A non-parametric statistical framework is outlined to display the result of brain MRE exams in terms of an effect size. This approach is built upon a neural network-based MRE inversion that estimates the posterior cumulative distribution function at each voxel, and accounts for age, sex, spatial variation in stiffness, and uncertainty in property estimation. Using these effect size maps but no summary statistics, neuoradiologists were able to diagnose normal pressure hydrocephalus with 70% sensitivity and 100% positive predictive value.
This abstract and the presentation materials are available to members only; a login is required.