, 440, 456, 457, 465,
talc 415, 418 530, 536, 537
tamoxifen 191, 391, 522, 523, 525, 528 traditional serrated adenoma 348
tanning devices 79, 375, 377, 380, 559, 560 TRAF3 318
Tao brush 417 transcription factors 181–184, 186, 187
taxation 242, 500–502, 508, 509, 559, 569 trans fat 97, 98, 559
telbivudine 364 transforming growth factor β (TGF-β) 215, 218, 387
telomerase 149, 457 transitions in human development 26, 27, 35, 36,
telomeres 149, 346, 457 40–43
temozolomide 175 transvaginal ultrasound 416
tenofovir 362, 364 trastuzumab 324, 340
teratoma 431, 436 trichloroethylene 129, 450, 451, 473
TERC 457 trigonelline 204
TERF2IP 376 trihalomethanes 120
TERT 356, 376, 377, 455–457 triple-negative breast cancer 382, 383, 385, 387,
389, 391
TERT-CLPTM1L 300, 301
tubal ligation 414, 415, 418
testicular cancer 121, 123, 157, 196, 430–437, 563
tumour antigens 216
testicular dysgenesis syndrome 430, 431, 434
tumour-associated macrophages (TAMs) 184, 215,
testicular germ cell tumours 430–432, 434–436
216, 218, 219
testosterone 194, 196, 387, 388, 437
tumour-educated platelets 550, 553, 554
TET2 474 tumour-infiltrating lymphocytes 217
tetrabromobisphenol A 233 tumour microenvironment 148, 150, 160, 184,
tetrachloroazobenzene 233 215–219, 541, 553
2,3,7,8-tetrachlorodibenzo-para-dioxin (TCDD) 124, tumour necrosis factor α (TNF-α) 80, 181, 183, 184,
229 186, 387
610
tumour suppression 149, 151 vinyl chloride 123, 129
tumour suppressor genes 158–160, 177, 208, 212, viral hepatitis control 355, 362, 364, 365, 570
303, 318, 335, 346, 356, 370, 398, 404, 425, 441, visual inspection of the cervix with acetic acid (VIA)
448, 449 263, 546
tylosis 324 visual oral examination 319, 320
TYR 376, 379 vitamin A 526
tyrosine kinase inhibitor 481 vitamin D 27, 94, 96, 526
vitamin E 526, 527
vitamins 92, 94, 96
U von Hippel–Lindau syndrome 448
UGT1A 444 voriconazole 82
ultrasonography 361, 363, 365, 371, 462 vulvar cancer 513, 516
ultraviolet lamps 78
ultraviolet radiation 77–82, 160, 170, 175, 375–377,
379, 380, 559 W
United Nations Declaration on the Rights of
Indigenous Peoples 292 waist circumference 106, 107, 407, 559
United Nations Development Programme 35 Warburg effect 202
United States Preventive Services Task Force 279, water contamination 120, 121, 291, 440, 442, 443, 567
305, 422, 426, 526 water pipe smoking 51, 53, 57, 262
United States Surgeon General 53 weight loss 390, 408, 511
universal health coverage 20, 242, 478, 485, 502, 571 welding 79
unopposed estrogen therapy 191 WHO Framework Convention on Tobacco Control
upper urinary tract urothelial carcinomas 175, 178, 54, 56, 491, 499, 563, 564, 569
440, 441 WHO Global Burden of Disease Study 134
urbanization 247, 253–255, 510 whole-genome sequencing 156, 553
uterine cancer 260, 403–408 whole grains 92, 93, 95, 97, 98, 350, 507
uterine lavage 417 WHO MPOWER package 54, 58, 499
WHO Report on Cancer 485
Wilms tumour 447
V wood dust 129
World Cancer Research Fund 68, 92–94, 101, 144,
vaccination 63, 64, 259, 263, 320, 355, 362, 364,
145, 506, 509, 567
399, 499–502, 504, 513–519, 570
World Code Against Cancer 295
vaccines 63, 64, 65, 66, 500, 513–515, 517–519
vaginal cancer 513, 516
varenicline 491
X
variants of uncertain significance 532, 536
varicella zoster virus 458, 459 xeroderma pigmentosum 82, 171
vascular endothelial growth factor receptor 2 X-radiation 84
(VEGFR2) 340
vascular endothelial growth factor (VEGF) 184
vegetables 92, 93, 95, 97, 98, 99, 185, 186, 326, Y
350, 507, 526, 559
VHL 448, 449 yolk sac tumour 431

