Cervical artery tortuosity—a reliable semi-automated magnetic resonance-based method

Background: Assessments of subclinical connective tissue disorders depend on complex approaches, emphasizing the need for more accessible methods applicable to clinical routine. Therefore, we aimed to establish a reliable approach assessing cervical vessel tortuosity, which is known to be associated with such disorders. Methods: Magnetic resonance angiography (MRA) images of ReSect study participants [single-center prospective cohort of spontaneous cervical artery dissection (sCeAD) patients] were used. Each patient underwent the same magnetic resonance imaging (MRI) protocol. The segmentation procedure was done using MATrix LABoratory 9.4 [up-sampling of raw MRA images, distance metric (DM) calculation], ITK-SNAP [region of interest (ROI) determination, vessel segmentation] and Vascular Modelling ToolKit (centerline determination). To assess inter-user variability and validity, we (I) had two blinded independent users segment all arteries and we (II) compared the results of our method to visual appraisal of vessel tortuosity done by two blinded expert neuro-radiologists. Results: A total of 526 extracranial cervical arteries were available for analysis. The inter-user variability of our method users was below 0.5% throughout. Overall, our method outperformed the visual tortuosity appraisal, as the visual grading underestimated the DM in 38.8% subjects when tasked to assess overall cervical artery tortuosity (both vertebral and internal carotid arteries) and in 16.6% and 33.3% respectively if tasked to grade anterior or posterior circulation separately. Conclusions: We present a reliable method to assess cervical artery tortuosity derived from MRA images applicable in clinical routine and future research investigating the potential correlation of sCeAD and connective tissue disorder.