Fax: (713) 745-1462
The impact of ethnicity on the presentation and prognosis of patients with gastric adenocarcinoma†
Results from the National Cancer Data Base
Version of Record online: 13 JUN 2008
Copyright © 2008 American Cancer Society
Volume 113, Issue 3, pages 461–469, 1 August 2008
How to Cite
Al-Refaie, W. B., Tseng, J. F., Gay, G., Patel-Parekh, L., Mansfield, P. F., Pisters, P. W. T., Yao, J. C. and Feig, B. W. (2008), The impact of ethnicity on the presentation and prognosis of patients with gastric adenocarcinoma. Cancer, 113: 461–469. doi: 10.1002/cncr.23572
Presented as a Merit Award oral presentation at the 2006 Annual Gastrointestinal Cancer Symposium, San Francisco, California.
- Issue online: 18 JUL 2008
- Version of Record online: 13 JUN 2008
- Manuscript Accepted: 25 MAR 2008
- Manuscript Revised: 19 MAR 2008
- Manuscript Received: 27 AUG 2007
- gastric adenocarcinoma;
- National Cancer Data Base
Regional-based studies have indicated that ethnicity is associated with presentation and outcome in patients with gastric adenocarcinoma. To validate this observation in a large cohort, the authors of this report used the National Cancer Data Base (NCDB) to determine whether self-reported ethnicity influences presentation and survival in this patient population.
Patient demographics, tumor-related features, and treatment-related features were analyzed by ethnicity. Univariate analyses were performed using the chi-square test. Overall median and relative survival rates were examined by using the Kaplan-Meier method. Cox proportional-hazards models were used to identify the predictors of survival outcomes.
Between 1995 and 2002, 81,095 cases of gastric adenocarcinoma were entered into the NCDB. There were 57,943 white patients (71.5%), 11,094 African-American patients (13.7%), 5665 Hispanic patients (7%), 4736 Asian/Pacific Islander (API) patients (5.8%), and 1657 patients of other ethnicities (2%). Significant differences were observed according to ethnicity among the variables that were compared (all P < .01). In patients with stage I and II disease, the 5-year relative survival rates for APIs (stage I, 77.2%; stage II, 48%) were more favorable than for whites (stage I, 58.7%; stage II, 32.8%), African Americans (stage I, 55.9%; stage II, 37.9%), and Hispanics (stage I, 60.8%; stage II, 39.3%). The overall median survival of APIs was more favorable than that of others (P < .01). Predictors of a better outcome were Asian race, female sex, younger age, earlier stage, lower grade, distal tumors, multimodality treatment, and care at a teaching hospital.
Ethnicity was associated with differences in presentation and outcome of patients with gastric adenocarcinoma. APIs had a more favorable outcome than patients of other ethnicities. Further studies should target underlying biologic and socioeconomic factors to explain these differences. Cancer 2008. © 2008 American Cancer Society.
Although the incidence of gastric adenocarcinoma continues to decline, it remains a common cause of cancer-related deaths in the world. In 2005, the American Cancer Society estimated that 21,860 new cases of adenocarcinoma of the stomach would be diagnosed and that 11,550 patients would die from this disease in the U.S. in 2005. These findings makes gastric cancer the 14th most common cancer and the eighth leading cause of cancer death in the U.S.1
Several institutional and regional studies have evaluated the relation between ethnicity and outcome among patients with gastric adenocarcinoma.2–7 Over the past 5 years, 2 population-based analyses, 1 conducted in British Columbia, Canada, and the other in southern California, revealed that self-identified Asian race/ethnicity, including patients of Asian descent, had distinct tumor-related features and a more favorable outcome in compared with individuals from other ethnic backgrounds with the disease.3, 5 Furthermore, a recent retrospective study from the University of Texas M. D. Anderson Cancer Center demonstrated that Hispanic patients with gastric adenocarcinoma displayed unique clinicopathologic features.7 Although Hispanic ethnicity did not influence survival in the latter study, Asian race remained a favorable predictor of outcome. Stage migration,8–11 treatment differences,12–14 and differences in disease behavior15–19 have been proposed to explain the differences in outcomes between Japanese patients and Western patients.
Although the majority of North American regional-based studies have focused on differences in outcomes between Asian and non-Asian patients, to our knowledge, patterns of disease presentation and survival rates of other ethnicities have not been evaluated in the setting of large-scale U.S. databases. Furthermore, there is a paucity of epidemiological studies evaluating these data in patients of varied ethnicities among Western populations. Therefore, we performed a large, hospital-based analysis using the National Cancer Data Base (NCDB) to determine whether self-reported ethnicity is associated with the presentation, treatment received, and prognosis of patients with gastric adenocarcinoma.
