Image review

The thin-section images at both low- and high-tube potentials (80 and 140 Sn kV) were de-identified and exported offline. A radiologist (R.S. with 5-year radiology experience) reviewed the images and radiology reports to classify each of the 170 ICA/CCA into those without luminal narrowing and those with different luminal stenosis grades according to the NASCET method.^16,17 Luminal stenosis was graded as mild, moderate, and severe based on <50%, 50%-70%, and >70% ICA/CCA luminal narrowing compared to the distal luminal dimension of normal ICA. Cases with narrowing of distal ICA due to severe stenosis were instead graded as severe stenosis (near-occlusion). For each ICA/CCA, the annotations were performed at a single image or section with maximum luminal stenosis and all images spanning the entire length of luminal stenosis. The ROI included both the lumen and vessel wall or plaques if present. The presence of any motion or metal streak artifacts in the region of luminal stenosis was recorded.

Each patient’s head CT and/or MRI exams were reviewed to record evidence of cerebrovascular ischemic stroke concordant with the vascular territory of the analyzed vessel. We recorded the history of any revascularization (endarterectomy), including carotid endarterectomy or ICA/CCA stenting following DECTA. The indication for revascularization therapy included severe, symptomatic ICA/CCA stenosis (20/25 patients) in patients with at least 5-year life expectancy and mild or moderate ICA/CCA stenosis and plaques with a history of embolic strokes (5/25 patients). To perform reliability analysis, a second radiologist (M.K.K., 21-year post-radiology residency experience) performed manual segmentation of 50 ICA/CCA in a randomly selected subset of 25 patients.