Tripartite strategy for dual reduction of radiation and iodine dose in obese CCTA: High-iodine contrast, 80 kVp, and deep learning reconstruction

Coronary CT angiography (CCTA) in obese patients is challenging due to high radiation exposure, elevated iodine load, and increased image noise. We evaluated a tripartite low-dose strategy combining high-iodine-concentration contrast media (HCM, 400 mgI/mL), 80-kVp scanning, and deep learning image reconstruction (DLIR) in obese patients (BMI ≥30 kg/m2) to simultaneously reduce radiation and iodine dose while maintaining diagnostic quality. In this prospective, single-center trial, 100 obese patients were randomized to either a low-dose group (80 kVp, HCM at 0.6 mL/kg, DLIR reconstruction; n = 50) or a standard-dose group (120 kVp, 350 mgI/mL contrast at 0.9 mL/kg, adaptive statistical iterative reconstruction [ASiR]-V 50% reconstruction; n = 50). Low-dose scans were also reconstructed with ASiR-V 50% for internal comparison. Objective (noise, signal-to-noise ratio, contrast-to-noise ratio) and subjective (5-point scale by blinded radiologists) image quality, radiation dose, total iodine load, and diagnostic performance (vs invasive coronary angiography in a 20-patient subgroup) for detecting ≥50% stenosis were compared. The low-dose protocol significantly reduced effective radiation dose by 48.6% (1.06 ± 0.26 mSv vs 2.27 ± 0.42 mSv; P <.001) and total iodine load by 23.8% (19.2 ± 2.1 gI vs 25.2 ± 2.4 gI; P <.001) compared to standard protocol. DLIR at 80-kVp significantly reduced image noise (aortic root: 16.5 ± 3.0 HU vs 25.1 ± 4.0 HU for ASiR-V at 80 kVp; P <.001) and improved contrast-to-noise ratio (e.g., proximal left anterior descending: +41.6%, P <.001) and signal-to-noise ratio (e.g., distal right coronary artery: +63.8%, P <.001). Subjective image quality scores were significantly higher for DLIR (mean 4.4 ± 0.5) versus low-dose ASiR-V (3.0 ± 0.7) and standard-dose ASiR-V (4.1 ± 0.7; P <.001), with excellent inter-observer agreement (κ = 0.88). 94% of DLIR datasets received optimal noise scores (≥4/5) versus 0% for low-dose ASiR-V. Against invasive coronary angiography, the low-dose protocol achieved 100% patient-level sensitivity and negative predictive value, with 91.7% vessel-level accuracy. Combining HCM, 80-kVp scanning, and DLIR significantly reduces radiation exposure and iodine load in obese CCTA patients without compromising image quality or diagnostic accuracy compared to standard protocols, offering a safer alternative, particularly for patients with renal impairment.