A101 Optimizing CT angiographic stent imaging: influence of tube voltage and reconstruction parameters on objective image quality

Le résumé fourni par la source

Background/Introduction Stents cause blooming artifacts and artificial-lumen-narrowing (ALN) in CT-angiography. ALN and contrast-to-noise-ratio (CNR) are influenced by tube-voltage and reconstruction parameters, which may interact. Objective/Purpose To determine the optimal combination of tube-voltage and reconstruction parameters and their relative impact on objective image quality in carotid, flow-diverting, and intracranial stents. Methods/Case Description Stent-bearing, contrast-agent filled tubes were placed in a body donor at cervical carotid (Boston Wallstent), basilar (Terumo FRED), and middle cerebral artery (Acandis Accero) levels. Images acquired at six tube-voltages (70-120 kV; identical dose) were reconstructed including all combinations (600) of two matrix sizes (512×512/1024×1024), overlapping/non-overlapping increments, five vascular kernels (Hv36-Hv59), and five iterative reconstruction (IR) levels. ALN, CNR, and an imaging-quality-score (IQS; equal weighting of ALN and CNR) were calculated. Results/Findings ALN was mainly influenced by kernel, tube-voltage, and IR (partial η2: 0.759-0.924; 0.128-0.452; 0.054-0.283), with sharp kernels, low tube-voltage and higher IR-levels reducing ALN. Higher-resolution matrices reduced ALN in carotid and flow-diverting stents (partial η2: 0.278-0.377). CNR was strongly influenced by kernel, IR, and tube-voltage (partial η2: 0.986-0.987; 0.941-0.956; 0.883-0.936). Soft kernels, higher IR-levels, and lower tube-voltages increased CNR. IQS was predominantly affected by kernel, followed by tube-voltage, IR, and matrix-size (partial η2: 0.832-0.872; 0.322-0.445; 0.133-0.246; 0.106-0.220). The combination of Hv59, 80kV, IR-level 5, and 1024×1024 matrix yielded the highest IQS, across all stents. All results were statistically significant (p<0.01). Reconstructions with highest (top row; Hv59, 80kV, IR-level 5, 1024x1024 matrix, overlapping increment) and lowest imaging-quality-scores (bottom row; Hv36, 90kV, IR-level 1, 512x512 matrix, non-overlapping increment). a&d: Boston Wallstent, b&e: Terumo FRED, c&f: Acandis Accero. Discussion/Conclusions/Learning Points Low tube-voltage, sharp kernels, high IR-levels, and high-resolution matrices optimize CT-angiographic stent imaging by minimizing artificial-lumen-narrowing while preserving CNR.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.