Original Article

Exploring the Rising Incidence of Neuroendocrine Tumors:

A Population-Based Analysis of Epidemiology, Metastatic
Presentation, and Outcomes
Julie Hallet, MD1,2; Calvin How Lim Law, MD, MPH1,2; Moises Cukier, MD1,2; Refik Saskin, MSc3; Ning Liu, MSc3;
and Simron Singh, MD, MPH4,5

BACKGROUND: An increased incidence of neuroendocrine tumors (NETs) has been reported worldwide, but the reasons underlying
this rise have not been identified. By assessing patterns of metastatic presentation, this study sought to examine the epidemiologic
characteristics of NETs and the contribution of early-stage detection to the rising incidence. METHODS: A population-based retro-
spective cohort study was conducted with prospectively maintained databases linked at the Institute for Clinical Evaluative Sciences.
Adult patients with a NET diagnosis from 1994 to 2009 in Ontario, Canada were included. The main outcomes included the overall
and site-specific incidence, proportion of metastatic disease, overall survival (OS), and recurrence-free survival (RFS). RESULTS: Five
thousand six hundred nineteen NET cases were identified. The incidence of NETs increased from 2.48 to 5.86 per 100,000 per year.
Metastases were found in 20.8% at presentation and in another 38% after the initial diagnosis. The proportion of metastases at pre-
sentation decreased from 1994 to 2009 (from 29% to 13%). Therefore, although the incidence of all NETs increased, the overall inci-
dence of metastases did not change (0.63-0.69 per 100,000 per year). The 10-year OS rate was 46.5%, and the RFS rate was 64.6%.
In addition to the primary tumor site, independent predictors of worse OS included an advanced age (P <.0001), male sex
(P <.0001), a low socioeconomic status (P <.0001), and rural living (P 5 0.049). CONCLUSIONS: The incidence of NETs has markedly
increased over the course of 15 years. This is the first study to provide evidence suggesting that the increase in the incidence of NETs
may be due to increased detection. In addition to tumor characteristics, low income and rural residency portend worse survival for
patients with NETs. Cancer 2015;121:589-97. V C 2014 American Cancer Society.

KEYWORDS: carcinoid, detection, epidemiology, incidence, neuroendocrine.

INTRODUCTION
Neuroendocrine tumors (NETs) are a group of rare cancers accounting for 0.46% of gastrointestinal and bronchopulmo-
nary malignancies.1-4 They are a heterogeneous group of malignancies most commonly found in the gastrointestinal sys-
tem, but they can also originate in other areas, including the pancreas, lungs, ovaries, thyroid, pituitary, and adrenal
glands.1,4,5 They exhibit a wide range of clinical behaviors due to the secretion of hormones, most often serotonin, which
creates nonspecific but debilitating systemic symptoms such as diarrhea, bronchospasm, flushing, and cardiac valve
disease.2
Despite their identification more than a century ago, NETs remain a poorly understood disease. Because of their rar-
ity, tumor heterogeneity, nonspecific presentation symptoms, and unique indolent biology as well as a lack of awareness,
patients with NETs can suffer delays in diagnosis of up to 7 years.1,6,7 As a result, they often present at an advanced stage
when a cure is no longer possible.1,7 In such metastatic presentations, the combination of slow progression and hormonal
production may produce debilitating symptoms with a potentially significant impact that leads to deteriorating quality of
life as well as health care resource consumption. Recently, international collaborations have resulted in promising random-
ized controlled trials that revealed significant benefits of somatostatin analog and targeted biologic therapies, which offer
NET patients improved progression-free survival.8,9 However, much remains to be learned about these malignancies to

Corresponding author: Julie Hallet, MD, Odette Cancer Centre, Sunnybrook Health Sciences Centre, 2075 Bayview Avenue, Suite T2-063, Toronto, Ontario M4N
3M5, Canada; Fax: (416) 480-6002; julie.hallet@sunnybrook.ca
1
Division of General Surgery, Odette Cancer Centre, Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada; 2Division of General Surgery, University of To-
ronto, Toronto, Ontario, Canada; 3Institute for Clinical Evaluative Sciences, Toronto, Ontario, Canada; 4Division of Medical Oncology, Odette Cancer Centre, Sunny-
brook Health Sciences Centre, Toronto, Ontario, Canada; 5Department of Medicine, University of Toronto, Toronto, Ontario, Canada.

Part of this work was presented as a poster presentation at the 2012 Gastrointestinal Symposium and at the 2014 Annual Meeting of the American Society of
Clinical Oncology and as a podium presentation at the 2014 Canadian Surgery Forum.

DOI: 10.1002/cncr.29099, Received: August 19, 2014; Revised: September 11, 2014; Accepted: September 17, 2014, Published online October 13, 2014 in Wiley
Online Library (wileyonlinelibrary.com)

Cancer February 15, 2015 589

Original Article
improve survival, which has been stagnating for the past 3
decades.1-3 A better understanding of NETs becomes
even more important and relevant as they are now more
prevalent than esophageal, gastric, pancreatic, or hepato-
biliary cancers, perhaps due to their indolent behavior.1
Previous epidemiological assessments of NETs have
reported a rise in their incidence over the past deca-
des.1,10,11 Although hypotheses of overdetection or
changes in tumor biology have been formulated to explain
this observation,12 no data currently exist to support
either.
This study used a population-based approach to
examine the epidemiologic characteristics of NETs and to
determine whether the changes in their incidence are
related to early detection as demonstrated by patterns of
metastatic presentation. This richly detailed population
data set could provide the first ever data and evidence that
could help us to understand why there has been this
observed rise in the incidence of NETs. We hypothesized
that the incidence of NETs rising over the study period
with a concomitant decrease in metastatic presentation
would bring light to the role of increased detection in this
phenomenon.
MATERIALS AND METHODS
The study was approved by the research ethics board of
Sunnybrook Health Sciences Centre (Toronto, Ontario,
Canada).
Design
A retrospective, population-based cohort study was con-
ducted with data linked from prospectively maintained
administrative databases stored at the Institute for Clinical
Evaluative Sciences (ICES) in Toronto, Ontario.13
Study Population
This population-based study was conducted with all
patients with a valid Ontario Health Insurance Plan
(OHIP) number from 1994 to 2009. During the study
period, the population in Ontario grew from
8,054,030 in 1994 to 10,004,048 in 2009. Under the
Canada Health Act, the Ontario population benefits
from universally accessible and funded health care
through OHIP.14 All residents of Ontario are eligible for
OHIP after they have resided in the province for 3
months.
Data Sources
Three administrative population-based data sets were
linked at ICES. The Ontario Cancer Registry (OCR) is a
registry of all Ontario residents diagnosed with cancer
590
since 1964, and it receives hospital discharge records, pa-
thology reports, death certificates, and reports from re-
gional cancer centers in the province of Ontario.15 All
medical reports containing a cancer diagnosis are legally
subjected to a mandatory report to the OCR. The reliabil-
ity of data contained in the OCR has been previously
ascertained.16-18 Diagnoses are classified according to
International Classification of Diseases, Ninth Revision
(ICD-9), or International Classification of Diseases, Tenth
Revision (ICD-10), for the primary disease site and
according to International Classification of Diseases for On-
cology (ICD-O) for morphology.19,20 The Registered Per-
sons Database (RPDB) collects data on all individuals
eligible for OHIP. It contains demographic information,
including age, sex, postal code, socioeconomic status, and
date of death. Each individual is identified at ICES by a
unique numeric encrypted identifier that can be used to
link these data sources.
Cohort
Adults (>18 years old) with a diagnosis of NET in the
OCR were identified with ICD-O codes. Carcinoid
tumors of uncertain malignant potential, enterochromaf-
fin cell carcinoid, enterochromaffin-like cell tumors, gob-
let cell carcinoid, tubular carcinoid, neuroendocrine
carcinoma, atypical carcinoid, islet cell tumors, insuli-
noma, glucagonoma, gastrinoma, VIPoma, somatostati-
noma, enteroglucagonoma, and mixed islet cell and
exocrine adenocarcinoma (ICD-O: 8240, 8241, 8242,
8243, 8245, 8246, 8249, 8150, 8152, 8153, 8154, 8155,
8156, and 8157) were included. Small cell and large cell
lung carcinoma, pheochromocytoma, paraganglioma,
extra-adrenal paraganglioma, medullary thyroid carci-
noma, and Merckel cell carcinoma (ICD-O: 8002, 8040,
8041, 8042, 8043, 8045, 8013, 8247, 8700, 8680, 8693,
and 8510) were excluded because they represent biologi-
cally different diseases.
Demographic Characteristics
Age and sex were obtained from the RPDB. Socioeco-
nomic status was assessed through income quintiles based
on the median income of a patient’s postal code of resi-
dence.21 The following primary NET sites were
described: stomach, small intestine (including duode-
num), large intestine (including appendix), rectum, pan-
creas, bronchus and lungs (ICD-9: 151, 152, 153, 15,
157, and 162), and others (other ICD-9 codes). Rare
NET sites (eg, biliary, prostate, breast, head, and neck)
were included under the other category because a small
number of cases were anticipated. Metastatic disease was
Cancer February 15, 2015
 Neuroendocrine Tumor Epidemiology/Hallet et al

TABLE 1. Demographics and Distribution of NETs Among Adult Patients in Ontario by Primary Tumor Site,
1994-2009

All NETs Stomach Small Intestine Pancreas Large Intestine Rectum Bronchus, Lung Other NETs

