Half-dose contrast media protocol using 70 kVp abdominal dynamic CT with super-resolution deep learning reconstruction: Evaluation of image quality and contrast performance.

OBJECTIVE: To evaluate contrast enhancement and image quality in 70 kVp abdominal dynamic CT using super-resolution deep learning reconstruction (SR-DLR) with reduced iodine contrast media (CM). METHODS: A total of 600 patients who underwent abdominal dynamic CT using SR-DLR were retrospectively investigated (estimate glomerular filtration rate [eGFR] ≤ 45 mL/min/1.73 m2, 70 kVp, iodine 300 mgI/kg, n = 28; eGFR > 45, 70 kVp, iodine 400 mgI/kg, n = 402; and institutional routine protocol, 100 kVp, iodine 600 mgI/kg, n = 170). Patients were selected by propensity score matching, and the final analysis included 70 kVp-300 mgI/kg (n = 24), 70 kVp-400 mgI/kg (n = 48), and 100 kVp-600 mgI/kg (n = 48). Contrast enhancement, contrast-to-noise ratio (CNR), and signal-to-noise ratio (SNR) were analyzed for the liver, pancreas, portal vein, hepatic artery, and aorta during arterial and portal phases. RESULTS: Liver enhancement (portal phase) exceeded 60 HU in all groups and was higher in the 70 kVp-400 mgI/kg group than in the comparator. CNR and SNR showed no substantial differences in the 70 kVp-300 mgI/kg group and tended to be higher in the 70 kVp-400 mgI/kg group. Enhancement, CNR, and SNR in evaluated vessels tended to be higher in the 70 kVp-400 mgI/kg group compared with the 100 kVp-600 mgI/kg group. CONCLUSION: Using 70 kVp abdominal dynamic CT with SR-DLR enables CM reduction while maintaining liver enhancement and overall image quality.