Non-invasive derivation of instantaneous free-wave ratio from invasive coronary angiography using a new deep learning artificial intelligence model and comparison with human operators’ performance
See where this sits in the topic map →Summary AI-generated
- TL;DR
- This study develops and evaluates a deep learning model to non-invasively classify coronary lesions using standard angiography, comparing its performance against experienced human operators.
- Problem
- Invasive coronary physiology assessments are underused, costly, and carry patient risks, while non-invasive alternatives derived from angiography—particularly for the instantaneous wave-free ratio (iFR)—remain rarely explored.
- Method
- Using a single-center retrospective dataset, the researchers trained a validated encoder-decoder convolutional neural network to segment vessels and tested three AI models combining transformers and EfficientNet architectures to classify lesions as positive or negative.
- Results
- Across 250 measurements, the unified AI model achieved an overall accuracy of 72% and a high negative predictive value of 90%, significantly outperforming human operators in the left anterior descending artery (78% vs. 60–64%).
- Contributions
- The study successfully unifies specialized vessel-based AI models—combining transformer and convolutional neural network approaches—for binary iFR lesion classification using ten-fold patient-level cross-validation.
- Limitations
- Not specified in the abstract.
- Takeaways
- The proposed AI model mildly outperformed interventional cardiologists in binary iFR lesion classification, supporting the need for further validation studies.
- Applications
- Non-invasive physiological assessment of coronary lesions from standard invasive coronary angiography.
- Topics
- Healthcare, Artificial Intelligence, Deep Learning, Computer Vision, Cardiovascular Imaging
- For industry
- Healthcare and Medical Technology
- Why it matters
- Demonstrating that AI can match or outperform cardiologists in non-invasive coronary physiology assessment could help reduce the risks and costs associated with invasive procedures.
Abstract
Invasive coronary physiology is underused and carries risks/costs. Artificial Intelligence (AI) might enable non-invasive physiology from invasive coronary angiography (CAG), possibly outperforming humans, but has seldom been explored, especially for instantaneous wave-free Ratio (iFR). We aimed to develop binary iFR lesion classification AI models and compare them with human performance. single-center retrospective study of patients undergoing CAG and iFR. A validated encoder-decoder convolutional neural network (CNN) performed segmentation. Manual annotation of target vessel and pressure sensor location on a segmented telediastolic frame followed. Three AI models classified lesions as positive (≤ 0.89) or negative (> 0.89). Model 1 uses preprocessed vessel diameters with a transformer. Models 2/3 are EfficientNet-B5 CNNs using concatenated angiography and segmentation - Model 3 employs class-frequency-weighted Cross-Entropy Loss. Previous findings demonstrated Model 3's superiority for left anterior descending (LAD) and Model 1's for circumflex (Cx)/right coronary artery (RCA) - they were therefore unified into a vessel-based model. Ten-fold patient-level cross-validation enabled full sample training/testing. Three experienced operators performed binary iFR classification using single frames of raw/segmented images. Comparison metrics were accuracy, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). Across 250 measurements, AI accuracy was 72%, PPV 48%, NPV 90%, sensitivity 77%, and specificity 71%. Human accuracy ranged from 54 to 74%. NPV was high for the Cx/RCA (AI: 96/98%; operators: 94/97%), but AI significantly outperformed humans in the LAD (78% vs. 60-64%). An AI model capable of binary iFR lesions classification mildly outperformed interventional cardiologists, supporting further validation studies.