Distribution, n (%) 5619 (100) 282 (5.0) 1015 (18.2) 531 (9.4) 727 (12.9) 690 (12.3) 1403 (25.0) 971 (17.3)
Age at NET diagnosis, 60.9 6 14.5 63.9 6 13.6 63.8 6 13.2 56.7 6 13.3 58.8 6 15.9 56.7 6 12.8 61.3 6 14.6 63.1 6 14.9
mean 6 SD
Male sex, n (%) 2784 (49.5) 131 (46.5) 558 (55.0) 303 (57.1) 377 (51.9) 351 (50.9) 591 (42.1) 473 (48.7)
Income quintile, na (%)
1 (lowest) 1072 (19.1) 58 (20.6) 202 (19.9) 92 (17.3) 137 (18.8) 116 (16.8) 281 (20.0) 186 (19.2)
2 1081 (19.2) 53 (18.8) 183 (18.0) 101 (19.0) 136 (18.7) 121 (17.5) 282 (20.1) 205 (21.1)
3 1123 (20.0) 62 (22.0) 197 (19.4) 105 (19.8) 148 (20.4) 137 (19.9) 280 (20.0) 194 (20.0)
4 1135 (20.2) 52 (18.4) 212 (20.9) 100 (18.8) 137 (18.8) 167 (24.2) 276 (19.7) 191 (19.7)
5 (highest) 1192 (21.2) 56 (19.9) 219 (21.6) 133 (25.0) 166 (22.8) 148 (21.4) 280 (20.0) 190 (19.6)

Abbreviations: NET, neuroendocrine tumor; SD, standard deviation.

a
Accounting for missing data.

Figure 1. Incidence of neuroendocrine tumors (NETs) in the adult population in Ontario, 1994-2009: (A) the incidence of NETs
versus the incidence of all other cancers, (B) the incidence of NETs by sex, (C) the incidence of NETs by age group (in years),
and (D) the incidence of NETs by income quintile.

captured with ICD-9 and ICD-10 codes (ICD-9: 196,
197, 198, 199, and 209; ICD-10: C77, 78, and 79) and
was characterized as at presentation (metastasis code dur-
ing the same episode as the first NET diagnosis) or after
initial diagnosis (metastasis code after the episode of the
first NET diagnosis).
Outcome Measures
The incidence of NETs—overall and by the primary
NET site—was determined for the defined adult Ontario
population for each year from 1994 to 2009. The date of
diagnosis was obtained from the OCR. The prevalence of
metastatic disease was determined for the whole cohort

Cancer February 15, 2015 591

Original Article

TABLE 2. Incidence of NETs per 100,000 per Year in the Adult Population in Ontario by Sex and Primary
Tumor Site, 1994-2009

All NETs
Small Large Bronchus, Other
Year All Male Female Stomach Intestine Pancreas Intestine Rectum Lungs NETs All Cancers

