A phase II trial of neoadjuvant cisplatin-fluorouracil followed by postoperative intraperitoneal floxuridine-leucovorin in patients with locally advanced gastric cancer.

BACKGROUND: The aim of the study was to evaluate the efficacy and toxicity of neoadjuvant chemotherapy with intravenous (i.v.) cisplatin and fluorouracil (5-FU), surgery and postoperative intraperitoneal (i.p.) floxuridine (FUdR) and leucovorin (LV) in patients with locally advanced gastric cancer. PATIENTS AND METHODS: Preoperative staging was confirmed by laparoscopy (LAP). Two cycles of i.v. cisplatin (20 mg/m(2)/day, rapid infusion) and 5-FU (1000 mg/m(2), continuous 24-h infusion), given on days 1-5 and 29-34, were followed by a radical gastrectomy and a D2 lymphadenectomy. Patients having R0 resections were to receive three cycles of i.p. FUdR (1000 mg/m(2)) and LV (240 mg/m(2)), given on days 1-3, 15-17 and 29-31. Intraperitoneal chemotherapy was begun 5-10 days from surgery. RESULTS: Thirty-eight patients were treated. Both preoperative and postoperative chemotherapy were well tolerated. T stage downstaging (pretreatment LAP versus surgical pathological stage) was seen in 23% of patients. The R0 resection rate was 84%. Neither an increase in postoperative morbidity nor operative mortality was noted. With a median follow-up of 43.0 months, 15 patients (39.5%) are still alive (median survival 30.3 months). Good pathologic response, seen in five patients (15%), was associated with better survival (P = 0.053). Peritoneal and hepatic failures were found in 22% and 9% of patients, respectively. Quality of life seemed to be preserved. CONCLUSIONS: Neoadjuvant cisplatin/5-FU followed by postoperative i.p. FUdR/LV can be safely delivered to patients undergoing radical gastrectomy and D2 lymphadenectomy. The R0 resection and the survival rates are encouraging. An association between pathologic response and patient outcome was suggested.

[Influence of laparoscopic staging on therapy of stomach carcinoma]. / Einfluss des laparoskopischen Staging auf die Therapie des Magenkarzinoms.

Surgical staging is the gold standard in assessing the extent of disease in gastric cancer. The introduction of video-assisted laparoscopy has allowed the surgeon the opportunity to maintain high quality with low morbidity and reduced length of stay for the patient. The body of evidence supporting its use in routine staging of gastric cancer is significant and now the potential to expand the role of laparoscopy into the area of new therapeutic procedures can be explored.

Analysis of 154 actual five-year survivors of gastric cancer.

Gastric cancer patients in the United States have a poor prognosis with a collective 5-year survival rate of less than 15%. We identified a subset of actual 5-year survivors (long-term survivors) and analyzed clinicopathologic variables predictive of recurrence and survival beyond the 5-year mark. A review of our prospective database from July 1985 to February 1993 revealed 154 patients who were long-term survivors and 280 patients who died of disease prior to 5 years (short-term survivors) following curative resection (R0). Tumor (T) stage, nodal (N) status, tumor location, and median number of positive nodes were compared between the two groups. Univariate and multivariate analysis of disease-free and greater than 5-year disease-specific survival was performed within the long-term survivors. Among the long-term survivors, 29% were classified as "early gastric cancers" (T1NX). The median number of positive nodes (0 vs. 5; P <0.001) and percentage of lesions that were T1/T2 (60% vs. 19%; P <0.001), node negative (58% vs. 15%; P <0.001), or proximal (40% vs. 65%; P <0.001) was significantly different in long-term survivors vs. short-term survivors, respectively. Of the 154 five-year survivors, gastric cancer recurred in 23, and 13 patients (8%) died of the disease at a median of 84 months from the original diagnosis. On univariate and multivariate analysis of prognostic factors in the long-term survivors, only the Lauren histologic classification predicted disease-specific and disease-free survival with diffuse histologic types faring significantly less well. T stage and N status are powerful prognostic factors of outcome within the first 5 years after curative resection of gastric carcinoma. However, the Lauren histologic type emerges as the dominant predictor of outcome once a patient with gastric cancer has survived for 5 years or more.

Lymph node staging in gastric cancer: is location more important than Number? An analysis of 1,038 patients.

