A recently published review comprehensively discusses the impact of artificial intelligence (AI) on clinical practice in interventional cardiology (IC), with a focus on recent advances.
Although the development of AI is still in its infancy, new technologies promise significant improvements in patient safety, risk stratification and therapeutic outcomes. Key goals include the integration of multiple cardiac imaging modalities, the establishment of online decision support systems and the creation of automated medical systems to provide electronic health records. The use of AI in IC can be divided into two main areas: virtual (medical imaging, decision making) and physical (robotic interventional procedures). Numerous studies have demonstrated the potential of AI in automatically interpreting and analyzing various cardiac modalities, which significantly improves the therapeutic process.
Artificial intelligence (AI), in particular machine learning (ML), enables real-time processing and analysis of large amounts of medical data and will revolutionize the healthcare system. AI is developing rapidly, particularly in the field of cardiology, from electrocardiogram (ECG) interpretation to clinical decision support systems for cardiology interventions. The majority of AI/ML-based devices approved by the US Food and Drug Administration (FDA) are related to radiology and cardiology. These devices enable cardiologists to implement a complex approach to heart disease by supporting early diagnosis, patient risk stratification prior to targeted interventions, and overall improvement in quality of care. The use of AI in IC covers every step of the therapeutic process, including initial assessment of chest pain and/or cardiogenic shock in the hospital, planning the intervention strategy for better navigation and guidance, and predicting patient risk and potential outcomes. The specific nature of IC provides clinicians with many imaging modalities, including anatomic and functional assessments of structural heart disease. Therefore, AI is seen as a promising technological tool that will have a significant impact on image reconstruction, analysis and interpretation, leading to an improvement in the availability and quality of health data and further advances in analytical techniques.
Methodology of the review
The methodology of this systematic review is based on the PRISMA statement. Recent publications, reports, protocols and reviews from the Scopus and Web of Science databases were considered. The keywords “artificial intelligence, machine learning, augmented reality, mixed reality, virtual reality, metaverse, cardiology, interventional cardiology, segmentation, segmentation algorithms, classification algorithms, ethics, AI ethics” and their variations were identified. In the first step, characteristics of the material such as title and abstract were evaluated, taking into account exclusion criteria (e.g. dissertations and non-cardiologically relevant material were removed, while full-text articles in English were considered). Subsequently, articles and technical reports that met the criteria were retrieved and analyzed. A total of 100 documents were considered.
Application of artificial intelligence
Artificial neural networks (ANNs): ANNs are interconnected nodes that model biological neurons as weights between nodes. They improve the diagnosis and treatment of cardiovascular diseases by automating the analysis of echocardiography and cardiac CT images, which increases accuracy and reduces detection time. ANNs learn from large amounts of data and predict outcomes based on patterns, which significantly helps in the early detection of diseases. Despite their advantages, ANNs face challenges such as the need for large training data and the risk of overfitting. They have been successfully used in the automated measurement of ejection fraction and left ventricular longitudinal strain with high accuracy and the differentiation of hypertrophic cardiomyopathy from physiological hypertrophy.
Recurrent Neural Networks (RNNs): RNNs manage and interpret sequential data such as ECG recordings and continuous health monitoring of patients, predict intervention outcomes and help plan effective treatments. RNN-based solutions such as DeepHeart predict cardiovascular risks using data from wearable devices. RNNs also automate the selection of myocardial inversion time, making the diagnostic process more efficient.
Convolutional Neural Networks (CNNs): CNNs process complex cardiovascular images and improve diagnostic accuracy and treatment effectiveness. They are crucial for analyzing angiograms and echocardiograms and identifying heart disease patterns. CNNs have demonstrated success in transcatheter aortic valve implantation, echocardiogram classification and ventricular segmentation and have had a significant impact on medical training and activity recognition of surgeons.
Spiking Neural Networks (SNNs): SNNs, brain-inspired networks for analyzing dynamic data and time-dependent information, are particularly effective in analyzing ECG signals. They identify subtle anomalies for early arrhythmia detection and support rapid interventions. The precision of SNNs in classifying heartbeats and detecting extra ventricular beats underlines their adaptability in different clinical contexts.
Deep Neural Networks (DNNs): DNNs, with multiple layers between input and output, decipher complex patterns in large data sets and are indispensable tools in modern medical analysis. They recognize subtle patterns in diagnostic images and patient records and improve the assessment of cardiology interventions. DNN-based methods accurately predict multiple medical events, assess the severity of coronary artery stenoses and improve diagnostic quality by generating new data and reducing noise in CT images.
Ethical implications of AI in interventional cardiology
The use of AI in cardiac interventions requires rigorous ethical scrutiny, taking into account institutional norms and detailed ethical practices. Cardiac interventions, often in life-threatening situations, present ethical dilemmas regarding resuscitation and legal implications. AI recommendations increase the complexity of liability and decision-making. Emerging digital twin technology, representing physical systems in real time, promises to analyze complex data sets and suggest treatment pathways. Issues of ownership, control and decision making regarding digital twins need to be clarified, requiring accessible ethical protocols.
The regulation of AI in healthcare is evolving, but focuses on Explainable AI (XAI) and Trustworthy AI (TAI). National and international regulations, such as the European Assessment List for Trustworthy Artificial Intelligence (ALTAI), are trying to keep pace with rapid technological developments. Ethical and technical considerations are crucial for the integration of AI into IC.
Future approach: Augmented reality and 3D visualization supported by AI
The integration of AI into immersive technologies is crucial for handling complex medical data and 3D representations. Current developments enable 3D reconstruction of organs, which is of great importance for clinical practice and training. AI-driven simulations will train interventional cardiologists in a safe environment and improve their skills. Immersive technologies, combined with AI, will facilitate remote multidisciplinary cardiac team meetings, overcoming geographical barriers and improving healthcare delivery.
Discussion and conclusions
The transformative role of AI in IC improves diagnostic accuracy, treatment outcomes, remote monitoring and training. AI-driven robotic systems support precise movements during procedures, improve outcomes and reduce physician fatigue. However, limitations such as data dependency and lack of transparency must be considered. Ethical considerations and an understanding of AI mechanisms are essential for effective integration into the healthcare system. The combination of AI and IC promises to increase the efficiency and accuracy of cardiovascular imaging while reducing costs. Despite challenges in full integration and clinical application, AI offers enormous potential to improve healthcare. Future developments in AI-driven simulations and immersive technologies will revolutionize IC, provide personalized, interactive and efficient solutions, and ultimately transform cardiology and improve healthcare.
Quelle: Rudnicka Z, Pręgowska A, Glądys K, et al.: Advancements in artificial intelligence-driven techniques for interventional cardiology. Cardiol J 2024; 31(2): 321–341. doi: 10.5603/cj.98650. Epub 2024 Jan 22. PMID: 38247435; PMCID: PMC11076027.
CARDIOVASC 2024; 23(2): 29–30
