Artificial Intelligence Based Multimodality Imaging: A New Frontier in Coronary Artery Disease Management

Abstract

Coronary artery disease (CAD) represents one of the most important causes of death around the world. Multimodality imaging plays a fundamental role in both diagnosis and risk stratification of acute and chronic CAD. For example, the role of Coronary Computed Tomography Angiography (CCTA) has become increasingly important to rule out CAD according to the latest guidelines. These changes and others will likely increase the request for appropriate imaging tests in the future. In this setting, artificial intelligence (AI) will play a pivotal role in echocardiography, CCTA, cardiac magnetic resonance and nuclear imaging, making multimodality imaging more efficient and reliable for clinicians, as well as more sustainable for healthcare systems. Furthermore, AI can assist clinicians in identifying early predictors of adverse outcome that human eyes cannot see in the fog of “big data.” AI algorithms applied to multimodality imaging will play a fundamental role in the management of patients with suspected or established CAD. This study aims to provide a comprehensive overview of current and future AI applications to the field of multimodality imaging of ischemic heart disease.

Artificial Intelligence as a Diagnostic Tool in Non-Invasive Imaging in the Assessment of Coronary Artery Disease

“…Secondly, there is no known randomized trial to date showing that improved phenotyping and personalized medicine based on this leads to better outcomes. At present, the majority of newly developed ML models are validated in single-centre studies with a restricted number of cases [ 14 ]. Early-stage evaluation studies are needed to investigate whether AI can lead to better care and outcomes whilst lowering costs [ 90 ].…”

“…Because of the nature of these datasets, information extracted can be skewed and unrepresentative of the wider population, thus resulting in considerable bias and lack of generalizability [ 15 ]. An example of this is the underrepresentation of ethnic minorities in datasets but also clinical trials [ 14 , 94 , 95 ]. This is a crucial problem, as it can lead to these groups potentially missing out on life-saving treatment and novel technology, with adverse outcomes [ 14 ].…”

“…An example of this is the underrepresentation of ethnic minorities in datasets but also clinical trials [ 14 , 94 , 95 ]. This is a crucial problem, as it can lead to these groups potentially missing out on life-saving treatment and novel technology, with adverse outcomes [ 14 ]. In fact, disparities in cardiovascular outcomes have been reported in the last few decades with regards to ethnic minorities [ 96 , 97 , 98 ].…”

“…In the nature of its complexity, multiple variables, and the large number of patients it affects annually, the field of coronary disease represents an ongoing minefield for AI development and is likely to be a key area where the impact of AI will be felt. Previous studies such as Maragna et al [ 14 ] and Gautam et al [ 15 ] have discussed the applications of AI in CAD, including an overview of multimodality cardiac imaging, with a discussion surrounding the strengths and pitfalls of this new technology. In this paper, we aim to review the current literature on applications of AI in diagnosing CAD, with an emphasis on non-invasive cardiac imaging modalities, and present a discussion on the use of AI and MRI in prognostication.…”

Non-invasive Multimodality Imaging of Coronary Vulnerable Patient

“…It is important that one keeps in mind that modalities scans are more than plain images; they are data. The analysis of such data using artificial intelligence is currently revolutionizing medical imaging ( 52 ). Big data include enormous numbers of predictors and outcomes with complex non-linear links, and conventional statistics usually fail to analysis them.…”