Li Defu, Guan Hanxiong, Wang Yujin, Zhu Tingting
Department of Radiology, Fuyong People's Hospital of Baoan District, Shenzhen, China.
Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China.
Quant Imaging Med Surg. 2025 Feb 1;15(2):1139-1150. doi: 10.21037/qims-24-1031. Epub 2025 Jan 22.
Accurate diagnosis of coronary artery disease is essential for preventing serious cardiovascular events. Although coronary computed tomography angiography (CCTA) is widely used in the clinic, it is limited because it only provides anatomical information, which makes differentiating in-depth between subtypes of noncalcified plaques and assessing the inflammatory state of coronary vessels difficult. Fractional flow reserve with computed tomography (FFR-CT) can be combined with CCTA to form a hybrid anatomic-physiologic diagnostic strategy. This study aimed to improve the recognition of stable and unstable angina with quantitative plaque characteristics, fat attenuation index (FAI), and fractional flow reserve with FFR-CT using a coronary artificial intelligence (AI)-assisted diagnostic system.
In this retrospective case-control study, 215 and 202 patients with stable and unstable angina pectoris, respectively, who were treated at our hospital between January 2015 and August 2023, were enrolled. Propensity score matching was used to reduce clinical baseline data bias. Binary logistic regression was used to determine the risk factors for unstable angina pectoris. The diagnostic efficacy of quantitative plaque characteristics, pericoronary FAI, FFR-CT, and their combined models in differentiating stable and unstable angina pectoris was determined using the area under the receiver operating characteristic (ROC) curve.
This study included 168 pairs of patients with stable or unstable angina. Patients with unstable angina had a significantly greater pericoronary FAI volume and percentage of, lipid, and fibrolipid components within the total plaque (all P<0.001) and a significantly smaller percentage of calcification components (P<0.001), FFR-CT (P=0.003), and lumen area at the narrowest point of the stenosis(P=0.003) than those with stable angina. Independent risk factors for unstable angina were FAI >-82 Hounsfield units (HU) and total intraplaque lipid component percentage >1.2% (P=0.003 and 0.009, respectively). The area under the curve (AUC) of the ROC regarding pericoronary FAI differentiating between stable and unstable angina was 0.631 (P<0.001). In contrast, the AUC of the combined model of FFR-CT, plaque characteristics, and pericoronary FAI was 0.698 (P<0.001). The AUC value of the combined model was significantly higher than that of the diagnostic model using a single index (all, P<0.001).
AI-assisted diagnostic systems could provide new methods to differentiate between stable and unstable angina. Patients with FAI >-82 HU and total intraplaque lipid component percentage >1.2% had a significantly increased risk of unstable angina, a finding that may be informative for clinical decision-making.
准确诊断冠状动脉疾病对于预防严重心血管事件至关重要。尽管冠状动脉计算机断层扫描血管造影（CCTA）在临床上广泛应用，但其存在局限性，因为它仅提供解剖学信息，这使得区分非钙化斑块亚型以及评估冠状动脉血管的炎症状态变得困难。计算机断层扫描血流储备分数（FFR-CT）可与CCTA相结合，形成一种解剖-生理混合诊断策略。本研究旨在利用冠状动脉人工智能（AI）辅助诊断系统，通过定量斑块特征、脂肪衰减指数（FAI）和FFR-CT血流储备分数来提高对稳定型和不稳定型心绞痛的识别。
在这项回顾性病例对照研究中，纳入了2015年1月至2023年8月在我院接受治疗的分别患有稳定型和不稳定型心绞痛的215例和202例患者。采用倾向评分匹配来减少临床基线数据偏差。使用二元逻辑回归确定不稳定型心绞痛的危险因素。使用受试者操作特征（ROC）曲线下面积来确定定量斑块特征、冠状动脉周围FAI、FFR-CT及其联合模型在区分稳定型和不稳定型心绞痛方面的诊断效能。
本研究纳入了168对稳定型或不稳定型心绞痛患者。与稳定型心绞痛患者相比，不稳定型心绞痛患者的冠状动脉周围FAI体积、总斑块内脂质和纤维脂质成分的百分比显著更高（均P<0.001），而钙化成分的百分比、FFR-CT（P=0.003）以及狭窄最窄处的管腔面积（P=0.003）显著更小。不稳定型心绞痛的独立危险因素为FAI>-82亨氏单位（HU）和总斑块内脂质成分百分比>1.2%（分别为P=0.003和0.009）。冠状动脉周围FAI区分稳定型和不稳定型心绞痛的ROC曲线下面积（AUC）为0.631（P<0.001）。相比之下，FFR-CT、斑块特征和冠状动脉周围FAI联合模型的AUC为0.698（P<0.001）。联合模型的AUC值显著高于使用单一指标的诊断模型（均P<0.001）。
AI辅助诊断系统可为区分稳定型和不稳定型心绞痛提供新方法。FAI>-82 HU且总斑块内脂质成分百分比>1.2%的患者发生不稳定型心绞痛的风险显著增加，这一发现可能为临床决策提供参考。