World Cancer Report
Cancer research for cancer prevention
Cancer is the second most common cause of death globally, accounting for an estimated 9.6 million deaths in 2018.
The 2017 World Health Assembly requested WHO, in collaboration with IARC, to provide a global perspective on
all measures that are recognized to limit the burden of cancer. The outcome of this charge – the WHO Report on
Cancer: Setting priorities, investing wisely and providing care for all – complements the IARC World Cancer Report
by synthesizing evidence to translate the latest knowledge into actionable policies to support governments.
— Dr Tedros Adhanom Ghebreyesus, Director-General, WHO
In 2014, World Cancer Report established that it is implausible to treat our way out of the coming cancer burden:
prevention is the only option. Accordingly, this new World Cancer Report is totally focused on prevention, and it
is the most comprehensive overview of relevant research currently available.
— Dr Christopher P. Wild, IARC Director 2009–2018
This new World Cancer Report provides investigators with detailed information across a multidisciplinary spec-
trum. Equally, World Cancer Report provides people in the wider community, no matter where they are located
worldwide, with insights into how the cancer types that have for so long affected their communities may now have
a lesser impact than was previously thought.
— Dr Elisabete Weiderpass, Director, IARC
“Cancer research for cancer prevention” is not simply a way to describe a particular field of investigation. Far
more importantly, these words identify a pathway that may materially reduce the acknowledged burden of cancer
faced by humanity. There is, in fact, no other way.
— Professor Bernard W. Stewart, University of New South Wales, Sydney
Highlights of this World Cancer Report include:

• Although excess body fatness increases the risk of cancers at various organ sites, including the colon and
rectum, the risk may be reduced by intentional weight loss.
• Cancer-causing pollution of air and water are amenable to intervention by technological and regulatory means.
• Cervical cancer may be eliminated as a public health problem by vaccination against human papillomavirus
(HPV) infection, even in low-income countries where cervical cancer is the major cancer type.
• In most countries, socioeconomic disparities limit the impact of proven preventive interventions.
• Individual susceptibility to particular cancers is increasingly understood from molecular technology.

World Cancer Report 2020

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

World Cancer Report 2020

Uploaded by

Copyright:

Available Formats

World

Edited by CHRISTOPHER P. WILD,

Edited by CHRISTOPHER P. WILD,

© International Agency for Research on Cancer 2020

Sales, rights and permissions.

IARC Library Cataloguing in Publication Data

World Cancer Report

Editors Christopher P. Wild

Programme Officer Véronique Chajès

Jason Gurney Béatrice Lauby-Secretan Josiah Ochieng

Rolando Herrero He Li Graham Pawelec

Rayjean J. Hung Terry Lichtor Neil Pearce

David J. Hunter Martha S. Linet David H. Phillips

Ivano Iavarone Johan P. Mackenbach Sydney E. Philpott-Streiff

André M. Ilbawi Núria Malats Martyn Plummer

Lisa Iversen Reza Malekzadeh Igor Pogribny

Charles W. Jameson Mohandas K. Mallath Kornelia Polyak

Aoife Ryan Alexandra G. Smith Christine Varon

Known causes of human cancer by organ site 45 3.5 Inflammation 181

2 Causes of cancer, including 49 3.6 Reproductive and hormonal factors 189

2.1 Tobacco products 50 3.7 Metabolic change and metabolomics 200

4.5 Variations in implementation of cancer 266 5.16 Kidney cancer 447

Dr Tedros Adhanom Ghebreyesus

World Cancer Report ix

World Cancer Report xi

The scope of this World Cancer Report

What information is provided in World Cancer Report?

World Cancer Report xiii

Where to from here?

Christopher P. Wild, Elisabete Weiderpass, and Bernard W. Stewart (Editors)

The burden and prevention of premature

or medium levels of the Human NCDs have become the lead-

No data Not applicable

No data Not applicable

Low / Medium HDI High / Very High HDI

Change in Risk of dying (%),30-69 years old

2000 2015 Reducing the risk

Global trends in cancer incidence

Isabelle Soerjomataram Bernard W. Stewart (reviewer)

●● An increase in prostate can- This chapter reviews the incidence

Chapter 1.2 • Global trends in cancer incidence and mortality 23

Chapter 1.2 • Global trends in cancer incidence and mortality 25

Chapter 1.2 • Global trends in cancer incidence and mortality 27

Chapter 1.2 • Global trends in cancer incidence and mortality 29

Chapter 1.2 • Global trends in cancer incidence and mortality 31

Chapter 1.2 • Global trends in cancer incidence and mortality 33

Transitions in human development

Miranda M. Fidler-Benaoudia Bernard W. Stewart (reviewer)

year by 2040, a 50% increase on health and disease patterns are

Cancer profile by HDI

Known causes of human cancer by organ site

All cancer sites (combined) 2,3,7,8-Tetrachlorodibenzo-para-dioxin

Anus Human immunodeficiency virus type 1

Biliary tract 1,2-Dichloropropane

Bladder Aluminium production

Brain and central nervous system X-radiation, γ-radiation

Breast Alcoholic beverages

Cervix Diethylstilbestrol (exposure in utero)

Colon and rectum Alcoholic beverages

Known causes of human cancer by organ site 45

Organ site Agent

Corpus uteri (endometrium) Estrogen menopausal therapy

Endothelium (Kaposi sarcoma) Human immunodeficiency virus type 1

Eye Human immunodeficiency virus type 1