Clinical Significance of Total Gastrectomy for Proximal Gastric Cancer

Transcription

1 Clinical Significance of Total Gastrectomy for Proximal Gastric Cancer AKIRA OOKI, KEISHI YAMASHITA, SHIRO KIKUCHI, SHINICHI SAKURAMOTO, NATSUYA KATADA, NOBUE HUTAWATARI and MASAHIKO WATANABE Department of Surgery, Kitasato University Hospital, Kitasato , Sagamihara , Kanagawa, Japan Abstract. Background: The optimal surgical strategy for proximal gastric cancer (PGC) remains controversial. The principal difference between total gastrectomy (TG) and proximal gastrectomy (PG) is whether lymph node (LN) 4d is dissected by radical lymphadenectomy. Patients and Methods: The subjects were 243 patients with PGC who had undergone TG between 1990 and The incidences of metastases to individual LNs were assessed. The 5-year disease-specific survival (DSS) was examined in 179 patients who had undergone TG and an additional 37 patients who had undergone PG between 1990 and Results: When PGC was confined to the muscularis propria (mp), no patient had LN 4d or 8 metastases although these nodes are considered regional LNs, and the 5-year DSS rate did not significantly differ between TG and PG. When the PGC extended beyond the mp, the patients with LN 4d or 8 metastases in N2 disease had very poor outcomes, similar to those with N3 disease. Multivariate analysis revealed that lymph node metastatic density over 40% (ND40) was the strongest independent prognostic factor (p=0.001, HR=6.1). ND40 was significantly associated with LN 4d or 8 metastases in N2 disease (p=0.047). Conclusion: LN 4d or 8 metastasis is equivalent to advanced LN metastasis and the presence of such metastasis indicates a poor prognosis, regardless of LN dissection, thus, PG might therefore be an adequate procedure for LN dissection in patients with PGC. Gastric cancer is the fourth most common cancer and the second leading cause of cancer-related death worldwide (1). The predominant site of occurrence has shifted from the distal to the proximal stomach over the past several decades (1, 2). Correspondence to: Masahiko Watanabe, MD, Ph.D., FACS, Professor of Department of Surgery, Medical School, Kitasato University, Kitasato , Sagamihara , Kanagawa, Japan Tel: , Fax: , gekaw@med.kitasato-u.ac.jp Key Words: Proximal gastric cancer, lymph node metastatic density, prognosis. Surgical resection remains the gold standard for curative treatment. Preservation of function by reducing the extent of gastrectomy is also an important goal of therapy (3). In distal gastric cancer, consensus has been reached that curative resection can be achieved by subtotal gastrectomy, considered equivalent to total gastrectomy (TG), even in advanced disease (4, 5). In contrast, the procedure of choice for proximal gastric cancer (PGC) remains controversial. As for symptoms and nutritional status after gastrectomy, the available evidence suggests that proximal gastrectomy (PG) with improved reconstruction is superior to TG (3, 6-8), but which procedure has better outcomes is unclear (9-15). In general, PG is considered effective for early PGC (16), whereas TG is recommended in order to dissect the regional lymph nodes (LNs) for advanced PGC (15, 17). If an adequate distal surgical margin can be obtained, the difference between TG and PG is whether the LNs above and below the pylorus (LN 5 and 6) and the LN along the right gastroepiploic vessels (LN 4d) are dissected by radical lymphadenectomy (12, 14). LN 5 and 6 are classified as N3, extra-regional LNs, according to the 13th edition of the Japanese Classification of Gastric Carcinoma (JCGC) (18). The elucidation of the optimal strategy for PGC therefore requires a comprehensive assessment of the incidence of LN 4d metastasis, defined as N2, and the relationships of such metastasis to outcomes and biological characteristics. The patterns of LN metastases in patients with primary PGC were therefore examined to clarify the optimal surgical strategy for PGC. Patients and Methods Registration of patients. Between February 1990 and August 2006, 316 patients with primary PGC had undergone TG with modified D2 or more extensive dissection at Kitasato University Hospital. The incidence of metastases to individual LNs was retrospectively studied in 243 patients who had undergone TG and fulfilled the eligibility requirements described below. Additionally the 5-year disease-specific survival (DSS) of the 179 patients who had undergone TG and an additional 37 patients who had undergone PG at Kitasato University Hospital between February 1990 and March 2002, was assessed. All the 37 patients had PGC confined to the muscularis propria (mp) /2008 $

