Clinical-radiomic Analysis And Model For Evaluating The Treatment Response To First TACE In Hepatocellular Carcinoma

Background and aimThe incidence and mortality of hepatocellular cancer(HCC)have been increasing recently.Many patients with HCC are diagnosed at the intermediate/late stage because of no specific symptoms in the early stage.According to the European and American guidelines for the management of HCC,transarterial chemoembolization(TACE)is recommended as the first-line therapy for patients with intermediate-stage HCC,but only half of HCC patients show objective response to first TACE.Notably,the objective response to first TACE is one of the significant prognosis factors of HCC patients undergoing TACE.Patients who have an objective response after first treatment can often benefit from TACE,while patients who have a non-objective response will suffer from liver damage after first treatment.Therefore,it is extremely important to evaluate the treatment response to first TACE accurately and early.In this study,we aim to identify the significant clinical and radiomics features,then develop a predictive model for evaluating treatment response to first TACE in HCC. MethodsWe retrospectively collected and analyzed the related information,including clinical features,radiological data and follow-up,from HCC patients who received TACE treatment in our hospital and other two medical centers between January 2010 and December 2014.Based on the inclusion and exclusion criteria,patients from our Hospital were identified as the primary cohort,which was randomly divided into a training cohort and an internal validation cohort,based on a random split-sample(3:1)approach.The external validation cohort comprised those patients who fulfilled the selection criteria from two other medical centers.A predictive CR model/nomogram based on RSS and clinical factors was developed using multivariate logistic regression.Calibration curves and area under the receiver operating characteristic curves(AUCs)were used to evaluate the model’s performance.This model was further validated with an independent external cohort.ResultsA total of 473 liver cancer patients in our hospital who met the enrollment and exclusion criteria were randomly divided into training cohort(N = 355)and internal verification cohort(N = 118)at a ratio of 3:1 in this study.Meantime,122 patients who fulfilled the selection criteria from two other medical centers as an external validation cohort.In the training cohort,univariate analysis suggested the antiviral treatment(P =0.048),preoperative Alpha-fetoprotein(AFP)(P < 0.001),Barcelona Clinic Liver Cancer staging classification(BCLC)B sub-classification(P = 0.003),tumor size(P = 0.003),tumor location(P < 0.001),tumor number(P = 0.001)and arterial enhancement(P <0.001)were related to the objective response to first TACE.The least absolute shrinkage and selection operator(LASSO)method is used to identify the significant radiomic characteristics of the treatment response to first TACE.The RSS-TP 10 mm included 18 radiomic features(14 features from target tumor at arterial phase,3 features target tumor at non-contrast phase and only one feature from peritumoral at arterial phase).In addition,patients with lower scores(radiomic signature score [RSS] ≤ 2.3)usually have worse treatment response to first TACE.The RSS-TP 10 mm was significantly correlated with the treatment response to first TACE(P < 0.001).The final clinical-radiomic(CR)model consisted of five independent predictors,including RSS-TP 10 mm(P < 0.001),AFP(P= 0.004),BCLC B subclassification(P = 0.010),tumor location(P = 0.039),and arterial hyperenhancement(P = 0.050).The internal and external validation results demonstrated the high-performance level of this model,with internal and external AUCs of 0.94 and0.90,respectively.Compared with other models respectively based on clinical characteristics(AUC = 0.73)or significant radiomic characteristics(AUC = 0.80),the CR model showed a higher performance.In addition,the predicted objective response via the CR model was associated with improved survival in the external validation cohort(hazard ratio [HR],2.43;95% confidence interval,1.60 to 3.69;P < 0.001).Other prognostic factors included AFP(P = 0.014),BCLC B subclassification(P < 0.001),arterial hyperenhancement(P = 0.029),and salvage treatments after failure of TACE(P= 0.047).The predicted treatment response also allowed for significant discrimination between the Kaplan-Meier curves of each BCLC B subclassification.Conclusions A CR model was developed and validated for evaluating treatment response to first TACE in HCC.It had an excellent performance in predicting the first TACE response in patients with HCC and could provide a robust predictive tool to assist with the selection of patients for TACE.