CEBP 09 AM Call for Abstracts w/deadline 09 AM Call for Abstracts w/deadline
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Fesinmeyer, M. D.
Right arrow Articles by Bowen, D. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Fesinmeyer, M. D.
Right arrow Articles by Bowen, D. J.
Cancer Epidemiology Biomarkers & Prevention Vol. 14, 1766-1773, July 2005
© 2005 American Association for Cancer Research

Differences in Survival by Histologic Type of Pancreatic Cancer

Megan Dann Fesinmeyer1,2, Melissa A. Austin1,2,3, Christopher I. Li2,3, Anneclaire J. De Roos2,3 and Deborah J. Bowen1,2

1 Institute for Public Health Genetics, University of Washington; 2 Division of Public Health Sciences, Fred Hutchinson Cancer Research Center; and 3 Department of Epidemiology, University of Washington, Seattle, Washington

Requests for reprints: Melissa A. Austin, Department of Epidemiology and Institute for Public Health Genetics, School of Public Health and Community Medicine, University of Washington, HSB F-363, Box 357236, 1959 Northeast Pacific Avenue, Seattle, WA 98195-7236. Phone: 206-543-0709; Fax: 206-685-9651. E-mail: maustin{at}u.washington.edu


   Abstract
Top
 Abstract
Introduction
 Patients and Methods
 Results
 Discussion
 References
 
Objective: Although pancreatic cancer has an extremely high case fatality rate, little is known about differences in mortality by histologic types. We examined median survival and risk of mortality for endocrine pancreatic tumors and two types of exocrine tumors, adenocarcinomas, and mucinous tumors.

Method: This analysis included 35,276 pancreatic cancer cases reported to the nine population-based cancer registries participating in the Surveillance, Epidemiology, and End Results program from 1973 to 2000. Survival among cases with pancreatic adenocarcinomas, mucinous tumors, and endocrine tumors were compared using Kaplan-Meier plots. Comparative risks of mortality were evaluated using multivariate adjusted Cox regression models.

Results: Endocrine pancreatic cancer cases had a median survival of 27 months compared with a median survival of 4 months for adenocarcinoma and mucinous tumor cases. Compared with adenocarcinoma cases, endocrine tumor cases had a 0.28-fold lower risk of mortality [95% confidence interval (95% CI), 0.26-0.30], and mucinous tumor cases had a 0.88-fold lower risk (95% CI, 0.84-0.91). These results were similar for men and women. Within histologic types, advanced tumor stage, older diagnosis age, surgery, and Black race were associated with increased risks of mortality, whereas female sex and more recent year of diagnosis were associated with decreased risks.

Conclusion: This study confirms the clinical observation that patients with endocrine pancreatic cancer survive longer than patients with exocrine tumors. A better understanding of these differences could contribute to identifying the underlying causes of pancreatic cancer and to improving survival rates across all histologic types.


   Introduction
Top
 Abstract
 Introduction
Patients and Methods
 Results
 Discussion
 References
 
Pancreatic cancer is a leading cause of cancer mortality in the United States; it is the fourth and fifth most common cause of cancer death among men and women, respectively (1). Although several risk factors for pancreatic cancer have been identified, including smoking (2), age (3), family history (4), and diabetes (5), this knowledge has not led to substantial improvements in the early detection or survival of pancreatic cancer cases (6). In addition, most studies of pancreatic cancer survival focus on survival after surgery and are thus not generalizable to all cases.

Similar to cancers arising in other organs, pancreatic tumors are classified according to their cell type of origin, structure, and behavior. The pancreas is a multifunctional organ consisting of a variety of cell types (6). The exocrine pancreas is comprised of duct cells and acinar cells that produce a combination of gastric enzymes needed for digestion (7). In contrast, the endocrine pancreas contains ß cells and other types of endocrine cells that produce hormones, including insulin, that are dispersed throughout the body and that are required for glucose metabolism (8).

The majority of exocrine pancreatic cancers are adenocarcinomas. These tumors originate in the epithelial cells lining the pancreatic duct, form gland-like structures, and account for ~90% of all pancreatic cancers (9, 10). Mucinous tumors, another type of exocrine pancreatic cancer, is the second most common histologic type of pancreatic cancer, accounting for <10% of all tumors. Mucinous tumors are usually less invasive at diagnosis than other types of exocrine pancreatic cancer (11). These tumors form sac-like structures, secrete mucin, and typically arise from the pancreatic ductal epithelium (11). Cancers of the endocrine pancreas are less common, occurring at a rate of about five per million person-years, and accounting for <5% of all pancreatic cancers (12). These tumors arise from pancreatic islet cells, including ß cells and {alpha} cells (8). Rather than ablating endocrine function in the pancreas, these tumors often produce pancreatic hormones at high levels (13). Tumors of this type include insulinomas and gastrinomas and are often detected due to the symptoms caused by excessive hormone levels (12).

