Chronic Neurological and Psychiatric Outcome in TBI Patients Assessed by A classification System Integrating Symptoms and Radiological Evidence

BackgroundTraumatic brain injury (TBI) can lead to long-term cognitive and psychiatric consequences such as intellectual disability, depression, and anxiety. The mental outcomes result from a complex interplay between neurobiological, structural, and clinical factors that are unlikely to be fully reflected by a single measure focused on only one aspect. Profiling mental outcomes is essential for identifying TBI patients at risk of cognitive-psychiatric complications, ideally through a tailored assessment system that integrates both subjective and objective indicators, yet such a system is currently absent in clinical settings. Mental Disability Classification (MDC) is a TBI-specific measurement system for mental outcome, widely practiced by forensic psychiatrists in China, that incorporates both symptomatic and radiological indicators of mental impairment. Despite its advantage in integrating multiple types of evidence, its clinical application remains limited due to reliance on expert judgment and the absence of comprehensive evaluation. We aimed to establish and improve the MDCs utility in clinical research by constructing prediction models, and profile the mental prognosis landscape of TBI patients through systematically evaluating the MDC system. MethodsWe performed a multicenter and longitudinal study of 499 TBI patients from four cohorts in China. SYSU-TBI1 cohort consisted of 356 TBI patients recruited from 2017 to 2019, from whom we collected 189 features including both acute and chronic phase variables concerning neurological and psychiatric symptoms, radiological characteristics, as well as demographic variables. SYSU-TBI1 cohort was used to construct prediction models for MDC using random forest and Least Absolute Shrinkage and Selection Operator (LASSO) methods, through which MDCs key chronic dimensional features were also identified. SYSU-TBI1 cohort was further used to identify acute phase features associated with MDC and MDCs key chronic dimensional features by regression models. SYSU-TBI2 cohort consisted of 84 TBI patients recruited from 2021 to 2023 and was used to validate the accuracy of MDC prediction model, alongside two external cohorts, ov cohort (N=11) and Kangning cohort (N=48). SYSU-TBI2 cohort was further used, in combined with 48 additionally collected matched healthy controls, to identify in-vivo plasma proteins associated with MDC and its key chronic dimensional features. FindingsChronically, MDC-not-mild TBI patients has significantly more cognitive impairment (organic intellectual deficiency (OR=49.452 (20.366 - 148.508)), encephalomalacia (OR=3.495 (1.616 - 8.732)), radiologically evident skull defect and cranioplasty (OR=3.505 (1.937 - 6.471))), but less psychiatric impairment (postconcussional syndrome, OR=0.022 (0.006 - 0.056)) compared to MDC-mild patients (PFDR < 0.05). The best prediction model for MDC (R2 ranging from 0.71 to 0.79) consisted nine psychiatric-cognitive symptomatic predictors (e.g., emotional dysregulation), one radiological predictor and two predictors concerning living and social function at chronic phase, which were recognized as the 12 key chronic dimensional features underlying TBI mental prognosis measured by MDC. Acute phase brain injuries in the left hemisphere exerted a significantly stronger influence on chronic phase mental prognosis captured by the MDC than those in the right hemisphere (P = 3.11x10-9). In particular, acute phase injuries to the parietal (PL_vs_R_FDR = 6.71x10-7) and temporal (PL_vs_R_FDR = 4.72x10-3) lobes demonstrated the most marked left-right disparities, with cognitive outcomes being the most influenced aspect by such left-bias. Chronic phase Plasma GFAP, NfL, and IL-6 protein levels were significantly associated with MDC (PFDR < 0.05), with the association of NfL mediated by recovery time (nominally significant, Pinteraction = 0.019). InterpretationOur comprehensive evaluation of MDC as a TBI mental prognostic measurement system revealed heterogenous landscape of mental outcome of TBI. Such evaluation, along with the development of the predictive model, not only facilitated MDCs clinical implementation but also identified potential targets for early intervention for mental complications in TBI. FundingSTI2030-Major Projects, 2021ZD0202000.