IVIM

IVIM, or intravoxel incoherent motion, is a diffusion MRI model that separates the signal attenuation caused by molecular diffusion from that caused by microvascular perfusion within a voxel. Introduced by Le Bihan and colleagues in the late 1980s, IVIM enables estimation of diffusion and perfusion-related parameters without the use of contrast agents.

The IVIM model describes the normalized diffusion-weighted signal S(b)/S0 as S(b)/S0 = (1 - f) exp(-b D) + f exp(-b D*), where b is the diffusion weighting, D is the true tissue diffusion coefficient, D* is the pseudodiffusion coefficient related to capillary blood flow, and f is the perfusion fraction representing the fraction of signal from the fast, perfusion-related compartment. In general, D is on the order of 10^-3 mm^2/s, D* is larger (often 10^-3 to 10^-2 mm^2/s), and f ranges from a few percent to a few tens of percent, depending on tissue and pathology.

Acquisition relies on diffusion-weighted imaging with multiple b-values, including low values to capture perfusion effects and higher values to quantify diffusion. Parameter estimation can be performed with a segmented approach—estimate D from high b-values, then fit f and D*—or via full nonlinear fitting, often with constraints to improve stability. Robust fitting requires adequate signal-to-noise ratio and careful selection of b-values.

Applications of IVIM span liver fibrosis assessment, tumor characterization and perfusion estimation, renal and brain tissue evaluation, and stroke research. Limitations include sensitivity to motion and noise, partial volume effects, and dependence on fitting strategies and acquisition design, which can affect reproducibility across scanners and protocols.