Although disparities in survival time between endocrine and exocrine pancreatic cancer have been noted clinically, these differences have not been quantified. Furthermore, survival differences among subtypes of exocrine pancreatic cancer (adenocarcinomas and mucinous tumors) have not been reported. The few studies that have evaluated pancreatic cancer survival by histologic type were not population-based and were restricted to pancreatic cancer cases receiving surgery thus limiting their generalizability (14-16). The purpose of this study was to quantify the median survival times and compare the risks of mortality for cases diagnosed with adenocarcinomas, mucinous tumors, and endocrine carcinomas of the pancreas using population-based data from the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) program from 1973 to 2000. In addition, we sought to identify risk factors associated with mortality within each histologic type.


   Patients and Methods
Top
 Abstract
 Introduction
 Patients and Methods
Results
 Discussion
 References
 
Case Selection
Pancreatic cancer cases diagnosed from January 1973 to December 2000 were ascertained from the nine population-based cancer registries that participated in the SEER program over this period (17). These registries serve the states of Connecticut, Hawaii, Iowa, New Mexico, and Utah and the metropolitan areas of Atlanta, Detroit, San Francisco-Oakland, and Seattle-Puget Sound. All registries contain data on cancers diagnosed from January 1973 onward, with the exception of the Seattle-Puget Sound and Atlanta registries, which contain data on cancers diagnosed from 1974 onward and 1975 onward, respectively. It is estimated that >95% of all incident cancer cases in the populations under surveillance are ascertained. Further operational details and methods used by the SEER program are provided elsewhere (18).

In total, 64,649 potentially eligible cases were identified from the SEER registries. Those with an unknown survival time (n = 2,691), a diagnosis age of <20 years (n = 32), nonmalignant disease (n = 47), unstaged tumors (n = 10,751), rare histologic types of pancreatic cancer based on their four-digit International Classification of Diseases for Oncology, 3rd Edition (ICD-O-3) histology code (n = 1,397), or a nonspecific histology (n = 14,455) were excluded. The remaining 35,276 cases of pancreatic cancer that were included in the analysis were divided into three histologic type groups based on their ICD-O-3 codes: (a) adenocarcinomas (n = 31,357; 89% of total; ICD-O-3 codes 8140, 8141, 8143, 8144, 8190, 8211, 8261, 8262, 8263, 8290, 8440, 8503, 8560, and 8570), (b) mucinous tumors (n = 2,865; 8% of total; ICD-O-3 codes 8470, 8471, 8472, 8473, 8480, and 8481), and (c) endocrine tumors (n = 1,054; 3% of total; ICD-O-3 codes 8150, 8151, 8152, 8153, 8154, 8155, and 8246). It is worth noting that 93% of the mucinous tumor cases were categorized as mucinous adenocarcinoma or mucin-producing adenocarcinomas (ICD-O-3 codes 8480 and 8481, respectively).

Survival and Mortality Analyses between Histologic Types
Survival time was measured in months from date of diagnosis until date of death. Nondeaths were censored at the follow-up cutoff date for the SEER data set (December 31, 2000), or the date the case was last known to be alive. SEER registries obtain survival time for cancer cases by reviewing medical records and death certificates to obtain dates of diagnosis, death, and last known follow-up. Because pancreatic cancer is often fatal at diagnosis, a large number of cases had a SEER-assigned survival time of 0 month (n = 3,080; 8.6% of total), indicating that the case survived for <1 month. To minimize any bias that could result from excluding these cases from the analysis, cases with a survival time of zero were each assigned a survival time of 0.5 month.

We used Stata version 8.2 (Stata Corp., College Station, TX; ref. 19) for all descriptive and statistical analyses. To compare survival in the three histologic types, we plotted Kaplan-Meier curves for 5-year survival. The assumption of proportional hazards was confirmed using plots of the log cumulative hazards for each histologic type (data not shown). Cox regression was then used to compare risks of mortality for the three case groups and to calculate hazard ratios (HR) and 95% confidence intervals (95% CI; ref. 20). In these analyses, cases with adenocarcinoma served as the reference group.

Several potential confounding variables were defined a priori and were included in the multivariate Cox model, coded as categorical variables. These variables were age at diagnosis (<50, 50-59, 60-69, 70-79, and >80 years), year of diagnosis (1973-1976, 1977-1980, 1981-1984, 1985-1988, 1989-1992, 1993-1996, and 1997-2000), SEER historic stage (local, regional, or distant), race (White, Black, Asian/Pacific Islander, and other/unknown), SEER registry, and gender. Another measure of tumor progression, tumor grade, was not included in these analyses due to the large percentage of cases with ungraded tumors (50-79% of all cases with different histology types in this data set). Because previous studies found differences in pancreatic cancer mortality for men and women (21, 22), we tested for effect modification by gender. Gender was found to be an effect modifier of the association between histologic type and risk of mortality by likelihood ratio testing (P < 0.001), comparing a model containing a multiplicative interaction term for gender and histology with a model without the interaction term. Thus, the mortality analyses were repeated stratified by gender.

Mortality Analyses Stratified by Histologic Type
We used Cox regression to identify risk factors associated with mortality within each of the three histologic types. The risk factors examined were tumor stage, age at diagnosis, calendar year at diagnosis, gender, and race in multivariate adjusted models, using the definitions of covariates provided above. Test for trend analyses for ordered variables (tumor stage, age at diagnosis, and calendar year at diagnosis) were done using grouped linear variables in the multivariate-adjusted models.

