Advancing Artificial Intelligence (AI) Toward Precision Medicine in Traumatic Brain Injury: A Collaboration by DHA, DOE, TRACK-TBI, and the CARE Consortium
Abstract
Background and Reasoning for Proposed Project: Traumatic brain injury (TBI) is a leading cause of death and disability in the U.S. and globally. Roughly 150 Americans die from TBI-related injuries each day. TBI often leads to significant serious impairments in thinking, behavior, and life function. While severe forms of TBI are associated with a higher risk of death and disability, mild TBI (mTBI, or concussion) accounts for the overwhelming majority (>80%) of all TBIs. mTBI affects an estimated 3.8 million individuals in the U.S. and 55.9 million globally each year, with particularly heightened risk among military Service Members and contact sport (e.g., football, etc.) athletes. According to the Department of Defense (DoD), from 2000 through 2020, the burden of military TBI worldwide was over 430,000 cases, with more than 80% classified as non-penetrating mTBI. Despite the label of “mild” brain injury, the changes in brain structure and function caused from mTBI can lead to significant disability. In fact, evidence suggests that up to 20% of individuals with mTBI experience persistent symptoms that negatively impact their academic, vocational, and social function years after injury. Problem to Be Addressed: Our ability to accurately predict which patients will do well after TBI versus which are at risk of poorer recovery is limited. We are also as yet unable to determine which treatments may be best suited for an individual patient. There is pressing need to leverage modern advances in data science (i.e., artificial intelligence (AI)) to analyze complex data from large-scale TBI studies. We propose continuation of a unique public-private collaboration that capitalizes on the rich data and expertise of clinician-scientists from the Transforming Research and Clinical Knowledge in Traumatic Brain Injury (TRACK-TBI) study and the Concussion Assessment, Research and Education (CARE) Consortium and data scientists from the U.S. Department of Energy (DoE). Together, TRACK-TBI and CARE represent the largest prospective precision medicine studies of TBI in civilians, military personnel, and athletes. This project will leverage the TRACK-TBI and CARE datasets and computational innovations from the DoE National Labs to improve care and outcomes for individuals affected by TBI by achieving the following aims: Aim 1: Curate, harmonize and cross-map common data elements (CDEs) and unique data elements (UDEs) from the CARE Consortium (sport, military TBI) and TRACK-TBI (civilian TBI) datasets into a harmonized and interrogatable dataset to enable advanced analytics of injury, recovery, and outcome across TBI cohorts. Aim 2: Apply advanced data analytical techniques (e.g., machine learning, AI) to interrogate and compare acute injury characteristics (including subject phenotypes), outcomes, and recovery in civilians, military service academy cadets/MIDN, and athletes affected by mTBI. Aim 3: Develop and apply data-driven, machine learning methods to better predict heterogeneous outcomes after mTBI, including identifying factors associated with the quality of recovery. The fitted models will aid in stratifying individuals with mTBI who would most benefit from emerging treatment trials. Individuals to Be Helped: Military Service Members who have sustained TBI. Knowledge gained will more broadly impact clinical practice guidelines for TBI across military, civilian, and sport medicine systems of care for TBI, with findings of interest to military and civilian clinicians and the lay public. Potential Clinical Applications, Benefits and Risks: TRACK-TBI and CARE have yielded evidence that informs our understanding of recovery after TBI/concussion and factors associated with outcome and risk after injury. The current project will have direct translational impact on current practices in the diagnosis, prognosis, and management (e.g., return to activity) of TBI in military, sport, and civilian medicine. This work will
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Dec 05, 2021
- Source ID
- W81XWH2110890
Entities
People
- Michael McCrea
Organizations
- Medical College of Wisconsin
- United States Army