Related Experiment Video
Updated: May 3, 2026
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
Transformer and Attention-Based Architectures for Segmentation of Coronary Arterial Walls in Intravascular
Vandana Kumari1, Alok Katiyar1, Mrinalini Bhagawati2
1School of Computer Science and Engineering, Galgotias University, Greater Noida 201310, India.
Insights
Transformer models enhance coronary artery disease diagnosis by improving intravascular ultrasound (IVUS) wall segmentation. While UNet and transformer models show promise, addressing AI biases and integrating explainable AI (XAI) are crucial for clinical use.
Area of Science:
- Artificial Intelligence in Medical Imaging
- Cardiovascular Disease Diagnostics
- Biomedical Signal Processing
Background:
- Coronary artery disease (CAD) remains a leading global cause of mortality, underscoring the need for accurate diagnostic tools.
- Intravascular ultrasound (IVUS) offers detailed coronary artery visualization, but precise wall segmentation and plaque quantification methods are still developing.
- Current diagnostic approaches require enhancement for early and accurate detection of CAD.
Purpose of the Study:
- To explore the application of transformer models for improved wall segmentation in IVUS scans.
- To assess inherent biases within artificial intelligence (AI) systems used for IVUS scan analysis.
- To investigate the integration of explainable AI (XAI) into deep learning (DL) models for IVUS wall segmentation.
Main Methods:
- A systematic review using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework identified transformer-based coronary wall segmentation strategies.
- Analysis focused on evaluating the characteristics, scientific validity, and clinical relevance of identified methods.
- Biases in DL systems for IVUS wall segmentation were investigated, and XAI concepts were incorporated into DL structures.
Main Results:
- The UNet and transformer-based model demonstrated efficiency in segmenting coronary walls from IVUS scans due to its multi-scale feature extraction and image reconstruction capabilities.
- This approach effectively utilizes encoders, decoders, and skip connections for robust segmentation.
- The study identified the UNet and transformer-based model as a promising technique for IVUS analysis.
Conclusions:
- A notable lack of incentives exists for adopting explainable AI (XAI) and pruned AI (PAI) models in clinical practice.
- No current UNet-based systems have achieved a bias-free configuration for IVUS analysis.
- Transitioning from theoretical research to practical clinical evaluation and deployment is essential for advancing CAD diagnostics.
Abstract:
Background: The leading global cause of death is coronary artery disease (CAD), necessitating early and precise diagnosis. Intravascular ultrasound (IVUS) is a sophisticated imaging technique that provides detailed visualization of coronary arteries. However, the methods for segmenting walls in the IVUS scan into internal wall structures and quantifying plaque are still evolving. This study explores the use of transformers and attention-based models to improve diagnostic accuracy for wall segmentation in IVUS scans. Thus, the objective is to explore the application of transformer models for wall segmentation in IVUS scans to assess their inherent biases in artificial intelligence systems for improving diagnostic accuracy. Methods: By employing the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework, we pinpointed and examined the top strategies for coronary wall segmentation using transformer-based techniques, assessing their traits, scientific soundness, and clinical relevancy. Coronary artery wall thickness is determined by using the boundaries (inner: lumen-intima and outer: media-adventitia) through cross-sectional IVUS scans. Additionally, it is the first to investigate biases in deep learning (DL) systems that are associated with IVUS scan wall segmentation. Finally, the study incorporates explainable AI (XAI) concepts into the DL structure for IVUS scan wall segmentation. Findings: Because of its capacity to automatically extract features at numerous scales in encoders, rebuild segmented pictures via decoders, and fuse variations through skip connections, the UNet and transformer-based model stands out as an efficient technique for segmenting coronary walls in IVUS scans. Conclusions: The investigation underscores a deficiency in incentives for embracing XAI and pruned AI (PAI) models, with no UNet systems attaining a bias-free configuration. Shifting from theoretical study to practical usage is crucial to bolstering clinical evaluation and deployment.