Analysis of Surgery Subgroup
Given the generally short survival time of pancreatic cancer, cases that survive long enough to receive surgical treatment are highly selected and are thus not likely to represent all pancreatic cancer cases. However, for comparison with other studies, we conducted an analysis of the association between surgery and pancreatic tumor survival. The three surgery classification groups included in the analysis were (a) any partial or total pancreatectomy, including the Whipple procedure (ref. 6; a total or partial pancreatectomy with a subtotal gastrectomy and duodenectomy, either with or without a splenectomy); (b) other cancer-directed surgery (including local excision of tumor); and (c) no cancer-directed surgery or no surgery (including cases who only received biopsies). HRs were calculated comparing recipients of pancreatectomy and other cancer-directed surgery with patients who did not receive surgery. Subjects with an unknown surgery status (n = 14,692) were excluded from this analysis.


   Results
Top
 Abstract
 Introduction
 Patients and Methods
 Results
Discussion
 References
 
Demographic characteristics of the three pancreatic cancer histologic types are presented in Table 1. The mean age at diagnosis was similar for both adenocarcinomas and mucinous tumors (67.3 and 66.7 years, respectively), whereas endocrine tumor patients had a somewhat younger mean age at diagnosis (58.0 years). In general, the number of cases diagnosed with each histologic type increased over the course of our study. Nearly equal numbers of women and men were diagnosed with mucinous tumors, whereas more men than women were diagnosed with endocrine tumors and adenocarcinomas. Distributions by race were similar within histologic types, although a slightly lower proportion of endocrine tumor cases were Black. In general, the proportion of cases with each histologic type were similarly distributed throughout the SEER sites. Distributions of localized, regional, and distant tumor stages were also similar within each of the three groups, with more than two thirds of the cases having distant disease.


View this table:
[in this window]
[in a new window]
  		Table 1. Demographic characteristics of pancreatic cancer cases by histologic type

 
Finally, surgery status was known for 58.3% of cases, although this proportion varied by histologic type as surgery status was known for 57.8% of the adenocarcinoma cases, 58.5% of the mucinous tumor cases, and 75.7% of the endocrine tumor cases (Table 1). The percentage of cases that received either a pancreatectomy or other cancer-directed surgery also differed by histologic type, ranging from 11.1% of adenocarcinoma cases to 17.6% of mucinous tumor cases and to 29.2% of endocrine tumor cases.

Survival and Mortality between Histologic Types
At all time points in the follow-up period, survival rates among endocrine pancreatic cancer cases were higher than for either adenocarcinoma or mucinous tumor cases (Fig. 1). For example, at 24 months, 51.6% of endocrine cases were alive compared with only 4.6% of adenocarcinoma cases and 8.6% of mucinous tumor cases. Median survival of endocrine pancreatic cancer cases was ~2 years longer than for adenocarcinoma or the mucinous tumor cases (27 versus 4 months, respectively, Table 2). Compared with adenocarcinoma cases, patients with endocrine tumors and mucinous tumors had reduced adjusted risks of mortality (HR, 0.28; 95% CI, 0.26-0.30 and HR, 0.88; 95% CI, 0.84-0.91, respectively).



View larger version (14K):
[in this window]
[in a new window]
  		Figure 1. Kaplan-Meier curves comparing percent survival for 5 years in the pancreatic cancer histologic types: adenocarcinoma (n = 31,357, solid line), mucinous tumors (n = 2,865, dotted line), and endocrine tumors (n = 1,054, dashed line).

 

View this table:
[in this window]
[in a new window]
  		Table 2. Median survival times and adjusted HRs for mortality by histologic type of pancreatic cancer and gender

 
Among mucinous tumor cases, the relative decrease in adjusted risk of mortality compared with adenocarcinoma cases was slightly greater among women (HR, 0.84; 95% CI, 0.79-0.89; Table 2) than among men (HR, 0.91; 95% CI, 0.86-0.97), although this difference was within the limits of chance. In endocrine tumor cases, the adjusted HRs were nearly identical in both genders (men: HR, 0.28; 95% CI, 0.26-0.31 and women: HR, 0.27; 95% CI, 0.24-0.30). However, women with endocrine tumors had a substantially longer median survival compared with men (37 versus 23 months).

Mortality within the Adenocarcinoma Cases
Table 3A presents median survival times and adjusted HRs for mortality among adenocarcinoma cases. Age at diagnosis, year of diagnosis, gender, race, tumor stage, and surgery were all related to risk of mortality in this group. Specifically, risk of mortality increased in each ascending category of age at diagnosis (Ptrend < 0.001), with corresponding decreases in median survival time. Patients diagnosed with adenocarcinoma in later calendar years had a decreased risk of mortality compared with cases diagnosed in the earliest time interval. This risk reduction increased in each time interval, culminating with a 27% reduction in risk of mortality in cases diagnosed in the 1997 to 2000 interval compared with those diagnosed between 1973 and 1976 (Ptrend < 0.001).


View this table:
[in this window]
[in a new window]
  		Table 3.