1994 2.46 (198) 2.6 (102) 2.32 (96) 0.07 (6) 0.42 (34) 0.1 (8) 0.37 (30) 0.22 (18) 0.83 (67) 0.43 (35) 541.38 (43,603)
1995 2.35 (192) 2.5 (99) 2.22 (93) 0.11 (9) 0.4 (33) 0.29 (24) 0.27 (22) 0.04 (3) 0.72 (59) 0.52 (42) 532.24 (43,404)
1996 2.84 (234) 3.02 (121) 2.66 (113) 0.07 (6) 0.53 (44) 0.32 (26) 0.35 (29) 0.27 (22) 0.70 (58) 0.59 (49) 542.67 (44,783)
1997 2.89 (242) 2.55 (104) 3.21 (138) 0.18 (15) 0.44 (37) 0.27 (23) 0.37 (31) 0.19 (16) 0.92 (77) 0.51 (43) 557.68 (46,694)
1998 3.29 (279) 3.44 (142) 3.14 (137) 0.12 (10) 0.58 (49) 0.24 (20) 0.42 (36) 0.24 (20) 1.05 (89) 0.65 (55) 567.92 (48,190)
1999 3.03 (261) 2.91 (122) 3.15 (139) 0.07 (6) 0.49 (42) 0.26 (22) 0.33 (28) 0.19 (16) 1.03 (89) 0.67 (58) 578.73 (49,817)
2000 3.44 (301) 4.02 (172) 2.87 (129) 0.13 (11) 0.68 (60) 0.35 (31) 0.3 (26) 0.43 (38) 0.90 (79) 0.64 (56) 586.86 (51,421)
2001 3.46 (309) 3.98 (174) 2.95 (135) 0.16 (14) 0.57 (51) 0.34 (30) 0.55 (49) 0.3 (27) 0.91 (81) 0.64 (57) 594.19 (53,141)
2002 3.23 (295) 3.14 (140) 3.32 (155) 0.14 (13) 0.55 (50) 0.38 (35) 0.49 (45) 0.39 (36) 0.78 (71) 0.49 (45) 587.88 (53,642)
2003 3.50 (325) 3.6 (163) 3.41 (162) 0.26 (24) 0.7 (65) 0.31 (29) 0.45 (42) 0.37 (34) 0.83 (77) 0.58 (54) 584.81 (54,264)
2004 4.12 (389) 4.15 (191) 4.1 (198) 0.24 (23) 0.84 (79) 0.31 (29) 0.55 (52) 0.49 (46) 1.07 (101) 0.63 (59) 601.56 (56,737)
2005 4.34 (415) 4.73 (221) 3.96 (194) 0.21 (20) 0.87 (83) 0.36 (34) 0.54 (52) 0.62 (59) 1.07 (102) 0.68 (65) 608.98 (58,296)
2006 4.92 (478) 4.77 (226) 5.07 (252) 0.31 (30) 0.93 (90) 0.44 (43) 0.69 (66) 0.64 (62) 1.07 (104) 0.85 (83) 615.83 (59,813)
2007 5.17 (509) 5.12 (246) 5.21 (263) 0.32 (32) 0.89 (88) 0.58 (57) 0.75 (74) 0.89 (88) 0.96 (95) 0.76 (75) 629.19 (61,991)
2008 5.96 (596) 5.76 (281) 6.15 (315) 0.34 (34) 1.07 (107) 0.59 (59) 0.75 (75) 1.087 (107) 1.24 (124) 0.90 (90) 615.54 (61,579)
2009 5.86 (596) 5.64 (280) 6.07 (316) 0.29 (29) 1.01 (103) 0.6 (61) 0.69 (70) 0.96 (98) 1.28 (130) 1.03 (105) 611.42 (62,174)
Absolute 3.40 (398) 3.04 (178) 3.75 (220) 0.22 (23) 0.59 (69) 0.5 (53) 0.32 (40) 0.74 (80) 0.45 (63) 0.60 (70) 70.04 (18,571)
difference
% of change 138 117 162 314 140 119 86 336 542 139 13

Abbreviation: NET, neuroendocrine tumor.

Data are presented as incidences (with absolute counts in parentheses) except for the last row.

and by the primary NET site. The 3-, 5-, and 10-year
overall survival (OS) rates were computed with the date of
death according to the RPDB as of March 31, 2010.
Patients with a diagnosis of non-NET cancer within 60
days of their NET diagnosis were excluded from that sur-
vival measure. Recurrence-free survival (RFS), defined as
the occurrence of metastatic disease after the initial diag-
nosis, was examined. The end of follow-up was defined as
the date of death, the date of last contact, or 10 years after
the diagnosis as of March 31, 2010.
Statistical Analysis
Statistical analyses were performed with SAS 9.2 (SAS
Institute, Inc, Cary, NC). Demographic characteristics of
the study cohort were reported as means and standard
deviations for continuous variables and as proportions
and absolute counts for categorical variables. The chi-
square statistic was used to compare proportions between
sex and primary tumor location groups. The incidence
was computed with the defined adult Ontario population
as the denominator and was reported as the incidence per
100,000 per year with the incident case count for that
year. OS and RFS were estimated with Kaplan-Meier
analyses, and comparisons were performed with a log-
rank test. Predictors of 10-year OS and RFS were deter-
mined through Cox regression analyses and were pre-
sented as hazard ratios (HRs) with 95% confidence
intervals (CIs). Statistical significance was set at P < .05.
RESULTS
During the study period, 5619 NET cases were identified
in Ontario. The baseline characteristics are presented in
Table 1. The mean age at diagnosis was 60.9 years, and
49.5% of the patients were male.
Incidence
From 1994 to 2009, the incidence increased from 2.48
(95% CI, 2.13-2.83) to 5.86 cases (95% CI, 5.40-6.35) per
100,000 per year; this represented a 2.36-fold increase. The
incidence of NETs increased steadily among all age groups
but varied according to the primary site of NETs. During
the study period, the incidence of all cancers in Ontario went
from 541.38 to 611.42 per 100,000 per year (or a 1.13-fold
increase; Fig. 1A). There was a progressive increase in inci-
dence in almost all primary NET site groups (Table 2). Simi-
lar increases in incidence were observed between sexes, age
groups, and income quintiles (Fig. 1B-D).
Tumor Characteristics and Metastatic Disease
Overall, the most frequent primary NET sites were bron-
chopulmonary sites (25.0%) and the small intestine
(18.1%). The proportion of patients presenting with met-
astatic disease in the cohort during the study period is
depicted in Figure 2A. Overall, 20.8% of the patients
with NETs (n 5 1166) presented with metastatic disease
at the time of diagnosis, and an additional 38.0%
(n 5 2133) presented with metastases during follow-up.

