Modeling of valve-in-valve transcatheter aortic valve implantation after aortic root replacement using a 3-dimensional artificial intelligence algorithm.

Abstract

Aortic root replacement requires construction of a composite valve-graft and reimplantation of coronary arteries. This study assessed the feasibility of valve-in-valve transcatheter aortic valve implantation after aortic root replacement. A retrospective review was conducted on 74 consecutive patients who received a composite valve-graft at a single institution from 2019 to 2021. Forty patients had bioprosthetic valves with adequate postoperative gated computed tomographic angiography scans. Computational simulations of balloon and self-expanding transcatheter valve deployments were performed. The modeled coronary distances were compared with traditional, manually measured valve-to-coronary distances. There was a statistically significant difference in the modeled versus manual measurements of valve to coronary distances for all patients regardless of valve type or coronary artery analyzed (P < .05). Most patients are low risk for coronary obstruction per 3-dimensional modeling, including those with a valve-to-coronary distance <4 mm. Only 1 patient (2.5%) was at risk for coronary obstruction for the left coronary artery using a balloon valve. No other valve combination was considered high risk of coronary obstruction. Five patients (12.5%) were at risk for possible valve stent deformation at the outflow, due to angulation at the graft anastomosis. Following aortic root replacement, all patients were candidates for valve-in-valve procedure using 1 or both types of transcatheter heart valves. Self-expanding valves may be at higher risk for stent frame deformation at graft anastomotic lines and balloon-expandable valves may be at higher risk of coronary obstruction.