Agreement of intracranial vessel diameters measured on 2D and 3D digital subtraction angiography using an automatic windowing algorithm.

BACKGROUND AND PURPOSE: Precise vessel measurement plays a major role in size selection of stents used for the treatment of intracranial aneurysms and became even more critical after the introduction of flow diverter stents. We assessed agreement between intracranial vessel diameters of aneurysm patients measured on 2D digital subtraction (2D DSA) and 3D volume rendering digital subtraction angiography (3D DSA) images using an automatic windowing algorithm. MATERIALS AND METHODS: Ten patients with intracranial aneurysms were enrolled and 120 measurement points were selected on both 2D and 3D DSA images acquired by a biplane angiographic system. Automatic windowing was applied to the 3D DSA images. Inter-method agreement of vessel measurements on 2D and 3D DSA images was assessed by Bland Altman plots and intraclass correlation coefficients (ICC). Inter- and intra-rater agreement of measurements on 3D DSA images were assessed by ICCs. RESULTS: The mean differences between measurements on 2D and 3D DSA images were 0.14mm for the ICA, and 0.18mm for the ACA and MCA, which is about the size of one 3D DSA image voxel. For ICA measurements, inter-method, inter-rater and intra-rater agreements were good or excellent (consistency and absolute ICC≥0.95). For ACA and MCA measurements, the inter-method, inter-rater and intra-rater agreements were also good or excellent (consistency ICC=0.94, 0.89 and 0.93, absolute ICC=0.83, 0.84 and 0.85 respectively). CONCLUSIONS: Vessel diameters may be measured on 3D DSA images with sufficient reliability for clinical use when applying an automatic windowing algorithm.