2 Table Ⅰ. Incidence of metastases to individual lymph nodes in 243 patients who had undergone TG. Incidence of individual lymph node metastasis (% ) Station number m (n=40) sm (n=45) mp (n=27) ss (n=82) se (n=41) si (n=8) Lymph node metastasis 0/40(0% ) 7/45(15.6% ) 8/27(29.6% ) 55/82(67.1% ) 31/41(75.6% ) 6/8(75.0% ) The first tier 1 0/40(0% ) 2/45(4.4% ) 1/27(3.7% ) 25/82(30.5% ) 19/41(46.3% ) 1/8(12.5% ) 2 0/40(0% ) 1/45(2.2% ) 1/27(3.7% ) 10/82(12.2% ) 10/41(24.4% ) 1/8(12.5% ) 3 0/40(0% ) 7/45(15.6% ) 7/27(25.9% ) 42/82(51.2% ) 24/41(58.5% ) 3/8(37.5% ) 4s 0/40(0% ) 2/45(4.4% ) 0/27(0% ) 8/82(9.8% ) 9/41(22.0% ) 1/8(12.5% ) The second tier 4d# 0/40(0%) 0/45(0%) 0/27(0%) 3/82(3.7% ) 8/41(19.5% ) 3/8(37.5% ) 7 0/40(0% ) 2/45(4.4% ) 2/27(7.4% ) 17/82(20.7% ) 10/41(24.4% ) 3/8(37.5% ) 8 0/40(0%) 0/41(0%) 0/25(0%) 3/82(3.7% ) 5/41(12.2% ) 1/8(12.5% ) 9 0/40(0% ) 1/41(2.4% ) 1/25(4.0% ) 5/82(6.1% ) 1/41(2.4% ) 2/8(25.0% ) 10 0/15(0% ) 1/27(3.7% ) 1/23(4.3% ) 7/82(8.5% ) 6/41(14.6% ) 3/8(37.5% ) 11 0/15(0% ) 0/27(0% ) 1/23(4.3% ) 8/82(9.8% ) 8/41(19.5% ) 2/8(25.0% ) The third tier 5# 0/40(0%) 0/45(0%) 0/27(0%) 2/82(2.4% ) 1/41(2.4% ) 0/8(0% ) 6# 0/40(0%) 0/45(0%) 0/27(0%) 0/82(0% ) 1/41(2.4% ) 1/8(12.5% ) 16 0/2(0% ) 0/1(0% ) 0/1(0% ) 1/18(5.6% ) 2/8(25.0% ) 1/4(25.0% ) Abbreviations: m, mucosa; sm, submucosa; mp, muscularis propria; ss, subserosa; se, serosa-exposed; si, serosa-infiltrating. # These lymph nodes may be left undissected by proximal gastrectomy. The stomach was separated into upper, middle, and lower third by dividing the lesser and major curvature into thirds and drawing lines between corresponding points in the two curvatures (18). PGC was defined as primary adenocarcinoma limited to the upper third of the stomach. Patients who had any of the following were excluded: lesions extending to the middle third of the stomach or to the esophagus, multiple carcinomas in the stomach, cancer in the residual stomach or distant metastasis. All the histopathological examinations were performed by experienced histopathologists according to the JCGC. The histopathological factors analyzed in this study were abstracted from the patients medical records. The nodal status was defined according to the JCGC, in which the LN are divided into the following different areas, designated as stations (Figure 1): LN 1, LN along the right cardia; 2, LN along the left cardia; 3, LN along the lesser curvature; 4s and 4d, LN along the left and right gastroepiploic vessels; 5, LN above the pylorus; 6, LN below the pylorus; 7, LN along the left gastric artery; 8, LN along the common hepatic artery; 9, LN along the celiac artery; 10, LN at the splenic hilum; 11, LN along the splenic artery; 12, LN along the hepatoduodenal ligament; 16, LN in the paraaortic region; 19, LN below the diaphragm and 20, LN near the esophageal hiatus. In PGC, these stations are further classified into three compartments (the first tier, LN 1, 2, 3, and 4s; the second tier, LN 4d, 7, 8, 9, 10, and 11 and the third tier, LN 5, 6, 12, 16, 19, and 20), which correspond to the N-factor classifications of N1, N2, and N3, respectively. Metastasis to any of the remaining stations is defined as distant metastasis (M). N3 and M disease are classified into stage IV according to the JCGC staging system, implying extraregional LNs metastases. The N-factor according to the International Union Against Cancer (UICC) staging system is defined on the basis of the