基于人工智能测量的冠状动脉钙化积分与CT-FFR及斑块特征的相关性

Correlation of coronary artery calcification score with CT fractional flow reserve andplaque characteristics measured by artificial intelligence

目的采用冠状动脉CT血管成像联合数坤科技智慧平台(AI)分析冠状动脉钙化积分(CACS)与无创血流储备分数(CTFFR)及斑块特征的相关性。方法根据AI测量的CACS数值,将2021年1月~2022年12月于蚌埠医科大学第一附属医院行冠状动脉CT血管成像检查的208例患者分为低度钙化组(n=73,0分<CACS<100分)、中度钙化组(n=64,100分≤CACS≤400分)和高度钙化组(n=71,CACS>400分)。比较各组一般临床资料及AI测量下的犯罪血管及犯罪斑块的特征参数,并分析特征参数与CACS组别的相关性,通过绘制ROC曲线计算曲线下面积(AUC)评估单一指标及联合指标对两个组别(低度钙化组-中度钙化组、中度钙化组-高度钙化组)的诊断效能。结果AI测量下,CT-FFR、斑块长度、斑块体积、最小管腔面积(MLA)在不同CACS组的差异均有统计学意义(P<0.05),斑块类型在低度钙化组-中度钙化组间差异有统计学意义(P<0.05),在中度钙化组-高度钙化组的差异无统计学意义(P>0.05)。多因素Logistic回归分析显示,年龄、CT-FFR、斑块体积、MLA是更高CACS组的危险因素,斑块体积与CACS严重程度呈正相关(r=0.437,P<0.001),CT-FFR、MLA与其呈负相关(r=-0.640,-0.658,P<0.001)。ROC曲线显示,在低钙化积分组-中钙化积分组,CT-FFR、斑块体积、MLA及联合指标的AUC值分别为0.731、0.678、0.748、0.824;在中钙化积分组-高钙化积分组,CT-FFR、斑块体积、MLA及联合指标的AUC值分别0.741、0.670、0.746、0.840。CT-FFR、斑块体积、MLA三个联合指标在两个组别的诊断效能均大于单一指标。结论AI测量下,CT-FFR、斑块体积、MLA在不同CACS组均有显著差异,是更高CACS组的危险因素,CT-FFR、MLA在不同CACS组均表现出良好的诊断效能,CT-FFR、斑块体积、MLA联合时诊断效能明显提高。

Objective To investigate the correlation between coronary artery calcium score(CACS)measured by coronary computed tomography(CCTA)combined with the Sukun Technology Intelligence Platform(AI),CT fractional flow reserve(CT-FFR),and plaque characteristics.Methods Based on the CACS values measured by AI,208 patients who underwent CCTA examination at the First Affiliated Hospital of Bengbu Medical College from January 2021 to December 2022 were divided into three groups:low calcification group(n=73):0<CACS<100;moderate calcification group(n=64):100≤CACS≤400,high calcification group(n=71):CACS>400.Comparison of general clinical data and characteristics of culprit vessels and culprit plaques measured by AI in different CACS groups were analyzed.The correlation between characteristic parameters and CACS groups was assessed,and the diagnostic efficiency of single and combined indicators for two groups(low calcification group vs.moderate calcification group,moderate calcification group vs.high calcification group)was evaluated by drawing ROC curves to calculate the AUC.Results Based on AI measurements,there were statistically significant differences(P<0.05)in CT-FFR,plaque length,plaque volume,and minimal lumen area(MLA)among different CACS groups.There was a significant difference in plaque type between the low calcification group and the moderate calcification group(P<0.05),but no statistical significance in differences between the moderate calcification group and the high calcification group(P>0.05).Multifactorial logistic regression analysis indicated that age,CT-FFR,plaque volume,and MLA were risk factors for higher CACS groups.Plaque volume was positively correlated with the severity of CACS(r=0.437,P<0.001),while CT-FFR and MLA were negatively correlated with it(r=-0.640,-0.658,P<0.001).The ROC curve showed that in the low calcification group to the moderate calcification group,the AUC values of CT-FFR,plaque volume,MLA and combined index are 0.731,0.678,0.748 and 0.824 respectively;in the moderate calcification group to the high calcification group,the AUC values of CT-FFR,plaque volume,MLA and combined index were 0.741,0.670,0.746 and 0.840 respectively.The diagnostic efficiency of the combined index of CT-FFR,plaque volume and MLA was greater than that of a single index in both groups.Conclusion AI measurements show significant differences in CT-FFR,plaque volume,and MLA among different CACS groups,indicating that they are risk factors for higher CACS groups.CT-FFR and MLA demonstrate good diagnostic performance across different CACS groups,while the combination of CT-FFR,plaque volume,and MLAsignificantly improves diagnostic efficacy.