Multimaterial Decomposition Algorithm for the Quantification of Liver Fat Content by Using Fast-Kilovolt-Peak Switching Dual-Energy CT: Clinical Evaluation

Purpose: To assess the clinical accuracy and reproducibility of liver fat quantification with the multimaterial decomposition (MMD) algorithm, comparing the performance of MMD with that of magnetic resonance (MR) spectroscopy by using liver biopsy as the reference standard. Materials and Methods: This prospective study was approved by the institutional ethics committee, and patients provided written informed consent. Thirty-three patients suspected of having hepatic steatosis underwent non-contrast material-enhanced and triple-phase dynamic contrast-enhanced dual-energy computed tomography (CT) (80 and 140 kVp) and single-voxel proton MR spectroscopy within 30 days before liver biopsy. Percentage fat volume fraction (FVF) images were generated by using the MMD algorithm on dual-energy CT data to measure hepatic fat content. FVFs determined by using dual-energy CT and percentage fat fractions (FFs) determined by using MR spectroscopy were compared with histologic steatosis grade (0-3, as defined by the nonalcoholic fatty liver disease activity score system) by using Jonck-heere-Terpstra trend tests and were compared with each other by using Bland-Altman analysis. Real non-contrast-enhanced FVFs were compared with triple-phase contrast-enhanced FVFs to determine the reproducibility of MMD by using Bland-Altman analyses. Results: Both dual-energy CT FVF and MR spectroscopy FF increased with increasing histologic steatosis grade (trend test, P <.001 for each). The Bland-Altman plot of dual-energy CT FVF and MR spectroscopy FF revealed a proportional bias, as indicated by the significant positive slope of the line regressing the difference on the average (P < .001). The 95% limits of agreement for the differences between real non-contrast- enhanced and contrast-enhanced FVFs were not greater than about 2%. Conclusion: The MMD algorithm quantifying hepatic fat in dual-energy CT images is accurate and reproducible across imaging phases. (C) RSNA, 2017