number of metastatic LNs (19). Modified D2 dissection involves removal of all of the first tier LNs plus 7 (D1+) or 7, 8, and 9 (D1+) (20). D2 dissection entails removal of all first and second tier LNs. All the patients with PGC extending beyond the mp had undergone TG with D2 or more extensive dissection. All the LNs, at least 15, were retrieved individually from the resected specimens, and were examined histologically to accurately evaluate node status (21). The incidence of metastases to individual LNs was defined as the percentage of patients with metastasis to the number of patients who had undergone individual LNs dissection. Lymph node metastatic density (NDX) was defined as the percentage (X) of metastatic LNs to the number of dissected LNs. Statistical analysis. The Fisher s exact test, Chi-square test, or Mann-Whitney U-test were used to statistically analyze the categorical variables, and the unpaired Student s t-test was used for the continuous variables. The Kaplan-Meier method was used to estimate the cumulative survival rates and the differences in survival rates were assessed with the use of the log-rank test (22). The 5-year DSS was measured from the date of surgery to the date of death or the last follow-up. The data for patients who died from causes other than gastric cancer (n=7) or who survived for over 60 months were censored at the time of death or at 60 months, respectively. The variables suggested to be prognostic factors on univariate analysis (p<0.05) were subjected to multivariate analysis using the Cox proportional-hazards regression model (23). p<0.05 was considered to indicate statistical significance. All the statistical analyses were conducted with the SAS software package StatView, version 5.0 (SAS Institute, Cary, NC, USA). 2876

3 Ooki et al: Total Gastrectomy for Proximal Gastric Cancer Figure 1. Anatomical distribution of lymph nodes defined according to the JCGC in PGC. Figures in the circles indicate station numbers of the LNs. The most common channel for lymphatic drainage in PGC is thought to be the left gastric artery channel (LN 1 or 3 to LN 7), while the right gastric artery channel (LN 5 or 8 to LN 12) and the right gastroepiploic artery channel (LN 4d to LN 6) are relatively rare. Results Patient characteristics and distribution of individual LNs metastases. In the 243 patients with PGC who had undergone TG, the mean numbers of dissected LNs and metastatic LNs were 45 (range: ) and 2.8 (range: 0-61), respectively. The disease stage according to the JCGC staging system was IA in 78 patients, IB in 48, II in 48, IIIA in 37, IIIB in 15 and IV in 17. One patient had a positive surgical margin. Reconstruction was carried out by Roux-en- Y anastomosis. The incidence of individual LN metastases tended to increase with deeper invasion (Table Ⅰ). When the depth of invasion was confined to the mp, there was no metastasis to LN 4d, 5, 6, or 8. 5-Year DSS rates following PG or TG in patients with PGC confined to the mp. For the 179 patients who had undergone TG and an additional 37 patients who had undergone PG between February 1990 and March 2002, the 5-year DSS rate did not differ significantly between the patients with mucosal (m) and mp invasion (100% and 92%, respectively), but did differ significantly between those with mp and subserosal (ss) invasion (92% and 56%, respectively, p=0.002, Figure 2A). Out of the 118 patients who had PGC confined to the mp, 37 had undergone PG and 81 TG. The clinicopathological findings of these patients are shown according to PG and TG in Table Ⅱ. Splenectomy and the extent of dissection significantly differed between PG and TG, but the 5-year DSS rate did not (Figure 2B). 