Survival Analysis Of Children And Adults With Liver Tumors After Hepatectom

Objective Hepatoblastoma(HB)is the most common primary liver malignant tumor in children,ranking third in the incidence of solid tumors in children.Surgical treatment is the primary treatment for early hepatoblastoma.HB has a high degree of malignancy and poor prognosis.This study investigated the prognostic factors of pediatric hepatoblastoma undergoing radical resection.Methods 75 cases of HB radical resection in our hospital were retrospectively analyzed.Univariate and multivariate COX regression analysis was performed on clinicopathological data and follow-up data of children undergoing radical hepatocellular tumor surgery to explore the risk factors for disease-free survival(DFS)and overall survival(OS)in the study.KM(Kaplan-Meier)survival analysis was used to analyze the influence of different risk factors on DFS and OS,and the log-rank test was used to compare the survival between groups.Results There were 75 HB children,47 males,and 28 females.1-year,3-year,and 5year overall survival rates were 94.6%,90.7%,and 89.3%respectively in this study.Univariate COX regression analysis for predicting DFS showed that age,platelet count,microvascular invasion,and clinical stage were correlated with DFS of hepatoblastoma(P<0.05).Multivariate COX regression analysis and KM survival analysis showed that age>2 years,microvascular invasion,and Pretext III-IV were independent risk factors for predicting DFS(P<0.05).Univariate COX regression analysis for predicting OS showed that age and pathological type were correlated with the overall survival of hepatoblastoma(P<0.05).Multivariate COX regression analysis and KM survival analysis showed that age>2 years and the pathological type was mixed hepatoblastoma were independent risk factors for predicting postoperative OS(P<0.05).Conclusion Age>2 years,microvascular invasion,and Pretext III-IV were independent risk factors for predicting DFS in HB patients undergoing radical resection.Age>2 years and mixed hepatoblastoma were independent risk factors for predicting postoperative OS.Objective Pediatric hepatocellular carcinoma is the second most common malignant liver tumor in children after hepatoblastoma.The incidence rate of hepatocellular carcinoma in children is 0.3-0.45 per million per year.The incidence rate is relatively low and the number of cases is relatively small.Adult liver cancer was studied to explore the value of the multimodal radiomics model based on CT and MRI in predicting the prognosis of hepatocellular carcinoma resection and its possible use as a prognostic indicator for pediatric tumor evaluation.Methods 103 cases of hepatocellular carcinoma(HCC)confirmed by histopathology were retrospectively studied in our hospital.Patients were randomly divided into the training group(73 cases)and the validation group(30 cases)in a 7:3 ratio.1217 radiomic features were extracted from tumor areas on CT and MRI images of patients at each stage.COX regression analysis of single factor,spearman correlation analysis,pearson correlation analysis,minimum absolute shrinkage,and selection operator COX regression analysis were used to reduce the dimensionality of CT and MRI radiomics features and the combination of CT and MRI radiomics features in the training group,and the features most related to DFS and OS were screened out.Computing radiomics score(Radscore)based on selected features.COX proportional risk regression model was established to predict postoperative DFS and OS of HCC patients.The C-index value was calculated to evaluate the predictive performance of the model and verified in the validation group.Radscores were divided into a high-risk group and a low-risk group according to the optimal truncation value of each score.KM survival analysis method was used to analyze the influence of the high-risk groups on DFS and OS,and log-rank was used to test whether the survival difference was statistically significant.The individual survival assessment was constructed based on the multimodal radiomics score and clinical risk factors,and the calibration curve was used to evaluate the model consistency.Results Radiomic score was an independent risk factor for predicting DFS and OS after liver cancer operation(P<0.05).KM survival analysis showed that the postoperative survival of patients in the high-risk group was significantly lower than that in the low-risk group,and the difference was statistically significant(P<0.05).CT+MRI+Clinical multimodal radiomics model showed the best efficacy in predicting DFS and OS of HCC.In predicting postoperative DFS,the C-index of the training group was 0.858(95%CI 0.811-0.905),and that of the validation group was 0.704(95%CI 0.56-0.845).In predicting postoperative OS,the C-index of the training group was 0.893(95%CI 0.8460.940)and that of the validation group was 0.738(95%CI 0.575-0.901).The multimodal radiomics nomogram has a certain value in predicting postoperative survival of HCC patients,and the calibration curves show good consistency.Conclusion The multimodal radiomics model based on CT and MRI radiomics features and clinical risk factors is effective in predicting the postoperative survival of HCC patients.Objective HCC has a low incidence in children,and the number of cases is relatively small.Adult liver cancer was studied to explore the value of radiomics features based on enhanced CT in predicting extrahepatic metastasis after hepatocellular carcinoma and its possible use as a prognostic indicator for pediatric tumor evaluation.Methods a total of 277 HCC patients from our