A LODDS-based nomogram for overall and cancer-specific survival in stage III-IV gastric signet ring cell carcinoma.

OBJECTIVE

Gastric cancer is a serious human chronic disease. The gastric signet ring cell carcinoma (GSRCC) is the most-dangerous subtype with several acute complications, including gastrointestinal hemorrhage, gastric perforation, pyloric obstruction and so on. This study aimed to compare the predictive efficiency of positive lymph nodes (PLN) and log odds of positive lymph nodes (LODDS) for survival and to establish a LODDS-based nomogram model in stage III-IV GSRCC.

METHODS

Stage III-IV GSRCC patients were acquired between 2015 and 2019 from SEER dataset and the affiliated Yixing hospital of jiangsu university, serving as training and validation datasets respectively. The X-tile software was used to identify cut-off values while their relationship with clinical features was explored by chi-square test. The Kaplan-Meier analysis was applied for survival curve while cox regression analysis was performed for independent risk factors. The nomogram model was built with ROC and calibration curves for verification.

RESULTS

A total of 585 stage III-IV GSRCC patients were included in this study with 536 patients for training and 49 patients for validation. The LODDS showed better predictive efficiency for overall survival (OS) and cancer-specific survival (CSS) than PLN. The LODDS, M stage and chemotherapy status were independent factors for both OS and CSS, with LODDS contribution accounting for 31.47% in OS and 30.39% in CSS. A LODDS-based nomogram was built with accurate efficiency in stage III-IV GSRCC. The 1-year, 2-year, 3-year OS area under curve (AUC) values were 0.755, 0.795, 0.759 for internal and 0.776, 0.756, 0.816 for external verification while 1-year, 2-year, 3-year CSS AUC values were 0.745, 0.803, 0.770 for internal and 0.796, 0.762, 0.820 for external verification.

CONCLUSION

LODDS is an independent risk factor in stage III-IV GSRCC. The LODDS-based nomogram model showed excellent predictive efficiency, providing a novel insight for early diagnosis and precise therapies of stage III-IV GSRCC.