Related Experiment Video
Updated: Jan 17, 2026
Anatomical Reconstructions of the Human Cardiac Venous System using Contrast-computed Tomography of Perfusion-fixed Specimens
Published on: April 18, 2013
Predictive Performance and Rater Reliability of CT-derived Virtual Reality Models for PDA Stent Planning in Patients
Ashish Saini1,2, Sassan Hashemi3, William McEachern4,5
1Division of Cardiology, Children's Healthcare of Atlanta, 2220 N Druid Hills Rd NE, Atlanta, GA, 30329, USA. sainia@kidsheart.com.
Insights
Generating 3D virtual reality models for transcatheter patent ductus arteriosus (PDA) stenting is feasible. These models accurately predict vascular access and aid in stent selection for congenital heart disease (CHD) patients.
Area of Science:
- Medical Imaging
- Cardiovascular Surgery
- Virtual Reality Applications
Background:
- Transcatheter patent ductus arteriosus (PDA) stenting is crucial for ductal-dependent cyanotic congenital heart disease (CHD).
- Accurate pre-procedural planning is essential for successful PDA stenting outcomes.
- Current planning methods may have limitations in predicting optimal vascular access and stent size.
Purpose of the Study:
- To assess the feasibility of creating 3D virtual reality (VR) models for PDA stenting.
- To evaluate the predictive accuracy of VR models for vascular access and stent size.
- To examine the inter- and intra-rater reliability of ductal measurements using VR models.
Main Methods:
- Retrospective study of 42 infants with ductal-dependent CHD undergoing PDA stenting.
- Generation of 3D VR models of the PDA and potential stent sizes from pre-procedural cardiac CT scans.
- Blinded prediction of vascular access and stent size by an interventional cardiologist using VR models.
- Comparison of VR predictions with actual procedural outcomes.
- Independent VR measurements of ductal dimensions (TL, SL, CI) by two blinded operators, with reliability assessed by ICCs.
Main Results:
- Successful generation of 3D VR models for all patients.
- VR-predicted vascular access matched actual access in 85.7% of cases.
- Stent size prediction was within one size category of the deployed stent in 88.1% of cases.
- Excellent intra-rater (ICCs > 0.926) and inter-rater (ICCs > 0.977) reliability for VR ductal measurements.
Conclusions:
- High-fidelity 3D VR models are feasible for planning transcatheter PDA stenting.
- VR modeling demonstrates high predictive performance for vascular access and good accuracy for stent size selection.
- VR-based ductal measurements exhibit excellent rater reliability, supporting its integration into pre-procedural planning workflows.
Abstract:
Transcatheter patent ductus arteriosus (PDA) stenting is a key palliative strategy for ductal-dependent cyanotic congenital heart disease (CHD). This study assessed the feasibility of generating 3-dimensional (3D) virtual reality (VR) models, evaluated their predictive performance for predicting vascular access and stent size, and examined inter- and intra-rater reliability of VR ductal measurements. This retrospective study included 42 infants (January 2021-December 2024) with pre-procedural cardiac CT (CCT) who underwent PDA stenting for ductal-dependent pulmonary circulation. 3D VR PDA and PDA stent models in 4 mm length increments were generated using commercial software. A blinded interventional cardiologist predicted optimal vascular access and stent size. VR-based recommendations were compared to procedural outcomes. Two blinded operators independently measured total length (TL), straight length (SL), and ductal curvature index (CI) in VR; reliability was assessed by intraclass correlation coefficients (ICCs). VR models were successfully generated for all patients. The mean age and weight at procedure were 8.1 ± 4.7 days and 3.2 ± 0.6 kg, respectively. VR-predicted vascular access matched actual access in 36 (85.7%) cases. Stent size prediction matched deployed stents in 15 (35.7%) cases and was ± one size category in an additional 22 (52.4%) cases (88.1% total cases within one size category of deployed stent). Intra-rater ICCs were > 0.926; inter-rater ICCs were > 0.977. In conclusion, high-fidelity VR models are feasible for PDA stent planning, demonstrate high predictive performance for vascular access, and have excellent rater reliability for ductal measurements. This supports integrating VR modeling into pre-procedural workflows to improve planning.
