Poly-omic predictors of symptom duration and recovery for adolescent concussion - PROJECT SUMMARY There are nearly three million mild traumatic brain injuries (mTBIs) in the U.S. each year, and most occur in patients less than 21 years of age. Clinical assessment of mTBI relies on symptom surveys that cannot accurately predict the duration of symptoms or objectively identify brain recovery. A biologic test would allow physicians to provide individualized recommendations for school and athletics participation, prescribe timely pharmacologic treatments, or initiate early psychosocial services in patients at risk for persistent post- concussion symptoms (PPCS). Non-coding ribonucleic acids (ncRNAs), such as microRNAs, are epi- transcriptional molecules that are altered in patients with mTBI. They can be measured in peripheral biofluids such as serum, or even saliva. Our previous research demonstrates that ncRNA changes in cerebrospinal fluid are reflected in saliva, and that saliva ncRNA levels can predict PPCS. Validation of these findings in a large, independent cohort could yield a biologic measure of PPCS risk (Aim 1), and guide individualized clinical management decisions (Aim 2). This scientific premise forms the basis for our proposed multi-center study. We will enroll 750 adolescents (ages 13-18 years) with mTBI, defined by the World Health Organization and Berlin Consensus Criteria. We will measure levels of saliva ncRNAs enriched in neuronal and glial exosomes at acute (<48 hours), sub-acute (7 days), and chronic (30 days) post-injury time points. PPCS will be defined by persistence of ≥ 3 symptoms on day 30 (compared with pre-injury state, determined by the Post-Concussion Symptom Inventory; PCSI). In 250 participants (training set), we will use a LASSO technique to refine a multivariate model, that employs acute and sub-acute ncRNA levels, along with clinical, social, and psychologic factors, to predict PPCS (while controlling for biologic covariates). Accuracy of the model will be externally validated in the remaining 500 participants (test set). Sensitivity and specificity will be compared to the validated “5P” clinical prediction tool. We will also examine the relationships between concussive symptom phenotypes and ncRNA levels with a factor analysis and hierarchical clustering. In Aim 2, we will use LASSO in a training set (n=250) to refine a second multivariate model, that uses acute and chronic ncRNA levels, along with clinical, social, and psychologic factors to identify concussion recovery. Recovery will be defined by self-report of “no difference from pre-injury” on the PCSI. Accuracy of the model will be externally validated in the test set, and compared to the accuracy of reaction time performance across acute and chronic time points. Our multi-disciplinary team includes experts in pediatrics, neurology, molecular biology, psychology, and emergency medicine with a published track record of collaboration and the expertise necessary for this proposal’s success. The study will yield an objective measure of PPCS risk, concussion phenotype, and clinical recovery. When paired with medical, social, and psychologic assessments, this technology will allow researchers to study mTBI therapies in biologically-defined patient subsets and personalize concussion care.