MATERIALS AND METHODS
The NCDB is a joint project of the Commission on Cancer of the American College of Surgeons and the American Cancer Society. The NCDB was established in 1989 to serve as a comprehensive clinical surveillance resource for cancer care in the U.S.. Approximately 1430 hospitals currently participate in the approvals program and respond to annual calls for data to be delivered to the NCDB. Hospital cancer registrars abstract each case according to a standardized set of data elements and definitions. Three types of hospitals contribute to the NCDB including: teaching hospital cancer programs, community hospital comprehensive cancer programs, and community hospital cancer programs.
Case Selection Criteria
In total, 81,095 cases of gastric adenocarcinoma (International Classification of Disease for Oncology codes C16.0 to C16.9) were reported to the NCDB from 1995 to 2002. Tumor histology included adenocarcinoma, not otherwise specified (NOS); intestinal type adenocarcinoma; carcinoma, NOS; mucinous adenocarcinoma; signet ring cell carcinoma; and other. Patients with other gastric neoplasms, such as neuroendocrine tumors, lymphoma, or sarcoma, were excluded from this analysis.
For each ethnicity (whites, African Americans, Hispanics, Asians/Pacific Island descendants [APIs], and others), we analyzed patient demographics, tumor-related clinicopathologic features, operations performed, and other treatment modalities. Because other ethnicities accounted for <1700 of 81,095 patients (2%), these groups were not included in the analyses. Factors that were included in patient demographics were race/ethnicity, age, and sex. For each ethnicity, age was divided into 3 categories: age <50 years, ages 50 to 69 years and age ≥70 years, because this age categorization represents young, middle aged, and elderly individuals. Reporting hospitals where patients received their care were grouped as described above.
Tumor-related clinicopathologic features included disease stage at presentation, primary tumor location, and histologic grade of differentiation. Stages at presentation were divided into I, II, III and IV and were defined according to the fourth edition (for the years 1995-1997) and the fifth edition (for the years 1998-2002) of the American Joint Committee on Cancer (AJCC) staging system.20 Primary tumor locations were divided into cardia, fundus, body, antrum, pylorus, lesser curvature, greater curvature, overlapping lesions, and others. To be consistent with other publications, these primary tumor locations were categorized as cardia, fundus and body, antrum and pylorus, lesser curvature, greater curvature, and overlapping lesions. Histologic grade was divided into 1 (well differentiated), 2 (moderately differentiated), 3 (poorly differentiated), and 4 (undifferentiated).
Types of treatment received treatment by patients with gastric adenocarcinoma included types of gastric resection, other surgical, and nonsurgical therapeutic modalities. Types of gastric resection were local, proximal gastrectomy, distal gastrectomy (including antrectomy, near total, and total gastrectomy), gastrectomy with en-bloc resection of other organs, and other. Treatment modalities included 1)either chemotherapy only or radiotherapy only (no surgery), 2) combined chemoradiation without surgery, 3) surgical treatment only, 4) surgical with single-modality treatment (chemotherapy or radiotherapy), or 5) surgical resection, including chemoradiation therapy and others.
The American College of Surgeons has executed a Business Associate Agreement that includes a data-use agreement with each of its Commission on Cancer-approved hospitals. Data reported to the NCDB are retrospective and include no patient or physician names. Analyses are reported only at the aggregate level and are used to assist hospital cancer programs with quality assurance rather than to make decisions about individuals and their care. Results reported in this study are in compliance with the privacy requirements of the Health Insurance Portability and Accountability Act of 1996 as reported in the Standards for Privacy of Individually Identifiable Health Information; Final Rule (Title 45, Code of Federal Regulations, Parts 160 and 164).
All analyses were performed using the SPSS statistical software package (SPSS for Windows, version 14.0; SPSS Inc, Chicago, Ill). The chi-square test was used for comparisons of proportions across levels of categorical variables. Results were based on 2-sided tests with a significance level of .01. Tests were adjusted for all pairwise comparisons using the Bonferroni correction. We used the Bonferroni correction with comparison of ethnicity as a safeguard against multiple tests of statistical significance on the same data, in which 1 of every 100 hypothesis tests will appear to be significant at the α = .01 level purely because of chance.
