Abstract
Byline: Nicolas D Garzon Mancera, Vascular Neurology, Columbia Univ in the City of New York, New York City, NY; Farid Khasiyev, SLU Sch of Medicine, St. Louis, MT; Antonio J Spagnolo Allende, Vascular Neurology, Columbia Univ in the City of New York, New York City, NY; Tatjana Rundek, MILLER SCHL OF MEDICINE UNIV M, Miami, FL; Clinton B Wright, NATIONAL INSTITUTES OF HEALTH, Rockville, MD; Mitchell Elkind; Ralph L Sacco, UNIVERSITY OF MIAMI, Miami, FL; Victor J Del Brutto, Univ of Miami, Miami, FL; OSCAR DEL BRUTTO, CLINICA KENNEDY, Guayaquil, Ecuador; Jose Gutierrez, COLUMBIA UNIVERSITY MEDICAL CE, New York, NY Introduction: Dolichoectasia of brain arteries has significant clinical implications. It can be measured with automated imaging software such as LKEB Automated Vessel Analysis (LAVA). However, this is not currently used as part of standard clinical practice. Non-automated alternatives such as a Picture Archiving Communication System like RadiAnt are more widely accessible in clinical settings Method: In this validation study, we included 600 participants of the Northern Manhattan Study (NOMAS) and 260 participants of the Atahualpa project. Both studies obtained 1.5T, 3-dimensional time-of-flight brain MRAs. Using RadiAnt (version 2022.1), three independent readers (general practitioner (GP), neurology resident (NR), and vascular neurologist (VN)) measured the diameter of the internal carotid artery (ICA); the most proximal 5 mm of the middle cerebral artery (MCA) and anterior cerebral artery (ACA); and the most proximal portion of the basilar artery (BA). The same vessels were also measured by LAVA. We calculated the intraclass correlation of each rater's diameters versus those obtained with LAVA (version 5.10.0). GP and NR measured the NOMAS cohort. VN measured the Atahualpa project cohort. Results: The intraclass correlations between diameters obtained by GP and LAVA were stronger for the BA, left ICA, and right ICA (table 1). For VN and LAVA, the same correlations were also significant. For NR and LAVA, correlations were more significant for BA and left MCA. Rater-LAVA correlations for the remaining vessels failed to reach 60%. There is a strong correlation between diameters obtained by GP and NR for the BA and right ACA. Conclusion: Our results suggest that manual measurements of the diameter of the ICAs and BA regardless of training status are reliable and could be used to identify dilated brain arteries using diameter cutoffs for possible interventions. The associations with the remaining vessels were too variable to support the same conclusion.
