Mortality calculator as a possible prognostic predictor for overall survival after gastrectomy in elderly patients with gastric cancer

Abstract Background The number of elderly patients with gastric cancer (elderGC) has been increasing. Most of elderly patients were associated with reduced physiological functions, which sometimes constitute an obstacle to safe surgical treatments. The risk calculator of National Clinical Database (NRC), a Japanese surgical big database, provides mortality and morbidity as surgical-related risks. The purpose of this study is to investigate clinical significance of operative mortality calculated by NRC (NRC-mortality) during long-term follow-up after gastrectomy for elderGC.Methods We enrolled 73 patients aged 80 or over who underwent gastrectomy at our institution. Their surgical risk was evaluated based on the NRC-mortality. Several clinicopathological factors including NRC-mortality were selected and analyzed as possible prognostic factors for elderGC after gastrectomy. Statistical analysis was performing using the log-rank test and Cox proportional hazard model.Results NRC-mortality ranged 0.5 to 10.6%, and median value was 1.7%. Dividing elderGC into high- (1.7% or more, n=38) and low- (less than 1.7%, n=35) mortality groups, high-mortality group showed a significantly poor prognosis in overall survival (OS) than the low-mortality group, whereas there was no difference between the two groups in disease specific survival (DSS). In the analysis of Cox proportional hazard model, multivariate analysis revealed that NRC-mortality was an independent prognostic factor as well as neutrophil-lymphocyte ratio and surgical procedure in OS. In contrast, PS and pStage were independent prognostic factors in DSS, but not NRC-mortality.Conclusions The NRC-mortality might be clinical useful for not only predicting surgical mortality but also OS after gastrectomy in elderGC.