Dependence of intravoxel incoherent motion diffusion MR threshold b-value selection for separating perfusion and diffusion compartments and liver fibrosis diagnostic performance

Background Intravoxel incoherent motion (IVIM) tissue parameters depend on the threshold b-value. Purpose To explore how threshold b-value impacts PF (f), D-slow (D), and D-fast (D*) values and their performance for liver fibrosis detection. Material and Methods Fifteen healthy volunteers and 33 hepatitis B patients were included. With a 1.5-T magnetic resonance (MR) scanner and respiration gating, IVIM data were acquired with ten b-values of 10, 20, 40, 60, 80, 100, 150, 200, 400, and 800 s/mm(2). Signal measurement was performed on the right liver. Segmented-unconstrained analysis was used to compute IVIM parameters and six threshold b-values in the range of 40-200 s/mm(2) were compared. PF, D-slow, and D-fast values were placed along the x-axis, y-axis, and z-axis, and a plane was defined to separate volunteers from patients. Results Higher threshold b-values were associated with higher PF measurement; while lower threshold b-values led to higher D-slow and D-fast measurements. The dependence of PF, D-slow, and D-fast on threshold b-value differed between healthy livers and fibrotic livers; with the healthy livers showing a higher dependence. Threshold b-value = 60 s/mm(2) showed the largest mean distance between healthy liver datapoints vs. fibrotic liver datapoints, and a classification and regression tree showed that a combination of PF (PF < 9.5%), D-slow (D-slow < 1.239 x 10(-3) mm(2)/s), and D-fast (D-fast < 20.85 x 10(-3) mm(2)/s) differentiated healthy individuals and all individual fibrotic livers with an area under the curve of logistic regression (AUC) of 1. Conclusion For segmented-unconstrained analysis, the selection of threshold b-value = 60 s/mm(2) improves IVIM differentiation between healthy livers and fibrotic livers.