Related Experiment Video
Updated: Jul 16, 2026
Imaging In-Stent Restenosis: An Inexpensive, Reliable, and Rapid Preclinical Model
Published on: September 14, 2009
Comparison of Superresolution Deep Learning Reconstruction and Hybrid Iterative Reconstruction for Coronary Stent
Cheng Xu1, Limiao Zou1, Shaofei He2
1Department of Radiology, State Key Laboratory of Complex Severe and Rare Diseases, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, No. 1, Shuaifuyuan, Dongcheng District, Beijing, 100730, China.
Abstract:
BACKGROUND. Conventional deep learning reconstruction (DLR) CT methods provide noise reduction but limited improvements in spatial resolution. OBJECTIVE. The purpose of this study was to compare a novel superresolution DLR (SR-DLR) algorithm and a hybrid iterative reconstruction (HIR) algorithm in terms of stent visualization and diagnostic performance for in-stent restenosis on coronary CTA, with invasive coronary angiography (ICA) used as the reference. METHODS. This prospective study enrolled patients undergoing coronary CTA at 11 centers in China from September 2023 to November 2024. Participants underwent coronary CTA performed using a normal-resolution 320-MDCT scanner with reconstruction of HIR and SR-DLR images. The final study sample included participants with a coronary stent who underwent ICA within 2 months after CTA. The stent CNR (CNRstent), stent-lumen attenuation increase ratio (SAIR), and stent edge sharpness were calculated. Two radiologists (reader 1 [R1], reader 2 [R2]) independently assessed stents for subjective image quality (scale of 1-5, with 5 denoting highest quality), diagnostic confidence (scale of 1-5, with 5 indicating greatest confidence), and in-stent restenosis (defined as ≥ 50% stenosis). HIR and SR-DLR were compared using ICA as the reference standard for in-stent restenosis. RESULTS. The analysis included 73 participants (62 men and 11 women; mean age, 65.5 ± 10.3 [SD] years) with 110 stents. SR-DLR, compared with HIR, showed greater mean CNRstent (44.0 ± 20.1 [SD] vs 33.6 ± 15.7), lower mean SAIR (0.44 ± 0.36 [SD] vs 0.71 ± 0.55), and greater mean edge sharpness (469 ± 261 [SD] vs 221 ± 130 HU/mm) (all p < .05). SR-DLR, compared with HIR, showed greater subjective image quality (both readers: median score, 4 vs 3) and greater diagnostic confidence (both readers: median score, 4 vs 3) (all p < .001). SR-DLR, compared with HR, showed greater accuracy for in-stent restenosis among all stents in per-stent analysis (R1: 89.1% vs 79.1%; R2: 88.2% vs 77.3%) and per-patient analysis (R1: 89.0% vs 79.5%; R2: 87.7% vs 72.6%) and among stents with a diameter of 3.0 mm or less in per-stent analysis (R1: 91.5% vs 81.4%; R2: 93.2% vs 81.4%) and per-patient analysis (R1: 90.7% vs 76.7%; R2: 93.0% vs 76.7%). CONCLUSION. SR-DLR, compared with HIR, yielded objective and subjective improvements in stent visualization, including differences attributable to reduced blooming artifacts, and greater diagnostic performance for in-stent restenosis. CLINICAL IMPACT. The novel SR-DLR algorithm improves the performance of coronary CTA obtained using conventional 320-MDCT scanners and could support the expanded use of CTA for stent evaluation. TRIAL REGISTRATION. Chinese Clinical Trial Registry ChiCTR2300075364.
