Left–Right Symmetry of the Carotid Bifurcation: An Imaging Analysis of 141 Reconstructed Models from the Open-Access Carotid Bifurcation Database

Abstract

Left-sided carotid arteries carry a disproportionate burden of atherosclerotic plaque, intraplaque hemorrhage, and ipsilateral ischemic stroke, yet whether the underlying bifurcation geometry is itself left–right asymmetric is unresolved. We tested whether reconstructed carotid bifurcation lumen geometry differs systematically between sides. We performed a cross-sectional secondary analysis of the open-access Carotid Bifurcation Database, a Zenodo deposit (version 2.0.0) of patient-specific three-dimensional surface models reconstructed from computed tomography angiography acquired in routine stroke diagnostics at two German centers. The analytic sample was 141 carotid case-sides (70 left and 71 right) from 78 patients. The primary outcome was lumen mesh triangle count, a resolution-standardized proxy for reconstructed surface complexity, with lumen surface area as a co-primary measure. The primary analysis was a paired Wilcoxon signed-rank test of the left-minus-right difference in the 63 patients with bilateral models, with 95% confidence intervals (CIs) from a 10,000-iteration bootstrap. Per-case asymmetry indices, plaque-surface geometry, and a NASCET-equivalent stenosis severity measured from the released centerlines were secondary outcomes, with Benjamini–Hochberg correction across nine tests. Lumen triangle count did not differ between sides (median paired difference, left − right, 944; 95% CI, − 5,276 to 3,200; p = 0.753), nor did lumen surface area (median difference: 32.6 mm2; 95% CI, − 68.1 to 99.2; p = 0.617). Per-case asymmetry indices were centered near zero for triangle count (median: 0.017; p = 0.701) and surface area (median: 0.015; p = 0.632). Plaque triangle count showed no laterality (median difference: 332; p = 0.958), nor did NASCET-equivalent stenosis severity (paired median difference: 0%; 95% CI, − 2.5 to 4.8; p = 0.453). Sensitivity analyses were concordant. No left–right differences were identified in reconstructed carotid bifurcation lumen geometry, plaque-surface geometry, or stenosis severity. The established left-sided excess of plaque vulnerability is therefore unlikely to arise from gross bifurcation lumen asymmetry, and these models provide a quantified baseline for hemodynamic analysis.