OBJECTIVE: To compare the impact of staging systems on the survival of 1,038 patients with gastric cancer undergoing resection for cure in a North American center. SUMMARY BACKGROUND DATA: In 1997, the American Joint Committee on Cancer (AJCC) and the Union Internationale Contre le Cancer redefined N stage in gastric cancer. The number of involved nodes rather than their location defines N, and a minimum of 15 examined lymph nodes is recommended for adequate staging. In the 1988 AJCC N-staging system, N1 and N2 node metastases were defined as within 3 cm or more than 3 cm of the primary; the 1997 AJCC N stages were defined as N1 = 1 to 6 positive nodes, N2 = 7 to 15 positive nodes, and N3 = more than 15 positive nodes. METHODS: Between 1985 and 1999, 1,038 patients underwent an R0 resection. Median and 5-year survival rates were compared and the Kaplan-Meier method was used to estimate median survival. RESULTS: The location of positive nodes did not significantly affect median survival when analyzed by the number of positive nodes. In contrast, the number of positive lymph nodes had a profound influence on survival. The new N categories served as a better discriminator of median survival when 15 or more nodes were examined. Survival estimates for stages II, IIIA, and IIIB were significantly influenced by examining 15 or more nodes. CONCLUSION: The number of positive nodes best defines the prognostic influence of metastatic lymph nodes in gastric cancer. Survival estimates based on the number of involved nodes are better represented when at least 15 nodes are examined.

Gastric cancer in young patients: demographic, clinicopathological, and prognostic factors in 92 patients.

BACKGROUND: This investigation was undertaken to define the demographic, clinicopathological, and prognostic factors relevant to young patients with gastric adenocarcinoma. METHODS: A prospective database of all patients with gastric cancer who presented to Memorial Sloan-Kettering Cancer Center was started in 1985. Clinical, pathological, and operative records and follow-up data on 92 patients, 40 years of age or younger, with a primary diagnosis of gastric cancer were reviewed. RESULTS: The mean patient age was 35 +/- 4.9 years (range, 17-40 years), and 52 were male. The male-to-female ratio of patients younger than 30 was 0.85/1; whereas in those older than 30, the ratio was 1.45/1. Sixty-six percent of the patients were white, 15% Asian, 11% Hispanic, and 8% were black American. Nineteen percent of patients reported a family history of gastric cancer. Sixty-six patients (71%) presented with stage III or IV disease, whereas 13 patients, each, presented with stage I or II disease. Poorly differentiated lesions were present in 71%. Resection with curative intent was undertaken in 47 patients, and resection with palliative intent was performed in 24 patients. Tumor site (proximal vs. distal vs. linitus plastica), advanced T stage, and the presence of nodal disease were significant predictors of disease-free survival on both univariate and multivariate analyses. The mean survival time and disease-specific 5-year survival rates for individual Union International Contre le Cancer tumor stages were similar to those observed in older populations of patients with gastric cancer; and eight patients, who presented with early (T1/T2) node-negative tumors, are alive and well a minimum of 60 months after resection. CONCLUSIONS: The high frequency of a positive family history in young patients suggests an opportunity to identify a high-risk population for screening.

Is gastric carcinoma different between Japan and the United States?

BACKGROUND: Analyses of surgical results for gastric carcinoma often lead to the conclusion that gastric carcinoma occurring in Japan is different from that diagnosed in the U.S. METHODS: To elucidate factors that might explain the differences in surgical results between the two countries, the authors compared data from a cancer center and a university hospital in Japan and a specialist cancer hospital in the U.S (Memorial Sloan-Kettering Cancer Center [MSKCC]). RESULTS: The mean age and body mass index were significantly greater in patients in the U.S. The N category appeared to be determined less accurately at MSKCC compared with the Japanese centers. The occurrence of early gastric carcinoma was not confined to Japanese patients because 20% of U.S. patients who underwent surgery were determined to have early stage disease. However, mucosal (in situ) carcinoma was detected rarely, and the proportion of advanced stage disease was greater in the U.S. Lesions in the upper gastric body, including the gastroesophageal junction, occurred in > 50% of cases at MSKCC but in only 20% of cases at the Japanese centers (P < 0.001). D2 lymph node dissection was possible with low morbidity and minimum mortality (31% and 3%, respectively, at MSKCC). The 5-year survival rates, stratified by tumor location and T category, revealed more similar results between Japan and the U.S. than had been reported previously. The marked difference between Japanese and American institutions only was observed for T1 and T2 tumors occurring in the lower gastric body and for T3 tumors occurring in the middle and upper third of the stomach. CONCLUSIONS: Based on the findings of the current study, it would appear that the more favorable outcome noted for gastric carcinoma patients in Japan primarily is explained by the differences in tumor location, a greater frequency of early stage disease, and more accurate staging compared with gastric carcinoma patients in the U.S. Results of gastric carcinoma treatment comparable to those obtained in Japan can be obtained in Western centers.

