Journal of Gastroenterologyand Hepatology (1991)6, 325-344 ADONIS 081593199100053J

WORKING PARTY REPORT T O T H E WORLD CONGRESSES OF

GASTROENTEROLOGY, SYDNEY 1990

Clinicopathological staging for colorectal cancer:

An International Documentation System (IDS) and an International
Comprehensive Anatomical Terminology (ICAT)

L. P. FIELDING,’.** P.A. ARSENAULT,’ P. H. CHAPUISY30. DENT: B. GATHRIGHT:

J. D. HARDCASTLEP P. HERMANEK? J. R. JASS8 AND R. C. NEWLAND9
’ Yale University School of Medicine,Department of Surgery, St Mary’s Hospital, Waterbury, Connecticut, U S A ,
Colon and Rectal Unit, Concord Hospital, Sydney, Australia, Department of Sociology, Australian National
University, Canberra, Australia, ’Department of Colon and Rectal Surgery, The Ochsner Clinic, New Orleans,
Louisiana, U S A , Department of Surgery, University Hospital, Nottingham, UK, Chirurgische Klinik der
Universitaet Erlangen-Nuernberg, UK,Erlangen, Germany, Department of Pathology, University of Auckland
School of Medicine, Auckland, N e w Zealand, and Department of Anatomic Pathology, Concord Hospital, Sydney,
Australia

Abstract The purpose of tumour staging for colorectal cancer (CRC) is to help define clinical management, facilitate
communication between physicians, provide a basis for stratification and analysis of treatment results in prospective studies,
and provide some prognostic information for patients and their families. The World Congresses of Gastroenterology, Digestive
Endoscopy, and Coloproctology, Working Party on staging for CRC studied six commonly used systems to review their
strengths and weaknesses. Although it was concluded that defining a new staging system was unnecessary, it was recognized
that there is a need to define a terminology to describe the full anatomic extent of spread of CRC. Furthermore, we note that
there are several additional features, derived from both clinical and pathology information, which have had prognostic
significance shown by appropriately constructed multivariate analyses and which can be used to formulate a more accurate
prognostic index than that provided by a description of anatomical tumour spread.
Thus the Working Party came to two principal conclusions. First, a standard format should be adopted for the collection of
the essential data required for prospective studies, and we recommend the ‘International Documentation System (IDS) for
CRC’ for this purpose. Second, a nomenclature which describes the full anatomical extent of tumour spread and residual
tumour status in CRC has been defined and should be adopted, from which all currently used staging systems can be derived.
We have called this nomenclature the ‘International Comprehensive Anatomical Terminology (ICAT) for CRC‘.
In the event that these recommendations are adopted, we envision that there will be improved clarity in the documentatioii of
treatment outcome for patients with CRC and improved communication of results derived from prospective studies.
Furthermore, an acceptance of IDS and ICAT would set the scene to develop a prognostic index for individual patients with
CRC by the expansion of anatomical clinicopathology staging information to include additional factors which have independent
prognostic significance.

Key words: colorectal cancer, multivariate analytical methods, prognosis, prognostic index, staging system.

HISTORY OF COLORECTAL CANCER
STAGING
Notwithstanding the preliminary observations at the turn of
the century by Cloggs’ and Miles2on the spread of colorectal
cancer (CRC), Lockhart-Mummery in 19263 is usually
acknowledged as the first to develop a staging classification
*Chairman.
Correspondence: L. P. Fielding, MB, FRCS, FRACS, Department of Surgery, St Mary’s Hospital, 56 Franklin Street, Waterbury, CT
06706, USA.
Accepted for publication 21 January 1991.
for this tumour. This was a simple ‘ABC’ system based on
his operative findings in patients with rectal cancer treated
by perineal excision. Dukes in 19324refined these observa-
tions, which had been made at operation, by giving his-
tological definition to each stage group depending on the
depth of tumour penetration into the bowel wall and the
existence of lymph node metastases. In 1935 Gabriel, coI-
laborating with Dukes and Bussey,’ refined the lymph node
326
positive group (C cases) by subdividing them (C1 and C2)
according to the level of nodal involvement in the resection
specimen, and in 1939 Simpson and Mayo6 applied this
classification to colon cancer. The validity of this method as
a prognostic system was confirmed by Dukes and Bussey in
1958’ in their analysis of survival of 2447 patients with rectal
cancer treated surgically.
The first important departure from the Dukes system was
made at the Mayo Clinic to evaluate the use of a restorative
resection in patients with tumours of the sigmoid colon and
rectum.* This operation was not generally accepted within
the United States, especially by Astler and Coller who in
19549 argued against its use for patients with tumours
arising in the extraperitoneal rectum. Their criticism that
this was in inadequate operation for rectal cancer was high-
lighted by dividing Dukes C cases into two groups according
to the depth of direct spread of the tumour through the
bowel wall. These and other modifications’’ of the Dukes
classification are all examples of anatomical staging systems
based solely on the pathologist’s examination of the surgical
specimen.
So-called clinicopathological (CP) staging of CRC deviates
fundamentally from that derived from pathology informa-
tion alone, in that a CP method takes into account the
presence or absence of tumour beyond the operative field.
This method of staging was introduced by Rupert Turnbull
and his colleagues” to describe his results for ‘no touch’
colectomy for colon cancer. He introduced a ‘D’ stage for
patients with tumours which at operation were considered to
have invaded adjacent organs as well as for those with
distant metastases.
In 1971, refinements to CP staging were introduced at
Concord Hosptial, Sydney. Data has been collected prospec-
tively since that time,”-14 and the Australian Clinicopath-
ological Staging System (ACPS)” represents a simplified
version of this staging method which has been extensively
reviewed and statistically validated using both univariate16
and multivariate statistical techniques. l7 A similar approach
was developed independently at the University of Erlangen,
Germany” and designated ‘Erlangen Prognostic Groups’
which had been used since 1969. The pTNM system used
by the Union Internationale Contre le Cancer (UICC)19
describes a four-stage clinicopathological system. However,
the principal difference between the Concord Hospital stag-
ing system and TNM lies in the obligatory search for
evidence of tumour transection in the operative specimen, in
the Concord system. The American Joint Committee on
Cancer (AJCC) and the UICC introduced in 1987 an option-
al residual tumour (R) classification.19320 This, combined
with the pTNM stages, permits direct comparisons to be
made of the two systems. In 1983, the Japanese Research
Society for Cancer of the Colon and Rectum reported (in
English) a very detailed CP system similar in principle to the
pTNM method but subdivided into five stages which had
been used in Japan since 1977.2’
Although beyond the scope of the Working Party, it
should be mentioned that in addition there is also a purely
clinical classification of tumour spread, which can be used
before treatment. This was developed by the UICC and
L. P. Fielding et a\.
AJCC and follows the same criteria as pTNM, but is
designated TNM or cTNM, and was devised to assist in the
selection of primary treatment (especially neo-adjuvant
radio- andor chemotherapy), and for comparison of results
of surgical and non-surgical (non-resective) treatments. 19320
,It is clear from this short review that despite some
similarities between staging systems, meaningful direct
comparisons cannot readily be made.
It is with this background that the World Congresses of
Gastroenterology, Digestive Endoscopy, and Coloproc-
tology (WCOGDEC) arranged this working party to study
the subject of clinicopathology staging of CRC and to make
recommendations in the hope that some type of unifying
approach might be derived.
OBJECTIVES AND METHODS
Having been asked to consider the subject of clinicopatho-
logical staging of large bowel cancer, the Working Party first
set out to identify a number of objectives which it felt could
be achieved in the time and with the financial resources
available. Those objectives selected were as follows: (i) to
identify those features of clinical information and histopath-
ology analysis which should be recorded in all cases of large
bowel cancer; (ii) to establish an internationally accepted
language by which the features in Objective 1 should be
described; and (iii) to consider the requirements of a new
clinicopathological staging system, or a multifactorial prog-
nostic index.
Working Party conduct
Although we recognized that a meeting of the whole Work-
ing Party would be desirable, it was unclear when we started
our work whether we would be able to arrange such a
meeting because of the expense and travel time involved for
a group so widely distributed around the globe.
However, and additional grant from the Australian Can-
cer Society made a meeting of the group possible. As a first
step we agreed to communicate in a series of agendas by
facsimile (FAX) machine with an occasional telephone call
to supplement the system.
In the first ‘FAX Round’ we identified a series of pro-
p o s a l to~ which
~ ~ ~each
~ ~participant responded with analysis
and recommendations. These papers were then collated by
the Chairman and our research assistant and all responses
for a given proposal were distributed to the members of the
Working Party with comment and conclusions by the Chair-
man. Thus, each item was cycled through the system until
one of two outcomes was achieved: (i) consensus opinion
was generated; or (ii) divergence of opinion was identified.
In the latter case, the Chairman arbitrated with the under-
standing that the recommendations were not ‘binding’ on
the group. The great majority of all these ‘proposals’ thus
reached resolution (61/69,88’/0) after seven rounds of FAX.
The remaining eight proposals which had not been agreed
upon formed the initial part of the agenda of the meeting of
the Working Party, in May 1990.
Colmectalcancer staging
Gradually as the ,proposals were modified and accepted,
we were able to create a list of items with mutually exclusive
alternative responses which, with the addition of some
comments and definitions, form a summary of our delibera-
tions and a framework from which future documentation
might be generated. This summary format is shown in
Appendices 1-111.
The distribution to all members of the Working Party of
everyone’s opinions on all proposals worked very well from a
functional point of view, although it was a time-consuming
exercise. However the benefit of this extensive review was
evident when the group eventually met bccause there was a
great deal of common ground which allowed us to resolve
the few remaining points of disagreement. The three-day
meeting was attended by all the Working Party except for
Professor Hardcastle and Dr Gathright. However, during
the conference, there was telephone communication with
these two members who agreed with the outcome of the
outstanding issues discussed. In addition, two additional
‘resource’ physicians were invited to attend: Dr Robert
Hutter, on behalf of the College of American Pathologists
and the American Joint Committee on Cancer, and Dr
Stanley Goldberg, both physicians having a long-standing
interest in the subject of staging for CRC.
The meeting itself was used to resolve outstanding issues,
write the summary paper for distribution at the time of the
WCOGDEC meeting in Sydney, plan the full paper for
publication and arrange the conduct of the Working Party
discussion for the congress. After the meeting, each member
of the Working Party wrote a section for the full version of
the paper, which was then edited into a single contribution
by the Chairman.
For the purposes of our review, we analysed the details, as
well as the strengths and weaknesses, of the six most com-
monly used staging systems, namely: The Dukes System,7
the Astler-Coller modification of the Dukes System? the
Concord Hospital Clinicopathological System,14 the Aus-
tralian Clinicopathological Staging System (ACPS),” the
UICUAJCC pTNM and the Japanese Research
Society Staging System.”
Statistical methods
Statistical considerations are fundamental to the develop-
ment of either a staging system or a prognostic index for
CRC. The two principal purposes of a staging system are:
to categorize patients into relatively homogeneous groups
(stages) according to their average expectation of life after
treatment; and to allocate patients to treatment regimens
according to the anatomical extent and histological charac-
teristics of their tumours. By contrast, a prognostic index
gives a specific prediction of outcome, within a margin of
probable error, for an individual patient based on the infor-
mation available at a particular time. In general, a prognos-
tic index will incorporate a wider range of tumour and
patient characteristics than are used for staging.
The key requirements for a staging system are that: (i) the
stages should be based on clinical and pathology criteria
which are widely accepted and clearly defined, and which
327
may be recorded with high reproducibility; and (ii) each
successive stage should be associated with a decrement in
survival as shown in large, unselected series of patients who
have been followed after treatment for several years. Patient
survival is measured from the time of operation or other
treatment or decision not to treat, to an outcome event,
which may be death from any cause or from CRC or, in
more specialized studies, the time of diagnosis of local
recurrence’ or distant metastasis, or metachronous CRC or
some other A range of statistical techniques is
available for examining the relationship between survival,
and patient and tumour characteristics.
Univariate statistical techniques applicable to such data
include life-table survival analysis25326and the Kaplan-
Meier estimator27 and their associated tests of statistical
significance for differences in survival. An important aspect
of these methods is that they permit the analysis of data from
‘right censored’ cases, that is patients who have been lost to
follow-up and those who, at the close of follow-up, had not
experienced the outcome event.
These methods have been used effectively in the evalua-
tion of CRC staging systems to plot the survival function for
patients within stages and to evaluate differences in survival
between stages. However univariate analysis cannot cope
with the ‘confounding’ effects of variables such as age and
sex, which are related to survival and which must be con-
trolled before any attempt is made to evaluate the association
between survival and other predictor variables, such as
tumour stage. Nor can univariate methods deal with ‘inter-
action’ among predictor variables, for example where the
effect of a third variable on survival differs in magnitude
from one stage to another within the staging system.
The method of ‘stratified’ univariate analysis attempts to
assess the impact of confounding and of interaction between
variables. Furthermore, stratification also attempts to evalu-
ate the effects of multiple predictor variables by separating
patients into subgroups or ‘strata’ on two or more predictors
and making between-strata comparisons of within-stratum
univariate analyses. For example, to examine the effect of
venous invasion on survival, controlling for sex, one would
use the Kaplan-Meier method to compare survival between
male patients with and without venous invasion, and then
compare this with the survival analyses for female patients
with and without venous invasion. However, a large number
of patients is needed when stratification is applied to more
than two or three predictor variables, because there need
to be sufficient numbers of patients in each substratum
for valid analysis. Furthermore, the number of categories
into which predictors are coded must be severely limited
for the same reason. Recent developments in multivariate
mathematical modelling have largely supplanted stratified
analysis.
Mathematical models, of which the nonparametric Cox
proportional hazards regression model is one example,18
provide a means of simultaneously evaluating the independ-
ent effects of several predictor variables on survival when
the data include censored cases. Such a model can show
whether a particular clinical or pathological variable is asso-
ciated with survival after the effects of other variables in the
328 L.P. Faekfing et al.