592 Cancer February 15, 2015

 Neuroendocrine Tumor Epidemiology/Hallet et al

TABLE 3. Overall Survival for Patients With NETs

Among the Adult Population in Ontario, 1994-2009

Overall Survival (%)

1 Year 3 Years 5 Years 10 Years

All patients 80.8 68.3 61.0 46.5

Sex
Female 83.3 71.8 65.2 50.6
Male 78.0 64.7 56.6 42.3
Age at NET diagnosis
19-50 y 92.0 83.6 78.9 69.7
51-60 y 84.2 73.0 67.1 53.6
61-70 y 79.8 67.2 59.6 43.1
71 y 67.5 50.2 39.3 20.4
Income quintile
1 (lowest) 77.1 62.5 55.7 40.3
2 80.3 67.6 59.6 45.9
3 81.6 70.0 62.2 49.9
4 81.2 70.5 63.8 46.8
5 (highest) 82.7 70.3 63.2 48.9
Primary NET site
Stomach 82.1 76.2 67.4 49.7
Small intestine 90.7 81.7 73.4 51.2
Pancreas 78.0 60.7 48.8 30.2
Large intestine 82.2 70.7 64.3 48.3
Rectum 95.3 89.2 87.2 84.0
Bronchus, lung 77.0 64.9 59.7 49.7
Others 66.8 46.8 36.7 23.1

Abbreviation: NET, neuroendocrine tumor.

and 46.5%, respectively (Table 3). OS differed signifi-

Figure 2. Patterns of metastases in neuroendocrine tumors
(NETs) in the adult population in Ontario by the primary tu-
mor site, 1994-2009: (A) the distribution of metastatic NETs
over the course of the disease according to the primary tu-
mor site, (B) the proportion of metastases at presentation for
NETs, and (C) the incidence of NETs metastatic at presenta-
tion versus the incidence of all NETs.
The proportion of patients presenting with metastatic dis-
ease at the time of diagnosis decreased from 29% (n 5 51/
177) in 1994 to 13% (n 5 70/545) in 2009. As a result of
the increased incidence of all NETs and the decreased
proportion of metastatic presentation, the incidence of
metastatic NETs at presentation remained stable (Fig.
2B,C).
OS and RFS
The median follow-up was 3.8 years. The 3-, 5-, and 10-
year OS rates for patients with NETs were 68.3%, 61.0%,
cantly according to the primary tumor site (P < .0001;
Table 3) as well as sex, age group, and income quintile
(P < .0001; Fig. 3). A metastatic status was associated
with worse survival, with 10-year OS rates of 68.2% for
nonmetastatic disease, 17.5% for metastases at presenta-
tion, and 18.7% for metastases after the initial diagnosis
(P < .0001). Independent predictors of 10-year mortality
included age (HR, 1.04; P < .0001), male sex (HR, 1.3;
P < .0001), socioeconomic status (lowest income quin-
tile: HR, 1.25; P 5 .001), rural living (HR, 1.13;
P 5 .049), and NET primary tumor sites (Table 4).
The 3-, 5-, and 10-year RFS rates were 75.0%,
71.0%, and 64.6%, respectively (Table 5 and Fig. 4). In
addition to the primary tumor site, advancing age was asso-
ciated with higher recurrence rates (P 5 .0003; Table 4).
DISCUSSION
This study represents one of the largest and most detailed
cohort analyses of the epidemiology and outcomes of
NETs performed for this unique malignancy. Over a 15-
year period, the incidence of NETs increased almost 2.5-
fold and reached 5.86 per 100,000 per year in 2009.
Overall, 1 in 5 patients initially presented with metastatic
disease. Interestingly, this figure decreased by close to
50% from 1994 to 2009 (from 29% to 13%) while the

Cancer February 15, 2015 593

Original Article

Figure 3. Overall survival for adult patients with neuroendocrine tumors (NETs) in Ontario with an adjustment for the age at diag-
nosis, 1994-2009: (A) overall survival for all patients with NETs, (B) overall survival for patients with NETs by sex, (C) overall sur-
vival for patients with NETs by age group, and (D) overall survival for patients with NETs by income quintile.

