Kahmann Jannik, Nörenberg Dominik, Papavassiliu Theano, Schoenberg Stefan O, Froelich Matthias F, Ayx Isabelle
Department of Radiology and Nuclear Medicine, University Medical Center Mannheim, Heidelberg University, Mannheim, Germany.
First Department of Medicine-Cardiology, University Medical Center Mannheim, Heidelberg University, Mannheim, Germany.
Front Cardiovasc Med. 2024 Dec 5;11:1499219. doi: 10.3389/fcvm.2024.1499219. eCollection 2024.
Recent research highlights the role of pericoronary adipose tissue (PCAT) in coronary artery disease (CAD) development. PCAT has been recognized as a metabolically active tissue involved in local inflammation and oxidative stress, potentially impacting CAD initiation and progression. Radiomics texture analysis shows promising results to better understand the link between PCAT quality and CAD risk. Photon-counting CT (PCCT) offers improved feature stability and holds the potential for advancing radiomics analysis in CAD research.
In this retrospective, single-center, ethic committee-approved study, PCAT of the left descending artery (LAD) and right coronary artery (RCA) was manually segmented and radiomic features were extracted using pyradiomics. The study population consisted of one group of patients with CAD and plaques exclusively located in the left coronary artery and another group without CAD. Mean and standard deviation were calculated using R Statistics. Random forest feature selection was performed to identify differentiating features between the four sets CAD-LAD, CAD-RCA, non-CAD-LAD and non-CAD-RCA.
36 patients were enrolled in this study (16 female, mean age 56 years). The feature "original_glszm_GrayLevelNonUniformity" measuring the gray-level variability was identified as the most potent differentiator between CAD-LAD and non-CAD-LAD, as well as CAD-RCA and non-CAD-RCA with the greatest differentiating capability for the LAD comparison. The feature showed little differentiating power between CAD-LAD and CAD-RCA and virtually none between non-CAD-LAD and non-CAD-RCA. The mean values were consistently lower in LAD-PCAT and exhibited patient-specific reductions in CAD patients (155.16 for CAD-LAD, 163.21 for non-CAD-LAD, 189.13 for CAD-RCA and 215.40 for non-CAD-RCA).
Radiomics analysis revealed differences in PCAT texture of patients with and without CAD with a potentially more homogeneous pattern in CAD-affected patients. These changes related to plaques in the left coronary artery also seemed to occur in the unaffected RCA-PCAT, although to a slightly lesser extent.
近期研究突显了冠状动脉周围脂肪组织（PCAT）在冠状动脉疾病（CAD）发展中的作用。PCAT已被视为一种参与局部炎症和氧化应激的代谢活跃组织，可能影响CAD的发生和进展。放射组学纹理分析在更好地理解PCAT质量与CAD风险之间的联系方面显示出有前景的结果。光子计数CT（PCCT）提供了更高的特征稳定性，并具有推进CAD研究中放射组学分析的潜力。
在这项经伦理委员会批准的单中心回顾性研究中，对左前降支（LAD）和右冠状动脉（RCA）的PCAT进行手动分割，并使用pyradiomics提取放射组学特征。研究人群包括一组CAD患者且斑块仅位于左冠状动脉，以及另一组无CAD的患者。使用R统计软件计算均值和标准差。进行随机森林特征选择以识别CAD-LAD、CAD-RCA、非CAD-LAD和非CAD-RCA这四组之间的差异特征。
本研究纳入了36例患者（16例女性，平均年龄56岁）。测量灰度级变异性的特征“original_glszm_GrayLevelNonUniformity”被确定为CAD-LAD与非CAD-LAD以及CAD-RCA与非CAD-RCA之间最有效的区分特征，在LAD比较中具有最大的区分能力。该特征在CAD-LAD和CAD-RCA之间的区分能力较弱，在非CAD-LAD和非CAD-RCA之间几乎没有区分能力。LAD-PCAT中的均值始终较低，且CAD患者中呈现出个体特异性降低（CAD-LAD为155.16，非CAD-LAD为163.21，CAD-RCA为189.13，非CAD-RCA为215.40）。
放射组学分析揭示了有CAD和无CAD患者的PCAT纹理存在差异，CAD患者的纹理模式可能更均匀。与左冠状动脉斑块相关的这些变化似乎在未受影响的RCA-PCAT中也会发生，尽管程度稍小。