5-Year DSS rates of patients according to metastases to individual LNs and of patients with LN 4d or 8 metastasis in N2 disease. None of the patients in this series underwent PG for PGC extending beyond the mp. It was therefore important to evaluate the outcomes of patients according to metastases to individual LNs, especially LN 4d, to clarify the indications for PG in patients with PGC extending beyond the mp. Out of the 216 patients with PGC between February 1990 and March 2002, 84 had LN metastases. Table Ⅲ depicts the 5-year DSS rates according to metastasis to individual LNs grouped according to N-factor. For the patients with N1 disease, the 5-year DSS rates were similar (about 60% ) regardless of which nodes were involved. In contrast, the 5-year DSS rates were 0% for the patients with LN 4d or 8 metastasis in N2 disease and in those with LN 5 or 6 metastasis in N3 disease. Out of the 84 patients with LN metastasis, PGC extending beyond the mp was N1 disease in 34 patients, N2 disease in 2877

4 Figure 2. 5-Year DSS rates according to the depth of invasion (A) and according to PG and TG (B). 30 and N3 disease in 9. Interestingly, in the N2 disease, the 5-year DSS rate of the patients who were positive for LN 4d or 8 metastasis (n=11) was significantly poorer than that of the patients who were negative for LN 4d and 8 metastases (n=19, p=0.001, Figure 3). Moreover, the 5-year DSS of patients who were positive for LN 4d or 8 metastasis (n=11) was similar to that of the patients with N3 disease (n=9) and the 5-year DSS of the patients who were negative for LN 4d 2878

5 Ooki et al: Total Gastrectomy for Proximal Gastric Cancer Table Ⅱ. Comparison of clinicopathological findings for PG and TG in 118 patients with PGC confined to the muscularis propria. Variable Proximal Total P-value gastrectomy gastrectomy (n=37) (n=81) DSS (months) 60±0.0 59±3.1 * Gender Male Female Age (years) 57.4± ±11.0 Size of tumor (cm) 3.8± ±2.3 Differentiation Well/Moderate Poor Histological type Intestinal type Diffuse type 7 9 Depth of invasion m sm mp 5 19 Lymph node metastasis Absence Presence 2 8 Stage Stage I Stage II 1 5 Stage III 0 1 Surgical margin Negative Positive 0 0 Splenectomy <0.001 Absence Presence 7 51 Extent of dissection <0.001 Modified D D2 or more extensive dissection 0 51 Number of removed lymph nodes 28.6± ± DSS, disease-specific survival;, not significant. *Log-rank test. and 8 metastases (n=19) was similar to that of the patients with N1 disease (n=34). Univariate and multivariate prognostic analyses of patients with stage II to IV disease. Out of the 216 patients, 88 had stage IA disease and 38 had stage IB disease. Both of these groups had excellent 5-year DSS rates (100% and 95%, respectively). Such patients were excluded from our analysis of prognostic factors. The remaining 90 patients with stage II to IV disease were included in the analysis. The JCGC TNM stage (p<0.001), ND40 (p<0.001), N-factor according to the UICC staging system (p<0.001) and age (p=0.04) were significantly associated with poor outcomes on univariate analysis (Table Ⅳ and Figure 4). Out of these factors, only ND40 had a 5-year DSS rate of 0% and was therefore considered a non-curative factor (Figure 4B). The variables suggested to be significant prognostic factors on univariate analysis were subjected to multivariate analysis. The results revealed that ND40 (p=0.001, HR=6.10), JCGC TNM stage (p=0.03), and age (p=0.04, HR=2.00) were independent prognostic factors in PGC (Table Ⅴ). Table Ⅵ shows the association between the independent prognostic factors and LN 4d or 8 metastasis in the patients with N2 disease. The ND40 was found to be significantly associated with LN 4d or 8 metastasis in N2 disease (p=0.047). Discussion In the present study, no PGC patient with tumors confined to the mp had LN 4d, 5, 6, or 8 metastasis, consistent with the results of previous studies (12, 24). Moreover, no significant difference was found in survival between PG and TG when the PGC was confined to the mp (Figure 2B). These findings supported the recommendation that PG could be approved as adequate gastrectomy, with radical lymphadenectomy, for patients with PGC confined to the mp (24). When the PGC extended beyond the mp, the patients with LN 4d or 8 metastasis in N2 disease had very poor outcomes, similar to those with N3 disease (Figure 3). Accumulating evidence suggests that the ND-factor is