To discover data-driven neuroanatomical subtype progression profiles of primary progressive aphasia (PPA) using machine learning and to characterize the diversity and complexity of PPA.
Key Findings:
Identified four neuroanatomical subtypes: S1 (left temporal), S2 (insula), S3 (temporoparietal), and S4 (frontoparietal), with subtype assignment stable for 84% of patients at first follow-up.
S1 strongly correlated with the semantic variant, while S2, S3, and S4 showed mixed associations with logopenic and non-fluent/agrammatic variants, indicating the complexity of these relationships.
Stage assignment was stable for 91.9%.
Interpretation:
Distinct neuroanatomical patterns exist within the PPA spectrum, but associations are complex, particularly for non-fluent/agrammatic and logopenic variants, indicating a need for nuanced understanding in clinical settings.
Limitations:
The study's findings may not fully conform to traditional clinico-anatomical correlations, which could limit their applicability.
The non-fluent/agrammatic and logopenic variants exhibited noisy associations, complicating interpretation.
Conclusion:
The study highlights the heterogeneity of PPA and the potential for machine learning to enhance understanding of its neuroanatomical profiles, which may inform clinical decision-making.
by Beatrice Taylor, Martina Bocchetta, Cameron Shand, Emily G Todd, Anthipa Chokesuwattanaskul, Sebastian J Crutch, Jason D Warren, Jonathan D Rohrer, Chris J D Hardy, Neil P Oxtoby
First CAR-T approach for Alzheimer’s reduces amyloid plaques in mice, while cell and gene therapies advance across blood cancers, retinal disease, amyloidosis, and FSHD