 
The median survival times for men and women differed slightly (3 months for men and 4 months for women), and the adjusted reduction in risk associated with female gender was 8% in adenocarcinoma cases (P < 0.001). Compared with Whites, who had a median survival of 4 months, Blacks had a median survival of 3 months and an 1.11-fold (95% CI, 1.07-1.15) increased adjusted risk of mortality. Asian/Pacific Islanders were not different from Whites with respect to survival or risk of mortality. Compared with patients with localized adenocarcinoma, patients diagnosed with regional or distant adenocarcinoma had 1.17-fold (95% CI, 1.12-1.22) and 2.43-fold (95% CI, 2.33-2.54) increases in risk of mortality, respectively (Ptrend < 0.001).

Compared with adenocarcinoma cases who received no cancer-directed surgery, who had a median survival of 4 months, cases who received a partial or complete pancreatectomy had a median survival of 13 months (the longest median survival of any of the adenocarcinoma strata analyzed) and an 0.53-fold (95% CI, 0.50-0.57) reduced adjusted risk of mortality. Cases receiving other cancer-directed surgery also experienced improvements in survival time compared with cases receiving no surgery, with a median survival of 8 months and a 0.70-fold (95% CI, 0.58-0.83) decreased adjusted risk of mortality.

Mortality within Mucinous Tumor Cases
Table 3B presents median survival times and adjusted HRs for mortality among mucinous tumor cases. Similar to adenocarcinoma cases, age at diagnosis, year of diagnosis, gender, race, tumor stage, and surgery were all related to risk of mortality.

Cases with mucinous tumors diagnosed in all age groups after age 50 had increased risks of mortality compared with cases diagnosed before age 50 (Ptrend < 0.001). Cases diagnosed at age 80 or older had the greatest risk of mortality (HR, 2.55; 95% CI, 2.12-3.07). Risk of mortality decreased for cases in each ascending year of diagnosis, with the greatest reduction observed among cases diagnosed from 1997 to 2000, who had a 0.68-fold (95% CI, 0.59-0.80) reduction in risk of mortality compared with patients diagnosed from 1973 to 1976 (Ptrend < 0.001). The median survival times for men and women differed slightly for mucinous tumors (3 months for men and 4 months for women), and females in this group had a 0.87-fold (95% CI, 0.80-0.94) adjusted reduction in risk of mortality. Compared with Whites, who had a median survival of 4 months, Blacks had a shorter median survival of 3 months, and a 1.33-fold (95% CI, 1.16-1.53) increased adjusted risk of mortality. Asian or Pacific Islander cases had a slightly longer median survival of 6 months, but their adjusted reduction in risk of mortality was not different from that of Whites. The risk of mortality associated with stage increased in each higher stage and was greatest in cases with distant tumors, who had a 4.64-fold (95% CI, 3.90-5.52) increased risk of mortality compared with cases with localized tumors (Ptrend < 0.001).

Compared with cases who received no cancer-directed surgery, who had a median survival of 4 months, cases who received a partial or complete pancreatectomy had a median survival of 24 months and a 0.39-fold (95% CI, 0.32-0.47) reduced adjusted risk of mortality. Cases receiving other cancer-directed surgeries experienced similar improvements in survival time and risk of mortality compared with patients receiving no surgery.

Mortality within the Endocrine Tumor Cases
Table 3C presents median survival times and adjusted HRs for mortality among endocrine tumor cases. Age at diagnosis, year of diagnosis, gender, race, tumor stage, and surgery were again all related to risk of mortality in this group. The risk of death among cases with endocrine tumors increased with age (Ptrend < 0.001), culminating in a 5.79-fold (95% CI, 4.06-8.27) increased risk of mortality in patients diagnosed at age 80 or older compared with patients diagnosed before age 50.

Median survival durations and the magnitudes of the HRs tended to increase over time compared with cases diagnosed from 1973 to 1976 (Ptrend < 0.001). The median survival was more than a year longer among women with endocrine tumors compared with men with these tumors (37 versus 23 months, respectively), and females had a 0.81-fold (95% CI, 0.70-0.94) adjusted reduction in risk of mortality. Race was not associated with adjusted risk of mortality in endocrine tumor cases, although median survival differed somewhat by race. Distant endocrine tumors had a 3.33-fold (95% CI, 2.50-4.45) increased risk of mortality compared with localized tumors, but no increase in risk was found for regional tumors. However, median survival was 29 months shorter (70 versus 99 months) for regional tumors compared with localized tumors, and advanced tumor stage was associated with increased adjusted risk of mortality (Ptrend < 0.001).

Compared with subjects who received no cancer-directed surgery, who had a median survival of 21 months, cases who received a partial or complete pancreatectomy had a median survival of 97 months and a 0.48-fold (95% CI, 0.36-0.65) reduced adjusted risk of mortality. Patients receiving other cancer-directed surgery experienced even better survival and reductions in mortality: fewer than 50% of these cases died before the end of the study period and had a 0.31-fold (95% CI, 0.19-0.51) reduced adjusted risk of mortality compared with cases receiving no surgery.