model have been statistically controlled. It will also show the hand column of each part of Appendices I and 11. Some
relative risk associated with each predictor variable. Further- features often recorded as a ‘routine’ have been excluded
more, the Cox regression model allows the inclusion of from our recommendations and these will be discussed.
different types of predictors: continuous variables (such as In addition, the Working Party derived an ‘International
age in single years); ranked items (such as the six levels of Comprehensive Anatomical Terminology’ (ICAT) (which is
direct spread of a primary tumour); and dichotomous vari- a small subset of the information in the IDS), to facilitate the
ables (such as the presence or absence of venous invasion). full description of the anatomical spread of CRC (Table 2).
However it must be emphasized that modelling is by no Furthermore, ICAT provides the descriptive elements from
means a mechanical procedure where variables are simply which staging according to all the commonly used systems
‘plugged’ into any available computer program. The model- can be generated. By assigning a number to each ICAT line
ling method most appropriate to the problem must be item, a matrix has been developed defining the stages in
chosen carefully. Each variable being considered for inclu- each system (Appendix 111).
sion in the model must be examined closely in relation both The principal features of information required for follow-
to the outcome variable and to other variables already in the up data in prospective studies is outlined in Table 3. The
model. Confounding and interaction effects need to be specific items for a particular study would be added to the
clearly understood, and whatever assumptions the model follow-up document. Each of these results will be consid-
makes about the data need to be tested and found to be ered in turn.
satisfactory.
An extension of mathematical modelling can generate a
prognostic index, which is a figure derived from a formula Table 1 International Documentation System (IDS) for
which combines an individual patient’s scores on several colorectal cancer
independent predictors, to express the probability that the
patient will reach a specified outcome. This approach can Information type Clinical features Pathology features
encompass a variety of outcomes such as overall survival to a
given number of years, tumour-free survival, or survival to Basic information Country Number of primary
local or distant tumour recurrence. A prognostic index may Hospital (name/ tumours
be derived from the Cox proportional hazards model, or code) Tumour
from a model based on Bayes’ t h e ~ r e m , * ~or- ~from
~ other Patient identification measurements
(namdcode) Appearance of
mathematical modelling appro ache^.^^ Patient race serosal surface
The accumulation of large numbers of CRC patients with Past tumour history Associated pathology
long follow-up periods, coupled with the recent develop- Tumour type
ment of fast, powerful microcomputers and sophisticated
statistical software, has resulted in rapid growth in our Variables of proven Surgeon (namdcode) Extent of direct
knowledge of CRC survival. These statistical advances have prognostic Patient gender and spread
led to significant refinements in the staging of CRC and significance age Regional nodal
provide the basis for the future development of prognostic Presentation status
indices in this disease. Anatomic extent of Local residual
Thus the report the Working Party has produced has, in tumour tumour
Residual tumour Distant metastasis
large part, been possible by the acceptance of these statisti-
status
cal principles. It needs to be recognized, however, that Venous involvement
statistical understanding in this field continues to evolve, Histology of
and we are certainly not at the point of having a single, infiltrating margin
universally accepted, mechanical procedure which may be Tumour grade
applied in all circumstances.
Information of Pre-operative Tumour perforation
probable treatment Inflammatory cell
RESULTS prognostic Anatomic site of infiltrate
significance primary Lymphoid
Tumour mobility aggregates
Our findings, summarized in Tables 1, 2, and 3 and in Technique of
Appendices I, 11, and 111, bring together the information tumour
which we recommend for collection for all patients with mobilization
CRC who are being entered into prospective studies. Our Tumour perforation
recommended International Documentation System (IDS) Surgical procedure
(summarized in Table 1 and presented in detail in Appendices Resection of distant
I and 11) contains three types of information: (i) basic patient metastases
Postoperative
information; (ii) variables of proven prognostic significance;
treatments
and (iii) information of probable prognostic significance.
Each category is shown by a key classification in the left-
Colorectal cancer staging 329

Table 2 International Comprehensive Anatomical Terminology Table 3 Basic follow-up information

(ICAT) for colorectal cancer
1. Date of last follow-up -
Microscopic description of t u m r depth
1. Primary tumour cannot be assessed 2. Last known follow-upstatus - Alive
2. No evidence of primary tumour - Deceased with unknown
3. Severe dysplasia; carcinoma in situ colorectal carcinoma status
4. Tumour invades submucosa - Deceased with colorectal
5 . Tumour invades muscularis propria carcinoma
6. Tumour invades through muscularis propria into the - Deceased with other causes
subserosal connective tissue or non-peritonealizedpericolic or - Lost to follow-up
perirectal tissue 3. If alive, condition at last - No tumour recurrence
7. Tumour directly invades other organs or structures follow-up - Local recurrence only
8. Tumour to and invading free (serosal) surface of the - Distant metastases only
specimen - Both local and distant
recurrence
Regional lymph node status
9. Cannot be assessed
4. Additional treatment (type - Radiation adjuvant
10. Number of lymph nodes examined
and start date) - Radiation therapeutic
11. Number of nodes positive for tumour
- Chemotherapy adjuvant
- Chemotherapytherapeutic
12. line 11 = 0
13. line 11 = 1-3 positive nodes 5. Identification of new primary - Colorectal
14. line 11 = 3 positive nodes (site and date of diagnosis) - Other

Status of nodes on vascular trunk

15. Not recorded The full details of follow-up information will depend upon the
16. Negative for tumour study being undertaken and the outcome being assessed.
17. Positive for tumour

