Reproducibility and workflow efficiency of a novel software for quantitative coronary plaque characterization using...

The clinical adoption of quantitative coronary plaque characterization with coronary CT angiography (CCTA) is currently limited by time demands and lack of standardization across software platforms. This study aimed to evaluate the reproducibility and efficiency of a novel plaque quantification tool and to compare results between energy-integrating detector (EID-CT) and photon-counting detector (PCD)-CT.

In total, 170 patients were retrospectively enrolled in the current study: 86 patients were scanned using conventional EID-CT, while 84 using a dual-source PCD-CT. Quantitative plaque analysis was carried out using dedicated plaque analysis tools: syngo.CT Coronary Cockpit (Software 1, currently not commercially available) and MEDIS QAngioCT (Software 2) by three experienced readers. Total plaque volume (TPV) was quantified both automatically and manually.

Among the 550 lesions analyzed, median TPV 67.0 mm³ [IQR: 31.4 to 148.7] using Software 1 and 50.8 mm³ [IQR: 23.7 to 110.0] using Software 2 (p < 0.001). Software1 demonstrated shorter measurement times than Software 2 (3.9 ± 2.5 vs. 22.9 ± 8.8 min; p < 0.001) and showed excellent intra- and inter-observer reproducibility for EID-CT and PCD-CT (ICC:0.995-0.998), whereas reproducibility using Software 2 was lower (ICC:0.845-0.880). PCD-CT showed improved agreement in TPV assessment, with a lower mean bias between software tools (PCD-CT:15.0 mm³ vs. EID-CT:45.3 mm³), and a smaller absolute difference between automated and manual measurements using Software 1, compared to EID-CT (PCD-CT:2.5 ± 23.4 mm³ vs. EID-CT:6.6 ± 28 mm³).

The prototype software offers a promising approach for rapid, reproducible quantification of coronary atherosclerosis, with PCD-CT technology further reducing manual corrections and inter-platform variability.