   Discussion
Top
 Abstract
 Introduction
 Patients and Methods
 Results
 Discussion
References
 
The analyses presented here quantitatively show that survival time among pancreatic cancer cases varies by histologic type in a large, population-based setting. Specifically, patients diagnosed with cancer of the endocrine pancreas survive ~2 years longer, on average, than those diagnosed with exocrine pancreatic cancer, including both adenocarcinoma and mucinous tumor cases. This result concurs with qualitative clinical observations of decreased mortality among endocrine pancreatic cancer cases compared with exocrine pancreatic cancer cases. After adjustment for age at diagnosis, gender, diagnosis year, stage, race, and SEER registry, there was a 0.28-fold decreased adjusted risk of mortality for endocrine tumors, and a 0.88-fold decreased adjusted risk of mortality for mucinous tumors compared with adenocarcinomas, although mucinous cases did not have a longer unadjusted median survival time. These decreased adjusted risks of mortality were similar among men and women, although the median survival time for women with endocrine cancer was considerably longer than it was for men (37 versus 23 months, respectively).

Whereas there are many published accounts of survival in specific patient subgroups (14-16, 23), we are not aware of any previous, population-based reports comparing survival between endocrine and exocrine pancreatic tumors. A previous analysis of data on 100,313 pancreatic cancer cases in the National Cancer Database estimated 5-year survival rates for different tumor stages and courses of treatment, but the National Cancer Database is not population-based, and that analysis did not examine survival within or between histologic types (24). Although differences in methodology preclude direct comparison between the National Cancer Database study and the results presented here, both studies found that surgery is associated with increased survival and documented that more advanced tumors are associated with decreased survival.

At this time, there is no known explanation for the difference in survival time between cases with exocrine and endocrine tumors, but tumors of the exocrine pancreas may be more lethal than endocrine tumors due to their cellular or molecular behavior. For example, pancreatic adenocarcinomas often block the pancreatic duct, resulting in jaundice and cachexia (25). In contrast, endocrine pancreatic tumors often overproduce normally occurring substances such as insulin and glucagons (13). How these histologic differences influence survival remains to be determined by basic science research studies.

Our observation that risks of mortality declined from 1973 to 2000 in each histologic type of pancreatic cancer suggests that some progress has been made in the diagnosis and/or treatment of the disease over this time period. For adenocarcinomas, this improvement has been modest, with only a slight increase in median survival time over the 28-year period (3 months in 1973-1976 versus 4 months in 1997-2000). The statistically significant decrease in risk of mortality over this same time period may be due to the large sample size in this analysis and may not represent a clinically significant improvement. A similar trend towards slight improvement in survival and risk of mortality over time was also seen in mucinous tumors. However, for endocrine tumors, increased survival times and reductions in risk of mortality over time have been more dramatic, as median survival times have improved from 11 months in 1973 to 1976 to 26 months in 1997 to 2000.

It is unclear why women with endocrine tumors survive >1 year longer than their male counterparts (37 versus 23 months). This finding suggests that hormonal factors may affect endocrine pancreatic cancer survival, as they are also related to the incidence of several other types of cancer(26, 27). In addition, several studies have found an association between hormones and reproductive factors and the incidence of exocrine pancreatic cancer. In the Nurse's Health Study, increased parity was significantly associated with a reduced risk of exocrine pancreatic cancer, with an average of 10% reduction in risk associated with each additional birth (28). In the Enhanced Cancer Surveillance project based in Canada, use of oral contraceptives was associated with a 0.36-fold (95% CI, 0.13-0.96) reduction in risk of exocrine pancreatic cancer (29). No epidemiologic study to date has investigated the role of hormonal factors in relation to endocrine pancreatic cancer risk, and no study has addressed the association between hormonal factors and pancreatic cancer survival. Furthermore, no prior studies have compared survival in men versus women diagnosed with different pancreatic cancer histologies. Thus, additional investigations aimed at furthering our understanding of the etiology of endocrine pancreatic tumors, especially differential risk by gender, are needed.

This analysis provides evidence of increased mortality risk among Blacks with exocrine pancreatic cancer compared with Whites. Previous analyses have not found consistent differences in pancreatic cancer mortality by race, including two studies using SEER data (30, 31). Unlike the results reported here, neither of these studies used SEER data from all nine registries in operation from 1973 to 2000. One of these studies found a higher incidence of pancreatic cancer in Blacks than Whites but reported no difference in risk of mortality (30). This is in contrast to our findings of a HR of 1.11 (95% CI, 1.07-1.15) and 1.33 (95% CI, 1.16-1.53) for Blacks compared with Whites with adenocarcinoma and mucinous tumors, respectively. The difference may be attributable to the fact that the previous study excluded all cases with unknown treatment, which constituted 41.6% of our sample. In a study using SEER centers with a large proportion of Asians, unadjusted median survival was longer in Asians than Whites, a finding that is not consistent with our results. This difference could be due to the inclusion of cases with unknown tumor stage in that study (11.1% of their sample; ref. 31), whereas we excluded such cases. Also in that study, median survival time in pancreatic cancer cases was slightly shorter among Black women than White women after adjustment for age, diagnosis year, tumor stage, tumor grade, and morphology, in agreement with our unadjusted median survival results. If the increased mortality risk for pancreatic cancer in Blacks is confirmed in future studies, there are at least two possible explanations to consider.

