Diffuse MRI Edema Predicts Relapse in Cerebral Amyloid Angiopathy-Related Inflammation.

OBJECTIVE: To identify MRI predictors of relapse and quantify relapse and mortality risk in cerebral amyloid angiopathy-related inflammation (CAA-RI). A secondary objective was to assess the association between mycophenolate mofetil use and relapse risk. METHODS: We performed a retrospective cohort study of 36 patients with CAA-RI treated at the Cleveland Clinic from January 2020 to December 2024. Relapse was defined as recurrent vasogenic edema on MRI after corticosteroid treatment. Edema was classified as focal (1 lobar region), multifocal (2-3 regions), or diffuse (≥ 4 regions). Cause-specific Cox models and Gray's test evaluated predictors of relapse and mortality, and the effect of mycophenolate on relapse. RESULTS: Among 36 patients (mean age 71.6 ± 8.2 years; 66.7% female), 11 (30.6%) relapsed, 9 (25.0%) died without having relapse, and 16 (44.4%) had neither event after a median prednisone taper of 4.5 months. 31 patients had follow-up data for time-to-event analyses, with a median follow-up time of 18.7 months. Cumulative relapse rates were 22.4%, 26.3%, and 44.0% at 6, 12, and 24 months, respectively. Diffuse edema strongly predicted relapse (hazard ratio [HR] = 12.26; 95% confidence interval [CI] = 1.95-76.94; p = 0.007). Mycophenolate was used in 11 patients; relapse occurred in 18.2% versus 30.0% without mycophenolate treatment. Initiation of mycophenolate at steroid taper in high-relapse-risk patients was associated with a statistically nonsignificant reduction in relapse risk (HR = 0.36; 95% CI = 0.07-1.09, p = 0.231). INTERPRETATION: Relapse in CAA-RI is common and extends beyond the early post-treatment phase. Diffuse edema on baseline MRI strongly predicts recurrence, enabling effective risk stratification. Mycophenolate shows promise as a steroid-sparing therapy for preventing relapse in high-risk patients and merits prospective evaluation.