Diagnostic reference levels in interventional neuroradiology: a scoping review

To review the literature on diagnostic reference levels (DRLs) in interventional neuroradiology (INR), summarise reported dose values, and examine the methodologies used for their establishment.

A scoping review was conducted using SCOPUS, Web of Science, PubMed, and ProQuest. Studies reporting DRLs for INR diagnostic procedures (cerebral angiography, (CA)) and therapeutic procedures (stroke thrombectomy, (ST); aneurysm coiling, (AC); arteriovenous malformation/fistula (AVM/AVF) embolisation) were included. Extracted data comprised dose metrics, sample size, percentile definition, procedure classification, and statistical approaches used for DRL derivation.

Thirty-nine studies reported DRLs for air kerma-area product (P_KA), fluoroscopy time (FT), and reference air kerma (RAK). Most studies defined DRLs using the 75th percentile, although variations were observed in percentile selection, procedure grouping, and inclusion criteria. Considerable heterogeneity in sample sizes and data collection methods was identified. Reported DRLs varied widely: for CA, P_KA 41-256.65 Gycm², FT 6-20 min, and RAK 289-921 mGy; for ST, P_KA 110-225.1 Gycm², FT 30-45 min, and RAK 730-1590 mGy; for AC, P_KA 52.1-487.4 Gycm², FT 16-90 min, and RAK 505-4750 mGy; and for AVM/AVF embolisation, P_KA 206.4-550 Gycm², FT 59-135 min, and RAK 2350-6000 mGy.

DRLs in INR show substantial variability, partly driven by methodological inconsistencies. Greater standardisation of DRL derivation and reporting is needed to support harmonisation and optimisation.

Question How does the lack of international consensus on interventional neuroradiology (INR) diagnostic reference levels (DRLs), alongside inconsistent reporting, hinder benchmarking, optimisation, and radiation protection? Findings DRLs are reported for major INR procedures, but vary widely across studies and procedure types. Clinical relevance Differences in dose metrics, procedure classification, and data collection hinder comparison and benchmarking between centres. Standardised methods and harmonised reporting are crucial for effective dose optimisation and radiation protection in INR. Consistency in deriving DRLs would enable reliable benchmarking and support future registry-based initiatives.