Large vessel occlusion incidence score (LVO-IS): the development and external validation of a model to estimate the proportion of large vessel occlusion among admitted ischemic stroke patients without CTA
Year: 2026
Type: article
Study design: Observational Study
Abstract: Background and purpose Accurate estimates of the region-specific incidence of large vessel occlusion (LVO) acute ischemic stroke (AIS) are essential for identifying disparities in mechanical thrombectomy access and guiding public health planning. Herein, we aim to derive, test, and validate the LVO-IS (Large Vessel Occlusion Incidence Score) as an epidemiological tool using widely available clinical and radiographic data points to estimate the proportion of LVO among ischemic stroke patients. Methods We selected three variables for our model, which included presence of cortical infarct on CT or MRI, NIHSS, and VAN score at presentation. CTA was used as ground truth for LVO presence, and LVO was defined as occlusions in ICA, M1 or M2, A1, P1, or basilar artery. The model was built using multivariable logistic regression in a cohort of patients with AIS at two comprehensive stroke centers in the greater Houston area, then validated in an external cohort of patients from Medical City, Pasig City, Philippines from January to October 2022. Results Among the 338 patients that met inclusion criteria in the Houston cohort the LVO-IS score, which included dichotomized NIHSS (≥10), VAN (Visual defect, aphasia, or neglect), positive or negative, and presence of cortical changes on CT or MRI, showed high classification performance (AUC 0.85). Among the 200 patients that met criteria in the Philippines, the previous regression was used and showed high classification performance (AUC 0.88). Sensitivity was 42.3%, and specificity was 96.6%. Conclusion Using widely available clinical data and non-contrast CT Head features, the proportion of LVO among admitted ischemic stroke can be estimated with high diagnostic accuracy. Implementation of this model, with local recalibration, may help under resourced settings without CTA to estimate their regional LVO burden.
Source: Frontiers in Neurology
Authors: Ananya Iyyangar(Data curation, Writing – original draft), Kimberly Geronimo(Data curation, Methodology, Writing – original draft), Deepa Dongarwar(Data curation, Formal analysis, Methodology, Writing – original draft), Amanda Jagolino-Cole(Writing – original draft, Writing – review & editing), Anjan Ballekere(Data curation, Writing – original draft), Ngoc Mai Le(Data curation, Writing – original draft), Hussain Azeem(Data curation, Writing – original draft), Prasen Marella(Writing – review & editing), Rania Abdelkhaleq(Writing – original draft), Fawaz Al-Mufti(Writing – original draft), Kaiz S. Asif(Writing – original draft), Dileep R. Yavagal(Writing – original draft), Santiago Ortega-Gutierrez(Writing – original draft), Jo Ann Soliven(Conceptualization, Data curation, Methodology, Writing – original draft), Sunil A. Sheth(Methodology, Writing – review & editing), Sushanth Rao Aroor(Conceptualization, Data curation, Formal analysis, Methodology, Writing – original draft, Writing – review & editing)
Language: English
Subfield: Clinical Neurology
Field: Medicine
Domain: Health Sciences
Open Access status: gold
Estimated APC paid: $3,801
OpenAlex ID: W7236686911
