Diagnostic Performance of TOF, 4D MRA, Arterial Spin-Labeling, and Susceptibility-Weighted Angiography Sequences in the Post-Radiosurgery Monitoring of Brain AVMs.

Abstract

Background: Brain AVMs are congenital anomalies of the cerebrovascular system, often discovered incidentally or through symptomatic presentations such as intracranial hemorrhage, seizure, headache, or neurologic deficits. Various treatment modalities exist for AVMs, including radiosurgery, a treatment technique that is noninvasive and efficient. Accurate imaging is crucial for risk assessment, treatment planning, and monitoring of these patients before and after radiosurgery.

Purpose: Currently, DSA is the preferred imaging technique. Despite its efficacy, DSA is notably invasive, presenting inherent risks to the patients. This systematic review and meta-analysis aimed to evaluate the efficacy of MRI sequences for monitoring brain AVMs after radiosurgery.

Data source: We performed a comprehensive search of PubMed, Scopus, Web of Science, and EMBASE databases and a methodologic quality assessment with the QUADAS-2 checklist diagnostic test accuracy.

Study selection: According to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, 3,220 abstracts were screened, 98 articles were reviewed in full text, and 14 articles met the inclusion criteria.

Data analysis: We used the bivariate random-effects meta-analysis model with STATA/MP 17 software for data analysis.

Data synthesis: No publication bias was detected. Fourteen studies were eligible for qualitative and quantitative analysis. MRI offers high sensitivity (85%) and specificity (99%) in detecting residual AVMs. Different MRI sequences, including 3D TOF-MRA, 4D MRA, and arterial spin-labeling (ASL) demonstrated varying diagnostic accuracies with areas under the curve of 0.92, 0.97, and 0.96, respectively. 4D MRA had a sensitivity of 72% and specificity of 99%, ASL showed a sensitivity of 90% and specificity of 92%, while 3D TOF-MRA had 90% sensitivity and 87% specificity.

Limitations: Meta-regression did not fully explain the sources of heterogeneity. Only 1 study assessed the susceptibility-weighted angiography (SWAN) method, and most studies involved small participant groups with varied MR techniques and sequences. Additionally, the retrospective nature of most studies may introduce bias, warranting cautious interpretation of the results.

Conclusions: MRI sequences show acceptable diagnostic performance in postradiosurgery monitoring of brain AVMs, with ASL and 4D MRA showing acceptable diagnostic accuracy. Combining different MRI sequences may further enhance diagnostic reliability. However, further investigation is needed to assess whether MRI sequences can serve as a feasible substitute for DSA, considering their risk-benefit profile, with the potential to establish them as the recommended standard.