Development And Validation Of A Predictive Model For The Survival Of Early-onset Gastric Cancer Patients

Objective:The aim of the study was to identify risk factors affecting overall survival(OS)of non-metastatic early-onset gastric cancer(EOGC)(age at diagnosis<40 years old),and to establish a nomogram for prognostic prediction using data from the Surveillance,Epidemiology and End Results(SEER)database.Furthermore,this study sought to externally validate this nomogram in an independent patient cohort.On the other hand,we aimed to investigate the relationship between age and the rate of lymph node metastasis.Material and methods:Patients with non-metastatic EOGC in SEER database(2006-2015)were selected as a training set.Patients with non-metastatic EOGC in our center(2009-2015)were selected as a validation set.The univariate and multivariate Cox proportional hazards regression model was used to identify factors predictive of OS in the training set.The independent predictive factors were subsequently integrated into a nomogram.In the training set and validation set,the bootstrap method was used to calculate the C index of the nomogram model.The calibration curve was utilized to evaluate the predictive accuracy of the nomogram.Furthermore,the predictive accuracy of the nomogram was compared to that of the eighth American Cancer Joint Committee(AJCC)-Tumor-node-metastasis(TNM)staging.Patients with resectable and non-metastatic EOGC with at least one lymph node harvested were retrieved from the SEER database(2006-2015).Logistic regression was utilized to evaluate the relationship between the age and lymph node metastasis.Results:The training set and validation set contains 559 and 201 patients with non-metastatic EOGC,respectively.Discrepancy was detected between the two groups in terms of race,tumor size,tumor subsite,Lauren classification,depth of invasion,lymph node metastasis,tumor stage and surgical therapy(p<0.05).The multivariate Cox model identified race,tumor subsite,tumor size,depth of invasion,lymph node metastasis,number of examined lymph nodes,and surgery as independent covariates associated with OS.The nomogram exhibited superior discriminative power than the eighth AJCC-TNM staging system in both the training set[Harrell’s concordance index(C index):0.762 vs.0.635,p<0.001]and validation set(C index:0.805 vs.0.712,p=0.176).Calibration of the nomogram was good in both cohorts.A total of 12918 patients with resectable and non-metastatic gastric cancer with at least one lymph node harvested were retrieved from the SEER database.In the logistic regression model,age was identified as an independent risk factor of lymph node metastasis,while there was no association between the age and lymph node metastasis in early gastric cancer.Conclusion:We developed a nomogram with great accuracy of survival prediction for patients with non-metastatic EOGC,which was superior to that of the 8th AJCC-TNM staging system.This nomogram model could visually predict the patient survival and individualize the treatment and follow-up strategies.EOGC was more likely to have lymph node metastasis in the whole cohort,but not in the early gastric cancer.