Copyright: Clinics Cardive Publishing (Pty) Ltd. publisher
of Cardiovascular Journal of Africa.
Abstract
Objective:
Coronary artery disease (CAD) is prevalent in patients with severe aortic stenosis (AS) undergoing transcatheter aortic valve implantation (TAVI). Computed tomography-derived fractional flow reserve (CT-FFR) may provide non-invasive functional assessment using existing TAVI planning imaging to evaluate CT-FFR diagnostic performance compared with invasive coronary angiography (ICA) for identifying functionally significant CAD in TAVI candidates.
Methods: This single-centre retrospective study included 37 patients with severe AS undergoing TAVI (June 2016–June 2023). All underwent pre-procedural coronary CT angiography and ICA. CT-FFR was retrospectively analysed using deep learning-based software (threshold < 0.81 for significant stenosis). Diagnostic performance was assessed per-vessel (52 coronary arteries) using ICA as reference standard.
Results: Mean age was 79.2 ± 8.4 years; 56.7% were male. ICA identified significant CAD in 14 patients (37.8%), involving 21 lesions. CT-FFR successfully analysed all 52 vessels in the final cohort (mean value 0.86 ± 0.11), identifying 20 lesions (38.4%) as functionally significant; however, 8 of 62 initially screened patients (12.9%) were excluded due to insufficient CT image quality, reflecting a relevant limitation of clinical feasibility. CT-FFR demonstrated 80.9% sensitivity (95% CI: 58.1–94.6%), 95.3% specificity (95% CI: 78.6–99.2%), positive predictive value 85.0%, negative predictive value 90.6%, and diagnostic accuracy 88.5%. No significant difference existed between CT-FFR and ICA classifications (p > 0.05). At 30 days, no deaths, myocardial infarctions, or strokes occurred.
Conclusions: CT-FFR provides good diagnostic performance for detecting haemodynamically significant CAD in TAVI candidates. Integration into routine workflow may reduce purely diagnostic invasive procedures while maintaining accuracy. Larger prospective studies are needed to validate these findings.
Methods: This single-centre retrospective study included 37 patients with severe AS undergoing TAVI (June 2016–June 2023). All underwent pre-procedural coronary CT angiography and ICA. CT-FFR was retrospectively analysed using deep learning-based software (threshold < 0.81 for significant stenosis). Diagnostic performance was assessed per-vessel (52 coronary arteries) using ICA as reference standard.
Results: Mean age was 79.2 ± 8.4 years; 56.7% were male. ICA identified significant CAD in 14 patients (37.8%), involving 21 lesions. CT-FFR successfully analysed all 52 vessels in the final cohort (mean value 0.86 ± 0.11), identifying 20 lesions (38.4%) as functionally significant; however, 8 of 62 initially screened patients (12.9%) were excluded due to insufficient CT image quality, reflecting a relevant limitation of clinical feasibility. CT-FFR demonstrated 80.9% sensitivity (95% CI: 58.1–94.6%), 95.3% specificity (95% CI: 78.6–99.2%), positive predictive value 85.0%, negative predictive value 90.6%, and diagnostic accuracy 88.5%. No significant difference existed between CT-FFR and ICA classifications (p > 0.05). At 30 days, no deaths, myocardial infarctions, or strokes occurred.
Conclusions: CT-FFR provides good diagnostic performance for detecting haemodynamically significant CAD in TAVI candidates. Integration into routine workflow may reduce purely diagnostic invasive procedures while maintaining accuracy. Larger prospective studies are needed to validate these findings.
Keywords:
transcatheter aortic valve implantation, aortic stenosis, coronary artery disease, CT-derived fractional flow reserve, functional assessment
Submitted: January 21, 2026;
Accepted: April 28, 2026;
Published: August 28, 2026
Cardiovasc J Afr 2025; 37: 375-380
Volume 37, Issue 3
Cardiovasc J Afr 2025; 37: 375-380
Volume 37, Issue 3
DOI Citation Reference: dx.doi.org/10.5830/CVJA-2026-032