Ultra-High-Resolution Computed Tomography Angiography for Assessment of Coronary Artery Stenosis.

BACKGROUND: Limitations of coronary computed tomography (CTA) include false-positive stenosis at calcified lesions and assessment of in-stent patency. A prototype of ultra-high resolution computed tomography (U-HRCT: 1,792 channels and 0.25-mm slice thickness×128 rows) with improved spatial resolution was developed. We assessed the diagnostic accuracy of coronary artery stenosis using U-HRCT.Methods and Results:Seventy-nine consecutive patients who underwent CTA using U-HRCT were prospectively included. Coronary artery stenosis was graded from 0 (no plaque) to 5 (occlusion). Stenosis grading at 102 calcified lesions was compared between U-HRCT and conventional-resolution CT (CRCT: 896 channels and 0.5-mm slice thickness×320 rows). Median stenosis grading at calcified plaque was significantly improved on U-HRCT compared with CRCT (1; IQR, 1-2 vs. 2; IQR, 1-3, P<0.0001). Assessability of in-stent lumen was evaluated on U-HRCT in 79 stents. Stent strut thickness and luminal diameter were quantitatively compared between U-HRCT and CRCT. Of 79 stents, 83.5% were assessable on U-HRCT; 80% of stents with diameter 2.5 mm were regarded as assessable. On U-HRCT, stent struts were significantly thinner (median, 0.78 mm; IQR, 0.7-0.83 mm vs. 0.83 mm; IQR, 0.75-0.92 mm, P=0.0036), and in-stent lumens were significantly larger (median, 2.08 mm; IQR, 1.55-2.51 mm vs. 1.74 mm; IQR, 1.31-2.06 mm, P<0.0001) than on CRCT. CONCLUSIONS: U-HRCT with improved spatial resolution visualized calcified lesions with fewer artifacts. The in-stent lumen of stents with diameter ≥2.5 mm was assessable on U-HRCT.

Estimating subjective evaluation of low-contrast resolution using convolutional neural networks.

To develop a convolutional neural network-based method for the subjective evaluation of computed tomography (CT) images having low-contrast resolution due to imaging conditions and nonlinear image processing. Four radiological technologists visually evaluated CT images that were reconstructed using three nonlinear noise reduction processes (AIDR 3D, AIDR 3D Enhanced, AiCE) on a CT system manufactured by CANON. The visual evaluation consisted of two items: low contrast detectability (score: 0-9) and texture pattern (score: 1-5). Four AI models with different convolutional and max pooling layers were constructed and trained on pairs of CANON CT images and average visual assessment scores of four radiological technologists. CANON CT images not used for training were used to evaluate prediction performance. In addition, CT images scanned with a SIEMENS CT system were input to each AI model for external validation. The mean absolute error and correlation coefficients were used as evaluation metrics. Our proposed AI model can evaluate low-contrast detectability and texture patterns with high accuracy, which varies with the dose administered and the nonlinear noise reduction process. The proposed AI model is also expected to be suitable for upcoming reconstruction algorithms that will be released in the future.