Prediction Model of Moderate and Severe Abdominal Pain after TACE for Hepatocellular Carcinoma Based on CT Imaging and Clinical Multivariate Regression Analysis
Objective To investigate the risk factors for acute severe abdominal pain after hepatocellular carcinoma(HCC)patients with hepatocellular carcinoma(HCC)after hepatic artery catheter chemoembolization(TACE),and to construct radiomics models and logistic regression models,respectively.Methods A retrospective analysis was performed for 128 patients with pathologically confirmed hepatocellular carcinoma and TACE surgery,and their imaging data and clinical data were collected.According to the 8:2 divided into training set and verification set,according to the BPI scoring standard,"0～3"score no or mild pain was negative group,a total of 57 cases,"4～10"moderate and severe abdominal pain was divided into positive group,a total of 71 cases,compared the clinical and imaging data of patients with moderate and severe pain with no or mild pain,radiomics and multivariate logistic regression analysis were used to screen risk factors,and a predictive model was established.Subject-operated(ROC)curve analysis predicts potency.Results The radiomics model:The models of arterial phase,portal phase,delayed phase,arterial phase+portal phase,arterial phase+delay period,portal phase+delay period and arterial phase+portal phase+delay period were analyzed,and the results showed that the AUC value of the training set of the combined omics model of arterial phase+portal phase+delay period was 0.79 and the AUC value of the validation set was 0.78,which was higher than that of the other six omics models.Logistic regression analysis of clinical factors showed the history of TACE,the history of abdominal pain after TACE,whether the distance between the tumor and the liver capsule was＞1cm,whether the maximum diameter of the tumor was＞5cm,and whether the number of tumors was＞2.Whether there is vascular invasion is its independent risk factor,and the AUC value of the prediction model established by clinical factor logistic regression analysis to screen independent factors is 0.85.Conclusion The combined model of arterial phase+portal phase+delayed phase in radiomics has certain clinical value for predicting the occurrence of abdominal pain after TACE surgery for hepatocellular carcinoma,which provides certain value for future radiomics for the prediction of abdominal pain after TACE surgery for hepatocellular carcinoma,and the clinical multivariate regression analysis model can provide more clinical factors and has good efficacy,and the omics model and the clinical multivariate regression model complement each other,and the prediction model is established to provide better personalized pain management for the clinic.Alleviates the patient's suffering.
NETL
NSTL
万方数据
维普
