Analysis Of Prognostic Factors And Construction Of Nomogram Models For PM0 Stage Diffuse-Type Gastric Adenocacinoma Based On The SEER Database

Objective: The clinicopathological traits and prognostic associated factors of patients with primary p M0 stage diffuse-type gastric adenocarcinoma(D-GAC)in the Surveillance,Epidemiology,and End Results(SEER)database were retrospectively analyzed,and the visual prognostic Nomogram model was established to evaluate the1-year,3-year,and 5-year overall survival(OS),in order to provide a more significant reference for each patient’s individualized prognostic evaluation.Methods: We screened the SEER database and downloaded data from primary D-GAC patients between 2004 and 2015.Univariate and multivariate COX proportional hazards regression analysis was used to identify independent prognostic factors.Patients were divided into the high-risk group(n=1381)and the low-risk group(n=1384)by the median risk score(median risk score=1.450969261).The survival differences between the two groups were examined using the Kaplan-Meier curves.The clinical predictive capacity of the risk score model and each independent prognostic factor was evaluated using time-dependent receiver operating characteristic(t-ROC)curves.Then independent prognostic factors with good predictive power were included in the nomogram,whose prediction accuracy was assessed using the ROC curves and the calibration curves.Finally,an accurate and discriminable prognostic nomogram model was created.Results: A total of 2765 patients in the 17 registry study data from the SEER database met our inclusion and exclusion criteria and were included in the analysis.The results of the univariate COX regression analysis showed that 10 indicators,including race,age,marital status,tumor size,T stage,N stage,TNM stage,the number of regional nodes examined,r N stage,and time from diagnosis to surgery were significantly associated with OS in primary p M0 stage D-GAC patients(P<0.05).Multivariate COX regression analysis showed that race,age,marital status,tumor size,T stage,N stage,the number of regional nodes examined,r N stage,and time from diagnosis to surgery were all independent prognostic factors affecting OS in primary p M0 stage D-GAC patients(P<0.05).The Kaplan-Meier survival analysis and log-rank test showed that overall survival was considerably shorter in the high-risk group compared with the low-risk group(P<0.001),which had significantly lower the 1-year,3-year,and 5-year OS(P=0.032).The risk score(AUC=0.827),r N stage(AUC=0.757),T stage(AUC=0.751),N stage(AUC=0.735),and tumor size(AUC=0.685)all displayed good accuracy in predicting OS,according to the time-dependent ROC curves.Higher risk scores were displayed in patients who were older,in later T and N stages,and in higher r N stages.On the basis of race,age,tumor size,T stage,N stage,and r N stage,the predictive nomogram model was created for the 1-year,3-year,and 5-year OS.The calibration curves also exhibited adequate fit and ideal net benefit in prediction and clinical application.The nomogram showed favorable sensitivity,with the area under the ROC curves(AUC)being 0.751,0.827,and 0.841,respectively.Conclusions: In the SEER data set of this study,patients with a white population,older age,larger tumor size,and later T stages,N stages,and r N stages showed higher risk scores and worse prognosis.The visualized nomogram constructed has high accuracy and clinical application in predicting 1-year,3-year,and 5-year OS of primary p M0 stage D-GAC patients,which will help clinicians make an individualized prognostic assessment of D-GAC patients with non-distant metastases.