Alzheimer's Imaging Consortium.

Background: Cavum Septum Pellucidum [CSP] is commonly observed on neuroimaging in individuals exposed to repetitive head impacts [RHI] and in post-mortem examination in Chronic Traumatic Encephalopathy [CTE]. A CSP is proposed as a potential biomarker for CTE, yet prevalence across neurodegenerative diseases and its clinical implications are largely unknown. We assessed CSP prevalence and clinical associations in RHI-exposed individuals in comparison to veterans with a history of traumatic brain injury [TBI], individuals with a neurodegenerative disease (i.e. Alzheimer's Disease [AD] or Frontotemporal dementia [FTD]) and Cognitively Unimpaired individuals [CU].

Method: The group-of-interest, i.e., individuals exposed to RHI in contact sports or military service (n = 66), was compared against age- and sex-matched ADNI-DOD participants with TBI (n = 62) and non-exposed participants of the Amsterdam Dementia Cohort (AD, n = 30; FTD, n = 25; CU, n = 25). Structural 3D brain MRI scans were visually rated on CSP grade (ranging 0-4, Figure 1) according to established criteria by two independent raters without access to clinical information. A CSP scored at least grade 2 was considered abnormal. If scores between raters differed, scans were discussed to reach consensus. Inter-rater reliability was assessed with Cohens' weighted Kappa (κ). We investigated group differences in CSP grade as well as associations between CSP grade and neuropsychiatric symptoms (using the Neuropsychiatric Inventory [NPI]) and CTE probability (using the Traumatic Encephalopathy Syndrome [TES] criteria).

Result: Inter-rater reliability was substantial (κ = 0.712). Prevalence of an abnormal CSP differed between groups (χ2 = 11.72, p = .020). An abnormal CSP was observed most often in the RHI group (43%), followed by TBI (31%, OR = 0.589, p = .158), and significantly less in AD (16%, OR = 0.255, p = .014), FTD (17%, OR = 0.267, p = .029), and SCD (14%, OR = 0.222, p = .012) compared to RHI (Figure 2). Across groups, CSP grade was not associated with severity of neuropsychiatric symptoms (F = 1.7, p = 0.151). An abnormal CSP was observed more in RHI-exposed individuals with probable (57%) or possible (55%) CTE compared to suggestive of CTE (36%) or no TES (36%).

Conclusion: A CSP was more prevalent in RHI-exposed individuals and veterans with TBI compared to patients with a neurodegenerative disease or CU individuals. Presence of a CSP on MRI may be indicative of head impact exposure, especially repetitive impacts.