Abstract/Description

Title

AI-Guided Simulation and Single-Cusp CLEVE to Enable High-Risk Valve-in-Valve TAVR

Authors

Joshua Dolsen MD, Omar Saleh BSc, Nicholas Valle DO, Matthew Summers MD,

Introduction

Acute coronary obstruction is a rare but potentially fatal complication of transcatheter aortic valve replacement (TAVR), with a markedly increased risk in valve-in-valve (ViV) procedures due to anatomical constraints. The risk is further heightened in patients with complex anatomy, such as low coronary ostia or narrow sinuses of Valsalva. Despite advances in imaging and procedural techniques, pre-procedural risk stratification has remained largely static, relying primarily on pre-procedural imaging with Computed Tomography (CT). Recently, artificial intelligence (AI)-driven simulation has emerged as a promising adjunct to traditional planning, enabling more precise prediction of coronary obstruction risk. Here, we present a case of ViV-TAVR complicated by high-risk coronary anatomy where AI-driven simulation was pivotal in guiding a successful intervention.

Case Presentation:

A 70-year-old female with severe, symptomatic aortic stenosis due to degeneration of a 23 mm surgical bioprosthetic aortic valve elected for ViV-TAVR. CT demonstrated high-risk anatomy with right and left coronary heights of 5.15 mm and 6.76 mm, respectively. DASI® AI simulation compared valve options and demonstrated a more favorable leaflet-to-coronary distance/coronary diameter (DLC/d) ratio for a 26 mm self-expanding Medtronic Evolut FX+® compared with a 26 mm balloon-expandable Edwards Sapien®. Despite this, the right cusp remained at high risk. To mitigate obstruction, single-cusp CLEVE was performed with electrosurgical perforation and sequential balloon dilation, followed by valve deployment and bioprosthetic valve fracture.

Results:

Immediate post-procedural echocardiography demonstrated improved hemodynamics (peak velocity 4.3 m/s to 1.5 m/s, mean gradient 40 mmHg to 4 mmHg) with preserved coronary flow. At 2 months, valve function remained stable (mean gradient 9 mmHg, valve area 1.61 cm2, peak velocity 2.1 m/s) without paravalvular leak or coronary compromise.

Discussion

Traditional CT-based pre-procedural planning for TAVR has historically relied on a limited set of anatomical measurements including coronary height, virtual transcatheter valve-to-coronary distance (VTc), and virtual transcatheter valve-to-sinotubular junction distance (VTstj). While useful, these metrics provide an incomplete picture of the complex interactions between the implanted valve, the native anatomy, and the coronary arteries.

The application of AI-based simulation and computational modeling, as demonstrated in this case, introduces new anatomical and hemodynamic relationships into pre-procedural planning. By predicting factors such as leaflet-to-coronary distance and simulating the final position and interaction of specific valve types, this technology allows for a more individualized risk assessment and enables interventional teams to make more informed decisions, ranging from optimal valve selection to proactive planning for

adjunctive leaflet modification procedures. While this case highlights the potential benefits, larger prospective studies are necessary to formally evaluate the impact of this technology on hard cardiovascular endpoints.

Conclusions:

In this case of a high-risk ViV-TAVR, AI-based pre-procedural simulation and computational modeling were instrumental in identifying the optimal valve choice and guiding a complex intervention that included prophylactic leaflet modification. The successful outcome underscores the potential of this emergent technology to enhance patient safety and procedural success in anatomically challenging TAVR cases.

Presenting Author Name/s

Joshua Dolsen MD

Faculty Advisor/Mentor

Matthew Summers, MD

Faculty Advisor/Mentor Email

mrsumme1@sentara.com

Faculty Advisor/Mentor Department

Structural Cardiology

College/School/Affiliation

Eastern Virginia Medical School (EVMS)

Share

COinS
 

AI-Guided Simulation and Single-Cusp CLEVE to Enable High-Risk Valve-in-Valve TAVR

Title

AI-Guided Simulation and Single-Cusp CLEVE to Enable High-Risk Valve-in-Valve TAVR

Authors

Joshua Dolsen MD, Omar Saleh BSc, Nicholas Valle DO, Matthew Summers MD,

Introduction

Acute coronary obstruction is a rare but potentially fatal complication of transcatheter aortic valve replacement (TAVR), with a markedly increased risk in valve-in-valve (ViV) procedures due to anatomical constraints. The risk is further heightened in patients with complex anatomy, such as low coronary ostia or narrow sinuses of Valsalva. Despite advances in imaging and procedural techniques, pre-procedural risk stratification has remained largely static, relying primarily on pre-procedural imaging with Computed Tomography (CT). Recently, artificial intelligence (AI)-driven simulation has emerged as a promising adjunct to traditional planning, enabling more precise prediction of coronary obstruction risk. Here, we present a case of ViV-TAVR complicated by high-risk coronary anatomy where AI-driven simulation was pivotal in guiding a successful intervention.

Case Presentation:

A 70-year-old female with severe, symptomatic aortic stenosis due to degeneration of a 23 mm surgical bioprosthetic aortic valve elected for ViV-TAVR. CT demonstrated high-risk anatomy with right and left coronary heights of 5.15 mm and 6.76 mm, respectively. DASI® AI simulation compared valve options and demonstrated a more favorable leaflet-to-coronary distance/coronary diameter (DLC/d) ratio for a 26 mm self-expanding Medtronic Evolut FX+® compared with a 26 mm balloon-expandable Edwards Sapien®. Despite this, the right cusp remained at high risk. To mitigate obstruction, single-cusp CLEVE was performed with electrosurgical perforation and sequential balloon dilation, followed by valve deployment and bioprosthetic valve fracture.

Results:

Immediate post-procedural echocardiography demonstrated improved hemodynamics (peak velocity 4.3 m/s to 1.5 m/s, mean gradient 40 mmHg to 4 mmHg) with preserved coronary flow. At 2 months, valve function remained stable (mean gradient 9 mmHg, valve area 1.61 cm2, peak velocity 2.1 m/s) without paravalvular leak or coronary compromise.

Discussion

Traditional CT-based pre-procedural planning for TAVR has historically relied on a limited set of anatomical measurements including coronary height, virtual transcatheter valve-to-coronary distance (VTc), and virtual transcatheter valve-to-sinotubular junction distance (VTstj). While useful, these metrics provide an incomplete picture of the complex interactions between the implanted valve, the native anatomy, and the coronary arteries.

The application of AI-based simulation and computational modeling, as demonstrated in this case, introduces new anatomical and hemodynamic relationships into pre-procedural planning. By predicting factors such as leaflet-to-coronary distance and simulating the final position and interaction of specific valve types, this technology allows for a more individualized risk assessment and enables interventional teams to make more informed decisions, ranging from optimal valve selection to proactive planning for

adjunctive leaflet modification procedures. While this case highlights the potential benefits, larger prospective studies are necessary to formally evaluate the impact of this technology on hard cardiovascular endpoints.

Conclusions:

In this case of a high-risk ViV-TAVR, AI-based pre-procedural simulation and computational modeling were instrumental in identifying the optimal valve choice and guiding a complex intervention that included prophylactic leaflet modification. The successful outcome underscores the potential of this emergent technology to enhance patient safety and procedural success in anatomically challenging TAVR cases.