overall incidence of NETs was rising. In addition to the
primary tumor site and metastatic status, advancing age,
male sex, low socioeconomic status, and rural area living
were also associated with worse survival.
The epidemiology of NETs has been examined in
previous national and regional registries, with the largest
ones using SEER and Norway National Cancer Registry
data.1,3,10 Although the increase in the incidence of NETs
appears consistent among studies,1,11,22 the magnitude of
the change that we are reporting (from 2.46 to 5.86 cases
per 100,000 per year) was observed only in the latest
SEER analysis (from 1.09 to 5.25 cases per year per
100,000).1 In Europe and Asia, the incidence appears sig-
nificantly lower and ranges from 1.1 to 3.24 cases per
100,000 per year.10,22-25 This may be due to older data
and differences in data registration, but it could also reveal
variability in environmental factors and tumorigenesis.
Health care resource utilization leading to differences in
the frequency of diagnosis is also likely to play a role in
explaining regional and national differences in the inci-
dence of NETs.
Improvements in diagnostic imaging, particularly
computed tomography and gastrointestinal endoscopy,
leading to the incidental identification of asymptomatic
NETs have been suggested to explain the incidence
change.12 In addition, an increased detection rate could
suggest the identification of earlier stage lesions and
thereby lead to the diagnosis of perhaps less aggressive
tumors that may not have been revealed otherwise. No
study has formally looked at the use of diagnostic modal-
ities with respect to diagnoses of NETs. Directly exploring
this increased detection hypothesis, we observed a
decrease in the proportion of NETs presenting with me-
tastases while the overall incidence was rising. Thus, the
increased incidence of all NETs was accompanied by a sta-
ble incidence of advanced NETs. For the first time, the
current study examined the rate of metastatic disease at
presentation with respect to the overall incidence of NETs
to better understand the relationship between the reported
increase in NETs and earlier detection. The decrease in
metastatic presentation in the midst of a rising incidence
of all NETs suggests that earlier detection of NETs is

594 Cancer February 15, 2015

 Neuroendocrine Tumor Epidemiology/Hallet et al

TABLE 4. Predictors of 10-Year Overall and Recurrence-Free Survival for Patients With NETs Among the
Adult Population in Ontario, 1994-2009

10-Year Mortality Recurrence at the End of Follow-Up

Hazard Ratio (95% Hazard Ratio (95%

Variable Confidence Interval) P Confidence Interval) P

Male sex 1.30 (1.19-1.41) <.0001 1.11 (0.99-1.24) .07

Age 1.04 (1.038-1.045) <.0001 1.01 (1.003-1.011) .0003
Income quintilea
1 (lowest) 1.25 (1.09-1.43) .001 0.92 (0.77-1.10) .39
2 1.09 (0.95-1.25) .22 1.01 (0.85-1.20) .92
3 0.98 (0.86-1.13) .80 0.94 (0.79-1.12) .50
4 1.09 (0.95-1.25) .23 0.97 (0.82-1.15) .73
Rural livingb 1.13 (1.00-1.27) .049 1.03 (0.87-1.21) .74
Primary NET sitec
Stomach 0.82 (0.65-1.03) .08 0.66 (0.47-0.93) .02
Pancreas 1.40 (1.20-1.63) <.0001 2.23 (1.86-2.69) <.0001
Small intestine 0.59 (0.51-0.68) <.0001 1.08 (0.9-1.30) .41
Large intestine 0.94 (0.81-1.09) .42 1.09 (0.89-1.33) .42
Rectum 0.34 (0.27-0.44) <.0001 0.36 (0.27-0.49) <.0001
Other NETs 1.59 (1.41-1.79) <.0001 1.75 (1.48-2.06) <.0001

Abbreviation: NET, neuroendocrine tumor.

a
The reference is the fifth income quintile (highest).
b
Living in a community with <10,000 people.
c
The reference is a bronchus and lung site.

TABLE 5. Recurrence-Free Survival for Patients Whether this is due to more frequent use of diagnos-
With NETs Among the Adult Population in Ontario, tic modalities or alterations in pathological disease defini-
1994-2009 tion is yet to be defined. Moreover, the identification of
Recurrence-Free Survival (%) socioeconomic outcome disparities outlines the possible
implication of access to care and health care delivery vari-
1 Year 3 Years 5 Years 10 Years ability in patterns of diagnosis and changes in incidence.
Indeed, NET patients often lack a clear pathway in their
All patients 81.4 75.0 71.0 64.6
Sex
cancer journey, and this is characterized by difficulties in
Female 81.6 76.5 73.0 67.5 obtaining a diagnosis.26 These observations warrant fur-
Male 81.2 73.5 69.0 61.6 ther investigation of health care delivery to improve out-
Age at NET diagnosis
19-50 y 86.0 80.4 76.5 69.8 comes for patients with NETs through a chance at an
51-60 y 77.8 70.0 66.5 61.1 timely diagnosis for all NET patients. Indeed, improve-
61-70 y 81.3 74.4 70.6 63.1
71 y 80.0 74.6 69.6 63.9 ments in diagnostic processes and access to a timely diag-
Income quintile nosis have previously been identified as priorities in NET
1 (lowest) 82.8 76.6 72.6 64.3
2 80.3 73.0 69.1 64.0
research.6
3 82.1 75.4 71.2 65.6 Identified predictors of OS and RFS offer insight
4 81.3 76.0 71.8 64.6
5 (highest) 80.6 74.5 70.8 64.4
into variables to examine in future studies to better under-
Primary NET site stand the behavior of NETs, achieve a timely diagnosis,
Stomach 88.6 84.7 82.1 77.0 and improve management. Worse OS was associated not
Small intestine 80.9 76.0 72.9 64.5
Pancreas 71.6 56.5 46.4 37.1 only with nonmodifiable variables such as male sex and
Large intestine 80.4 76.3 73.1 67.9 advanced age but also with low socioeconomic status and
Rectum 93.5 91.5 90.3 88.8
Bronchus, lung 83.0 77.5 74.8 69.4 rural residency. As previously mentioned, these differen-
Others 74.5 63.7 55.7 45.9 ces may be underlined by variations in access to care in
Abbreviation: NET, neuroendocrine tumor.
populations having lower incomes or living in rural areas
further away from major cancer centers.27 Rural and
taking place. These results shed light on the observations urban patients may also have different environmental
made by previous epidemiology analyses around the exposures and risk factors, although this has not been
world.1,10,23-25 examined to date.

