A population-based study of tumor-node relationship,resection margins,and surgeon volume on gastric cancer survival

BACKGROUND: Gastric cancer surgery literature is conflicting. Two European level I randomized controlled trials refute Asian lesser level evidence promoting more radical resections. Population-based study evidence is undefined. METHODS: Using this study design we examined the overall survival, the tumor-node relationship, margins, and surgeon volume on gastric cancer survival in a Canadian province. RESULTS: Between 1991 and 1997, 577 (71 +/- 13 years 60% male) gastric adenocarcinomas were diagnosed in Northern Alberta (population 1.7 million). Respectively, median survival in months for stage I (n = 67) was 77, stage II (n = 55) 75, stage III (n = 155) 12, stage IV (n = 235) 3, and 65 unstaged (n = 65) 4. Five-year survival for T1N0 (n = 28) was 68% versus T1N1 (n = 7) 71% (P = 0.80); for T2N0 (n = 29) 58% versus T2N1 (n = 19) 58% versus T2N3 (n = 7) 29% (P = 0.08); for T3N0 (n = 33) 57%, versus T3N1 (n = 98) 9% versus T3N2 (n = 47) 0% versus T3N3 (n = 8) 0% (P < 0.0001). Median gastrectomy survival (months) in stage III was 15 months margin negative versus 8 months margin positive versus 6 bypass and 5 for no surgery (P = 0.0004). In stage IV it was margin positive 8 versus margin negative 6 (nonsignificant), bypass 3 versus no surgery 2. Five-year survival for surgeons doing fewer than 20 gastrectomies (n = 196 patients) was 29% (median 1.4 years) versus 35% (median 2.3 years; n = 72 patients) for surgeons doing 20 or more (n = 4; P = 0.325). CONCLUSIONS: From these population data we conclude that (1) few patients present with "curable" gastric cancer, (2) node negative or small gastric cancer survival is not influenced by nodal stage, (3) positive margin resection survival is better than bypass or no surgery in stage IV but not stage III disease, and (4) surgeon volume does not appear to influence patient survival.