Apical node status

18. Not recorded and by histopathology examination if the tumour is resected;
19. Negative for tumour the details are listed below. However, for patients who d o
20. Positive for tumour not have tumour resection, the locoregional spread of tumour
can, to some extent, be assessed clinically by endoscopy,
Distant metastasis: status before definitive treatment computerized tomography (CT scanning), endoluminal son-
21. Cannot be assessed
ography, double contrast radiography, and, in the lower
22. None
23. Present rectum, by digital e~amination.~’ However, the Working
Party felt that a detailed review of these purely clinical
Residual tumour: status after definitive treatment evaluations in terms of their prognostic efficacy was beyond
24. Cannot be assessed its terms of reference.
25. None For patients who have resective surgery, survival varies
26. Locally in line of bowel resection only (shown histologically) from one surgeon to another, given a similar case and
27. Distant only (histologically or clinically) thus the ‘surgeon’ should be identified as a prognostic
28. Both local and distant variable. It is also known that the sex and age of the patient
influences outcome. l7
The different forms of clinical presentation have a sub-
Appendix I: Clinical features, (A) Pretreatment stantial effect on long-term prognosis. Patients diagnosed as
information part of a surveillance programme or as a coincidental finding
during an investigation of another disease are often at a less
advanced stage and may thus have a better prognosis.
Basic patient infomation (Appendix I : Items 1-5) However, this prolonged survival could be the result of lead-
The country in which the operation has been carried out, the time bias and proof as to whether screen-detected cancers have
hospital name and code, patient identification code, race, an intrinsically improved survival must await the results of
and past history of bowel tumours were considered neces- the randomized controlled trials currently in progress.37J8 I t
sary features for patient identification and for the analysis of is also important to note that patients identified by screening
results. methods are significantly more likely to be treated by
endoscopic polypectomy than are those presenting with
Variables of proven prognostic significance (Appendix I: symptoms. 37
Items 6-9) Multivariate analysis of staging criteria has been the
The anatomical extent of tumour is the most important essential step from which prognostic factors in CRC have
cluster of prognostic factors and is assessed both clinically been derived. However these analyses have included only
330
patients with symptomatic tumours. Thus we need to recog-
nize that asymptomatic patients need to be considered as a
separate group until their prognostic factors can be studied
in more detail.
It is important to identify individuals presenting as an
emergency with bowel obstruction or perforation, because
this type of presentation is associated with impaired progno-
si~.‘’,~~The definition of emergency presentation is the need
for urgent surgery within 48 h of admission.
The clinical definition of ‘perforation’ is relatively straight-
forward. It is a local abscess originating through the tumour,
or a defect giving rise to faecal peritonitis. Very occasionally,
the gut proximal to the tumour (usually the caecum) is
ruptured as a consequence of unrelieved, more distal ob-
struction. All these patients have a poor prognosis.
The definition of ‘bowel obstruction’ is difficult. It is clear
that as a group, this presentation has a diminished survival
both in the short and the long term. However, it is also clear
that these patients do not form a homogeneous group.
We believe that the diminished prognosis is more closely
related to the physiological impact of bowel obstruction than
it is indicated by the macroscopic description of the degree
of bowel distension. Therefore the definitions used to describe
the spectrum of conditions are on a three-point scale related
to these physiological effects, as follows: (i) obstruction to
solids only; little or no physiological effects; (ii) complete
obstruction to solids and gas; local bowel wall oedema, local
fluid sequestration; and (iii) obstruction to solids and gas
with systemic effects; the above physiological changes plus a
systemic effect on the cardiovascular system giving rise to
increasing degrees of circulatory decompensation.
Infmmation of probable prognostic significance
(Appendix I: Items 10-1 I)
It is very likely that both pre-operative treatment with
radiation therapy, as well as local and systemic chemo-
therapy, have some effect on prognosis. Therefore if these
modalities have been used, such patients need to be identi-
fied and analysed separately.
The site of a tumour within the colon should be defined
although there is much debate as to whether one site or
another leads to a better intrinsic survival. However, rectal
tumours, defined as lesions which have a lower border of the
tumour 16 cm or less from the anal verge, appear to have an
intrinsically poorer prognosis than colon adenocarcinoma.
Appendix I: Clinical features, (B)Post-treatment
information
Variablesof proven prognostic significance (Appendix I:
Items 12-17)
The ‘start date’ from which outcome analysis (e.g. survival)
is derived has been termed the ‘definitive treatment start
date’. Although for many this will be resectional surgery, for
some it will be pre-operative irradiation. The timing of
surgery (elective, urgent, emergency) reflects the form of
presentation, which has been demonstrated to have prognos-
tic si@cance. The presence or absence of distant meta-
L. P. Fielding et al.
stases is the most powerful of all prognostic information. A
chest X-ray is necessary in all patients with CRC to exclude
pulmonary metastases (more common in patients with rectal
than with colon cancer). The liver should be scanned in all
patients; intra-operative ultrasound gives the most accurate
information about the presence or absence of liver meta-
stasis. Histological confiiation of suspicious lesions
should, if at all possible, be obtained by biopsy or by needle
aspiration cytology. Once the pre-operative investigations,
imaging, and operative findings are available, the patient’s
liver status (number of tumour nodules in the right lobe and
in the left lobe of the liver), and the tumour burden (0,
< 25%; 25-50%; 51-75%; > 75%) should be documented.
After definitive treatment, a statement on the presence or
absence of residual turnour is, of course, a baseline clinical
feature, which will very strongly influence patient prognosis.
If structures to which a turnour is attached are not resected,
a biopsy should be taken from the area of adherence,
because fixation is often the result of an inflammatory
reaction rather than tumour, a distinction which may not be
appreciated by the surgeon at operation.
Information of probable prognostic significance
(Appendix I: Items 18-29
The mobility of a tumour should be assessed by palpation,
and should be described as ‘mobile, tethered, or fixed’,
because this feature has been shown to have prognostic
~ignificance.~’ The technique for tumour mobilization (con-
ventional versus Turnbull ‘no-touch‘ method) has been sug-
gested to influence the incidence of hepatic metastasi~.~’,~~
Tumour perforation (nondspontaneoudiatrogenic)has simi-
larly been reported in some series to be of prognostic
~ignificance.~~-~
Of course it is well known that the type 01 procedure
(non-resecting treatmendlocal surgery/limited resectiod
radical resection) has a major impact on prognosis. The
formerly used terminology was ‘palliative’ versus ‘curative’,
but we recommend the nomenclature just described, to
avoid the presumptions inherent in the more traditional
words.
The Working Party recognized that many alternative
names for operations have been used in the treatment of
CRC. We have suggested a comprehensive list of alterna-
tives to cover the areas of ‘non-resecting treatment’, ‘local
surgery’, ‘limited resection’, and ‘radical resection’, which
should accommodate all possible combinations. The resection
of adjacent organs attached to the tumour mass, either as
separate entities, or in an ‘en bloc’ fashion, is generally
considered to be of prognostic significance. It is recognized
that a small cohort of patients may have liver metastases
resected, followed by long-term survival, and thus this
information needs to be recorded. If this resection is not
carried out at the time of initial treatment, it may then be
added to the follow-up data if the procedure is undertaken
at a later date. Similarly, the use of postoperative non-
surgical treatments such as radiotherapy, local or systemic
chemotherapy, should be recorded in order to identify these
subsets of patients so that only appropriate analyses might
be undertaken.
Cohectal cancer staging
Appendix 11: Pathology features of the tumour
Basic patient infnmation (Appendix II: Items 1-5)
In the event that more than one primary carcinoma of the
bowel is present, the number of such tumours should be
given, and the stage recorded for the most advanced lesion.
Macroscopic measurements are necessary and special note
should be made of whether the serosal surface of the tumour
specimen is involved macroscopically with tumour. It is
customary to record the presence or absence of associated
pathologies which have malignant potential, as well as the
tumour type according to the World Health Organization
Classifi~ation.~~
Multivariate studies have demonstrated that defining
tumour type (adenocarcinoma, mucinous adenocarcinoma,
signet ring cell carcinoma, or others) conveys no independent
prognostic information once other factors, notably anatomi-
cal spread, have been considered.& However, it is so usual
to collect this information that this item has been included.
Variables of proven prognostic significance (Appendix II:
Items 6-15)
For all patients, every attempt should be made to make four
types of pathology assessment; (i) the depth of invasion of
the primary tumour; (ii) histological confirmation of lymph
node status; (iii) evidence of local residual tumour; and (iv)
evidence of distant metastasis (preferably by a biopsy sample).
Apart from the special circumstances where a tumour has
been removed by endoscopy or local surgical excision, most
invasive tumours will be resected by the removal of a length
of bowel and its attached mesentery.
By careful slicing of the tumour and thoughtful selection
of the areas to be sampled, pathologists should be certain
that the histological assessment of depth of tumour invasion
is made at the point of greatest tumour penetration. Large
area sections (whole-mount giant sections) have been found
to be especially suitable for this purpose, as well as for
measurements of the depth of invasion beyond the mus-
cularis pr~pria.~’.~’ The language being used to define each
line item by which the depth of penetration of the tumour
should be described, occupied much of the discussion time
of the Working Party. The two principal areas of difficulty
were: first, the description of superficial neoplastic change;
and second, the identification and description of residual
tumour, both local tumour spread beyond the confines of
the resected specimen and metastases.
