Adaptive Learning Algorithms for Low Dose Optimization in Coronary Arteries Angiography: A Comprehensive Review

Abstract

Abstract

Objective: Coronary artery angiography plays a pivotal role in cardiovascular disease diagnosis and treatment, but concerns regarding patient safety due to ionizing radiation necessitate innovative approaches. The article explores the integration of adaptive learning algorithms to optimize low-dose imaging in coronary artery angiography.

Method: Articles are selected on the basis of inclusion criteria that mentions studies in between the time span of 2018 to 2022 emphasizing the detailed algorithmic studies of Low dose optimization of coronary arteries angiography and techniques used in it, mentioned total 175 studies were included in initial studies that were reduced to ten final selected studies.

Results: The extracted data shows a comprehensive data on various techniques that are used for low dose CAA, advancements in image segmentation, noise reduction, and operator dose reduction highlight the potential of machine learning techniques. Innovative methods such as Model-Based Deep Learning (MBDL) and Self-Attention Generative Adversarial Networks (SAGAN) demonstrate efficient reconstruction capabilities. Application of such algorithms include automate segmentation, lesion detection, and real-time image analysis, optimizing dose parameters based on patient-specific factors, thus prioritizing patient safety and treatment effectiveness while revolutionizing medical imaging. Then there are possible limitations of the Algorithmic approach to reduce radiation dose for patient that include concerns include data heterogeneity, lack of diversity, and challenges in data collection and privacy protection. Addressing these limitations is crucial for enhancing the reliability of AI-driven dose optimization methods.

Conclusion: This comprehensive review provides valuable insights into the potential of adaptive learning algorithms for low-dose optimization in coronary artery angiography. It underscores the importance of safer imaging practices without compromising diagnostic efficacy. And the future lies in exploring adaptive learning algorithms, integrating patient-specific data, and real-time adaptability during procedures. Validation studies and collaboration with healthcare institutions are essential for successful integration into clinical practice.