A Magnetic Resonance Imaging-Based Classifier to Accurately Diagnose Persistent Post-Traumatic Headache and to Differentiate It from Chronic Migraine

Project: Research project

Project Details

Description

Topic Area: Chronic migraine and post-traumatic headaches.


Importance of Problem Addressed by This Proposed Research: Traumatic brain injury (TBI) is one of the signature injuries sustained by U.S. Soldiers during modern wars. Twenty percent of U.S. Soldiers who return from recent wars have sustained TBI. In addition, TBI is a substantial public health problem among civilians, with 1.7 million Americans seeking medical attention annually for TBI and 1.6-3.8 million sustaining sports-related mild traumatic brain injuries (mTBI) each year. mTBI accounts for 80% of all TBIs in the military and civilian populations. Post-traumatic headache (PTH) is the most common symptom following mTBI, afflicting up to 90% of people with mTBI. A substantial proportion of PTH persists beyond the first 3 months following the TBI, thus becoming "persistent post-traumatic headache" (PPTH), a chronic and disabling problem for the Soldier or civilian with mTBI. PTH is a frequent indication for medical evacuation of Soldiers from recent wars, and only 18% of these Soldiers return to duty. Among the civilian population, up to 60% of patients with PPTH require long-term disability benefits, quit work, are dismissed from work, or require reduced work hours.


Despite the high prevalence of PPTH and the substantial burden it places on military and civilian U.S. populations, the ability to accurately diagnose PPTH is often fraught with difficulty. According to current diagnostic criteria for PPTH (International Classification of Headache Disorders, 3rd edition), the only feature that differentiates PPTH from other headache types is the interval between sustaining a TBI and onset of the headaches. Diagnostic criteria stipulate that headaches must begin within 7 days of TBI. There are no other known features that differentiate PPTH from other headache types. These diagnostic criteria often result in substantial diagnostic difficulty since: (1) the headache and associated symptoms of PPTH are commonly indistinguishable from chronic migraine (CM); (2) a substantial proportion of presumed PTHs start after 7 days from TBI; (3) migraine is common (1-year prevalence of 12% in the general population; five times more common among U.S. Soldiers), often present in Soldiers and civilians who subsequently sustain a TBI. Thus, although it is straightforward to diagnose PPTH attributed to mTBI in a patient who never had headaches but then reliably reports having developed headaches within 7 days of a TBI, the typical clinical presentation is much more complex. The inability to accurately diagnose PPTH attributed to mTBI and to differentiate it from CM is a substantial limitation for patient care and for conducting research on PPTH that will lead to a better description of PPTH mechanisms and to better PPTH treatments.


How the Proposed Research Addresses This Critical Problem: For this investigation, we will use state-of-the-art mathematical methods based on machine learning algorithms to build automated diagnostic models that accurately diagnose PPTH attributed to mTBI and differentiate it from CM. The diagnostic model will be constructed using the optimal combination of patient symptoms with brain structure and function collected via advanced magnetic resonance imaging (MRI) techniques. Although there are usually very few to no differences between the symptoms of PPTH and CM according to typically collected clinical information, in this study we will collect very detailed clinical symptom data that extend beyond the usually collected information (e.g., prominent symptom exacerbation with exertion; symptoms of orthostatic intolerance). Similarly, even though CM and PPTH due to mTBI cannot typically be differentiated by results from clinical MRI, this investigation will use advanced methods of studying brain structure and function (e.g., cortical thickness measures, regional brain volumes, structural connectivity, functional connectivity) that we believe will differentiate PPTH from CM. This study is designed with the specific intent of using MRI sequences that could easily be used by scientists and clinicians at other centers, that do not require contrast administration, do not add substantial costs to clinical MRI scans, do not take long to acquire, and do not require patients to perform tasks during collection of MRI data.


Impact/Applicability of Research: This study would build automated data models that will accurately diagnose PPTH attributed to mTBI and differentiate it from CM. These models will ultimately be disseminated to other scientists and clinicians for use as computer-aided diagnostic tools. The accurate identification of PPTH will enhance the ability to manage PPTH patients and to conduct PPTH research.

StatusActive
Effective start/end date1/1/14 → …

Funding

  • Congressionally Directed Medical Research Programs: $1,570,257.00

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.