We concluded that the first of these concerns should be
resolved by amalgamating into a single group the following:
severe dysplasia; carcinoma in siru; and superficial mucosal
cancer. While pathologists may be able to distinguish be-
tween these entities, their biological behaviour is such as to
justify their grouping into a single category. Furthermore,
there has been a reluctance to use the term ‘mucosal cancer’
lest clinicians overtreat this ‘non-metastasizing’ neoplasm.
Second, the recognition that a tumour has spread beyond
the confines of a resected specimen, and/or the presence of
distant metastases, are both very powerful prognostic state-
ments and should be accommodated in the staging systems
33 1
used in clinical practice and for research. l6 In addition, the
Working Party recommends that a distinction be made
between ‘adjacent organ invasion’, and ‘free serosal surface
invasion’. These two features are combined in the pT4 term
in the UICC/AJCC classification. However, unpublished
data from the Concord Hospital and from the Erlangen
Registry of Colorectal Cancer suggest that invasion of adja-
cent organs carries a more favourable prognosis than does
perforation of a free peritoneal surface.
In 1960 Dukes pointed out that rectal carcinomas with
invasion of the perirectal tissue have different prognoses
according to the extent of this extrarectal invasion.* Of
course there is a correlation here with the distance between
tumour and lateral (deep) line of resection.
The Concord Hospital study has shown that direct spread
of tumour from muscularis propria into the surrounding
tissues has no statistically significant impact on prognosis
provided that there is no involvement of a free mesothelial
surface or line of resection by tumour and that there are no
known metastases. This finding suggested that when clin-
icopathology staging is used, the extent of tumour clearance
from the lateral line of resection may be of relatively little
importance. Further studies using more precise measure-
ments of spread and lateral clearance will be needed to
confirm this o b s e r v a t i ~ n . ~ ~ ~ ~ ~ - ~ ~
For an adequate description of lymphatic spread in the
mesocolon or mesorectum of a resected specimen, a state-
ment on the number of lymph nodes examined and the
number involved with tumour is needed, as well as a
statement of their localization, namely in the perirectal or
pericolic tissue, on a named vascular trunk and/or apical in
the specimen.53
To yield an appropriate lymph node harvest, the meso-
colon, or the mesorectum, must be carefully studied. The
number of lymph nodes found in a resected specimen
depends on the size of the mesentery resected, on the care
taken by the pathologist, and also on the examination
method employed. Institutions which are especially hterested
in this topic report an average of 31 lymph nodes found in
colon carcinoma specimens, while 24 is the number found
in rectal carcinoma^.'^ Furthermore, the average number of
lymph nodes found after specimen fat clearance techniques
have been used is 47 nodes in colon cancer and 45 for rectal
~arcinoma.’~ However, we must ask the question, ‘Is it
necessary to study this number of lymph nodes in order to
assess lymph node status reliably in a specimen?’ Because
lymph node status is of such critical importance for progno-
sis,33*46,54-56
the Working Party was very concerned to for-
mulate clear guidelines for the statement that a specimen is
‘lymph node negative for tumour’. Thus the Working Party
has agreed on the following language:
Before deeming a radical resection to be without lymph
node metastasis, it is recommended that at least 12
lymph nodes be examined. If 12 nodes are not identi-
fied by traditional methods, further dissection or speci-
men clearance methods should be con~idered.’~-~ The
Working Party recognized, however, that not all speci-
mens will contain this number of lymph nodes, and this
is particularly true of the patients who have received
332
pre-operative irradiation therapy. Nevertheless, the
great majority of all radically resected specimens will
contain at least 12 lymph nodes, allowing the patholo-
gist to make this important determination.
Although an increased number of positive lymph nodes
adversely influences patient prognosis, outcome is also influ-
enced by the presence of nodal involvement along major
named vascular and also the presence of involved
apical nodes in the specimen.‘6i49@4 Whether these latter two
features are equivalent to each other is not known, and
therefore we have recommended that where possible, both
items of information be collected.
In addition to the features of anatomical extent of disease,
the Working Party identified a number of other pathological
items where there is sufficiently strong data to determine
that these features are of prognostic significance.
Tumour transected, shown histologically Clearly,
residual tumour at the site of the resection worsens progno-
sis. The pathologist therefore needs to assess the proximal
and distal lines of resection and in particular, the lateral
(deep) line of resection in order to make this determination.
The Working Party concluded that tumour cells must be
demonstrated in a line ofresection in order for this statement
to become positive. The coating of the deep lateral surface of
the specimen with Indian ink may be helpful in this regard.
While the histological examination of the lateral (deep)
line of resection is mandatory, it is not necessary to examine
the proximal and distal resection lines histologically, unless
(a) the margins are close (< 5 cm on measurement on the
fresh, unstretched specimen; or (b) the carcinoma shows
extensive lymphatic or venous invasion. The assessment of
the distal line of resection includes the histological examina-
tion of the ‘doughnut’ in case of stapled anastomosis.
On the rare occasion that distant metastases are excised,
the presence or absence of tumour in the line of resection
influences long-term patient s ~ r v i v a I . ~ ~ - ~ ~
Venous spread It is agreed that the incidence of venous
invasion increases with advancing and is associated
with distant spread of t ~ m o u r . ~An
’ independent prognostic
effect in multivariate analysis has been demonstrated by
some groups,17i68but not others.46 A significant influence
within specific substages has also been reported.71Although
the assessment of venous invasion is subjective, the observa-
tion warrants recording, in view of the preceding positive
findings. This histological evaluation may be assisted by the
use of elastic stains.
Histologic puttmi of infiltrating margin Classification
of the advancing turnour margin as ‘expanding’ or ‘infiltrating’
has been shown to provide independent prognostic informa-
tion in a number of s t ~ d i e s . ~ ~ Although
, ~ ’ , ~ ~ acceptable
levels of reproducibility have been demonstrated, judge-
ment is needed to assign cases according to the predominant
pattern of the tumour margin. Approximately 20-25% of
cases have an ‘infiltrating’ margin by this assessment in the
reports to date.
L. P . Fielding et al.
Tumour grade The long-established discipline of tumour
grading of adenocarcinoma and mucinous adenocarcinoma
into ‘well’, ‘moderately’, and ‘poorly’ differentiated neo-
plasms provides prognostic information in univariate analy-
sis,4,14,74-77 and in some multivariate s t ~ d i e s , bu’ t~ ~ ~ ~ ~ ~
not in other^.^',^^,^^ This assessment is subjective, and al-
though certain levels of reproducibility have been achieved,
this feature is open to very wide interobserver
Thus there needs to be substantial reduction in observer
variation before this feature can be fully utilized in multi-
centre studies.
Information of probable prognostic significance
(Appendix 11: Items 16-1 9)
Tumour perforation, and the tumour status of the distal
‘doughnut’ of tissue removed from a staple machine, are
both likely to be of some prognostic significance. Further-
more, the presence of a conspicuous mixed inflammatory
cell infiltrate influences prognosis: a distinctive and delicate
connective tissue mantle with lymphocytes and other cells
on the deepest point of invasion confers an improved prog-
nosis, whereas the absence of it with peritumoral fibrosis
(desmoplastic reaction) and no significant cellular infiltrate
affects survival a d v e r ~ e l y .In~ addition,
~ ~ ~ ~ >the
~ ~presence of
lymphoid aggregates in the normal surrounding tissue is of
some prognostic benefit.83Multivariate analyses have shown
that these last two features have an independent association
with but assessment is subjective and reproduc-
ibility may be a problem. Therefore, the Working Party felt
it wise to list these features as of probable prognostic
significance because they have not yet been assessed under
conditions of general use.
Features nor included in Appendix 1-111
Clinical information
The overall general health of a patient clearly influences
survival and with increasing age this factor becomes more
relevant, particularly in patients with distant metastasis, and
those in whom treatment has been unable to achieve removal
of all macroscopic tumour. The presence of comorbid dis-
ease (diabetes, cardiac or pulmonary disease, substantial
recent weight loss, or the prescription of steroids or other
immunosuppressants, etc.) and differences in functional
status will thus influence outcome. Although the American
Joint Committee on Cancer (AJCC) scale is preferred (the
alternatives being the Eastern Co-operative Oncology Group
(ECOG) scale, or the Karnofsky ~ c a l e ~ ’ an ~ independent
~~~~~),
effect on prognosis has not been demonstrated in mul-
tivariate analyses. Thus, a majority of the Working Party
considered that this item would represent too detailed a
listing for general use, although being a suitable and neces-
sary item for some prospective randomized trials.
The recording of cancer-related diseases in the patient’s
parents or siblings was felt to be of research rather than of
clinical relevance, and similarly, we chose not to record
information on the duration of symptoms.
 Colorectal cancer staging
The Working Party considered including a section on the
pre-operative clinical evaluation of patients with rectal can-
cer with special reference to clinical and imaging methods to
determine pre-operatively the depth of tumour invasion.
However, we felt that pretherapeutic or clinical evaluation
alone was beyond our brief. Similarly, various methods for
clinically evaluating the colon (flexible sigmoidoscopy,
colonoscopy, and barium enema) were not included.
We debated the need to include the ‘origin’of information
because of differences in the sensitivity and specificity of
each test from which data is derived. For example, the
evaluation of the liver and a statement that this organ has or
has not liver metastasis depends on the method chosen.
Thus, liver function tests, simple manual palpation at
time of surgery, computerized tomography scan, magnetic
resonance imaging scan, transcutaneous ultrasound, or
intraoperative ultrasound will identify patients with liver
metastasis to substantially different rates. Although this is
an important subject in relation to the reliability of informa-
tion data gathering,86 the Working Party felt that it would