Cancer February 15, 2015 595

Original Article
Figure 4. Recurrence-free survival for adult patients with
neuroendocrine tumors in Ontario, 1994-2009.
Despite the inherent challenges associated with
population-based studies, this examination provides a
long-term detailed and robust appraisal of NETs. It is
acknowledged that this study has limitations, the most
important being the identification of NETs through
ICD-O codes, which are complex and not ideally suited
for the identification of a heterogeneous disease whose
definition may have changed over time. This study also
relies on pathology diagnoses, which could have led to an
underestimation of the actual incidence, an issue for all
previous studies as well. We have chosen to use metastases
as a surrogate for advanced disease rather than TNM stag-
ing. Indeed, NETs currently lack a unified specific staging
system beyond the World Health Organization grade clas-
sification. This decision relies on the unique biology of
NETs and previous observations of metastatic status as
the most significant prognostic factor (as opposed to
nodal disease).1 We acknowledge that regional nodal dis-
ease is also relevant in NETs, but with respect to symp-
toms rather than survival.28 Finally, our OS analysis
defined the last follow-up and, therefore, censoring as the
date of death, March 31, 2010, or the date of last contact
with the health care system. We acknowledge that this
later censoring criterion does not account for patients who
may have moved to another jurisdiction and are still alive:
this decision was made to obtain a conservative estimate
of survival rather than an overestimation.
However, this study provides a uniquely detailed
assessment of the NET landscape to improve our under-
standing of the disease and guide future research efforts
and awareness initiatives. It distinguishes itself from previ-
ous work by providing an exploration of the underlying
mechanism of the increased NET incidence through an
596
analysis of patterns of metastatic presentation over time. It
also offers a detailed description of patterns and predictors
of recurrence in a large cohort of patients with NETs.
In conclusion, the incidence of NETs has markedly
increased over a 15-year period, and this has been paral-
leled by a decreased proportion with metastatic presenta-
tion. This points toward increased detection as a possible
explanation for the significant rise in incidence. Beyond
the primary tumor site and metastatic disease, the socioe-
conomic factors of income and rural residence are signifi-
cantly associated with NET outcomes.
In an effort to raise awareness, promote timely
diagnoses, and improve the management of NETs, par-
ticular attention should be paid to factors associated with
worse outcomes, especially low socioeconomic status
and rural residency. This highlights the need for exami-
nations of health care delivery patterns in the diagnosis
and management of NETs. We believe that the findings
of this unique study will help to further increase our
understanding of the incidence of NETs and shape
future processes of care strategies for the management of
patients with NETs.
FUNDING SUPPORT
This study was funded by an unrestricted grant from the Ontario
Institute for Cancer Research.
CONFLICT OF INTEREST DISCLOSURES
Simron Singh reports grants, personal fees, and nonfinancial sup-
port from Novartis and personal fees from Pfizer outside the sub-
mitted work.
REFERENCES
1. Yao JC, Hassan M, Phan A, et al. One hundred years after
“carcinoid”: epidemiology of and prognostic factors for neuroendo-
crine tumors in 35,825 cases in the United States. J Clin Oncol.
2008;26:3063-3072.
2. Modlin IM, Kidd M, Latich I, Zikusoka MN, Shapiro MD. Current
status of gastrointestinal carcinoids. Gastroenterology. 2005;128:1717-
1751.
3. Modlin IM, Lye KD, Kidd M. A 5-decade analysis of 13,715 carci-
noid tumors. Cancer. 2003;97:934-959.
4. Modlin IM, Champaneria MC, Chan AK, Kidd M. A three-decade
analysis of 3,911 small intestinal neuroendocrine tumors: the rapid
pace of no progress. Am J Gastroenterol. 2007;102:1464-1473.
5. Fraenkel M, Kim MK, Faggiano A, Valk GD. Epidemiology of gas-
troenteropancreatic neuroendocrine tumours. Best Pract Res Clin Gas-
troenterol. 2012;26:691-703.
6. Modlin IM, Moss SF, Chung DC, Jensen RT, Snyderwine E. Pri-
orities for improving the management of gastroenteropancreatic
neuroendocrine tumors. J Natl Cancer Inst. 2008;100:1282-1289.
7. Modlin IM, Oberg K, Chung DC, et al. Gastroenteropancreatic
neuroendocrine tumours. Lancet Oncol. 2008;9:61-72.
8. Rinke A, Muller HH, Schade-Brittinger C, et al. Placebo-controlled,
double-blind, prospective, randomized study on the effect of octreo-
tide LAR in the control of tumor growth in patients with metastatic
neuroendocrine midgut tumors: a report from the PROMID Study
Group. J Clin Oncol. 2009;27:4656-4663.
Cancer February 15, 2015