one of the most important prognostic factors in gastric cancer (25-31). The present study demonstrated that ND40 was an independent prognostic and non-curative factor in PGC and was significantly associated with LN 4d or 8 metastasis in N2 disease (Table Ⅵ). Interestingly, the association between ND40 and LN 4d or 8 metastasis was observed even in the patients with disease more advanced than N2 (p=0.002) and 8 (89% ) out of 9 patients with ND40 had LN 4d or 8 metastases. Taken together, these findings suggested that LN 4d or 8 metastasis represents advanced LN metastasis status and is directly associated with the outcome in PGC and the presence of such metastasis indicates a poor prognosis, regardless of LN dissection. We therefore believe that TG is not indicated in order to dissect LN 4d as radical lymphadenectomy, even in those patients who have PGC extending beyond the mp. These findings could be interpreted in two ways. One possibility is that LN 4d or 8 metastasis is a parameter for the presence of extraregional LNs metastases. Yoshikawa et al. have reported that LN 16 metastasis was observed in 20% of patients with presumed N2 disease when D2 dissection with paraaortic lymphadenectomy was performed instead of D2 dissection for the patients with PGC extending beyond the mp, and suggested that extended surgery provided more accurate information on LNs metastases (32). In the present study, 7 (64% ) out of the 11 patients with LN 4d or 8 metastasis who underwent lymphadenectomy more extensive than D2 were shifted from N2 disease presumed by D2 dissection to N3 or M disease. If lymphadenectomy more aggressive than D2 2879

6 Figure 3. 5-year DSS rates according to the positivity (n=11) or negativity (n=19) of LN 4d or 8 metastasis in patients with N2 disease who had PGC extending beyond the mp. Table Ⅲ. 5-Year disease-specific survival rates in 84 patients with lymph node metastasis. Individual lymph nodes grouped according to N-factor. 5-Year DSS N1 (n=40) N2 (n=34) N3 or M (n=10) s 4d Others 60% 57% 60% 60% 0% 35% 0% 67% 22% 17% 0% 0% 13% (n=15) (n=7) (n=30) (n=5) (n=8) (n=23) (n=4) (n=3) (n=9) (n=12) (n=3) (n=2) (n=8) 5-Year DSS, 5-year disease-specific survival. There were 11 patients with LN 4d or 8 metastasis in N2 disease. dissection had been performed, the patients with LN 4d or 8 metastasis might have been found to have extraregional LNs metastases. The other possibility is that LN 4d and 8 themselves are extraregional LNs. The most common channel for lymphatic drainage in PGC is thought to be the left gastric artery channel (LN 1 or 3 to LN 7), while that by the right gastric artery (RGA) channel (LN 5 or 8 to LN 12) and the right gastroepiploic artery (RGEA) channel (LN 4d to LN 6) are relatively rare (33). The present study was also consistent with this finding (Table I). Out of the LNs flowing by the RGA or RGEA channel, only LN 4d and 8 are not extraregional LNs according to the JCGC (Figure 1), but the outcomes of the patients with such metastasis were similar to those of the patients with the remaining LNs metastasis. The RGA and RGEA channels might therefore be extraregional channels for lymphatic drainage in PGC. Although 64 patients with PGC extending to the middle third of the stomach were excluded from the present study, these patients were independently investigated to elucidate whether these patients differ from those with PGC limited to the upper third of the stomach from the frequent and prognostic points of view. Metastases to LN 4d, 5, or 6 were found even when the depth of invasion was sm. The total incidences of LN 4d, 5, and 6 metastases were 54.7% (n=35), 14.1% (n=9), and 18.8% ( n=12), respectively. Interestingly, the 5-year DSS rate was 100% for the 6 patients with LN 4d, 5, or 6 metastases whose tumors were confined to the mp, but markedly decreased to 13.3% (n=31), 37.5% (n=8), and 10.0% (n=11), respectively, in the patients whose tumors extended beyond the mp. These findings suggested that LN 4d, 5, or 6 metastases did not necessarily indicate 2880