be too cumbersome to include a more detailed description to
identify the ‘origin’ of data.
There is some evidence which suggests that the use of
blood during the perioperative period has a negative impact
on survival. However, the evidence remains c o n t r ~ v e r s i a l , ~ ~ - ~spread in CRC, because the most important predictors of
and the Working Party considered that this item would be
more appropriately collected as a part of research protocol.
Pathology factors
The time-honoured macroscopic description of tumours
(polypoid, ulcerating, plaque-like, fungating, or stenotic)
does not provide independent prognostic significance and
therefore has not been included in our recommendations.
The relation of a distal tumour to the pelvic peritoneal
reflection was not included because independent prognostic
significance is not sufficiently strong, although this feature
may influence certain technical aspects of treatment.
Small vessel and perineural invasion are not widely
accepted as useful prognostic variables. Nevertheless, lym-
phatic vessel invasion has been shown to be an independent
prognostic factor,69 and to reduce survival significantly
within specific substage^.^' Similarly, perineural invasion
has been put forward as an independent variable.68However,
both these assessments are subjective and time consuming,
and therefore the Working Party recommends that these
features not be routinely reported.
Possible future prognostic factors
The measurement of carcino-embryonic antigen (CEA) is
common in clinical practice today. A raised CEA level may
indicate the presence of metastatic disease at the time of
presentation. If normal, or becoming so after surgery, an
increasing CEA level may indicate the onset of metastatic
disease some 4-8 months before these lesions become clini-
cally evident. However, the Working Party felt that there
was insufficient evidence at present to class CEA level as an
independent prognostic factor.
Measurement of the DNA content of malignant nuclei
may be achieved by flow cytometry, and on this basis,
333
tumours may be classified as diploid or aneuploid. While
most univariate studies demonstrate poorer survival when
cancers are aneuploid, it appears that any independent effect
is weak if it exists at a11,73291and there are no grounds
to undertake DNA flow cytometry on a routine basis.
Immunohistochemical techniques have been used to show
loss of differentiation antigens and the acquisition of tumour
associated.antigens. Some of the findings have been l i k e d
with prognosis, but no important independent effects have
been demonstrated to date. This may, however, change with
the increasing availability of monoclonal antibodies to
oncogene products. Molecular biological techniques for
demonstrating loss of genetic material (e.g. anti-oncogenes
on Chromosomes 17 and 18) offer the promise of a new
understanding of tumour behaviour and may alter our ap-
proach to tumour classification in the future.92
Appendix I11 A: International Comprehensive
Anatomical Terminology (ICAT) for Colorectal
Cancer (CRC)
The Working Party recognized that a categorizationis needed
to describe the details of the anatomical extent of tumour
outcome are the anatomical extent of tumour at the time of
definitive treatment, and the residual tumour status following
therapy. 14318256Thus, we developed a nomenclature (a subset
of the data collected in the IDS) which has been called the
‘International Comprehensive Anatomical Terminology ( I C AT )
for Colorectal Cancer (CRC)’. The principal features of this
terminology are grouped into four areas: (i) microscopic
description of tumour depth; (ii) regional lymph node status
(numbers of lymph nodes with and without tumour; status
of nodes on vascular trunks; and apical node status); (iii)
distant metastasis status before definitive treatment; (iv)
residual tumour status after definitive treatment (Table 2
and Appendix I11 A).
There is a close correlation between anatomical extent of
tumour and residual tumour status. When the primary
tumour does not penetrate the muscularis propria, complete
removal is usual when appropriate surgical techniques are
used. Residual tumour in the area of a surgically excised
specimen is only expected when the lesion has invaded the
pericolic or perirectal tissues.
Agreement on the definition of the regional lymph nodes
is important in staging. In the 1987 UICCI9 and the 1988
AJCC recommendations” for colorectal carcinoma, there is
a separation between pericolic/perirectal lymph nodes and
those located along the course of a named vascular trunk,
that is along the ileocolic, right colic, middle colic, left colic,
inferior mesenteric and superior rectal arteries. Other
studies recommend a separate examination of apical nodes,
that is nodes at the highest surgical ligature on the specimen.
Para-aortic, other abdominal, pelvic and extra-abdominal
nodes are considered a form of ‘distant metastasis’.
Distant metastases may arise by a number of routes:
transvenously into the portal circulation (low rectal carcinoma
may spread into the systemic circulation directly); tumour
334
spreading on a free serosal surface of the specimen; and
through the lymphatic system. Statistically proven correlations
exisr between these three types of tumour spread: regional
lymph node metastases are increasingly observed with progres-
sive local tumour invasion; distant metastases become more
frequent with increasing depth of primary tumour invasion
and also with increasing lymph node i n v ~ l v e m e n t . ’ ~ ~ ~ ~The
Occasionally it is possible to completely remove a meta-
stasis (determined by clear tumour margins on histology),
rendering the patient ‘tumour free’, which may be associated
with long-term survival. However, this is only possible in a
small minority of patients with distant metastases.
Detection and significance of residual tumour
None of the staging systems in use before 1967 addressed the
issue of separately identifying patients who had tumour(s)
which had been incompletely fesected. The initiative in this
area came, in 1967, from Turnbull” who introduced a
‘Stage D’ category for tumours which were advanced at the
time of bowel resection, and included patients with clinically
demonstrable distant metastases, irremovable tumour be-
cause of ‘parietal’ invasion and adjacent organ attachment.
It should be noted in this context that Turnbull has been
criticized for the inclusion in a ‘D’ category of patients with
adjacent organ invasion, because it is well known that not all
these patients are incurable. However, it should be remem-
bered that Turnbull was comparing the potential impact of
his ‘no-touch’ technique with ‘conventional’ tumour
mobilization and wished to define a cohort of patients in
which there was a single vascular bed involved in the blood
supply to and from the t ~ m o u r . ~In ’ a later study the
definition of a ‘D stage’ was restricted to include only
tumours thought, at the time of bowel resection, to be
‘incurable’.l4
Thus, ‘incurable’ tumours can be defined as: ‘those neo-
plasms with distant metastases (histologically proven, or
with convincing clinical evidence) and those with tumour
demonstrable histologically in a line of resection (almost
always in a lateral (or deep) resection margin).This definition
recognizes the difficulty, in some cases, of distinguishing
between tumour and inflammatory tissue at the site of
attachment to adjacent structures. It should be noted that
using this definition, adjacent organ invasion does mt qualify a a tumour into six of 28 line items, it is possible to construct
tumour for the ‘incurable’ classification.
When ‘incurability’is defined as above, about one quarter
of tumours treated by bowel resection fall into this category.
In the Newland study,“ 23% of patients having bowel
resection were so designated with a corrected five-year
survival of 27%. Twenty-six per cent of these patients, that
is 6% of all resected cases, were so classified only because
tumour was demonstrated histologically in a line of resec-
tion. The survival of this group exceeded those with distant
metastases, however the difference was not statistically sig-
nificant (P = 0.3391). In another seriess6in which similar
definitions were used, patients with local residual tumour
had a significantly longer survival than those with distant
metastases (14.4 versus 7.9 months) The difference between
this and the Newland study may, in some measure, be
L. P. Fielding et al.
attributed to the use of postoperative radiotherapy in 30% of
patients with local residual tumour (Hermanek, 1990, pers.
comm.). Consistent with these findings, in an additional
study, is the high incidence of local tumour recurrence in
patients in whom tumour was demonstrated histologically
in a line of resection.”
~ ~ ~ and the AJCC have offered an ‘R’ classification
~ ~ UICC
to address the problem of ‘residual’ t u m ~ u rThe . ~ defini-
~ ~ ~ ~
tions are: RO: no residual tumour; R1:microscopic residual
tumour; R2: macroscopic residual tumour; RX:pre,senceof
residual tumour cannot be assessed. However, in the pTNM
system this R classification is optional. Current pTNM
stages therefore do not enable separate categorization of
tumours which have been transected during their removal.
However, by the combined use of the pTNM with the R
classification, the full objectives of clinicopathological stag-
ing can be achieved as already provided for in the Concord
Hospital system and in the Erlangen Prognostic Groups.
The concept of identifying and segregating patients with
‘residual tumour’ at the time of bowel resection, particularly
at the local site, may be regarded as the most significant
advance in staging since Dukes described his classification
of rectal carcinomas in 1932.4 Thus the Working Party
considers that information on the presence of residual
tumour at the time of definitive treatment both at the local
site of tumour resection as well as distant sites should be
recorded in every case, and hence these features have been
included in the proposed International Comprehensive Ana-
tomical Terminology.
Appendix I11 B: Matrix for staging system
conversion
The six most commonly used staging systems for colorectal
cancer differ with respect to consideration of distant meta-
stasis and of residual tumour status. The basic criteria for
describing the depth of invasion and the lymphatic spread
are very similar. However the allocation to the individual
stages is different. Thus, although there is a general trend in
which the earlier the stage number (0-V), or letter (A-D),
the better the prognosis, the stages are not comparable
directly.
Nevertheless, by clearly defining the anatomical extent of
any and all of these six staging systems. This form of
‘matrix’ for deriving different staging systems is highly
suited to computation, so that with the entry of information
in the six areas defined in Table2 (Appendix I11 A), a
software package could then undertake the conversion
(Appendix I11 B) to the staging system required by the
clinician, institution, or accrediting authority. Such a soft-
ware package is currently in Beta test version and will be
available in 1991.
Follow-up information
In many trials the follow-up data are very involved and
complex. Although this complexity may be needed in the