hospital were retrospectively selected and randomly divided into a training group(193 cases)and a validation group(84 cases)according to the ratio of 7:3.The radiomics features were extracted from enhanced CT images.We adopted the feature reduction algorithm based on random forest recursive elimination(RFC-RFE)and the minimum absolute contraction and selection operator(LASSO).Radiomics score was calculated based on selected features.Five classifiers(logistic regression algorithm classifier,support vector machine algorithm classifier,random forest algorithm classifier,and decision tree algorithm classifier,naive bayes algorithm classifier)are adopted in this study to construct radiomics models to predict the efficacy of extrahepatic metastasis.The clinical model was constructed based on clinical risk factors and the combined model was constructed based on radscore and clinical risk factors.The efficacy of each model in predicting hepatocellular carcinoma metastasis was observed according to the area under the ROC line(AUC),accuracy,and other indicators.In addition,we used the minority oversampling(SMOTE)amplifying data in the training group for training and validation again due to the frequency imbalance of extrahepatic metastasis.The experiment was also repeated by reducing the number of cases and redistributing the data set to observe whether the selected features and the constructed model were robust.The nomograms of the combined model and clinical model were drawn.Decision curves were used to evaluate the clinical usefulness of rosette in predicting extrahepatic metastasis of HCC.The rationality of the radiomics model was explained by the shapley additive explanation algorithm(SHAP).Results Radscore was an independent predictor of extrahepatic metastasis of HCC(P<0.05).The effectiveness of the radiomics model combined with SMOTE algorithm in predicting extrahepatic metastasis is obviously better than that of the non-SMOTE radiomics model.SMOTE algorithm,LASSO algorithm,RFC-RFE algorithm and logistic regression algorithm combined with SMOTE algorithm had the best performance in predicting extrahepatic metastasis of HCC.The AUC of the training and validation groups was 0.91/0.81(95%CI 0.89-0.94/0.70-0.91),respectively.Combining radiomics scores with clinical risk factors[body mass index(BMI),tumor diameter,microvascular invasion,neutrophil count,and T stage]to construct a combined model.The predictive efficacy of the combination model[AUC 0.90(95%CI 0.84-0.95)]was superior to that of the clinical model[AUC 0.85(95%CI 0.76-0.93)]and the radiomics model[AUC 0.78(95%CI 0.78)0.68-0.89)]in the validation group.The decision curve showed that the combined model had a higher net benefit than the clinical model and radiomics model in predicting extrahepatic metastasis of HCC.We verified the predictive ability of each group of models by reducing the number of cases and assigning data sets.Small differences were observed in AUC values,accuracy values,and other indicators of each classifier model,indicating that the radiomics features we screened and the radiomics model constructed were robust.The SHAP interpretation model showed that the radiomic features were highly correlated with the risk of extrahepatic metastasis,and the prediction model based on radiomic features was interpretable.Conclusion Machine learning models based on radiomics features can predict hepatocellular carcinoma metastasis after the operation and the models are robust.Objective HCC has a low incidence in children,and the number of cases is relatively small.This study took adult hepatocellular carcinoma as an example to explore the value of enhanced CT-based radiomics features in predicting HCC microvascular invasion and its potential as a prognostic indicator for pediatric cancer.Methods A total of 277 HCC patients in our hospital were selected for retrospective analysis.The patients were randomly divided into the training group(203 cases)and the validation group(68 cases)in a 3:1 ratio.Radiomic features were extracted from each tumor region on CT images of each patient based on the minimum redundancy maximum relevancy and the LASSO algorithm.The radiomics score was calculated and the radiomics model was constructed.The clinical model was constructed based on clinical risk factors and the combined model was constructed based on radscore and clinical risk factors,and draw the nomogram of the combined model.ROC curves were used to evaluate the efficacy of each model in predicting HCC microvascular invasion,and the delong tests were used to compare the efficacy of the combined model with the clinical model and radiomics model.The fitting degree of the combined model was evaluated by correction curve,and the net benefit of the combined model for predicting HCC microvascular invasion was evaluated by decision curve.Results Radiomics score is an independent factor for predicting microvascular invasion of HCC(P<0.05).The combined model showed good predictive efficacy,with an AUC of 0.80(95%CI 0.75-0.85)in the training group and 0.75(95%CI 0.65-0.85)in the validation group.The efficacy of the combined model in predicting microvascular invasion of HCC was significantly higher than the radiomics model and clinical model(P<0.05)in the training group.The decision curve showed that the combined model had a higher net benefit than the clinical model in predicting HCC microvascular invasion.Conclusion The prediction model constructed based on CT radiomics features and clinical risk factors of HCC patients has certain efficacy in preoperative prediction of HCC microvascular invasion.