Values and Count Optimization of IVIM diffusion Weighting Parameter

IntraVoxel Incoherent Motion (IVIM) based Magnetic Resonance Imaging (MRI) has gained popularity in the field of medical diagnosis in recent times. The IVIM principle distinguishes the effect of perfusion and diffusion in biological tissues in a non-invasive manner, without injecting any contrast agents to the body tissues. The pathological conditions can be quantitatively determined by IVIM parameters and for the proper estimation of these parameters, the MRI signal must be acquired with several diffusion weightings, b. The b value distributions used so far are heuristic and show broad differences between researchers. In this analysis, the effect of b values on the acquisition of IVIM data is evaluated and the optimal distribution of b values is determined using a genetic algorithmic method. Our result generates 10 b value distributions as {0, 10, 10, 20, 30, 70, 80, 110, 120 and 500} and this distribution has worked well and can greatly decrease the overall measurement error. The overall parameter efficiency relies heavily on the Signal to Noise(SNR): for a lower SNR value, the error on parameter estimation is far higher than the error for a higher SNR value.