Is lymphadenectomy necessary for early gastric cancer?

BACKGROUND: Although early gastric cancer (T1, NX) in Japan has been reported to have an excellent prognosis, the experience with this cancer in the United States is limited. The treatment of these tumors in Japan is becoming less aggressive as "good prognostic factors" are increasingly recognized. Our objective was to identify predictors of nodal disease and survival in a large cohort of Western patients with T1 tumors. METHODS: A retrospective review of our prospective data base from July 1985 to March 1998 revealed 165 patients undergoing surgical resection for T1 gastric tumors. Clinicopathological factors analyzed and compared included presence of positive nodes, tumor size (> or =4.5 vs. <4.5 cm), depth (mucosal vs. submucosal), grade (poor vs. moderate and well), and tumor site (proximal vs. distal), presence of venous or perineural invasion, and Lauren's classification. Factors predicting lymph node involvement and disease-specific survival were evaluated by univariate and multivariate analysis. RESULTS: Median follow-up time was 36 months. The actuarial 5-year survival was 88%. Thirteen patients (8%) died of disease. Lymph node involvement was present in 31 tumors (19%), with a 5-year survival of 91% with negative nodes vs. 78% with positive nodes. On univariate and multivariate analysis, the presence of tumor submucosal invasion (P<.05), venous invasion (P = .02), and size of 4.5 cm and larger (P = .02) was significantly associated with an increased risk for nodal positivity. On univariate analysis of survival, node-positive tumors (P = .02) and tumors 4.5 cm and larger (P = .008) were significantly associated with decreased survival. On multivariate analysis, only node-positive tumors were significantly (P = .01) associated with decreased survival. Those tumors that were limited to the mucosa and less than 4.5 cm in size (n = 47) had a 4% rate of positive nodes. In contrast, those tumors that were 4.5 cm and larger and had penetrated into the submucosa (n = 16) had a 56% chance of positive nodes. CONCLUSIONS: Early gastric carcinoma in North America has an excellent prognosis, similar to that in Japan. Tumors that are limited to the mucosa and smaller than 4.5 cm could be considered for limited resection without lymphadenectomy.

Preoperative biliary drainage: impact on intraoperative bile cultures and infectious morbidity and mortality after pancreaticoduodenectomy.

Whether it is necessary to perform biliary drainage for obstructive jaundice before performing pancreaticoduodenectomy remains controversial. Our aim was to determine the impact of preoperative biliary drainage on intraoperative bile cultures and postoperative infectious morbidity and mortality following pancreaticoduodenectomy. We retrospectively analyzed 161 consecutive patients undergoing pancreaticoduodenectomy in whom intraoperative bile cultures were performed. Microorganisms were isolated from 58% of these intraoperative bile cultures, with 70% of them being polymicrobial. Postoperative morbidity was 47% and mortality was 5%. Postoperative infectious complications occurred in 29%, most commonly wound infection (14%) and intra-abdominal abscess (12%). Eighty-nine percent of patients with intra-abdominal abscess (P = 0.003) and 87% with wound infection (P = 0.003) had positive intraoperative bile cultures. Microorganisms in the bile were predictive of microorganisms in intraabdominal abscess (100%) and wound infection (69%). Multivariate analysis of preoperative and intraoperative variables demonstrated that preoperative biliary drainage was associated with positive intraoperative bile cultures (P <0.001), postoperative infectious complications (P = 0.022), intra-abdominal abscess (P = 0.061), wound infection (P = 0.045), and death (P = 0. 021). Preoperative biliary drainage increases the risk of positive intraoperative bile cultures, postoperative infectious morbidity, and death. Positive intraoperative bile cultures are associated with postoperative infectious complications and have similar microorganism profiles. These data suggest that preoperative biliary drainage should be avoided in candidates for pancreaticoduodenectomy.

Association of preoperative biliary drainage with postoperative outcome following pancreaticoduodenectomy.

