Yu Shi1, Fei Xia2, Qiu-Ju Li1, Jia-Hui Li1, Bing Yu1, Yue Li3, He An1, Kevin J Glaser4, Shengzhen Tao4, Richard L Ehman4, Qi-Yong Guo1. 1. Department of Radiology, Shengjing Hospital of China Medical University, Shenyang, China. 2. Department of Infectious Disease, Shengjing Hospital of China Medical University, Shenyang, China. 3. Department of Pathology, Shengjing Hospital of China Medical University, Shenyang, China. 4. Department of Radiology, Mayo Clinic, Rochester, Minnesota, USA.
Abstract
OBJECTIVES: Magnetic resonance elastography (MRE) with three-dimensional spin-echo echo planar imaging (3D-SE-EPI) is a newly emerging noninvasive method for assessing liver fibrosis. We hypothesized that 3D-SE-EPI might have better diagnostic accuracy than conventional two-dimensional gradient-recalled echo (2D-GRE). METHODS: We prospectively included 179 consecutive patients with chronic hepatitis B (CHB) or C (CHC) who underwent both MRE and liver biopsy. Liver stiffness was measured by both 3D-SE-EPI and 2D-GRE for staging biopsy-proven liver fibrosis (using METAVIR scores). A receiver-operating characteristic analysis using the area under the receiver-operating characteristic curve (AUC) was used to compare the diagnostic performance in predicting liver fibrosis between these two techniques, and compared them to serum markers of fibrosis. RESULTS: The technical failure rate of 3D-SE-EPI (2.2%, n=4/179) was lower compared with 2D-GRE (8.3%, n=15/179). The stiffness measured by 3D-SE-EPI was slightly lower compared with 2D-GRE, with the mean difference of 0.57 kPa (Bland and Altman plot, 95% limits of agreement: -0.32 and 1.45 kPa). AUCs for the characterization of ≥F1, ≥F2, ≥F3, and F4 were 0.957 (95% confidence interval (CI): 0.913-0.983), 0.971 (0.932-0.991), 0.991 (0.961-0.999), and 0.979 (0.942-0.995) for 3D-SE-EPI, which was slightly higher compared with the AUCs for 2D-GRE at each fibrosis stage (0.948 (0.901-0.977), 0.959 (0.915-0.981), 0.979 (0.943-0.995), and 0.976 (0.938-0.994), respectively), although none reached statistical significance (P=0.160-0.585). In an "intention-to-diagnose" analysis, the diagnostic accuracy (the proportion of well-classified patients) by EPI (86.7-91.3%, n=169) was higher compared with GRE (80.9-82.1%, n=158) after applying optimal cutoffs. Both 3D-SE-EPI and 2D-GRE performed better than serum fibrosis markers. CONCLUSIONS: With respect to 2D-GRE, 3D-SE-EPI has the advantage of lower failure rate with equivalent high diagnostic performance for staging liver fibrosis in CHB/CHC patients, and thus more helpful for those challenging cases in 2D-GRE.
OBJECTIVES: Magnetic resonance elastography (MRE) with three-dimensional spin-echo echo planar imaging (3D-SE-EPI) is a newly emerging noninvasive method for assessing liver fibrosis. We hypothesized that 3D-SE-EPI might have better diagnostic accuracy than conventional two-dimensional gradient-recalled echo (2D-GRE). METHODS: We prospectively included 179 consecutive patients with chronic hepatitis B (CHB) or C (CHC) who underwent both MRE and liver biopsy. Liver stiffness was measured by both 3D-SE-EPI and 2D-GRE for staging biopsy-proven liver fibrosis (using METAVIR scores). A receiver-operating characteristic analysis using the area under the receiver-operating characteristic curve (AUC) was used to compare the diagnostic performance in predicting liver fibrosis between these two techniques, and compared them to serum markers of fibrosis. RESULTS: The technical failure rate of 3D-SE-EPI (2.2%, n=4/179) was lower compared with 2D-GRE (8.3%, n=15/179). The stiffness measured by 3D-SE-EPI was slightly lower compared with 2D-GRE, with the mean difference of 0.57 kPa (Bland and Altman plot, 95% limits of agreement: -0.32 and 1.45 kPa). AUCs for the characterization of ≥F1, ≥F2, ≥F3, and F4 were 0.957 (95% confidence interval (CI): 0.913-0.983), 0.971 (0.932-0.991), 0.991 (0.961-0.999), and 0.979 (0.942-0.995) for 3D-SE-EPI, which was slightly higher compared with the AUCs for 2D-GRE at each fibrosis stage (0.948 (0.901-0.977), 0.959 (0.915-0.981), 0.979 (0.943-0.995), and 0.976 (0.938-0.994), respectively), although none reached statistical significance (P=0.160-0.585). In an "intention-to-diagnose" analysis, the diagnostic accuracy (the proportion of well-classified patients) by EPI (86.7-91.3%, n=169) was higher compared with GRE (80.9-82.1%, n=158) after applying optimal cutoffs. Both 3D-SE-EPI and 2D-GRE performed better than serum fibrosis markers. CONCLUSIONS: With respect to 2D-GRE, 3D-SE-EPI has the advantage of lower failure rate with equivalent high diagnostic performance for staging liver fibrosis in CHB/CHCpatients, and thus more helpful for those challenging cases in 2D-GRE.
Authors: Ryan B Perumpail; Josh Levitsky; Yi Wang; Victoria S Lee; Jennifer Karp; Ning Jin; Guang-Yu Yang; Bradley D Bolster; Saurabh Shah; Sven Zuehlsdorff; Albert A Nemcek; Andrew C Larson; Frank H Miller; Reed A Omary Journal: Acad Radiol Date: 2012-09 Impact factor: 3.173
Authors: Meng Yin; Kevin J Glaser; Jayant A Talwalkar; Jun Chen; Armando Manduca; Richard L Ehman Journal: Radiology Date: 2015-07-08 Impact factor: 11.105
Authors: Paul Kennedy; Mathilde Wagner; Laurent Castéra; Cheng William Hong; Curtis L Johnson; Claude B Sirlin; Bachir Taouli Journal: Radiology Date: 2018-03 Impact factor: 11.105