Intraarterial contrast-enhanced ultrasound to predict the short-term tumour response of hepatocellular carcinoma to Transarterial chemoembolization with Lipiodol

Abstract Background Transarterial chemoembolization (TACE) is an effective locoregional therapy in hepatocellular carcinoma (HCC). However, it is difficult to predict the tumour response (TR) of TACE intraprocedurally. The aim of this study was to predict the TR after TACE (1–3 months) in HCC patients using intraprocedural intraarterial contrast enhanced ultrasound (IA-CEUS). Methods In this case-control study, consecutive patients who received TACE in our hospital from September 2018 to May 2019 were enrolled. IA-CEUS was performed before and after TACE. Postoperative contrast-enhanced liver MRI was performed 1–3 months after TACE as the gold standard. According to the modified Response Evaluation Criteria in Solid Tumours (mRECIST), ultrasonic manifestations were compared between the complete remission (CR) group and non-CR group by univariate and multivariate analyses. A logistic predictive model was established and validated, and its diagnostic efficiency was evaluated. Results Forty-four patients with sixty-one lesions were enrolled in the study. Multivariate analysis identified, the risk factors as a large lesion diameter (OR: 1.84; 95% confidence interval [CI]: 1.009, 3.080; P = 0.020), a larger dimension of non-enhancing area in superior mesenteric artery (SMA)-CEUS than the size in B-mode ultrasound preoperatively (OR: 3.379; 95% CI: 1.346,8.484; P = 0.010), presence of corona enhancement in hepatic artery (HA)-CEUS postoperatively (OR: 6.642; 95% CI: 1.214, 36.331; P = 0.029), and decreased corona enhancement thickness (per centimetre) postoperatively (OR: 0.025; 95% CI: 0.006,0.718; P = 0.025). The area under the receiver operating characteristic curve (AUROC) of the predictive model was 0.904 (95% CI: 0.804, 0.966; P < 0.001). The sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were 81.08, 91.67, 85.25, 93.75, and 75.86%, respectively. Leave-one-out cross-validation (LOOCV) showed that the accuracy was 77.05%. Conclusions Intraprocedural IA-CEUS can be used to predict the TR in HCC patients after TACE