OBJECTIVE: To determine whether preoperative biliary instrumentation and preoperative biliary drainage are associated with increased morbidity and mortality rates after pancreaticoduodenectomy. SUMMARY BACKGROUND DATA: Pancreaticoduodenectomy is accompanied by a considerable rate of postoperative complications and potential death. Controversy exists regarding the impact of preoperative biliary instrumentation and preoperative biliary drainage on morbidity and mortality rates after pancreaticoduodenectomy. METHODS: Two hundred forty consecutive cases of pancreaticoduodenectomy performed between January 1994 and January 1997 were analyzed. Multiple preoperative, intraoperative, and postoperative variables were examined. Pearson chi square analysis or Fisher's exact test, when appropriate, was used for univariate comparison of all variables. Logistic regression was used for multivariate analysis. RESULTS: One hundred seventy-five patients (73%) underwent preoperative biliary instrumentation (endoscopic, percutaneous, or surgical instrumentation). One hundred twenty-six patients (53%) underwent preoperative biliary drainage (endoscopic stents, percutaneous drains/stents, or surgical drainage). The overall postoperative morbidity rate after pancreaticoduodenectomy was 48% (114/240). Infectious complications occurred in 34% (81/240) of patients. Intraabdominal abscess occurred in 14% (33/240) of patients. The postoperative mortality rate was 5% (12/240). Preoperative biliary drainage was determined to be the only statistically significant variable associated with complications (p = 0.025), infectious complications (p = 0.014), intraabdominal abscess (p = 0.022), and postoperative death (p = 0.037). Preoperative biliary instrumentation alone was not associated with complications, infectious complications, intraabdominal abscess, or postoperative death. CONCLUSIONS: Preoperative biliary drainage, but not preoperative biliary instrumentation alone, is associated with increased morbidity and mortality rates in patients undergoing pancreaticoduodenectomy. This suggests that preoperative biliary drainage should be avoided whenever possible in patients with potentially resectable pancreatic and peripancreatic lesions. Such a change in current preoperative management may improve patient outcome after pancreaticoduodenectomy.

Effect of microscopic resection line disease on gastric cancer survival.

To study the effect of residual microscopic resection line disease in gastric cancer, we compared 47 patients with positive margins to 572 patients who underwent R0 resections using a multivariate analysis of factors affecting outcome. Although the presence of positive margins was a significant and independent predictor of outcome for the entire group (N = 619), this factor lost significance in patients who had undergone D2 or D3 lymph node dissections (N = 466). Subset analysis within the D2/D3 group determined that this finding was limited mainly to those patients with.5 positive nodes (N = 189). The survival of patients who had 5 positive nodes. We conclude that the significance of a positive microscopic margin in gastric cancer is dependent on the extent of disease. This factor is not predictive of outcome in patients who have undergone complete gross resection and have pathologically proved advanced nodal disease. Thus the goal in these cases should be an R0 resection when feasible but with the realization that the presence of >/=5 positive nodes (N2 disease according to the 1997 American Joint Committee on Cancer criteria) will mainly determine outcome and not microscopic residual cancer at the margin.

Extended lymphadenectomy is associated with a survival benefit for node-negative gastric cancer.

The purpose of this study was to determine whether extended lymph node (D2) dissection is associated with a survival benefit for patients with histologically node-negative gastric cancer at a single institution in the United States. Review of the prospective gastric database at Memorial Sloan-Kettering Cancer Center from July 1985 to August 1995 identified 774 patients who had undergone curative gastric resection. Of these, 247 patients (32%) were identified with histologically negative lymph nodes by hematoxylin-eosin staining. Data are expressed as median (range). Overall survival was compared by log-rank test. The overall 5-year survival rate for the entire cohort was 79%. The extent of lymph node dissection did not predict survival over the entire cohort. However, when stratified for tumor (T) stage, D2 dissection offered a survival advantage for T3 tumors. The 5-year survival rate for patients with T3 tumors undergoing a D2 dissection (n = 15) was 54% compared to 39% for those undergoing a Dl dissection (n = 53, P <0. 05). D2 dissection is associated with improved survival in advanced T stage, node-negative gastric cancer. With adequate staging, results of gastric resection for adenocarcinoma in Western countries begin to approximate those seen in Japan. Excision of N2 lymph nodes may also remove microscopic metastatic disease, contributing to the survival benefit.

Portal system thrombosis after splenectomy for neoplasm or chronic hematologic disorder: is routine surveillance imaging necessary?