Survival rates of patients with gastric adenocarcinomas for the 1995 through 1997 cohort were calculated. The date of diagnosis was defined as the starting point of survival and date of last contact, and vital status at that time was the endpoint. Patients who remained alive at the end of 1997 were censored. Five-year survival rates were available for the 1995 through 1997 cohort (n = 32,703 of the entire cohort of 81,095 patients). The relative survival rate was calculated by using the procedure described by Ederer et al21 (ie, the ratio of the observed survival rate to the aggregated survival rate of individuals of the same sex, age, and ethnic background). That rate represents the likelihood that a patient will not die from causes associated specifically with their cancer at some specified time after diagnosis.
Overall median survival rates were determined by using the Kaplan-Meier method for each ethnicity with differences assessed by the log-rank test; then, the rates were compared by using a chi-square test. A multivariate Cox proportional-hazards model was fitted to the data for cases diagnosed in 1995, 1996, and 1997. A forward stepwise method using the Wald statistical test was used to identify changes between steps. Initially, both Kaplan-Meier univariate approaches were used. Then, a univariate Wald test was used to provide guidance during the model-building process to identify possible variables that could be eliminated without affecting model performance. A reduced model was built without interaction terms, and the performance of the models was compared by using the Wald test again. The coefficients in the reduced model were significant; interactive terms did not contribute to the model. The significance level was set at .05 to identify the effects of potential prognostic factors (ethnicity, stage, grade, site, sex, age category, treatment, and hospital type) on survival based on the likelihood ratio test to identify significant simultaneous prognostic covariates, because they all were included in the regression model. We tested for collinearity by using the tolerance test. The tolerance statistics ranged from .924 to .990, indicating low multicollinearity. We assessed the proportional-hazards assumption through graphic inspection of the log-minus-log survival through the partial residuals. There was no indication of violation of the assumptions. We tested for interactions between ethnicity and the following factors: age category, primary tumor location in the stomach, histologic grade, and types of treatment. None of the interactions were statistically significant. Therefore, we did not include any interactions in our final model. In addition, and after removing ‘hospital setting’ from our model, we reran our Cox proportional-hazards model and observed that the model was not significantly different from the model that included hospital setting. To test the adequacy of our model, we used a Cox proportional-hazards model (forward step) to evaluate the following variables as predictors off survival: race, age category, sex, primary tumor location, histologic grade, AJCC stage, first course of treatment, and type of treating hospital. At each step, the stepwise methods added the variable with the highest significant score. The residual chi-square was computed. Then, the variable with the next highest score statistic was entered. The model coefficients were tested by using the −2 log-likelihood method to test the change from the prior step and the change from the previous block. The criterion for selection was based on the Wald statistic at a P < .05.
From 1995 through 2002, data for 81,095 patients with a diagnosis of gastric adenocarcinoma were entered into the NCDB. Of these 81,095 patients, 57,943 (71.5%) were white, 11,094 (13.7%) were African American, 5665 (7%) were Hispanic, 4736 (5.8%) were APIs, and 1657 (2%) were other (Table 1). The results indicated that Hispanics were more likely to be in the youngest age group (aged <50 years) at diagnosis than patients of all other ethnicities (P < .01), whereas whites were more likely to be older (aged ≥70 years) at diagnosis than patients of all other ethnicities (P < .01). Furthermore, white patients with gastric adenocarcinoma more often were men (P < .01). Whites were more likely to receive their healthcare at community cancer centers (P < .01) and comprehensive community cancer centers (P < .01) than African Americans, APIs, and others. African Americans were more likely to be treated at a teaching/research hospital (P < .01) than whites, Hispanics, and APIs. APIs received care at a comprehensive community cancer center more often (P < .01) than African Americans, Hispanics, and others.