First, poorer health outcomes for Black versus White cancer cases have been reported in many types of cancer, including colorectal, breast, lung, and prostate cancers (32). Furthermore, a higher proportion of Black Americans than White Americans have no health insurance (33), and uninsurance or underinsurance has been associated with decreased survival time in several cancer types (34-36). These trends might also be expected to apply to pancreatic cancer cases. However, SEER does not collect data on social or economic variables; thus, it was not possible to investigate the interaction of socioeconomic risk factors with race on survival in this data set. In addition, prognostic biomarker profiles may vary between Whites and Blacks. Cases with K-ras mutations in the paraortic lymph nodes have shorter survival times than cases with lymph nodes free of K-ras mutations (37). Although the proportion of cases with K-ras mutations was similar in Whites and Blacks in one study (38), the prognostic value of specific K-ras and other mutations in Blacks and Whites warrant investigation as possible mechanisms influencing survival.

In this study, among partial or complete pancreatectomy or other cancer-directed surgery recipients, median survival was considerably greater than among cases receiving no cancer-directed surgery, no surgery, and with unknown surgery status in all histologic types. Previous research on pancreatic cancer survival has focused on cohorts of patients who have undergone pancreatic resection, rather than all diagnosed cases. For example, a recent study of pancreatic adenocarcinoma patients who underwent pancreatic resection reported a median survival of 13.6 months and a 5-year survival of 14.6% (14). In that study, treatment type, large tumor size, advanced stage, and high grade were each associated with decreased 5-year survival. This result is similar to our finding that tumors diagnosed at an advanced stage were associated with an increased risk of mortality and concurs with our report of a 13-month median survival in adenocarcinoma cases undergoing pancreatectomy.

The survival analyses conducted here, comparing cancer-directed surgery recipients with patients receiving no surgery, suggest that surgery increases survival time. However, surgery is generally offered to patients who have a better prognosis and less advanced disease. Thus, even if surgery conveyed no true survival benefit, the subgroup of patients who receive surgery are likely to live longer than those patients who were not surgical candidates. Adjustment for stage may not be sufficient to remove this bias, because prognosis can vary within each tumor stage. In addition, the proportion of patients who did not receive surgery varied between histologic types; thus, a direct comparison of surgery effects between histologic types is probably not valid. Whereas this analysis may overestimate the quantitative benefit of surgery, the large reduction in risk of mortality and increase in median survival in all histologic types suggests that surgery provides a survival benefit to pancreatic cancer patients.

Another implication of the results presented here is that the high case fatality rate of pancreatic cancer may introduce bias into population-based studies of cancer survivors, caused by a combination of recruitment of cases with less rapidly fatal pancreatic cancer types, and the unavailability of patients with more advanced disease. In particular, studies that recruit survivors of pancreatic cancer are likely to overrepresent cases with endocrine tumors and underrepresent cases with adenocarcinoma. Recent data analyses from the Cancer Genetics Network, a multisite national registry, illustrate this important bias (39). The Cancer Genetics Network uses SEER to identify cancer cases from six different cancer sites: melanoma, pancreatic, kidney, lung, thyroid, and non-Hodgkin's lymphoma, using identical recruitment and data collection methods for all cancer sites. Cancer Genetics Network researchers found that response rates (number of completed interviews over number of possibly eligible respondents) vary dramatically among the six cancer sites, from a low of 23% for pancreatic cancer to 51% for kidney cancer to 74% for bladder cancer. A major reason for this variation is the number of deceased cancer patients by the time of attempted contact. In contrast, completion rates (number of completed interviews divided by the number of eligible, contacted respondents) among patients from the six cancer sites varied much less, with a range of 78% to 96%. These results underscore the importance of rapid case recruitment strategies in pancreatic cancer research to better understand the underlying causes of all forms of pancreatic cancer. Conversely, it is important for physicians to realize that there are less fatal forms of pancreatic cancer that may benefit from surgery. For example, endocrine tumor cases who do not receive surgery have a median survival time of 21 months, but median survival time increases to 97 months among those cases who receive a pancreatectomy.

Whereas the population-based design of this study enhances the generalizability of its results, the noncentralized nature of SEER data collection at different geographic sites is a limitation of this study. Reviews of pathology reports and the recording of histology codes was conducted by a large number of SEER staff over 28 years at nine sites, using medical records and death certificates with varying levels of completeness and accuracy. In addition, histopathologic diagnoses were made by a large number of pathologists and physicians, with perhaps different methods of diagnosing pancreatic tumors and classifying their histology. Thus, there is likely to be some degree of misclassification not only by histologic type but also for the other variables collected by SEER that were used in this analysis. This data set is also limited by the small number of available potential confounding variables that could be related to survival following a pancreatic cancer diagnosis such as smoking status, alcohol use, comorbidities such as diabetes, access to health care, and socioeconomic status. Furthermore, it is important to note that excluding 14,455 cases with no known histology could bias the results, if survival time or underlying histologic type is related to missing data for the histology variable. It is possible that cases who die quickly may be more likely to have an unrecorded histologic type because they undergo fewer diagnostic procedures before dying. Thus, excluding these cases from the survival analyses may result in an overestimation of survival time for pancreatic cancer cases overall. However, if the distribution of underlying histologic types is similar both in cases with known and unknown histology, then the between-histologic type comparisons are valid.

