Establishment And Validation Of The Prognostic Nomograms For Multifocal HCC And Multifocal ICCA Based On SEER Database

Background:The incidence and mortality rates of primary liver cancer remain high worldwide.Despite significant advances in screening and treatment methods,studies have shown that the proportion of patients with multifocal hepatocellular carcinoma and intrahepatic cholangiocarcinoma at initial diagnosis is still around 60%and 75%,respectively.Although there are some relevant studies on multifocal hepatocellular carcinoma and intrahepatic cholangiocarcinoma,large-scale clinical studies establishing a visualization model to predict the prognosis of relevant patients are relatively scarce.In this study,nomogram models was constructed to predict tumor-specific survival for patients with multifocal hepatocellular carcinoma and intrahepatic cholangiocarcinoma using data from the SEER database and was subsequently validated.Methods:This study screened relevant data from the SEER database for 5472 patients with multifocal hepatocellular carcinoma and 382 patients with multifocal intrahepatic cholangiocarcinoma.The data was randomly divided into training and validation groups in a 7:3 ratio.Univariate and multivariate Cox regression analyses were used to identify independent risk factors for CSS in patients.These independent risk factors were used to construct a nomogram prediction model and plot a nomogram for visualization.The constructed nomogram model was evaluated and validated using C-index,ROC curve,calibration curve,net reclassification index(NRI),integrated discrimination improvement(IDI)index,and DCA curve.In addition,this study used the nomogram model to calculate the specific score for each patient.Patients in each group were stratified according to their risk based on the obtained cutoff value and the effectiveness of the risk stratification of the constructed nomogram prediction model was evaluated.Results:Multivariate Cox analysis results for patients with multifocal hepatocellular carcinoma showed that age,race,marital status,maximum tumor diameter,vascular invasion status,AFP level,tumor pathological grade,AJCC stage,surgery,and chemotherapy were independent risk factors for CSS.Multivariate Cox analysis results for patients with multifocal intrahepatic cholangiocarcinoma showed that tumor vascular invasion status,tumor pathological grade,surgical treatment and chemotherapy were independent risk factors for CSS.Based on these results,this study established a traditional nomogram and its web version and then validated it.The C-index of the nomogram prediction model for patients with multifocal hepatocellular carcinoma in the training and validation groups were 0.729(95%CI 0.724-0.734)and 0.724(95%CI:0.716-0.732),respectively.The C-index of the nomogram prediction model for patients with multifocal intrahepatic cholangiocarcinoma in the training and validation groups were 0.704(95%CI:0.685-0.723)and 0.743(95%CI:0.715-0.771),respectively.ROC curve,calibration curve,NRI,IDI index and DCA curve all showed that the established model had good prognostic prediction ability and clinical practicality.Compared with the AJCC staging system,the nomogram prediction model had better ability to predict patient prognosis and stratify risk.Conclusion:This study established corresponding prognostic nomogram prediction models for patients with multifocal hepatocellular carcinoma and multifocal intrahepatic cholangiocarcinoma based on data from the SEER database.These models have high prognostic prediction efficiency and risk stratification ability and can provide clinical decision-making reference for relevant patients.