|Characteristic||No. of patients (%)||P*|
|Whites (n=57,943)||African-American (n=11,094)||Hispanics (n=5665)||Asians (n=4736)||Others (n=1657)|
|<50||4801 (8.3)||1715 (15.5)||1200 (21.2)||777 (16.4)||330 (19.9)||<.01|
|50–69||21,193 (36.6)||4364 (39.3)||2348 (41.4)||1909 (40.3)||700 (42.2)|
|>70||31,949 (55.1)||5015 (45.2)||2117 (37.4)||2050 (43.3)||627 (37.8)|
|AJCC stage at presentation|
|I||11,665 (20.2)||2271 (20.5)||1024 (18.1)||1128 (23.9)||337 (20.4)||<.01|
|II||8281 (14.3)||1379 (12.5)||746 (13.2)||628 (13.3)||245 (14.8)|
|III||14,306 (24.8)||2491 (22.5)||1384 (24.5)||1194 (25.3)||411 (24.8)|
|IV||23,506 (40.7)||4916 (44.5)||2492 (44.1)||1772 (37.5)||661 (40.0)|
|Primary tumor location|
|Cardia||21,705 (37.5)||1119 (10.1)||802 (4.2)||460 (9.7)||430 (26)||<.01|
|Fundus plus body||5770 (10)||1285 (11.6)||794 (14)||569 (12)||176 (10.6)|
|Antrum plus pylorus||9900 (17.1)||3308 (29.8)||1422 (25.1)||1551 (32.7)||310 (18.7)|
|Lesser curvature||4543 (7.8)||1362 (12.3)||1363 (12.3)||649 (13.7)||177 (10.7)|
|Greater curvature||2334 (4)||544 (4.9)||295 (5.2)||280 (5.9)||55 (3.3)|
|Overlapping lesions||4191 (7.2)||883 (8)||582 (10.3)||505 (10.7)||129 (7.8)|
|Others||9500 (16.4)||2592 (23.4)||1125 (19.9)||722 (15.2)||380 (22.9)|
|1||2349 (4.6)||506 (5.3)||171 (3.4)||149 (3.5)||63 (4.4)||<.01|
|2||14789 (29.1)||2912 (30.4)||1196 (24)||1118 (26)||336 (23.4)|
|3||32185 (63.3)||6014 (62.7)||3503 (70.3)||2938 (68.3)||931 (64.9)|
|4||1459 (2.9)||157 (1.6)||110 (2.2)||97 (2.3)||104 (7.3)|
|Type of treatments received|
|Chemotherapy only or radiotherapy only||7744 (13.4)||1405 (12.7)||792 (14)||472 (10)||75 (4.5)||<.01|
|Chemoradiation therapy||3655 (6.3)||394 (3.6)||229 (4)||137 (2.9)||664 (40.1)|
|Surgical resection only||23,094 (39.8)||4494 (40.4)||2089 (36.8)||2190 (46.2)||164 (9.9)|
|Surgery with single modality||5484 (9.5)||1108 (10)||700 (12.4)||600 (12.7)||187 (11.3)|
|Surgical resection with chemoradiotherapy||6194 (10.7)||961 (8.7)||604 (10.7)||560 (11.8)||337 (20.3)|
|No therapy||11,008 (19)||2605 (23.4)||1176 (20.8)||737 (15.6)|
|Others||764 (1.3)||129 (1.2)||75 (1.3)||40 (0.8)||210 (12.7)||2|
|Teaching/research||9907 (17.7)||1693 (15.7)||1157 (21)||675 (14.6)||187 (11.5)||<.01|
|Community cancer||24,041 (42.9)||3453 (32)||1905 (34.7)||1989 (42.9)||488 (30.1)|
|Comprehensive community cancer||20,600 (36.8)||5409 (50.1)||2285 (41.6)||1902 (41)||918 (56.7)|
|Others||1428 (2.6)||251 (2.3)||150 (2.7)||71 (1.5)||27 (1.7)|
Tumor-related Clinical Features
Stage at presentation
When the ethnic groups were compared from the standpoint of stage at presentation, African Americans and Hispanics (P < .01) were more likely to present with more advanced AJCC stage disease (stage IV), whereas APIs (P < .01) were more likely to present with AJCC stage I disease.
Location of primary tumor
Tumors of the cardia were more common in whites (P < .01) than in patients of other ethnicities. Tumors of the antrum (P < .01) and pylorus (P < .01) were more common in African Americans and APIs than in other ethnicities.
Histologic grade of differentiation
Whites (P < .01) and African Americans (P < .01) were more likely to present with grade 2 histologic differentiation than patients of the other ethnicities. Grade 3 histologic differentiation was more prevalent in Hispanics and APIs (P < .01) than in whites, African Americans, and others.
Types of gastric resection
Compared with the other ethnicities, whites were more likely to undergo upper proximal gastrectomy, esophagogastrectomy, and subtotal gastrectomy, including a portion of the esophagus (P < .01). APIs were more likely than the other ethnicities to require a distal gastrectomy (P < .01).
