X-ray coronary angiography is the clinical gold standard for coronary artery disease during real-time cardiac interventions, but provides only 2D projections of inherently 3D vessels. Existing learnin...
PPCAR-Net is a neural network that reconstructs 3‑D coronary artery centrelines and radii from a sparse set of X‑ray angiographic views. It predicts a branch‑structured centreline-and‑radius representation directly, avoiding explicit point matching or triangulation. The method refines geometry and radius estimates using projection‑guided refiners and achieves real‑time performance with strong centreline accuracy, especially for the right coronary artery.
The paper introduces a physics-informed deep learning framework that reconstructs 3D coronary geometry from dual-view angiography and predicts velocity and pressure fields using a decoupled network with embedded physical priors. Across 32 patients and four flow conditions, the model achieved a trans‑stenotic pressure‑drop error of 2.02% and velocity/pressure relative‑L2 errors of 0.054 and 0.023, respectively, while matching hospital‑measured FFR with 93.8% diagnostic accuracy. The pipeline completes the full angiography‑to‑hemodynamics conversion in about 20 minutes per patient and supports sparse‑data assimilation for revascularization planning.
The paper introduces VasoMIM, a vascular anatomy‑aware masked image modeling framework for X‑ray angiogram analysis, which uses an anatomy‑guided masking strategy and an anatomical consistency loss to learn robust vascular representations. It also presents XA‑170K, the largest pre‑training dataset for X‑ray angiograms. Experiments on six datasets and four downstream tasks show that VasoMIM outperforms existing methods and achieves state‑of‑the‑art performance.
arXiv:2610.20509v1 Announce Type: cross Abstract: Generative super-resolution (SR) can recover sharp detail in computed tomography (CT) images from severely downsampled observations, but the missing...
The paper presents an anatomy‑aligned surface learning framework for reconstructing patient‑specific 4D myocardial surfaces from sparsely sampled short‑axis cine MRI. By parameterizing epicardial and endocardial surfaces on a shared circumferential‑longitudinal UV domain, the method transforms irregular 3D reconstruction into structured coordinate‑field completion, enabling explicit correspondence across subjects and cardiac phases. Experiments on three public datasets show the approach outperforms mesh‑based and implicit methods, achieving Chamfer distances around 2.6–2.9 mm and preserving ventricular function with small errors in volume and ejection fraction.
Multi-view reasoning in coronary X-ray angiography is inherently a cross-projection geometric problem, yet automated report generation in this setting remains largely unexplored. The 3D vascular topology leads to projection-dependent branch overlap and foreshortening, rendering single-view modeling fundamentally incomplete and unstable for lesion localization and stenosis grading.
Patient-specific 4D myocardial reconstruction from cine MRI supports quantitative functional assessment, regional motion analysis, and simulation-based modeling. However, routinely acquired short-axis...
arXiv:2610.08983v1 Announce Type: new Abstract: Generative inpainting of brain MRI volumes is essential for synthesizing healthy tissue in pathological regions, improving the accuracy and reliability...
arXiv:2608. 12274v1 Announce Type: cross Abstract: Background: Accurate segmentation of the Left Anterior Descending (LAD) artery in 3D free-breathing, non-contrast CT is critical for cardiac dose sparing in thoracic radiotherapy.
X‑LMC is a spatiotemporal deep‑learning framework that automatically scores leptomeningeal collateral (LMC) status from time‑resolved biplane digital subtraction angiography (DSA). It uses a DINOv2 backbone to encode spatial frames, a token‑level cross‑view attention module to fuse orthogonal projections, and a recurrent network to model contrast bolus dynamics. On a multicenter dataset of 134 M1‑segment occlusion patients, X‑LMC achieved a Quadratic Weighted Kappa of 0.398 and a macro‑F1 of 0.711, outperforming static and other spatiotemporal baselines and matching clinical inter‑rater agreement.
ReG-SAM is a SAM-based framework designed for 2D vessel segmentation in medical images. It introduces reference graph prompt embeddings (GPEs) and vascular prototype embeddings (VPEs) to capture global spatial and fine-grained modality-specific vessel features, respectively. By building a modality-wise vascular database and learning these embeddings from reference masks, ReG-SAM consistently outperforms existing baselines across 19 datasets, especially on thin vessels.