# Predictors of Diagnostic Inaccuracy of Detecting Coronary Artery Stenosis by Preprocedural CT Angiography in Patients Prior to Transcatheter Aortic Valve Implantation

## TL;DR

This study identifies younger age and poor image quality as factors that lead to inaccurate detection of heart artery blockages using CT scans before aortic valve implantation.

## Contribution

The study identifies novel predictors of diagnostic inaccuracy in pre-TAVI CT angiography for coronary artery disease.

## Key findings

- Younger age is independently associated with inaccurate CT assessment of coronary artery stenosis.
- Poor CT image quality significantly increases the likelihood of diagnostic inaccuracy.
- Atrial fibrillation and scanner generation showed initial discriminative power in univariate analysis.

## Abstract

Background: The diagnostic performance of preprocedural CT angiography in detecting coronary artery disease (CAD) in patients scheduled for transcatheter aortic valve implantation (TAVI) has been reported. However, data on predictors of diagnostic inaccuracy are sparse. We sought to investigate clinical characteristics and imaging criteria that predict the inaccurate assessment of coronary artery stenosis based on pre-TAVI-CT. Methods: The patient- and vessel-level analysis of all CT datasets from 192 patients (mean age 82.1 ± 4.8 years; 63.5% female) without known CAD or severe renal dysfunction was performed retrospectively in a blinded fashion. Significant CAD was defined as a CAD-RADS™ 2.0 category ≥ 4 by CT. Invasive coronary angiography (ICA) served as the reference standard for relevant CAD (≥70% luminal diameter stenosis or fractional flow reserve ≤ 0.80). Pertinent clinical characteristics and imaging criteria of all true-positive (n = 71), false-positive (n = 30), false-negative (n = 4), and true-negative patient-level CT diagnoses (n = 87) for relevant stenosis according to ICA were assessed. Results: In the univariate per-patient analysis, the following parameters yielded discriminative power (p < 0.10) regarding inaccurate CAD assessment by pre-TAVI-CT: age, atrial fibrillation, scanner generation, and image quality. Factors independently associated with CT diagnostic inaccuracy were determined using multivariable logistic regression analysis: a younger age (odds ratio [OR] 0.87; 95% confidence interval [CI] 0.80 to 0.94; p < 0.01) and insufficient CT image quality (OR 0.6; CI 0.41 to 0.89; p < 0.01). Conclusions: Our results demonstrate younger age and poor CT image quality to predict less accurate CAD assessments by pre-TAVI-CT in comparison with ICA. Knowledge of these predictors may aid in more efficient coronary artery interpretations based on pre-TAVI-CT.