Differentiation of focal indeterminate marrow abnormalities with multiparametric MRI

Abstract

Purpose: To explore magnetic resonance imaging (MRI) parameters from intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI), multiecho Dixon imaging (ME-Dixon), and dynamic contrast-enhanced imaging (DCE) for differentiating focal indeterminate marrow abnormalities. Materials and Methods: Forty-two patients with 14 benign and 28 malignant focal marrow abnormalities were included. The following were independently analyzed by two readers: signal intensity (SI), contour, and margin on conventional MR images; SI on b-800 images (SIb-800), apparent diffusion coefficient (ADC), IVIM parameters (Dslow, Dfast, and f), fat fraction (Ff), and DCE parameters (time-to-signal intensity curve pattern, iAUC, Ktrans, kep, and ve). The MR characteristics and parameters from benign and malignant lesions were compared with a chi-squared test and the Mann–Whitney U-test, respectively. The area under receiver operating characteristic (ROC) curves (AUC) of each sequence were also compared. Interobserver agreements were assessed with Cohen's κ, and intraclass correlation coefficient (ICC). Results: ADC, Dslow, and Ff demonstrated a significant difference between benign and malignant marrow abnormalities for both readers (P < 0.001). SIb-800 and perfusion-related parameters from IVIM-DWI and DCE were not significantly different between the two groups (P = 0.145, 0.439, and 0.337 for reader 1, P = 0.378, 0.368, and 0.343 for reader 2, respectively). The AUCs of ADC, Dslow, and Ff were significantly higher for differentiating indeterminate marrow abnormalities in both readers (P < 0.001). Interobserver agreements were substantial in SIb-800, and ICCs were almost perfect for ADC, Dslow, f, and Ff, and substantial for iAUC, kep, Ktrans, ve, and Dfast. Conclusion: ADC, Dslow, and Ff may provide information for differentiating focal indeterminate abnormalities. Level of Evidence: 3. Technical Efficacy: Stage 2. J. MAGN. RESON. IMAGING 2017;46:49–60.