Abstract: Multiple sclerosis (MS) is a disease primarily characterized by lesions on MRI images. Detection of changes in MS lesions is essential for accurate characterization of the disease progression and assessment of the treatment response. We propose a new lesion change detection method that employs multi-scale radiomic features of longitudinal MR scans. On a dataset of 15 patients, the proposed method has statistically significantly higher performance than state-of-the-art methods.

Learning Objective 1: Understand the research problem of multiple sclerosis lesion change detection and learn the benefit of using the proposed method compared to existing strategies.

Learning Objective 2 (Optional): Learn about the challenges of lesion change detection, steps taken to solve these challenges by applying techniques used in other imaging problems, and its future directions and clinical applications.


Myra Cheng, Stanford University
Alfiia Galimzianova, Stanford University
Ziga Lesjak, University of Ljubljana
Ziga Spiclin, University of Ljubljana
Christopher Lock, Stanford University
Daniel Rubin, Stanford University
Imon Banerjee (Presenter)
Stanford University

Presentation Materials: