BACKGROUND AND PURPOSE: Brain arteriovenous malformations (bAVMs) carry a variable risk of hemorrhage that persists until complete obliteration is achieved. Gamma-knife radiosurgery (GKRS) is an established treatment, yet delayed treatment effects necessitate reliable follow-up imaging. Digital subtraction angiography (DSA) remains the gold standard, but non-invasive alternatives are desirable. Arterial spin labeling (ASL) MRI, which quantifies cerebral blood flow, has shown promise in detecting residual bAVM perfusion. This study aims to assess the diagnostic performance and prognostic value of ASL compared to conventional MRI sequences and DSA for detecting residual bAVM perfusion in adult patients following GKRS. MATERIALS AND METHODS: We conducted a retrospective, multicenter study of adult patients with bAVMs treated with GKRS between 2015 and 2024 at two tertiary centers. Patients who underwent at least one follow-up brain MRI including ASL perfusion sequences were eligible. Imaging results were compared to conventional MRI sequences and DSA, used as the reference standard. Diagnostic accuracy metrics were calculated. Quantitative ASL perfusion ratios were analyzed to assess their prognostic value for nidus obliteration over time. RESULTS: A total of 99 patients (mean age 49.4 ± 16.3 years, 51% female) underwent 348 follow-up MRI examinations, including 97 with ASL. ASL demonstrated higher sensitivity (88.9%) and specificity (88.6%) than conventional MRI sequences. Quantitative ASL ratios reliably distinguished residual shunting and predicted time to obliteration. Serial reductions in ASL ratios were associated with progressive bAVM involution (p < 0.01). CONCLUSION: ASL outperforms conventional MRI in detecting residual bAVM perfusion after GKRS and provides prognostic information on obliteration. Due to its ability to both identify incomplete obliteration and to predict complete nidal involution ASL may complement standard imaging to optimize DSA timing and reduce invasive procedures.
Grüter et al. (Wed,) studied this question.