Chronic traumatic encephalopathy (CTE) is a progressive neurodegenerative disease found in individuals with a history of repetitive head injury (RHI) received through playing contact sports. Currently, CTE can only be diagnosed after death through the identification of perivascular neuronal hyperphosphorylated tau (pTau) pathology, found at the depth of the cortical sulcus. Although the pathognomonic lesion is found among all cases, pTau deposition is patchy and heterogeneous among individuals. To determine whether the heterogeneity of CTE masks more subtle ordering or additional disease variants, we applied the unsupervised machine learning algorithm subtype and stage inference (SuStaIn) on fully quantitative pTau density data from 26 brain regions in 207 CTE cases and 75 control cases. SuStaIn identified three distinct pathological progression subtypes: 93 cases were classified as Subtype 1 (cortical predominant) and showed dense, rapidly progressing cortical pTau akin to the classically described CTE neuropathology; Subtype 2 (cortical sparing-hippocampal enhanced) had 105 cases that exhibited a reduced cortical pTau burden that progressed slower, but also had pronounced hippocampal involvement in a CTE-specific pattern; and Subtype 3 (copathology altered) consisted of 28 older cases with higher CTE stage, was enriched for comorbid pathologies (pTDP43, Aβ, hippocampal sclerosis, and arteriolosclerosis), and had enhanced pTau in regions highly related to those additional pathologies. Clinically, impaired performance in the Functional Activities Questionnaire (FAQ) correlated with pTau regional burden and severity only in cases with Subtype 1 pathology, suggesting the enhanced cortical pTau pathology was linked to greater functional decline. Overall, these findings help characterize the heterogeneity in CTE progression, validate key pathological variants, and will be crucial for refining diagnostic criteria and advancing in-life diagnosis.
Tuz‐Zahra et al. (Tue,) studied this question.