Computed tomography angiography (CTA) is a cornerstone of head and neck vascular assessment, but it requires the administration of iodinated contrast agents, which can cause adverse effects and are contraindicated in vulnerable patients. To enable contrast-free angiographic visualization, deep learning-based approaches have been proposed, which generate virtual CTA (vCTA) from non-contrast CT (NCCT). However, most existing methods adopt a single, global synthesis strategy, which often fails to delineate thin, spatially sparse vessels and degrades background anatomy. In this paper, we propose a novel localization-to-synthesis framework, called L2S-CTA. In the first stage, a localization network predicts a vascular map that localizes vasculature-related regions from NCCT. To promote vessel continuity, we introduce an inter-slice local attention (ILA) that aggregates contextual features from neighboring axial slices. In the second stage, we synthesize vCTA conditioned on the predicted localization prior, allowing the synthesizer to focus on vasculature-related regions. Experiments on both internal and external cohorts support the feasibility of the proposed framework for retrospective vCTA synthesis. The source code will be available at https://github.com/keunsoo-ko/L2S-CTA.
: Medial temporal lobe atrophy (MTA) remains an important imaging biomarker in brain magnetic resonance imaging (MRI) for diagnosing and evaluating Alzheimer's disease (AD). We compared the detection performance for determining significant MTA and the correlation with Scheltens' scale between two automatic volumetry software, NeuroQuant® and Aqua®, in non-AD and AD groups. Overall, 127 (non-AD, 54; AD, 73) individuals who underwent three-dimensional T1-weighted volumetric MRI in our hospital between January and August 2011 were enrolled and retrospectively analyzed using NeuroQuant® and Aqua®. The accuracy, sensitivity, and specificity for detecting significant MTA and the correlation for Scheltens' scale with the mean hippocampal volume of each software were evaluated. The study sample’s mean age was 78 (range, 59–97) years. Regarding accuracy and specificity, Aqua® was non-inferior to NeuroQuant®. Aqua® had significantly higher overall specificity than NeuroQuant® for detecting significant MTA. Additionally, the calculated mean hippocampal volume of Aqua® was more correlated with Scheltens' scale (Kendall's τ, - 0.48) than that of NeuroQuant® (Kendall's τ, - 0.20). The accuracy and specificity of Aqua® in detecting MTA are not inferior to those of NeuroQuant®. Additionally, Aqua® has a stronger correlation of mean hippocampal volume with Scheltens' scale than NeuroQuant®.