Semi-automated tortuosity measurements confirm generalizability of IMPERATIVE trial results to real-world patients with acute ischemic stroke undergoing thrombectomy

Abstract

BackgroundCriticism of clinical trials of endovascular therapy of acute ischemic stroke due to large vessel occlusion includes their lack of generalizability. We aimed to evaluate the impact of vessel tortuosity on the outcomes of large-bore and super-bore aspiration catheters in the Imperative Trial and to compare trial's selection of patients to a real-world setting.MethodsUsing baseline craniocervical angiography, we performed semi-automated analysis of various tortuosity characteristics. Comparison of tortuosity characteristics was made to a previously published cohort of 100 consecutive patients treated with thrombectomy (real-world cohort).ResultsOf the 249 Imperative Trial patients with anterior circulation strokes, 187 (89%) had complete tortuosity assessments from the aortic arch to the occlusion site. Tortuosity indexes for the common carotid, extracranial and intracranial internal carotid artery segments were similar for both cohorts (right side 0.18 ± 0.10, 0.17 ± 0.09, 0.45 ± 0.09 vs. 0.20 ± 0.09, 0.17 ± 0.09, 0.45 ± 0.09; left side: 0.12 ± 0.08, 0.19 ± 0.09, 0.44 ± 0.07 vs. 0.15 ± 0.08, 0.18 ± 0.08, 0.47 ± 0.07 in the Imperative Trial and in the real-world cohort, respectively). The proportion of patients with type 3 aortic arches was higher in the Imperative Trial than the real-word cohort (26% vs. 15%, p = .038).ConclusionsImperative trial patients treated with aspiration thrombectomy had similar vascular tortuosity characteristics compared to patients treated with thrombectomy in a real-world clinical setting. This confirms the generalizability of Imperative Trial findings to real-world clinical practice.

Keywords: CTA; Stroke; aspiration; thrombectomy; tortuosity.

Figure 1.

Image processing and semi-automated measurements of tortuosity. (a) An illustrative example of 3D reconstruction and tortuosity index (TI) calculated for the common carotid artery (CCA), extracranial internal carotid artery (ICA) and intracranial ICA segments. Here, bilateral TI measurements are shown. (b) Examples of type I, II, III and bovine arch anatomical variants. Dashed lines indicate the top of the arch and the level of the origin of the brachiocephalic artery. Solid lines correspond to the diameter of the left CCA. (c) Tortuosity index (TI) calculations for the CCA, extracranial and ICA artery segments from the Imperative Trial patients and previously published cohort of consecutive patients treated with thrombectomy in a real-world clinical setting (Mokin et al., Figure 3 histogram data) are shown. Whiskers represent the normal range of TI variation (excluding outliers more than 1.5× the IQR from the median), asterisks the outlier datapoints, closed bars the IQR of variation, and the solid bars within the box mark the median.

Figure 2.

Case example of aspiration thrombectomy and semi-automated measurements of tortuosity. (a) Craniocervical computed tomography angiography showing right middle cerebral artery M1 occlusion. The patient had type III arch variant. Note severe tortuosity of the cervical carotid segment. (b) Aspiration thrombectomy was performed and with a single pass, complete successful reperfusion was achieved. (c) 3D reconstruction for semi-automated measurements of tortuosity index were used for data analysis.