Segmentation of X-ray coronary angiography with an artificial intelligence deep learning model: Impact in operator...

Visual assessment of the percentage diameter stenosis (%DS_VE ) of lesions is essential in coronary angiography (CAG) interpretation. We have previously developed an artificial intelligence (AI) model capable of accurate CAG segmentation. We aim to compare operators' %DS_VE in angiography versus AI-segmented images.

Quantitative coronary analysis (QCA) %DS (%DS_QCA ) was previously performed in our published validation dataset. Operators were asked to estimate %DS_VE of lesions in angiography versus AI-segmented images in separate sessions and differences were assessed using angiography %DS_QCA as reference.

A total of 123 lesions were included. %DS_VE was significantly higher in both the angiography (77% ± 20% vs. 56% ± 13%, p < 0.001) and segmentation groups (59% ± 20% vs. 56% ± 13%, p < 0.001), with a much smaller absolute %DS difference in the latter. For lesions with %DS_QCA of 50%-70% (60% ± 5%), an even higher discrepancy was found (angiography: 83% ± 13% vs. 60% ± 5%, p < 0.001; segmentation: 63% ± 15% vs. 60% ± 5%, p < 0.001). Similar, less pronounced, findings were observed for %DS_QCA  < 50% lesions, but not %DS_QCA  > 70% lesions. Agreement between %DS_QCA /%DS_VE across %DS_QCA strata (<50%, 50%-70%, >70%) was approximately twice in the segmentation group (60.4% vs. 30.1%; p < 0.001). %DS_VE inter-operator differences were smaller with segmentation.

%DS_VE was much less discrepant with segmentation versus angiography. Overestimation of %DS_QCA  < 70% lesions with angiography was especially common. Segmentation may reduce %DS_VE overestimation and thus unwarranted revascularization.