Comparison of the diagnostic performance of 2D and 3D MR elastography in staging liver fibrosis

Objectives To compare the diagnostic performance and image quality of state-of-the-art 2D MR elastography (MRE) and 3D MRE in the basic application of liver fibrosis staging. Methods This retrospective study assessed data from 293 patients who underwent 2D and 3D MRE examinations. MRE image quality was assessed with a qualitative 2-point grading system by evaluating artifacts. Two experienced analysts independently measured mean liver stiffness values. The interobserver agreement of liver stiffness measurement was assessed by the intraclass correlation coefficient (ICC). The area under the receiver operating characteristic curve (AUC) was used to assess the diagnostic performance of 2D and 3D MRE and blood-based markers for fibrosis staging using the pathology-proven liver fibrosis stage as the gold standard. Results The image quality provided by 3D MRE was graded as significantly higher than that obtained with the 2D MRE method (p < 0.01). Interobserver agreement in liver stiffness measurements was higher for 3D MRE (ICC: 3D 0.979 vs 2D 0.955). The AUC values for discriminating >= F1, >= F2, >= F3, and F4 fibrosis for 3D MRE (0.89, 0.92, 0.95, and 0.93) were similar to those for 2D MRE (0.89, 0.91, 0.94, and 0.92). Both the 2D and 3D MRE methods provided superior accuracy to the blood-based biomarkers, including APRI, FIB-4, and Forns index, especially for >= F2, >= F3, and F4 fibrosis stages (all p < 0.01). Conclusions While 3D MRE offers certain advantages and opportunities for new applications of MRE, current widely deployed 2D MRE technology has comparable performance in the basic application of detecting and staging liver fibrosis.

Automated summary

3D MRE has higher image quality and interobserver agreement compared to 2D MRE, but similar diagnostic performance in detecting and staging liver fibrosis. Both 2D and 3D MRE methods outperform blood-based biomarkers in accuracy for >= F2, >= F3, and F4 fibrosis stages.