|
Activity Number:
|
514
- Recent Advances in Imaging Statistics: Bayesian Methods and Beyond
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 2, 2017 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Imaging
|
|
Abstract #324245
|
View Presentation
|
|
Title:
|
Experimental Design in Longitudinal MRI for Multiple Sclerosis
|
|
Author(s):
|
Menghan Hu* and Ani Eloyan and Matthew Schindler and Daniel Reich and Russell Taki Shionhara and Blake Dewey
|
|
Companies:
|
Brown University and Brown University and NIH, NINDS and NINDS and UPenn and Johns Hopkins Whiting School of Engineering
|
|
Keywords:
|
Imaging statistics ;
Multiple Sclerosis ;
longitudinal ;
sampling ;
structural MRI
|
|
Abstract:
|
Multiple sclerosis (MS) is a progressive neurological disease characterized by brain and spinal cord lesions. MRI is highly sensitive to detect inflammatory lesions, making it an essential clinical tool to monitor disease activity longitudinally. While MRI provides valuable data to identify disease biomarkers in MS, it is costly and time consuming. In this talk, I will describe data-driven sampling strategies for longitudinal MRI studies in MS that identify the minimum number of follow-up scans required for identifying a voxel-level biomarker of MRI intensity changes in incident MS lesions related to clinical information and therapeutic interventions. We propose models for biomarker identification that incorporate data with different number of followup studies. For each design strategy, principal component analysis (using intensity profiles from MRI data of four modalities) will be used to develop a voxel-level biomarker for determining the longitudinal change within the lesion voxel. Result shows that using four scans can lead to identification of a biomarker resulting in estimation of significant associations between the biomarker and disease modifying therapeutic interventions.
|
Authors who are presenting talks have a * after their name.