PURPOSE: This study was undertaken to assess the frequency and sequelae of portal system thrombosis (PST) after splenectomy in patients with cancer or chronic hematologic disorders and to determine whether routine surveillance imaging for this potential complication is warranted. METHOD: The radiology reports of 203 consecutive patients with cancer or chronic hematologic disorders who underwent splenectomy between January 1990 and January 1997 were reviewed. Imaging examinations and medical records were reviewed for those in whom PST was found after splenectomy. RESULTS: One hundred twenty-three patients (60.6%) underwent CT (n = 88), sonography (n = 10), or both (n = 24) after splenectomy; one other patient underwent MRI. Twelve of these patients (9.8%) had thrombosis of the splenic, portal, and/or superior mesenteric veins. Their underlying diseases were myelofibrosis/ myelodysplastic syndrome (n = 8), lymphoma (n = 3), and leukemia (n = 1). At follow-up imaging (obtained in 10 of the 12 patients), PST had resolved (n = 5), worsened (n = 2), improved (n = 1), remained unchanged (n = 1), or resulted in cavernous transformation of the portal vein (n = 1). Nine of 12 patients were symptomatic. No patient died of PST. CONCLUSION: PST was an uncommon and typically unsuspected finding after splenectomy in this patient population, and no serious sequelae of PST were found. Routine surveillance imaging for PST after splenectomy does not seem warranted, but in symptomatic patients (particularly those with myelofibrosis/myelodysplastic syndrome), a high clinical suspicion and a low threshold for obtaining imaging examinations are needed.

Gastric carcinoma.

Gastric cancer continues to be a major global health problem. In the American College of Surgeons Patient Care study on gastric cancer, 5-year disease-specific survival was 26%, with an overall survival of 14%. Improvements in survival will require both earlier diagnosis and new therapeutic strategies. The ability of surgical oncologists to understand the natural history of the disease, accurately define its extent, and provide the most effective treatment places them in a key position to see that improvements in outcome are brought about expeditiously.

Peritoneal lavage cytology in gastric cancer: an independent predictor of outcome.

BACKGROUND: The prevalence and significance of free cancer cells in the peritoneal cavity of clinically M0 gastric cancer patients is unknown. We reviewed our results with peritoneal washings to determine (1) the prevalence of positive cytology in M0 and M1 disease and (2) the influence of positive cytology on the pattern of failure and survival. METHODS: Laparoscopic washings were obtained from 127 patients with gastric cancer at Memorial Sloan-Kettering Cancer Center from December 1, 1990 to August 1, 1996. Cytology was performed by the Papanicolau technique. RESULTS: The prevalence of positive cytology was as follows: 0% (0/45) in T1/T2 M0 disease; 10% (3/31) in T3/T4 M0 disease; and 59% in M1 disease. The three M0 patients with positive cytology recurred intra-abdominally (median follow up of 8.5 months). Survival was significantly less compared with stage-matched controls with negative cytology resected for cure (P <.03), and the same as those patients with stage IV disease. CONCLUSION: Patients with positive lavage cytology are stage IV, even in the absence of macroscopic peritoneal disease. Laparoscopic lavage cytology is a rapid technique for identifying the subset of M0 patients who are unlikely to benefit from resection alone. Such patients require additional treatment strategies to improve survival.

Total gastrectomy is not necessary for proximal gastric cancer.

BACKGROUND: Although there is an increasing incidence of proximal gastric cancers in the United States, the appropriate extent of resection for proximal gastric cancer is not known. This study addresses whether the type of operation (total gastrectomy [TG] vs proximal gastrectomy [PG]) affects outcome for proximal gastric adenocarcinoma. METHODS: Review of the prospective gastric database at Memorial Sloan-Kettering Cancer Center from July 1985 to August 1995 identified 391 patients with proximal gastric cancer. Of those patients, 98 underwent curative TG or PG through an exclusively abdominal approach. Patients undergoing esophagogastrectomy (n = 293) were excluded from analysis. Data are expressed as medians and ranges. RESULTS: The length of hospital stay was the same for patients undergoing resection for PG (16.5 days [range 8 to 55]) and for TG (18 days [range 8 to 48]). In addition, hospital mortality rates for PG (6.0%) were similar to those for TG (3.0%). There was no significant difference in tumor differentiation and overall stage between the groups that underwent TG and those that underwent PG. There was no significant difference in time to recurrence between the two operative groups (PG, 15.7 months, versus TG, 18 months). In addition, there was no association between first site of recurrence and type of procedure. The overall 5-year survival rate for proximal gastric cancer was 43% (median survival 46 months), whereas the 5-year survival rate for TG was 41% (median survival 51 months; difference not significant). CONCLUSIONS: The extent of resection for proximal gastric cancer does not affect long-term outcome. TG and PG have similar overall survival rates and time and rate of recurrence, and both procedures can be accomplished safely.

Combined modality therapy of gastric cancer.

Most patients with adenocarcinoma of the stomach continue to present with advanced stages of disease. Although cure is not possible without complete resection, most patients will go on to recur and die of their disease. Despite the efforts of the last 3 to 4 decades, little progress has been made in improving disease specific survival. The latest efforts to treat gastric cancer have now focused on preoperative treatment modalities. This article reviews the rationale and current status of preoperative multimodality therapy for gastric cancer.

The role of dietary factors in the intestinal and diffuse histologic subtypes of gastric adenocarcinoma: a case-control study in the U.S.

BACKGROUND: The decline of gastric adenocarcinoma in the U. S. and the parallel decrease in the predominance of the intestinal type of gastric adenocarcinoma may reflect, in part, changes in diet over the last 60 years. Because the intestinal and diffuse types of gastric adenocarcinoma may be epidemiologically distinct, the authors hypothesized that different nutritional factors are associated with the different subtypes of gastric adenocarcinoma. METHODS: Ninety-one incident cases with a pathologic diagnosis of gastric adenocarcinoma and 132 cancer free controls were included in this study. All cases were defined as being either the intestinal or diffuse type of gastric adenocarcinoma. Epidemiologic data were collected by a modified National Cancer Institute Health Habits History Questionnaire. Nutritional and dietary factors were analyzed using a logistic regression model. RESULTS: Several dietary factors were significantly associated with both subtypes of gastric adenocarcinoma, including dietary intakes of fiber, oleic acid, potassium, and fruits. Almost all dietary factors and food groups unique to intestinal gastric adenocarcinoma were protective in nature (vitamin B6, folate, niacin, iron, noncitrus fruits, and raw fruit), except for a high intake of dietary calories, which was a risk factor for intestinal gastric adenocarcinoma. The unique factors found to be protective for diffuse disease were carbohydrate and vitamin C intake. CONCLUSIONS: This study suggests that dietary factors contribute to the carcinogenesis of gastric adenocarcinoma. It also appears that although dietary risk and protective factors are common to both the intestinal and diffuse types of this malignancy, protective dietary factors may play a more important role in preventing the intestinal type of gastric adenocarcinoma.

Proximal gastric cancers resected via a transabdominal-only approach. Results and comparisons to distal adenocarcinoma of the stomach.

OBJECTIVE: The purpose of this study is to compare the outcome of patients with proximal gastric cancer (PGC) treated by a transabdominal-only resection to that of patients with distal gastric cancer (DGC). SUMMARY BACKGROUND DATA: It has been suggested that PGC is inherently more aggressive than DGC. The worse survival of PGC compared with that of DGC may be in part, because of the difficulty distinguishing PGC from distal esophageal adenocarcinoma. By defining a subset of PGC resected using an transabdominal-only approach, one may discriminate true PGC from distal esophageal adenocarcinoma. This subset of patients is a more appropriate comparison group when analyzing outcome relative to patients with DGC. METHODS: A review of the prospective database for gastric adenocarcinoma at Memorial Sloan-Kettering Cancer Center between July 1985 and August 1995 identified 98 patients with PGC resection via a transabdominal-only approach. Of these, 65 underwent proximal gastrectomy and 33 underwent total gastrectomy. For DGC, 258 required a distal gastrectomy and 71 required total gastrectomy. RESULTS: The overall 5-year survival of patients with PGC was 42% (median survival, 47 months), whereas the 5-year survival for patients with DGC was 61% (median survival, 106 months, p = 0.03). Within each stage, there were no significant survival differences, but in all stages, survival was better for patients with DGC. More important, the site of the primary tumor appears to affect survival, with a worse outcome as the tumor moves proximally. CONCLUSIONS: Despite excluding distal esophageal cancers, survival for patients with PGC remains worse than for those with DGC. Late stage of presentation could not explain this difference. It appears that PGCs are inherently more aggressive than are DGCs. In addition, site of the primary tumor appears to affect outcome, with a trend toward a worse outcome as the tumor moves proximally.