9. Yao JC, Phan AT, Chang DZ, et al. Efficacy of RAD001 (everoli-
mus) and octreotide LAR in advanced low- to intermediate-grade
neuroendocrine tumors: results of a phase II study. J Clin Oncol.
2008;26:4311-4318.
10. Hauso O, Gustafsson BI, Kidd M, et al. Neuroendocrine tumor epi-
demiology. Cancer. 2008;113:2655-2664.
11. Ellis L, Shale MJ, Coleman MP. Carcinoid tumors of the gastroin-
testinal tract: trends in incidence in England since 1971. Am J Gas-
troenterol. 2010;105:2563-2569.
12. Zarate X, Williams N, Herrera MF. Pancreatic incidentalomas. Best
Pract Res Clin Endocrinol Metab. 2012;26:97-103.
13. Institute for Clinical Evaluative Sciences. http://www.ices.on.ca.
Accessed September 2014.
14. Health Canada. Canada Health Act. http://www.hc-sc.gc.ca/hcs-sss/
medi-assur/cha-lcs/index-eng.php. Accessed January 2014.
15. Robles SC, Marrett LD, Clarke EA, Risch HA. An application of
capture-recapture methods to the estimation of completeness of can-
cer registration. J Clin Epidemiol. 1988;41:495-501.
16. Clarke EA, Marrett LD, Kreiger N. Cancer registration in Ontario:
a computer approach. IARC Sci Publ. 1991:246-257.
17. Holowaty EJ, Dale D. The hospital-only project. Health Rep. 1993;
5:91-95.
18. Paszat LF, Brundage MD, Groome PA, Schulze K, Mackillop WJ. A
population-based study of rectal cancer: permanent colostomy as an
outcome. Radiat Oncol Biol. 1999;45:1185-1191.
19. World Health Organization. International Classification of Diseases
for Oncology (ICD-O). Geneva, Switzerland: World Health Organi-
zation; 1976.
Cancer February 15, 2015
Neuroendocrine Tumor Epidemiology/Hallet et al
20. World Health Organization. Manual of the International Statistical
Classification of Diseases, Injuries, and Causes of Death (9th Revi-
sion). Geneva, Switzerland: World Health Organization; 1977.
21. Wilkins R. Use of postal codes and addresses in the analysis of
health data. Health Rep. 1993;5:157-177.
22. Tsai HJ, Wu CC, Tsai CR, Lin SF, Chen LT, Chang JS. The epi-
demiology of neuroendocrine tumors in Taiwan: a nation-wide can-
cer registry-based study. PLoS One. 2013;8:e62487.
23. Westergaard T, Frisch M, Melbye M. Carcinoid tumors in Denmark
1978-1989 and the risk of subsequent cancers. A population-based
study. Cancer. 1995;76:106-109.
24. Quaedvlieg PF, Visser O, Lamers CB, Janssen-Heijen ML, Taal BG.
Epidemiology and survival in patients with carcinoid disease in the
Netherlands. An epidemiological study with 2391 patients. Ann
Oncol. 2001;12:1295-1300.
25. Hemminki K, Li X. Incidence trends and risk factors of carcinoid
tumors: a nationwide epidemiologic study from Sweden. Cancer.
2001;92:2204-2210.
26. Feinberg Y, Law C, Singh S, Wright FC. Patient experiences of hav-
ing a neuroendocrine tumour: a qualitative study. Eur J Oncol Nurs.
2013;17:541-545.
27. Teckle P, Hannaford P, Sutton M. Is the health of people living in
rural areas different from those in cities? Evidence from routine data
linked with the Scottish Health Survey. BMC Health Serv Res. 2012;
12:43.
28. Ohrvall U, Eriksson B, Juhlin C, et al. Method for dissection of
mesenteric metastases in mid-gut carcinoid tumors. World J Surg.
2000;24:1402-1408.
597