Types of received treatment modalities
APIs were less likely than the other ethnicities to receive a single nonsurgical treatment (P < .01) or to undergo gastrectomy only than patients in the other ethnic groups (P < .01). Whites were more likely to be treated with combined chemoradiation therapy than patients in the other ethnic groups (P < .01). Hispanics and APIs were more likely to undergo surgical resection plus receive a single treatment modality compared with the other ethnic groups (P < .01). African Americans were less likely to undergo any form of gastric resection only (P < .01) or surgical resection plus chemoradiation treatment (P < .01).
Survival data from the NCDB were available for cases of gastric adenocarcinoma that were diagnosed from 1995 through 1997 (n = 31,703). The overall median survival was 9.7 months for white patients, 9.6 months for African American patients, 10.8 months for Hispanic patients, and 14.8 months for API patients (Fig. 1). Of all the ethnic groups, APIs had the most favorable outcome (P < .01). Among the patients with stage I disease in this cohort, the 5-year relative survival rate for APIs (77.2%; 95% confidence interval [95% CI], 71%–83%) was more favorable than that for whites (58.7%; 95% CI, 56%–61%), African Americans (55.9%; 95% CI, 51%–61%), and Hispanics (60.8%; 95% CI, 52%–69%). Among the patients with stage II or III disease, the 5-year relative survival rate for APIs (stage II: 48%; 95% CI, 39%–57%; stage III: 19%; 95% CI, 15%–23%) was better than that for whites (stage II: 32.8%; 95% CI, 31%–35%; stage III: 13%; 95% CI, 12%–14%) (Fig. 2). Cox proportional-hazards regression analysis demonstrated that favorable predictors of outcome were Asian race, being a woman, younger age, earlier stage, lower grade, distal tumors, multimodality treatment, and care at a teaching hospital (P < .01) (Table 2).
|Primary tumor location|
|Chemotherapy or radiotherapy (no surgery)||1.46||1.39–1.54||<.0001|
|Combined surgery plus chemoradiotherapy||0.80||0.77–0.83||<.0001|
|Comprehensive community cancer||1.09||1.06–1.13||<.0001|
The current findings clearly demonstrate that ethnicity is associated with the presentation, treatment delivered, and outcome in patients with gastric adenocarcinoma. API patients as a group can expect the most favorable outcome, as indicated by the overall median survival and 5-year relative survival rates for API patients with stage I and II disease. In addition, we observed that African-American patients were the least likely of all ethnic groups to receive multimodality treatment and were more likely to receive no therapy.
In agreement with our findings, previous retrospective studies conducted in various regions of North America also have demonstrated superior outcomes for Asians with adenocarcinomas of the stomach. For example, in a 2003 study in which the records of 2043 patients with gastric cancer who were registered in the British Columbia Cancer Agency in Canada were reviewed, Gill et al3 observed that the Asian patients in their study were younger, showed a greater propensity for distal gastric cancer, and harbored signet ring cell histology at the time of presentation more often than non-Asian patients. In that study, the overall 5-year survival rate was 19.3% for Asians and 11.4% for non-Asians (P = .0016). However, Asian ethnicity was not identified as an independent predictor of survival outcome (hazards ratio [HR], 0.89; 95% CI, 0.74–1.08). Likewise, Theuer et al22 reviewed data on 3770 patients with gastric carcinoma who were registered in 3 Southern California cancer registries from 1984 to 1996. Those authors observed that Asians were more likely than non-Asians to have localized cancers and antral locations and were less likely to have distant metastases. Similarly, non-Asians were at a higher risk of death from their gastric malignancy compared with Asians (HR, 1.21; 95% CI, 1.04–1.42 [P < .05]). In another study that examined ethnicity-related differences in patients with gastric adenocarcinomas, Stemmermann et al23 observed that, after adjustments for age, gastric tumors in the cardia were more common in white men (41.8%) than in women of other ethnicities (13.4%; P = .0002). These findings come from an analysis of 532 patients with gastric adenocarcinoma recorded over 6 years in the Hawaii Tumor Registry. Likewise, after adjustments for age, the percentage of cancers of the cardia was higher among white Hawaiian women (22.4%) than among women of other ethnicities (7.3%; P = .08).
In agreement with several studies, we observed that white patients with gastric adenocarcinoma were more likely to be men and to present with proximal tumors, both of which were identified as predictors of a worse outcome on Cox regression multivariate analysis. Some researchers have suggested that proximal tumors constitute different underlying disease biology. Because our current study is representative nationally, these findings also may be attributed to the typically advanced nature of proximal tumors at presentation or to the possibility that these cancers constitute a different disease.6, 24, 25
Yao et al7 assessed the characteristics and outcome of 301 Hispanic patients with gastric adenocarcinoma among 1897 such patients who were seen at the University of Texas M. D. Anderson Cancer Center. In that study, Hispanics were more likely to be younger, to have distal tumors, and to develop peritoneal metastasis than other ethnicities. Hispanics also were less likely to undergo total gastrectomy and to develop liver metastasis than whites. A multivariate Cox regression analysis also demonstrated that Asian ethnicity was a favorable predictor of outcome.