In summary, this large, population-based, quantitative analysis confirms the clinical observation of increased median survival time in endocrine pancreatic cancer compared with either pancreatic adenocarcinoma or mucinous pancreatic tumors (27 versus 4 months, respectively). Furthermore, a larger percentage of endocrine than exocrine cases survived at all time points up to 5 years after diagnosis. Among patients with endocrine tumors, women survived substantially longer than men (37 versus 23 months, respectively). Within histologic types, advanced tumor stage, older diagnosis age, surgery, and Black race were associated with increased risk of mortality, whereas female sex and more recent year of diagnosis were associated with decreased risk. Although selection bias may be present, the results also suggest that surgery provides survival benefit. Better understanding of these differences in survival could contribute to identifying the underlying causes of pancreatic cancer and increasing survival among all histologic types.


   Footnotes
 
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received 2/16/05; revised 4/21/05; accepted 4/29/05.


   References
Top
 Abstract
 Introduction
 Patients and Methods
 Results
 Discussion
 References
 

  1. Jemal A, Tiwari RC, Murray T, et al. Cancer statistics, 2004. CA Cancer J Clin 2004;54:8–29.[Abstract/Free Full Text]
  2. Warshaw AL, Fernandez-del Castillo C. Pancreatic carcinoma. N Engl J Med 1992;326:455–65.[Medline]
  3. Ghadirian P, Lynch HT, Krewski D. Epidemiology of pancreatic cancer: an overview. Cancer Detect Prev 2003;27:87–93.[CrossRef][Medline]
  4. Lynch HT, Smyrk T, Kern SE, et al. Familial pancreatic cancer: a review. Semin Oncol 1996;23:251–75.[Medline]
  5. Everhart J, Wright D. Diabetes mellitus as a risk factor for pancreatic cancer. A meta-analysis. JAMA 1995;273:1605–9.[Abstract]
  6. Bastidas JA, Poen JC, Niederhuber JE. Pancreas. In: Abeloff MD, Armitage JO, Lichter AS, Niederhuber JE, editors. Clinical oncology. New York: Churchill Livingstone; 2000.
  7. Hollingsworth MA. Proteins expressed by pancreatic duct cells and their relatives. Ann N Y Acad Sci 1999;880:38–49.[Abstract/Free Full Text]
  8. Cotran RS, Kumar K, Collins T. Robbins pathologic basis of disease. Philadelphia: Saunders; 1999.
  9. Cowgill SM, Muscarella P. The genetics of pancreatic cancer. Am J Surg 2003;186:279–86.[CrossRef][Medline]
  10. Li D, Jiao L. Molecular epidemiology of pancreatic cancer. Int J Gastrointest Cancer 2003;33:3–14.[Medline]
  11. Lichtenstein DR, Carr-Locke DL. Mucin-secreting tumors of the pancreas. Gastrointest Endosc Clin N Am 1995;5:237–58.[Medline]
  12. Mullan MH, Gauger PG, Thompson NW. Endocrine tumours of the pancreas: review and recent advances. ANZ J Surg 2001;71:475–82.[Medline]
  13. Rindi G, Bordi C. Highlights of the biology of endocrine tumours of the gut and pancreas. Endocr Relat Cancer 2003;10:427–36.[Abstract]
  14. Cleary SP, Gryfe R, Guindi M, et al. Prognostic factors in resected pancreatic adenocarcinoma: analysis of actual 5-year survivors. J Am Coll Surg 2004;198:722–31.[CrossRef][Medline]
  15. Simunovic M, To T, Theriault M, Langer B. Relation between hospital surgical volume and outcome for pancreatic resection for neoplasm in a publicly funded health care system [see comment]. CMAJ 1999;160:643–8.[Abstract]
  16. Yeo CJ, Cameron JL, Sohn TA, et al. Six hundred fifty consecutive pancreaticoduodenectomies in the 1990s: pathology, complications, and outcomes. Ann Surg 1997;226:248–57; discussion 257–60.[CrossRef][Medline]
  17. Surveillance, Epidemiology, and End Results (SEER) Program (http://www.seer.cancer.gov) Public-Use Data (1973–2000), National Cancer Institute DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2003, based on the November 2002 submission; 2003.
  18. Surveillance, Epidemiology, and End Results (SEER) Program Website. http://www.seer.cancer.gov. 2004, accessed September 24, 2004.
  19. Stata version 8.2. College Station (TX): Stata Corporation.
  20. Hosmer DW, Lemeshow S. Applied survival analysis: regression modeling of time to event data. Wiley series in probability and statistics. New York: John C. Wiley & Sons; 1999.
  21. Sahmoun AE, D'Agostino RA Jr, Bell RA, Schwenke DC. International variation in pancreatic cancer mortality for the period 1955–1998. Eur J Epidemiol 2003;18:801–16.[CrossRef][Medline]