context of certain randomized studies, we recommend that
Colmectal cancer staging
such information be kept simple, with data being recorded
concerning four features; (i) follow-up status and date when
patient was last seen; (ii) if the patient is alive, condition at
last follow-up; (iii) additional treatments; (iv) identification
of new primary tumour(s) (Table 3). It may be useful to
identify the dates at which each of these features was ‘first
suspected’ and then ‘confirmed’. Histological definition of
these items is preferable, but unequivocal clinical evidence
is frequently the basis for the definition.
DISCUSSION
The first objective of the Working Party was to identify
those features of clinical information and histopathology
analysis which should be recorded in all cases of large bowel
carcinoma. The presented International Documentation
System (IDS) contains basic patient information, variables
of proven prognostic significance and information of prob-
able prognostic significance. IDS is subdivided into clinical
(pre- and post-treatment) and pathology data elements.
Comparisons of treatment results between institutions are
often impossible because the documentation systems are
different. Studies on prognosis of CRC show different results
because the factors analysed are different. Thus, the recom-
mended IDS would be a first step to a uniform international
documentation reporting system for this tumour leading
to standardization of biometric analysis of outcome in
colorectal carcinoma.
In the selection of features to be included in IDS, inde-
pendent prognostic significance, as evident from multi-
variate studies, was decisive. Features that can be assessed
only by time- and cost-consuming methods, as well as those
with a low level of reproducibility, were not included.
The Working Party recognized the need to clarify the
similarities and differences between commonly used staging
systems. This required careful definition of both pathology
and clinical features in these systems and clarification of the
way these components are combined to define tumour stages.
Three principal areas of difficulty were identified with
commonly used staging systems: first, the description of
superficial neoplastic change; second, the involvement of
adjacent organ($ and invasion of a free serosal surface; and
third, the identification and description of residual tumour
remaining after definitive treatment both locally beyond the
confines of the resected specimen and at distant site(s).
We concluded that those histological lesions determined
to be part of superficial neoplastic change (severe dysplasia,
carcinoma in situ, and superficial mucosal cancer) should be
categorized into a single entity because their biological
behaviour is not distinguishable.
By contrast the Working Party recommends that we
should distinguish between ‘adjacent organ invasion’ and
‘free serosal surface invasion’ because there are data demon-
strating a worse prognosis in the latter.I6
Finally, recognition that a tumour has spread beyond the
confines of a resection specimen andor remaining regional
or distant metastases are both very powerful prognostic
statements, which need to be accommodated in staging
335
systems. If these recommendations are accepted, then the
pTNM nomenclature will require the obligatory addition of
the residual tumour (R) classification to fully characterize
tumour prognosis. This objective can also be achieved by
the Cohcord Hospital and by the Australian Cfinicopatho-
logical Staging Systems (Appendix I11 B). However, this
objective is not achieved by the Dukes system (or its modifi-
cations), or by the Japanese Research Society Staging system,
or the UICC/AJCC pTNM system without the additional
residual (R) classification.
The Working Party concluded after reviewing the six
currently used methods that it would not be of value to
generate a further staging system. Those using each classifi-
cation have become very familiar with the details of their
respective methods, and the nomenclature has become so
cluttered with alphanumeric abbreviations, that to suggest a
further staging system appeared counterproductive. Further-
more, in the United States, the Commission on Cancer and
the Joint Commission on Accreditation of Hospital Organi-
zations (JCAHO) are working towards the introduction of
the pTNM staging system for institutional accreditation (Dr
Robert Hutter, pers. comm.).
For all these reasons, the Working Party identified as a
major objective the establishment of a terminology which
would enable users to derive any or all staging systems
according to the needs of the clinician, institution, or
accrediting authority.
The Working Party recommends that any case of CRC
after treatment should be described according to ICAT. The
specific ICAT findings, rather than summarizing stage
grouping, should be the basis for estimation of prognosis
and evaluation of treatment results. However, the assign-
ment to the stages or substages of the various systems may
be preferable for short data surveys and may be important
for institutions with smaller series of patients and for com-
parison with historical data.
Looking to the future, it was also apparent that there is a
growing body of additional clinical and pathology informa-
tion of predictive value beyond that derived from a compre-
hensive review of anatomical extent of tumour spread.
Thus, the integration of independent nonanatomic factors of
prognostic significance with clinicopathological staging
information leading to the development of a prognostic
index (PI) for colorectal carcinoma, is the next development
in tumour c~assification.~~’~~
The Working Party wishes to emphasize that to derive the
required accurate descriptive data for a comprehensive ana-
tomical description of colorectal cancer spread, as well as the
additional items needed for a prognostic index, requires
collaboration between surgeons and pathologists. The
achievement of a comprehensive view of tumour spread and
tumour prognosis is not possible without such co-operation.
RECOMMENDATIONS
(1) The Working Party has identified a minimal list of
clinical and pathology features which should be docu-
mented in prospective studies, and we recommend
336 L. P. Fieldinget al.

the ‘International Documentation System (IDS) for
CRC‘ for general use (Table 1; Appendices I and 11).
(2) The Working Party has demonstrated the need for a
standard terminology to fully describe the spread of
CRC. The newly defined International Comprehen-
sive Anatomical Terminology (ICAT) for colorectal
cancer satisfies this need (Table 2; Appendix 111). We
recommend that ICAT be used for world-wide docu-
mentation of patients with colorectal carcinoma from
which the six most commonly used staging systems
can be derived.
ACKNOWLEDGEMENTS
The Working Party was sponsored by the World Congresses
of Gastroenterology, Digestive Endoscopy and Coloproc-
tology (WCOGDEC), with an additional generous grant
from the Australian Cancer Society. The Working Party
gratefully acknowledges the assistance of Dr Robert Hutter
and Dr Stanley Goldberg.
REFERENCES
1. CL~GGS H. S . Cancer of the colon: A study of 72 cases. Lancet
1908; ii: 1007-12.
2. MILESW. E. The radical abdomino-perineal operation for
cancer of the rectum and the pelvic colon. BM3 1910; 2: 941-3.
3. LOCKHART-MUMMERY J. P. Two hundred cases of cancer of
the rectum treated by perheal excision. Brit. 3. Surg. 19261
1927; 14: 110-24.
4. DUKESC. E. The classifrcation of cancer of the rectum. 3.
Pathol. Bactm’ol., 1932; 35: 323-32.
5. GABRIELW. B., DUKESC. E. & BUSSEY H. J. R. Lymphatic
spread in cancer of the rectum.Brit. 3 . Swg. 1935; 23: 395-413.
6. SIMPSON W. C. & MAYOC. W. The mural penetration of the
carcinoma cell in the colon: Anatomic and clinical study.
Surg. Gynecol. Obster. 1939; 68: 872-7.
7. DUKESC. E. & BUSSEY H. J. R. The spread of rectal cancer
and its effect on prognosis. Brit. 3 . Cancer 1958; 12: 309-20.
8. KIRKLINJ. W., DOCKERTY M. B. & WAUGHJ. M. The role
of the peritoneal reflection in the prognosis of carcinoma of
the rectum and sigmoid colon. Surg. Gynecol. Obstet. 1949;
88: 326-31.
9. ASTLERV. B. & COLLERF. A. The prognostic significanceof
direct extension of carcinoma of the colon and rectum. Ann.
Surg. 1954; 139: 846-52.
10. GUNDERSON L. L. & SOSINH. Areas of failure found at re-
operation (second or symptomatic look) following ‘curative
surgery’ for adenocarcinoma of the rectum. Cancer 1974; 34:
1278-92.
11. TURNBULL R. B., KYLEK., WATSONF. R. & SPRAIT J.
Cancer of the colon: The influence of the no-touch isolation
technique on survival rates. Ann. Surg. 1967; 166: 420-7.
12. PHEILS M. T., BARNETT J. E., NEWLAND R. C. & MACPHER-
SON J. G. Colorectdl carcinoma: A prospective clinicopath-
ological study. Med. J . Aust. Spec. Suppl. 1976; 1: 17-21.
13. CHAPUIS P. H. & PHEILSM. T. Report of a multidisciplinary
research programme for colorectal cancer. Anticancer Research
1981; 1: 15-8.
14. NEWLAND R. C., CHAPUIS P. H., PHEILSM. T. & MACPHER-
SON J. G. The relationship of survival to staging and grading
of colorectal carcinoma. A prospective study of 503 cases.
Cancer 1981; 47: 1424-9.
15. DAVISN. C. & NEWLAND R. C. Terminology and classifica-
tion of colorectal carcinoma: The Australian clinico-patho-
logical staging system. Aust. N.Z. J. Surg. 1983; 53: 211-21.
16. NEWLAND R. C., CHAPUIS P. H. & SMYTHE. J. The prognos-
tic value of substaging colorectal carcinoma: A prospective
study of 1117 cases with standardised pathology. Cancer 1987;
60: 852-7.
17. CHAPUIS P. H.,DENT0. F. & FISHER R. et al. A multivariate
analysis of clinical and pathological variables in prognosis
after resection of large bowel cancer. Brit. J . Surg. 1985; 72:
698-702.
18. HERMANEK P., GALLF. P. & ALTENDORFA. Prognostic
grouping in colorectal carcinoma. J . Cancer Res. Clin. Oncol.
1980; 98: 185-93.
19. HERMANEK P. & SOBINL. H., eds, UICC TNM Classifica-
tion of Malignant Tumours, 4th edn, Springer-Verlag, Berlin,
1987.
20. BEAHRS0. H., HENSOND. E., HUTTERR. V. P. & MYERS
M. H., eds, A3CC Manual for Staging of Cancer, 3rd edn.
Lippincott, Philadelphia, 1988.
21. JINNAID. General rules for clinical and pathological studies
on cancer of the colon, rectum, and anus; Part I: Clinical
classification.Jap. J . Surg. 1983; 13: 557-73.
22. FIELDING L. P. Clinical pathologic staging of large bowel
cancer: A report of the ASCRS Committee. Dis. Colon Rectum
1988; 31: 204-9.
23. WILLIAMS N. S., JASSJ. R., HARDCASTLE J. D. Clinicopath-
ological assessment and staging of colorectal cancer. Brit. 3.
Surg. 1988; 75: 649-52.
24. LEE E. T. Statistical Methods for Survival Data Analysis.
Lifetime Learning Publications, Belmont, California, 1980.
25. BERKSONJ., GAGER. P. Calculation of survival rates for
cancer. Proc. Staff Meet. Mayo Clin. 1950; 25: 270-86.
26. CUTLERS. J. & EDERERF. Maximum utilization of the life
table in analyzing survival. 3. Chronic Dis. 1958; 8: 699-712.
27. KAPLANE. L. & MEIERP. Nonparametric estimation from
incomplete observations. J. Am. Stat. Assoc. 1958; 53: 457-81.
28. COX D. R. Regression models and life tables. J . Roy. Stat.
SOC.Series B 1972; 34: 187-220.
29. BAILEYN. J. T. The Mathematical Approach to Biology and
Medicine. Wiley, London, 1967, pp. 244,250.
30. ARMITACE P. Statistical Methods in Medical Research. Black-
