Gd-EOB-DTPA Enhanced MRI And Various Models Of Diffusion Weighted Imaging For Prediction Of The Early Therapeutic Response To TACE In Patients With Hepatocellular Carcinoma

Objective:To explore the value of Gadolinium Ethoxybenzyl Diethylene-triamine-pentaacetic-acid(Gd-EOB-DTPA)enhanced magnetic resonance imaging(MRI)and various models(including mono-exponential,bi-exponential and stretched exponential model)of diffusion weighted imaging(DWI)for predicting the early therapeutic response to transarterial chemoembolization(TACE)in patients with hepatocellular carcinoma(HCC).Materials and methods:A retrospective analysis was performed on 162 patients with 162 primary hepatocellular carcinoma(HCC)who received TACE treatment for the first time in Nanfang Hospital of Southern Medical University from January 2020 to December 2021,including 136 males and 26 females.All cases underwent Gd-EOB-DTPA enhanced MRI and 9-b-value diffusion weighted imaging(DWI)scans within 1 month before TACE and underwent CT or MRI enhanced scan within 3 months after TACE.A total of 162 HCC lesions were divided into a non-response group(including SD and PD)and a response group(including CR and PR)after TACE by modify response evaluation criteria in solid tumors(mRECIST).Clinical data(gender,age,AFP,diameter of tumor,liver function indexes,etc.)and MRI imaging characteristics(tumor margin,peritumoral enhancement,tumor rim enhancement,peritumoral hypointensity,etc.)were compared between the non-response and response group.Two radiologists with experience in abdominal radiology drew ROIs on axial DWI images with b=1000s/mm2.Whole-tumor ROIs were manually delineated to encompass as much of the entire HCC lesions in each slice as possible.Various models of DWI parameters are obtained:parameter of the mono-exponential included the apparent diffusion coefficient(ADC),parameters of the bi-exponential model included true diffusion coefficient(D),perfusion related diffusion coefficient(D*),perfusion fraction(f)values and parameters of stretched exponential model included distributed diffusion coefficient(DDC),distributed heterogeneity in voxels(α)values.Histogram parameters included percentiles(5th,25th,50th,75th,95th),mean,skewness and kurtosis.Histogram parameters were compared between HCCs with non-response and response group.For significant parameters,receiver operating characteristic(ROC)curves were further plotted to compare the diagnosis performance.Combined the parameters by binary logistic regression analysis,and finally calculated the diagnostic efficiency of the combined parameters by ROC curve and AUCs(area under the curve).DeLong test were used for statistical analysis of AUCs.Results:Confirmed by the mRECIST,there were 92 cases in the response group(including 62 cases of complete response and 32 cases of partial response)and 70 cases in the non-response group(including 38 cases of stable disease and 32 cases of progressive disease).The tumor size and non-smooth tumor margin were statistically significant(P<0.05).The 5th percentiles of ADC,the 5th and 25th percentiles of D and the mean,5th,25th,50th percentiles of DDC between the two groups were statistically significant(all P<0.05).The histogram parameters of D*,f and α showed no statistically significant differences between the two groups(all P>0.05).The largest AUC value of 0.793 of DDC(5th percentile)showed higher accuracy with a cut-off of 0.69x103mm2/s,the 5th percentile of DDC value provided a sensitivity of 95.2%and a specificity of 58.4%in the prediction of the early therapeutic response of TACE.The significant parameters were combined in different combinations,and the ROC curve analysis was performed after the combination.The AUC value of the combination of the DWI parameters histogram(ADC+D+DDC),the maximum diameter of the tumor and the tumor margin achieved the largest of 0.864,and the sensitivity was 92.8%,the specificity was 72.6%.According to DeLong test,there was significant difference of the AUC value between the combined parameters and others(P<0.05),while there was no significant difference of AUC of other parameters(P>0.05).The other combined parameters included D value histogram(5th+25th percentile of D),DDC value histogram(mean+5th+25th+50th percentile of DDC),the histogram of mutli-model DWI parameters(ADC+D+DDC value histogram).Conclusion:The histogram of multiple DWI models parameters(ADC value of mono-exponential model,D value of bi-exponential model,DDC value of stretched exponential)are beneficial to predict the early efficacy of HCC patients after TACE.In the single parameter,the 5th percentile of DDC value has the best diagnostic performance.When the DWI parameters of various models are combined with the Gd-EOB-DTPA enhanced MRI characteristics,the diagnostic performance is improved.