0669 MRI-Derived Perivascular Space Volumes in CPAP-Treated Obstructive Sleep Apnea

Abstract Introduction Obstructive sleep apnea (OSA) has been associated with cognitive decline and neurodegenerative risk, potentially through impaired glymphatic clearance. MRI-visible perivascular spaces (PVS) may relate to this dysfunction, and longitudinal studies suggest CPAP adherence may stabilize PVS progression, though evidence remains limited. We examined whether OSA severity relates to PVS burden, plasma AD biomarkers, cognition, and CSF volumetrics, and whether CPAP adherence alters these measures over time. We hypothesize that severity of OSA will be associated with greater baseline PVS burden and that PVS burden will be attenuated with consistent CPAP use. Methods Adults with newly diagnosed OSA and controls without OSA were enrolled in a prospective longitudinal neuroimaging study. Fourteen participants have completed baseline MRI (OSA n=8; controls n=6). Eleven participants (OSA n=8; controls n=3) have completed follow-up MRI; three participants are pending follow-up scans. Among OSA participants with follow-up MRI, 5 were CPAP-adherent and 3 were non-adherent based on device-recorded usage (≥4 h/night on ≥70% of nights). PVS segmentation was performed using a validated convolutional neural network–based pipeline applied to FreeSurfer-derived regions of interest including centrum semiovale, basal ganglia, and total white matter. PVS volumes were normalized to white matter volume. Paired t-tests and ANOVA were used in preliminary comparisons. Results Across participants with follow-up MRI, preliminary comparisons did not reveal consistent or interpretable patterns of change in normalized PVS burden within or between groups. Early numerical differences suggested that OSA participants, particularly CPAP-non-adherent individuals, may show greater increases in PVS burden than controls and CPAP-adherent individuals. However, these observations remain tentative given the small sample and incomplete dataset. Conclusion These early longitudinal observations do not yet clarify whether OSA severity or CPAP use influences short-term PVS burden. Preliminary patterns suggest possible differences in PVS by OSA severity and CPAP adherence, but the current dataset is still under active enrollment and not yet complete for statistical interpretation. While additional biomarkers and cognitive measures are being collected within the study, ongoing data collection and full statistical modeling of the complete cohort are required to understand the associations between OSA severity, PVS burden, and CPAP adherence. Support (if any) American Academy of Sleep Medicine