Anatomical Predictions using Subject-Specific Medical Data

Marianne Rakic, John Guttag, Adrian V. Dalca

Keywords: medical imaging, computer vision, prediction, registration, clinical, neural networks

Track: short paper

Paper Type: both

Abstract: Changes in brain anatomy can provide important insight for treatment design or scientific analyses. We present a method that predicts how brain anatomy for an individual will change over time. We model these changes through a diffeomorphic deformation field, and design a predictive function using convolutional neural networks. Given a predicted deformation field, a baseline scan can be warped to give a prediction of the brain scan at a future time. We demonstrate the method using the ADNI cohort, and analyze how performance is affected by model variants and the type of subject-specific information provided. We show that the model provides good predictions and that external clinical data can improve predictions.

TL;DR: We present a learning-based model based on diffeomorphic registration to predict the brain anatomy in patient follow-up MRI scan given just a baseline scan using population trends and subject-specific genetic and clinical information.

Presentation Upload Agreement: I agree that my presentation material (videos and slides) will be made public.