Radiation exposure during transvenous embolization of cerebrospinal fluid venous fistulae: determinants, benchmarks, and mitigation strategies.

Abstract

Transvenous embolization for cerebrospinal fluid-venous fistula (CVF)-related spontaneous intracranial hypotension is increasingly performed, yet procedure-level dosimetry and specific benchmarks are undefined. We quantified radiation exposure during CVF embolization, identified determinants of higher dose, and derived local reference levels using diagnostic-reference-level (DRL) methodology. This retrospective study analyzed a single center cohort of consecutive CVF embolizations (May 2023-September 2025). Studied dose indices were kerma-area product (KAP, Gy·cm²) and reference air kerma (K_a,r, Gy), partitioned by acquisition mode and plane. Associations with patient/procedural variables (body mass index (BMI), number of levels, navigation time, Onyx injection time, fluoroscopy time, plane use, cone-beam CT (CBCT)) were evaluated using non-parametric tests and Spearman correlations. DRLs were defined as the 75th percentile. Radiation-attributable skin effects were assessed at follow-up when indicated. 52 embolizations were performed in 43 patients. Median KAP was 288.3 Gy·cm² and median K_a,r 3.29 Gy. For all procedures, DRLs were KAP 515.7 Gy·cm² and K_a,r 4.76 Gy; for single-level cases, K_a,r 3.23 Gy. K_a,r >3 Gy occurred in 28/52 (53.8%) procedures. Dose accrued predominantly on the lateral plane and increased with higher BMI, more embolized levels, and longer navigation/injection times. CBCT use was not associated with higher dose. No radiation-attributable skin injury was documented on follow-up. Transvenous CVF embolization frequently incurs significant radiation exposure, mostly driven by lateral plane utilization and procedural complexity. We propose practical adjustments-limiting lateral plane utilization, digital zoom, and staging multilevel cases-that can plausibly reduce exposure.