The current study was unique, because it included a large and diverse population of patients drawn from hospital-based registries throughout the U.S. and from a contemporary time period. However, there were limitations. First, the adequacy of lymphadenectomy, which is considered an important element of staging and outcome in gastric cancer,26–28 was not evaluated in this study. Second, the follow-up data we examined in this study were for patients who were diagnosed between 1995 and 1997. Because the NCDB typically calls for follow-up data every 5 years, this represents the most up-to-date survival data that are available throughout the database. However, the relative survival analyses in the current study represented over 31,000 patients. Another potential limitation is that it was not possible to analyze the ethnic differences in our study from the standpoint of certain risk exposures that may be associated with ethnicity. Beyond this point, the current review was not designed as a case–control study to look for causation; rather, it was a retrospective and longitudinal examination.
Our study has several important implications. First, our results suggest that ethnic differences may have an impact on the receipt of adequate treatment for gastric adenocarcinoma. For example, low socioeconomic status and difficult access to healthcare may contribute to the finding that stage IV disease was more common at presentation in African-American patients than in patients of other ethnicities. The question of whether ethnic-, socioeconomic-, or treating facility-related factors have an impact on the delivery of multimodality treatment for patients with gastric cancer in the U.S. warrants further investigation. Second, we and others have demonstrated that ethnicity is associated with presentation, treatment, and outcome in patients with gastric adenocarcinoma.3, 5, 7, 29 Ethnicity may be an additional factor that should be integrated into the recently published prognostic nomogram developed by researchers at Memorial Sloan-Kettering Cancer Center that is used to determine gastric cancer disease-specific survival after total (R0) resection.28 The utility of prognostic variables should be maximized in discussions with our patients, in treatment decision making, and in designs of future therapy protocols. Third, in view of the consistent findings of favorable survival rates in the Asian population living in the U.S., studies are needed to address the controversies that still surround the survival differences between Japanese and Western patients with gastric cancer. To our knowledge to date, there is no clearly accepted explanation for these differences: The multitude of theories includes stage migration, extent of lymphadenectomy, and possible differences in tumor biology. Another important area of research that is used to address this controversy is the identification of molecular markers in gastric adenocarcinoma. Already, several investigators have evaluated the role of molecular makers, including p53, E-cadherin, c-erb, and c-myc, as prognostic tools in patients with gastric cancer.30–33 Evolving data from gene expression profiling experiments using combinational DNA microarray techniques and logistic regression models have demonstrated that such markers appear to predict outcome after gastrectomy in these patients.34 There is also a clear need for studies assessing the molecular and genetic differences among ethnic groups that can explain the various biologic characteristics and outcomes in patients with gastric adenocarcinomas. The information yielded by such studies will be valuable in guiding treatment.
In this large, hospital-based study from the NCDB, ethnicity was associated with differences in presentation and the outcome of patients with gastric adenocarcinomas. Among patients with stage I and II disease, API descents had a more favorable outcome than patients in other ethnic groups. Further studies should focus on identifying any underlying biologic and socioeconomic factors that may be responsible for these differences and their impact on the delivery of adequate care for patients with gastric cancer.
We thank Beth Notzon for her expert help in editing this article.
- 4Marked multi-ethnic variation of esophageal and gastric cardia carcinomas within the United States. Am J Gastroenterol. 2004; 99: 582–588., .Direct Link:
- 20Greene FL,Page DL,Fleming ID, et al, eds. AJCC Cancer Staging Manual,6th ed. New York, NY: Springer-Verlag; 2002.
- 24The National Cancer Data Base Report on poor survival of U.S. gastric carcinoma patients treated with gastrectomy: 5th edition. American Joint Committee on Cancer staging, proximal disease, and the “different disease” hypothesis. Cancer. 2000; 88: 921–932., , .
- 30The prognostic significance of p53, p27 kip1, p21 waf1, HER-2/neu, and Ki67 proteins expression in gastric cancer: a clinicopathological and immunohistochemical study of 121 Arab patients. J Surg Oncol. 2005; 91: 243–252., , , et al.