7 Ooki et al: Total Gastrectomy for Proximal Gastric Cancer Figure 4. Prognostic analysis of patients with PGC who had undergone gastrectomy. Kaplan-Meier curves for the 90 patients with PGC according to (A) the JCGC staging system (p<0.001), (B) lymph node metastatic density, ND40 (p<0.001), (C) N-factor according to the UICC (p<0.001) and (D) age (p=0.04). incurable disease in patients with PGC extending to the middle third of the stomach, and may therefore be an indication for PG in selected cases. In PGC extending to the middle third of the stomach, the RGEA channel and the RGA channel may serve as the main lymphatic drainage channels. This is supported by the good outcomes despite the high frequencies of LN metastases in the patients whose tumors were confined to the mp, but not in those whose tumors extended beyond the mp. We believe that this is the first study to examine the incidences of metastases to individual LNs in patients with PGC and the relationship of such metastases to the outcomes and biological characteristics. In conclusion, LN 4d or 8 metastasis is equivalent to advanced LN metastasis and the presence of such metastasis indicates a poor prognosis, regardless of LN dissection. We therefore believe that TG is not indicated in order to dissect LN 4d as radical lymphadenectomy. Our findings might provide a basis for justifying the performance of PG in patients with PGC, regardless of whether cancer is early stage or not, thereby improving postoperative function and patients quality of life without negatively affecting survival. Additional studies will be required to confirm these findings. References 1 Crew KD and Neugut AI: Epidemiology of gastric cancer. World J Gastroenterol 12: , Shang J and Pena AS: Multidisciplinary approach to understand the pathogenesis of gastric cancer. World J Gastroenterol 11: ,

8 Table Ⅳ. Distribution of clinicopathological factors and univariate analysis of prognostic factors in 90 patietns with stage II to IV disease. DSS Variable No. of Mean survival P-value* patients (months) Gender Male Female Age (years) 0.04 < > or = Size of tumor (cm) < > or = Differentiation Well/Moderate Poor Lymphatic permeation Absence 5 41 Presence Vascular permeation Absence 6 40 Presence N-factor according to the UICC <0.001 N N N N3 and M ND (% ) <0.001 < > or = Stage according to the JCGC <0.001 II IIIA IIIB IV DSS, disease-specific survival;, not significant; ND, lymph node metastatic density. *Log-rank test. Table Ⅴ. Multivariate analyisis of factors associated with diseasespecific survival in patients with stage II to IV disease. Variables Coefficient P-value HR 95% CI ND Stage 0.03 Stage II reference Stage IIIA Stage IIIB Stage IV Age N-factor according to UICC N0 reference N N N3-M CI, confidence interval; HR, Hazard ratio;, not significant. ND, lymph node metastatic density. Table Ⅵ. Association between independent factors and LN 4d or 8 metastasis in 30 patients with N2 disease and PGC extending beyond the muscularis propria. Variable LN 4d or 8 LN 4d and 8 P-value positivity (n=11) negativity (n=19) ND Absence 7 18 Presence 4 1 Age (years) < > or = Stage IIIA 3 11 IIIB 5 3 IV 3 5, not significant. 3 Katai H: Function-preserving surgery for gastric cancer. Int J Clin Oncol 11: , Gouzi JL, Huguier M, Fagniez PL et al: Total versus subtotal gastrectomy for adenocarcinoma of the gastric antrum. A French prospective controlled study. Ann Surg 209: , Bozzetti F, Marubini E, Bonfanti G et al: Subtotal versus total gastrectomy for gastric cancer: five-year survival rates in a multicenter randomized Italian trial. Italian Gastrointestinal Tumor Study Group. Ann Surg 230: , Hinoshita E, Takahashi I, Onohara T et al: The nutritional advantages of proximal gastrectomy for early gastric cancer. Hepatogastroenterology 48: , Shiraishi N, Adachi Y, Kitano S et al: Clinical outcome of proximal versus total gastrectomy for proximal gastric cancer. World J Surg 26: , Yoo CH, Sohn BH, Han WK et al: Proximal gastrectomy reconstructed by jejunal pouch interposition for upper third gastric cancer: