Construction And Clinical Applicability Evaluation Of A Novel Prognostic Assessment Model For Gastric Cancer Based On The Negative Lymph Node Count

PART 1 STUDY OF THE PROGNOSTIC VALUE OF THE NEGATIVE LYMPH NODE COUNT IN PATIENTS WITH GASTRIC ADENOCARCINOMA AFTER SURGERYObjective: To investigate the clinical value of negative lymph node count(NLNC)in assessing the prognosis of patients with gastric adenocarcinoma.Methods: Clinicopathological data of 11897 patients with gastric adenocarcinoma undergoing surgery registered in The Surveillance,Epidemiology,and End Results(SEER)program database from January 2004 to December 2019 were used.The X-tile software determined the optimal cut-off value for the NLNC,Kaplan Meier survival curve,and multivariate Cox proportional risk model were used to assess the prognostic factors of overall survival(OS)and cancer-specific survival(CSS)in gastric adenocarcinoma patients after surgery and at different pathological p T stage and p N stage were analyzed in subgroups,and receiver operating characteristic(ROC)curves were plotted for different groups of NLNC predicting 3-year,5-year and 10-year OS and CSS in gastric adenocarcinoma patients,respectively.Results: The best cut-off values of NLNC classified by X-tile software were 3 and13,so NLNC was divided into three subgroups of ≤3,4～13,and >13,named NLNC1,NLNC2,and NLNC3.Respectively,the 5-year OS of the three groups were 21.7%,39.5%,and 55.6%,and the 5-year CSS were 28.9%,47.7%,and 63.1%.The results of univariate and multivariate Cox analyses showed that NLNC was an independent prognostic factor for predicting OS in patients with gastric adenocarcinoma,NLNC2 vs NLNC1 [HR=0.687(95% CI,0.645-0.733)] and NLNC3 vs NLNC1 [HR=0.490(95%CI,0.458-0.524)],HR=0.687(0.645-0.733),which was similarly confirmed in predicting CSS for NLNC2 vs NLNC1 [HR=0.678(95% CI,0.631-0.729)] and NLNC3 vs NLNC1 [HR=0.469(95% CI,0.434-0.506)].Subgroup analysis showed that NLNC remained an independent prognostic factor at different p T and p N stages.The AUC for predicting 3-year OS,5-year OS,and 10-year OS in gastric adenocarcinoma patients were 0.625,0.612,and 0.571,respectively;and the AUC for predicting 3-year CSS,5-year CSS,and 10-year CSS were 0.629,0.615,and 0.579,respectively.all of these results suggest that NLNC has a significant effect on predicting the prognosis of patients with gastric adenocarcinoma and has some clinical value.Conclusion: This study showed that NLNC was an independent prognostic factor for OS and CSS in gastric adenocarcinoma patients,and with the increase of cleared and sent NLNC,OS,and CSS of gastric adenocarcinoma patients improved,which should be recommended for clinical application.PART 2 CONSTRUCTION OF A PROGNOSTIC MODEL FOR PATIENTS WITH GASTRIC ADENOCARCINOMA BASED ON THE NEGATIVE LYMPH NODE COUNT AND INTERNAL VALIDATIONObjective: A Nomogram model based on negative lymph node count(NLNC)for the prognosis of gastric adenocarcinoma patients has not been reported.In this study,we proposed to construct an NLNC-based line graph model to accurately predict overall survival(OS)in gastric cancer patients and to evaluate the clinical applicability of the model.Methods: The SEER cohort were divided into a modeling cohort and a validation cohort in a 7:3 ratio according to the random number method,with data from the modeling cohort(8329 cases)used only to construct the Nomogram prediction model and data from the validation cohort(3568 cases)used for internal validation.The prognosis-related variables were first screened in the modeling cohort using univariate Cox regression analysis,and then the screened predictors were included in the multifactor Cox regression for modeling,and the final prediction model was presented in the form of a Nomogram.It was also discriminated and calibrated in both cohorts.The clinical efficacy of the prediction model was assessed by subject operating characteristic(ROC)curves,calibration curves,and clinical decision curve analysis(DCA)curves,and compared with the 8th edition of the postoperative pathological tumor-node-metastasis(p TNM)staging system.Results: Cox proportional risk regression models showed that NLNC was an independent prognostic factor for postoperative prognosis in patients with gastric adenocarcinoma,as were age,race,tumor size,tumor site,p T stage,p N stage,and NLNC.In the training cohort,the Nomogram model predicted 3-year OS,5-year OS,and 10-year OS for gastric adenocarcinoma patients at 0.785,0.788,and 0.761,respectively.in the validation cohort,Nomogram predicted AUC values for 3-year OS(0.777 vs 0.746,P<0.001),5-year OS(0.783 vs 0.746,P<0.001),and 10-year OS(0.765vs 0.720,P<0.001),all showed that the newly constructed model had higher prediction accuracy than the AJCC/UICC p TNM Cancer Staging System(8th edition).The calibration curves showed that the Nomogram model predicted a high degree of patient survival and actual survival.the DCA decision curve of Nomogram was far from the baseline and above the p TNM staging indicating that the model was clinically applicable and superior to the p TNM staging system in terms of clinical application.Individual scores were calculated for all patients according to Nomogram,and based on the calculated optimal cut-off values of 93 and 173,we divided the patients into low-,intermediate-,and high-risk groups,and the prognosis of the three groups showed significant differences.Finally,we developed another web-based version of the dynamic column line graph to facilitate the prediction of survival at different times of the line.Conclusion: In this study,an individualized NLNC-based Nomogram model was constructed,which can accurately predict the 3-year,5-year,and 10-year OS and CSS of gastric adenocarcinoma patients and can be used for risk stratification of long-term prognosis of gastric adenocarcinoma patients,with potential clinical application.PART 3 EXTERNAL VALIDATION OF A POSTOPERATIVE OVERALL SURVIVAL PREDICTION MODEL FOR PATIENTS WITH GASTRIC ADENOCARCINOMAObjective: The Nomogram model for predicting overall survival(OS)of gastric adenocarcinoma patients,which has been constructed in Part II,was externally validated to evaluate the general applicability of the model.Methods: The external validation cohort was 608 patients with gastric adenocarcinoma who underwent surgery at the Second Hospital of Lanzhou University(recorded as an external validation cohort).The baseline data of the SEER cohort and the patients in our center were first compared and then validated on the external validation cohort using receiver operating characteristic(ROC)curves,calibration curves,and clinical decision curve analysis(DCA).Results: Data from our institution showed significant differences in OS among different groups of gastric adenocarcinoma patients with NLNC(x2=14.17,P<0.001),consistent with the results in Part I.In the external validation cohort,areas under the ROC(AUC)values for predicting 3-year and 5-year OS were 0.719 and 0.702,respectively,indicating good agreement with the model;the calibration curve showed good accuracy.The DCA decision curve was far from the baseline and higher than the p TNM staging system within a certain threshold probability,indicating that the model has good clinical application value.Conclusion: The study developed a Nomogram model for predicting OS after surgery in gastric adenocarcinoma patients based on NLNC with relevant clinicopathological features,and external validation showed that the model is universally applicable to gastric adenocarcinoma patients of different ethnicities.The easy-to-use nomogram may provide clinicians with suitable tools to design individualized and precise treatment strategies.