Background and Purpose Digital subtraction angiography (DSA) is considered the standard for characterisation of intracranial/spinal arteriovenous fistulas (AVFs). Although DSA offers superior spatial and temporal resolution, it is limited by 2D planar imaging. We directly paralleled DSA and intra-arterial cone-beam CT angiography (IACBCTA) modalities to assess the relative IACBCTA efficacy in the anatomical identification/localisation of AVFs and utility for surgical/endovascular treatment planning.

Materials and Methods Patients with intracranial/spinal AVFs underwent biplane DSA and IACBCTA procedures (AXIOM-Artis FlatDetector Biplane-Angiosuite, Siemens) at two institutions with identical contrast injection, radiation exposure, and postprocessing techniques. Images were reviewed retrospectively and independently by two neurointerventionalists on a de-identified PACS workstation. Observers were blinded to clinical information and reviewed both DSA and IACBCTA images in tandem, obtained from the same IA injection. Qualitative image analysis was performed based on the level of delineation on a scale of 1–3 (3: Excellent/Good with minimal attenuation; 2: Moderate, relevant visibility with restrictions; 1: Poor, nondiagnostic). The following parameters were scored: a) Arterial feeders, b) Venous drainers and course, c) Fistula site, d) Adjacent anatomical landmarks for cross-sectional localisation, and e) Overall diagnostic value for interpretation. Differences between the scores were defined as the IACBCTA efficacy value. For evaluation of IACBTA treatment planning efficacy, observers described the treatment strategy at the end of DSA and IACBCTA grading respectively and scored its adjunctive value: altered or more confident treatment plan versus no value. Wilcoxon signed rank test compared scores of image quality parameters. Interobserver agreement was assessed using Kendall W coefficient of concordance.

Results Thirty-two consecutive patients [22M/10F, mean age 60.9 ± 12.4 years (29–83 years)] were studied. Intracranial AVFs were classified as Cognard type I (n=3), type II (n=11), type III (n=3), type IV (n=1) and metameric dural/pial (n=5). Spinal AVFs were limited to Anson-Spetzler type I (n=9). Interobserver agreement of overall diagnostic value for IACBCTA was good (K = 0.59, P=0.001) while it was fair for IACBCTA overall efficacy value (K = 0.32, P=0.046); however, no significant difference was observed between two observers’ IACBCTA efficacy values (P=0.3). Both observers assigned significantly higher scores to IACBCTA for overall diagnostic value (both observers: P<0.0001). No significant differences were observed between DSA and IACBCTA scores while evaluating arterial feeders (observer 1: P=1.0, observer 2: P=0.5) or venous drainers (observer 1: P=0.08, observer 2: P=0.1). Both readers assigned significantly higher scores to IACBCTA when evaluating the fistula site (observer 1: P <0.0001, observer 2: P=0.0003) and adjacent anatomic landmarks/cross-sectional localisation (both observers: P<0.0001).

In 30/32 cases, both observers noted that IACBCTA did not alter the treatment plan; but provided a more confident endovascular or surgical treatment approach. Both observers altered their treatment plan from an endovascular approach to microsurgery in one case based on IACBCTA.

Conclusion IACBCTA adjunctively improves the anatomical delineation of AVFs, particularly in terms of fistula site and localisation; biplane DSA may be equally sensitive for arterial feeder and venous drainer identification. IACBCTA confers a more confident endovascular or surgical approach when performed in conjunction with DSA.

Disclosures A. Honarmand: None. J. Gemmete: None. M. Soltanolkotabi: None. M. Hurley: None. N. Chaudhary: None. A. Pandey: None. A. Shaibani: None. S. Ansari: None.