Deep Learning Algorithm-Reconstructed Triple-Rule-Out CT: Image Quality and Performance in Evaluating Coronary Lumen Stenosis.

To assess radiation burden and contrast medium load, image quality, and the performance of AI-reader and junior radiologists in evaluating coronary stenosis using 80-kVp triple-rule-out (TRO) CT with deep learning image reconstruction (DLIR), 159 patients scheduled for TRO were prospectively recruited and randomized into group A (n = 80; 80-kVp with DLIR) or group B (n = 79; 100-kVp with adaptive statistical iterative reconstruction-V [ASIR-V 60%]). Patients were stratified by BMI (< 25 kg/m2 or ≥ 25 kg/m2). Contrast injection rate and volume were tailored to body surface area and tube voltage. Image quality was evaluated for pulmonary, coronary, and thoracic-abdominal arteries, focusing on noise, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), edge rise slope (ERS), figure-of-merit (FOM), and subjective scoring. The collective diagnosis by two chief radiologists served as the consensus-based reference standard for evaluating AI-reader and junior radiologists' reading consistency on coronary stenosis. Group A showed significantly lower radiation dose (7.29 ± 1.59 mSv vs. 11.22 ± 3.12 mSv, P < 0.001) and contrast volume (51 ± 5 mL vs. 62 ± 7 mL, P < 0.001) compared to group B. Group A also demonstrated lower image noise, higher SNR and CNR, comparable or superior ERS, and higher FOM across nine artery sites. Subjective scores for sharpness, vascular contrast, small-branch visibility, and overall image quality were comparable or better in Group A. AI-reader and junior radiologists performed similarly in classifying coronary stenosis in both groups. Then, 80-kVp TRO-CT with DLIR reduced radiation by 35.1% and contrast by 18.8%, while maintaining excellent image quality and diagnostic performance across BMI groups, offering a superior dose-to-quality balance compared to conventional methods.