Objectives Traditional filtered back projection (FBP) play a major role in defining the quality and integrity of medical imaging using computed tomography. Recent studies demonstrated that sinogram affirmed iterative reconstructions (SAFIRE) can produce higher-resolution images with greater robustness for the reduction of various imaging artefacts. The aim of this study was to assess the diagnostic accuracy of in-stent restenosis (>50% luminal narrowing) using low-dose high-pitch dual-source CT coronary angiography (Flash CTCA) with SAFIRE in symptomatic patients referred for conventional coronary angiography (CCA).
Methods 137 stents in 70 patients (average heart rate was 57 ± 8 bpm), were prospectively evaluated. The interval between stenting and inclusion in the study was 21 ± 12 months. Before scheduled CCA, Flash CTCA was performed between September 2011 and December 2012. In-stent noise, signal-to-noise ratio (SNR) and stent-lumen attenuation increase ratio (SAIR), as well as subjective image quality score, were measured and compared between SAFIRE reconstruction (group A) and FBP reconstruction (group B). CCA was served as the standard of reference to further analyse accuracy of both groups on detecting in-stent restenosis.
Results Of the 137 stents, group A were superior to group B on in-stent noise (22.5 ± 8.6 vs. 36.1 ± 13.9; P < 0.05), SNR (20.1 ± 7.4 vs. 16.4 ± 6.3; P < 0.05), SAIR (21.7 ± 11.5 vs. 33.4 ± 24.1%; P < 0.05), and image quality score (3.2 ± 0.73 vs. 2.5 ± 1.1; P < 0.05). On a stent-based analysis, the sensitivity, specificity, positive predictive value, negative predictive value and accuracy of group A were 90%, 92%, 81%, 96% and 92% respectively, and those of group B were 87%, 86%, 71%, 94% and 86% respectively, which showing no significant difference (P > 0.05). However, in subgroup of smaller stent (< = 3 mm; n = 78), comparing SAFIRE with FBP, specificity (86% vs. 73%), positive predictive value (77% vs. 61%) and accuracy (86% vs. 76%) improved significantly (P < 0.05); Sensitivity (85% vs. 81%) and negative predictive value (92% vs. 88%) did not differ (P > 0.05). CCTA average effective dose was (1.41 ± 0.45) mSv.
Conclusions Low-dose high-pitch dual-source CT angiography can be performed well in the detection of in-stent patency. Iterative image reconstruction significantly improves diagnostic accuracy of in-stent restenosis even in smaller stents.