Virtual 2D angiography from four-dimensional digital subtraction angiography (4D-DSA): A feasibility study

Abstract

Background Digital subtraction angiography (DSA) remains the gold standard for angiographic evaluation of cerebrovascular pathology, however, multiple acquisitions requiring additional time and radiation are often needed. In contrast, 3D-DSA provides volumetric information from a single injection but neglects temporal information. Four-dimensional-DSA (4D-DSA) combines temporal information of 2D-DSA with volumetric information of 3D-DSA to provide time-resolved tomographic 3D reconstructions, potentially reducing procedure time and radiation. This work evaluates the diagnostic quality of virtual single-frame 4D-DSA relative to 2D-DSA images by assessing clinicians’ ability to evaluate cerebrovascular pathology. Methods Single-frame images of four projections from 4D-DSA and their corresponding 2D-DSA images (n = 15) were rated by two neurointerventional radiologists. Images were graded based on diagnostic quality (0 = non-diagnostic, 1 = poor, 2 = acceptable, 3 = good). Dose area product (DAP) for each case was recorded for all 2D-DSA, 4D-DSA acquisitions, and the overall procedure. Results The mean diagnostic quality of all four 4D-DSA projections from both raters was 1.75 while the mean of 2D-DSA projections was 2.8. Student’s t-test revealed significant difference in diagnostic quality between 4D-DSA and 2D-DSA at all four projections (p < 0.001). On average 4D-DSA acquisitions accounted for 30% dose compared to the overall average aggregated dose per procedure. Conclusions The difference in image quality between virtual single-frame 4D-DSA and their respective 2D-DSA images is statistically significant. Furthermore, 4D-DSA acquisitions require less radiation dose than conventional procedures with 2D-DSA acquisitions.