  22. Coughlin SS, Calle EE, Patel AV, Thun MJ. Predictors of pancreatic cancer mortality among a large cohort of United States adults. Cancer Causes Control 2000;11:915–23.[CrossRef][Medline]
  23. Maire F, Hammel P, Terris B, et al. Prognosis of malignant indraductal papillary mucinous tumours of the pancreas after surgical resection. Comparison with pancreatic ductal adenocarcinoma. Gut 2002;51:717–22.[Abstract/Free Full Text]
  24. Sener SF, Fremgen A, Menck HR, Winchester DP. Pancreatic cancer: a report of treatment and survival trends for 100,313 patients diagnosed from 1985–1995, using the National Cancer Database [see comment]. J Am Coll Surg 1999;189:1–7.[CrossRef][Medline]
  25. Li D, Xie K, Wolff R, Abbruzzese JL. Pancreatic cancer. Lancet 2004;363:1049–57.[CrossRef][Medline]
  26. Brawley OW. Hormonal prevention of prostate cancer. Urol Oncol 2003;21:67–72.[Medline]
  27. Collins J, Crosignani PG, The ESHRE Capri Workshop Group. Hormones and breast cancer. Hum Reprod Update 2004;10:281–93.[Abstract/Free Full Text]
  28. Skinner HG, Michaud DS, Colditz GA, et al. Parity, reproductive factors, and the risk of pancreatic cancer in women. Cancer Epidemiol Biomarkers Prev 2003;12:433–8.[Abstract/Free Full Text]
  29. Kreiger N, Lacroix J, Sloan M. Hormonal factors and pancreatic cancer in women. Ann Epidemiol 2001;11:563–7.[CrossRef][Medline]
  30. Chang KJ, Parasher G, Christie C, Largent J, Anton-Culver H. Risk of pancreatic adenocarcinoma: disparity between African Americans and other race/ethnic groups. Cancer 2005;103:349–57.[Medline]
  31. Longnecker DS, Karagas MR, Tosteson TD, Mott LA. Racial differences in pancreatic cancer: comparison of survival and histologic types of pancreatic carcinoma in Asians, Blacks, and Whites in the United States. Pancreas 2000;21:338–43.[CrossRef][Medline]
  32. Schwartz KL, Crossley-May H, Vigneau FD, Brown K, Banerjee M. Race, socioeconomic status and stage at diagnosis for five common malignancies. Cancer Causes Control 2003;14:761–6.[CrossRef][Medline]
  33. DeNavas-Walt C, Proctor BD, Mills RJ. Income, poverty, and health insurance coverage in the United States: 2003. Washington (DC): U.S. Department of Commerce, Economics, and Statistics Administration; 2004.
  34. Hall SE, Holman CD, Wisniewski ZS, Semmens J. Prostate cancer: socio-economic, geographical, and private-health insurance effects on care and survival. BJU Int 2005;95:51–8.[CrossRef][Medline]
  35. Kelz RR, Gimotty PA, Polsky D, Norman S, Fraker D, DeMichele A. Morbidity and mortality of colorectal carcinoma surgery differs by insurance status. Cancer 2004;101:2187–94.[Medline]
  36. Okunade AA, Karakus MC. Mortality from breast carcinoma among US women: the role and implications of socio-economics, heterogeneous insurance, screening mammography, and geography. Health Care Manag Sci 2003;6:237–48.[CrossRef][Medline]
  37. Niedergethmann M, Rexin M, Hildenbrand R, et al. Prognostic implications of routine, immunohistochemical, and molecular staging in resectable pancreatic adenocarcinoma. Am J Surg Pathol 2002;26:1578–87.[CrossRef][Medline]
  38. Pernick NL, Sarkar FH, Philip PA, et al. Clinicopathologic analysis of pancreatic adenocarcinoma in African Americans and Caucasians. Pancreas 2003;26:28–32.[CrossRef][Medline]
  39. Anton-Culver H, Ziogas A, Bowen DJ, et al. The Cancer Genetics Network: recruitment results and pilot studies. Community Genet 2003;6:171–7.[Medline]



This article has been cited by other articles:


Home page
 Ann OncolHome page
T. R. Halfdanarson, K. G. Rabe, J. Rubin, and G. M. Petersen
Pancreatic neuroendocrine tumors (PNETs): incidence, prognosis and recent trend toward improved survival
Ann. Onc., October 1, 2008; 19(10): 1727 - 1733.
[Abstract] [Full Text] [PDF]

Home page
 Endocr Relat CancerHome page
T. R Halfdanarson, J. Rubin, M. B Farnell, C. S Grant, and G. M Petersen
Pancreatic endocrine neoplasms: epidemiology and prognosis of pancreatic endocrine tumors
Endocr. Relat. Cancer, June 1, 2008; 15(2): 409 - 427.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
J. A. Zell, J. M. Rhee, A. Ziogas, S. M. Lipkin, and H. Anton-Culver
Race, Socioeconomic Status, Treatment, and Survival Time among Pancreatic Cancer Cases in California
Cancer Epidemiol. Biomarkers Prev., March 1, 2007; 16(3): 546 - 552.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Fesinmeyer, M. D.
Right arrow Articles by Bowen, D. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Fesinmeyer, M. D.
Right arrow Articles by Bowen, D. J.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online