Background: The dismal prognosis of gastric signet ring cell carcinoma (GSRC) is a global problem. The current study is conducted to comprehensively evaluate clinicopathological features and survival outcomes in GSRC patients stratified by anatomic subsites. Then predictive nomograms are constructed and validated to improve the effectiveness of personalized management.
Method: The patients diagnosed with GSRC were recruited from the online SEER database. The influence of anatomic subsites on overall survival (OS) and cancer-specific survival (CSS) was evaluated using multivariate Cox regression and Kaplan-Meier analysis. Then we employed propensity score matching (PSM) technique to decrease selection bias and balance patients’ epidemiological factors. Predictive nomograms were constructed and validated.
Results: Multivariate Cox regression demonstrated that the patients with overlapping gastric cancer (OGC) suffered the highest mortality risk for OS (HR, 1.29; 95%CI, 1.23-1.36; P<0.001) and CSS (HR, 1.33; 95%CI, 1.28-1.37; P<0.001). Age, TNM stage, tumor localization, tumor size, surgery and chemotherapy presented a highly significant relationship with OS and CSS. Following subgroup and PSM analysis, OGC patients were confirmed to have the worst OS and CSS. Then nomograms predicting 6 months, 12 months and 36 months OS and CSS were constructed. The calibration curves and reveiver operating characteristic curves demonstrated the great performance of the nomograms.
Conclusion: We identified anatomic subsites as a predictor of survival in those with GSRC. Patients with OGC suffered the highest mortality risk. The proposed nomograms allowed a relatively accurate survival prediction for GSRC patients.