Following a stroke, each patient has a specific capacity to recover function that likely hinges on their ability to mount a neural repair response. Knowing this capacity early on would provide opportunities for personalized medicine in stroke rehabilitation. Physicians and therapists currently have few methods to predict recovery, and therefore have little data to guide decisions on treatment approach and length of stay. This has led to suboptimal therapy delivery. There is, however, emerging evidence to suggest that biomarkers can accurately predict recovery potential in order to guide therapy. Early attempts to improve motor recovery prediction used clinical, neurophysiologic (TMS), and MRI biomarkers to develop algorithms with up to 80% accuracy. While these are helpful, they tell us little about the biology of stroke recovery and are difficult to implement in most rehabilitation facilities. We recently identified a panel of 5 microRNAs (miRNAs) collected from plasma on rehab admission that can predict motor recovery with 95% accuracy. These 5 miRNAs converge on axonal guidance and developmental biology pathways. We therefore suspect that they capture the potential to mount a neural repair response after stroke. The overall goal of the proposal is to refine this miRNA predictive panel using a larger number of samples from 3 studies, test it?s predictive / analytic qualities, and validate the predictive panel in a large prospective study. R61 Phase Specific Aims:
Aim #1 a. Analyze stroke samples from 3 independent studies and matched controls using discovery-based qRT-PCR.
Aim #1 b. Use machine learning to develop a miRNA panel with ? 80% predictive ability across the 3 studies to identify patients who achieve excellent, moderate, and poor motor recovery by 3-12 months, controlling for covariates.
Aim #1 c. Develop custom qRT-PCR cards that include ~10 miRNAs in the refined panel in triplicate along with endogenous controls. Perform analytic validation to gauge precision, accuracy, and ideal handling / storage. R33 Phase Specific Aims:
Aim #2 a. Prospectively validate the miRNA panel by collecting samples from patients with upper limb impairment admitted to 2 inpatient rehab facilities, testing sensitivity, specificity, and positive/negative predictive value for classifying excellent, moderate, and poor motor recovery at 6 months.
Aim #2 b Give the predicted recovery potential score to the treatment team upon patient discharge from the inpatient rehabilitation facility. Use regression and chi-squared tests to assess how having this information on the day of admission would have changed percent of time delivering specific therapies and recommended length of stay. Tailoring rehabilitation treatment to the individual?s biological capacity to recover may lead to greater independence and lower costs. Specifically, evidence suggests that delivering more restorative-type therapy to the subset of patients most likely to respond improves motor outcomes. Just as important, the miRNAs identified in the panel will become treatment targets for recovery-based therapeutics.

Public Health Relevance

This project will use qRT-PCR to analyze plasma samples collected from stroke patients to develop a microRNA-based motor recovery potential score. This score will undergo clinical validation as a tool to tailor rehabilitation therapy delivery for individual stroke patients at inpatient rehabilitation facilities based on their predicted level of motor recovery. Once validated, the score will be used in future studies to streamline therapy delivery and improve outcomes.

National Institute of Health (NIH)
National Institute of Neurological Disorders and Stroke (NINDS)
Project #
Application #
Study Section
National Institute of Neurological Disorders and Stroke Initial Review Group (NSD)
Program Officer
Pelleymounter, Mary A
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Georgetown University
Schools of Medicine
United States
Zip Code