Kravchenko Dmitrij, Vecsey-Nagy Milan, Varga-Szemes Akos, Hagar Muhammad Taha, Schoepf U Joseph, Gnasso Chiara, Zsarnóczay Emese, O'Doherty Jim, Caruso Damiano, Laghi Andrea, Emrich Tilman, Tremamunno Giuseppe
Department of Radiology and Radiological Science, Medical University of South Carolina, Charleston, SC, USA; Department of Diagnostic and Interventional Radiology, University Hospital Bonn, Bonn, Germany; Quantitative Imaging Laboratory Bonn (QILaB), Bonn, Germany.
Department of Radiology and Radiological Science, Medical University of South Carolina, Charleston, SC, USA; Heart and Vascular Centre, Semmelweis University, Budapest, Hungary.
Int J Cardiol. 2025 Feb 1;420:132749. doi: 10.1016/j.ijcard.2024.132749. Epub 2024 Nov 21.
The impact of novel photon-counting detector (PCD)-CT technology on in-vivo radiomics is not fully understood. This study aimed to compare the intra-individual stability and reproducibility of pericoronary adipose tissue (PCAT) radiomic features between PCD-CT and energy-integrating detector (EID)-CT in patients undergoing coronary CT angiography (CCTA) on both systems.
Patients undergoing clinically indicated CCTA on an EID-CT were prospectively enrolled for research PCD-CCTA within 30 days. Image acquisition parameters were standardized; PCD-CT datasets were reconstructed both down-sampled to 0.6 mm to match the clinical scan (PCD-CT) and at 0.2 mm ultrahigh-resolution mode (PCD-CT). Automatic PCAT segmentation was performed; a total of 110 radiomic feature classes were extracted and compared across the three datasets (EID-CT, PCD-CT, and PCD-CD). Feature stability was assessed using paired t-test filtered for false discoveries using Benjamini-Hochberg method, and reproducibility using intraclass correlation coefficient (ICC).
A total of 42 patients (34 male [81.0 %]; 67.9 ± 7.6 years) were included. Feature stability was 91 % for EID-CT vs. PCD-CT, but decreased for UHR datasets (EID-CT vs. PCD-CT: 55 %; PCD-CT vs. PCD-CT: 51 %). However, inter-scanner reproducibility was poor in both comparisons (EID-CT vs. PCD-CT median ICC: 0.43 [0.03-0.69]; EID-CT vs. PCD-CT: 0.29 [0.01-0.51]). Nevertheless, reproducibility improved within PCD-CT datasets (PCD-CT vs. PCD-CT: 0.72 [0.48-0.83]), regardless of the difference in slice thickness.
Most PCAT radiomic features remained stable between EID-CT and PCD-CT, although inter-scanner reproducibility was poor, emphasizing the significant impact of detector technology. Conversely, reproducibility of features within PCD-CT datasets showed more consistent results, even when comparing standard to UHR.
新型光子计数探测器（PCD）CT技术对体内放射组学的影响尚未完全明确。本研究旨在比较接受两种系统冠状动脉CT血管造影（CCTA）检查的患者中，PCD-CT与能量积分探测器（EID）-CT之间冠状动脉周围脂肪组织（PCAT）放射组学特征的个体内稳定性和可重复性。
在EID-CT上接受临床指征CCTA检查的患者在30天内被前瞻性纳入研究PCD-CCTA。图像采集参数标准化；PCD-CT数据集分别重建为下采样至0.6mm以匹配临床扫描（PCD-CT）和0.2mm超高分辨率模式（PCD-CT）。进行自动PCAT分割；共提取110个放射组学特征类别，并在三个数据集（EID-CT、PCD-CT和PCD-CD）之间进行比较。使用经Benjamini-Hochberg方法过滤错误发现的配对t检验评估特征稳定性，使用组内相关系数（ICC）评估可重复性。
共纳入42例患者（34例男性[81.0%]；67.9±7.6岁）。EID-CT与PCD-CT之间的特征稳定性为91%，但超高分辨率数据集的稳定性下降（EID-CT与PCD-CT：55%；PCD-CT与PCD-CT：51%）。然而，在两种比较中，扫描仪间的可重复性均较差（EID-CT与PCD-CT中位数ICC：0.43[0.03-0.69]；EID-CT与PCD-CT：0.29[0.01-0.51]）。尽管如此，PCD-CT数据集中的可重复性有所提高（PCD-CT与PCD-CT：0.72[0.48-0.83]），无论切片厚度有无差异。
尽管扫描仪间的可重复性较差，但大多数PCAT放射组学特征在EID-CT和PCD-CT之间保持稳定，这强调了探测器技术的重大影响。相反，即使在比较标准分辨率和超高分辨率时，PCD-CT数据集中特征的可重复性也显示出更一致的结果。