well Scientific Publications, Oxford, 1973, pp. 78-81, 144-6.
31. KNILL-JONES R. P., STERNR. B., TIRMES D. H., MAXWELL
J. D., THOMPSON R. P. H. & WILLIAMS R. Use of sequential
Bayesian model in diagnosis of jaundice by computer. BMJ
1973; 1: 530-4.
32. JENNEITB., TEASDALE G., BRAAKMAN R., MINDERHOUD J.
& KNILL-JONES R. Predicting outcome in individual patients
after severe head injury. Lancer 1976; i: 1031-4.
33. FIELDING L. P., PHILLIPSR. K. S., FRYJ. S. & HITTINGER
R. Prediction of outcome after curative resection for large
bowel cancer. Lancet 1986; ii: 904-7.
34. SILVERMAN D. T., MURRAYJ. L., SMARTC. R., BROWN
C. C. & MYERSM. H. Estimated mean survival times of
patients with colorectal cancer based on experience with 9745
patients. Amer. 3. Surg. 1977; 133: 289-97.
35. NICHOLLS R. J., YORK MASONA., MORSONB. C., DIXON
A. K. & KELSEY-FRY I. The clinical staging of rectal cancer.
Brit. 3. Surg. 1982; 69: 404-9.
36. FIELDING L. P., STEWART-BROWN S. & DUDLEYH. A.
Surgeon-related variables and the clinical trial. Lancet 1978;
ii: 778-9.
Colorectal cancer staging 337

37. HARDCASTLE J. D., CHAMBERLAIN J. & SHEFFIELD J. Ran- technique for the detection of lymph nodes in colorectal
domised, conyolled trial of faecal occult blood screening for cancer. Brit. 3.Surg. 1986; 73: 58-60.
colorectal cancer. Lancet 1989; i: 1160-4. 59. JASSJ. R., MILLERK. & NORTHOVER J. M.A. Fat clearance
38. KRONBORG O., FENGER C. & OLSENJ. Interim report on a method versus manual dissection of lymph nodes in speci-
randomized trial of screening for colorectal cancer with mens of rectal cancer. Int. J . Colorect. Dis. 1986; 1: 155-6.
hemoccult-11. In: Millar A. B., Chamberlain J., Day N. E., 60. SCOTTK. W. M. & GRACER. H. Detection of lymph node
Hakama M. & Prorok P., eds, Screening for Cancer. Cam- metastases in colorectal carcinoma before and after fat clear-
bridge University Press, 1991. ance. Brit. J . Surg. 1989; 76: 1165-7.
39. FREEDMAN L. S., MACASKILL P. & SMITHA. N. Multivariate 61. HERMANEK P. & ALTENDORF A. Classification of colorectal
analysis of prognostic factors for operable rectal cancer. Lan- carc&omas with regional lymphatic metastases. Path. Res.
cet 1984; ii: 733-6. Pract. 1981; 173: 1-11.
40. WIGGERST., JEEKELJ., ARENDSJ. W. et al. No-touch 62. HUTTERR. V. P. & SOBINL. H. Universal staging system for
isolation technique in colon cancer: A controlled prospective colorectal cancer: In reply. Arch. Pathol. Labor. Med. 1987;
trial. Brit. J . Surg. 1988; 75: 409-15. 111: 312-4.
41. FIELDING L. P. The portal vein and colorectal cancer. Brit. J . 63. ONDEROH., MAETANIS., NISHIKAWA T. & TOBET. The
Surg. 1988; 75: 402-3. reappraisal of prognostic classifications for colorectal cancer.
42. SLANETZ C. A. J. R. The effect of inadvertent intra-operative Dis. Colon Rectum 1989; 32: 609-14.
perforation on survival and recurrence in colorectal cancer. 64. PHILLIPSR. K. S., HITTINGER R., BLESOVSKY L., FRYJ. S.
Dis. Colon Rectum 1984; 27: 792-7. & FIELDING L. P. Large bowel cancer: Surgical pathology
43. RANBARGER K. R., JOHNSON W. D. & CHANG J. C. Prognos- and its relationship to survival. Brit. J. Surg. 1984; 71: 604-10.
tic Significance of surgical perforation of the rectum during 65. SCHEELE J., STANGL R. & ALTENDORF-HOFMANN A. Hepatic
abdominoperineal resection for rectal carcinoma. Amer. J. metastases from colorectal carcinoma: Impact of surgical
Surg. 1982; 143: 186-8. resection on the natural history. Brit. J. Surg. 1990; 77: 1241-6.
44. ZIRNGIBL H., HUSEMANN B. & HERMANEK P. Intra-operative 66. FORTNER J. G., MACLEANB. J., KIM D. K. et al. The
spillage of tumor cells in surgery for rectal cancer. Dis. Colon seventies evolution in liver surgery for cancer. Cancer 1981;
Rectum 1990; 33: 610-4. 47: 2162-6.
45. JASSJ. R. & SOBINL. H. Histological typing of intestinal 67. FORTNER J. G., SILVAJ. S., GOLDENR. B., Cox E. B. &
tumours (2nd ed.) W H O International Histological Classifica- MACLEAN B. J. Multivariate analysis of a personal series of
tion of Tumours. Springer-Verlag, Berlin, 1989. 247 patients with liver metastases from colorectal cancer.
46. JASSJ. R., LOVES. B. & NORTHOVER J. M. A. A new Ann. Surg. 1984; 199: 306-17.
prognostic classifcation of rectal cancer. Lancet 1987;i: 1303-5. 68. KNUDSEN J. B., NILSSON T., SPRECHLER M., JOHANSEN A. &
47. QIZILBASH A. H. Pathologic studies in colorectal cancer: A CHRISTENSEN N. Venous and nerve invasion as prognostic
guide to the surgical pathology examination of colorectal factors in postoperative survival of patients with resectable
specimens and review of features of prognostic significance. cancer of the rectum. Dis. Colon Rectum 1983; 26: 613-7.
Pathol. Ann. 1982; 17: 1-46. 69. MINSKY B. D., MIESC., RICHT. A. & TECHT A. Lymphatic
48. CHANK. W., BOEYJ. & WONGS. K. C. A method of vessel invasion is an independent prognostic factor for sur-
reporting radial invasion and surgical clearance of rectal vival in colorectal cancer. Int. J. Rad. Oncol. 1989; 17:
carcinoma. Histopathologv 1985; 9: 1319-27. 311-8.
49. DUKESC. E. The pathology of rectal cancer. In: Dukes C. E., 70. TALBOT I. C., RITCHIES., LEIGHTON M. H., HUGHES A. O.,
ed., Cancer of the Rectum. Livingstone, Edinburgh and BUSSEY H. J. R. & MORSON B. C. The clinical significance of
London, 1960. invasion of veins by rectal cancer. Brit. J. Surg. 1980; 67:
50. HEALDR. J., HUSBAND E. M. & RYALLR. D. H. The 439-42.
mesorectum in rectal cancer surgery: The clue to pelvic 71. HERMANEK P., GUGGENMOOS-HOLZMANN I. & GALLF. P.
recurrence? Brit. 3.Surg. 1982; 69: 613-6. Prognostic factors in rectal carcinoma: A contribution to the
51. QUIRKEP., DURDEYP., DIXONM. F. & WILLIAMS N. S. further development of tumor classification. Dis. Colon Rec-
Local recurrence of rectal adenocarcinoma due to inadequate tum 1989; 32: 593-9.
surgical resection: Histopathological study of lateral tumour 72. SCOTT N. A., RAINWATER L. M., WIELAND H. S. et al. The
spread and surgical excision. Lancet 1986; ii: 996-9. relative prognostic value of flow cytometric DNA analysis and
52. QUIRKE P. & DIXONM. F. The prediction of local recurrence conventional clinicopathologic criteria in patients with opera-
in rectal adenocarcinoma by histopathological examination. tive rectal carcinoma. Dis. Colon Rectum 1987; 30: 513-20.
Int.J. Cohect. Dis. 1988; 3: 127-31. 73. QUIRKE P., DIXONM. F., CLAYDEN A. D., et al. Prognostic
53. HERMANEK P., GIEDLJ. & DWORAK 0. Two programs for significance of DNA aneuploidy and cell proliferation in
examination of regional lymph nodes in colorectal carcinoma rectal adenocarcinomas.J . Pathol. 1987; 151: 285-91.
with regard to the new pN classification. Path. Res. Pract. 74. DUKESC. E. Histological grading of rectal cancer. Proc. R.
1989; 185: 867-73. SOC.Med. 1936; 30: 371-6.
54. Symposium staging of colorectal cancer. Int. 3. Colorect. Dis. 75. GRINELL R. S. The grading and prognosis of carcinoma of the
1987; 2: 123-38. colon and rectum. Ann. Surg. 1939; 109: 500-33.
5 5 . WIGGERS T., ARENDSJ. W.,SCHUTTEB., VOLOVICS L. & 76. ALDRIDGEM. C., PHILLIPS R. K. S., FRYJ. S. & FIELDING
BOSMAN F. T. A multivariate analysis of pathologic prognos- L. P. Influence of tumour site on presentation, management
tic indicators in large bowel cancer. Cancer 1988; 61: 386-95. and subsequent outcome in large bowel cancer. Brit. J . Surg.
56. HERMANEK P. Colorectal carcinoma. Histopathological diag- 1986; 73: 663-70.
nosis and staging. Clin. Gastroenterology 1989; 3: 5 11-29. 77. GRIFFINM. R., BERGSTRALH E. J., COFFEYR. J., BEART
57. PICKREN J. W. Nodal clearance and detection. J A M A 1975; R. W. & MELTONL. J. Predictors of survival after curative
231: 969-71. resection of carcinoma of the colon and rectum. Cancer 1987;
58. CAWTHORN S. J., GIBBSN. M. & MARKSC. G. Clearance 60:2318-24.
338 L. P. Fielding et al.

78. STEINBERG
S. M., BARWICK
K. W. & STABLEIND. M. cancer. N. Engl. 3.Med. 1985; 312: 1604-8.
Importance of tumor pathology and morphology in patients 87. FIELDING L. P. Red for danger: Blood transfusion and
with surgically resected colon cancer: Findings from the colorectal cancer. BM3 1985; 2: 841-2.
GasuointestinalTumor Study Group. Cmer 1986; 58: 1340-5. 88. VOOGTP. J., VAN DE VELDE c. J. H., BRAND A. et al. Peri-
79. JASSJ. R., ATKIN W. S., CUZICK J. et al. The grading of operative blood transfusion and cancer prognosis. Cancer
rectal cancer: Historical perspectivesand a multivariateanaly- 1987; 59: 836-43.
sis of 447 cases. HistopathologV 1986; 10: 437-59. 89. CROWSON M. C., HALLISSEY M. T., KIFF R. S., KINGSTON
80. BLENKINSOPP W. K., STEWART-BROWN S., BLESOVSKY L., R. D. & FIELDING J. W.L. Blood transfusion in colorectal
KEARNEY G., & FIELDING L. P. Histopathology reporting in cancer. Brit. 3. Surg. 1989; 76: 522-3.
large bowel cancer.3. Clin. Pathol. 1981; 34:509-13. 90. VENTEJ. P., WICGERS R.H., WEIDEMAN W. F., JECKEL J. &
81. THOMAS G. D. H., DIXONM. F., SMEETONN. G. & OBERTOP H. Peri-operative blood transfusion in colorectal
WILLIAMS N. S. Observer variation in the histological grading cancer. Eur. 3. Surg. Oncol. 1989; 15: 371-4.
of rectal carcinoma. 3. Clin. Pathol. 1983; 36: 385-91. 91. JASS J. R., MUKAWA K., GOH H. S., LOVES. B. &
82. ]ASS J . R. Lymphocytic infiltration and survival in rectal CAPELLARO D. Clinical importance of DNA content in rectal
cancer. 3. Clin. Pathol. 1986; 39: 585-9. cancer measured by flow cytometry.3. Clan. Pathol. 1989; 42,
83. HALVORSEN T. B. & SEIME. Association between invasive- 254-9.
ness, inflammatory reaction, desmoplasir and survival in 92. SOLOMON E. Colorectal cancer genes. Nature 1990; 343: 412-4.
colorectal cancer. 3. Clin. Pathol. 1989; 42: 162-6. 93. DIONNE L. The pattern of blood-borne metastasis from carci-
84. KARNOPSKY D. A. & BURCHENAL J. H. The clinical evalua- noma of rectum. Cancer 1965; 18: 775-81.
tion of chemotherapeutic agents in cancer. In: MacLeod 94. HERMANEK P. & KARRERK. Atlas of Colorectal Tumors.
C. M., ed., Chemotherapy of Cancer. Columbia University English edition edited by Sobin L. H., Butterworth Scientific,
Press, New York. 1949; pp. 191-205. London, 1983.
85. DAVIES R. J. Staging in cancer. In: Moossa A. R., Robson 95. FIELDING L. P. & BALLANTYNE G. H. Classification systems
M. C. & Schimpff S . C., eds, Comprehensive Textbook of for staging colorectal cancer. Problems Gen.Surg. 1987; 4(1):
Oncologv. Williams & Wilkins, Baltimore, 1986, p. 67. 39-53.
86. FEINSTEIN A. R., SOSIND. M. & WELLSC. K. The Will P., HUTTERR. V. P., SOBINL. H. Prognostic
96. HERMANEK
Rogers Phenomenon: Stage migration and new diagnostic grouping: The next step in tumor classification. 3. Cancer
techniques as a source of misleading statistics for survival in Res. Clin. Oncol. 1990; 116: 513-6.
APPENDIX I . C l i n i c a l f e a t u r e s A. P r e t r e a t m e n t

Item
Alternative first level Alternative second level Definitionslcomments
x Name

Basic p a t i e n t i n f o r m a t i o n
1 CounW Comment: Basic patient information for items 1-5

2 Hospital
(Name/&)

3 Patient identification
(Namelcode)

-Ahin
- Asian
- Caucasian
-Other
5 Pasthitory - Cobrectal cardnoma -NolYes Comment: Past b t o r y of t u m r may aftect long-term outcome.
~ Other m a l i t m u r -NolYes

Data of proven prognostic significance

6 Surgeonidenljficakw Comment: 'Surgeon' is a pmgroszC facbx.
(kdccw

7 Gender - Male I Female