prospective randomized study. World J Surg 29: , Kaibara N, Nishimura O, Nishidoi H et al: Proximal gastrectomy as the surgical procedure of choice for upper gastric carcinoma. J Surg Oncol 36: , Moreaux J and Msika S: Carcinoma of the gastric cardia: surgical management and long-term survival. World J Surg 12: , Jakl RJ, Miholic J, Koller R et al: Prognostic factors in adenocarcinoma of the cardia. Am J Surg 169: , Kitamura K, Nishida S, Yamamoto K et al: Lymph node metastasis in gastric cancer in the upper third of the stomachsurgical treatment on the basis of the anatomical distribution of positive node. Hepatogastroenterology 45: ,

9 Ooki et al: Total Gastrectomy for Proximal Gastric Cancer 13 Harrison LE, Karpeh MS and Brennan MF: Total gastrectomy is not necessary for proximal gastric cancer. Surgery 123: , Kobayashi T, Sugimura H and Kimura T: Total gastrectomy is not always necessary for advanced gastric cancer of the cardia. Dig Surg 19: 15-21, Kim JH, Park SS, Kim J et al: Surgical outcomes for gastric cancer in the upper third of the stomach. World J Surg 30: , Katai H, Sano T, Fukagawa T, Shinohara H et al: Prospective study of proximal gastrectomy for early gastric cancer in the upper third of the stomach. Br J Surg 90: , Archie V, Kauh J, Jones DV Jr et al: Gastric cancer: standards for the 21st century. Crit Rev Oncol Hematol 57: , Japanese Gastric Cancer Association: Japanese Classification of Gastric Carcinoma, 2nd English edition. Gastric Cancer 1: 10-24, Sobin LH and Wittekind CH: International Union Against Cancer (UICC): TNM Classification of Malignant Tumors, 5th edition. New York: John Wiley and Sons, Nakajima T: Gastric cancer treatment guidelines in Japan. Gastric Cancer 5: 1-5, Karpeh MS, Leon L, Klimstra D et al: Lymph node staging in gastric cancer: is location more important than number? An analysis of 1,038 patients. Ann Surg 232: , Peto R, Pike MC, Armitage P et al: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. analysis and examples. Br J Cancer 35: 1-39, Harrell FE Jr, Lee KL and Pollock BG: Regression models in clinical studies: determining relationships between predictors and response. J Natl Cancer Inst 80: , Isozaki H, Okajima K, Yamada S et al: Proximal subtotal gastrectomy for the treatment of carcinoma of the upper third of the stomach: its indications based on lymph node metastasis and perigastric lymphatic flow. Surg Today 25: 21-26, Lee JH, Noh SH, Lah KH et al: The prognosis of stage IV gastric carcinoma patients after curative resection. Hepatogastroenterology 48: , Bando E, Yonemura Y, Taniguchi K et al: Outcome of ratio of lymph node metastasis in gastric carcinoma. Ann Surg Oncol 9: , Nitti D, Marchet A, Olivieri M et al: Ratio between metastatic and examined lymph nodes is an independent prognostic factor after D2 resection for gastric cancer: analysis of a large European monoinstitutional experience. Ann Surg Oncol 10: , Rodriguez Santiago JM, Munoz E, Marti M et al: Metastatic lymph node ratio as a prognostic factor in gastric cancer. Eur J Surg Oncol 31: 59-66, Kunisaki C, Shimada H, Nomura M et al: Clinical impact of metastatic lymph node ratio in advanced gastric cancer. Anticancer Res 25: , Cheong JH, Hyung WJ, Shen JG et al: The N ratio predicts recurrence and poor prognosis in patients with node-positive early gastric cancer. Ann Surg Oncol 13: , Marchet A, Mocellin S, Ambrosi A et al: The ratio between metastatic and examined lymph nodes (N ratio) is an independent prognostic factor in gastric cancer regardless of the type of lymphadenectomy: results from an Italian multicentric study in 1853 patients. Ann Surg 245: , Yoshikawa T, Sasako M, Sano T et al: Stage migration caused by D2 dissection with para-aortic lymphadenectomy for gastric cancer from the results of a prospective randomized controlled trial. Br J Surg 93: , Kosaka T, Ueshige N, Sugaya J et al: Lymphatic routes of the stomach demonstrated by gastric carcinomas with solitary lymph node metastasis. Surg Today 29: , Received March 31, 2008 Revised June 30, 2008 Accepted July 28,