Author + information
- Javier Valle1,
- James Chen2,
- Adam Hansgen3,
- Dominik Wiktor4,
- Joseph Cleveland5,
- Jamaluddin Moloo6,
- John Messenger6 and
- John Carroll7
- 1Univ Colorado Hospital, Denver, Colorado, United States
- 2University of Colorado Hospital, Aurora, Colorado, United States
- 3University of Colorado Health Sciences Center, Aurora, Colorado, United States
- 4University of Colorado, Denver, Colorado, United States
- 5University of Colorado Anschutz Medical Cente, Aurora, Colorado, United States
- 6University of Colorado School of Medicine, Aurora, Colorado, United States
- 7University of Colorado Denver, Aurora, Colorado, United States
CT has been adopted as the optimal preprocedural imaging modality for TAVR, offering high resolution reconstructions of cardiovascular and valvular structures. However, quantitative analyses mainly rely on using planimetric techniques on volumetric CT images, with viewing planes subjectively selected to determine key measurements. This may result in significant intra- or inter-operator variability. It is unknown if 3D quantitative measurement may improve on this process.
3D aortic models were generated from CT scans at R-R 30% cardiac phase for each patient, followed by extraction of a series of key landmarks. 3D quantitative estimates are determined from these landmarks, yielding annular, STJ, and sinus of Valsalva (SOV) measurements. An optimal angiographic view (C-arm gantry angle) for valve implant demonstrating a collinear and symmetric view of cusp nadirs was calculated using an innovative direct-projection technique based on the 3D cusp contours. Quantitative measurements were compared between Direct 3D measurement and standard 2D interpretations (industry and institution), and required changes to the predicted optimal angiographic view were classified as no (< 3 degrees) slight (< 10 degrees) or significant (> 10 degrees).
3D aortic root model construction was performed in 73 TAVR patients. There was no significant difference in annular, SOJ, or SOV measurements by 3D vs 2D assessments (p>.05). Required changes in optimal angiographic view occurred in 29 (40%) cases (17 slight changes (23%), 12 significant changes (17%) and no changes in 44 (60%) cases. In comparison, industry predicted views required changes in 85% of cases. Root-mean-square (RMS) errors in differences between 3D predicted and actual working views were 4±5 degrees, and 10±7 for industry predicted views.
Direct 3D modeling is a viable alternative to standard 2D assessment for pre-procedure planning in TAVR. This technique offers reliable quantitative measurements of the aortic structures required for TAVR and accurate determinations of optimal angiographic views for valve implantation. Adoption of these techniques may enhance procedural efficiency and reduce contrast and radiation exposure through improved pre-procedural planning.
STRUCTURAL: Valvular Disease: Aortic