Neural network-based prognostic predictive tool for gastric cardiac cancer: the worldwide retrospective study

Abstract

Abstract Background: Gastric cancer (GC) is one of the most common cancers and the main cause of tumor-related death worldwide. Moreover, the incidence of gastric cardiac cancer (GCC) has increased obviously, with the potentially different prognosis from other sites of GC (non-gastric cardiac cancer, NGCC). We will analyze the prognosis between GCC and NGCC, and set up an effective prognostic model based on neural network for GCC. Methods: In the population-based cohort study, we firstly enrolled the clinical features from the Surveillance, Epidemiology and End Results (SEER) data (n=31397) as well as the public Chinese data from different hospitals (n=1049). Then according to the diagnostic time, the SEER data were then divided into two cohorts, the train cohort (patients were diagnosed as GCC in 2010-2014, n=4414) and the test cohort (diagnosed in 2015, n=957). Age, sex, pathology, tumor, node, and metastasis (TNM) stage, tumor size, surgery or not, radiotherapy or not, chemotherapy or not and history of malignancy were chosen as the predictive clinical features. Results: The prognosis of GCC patients in SEER database was worse than that of NGCC patients, while it was not worst in the Chinese data. The total of 5371 patients were used to conduct the development model, following inclusion and exclusion criteria. Neural network-based prognostic predictive model had a satisfactory performance for GCC overall survival (OS) prediction, which owned 0.7431 AUC in the train cohort (95% confidence intervals, CIs, 0.7423-0.7439) and 0.7419 in the test cohort (95% CIs, 0.7411-0.7428). Conclusion: GCC patients indeed have the different survival time compared with NGCC patients. And this neural network-based prognostic predictive tool is a novel and promising software for the outcome of GCC patients.