MRI-based Post-Concussive Syndrome (PCS) Index: No Differences between Athletes and Non-Athletes

Abstract Purpose: To evaluate for differences in patients with post-concussive syndrome from sports-related versus non-sports related mTBI using a machine-learning algorithm of MRI data. Methods: For this retrospective study, we explored the MRI records of 28 Athletes (17 M, 11F) and 16 non-Athletes (6 M, 10F) with a history of concussion and clinical PCS, and 73 no concussed reference controls (26 M,47F). All subjects were between 19 and 35-years in age. The MRI studies were done in a clinical 3 T MRI scanner. MPRAGE, DTI-FA and DTI-ADC images were used to extract radiomics features from automatically segmented MRI structures of the brain. After that, the radiomic features were processed to extract the MRI-PCS Index (Qmetrics Technologies, Rochester, NY). MRI-PCS Index is an Artificial-Intelligence (AI) derived holistic evaluation of brain health that ranges from 0.0 to 1.0. The index indicates the degree of signal/structural abnormalities found in the brain tissue of a concussed subject when compared to a non-concussed subject; hence it provides an objective measurement of PCS. Results: The sports played by the athletes were mainly football, hockey, lacrosse, and soccer while non-Athletes were mainly concussed in motor vehicle accidents. Athletes MRI-PCS Index was similar to non-Athletes (0.78 vs 0.78, p = 0.89) and both indexes were very different to the reference control (MRI-PCS Index = 0.23, IQR = 0.10–0.38, p < 0.001). Conclusions: The MRI-PCS Index, an objective measurement of structural brain health, indicated that subjects that suffer from PCS present similar MRI abnormality burden regardless of whether the mTBI was related to sports injuries or other trauma.