Vascular Surgery

Textbook outcome after EVAR, TEVAR, and complex EVAR strongly predicts 1‑year survival

This large VQI analysis evaluated elective EVAR, TEVAR, and fenestrated/branched EVAR for aneurysmal disease over two decades. Textbook outcome (TO) required no major complications, no perioperative death or reintervention, no prolonged stay, and discharge home. TO was achieved in about 78% of EVAR, 69% of TEVAR, and 61% of complex EVAR procedures. One-year survival was markedly higher with TO versus non-TO for all procedure types, with adjusted hazard ratios around 0.24–0.36. Risk factors for failing TO included age >75, impaired renal function, female sex, heart failure, and dependent functional status. These data support TO as a clinically meaningful composite quality metric and benchmark for endovascular aortic programs. 1

CTA plaque volume independently predicts recurrent events in symptomatic <70% carotid stenosis

This PARISK substudy examined whether CTA-derived plaque burden predicts recurrent ipsilateral TIA or stroke in symptomatic carotid stenosis <70%. Total plaque volume of the ipsilateral carotid artery was quantified using semiautomated segmentation on baseline CTA. Over a median 5.1 years, higher plaque volume independently increased recurrent ipsilateral event risk, with a modest hazard increase per 100 µL. Adding plaque burden to the ECST score improved model discrimination, increasing the C-statistic from 0.65 to 0.70. These findings support CTA plaque quantification as a practical tool when MRI is unavailable to refine recurrent risk assessment. 2

Braided carotid stent wire crossings act as focal thrombogenic niches in a carotid flow model

Investigators compared three clinically used carotid stents in a macrofluidic carotid model using real-time and electron microscopy. Braided devices (Casper, Wallstent) produced more platelet adhesion and thrombus than the laser-cut Protégé RX stent. Thrombus predominantly formed at wire intersections in braided stents, especially external wires contacting the vessel wall. Computational fluid dynamics showed these crossings create recirculation and stagnation zones, generating a prothrombotic shear environment. Aspirin and the αIIbβ3 inhibitor abciximab effectively prevented thrombus formation in this model. Results implicate stent design, particularly braided wire crossings, as a key determinant of intrinsic thrombogenicity. 3

Physics-based growth–remodeling plus ML to predict AAA shape and diameter evolution

This study used a physics-based growth and remodeling model to simulate abdominal aortic aneurysm shape evolution. Elastin damage and collagen production parameters were systematically varied, generating 200 distinct simulated aneurysm geometries. Kriging extended these simulations into a larger virtual cohort, which was combined with limited clinical imaging data. Multiple machine learning models were trained to rapidly predict future maximum diameter and growth rate from prior geometry. Recurrent architectures, including RNN and LSTM, achieved high R² values for diameter and growth-rate prediction. The framework illustrates how biomechanics-informed ML might ultimately personalize AAA surveillance and timing of intervention. 4