Abstract Details

Title A Data Driven Approach to Define Disease Course Categories in Multiple Sclerosis

Topic Multiple Sclerosis

Presentation(s) MS and CNS Inflammatory Disease Posters (7:00 AM-5:00 PM)

Poster/Presentation
Number 244

Objective To create categories of multiple sclerosis (MS) disease course using data from clinical performance and patient reported outcome measures followed by validation of categories using quantitative MRI.

Background

The contemporary classification of disease course based on clinical descriptions of relapses and disability accrual does not encompass quantitative determination of performance or patient-reported symptoms and disease status.

Design/Methods Data from 55,606 visits (16,399 unique patients) in the MS PATHS cohort were extracted. 20 features were used to create disease categories including: demographics, disease duration, Patient Determined Disease Steps (PDDS), manual dexterity test (MDT), walking speed test (WST), processing speed test (PST), and Neuro-QOL. Machine learning k-means clustering method was applied. Based on the sum of squared distances plot, 4 disease categories were derived. Validation was conducted using a machine learning model to predict categories included brain parenchymal fraction, T2 lesion volumes, gray/white matter fraction, thalamic volume, and cortical/deep gray matter volume.

Results The 4 disease categories were ranked by PDDS: Category A: 0.4 (SD 0.9), Category B: 0.8 (SD 1.0) Category C: 3.3 (SD 1.9), Category 4: 4.0 (SD 2.1). Stages A/B were predominately relapsing (91.2%/83.8%), while stages C/D were predominantly progressive (53.6%/55.9%). Relapses were reported in all categories including C/D (46%/27.7%). Category C appeared to have accumulated disability faster than D (disease duration of 15.3 [SD 10.4] vs. 23.2 [SD 12.4] years), and despite only 4 more years of disease duration than A had >3 points higher PDDS and 16 times more disabled patients (2.8% versus 46.3%). MDT, PST, WST, and MRI measures were worse across the categories as expected. Validation demonstrated that MRI measures predicted category with an accuracy of 0.802.

Conclusions

Data driven disease categories based on clinical/patient reported measures reflect biological disease burden on MRI. In data driven classification, in contrast to conventional disease courses, progression/relapses occurred across all disease categories.

Authors/Disclosures
Daniel Ontaneda, MD, PhD, FAAN (Cleveland Clinic) PRESENTER	Dr. Ontaneda has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Novartis. Dr. Ontaneda has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Genentech/Roche. Dr. Ontaneda has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Biogen Idec. Dr. Ontaneda has received personal compensation in the range of $500-$4,999 for serving as a Consultant for BMS. Dr. Ontaneda has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Sanofi. The institution of Dr. Ontaneda has received research support from NIH. The institution of Dr. Ontaneda has received research support from PCORI. The institution of Dr. Ontaneda has received research support from NMSS. The institution of Dr. Ontaneda has received research support from Genetech.
Jingan Qu	No disclosure on file
Marisa P. McGinley, DO (Cleveland Clinic)	Dr. McGinley has received personal compensation in the range of $5,000-$9,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Genentech. Dr. McGinley has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for EMD Serono. The institution of Dr. McGinley has received research support from Biogen. The institution of Dr. McGinley has received research support from Genentech. The institution of Dr. McGinley has received research support from NIH. The institution of Dr. McGinley has received research support from AHRQ. The institution of Dr. McGinley has received research support from EMD Serono.
	No disclosure on file
	No disclosure on file
David D. Schindler, BS (Motion Med LLC)	No disclosure on file
Robert A. Bermel, MD, FAAN (Cleveland Clinic)	Dr. Bermel has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Sanofi/Genzyme. Dr. Bermel has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Genentech/Roche. Dr. Bermel has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Novartis. Dr. Bermel has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for TG Therapeutics. The institution of Dr. Bermel has received research support from Biogen. The institution of Dr. Bermel has received research support from Roche. The institution of Dr. Bermel has received research support from Novartis. Dr. Bermel has received intellectual property interests from a discovery or technology relating to health care.
Jeffrey A. Cohen, MD (Cleveland Clinic)	Dr. Cohen has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Convelo. Dr. Cohen has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Astoria. Dr. Cohen has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Bristol Myers Squibb. Dr. Cohen has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Viatris. Dr. Cohen has received personal compensation in the range of $500-$4,999 for serving as a Consultant for PSI. Dr. Cohen has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Shionogi. Dr. Cohen has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Celltrion.
	No disclosure on file

��ɫ��

��ɫ��