Takafuji Masafumi, Kitagawa Kakuya, Mizutani Sachio, Hamaguchi Akane, Kisou Ryosuke, Iio Kotaro, Ichikawa Kazuhide, Izumi Daisuke, Sakuma Hajime
From the Department of Radiology, Mie University Graduate School of Medicine, 2-174 Edobashi, Tsu 514-8507, Japan (M.T., K.K., H.S.); and Departments of Radiology (M.T., S.M., A.H., R.K.) and Cardiology (K. Iio, K. Ichikawa, D.I.), Matsusaka Municipal Hospital, Matsusaka, Japan.
Radiol Cardiothorac Imaging. 2023 Aug 17;5(4):e230085. doi: 10.1148/ryct.230085. eCollection 2023 Aug.
To investigate image noise and edge sharpness of coronary CT angiography (CCTA) with super-resolution deep learning reconstruction (SR-DLR) compared with conventional DLR (C-DLR) and to evaluate agreement in stenosis grading using CCTA with that from invasive coronary angiography (ICA) as the reference standard.
This retrospective study included 58 patients (mean age, 69.0 years ± 12.8 [SD]; 38 men, 20 women) who underwent CCTA using 320-row CT between April and September 2022. All images were reconstructed with two different algorithms: SR-DLR and C-DLR. Image noise, signal-to-noise ratio, edge sharpness, full width at half maximum (FWHM) of stent, and agreement in stenosis grading with that from ICA were compared. Stenosis was visually graded from 0 to 5, with 5 indicating occlusion.
SR-DLR significantly decreased image noise by 31% compared with C-DLR (12.6 HU ± 2.3 vs 18.2 HU ± 1.9; < .001). Signal-to-noise ratio and edge sharpness were significantly improved by SR-DLR compared with C-DLR (signal-to-noise ratio, 38.7 ± 8.3 vs 26.2 ± 4.6; < .001; edge sharpness, 560 HU/mm ± 191 vs 463 HU/mm ± 164; < .001). The FWHM of stent was significantly thinner on SR-DLR (0.72 mm ± 0.22) than on C-DLR (1.01 mm ± 0.21; < .001). Agreement in stenosis grading between CCTA and ICA was improved on SR-DLR compared with C-DLR (weighted κ = 0.83 vs 0.77).
SR-DLR improved vessel sharpness, image noise, and accuracy of coronary stenosis grading compared with the C-DLR technique. CT Angiography, Cardiac, Coronary Arteries . © RSNA, 2023.
研究与传统深度学习重建（C-DLR）相比，超分辨率深度学习重建（SR-DLR）在冠状动脉CT血管造影（CCTA）中的图像噪声和边缘锐度，并以有创冠状动脉造影（ICA）为参考标准，评估CCTA在狭窄分级方面的一致性。
这项回顾性研究纳入了2022年4月至9月期间使用320排CT进行CCTA检查的58例患者（平均年龄69.0岁±12.8[标准差]；男性38例，女性20例）。所有图像均采用两种不同算法重建：SR-DLR和C-DLR。比较了图像噪声、信噪比、边缘锐度、支架半高宽（FWHM）以及与ICA在狭窄分级方面的一致性。狭窄程度通过视觉分级为0至5级，5级表示闭塞。
与C-DLR相比，SR-DLR显著降低了31%的图像噪声（12.6 HU±2.3对18.2 HU±1.9；P<.001）。与C-DLR相比，SR-DLR显著提高了信噪比和边缘锐度（信噪比，38.7±8.3对26.2±4.6；P<.001；边缘锐度，560 HU/mm±191对463 HU/mm±164；P<.001）。SR-DLR上支架的FWHM（0.72 mm±0.22）明显比C-DLR上的薄（1.01 mm±0.21；P<.001）。与C-DLR相比，SR-DLR提高了CCTA与ICA在狭窄分级方面的一致性（加权κ=0.83对0.77）。
与C-DLR技术相比，SR-DLR提高了血管锐度、图像噪声和冠状动脉狭窄分级的准确性。CT血管造影、心脏、冠状动脉。©RSNA，2023。
