Nature Communications 2021 Longitudinal Response Modeling

3D RP-Net

A longitudinal multi-task MRI model for treatment response prediction in rectal cancer.

3D RP-Net showcase image

Overview

3D RP-Net models treatment response from paired pre-treatment and post-treatment MRI rather than relying on a single static snapshot. By learning segmentation and response prediction together, it uses temporal imaging change as a signal for therapeutic assessment in rectal cancer.

This project represents the longitudinal branch of the lab’s imaging work: systems that not only see anatomy, but also track how tumors evolve during treatment.