Prediction Model for Intermediate‐Stage Hepatocellular Carcinoma Response to Transarterial Chemoembolization

Abstract

Background: The outcome of intermediate-stage hepatocellular carcinoma (HCC) treated with transarterial chemoembolization (TACE) is greatly heterogeneous. Current means for predicting HCC response to TACE are lacking. Purpose: To investigate whether the combination of parameters derived from amide proton transfer (APT) and intravoxel incoherent motion (IVIM) imaging, and morphological characteristics of tumor can establish a better prediction model than the univariant model for HCC response to TACE. Study Type: Prospective. Subjects: 56 patients with intermediate-stage HCC (50 males and six females). Field Strength/Sequences: 3.0T; T 2-weighted-fast spin echo, 3D liver acquisition with volume flex, single-shot fast spin echo-planar imaging (EPI), spin echo-EPI. Assessment: Pretreatment APT signal intensities (SIs), apparent diffusion coefficient (ADC), true molecular diffusion coefficient (D), pseudodiffusion coefficient (D*), and perfusion fraction (f) for tumor, peritumoral, and normal tissues were measured. Follow-up MRI scanning was performed, and the patients were classified as responders or nonresponders based on the modified Response Evaluation Criteria in Solid Tumors (mRECIST) criteria. Statistical Tests: The imaging parameters were compared among the three tissues and between the two groups using analysis of variance (ANOVA) or two-sample t-test. 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. Results: Based on the logistic regression analysis results, we established a prediction model that integrated the APT SI and D values in the tumor tissue and the tumor size. ROC 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. Data Conclusion: A prediction model incorporating pretreatment APT SI, D in the tumor tissue and tumor size may be useful for predicting the response of intermediate-stage HCC to TACE. Level of Evidence: 1 Technical Efficacy: Stage 1

HCC advances in diagnosis and prognosis: Digital and Imaging

“…This promising model was shown to be associated with recurrence‐free survival 30 . Another avenue of research is evaluating advanced quantitative methods such as intravoxel incoherent motion MRI focusing on the metabolic function of tumoral tissue 31 . Finally, like in most medical fields, machine learning is now being evaluated in predicting the response to TACE.…”

“…30 Another avenue of research is evaluating advanced quantitative methods such as intravoxel incoherent motion MRI focusing on the metabolic function of tumoral tissue. 31 Finally, like in most medical fields, machine learning is now being evaluated in predicting the response to TACE. One published study attempted to evaluate the performance of a CT-based neural network with encouraging results.…”

The Role of Non-Gaussian Models of Diffusion Weighted MRI in Hepatocellular Carcinoma: A Systematic Review

“…Among the 33 studies, 11 investigated IVIM for HCC diagnosis, namely for HCC detection respect to normal liver parenchyma or other types of hepatic lesions (either benign or malignant), and seven evaluated the power of IVIM parameters for HCC histological grading. Among the 15 remaining studies, six assessed the usefulness of IVIM for the response of HCC to therapy [40,42,48,66,69,73], five evaluated if IVIM could be associated with prognostic factors [51,52,56,58,67], and four explored IVIM model for multiple aims [37,41,53,64] (two investigated response to therapy and survival [53,64], one on grading and prognostic factors [41], and the remaining one on diagnosis and grading [37]).…”

“…Among studies assessing the usefulness of IVIM the response of HCC to therapy, Jia et al [40] found that D could be helpful in predicting HCC response to transarterial chemoembolization (TACE) and was a significant predictor of response to therapy both univariately and in multivariate analysis, including also parameters derived from amide proton transfer. Park et al [73] also assessed the usefulness of IVIM in the response of TACE in HCC and found that D* was able to distinguish patients with good lipiodol uptake from those with a poor one.…”

A Brief History and Future Prospects of CEST MRI in Clinical Non-Brain Tumor Imaging

“…Therefore, a fast image readout was chosen, such as echo planar imaging (EPI), rapid imaging with refocused echoes (RARE), fast spin echo (FSE), and/or fast imaging with steady-state precession (FISP) [ 6 , 34 , 35 , 36 ]. As shown in Table 1 , out of the three papers on liver and the two on lung, four used FSE [ 74 , 75 , 81 , 82 ] and one used EPI [ 73 ]. Besides a fast readout, respiration-gated design [ 81 , 82 ] or breath holding [ 73 , 74 ] are still required to reduce motion during acquisition.…”