Key result
Deep learning noise reduction boosts coronary CT image quality while preserving CAD diagnostic accuracy at ~90% AUC.
Why the study?
To evaluate the impact of a deep learning-based noise reduction technique on image quality and diagnostic accuracy for coronary artery assessment in pre-TAVI CT imaging.
Does a deep learning-based noise reduction technique improve image quality and diagnostic accuracy for evaluating coronary arteries in patients with severe aortic stenosis undergoing pre-TAVI CT scans?
Observational (n=200)
Does a deep learning-based noise reduction technique improve image quality and diagnostic accuracy for evaluating coronary arteries in patients with severe aortic stenosis undergoing pre-TAVI CT scans?
Effect estimate: AUC 90% (95% CI 88.5-91.3)
p-value: p=0.056
Deep learning denoising substantially improves CT image quality for coronary artery assessment in pre-TAVI patients without affecting diagnostic accuracy.
RATIONALE AND OBJECTIVES: To assess the impact of a deep learning-based noise reduction (DLD) technique on image quality and diagnostic accuracy for the evaluation of coronary arteries in transcatheter aortic valve implantation (TAVI) CT imaging. MATERIALS AND METHODS: Two hundred patients with severe aortic stenosis who underwent CT scans for pre-TAVI planning between October 2022 and April 2024 were retrospectively enrolled. Conventional images were reconstructed and denoised images were generated using dedicated software. Objective image quality was evaluated by measuring the mean Hounsfield unit (HU) and standard deviation (SD) in regions of interest within the aortic root, coronary arteries, and subcutaneous fat to calculate signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR). For subjective assessment, two readers used a 5-point Likert scoring system to evaluate sharpness, noise, contrast and overall image quality. The diagnostic performance of both datasets was assessed using invasive coronary angiography as reference standard. RESULTS: Denoised reconstructions showed significantly higher SNR (37.5±12.8 vs.12.3±4.1) and CNR (45.3±15.4 vs. 14.7±4.4), and lower noise (16.9±7.9 vs. 47.9±11.6 HU) (all p<0.001). Subjective assessment demonstrated that denoised images received the highest score for sharpness, noise, contrast and overall image quality (all p<0.001). For the evaluation of diagnostic accuracy, a total of 800 vessels and 1787 segments were analyzed. The per-segment diagnostic performance of the DLD for detection of CAD revealed an AUC of 90% (95% CI: 88.5-91.3), with accuracy of 93.9% (95% CI: 92.7-95), 85.7% (95% CI: 78.7-90.4) sensitivity and 94.7% (95% CI: 93.5-95.7) specificity, in the absence of a statistically significant difference compared with the evaluation performed on standard images (p=0.056). CONCLUSION: The DLD substantially improves image quality without affecting diagnostic accuracy for the evaluation of coronary arteries in patients undergoing pre-TAVI CT scans.
No takes yet. Share an insight, caveat, or question.
Lanzafame et al. (2025) conducted an observational in Severe aortic stenosis (n=200). Deep learning-based noise reduction (DLD) technique vs. Conventional standard images was evaluated on Per-segment diagnostic performance for detection of CAD (AUC) (AUC 90%, 95% CI 88.5-91.3, p=0.056). Deep learning-based noise reduction significantly improved image quality (SNR 37.5 vs 12.3; P<0.001) without compromising diagnostic accuracy for CAD detection (AUC 90%; P=0.056 vs standard images).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: