Analysis Of Independent Risk Factors For Tumor Deposits In Patients With Gastric Cancer And Construction And Validation Of A Nomogram

Objective:To explore the independent risk factors for tumor deposits in patients with gastric cancer,to construct a nomogram to predict the occurrence of tumor deposits in patients with gastric cancer,and to evaluate,internally and externally validate the nomogram.Methods:A retrospective case-control study method was used,and the samples were all from patients with gastric cancer admitted to the Affiliated Hospital of Yan’an University.According to the inclusion and exclusion criteria,the data of 822 patients were collected from January 2014 to December 2020 for the construction,evaluation and internal validation of the nomogram.Following the same criteria,the data of 117 patients were collected from January 2021 to December 2021 for the external validation of the nomogram.According to the presence or absence of tumor deposits detected by postoperative pathology,patients with gastric cancer were divided into tumor deposits-negative group and tumor deposits-positive group.Pearsonχ~2 test was used for univariate analysis,and binary logistic regression analysis was used for multivariate analysis to identify independent risk factors for tumor deposits in patients with gastric cancer.A nomogram for predicting the probability of tumor deposits in patients with gastric cancer was constructed.ROC curves were drawn and AUC values were calculated to evaluate the predictive power of the nomogram in the model building group and the external validation group.The internal verification adopted the Bootstrap self-sampling method,the sampling times were set to 1 000 times,and the C-index value tested the discrimination of the nomogram.The degree of calibration of the nomogram was assessed using the graphical calibration method and the Hosmer-Lemeshow goodness-of-fit test.P<0.05 indicated that the difference was statistically significant.Results:The positive rate of tumor deposits in patients with gastric cancer in the model construction group was 32.48%（267/822）.Male（OR=2.082,95%CI:1.027～4.221,P=0.042）,serum CA72-4≥6.9 U/ml（OR=2.486,95%CI:1.360～4.543,P=0.003）,poorly differentiated or undifferentiated tumor（OR=1.815,95%CI:1.047～3.146,P=0.034）,tumor invasion depth T₃～T₄（OR=3.636,95%CI:1.645～8.036,P=0.001）,lymph node metastasis（OR=6.932,95%CI:4.001～12.011,P<0.001）,lymphovascular invasion（OR=5.168,95%CI:2.769～9.645,P<0.001）and perineural invasion（OR=4.257,95%CI:2.440～7.427,P<0.001）were independent risk factors for tumor deposits in patients with gastric cancer.The independent risk factors were listed separately and quantified in the form of scores,and the cumulative scores of all independent risk factors were matched with the outcome scale to generate a nomogram predicting the probability of tumor deposits in patients with gastric cancer.The AUC value of the nomogram was 0.952,which was higher than the independent risk factors.The diagnostic cutoff value of the nomogram was 0.367,the sensitivity was 0.876,and the specificity was 0.892.The C-index value of the Bootstrap self-sampling method was 0.948.Both the model performance curve and the bias correction curve in the graphic calibration method were close to the ideal model curve.The result of Hosmer-Lemeshow goodness-of-fit test wasχ~2=14.047,P=0.081.The AUC value for the external validation of the nomogram was 0.905.Conclusion:The independent risk factors for tumor deposits in patients with gastric cancer are male,serum CA72-4≥6.9 U/ml,poorly differentiated or undifferentiated tumor,tumor invasion depth T₃～T₄,lymph node metastasis,lymphovascular invasion and perineural invasion.The nomogram constructed in this study has high predictive accuracy and could more reliably assess the risk of tumor deposits in patients with gastric cancer,reflecting good practical application value.