A nomogram for predicting lymph node metastasis in early gastric signet ring cell carcinoma

Abstract

Abstract Background: At present, the risk factors of lymph node metastasis in early gastric signet ring cell carcinoma (SRCC) are not clear. The lymph node metastasis (LNM) rate and prognosis of early gastric SRCC are better than those of other undifferentiated cancers. With the development of endoscopic technology, the five-year survival rate of endoscopic treatment of early gastric cancer is similar to that of traditional surgery, and its quality of life is better than that of traditional surgery. Therefore, the aim of this study is to develop a nomogram that can predict the SRCC, hoping to help clinicians choose the best treatment strategy. Methods: The data of 183 patients with early gastric SRCC who underwent radical gastrectomy with lymph node dissection in our hospital from January 2014 to June 2022 were retrospectively collected to establish a research cohort. The least absolute selection and shrinkage operator (Lasso) and multivariate regression analysis were used to identify the predictors of early gastric SRCC lymph node metastasis in the study cohort, and Nomogram was established. The receiver operating characteristic (ROC) curve, calibration curve and decision curve were used to evaluate the discrimination, accuracy and clinical practicability of the nomogram. Results: The overall incidence of lymph node metastasis was 21.9% (40/183). Multivariate logistic regression analysis showed that tumor size and lymphovascular invasion (LVI) were independent risk factors for lymph node metastasis. Lasso regression analysis showed that tumor size, depth of invasion, LV, E-cad, dMMR, CA242, NLR and macroscopic type were related to LNM. The basic model 1, which included tumor size and LVI, had an area under curve(AUC) of 0.741 for predicting LNM. The addition of depth of invasion to model 1 resulted in significant improvements in AUC (P=0.023) and net reclassification index (NRI) (P < 0.001).The inclusion of dMMR and CA242 also improved NRI (P < 0.001). When type_1 was included, the AUC (P=0.017), Integrated discrimination Improvement (IDI) (P=0.003) and NRI (P=0.032) of the model were significantly improved. Therefore, we finally included tumor size, LVI, depth of invasion, dMMR, CA242 and macroscopic type to establish the nomogram, which showed good discrimination (AUC=0.823, 95%CI: 0.757-0.889) and calibration. Decision curve analysis showed that the nomogram had good clinical performance. Conclusion: We developed a risk prediction model for lymph node status in early gastric signet ring cell carcinoma, which can be used for patient consultation and treatment decision-making.