The Predictive Value Of Multi-modality MRI In Hepatocellular Carcinoma Response To Transarterial Chemoembolization

BackgroundHepatocellular carcinoma is the most common primary malignant hepatic tumor worldwide and the third most common cause of cancer-related death.Transarterial chemoembolization is the standard treatment for patients with intermediate-stage HCC.But the outcome of intermediate-stage HCC treated with TACE is greatly heterogeneous.Current means for predicting HCC response to TACE are lacking.Intravoxel incoherent motion and amide proton transfer imaging are two techniques that have been previously applied to evaluate the therapeutic efficacy and prognosis of various malignant tumors.The development of these new technologies provides a new direction for predicting HCC response to TACE.ObjectiveTo investigate that the predictive value of multi-modality MRI of amide proton transfer(APT)and intravoxel incoherent motion(IVIM)imaging in HCC response to TACE.MethodsFrom September 2018 to February 2020,56 patients with intermediate-stage HCC were enrolled and scanned with conventional MRI sequences(coronal T2 WI sequence,axial T1 WI,T2WI sequence)and study sequences(axial IVIM sequence,axial APT sequence).Two abdominal radiologists read the film together and independently selected the corresponding regions of interest(ROI).Pretreatment APT signal intensity(SI),apparent diffusion coefficient(ADC),true molecular diffusion coefficient(D),pseudo-diffusion coefficient(D*)and perfusion fraction(f)for tumor,peritumoral and normal tissues were independently measured by two radiologists.The average values of two observers were included in the final data.Follow-up MR scanning was performed,and the patients were classified as responders or non-responders based on the modified Response Evaluation Criteria in Solid Tumors(m RECIST)criteria.The imaging parameters were compared among the three tissues using an analysis of variance.The imaging parameters were compared between the two groups using two sample t test.Univariate and multivariate logistic regression analyses were used to evaluate the relative value of clinical and laboratory variables and different parameters as risk factors for early non-responder.The prediction model’s variables were derived from univariate and multivariate logistic regression analyses.Receiver operating characteristic(ROC)curve analysis was used to explore the predictive performance.Results1.There were significant difference in APT SI,ADC,D and f values among the tumor,peritumoral and normal tissues(both P < 0.001).2.In the tumor tissue,the APT SI of the response group was significantly lower than that of the non-response group(P=0.009).The D value of the response group was significantly higher than that of the non-response group(P=0.001).In the peritumoral tissue,the APT SI of the response group was significantly lower than that of the non-response group(P=0.035)and the ADC and D values of the response group were significantly higher than those of the non-response group.3.Logistic regression analysis showed that only tumor size(P=0.038),APT value(P=0.034)and D value(P=0.038)were independent predictors of tumor response to TACE.So we established a prediction model that integrated the APT SI and D values in the tumor tissue and the tumor size.A concrete calculation formula was expressed as follows: logit(P)= 0.031 × tumor size + 4.637 × D –0.349 × APT SI.4.ROC curve analyses revealed that the model was better able to predict tumor response to TACE(area under the ROC curve = 0.851)than the individual parameters on their own.(Z=2.457,P=0.003;Z=1.967,P=0.047;Z=2.092,P=0.028 compared with ROC curves of tumor size,D and APT SIs of the tumor tissue,respectively).ConclusionIVIM combined with APT imaging is an effective method to predict the response of TACE in the treatment of hepatocellular carcinoma.The lower APT value and the higher D value suggest that TACE is more responsive.