8 Dateofbirth I~~/mfWearl Comment: sunrival is influenced by a p at presem~h~.

9 Uicalpresmtalion - Aqnmpto~tic - Populationsaeening program
- Eady detecbrVsuweiIlance
Fogram
- Coincidentalfinding during
investigationlor a n o h
disease
Comment: The rm(rivarialeanalyses conducted M far appeal
to indude only patients with syrrpmmatic bmou.
For lead-time bias. we should identify hese
asymptomaticpahnts.

- Symptomlk elective
- Symptomatic obstruction - To d ionly Comment: Bowel obstruction is an independentpmgmstic factor
-To d iand gas in several studies. There is TY) e s t a M i
- TOdi.g a @US ~ classificationk r 'severity of obsmction.' and thii
systemic effects is an arbitrary classification. It is likely that when
obstruction becomes severe. it is the added systemic
- Symptomatic pertoration efffecs which inducea negative effect on long-term
(* obsbuction) progncsis: hence thii classilkation.
D a t a of p r o b a b l e p r o g n o s t i c s l g n i f l c a n c e
-
10 Preoperativetreabmnt - Radiothew -NolYeS
- chemotharapylocal -NolYes
- chemtharapysystemic -NolYes

11 Anatomic site of - Appendix

tumr -cecum
- Asxnding colon
- Hepatic flexure Definition: The rectum is definedanatomically as the d istal large
- Transversa colon bowel commencing opposite the sacral promontory
- splenic Raxue and ending at the upper border of the anal canal. When
- Descending colon measuredfrom below with a rigid signmidoscope.
- S i i d colon the upper limit is 16cm from the anal verge. Ifhe lower
-Fkcm -em from anal verge margin of a lumour lies at or within 16an of the anal
- Colon Nos verge it is defined as a "rectal turnour' w
w
\o
APPENDIX I . C l i n i c a l f e a t u r e s 8. P o s t - t r e a t m e n t

Itern Alternative first level Alternativesecond level Definitionslcornrnents

x Name

D a t a of p r o v e n p r o g n o s t i c s i g n i f i c a n c e
12 Definitivetreatment (Day/mon
Wyear] Comment: Tme to outcome is calculatedfrom date of pincipal
star^ date bealment. ( i .surgerykliignosisletc.)

13 7imit-g of surgery - Elective Comment: Emergency pesentabbo at surgery carries a WDOB

- Urgent (<ah) pogmis. This is an arbitrary classification.
-Emergency (&h)

14 Siteofdistant - Canna1ba assesed Comment: The presenceor absence of distant metastasis is the
metastasis pincipal determinantof prognosis. Thii finding should
-No Iumour spread seen indude all available clinical and investigative
infortration.

- Willin abdomen - Distant nodes

- Liver
- PeriDneum
- Oher
-outsideabdomen -Lug
- Bone
- Brain
~ Oher

15 LjVBrstaLs - Right bbe deposih Comment: Rogrmisis rehted to number of depasilo rvhen liver
nun be^ 0 . 1 , 2 . 3 , 4 , 5 + is sole site of metastasis. Ttis infommbboshould indude
- Left bbe depmils peoperaliveinv&gaIims, W ng and operalive
Nunbets 0. 1,2,3.4,5+ Endings.

16 Liver Iumur burden -0% Comment: Clinical estimates alter investigationand operative
- < 25% findings.
- 2560%
- 51-75%
- ,75%

17 M i a 1 Iumour -None
- Locally wnQwus with original
bowel resectiononly
- Distantmetastasisonly
- Both local 8 distant Iumour
Data of p r o b a b l e p r o g n o s t i c s i g n i f i c a n c e

18 Tumwr mobiliIy - Mobile Comment: A clinically 'fixed' Iumour is a pogmstic factor in some
at surgery -Tethered S a d i .
- Fixed

19 Technique of Iumour - Convenional Comment: May influenceincidenceof liver melasrasis.

P
mobilization - Turnbull 'no-touch- s
20 Tumour perforated .None
.Spontaneous
- labogenic
APPENDIX I . C l i n i c a l features 8. Post-treatment, c o n t i n u e d

Item
Alternative first level Alternative second level Delinitionslcomments
: Name

D a t a of p r o b a b l e prognostic significance, cont.

21 Type ofprocedure .Non-resectingsurgery Definition: Surgery w i h u t removal 01 the pimary tumour. eg.
exploratory laparolwnyor bypass operation.

Definition: Local turmur ablation. eg. pdypectomy,disc excision,

e l a c ~ o c o a ~ i oeryotherapy.
n. laser.

- Limitedr e s e c h Definition: Removal01 tumou with or without limited regional

lymphadenectomy.

- Radical resection Definition: Removalof tunmu with formal regionallymphadenectuny.

22 Name of pocedwe ~ Elecrocoagulation

~ Laser ablation
-cryolheraw
- Polypectomy
-S w excision
- Disc excision
- Segmed exckion
- Right hemiictomy
- Tramverse mlectomy
- U i t & transversecolectomy
- Len e d e c m y
- Lana transvecse 0 0 l e ~ t 0 ~
-siaeolecbmy
- High ansrior resection Definition: Rectal resection with inlraabdominalanaslomosis.
- Low atwrior resection Definition: h b m o s i s below peribneal reflection.
- Cob-anal anastomosis
- Abdomitmperineal excision
- Subtotal colectomy Definition: Ueo-zipokl anastomosis.
.Total mlectomy DeWhn: be-mclpl anasbrmsir.
- Proctocolecbmy Definition: Permanent ileostomy 01 i b w n a l ana&mo&.

23 Adjacent organs - Not applicable

resected - Errbloc
- Separated
24 Liver or other metastasis - NolYes Comment: Certain p a h t s will do well after resectionof
resection m81ast85BS.

25 PostoperativeRx Radiotherapy -NOlYeS Comment: Affects prognosk in some studies.

Chemohrapy local -NolYeS
Chemotherapysystemic -NolYes
APPENDIX 11. P a t h o l o g y f e a t u r e s

Item Alternativefirst level Alternative second level Definitionslcomments

# Name

Basic p a t i e n t i n f o r m a t i o n
1 FBnnberotprimary Give number Comment: If there is mcfe man MB tumour pnrsent, he more
caranomas advanced tumour should be used for staging.
ofhe cola, L rectum
2 Measwemenis at t u m r *Is - Bowel wall barn- measurement c m Comment: BasicpatholoOyinfonmmn Measunnenkambeamade
- Max. tumour size - transverse - on he fresh. unsbetohed specimen.
Freshspecim
-lmgitudinal -cm Fixedspecknen
-m i C m
- Distaldearance margin C m

-No/Yes comment: Mayhavee(lectonpognis.

-m Comment: Delines subgroups *rhich may haw dinerent nahral

- Ulceralive colilis hstay to ‘ s p o r w CRC.
- Cmhnb disease
- Familii adenomatosis
- Radiation colitis
-Schistosomiasii
-Contigwusadenomawilhcancar -NolYes
- Separate adenoma - None
- Give number
5 Tumwr type - Menocarcinoma Definition: Amudnous adenocsrcinorna b demedasahmar in
- Mucinous ademcarcinoma whii>50%ismmposedofelhace*rlamucin.

-Signet ring ceU cardnoma Definition: A signet ring call carcinoma is a neoplasm in which >50%
- UndifterentiaBd is amposed of signet ring ~ o m l a h i n g imaml+ular
- Wler muan (highgrade ky definition).
D a t a of p r o v e n p r o g n o s t i c s i g n i f i c a n c e

6 MiuDsoDpic desaiption of - Primary t u m r cannot be assessed

tumour depth - No evidenca of primary tumour
- Severe dyspiasii Ca in situ
- Tumour invades submucosa
- Turnour invades musarlaris pop-
- Tumour invades through muscularis popria into
the subserosal connective tissue OT non-
peritonealizedp e r i i c or perirecraltissue
- Tumour invades directly into omer organs or
sbllchlres
- Tumour to and invading free (serosd) surfam

h status
7 Regional l ~ p node - ~egionallymph nodes cannot be assessed Comment Most important sngle prognosbc factor denved from pamobgy
- Number of regional lymph n&s examined informaton in the abseoce of restdual disease Thus,
P
- Number involvedwith turnour establishment of lymph node status IS of aibcal importance s
In radical speamens it IS recommended that at least 12 lymph
8 Status of nodes along m a p - Not r m d e d nodes be sought, paying paracular anennon to nodes deep to
named vascular bunk - Negative for turnour and draning the t u m r If a speamen IS deemed node
- Positive for tumour negaave. considerationshould be given to fat dearance of lhe
swamen when less than 12 nodes have been examined
APPENDIX 11. Patholonv features, c o n t i n u e d
Item
# Name Alternative first level Alternative second level Definitionslcomments

D a t a o f p r o v e n p r o g n o s t i c significance

Apical node Stabrs - Not recorded

- Negative for tumoul
- Positive foc turncur
Tumour wansected shown - cannot be assessed Definition: Tumour cells be demonstrated in the line of reseaion.
histokgiily - None
- Proximal line of resecki~ -No/Yes Comment: This feature is of a i t i importanca for patient prognosis.
- Lateral (deep) line of resection -NolYes In case of stapled anastomosis, for assessmentofthe d ~ t a l
- Distal line of resection - No IYes line of resection. the findings on the .doughnur (lm17)
are relevant.
Histologically proven distant -NolYes
metastases

Hilogy in line of resection - No metastasisexcised

of excised distant met&a& - No turncur in line of resetlion
- Tumow in line of resection
13 Venous involvement - Not specifled
- Intramural veins positive
- Extramuralveins positive
- Both inba- and extramural veins positive
14 Histologic partemof - Not recQded
infiltrating margin - Expanding(well circumsaibed)
- Diffusely infiltrating
15 Turnourgrade ~ well d*enliated
- Moderataly differentiated
- Poorly ditterenliated
Comment: The WHO dassifwtion (1W9) aNom foc well and
moderately ditterenliated turnours to be amalgamated into
Data of p r o b a b l e p r o g n o s t i c significance a 'low-grade' category. Poorly dilferentiated
Tumour perforation - Not specified adenocarcinoma,signet ring can, and undifterentiated
- None carcinomas are all cksified as 'high grade.'
- spontaneous
- Surgically induced
Involvementof distal .Not applicable
'dwghnur tissue - No tumou
- Tumour present
l
Mixed infhmmatory d - Not specifled
infiltrate - Not con~piarouS
- conspiarous
Ds(initi0n: Cellular inRlbate includinglymphocvtes in man&-like
Lymphoidaggregates - Not spedfled arrangement at infiltrating margin of the tumour.
- None
- Present
Definition: Nodular mllections of lymphocytes in tissue surrounding
the tumour.
w
ik
APPENDIX Ill. A. International comprehensive anatomical terminology (CAT) for COlOreCtal cancer and 6. Matrix for staging system conversion

Japanese Research
pTNM ACPS Concord Hospital Dukes a Bussey 1 9 5 8 AstlerColler
Line Societv
# Fealure in 1 cat pmu LnUU s.0. L k U swl LCr. Lh. sc.p. S(.W Urn8 st- Lillw--'
0 3 0 3 A-1 3 A 4.5 A 3 1 %
1 - Pnmarv tumour cannot b assessed pTx i 12 12 12 12 12 12
2 pT0 2 15.16 15.16 15.16 15.16 15.16 15.16
3 pTs 3 18.19 18.19 18.19 18.19 18.19 18.19
4 DT~ 4 n 22 22 22
5 5 25 25 B 6 4 B-1 45
6 6 I 4.5 12 12 II 6.8
7.8 12 A 4.5 A-2 4 15.16 15.16 12
15.16 12 12 18.19 1.10 15.16
9 18.19 15.16 15.16 18.19
7 12 n 18.19 18.19 C-1 44 8-2 6-8 22
13.15.16.18.19 22 22 13.14 12
8 14.15.16.16.19 II 64 25 25 15-17 15.16 Ill 7 or 44
1720 12 18.1s 18.19 12 13.14
15.16 8 6 4 A-3 5 15.16 15.16
21 18.19 12 12 C-2 44 Cl 4.5 18.19 18.19
9 22 22 15.16 15.16 13.14 13.14 z n
10 n 18.19 18.19 15-17 1517
11 111 1-8 22 22 20 18-20 N 4 4
12 24 13.14 25 25 13.14
13 25 1517 c26-8 17
14 2M8 18-20 c 1-8 6-1 6.7 13.14 20
22 13.14 12 1517 22
15-17 15.16 18-20
15 N 1-8 16-20 18.19
16 9.12-14 22 22 v 1-8
17 1517 25 25 9.12-14
18-20 15-17
23 D 1-9 52 8 18-20
18 9.12-14 12 23
19 1517 15.16
20 18-20 18.19
23 22
24-28 25

21 C1 1-8
22 13.14
23 15-17
18.19
22
25
24
25 c-2 1-3
26 13.i4
15-17
27 20
28 22
2s

D-1 1-8
9.12-14
15-17
Notes
18-20 r-.
22 ( 1 ) Distant metastasis (Line 21-23) is not considered
26 in Dukes-Bussey and Astler-Collar system. s
(2) Residual tumour status (Line 24-28) is considered in
D-2 1-8
ACPS stage and Concord Hospital system only. It may
9.12-14
15-17 be recorded in the TNM system by the additional
18-20 R classification.
23 (3) All data is of proven prognostic significance.
2+28 0
c