CT angiography for diagnosis of carotid near-occlusion: a digital subtraction angiography validation study

Diagnosing carotid near-occlusion is a difficult task—but it might get easier

“…Feature interpretation is arguably the current reference method as it was used to diagnose near-occlusion in the major RCTs [9] and in relevant prognostic studies [6], albeit Manrique-Zegarra did not use this approach [1]. Feature interpretation is performed by assessing various available features to determine if the distal ICA is small and if more proximal stenosis is the cause of the distal ICA collapse.…”

“…In our previous CTA analyses, we have focused on visual assessment of stenosis severity, distal ICA diameter, ICA side-to-side comparison, ICA-ECA comparison, and assessed Circle of Willis configuration (for the possibility of anatomical variants) [3-6, 8, 10-13]. Manrique-Zegarra assessed a limited version of this approach by visually assessing two features and using dichotomization [1]. We have previously assessed features on a spectrum (not dichotomized): A clearly "positive" or "negative" feature is weighted more in the analysis than barely "positive" or "negative" features.…”

“…In this issue, Manrique-Zegarra and co-workers present their comparison of CTA and DSA when diagnosing carotid nearocclusion [1]. As Manrique-Zegarra describes, this is a challenging diagnosis, but a relevant topic as separation of nearocclusion (especially those without full collapse, Fig.…”

“…As Manrique-Zegarra describes, this is a challenging diagnosis, but a relevant topic as separation of nearocclusion (especially those without full collapse, Fig. 1) and conventional stenosis affects management [1,2]. Indeed, 30% of ≥ 50% are near-occlusions but they are often underdiagnosed [3][4][5].…”

“…The suggested features on DSA have only been compared with feature interpretation in 32 of the cases from large RCTs, and having two or more criteria was 91% sensitive and 94% specific [9]. Manrique-Zegarra used this (≥ 2 DSA criteria) to define near-occlusion in their analysis [1]. With no clear thresholds, these RCT criteria are somewhat difficult to apply in routine practice.…”