Using CiteSpace, we identified 12 core authors contributing 39 papers, which account for 33% of the total publication volume following Price's Law, indicating that the field has yet to form a stable group of authors. Average Citations per Item (ACI) serves as an indicator to measure the impact of scientific literature, commonly utilized to quantify the average number of citations received by a scholar, a journal, or an article, and Table 2 displays the highly productive authors in this field with more than three publications, ranked by the number of citations.

Schoepf U. Joseph leads with four papers and 36 citations, primarily focusing on AI and coronary artery CT technologies (9) for optimizing preoperative assessment (10) and decision-making (11) for patients with severe AVS (12). The collaborative network among core authors indicates strong partnerships, particularly between Schoepf and Varga-Szemes, showing their pivotal role in advancing AI research in the field. From Figures 3, 4, a series of highly interconnected author clusters are evident, signifying the presence of stable research teams and cross-institutional collaborative projects.

Similar to the core author analysis, identifying principal research institutions focuses on assessing the institutions' publication output, citation frequency, and collaborative interactions. The University of California, San Francisco, Brigham & Women's Hospital, and Massachusetts General Hospital are the top research institutions, with the University of California leading in both the number of publications and citation impact in Table 3. These institutions are at the forefront of applying AI to improve the diagnosis and treatment of AVS, and their collaborations have significantly contributed to research advancements in this field.

As shown in Table 5, the most cited article is “Deep Learning-Based Algorithm for Detecting Aortic Stenosis Using Electrocardiography,” where researchers demonstrated a deep learning-based algorithm capable of detecting severe Aortic Stenosis with high precision using 12-lead and single-lead electrocardiograms (ECG). The second most cited article developed an artificial intelligence-based electrocardiogram (AI-ECG) employing convolutional neural networks to identify patients with moderate to severe Aortic Stenosis. The results indicate that this AI-ECG exhibits high accuracy and has the potential to serve as a powerful screening tool for AVS in the community. These key studies demonstrate AI's transformative potential in AS diagnostics.

The most influential journals in this field are Scientific Reports and Journal of the American Society of Echocardiography evaluated through metrics like the Journal Impact Factor (JIF) as shown in Table 6. These journals are central to disseminating research on AI applications in AVS, playing a vital role in the academic community's understanding of the potential of AI in clinical settings.

The most frequently occurring keywords in the AI-AVS literature are “machine learning”, “aortic stenosis” and “artificial intelligence” in Table 7 signaling research focuses on diagnostic improvements and predictive models. Clusters of co-occurring keywords (Figures 6, 7) show that research is centered on AI applications in medical imaging, diagnosis, and risk prediction.

As illustrated in Table 8, keywords such as “machine learning,” “implantation,” “management,” “artificial intelligence,” and “aortic stenosis” stand out due to their high frequency. These discoveries offer a lucid perspective on the research trends and core themes prevalent in the field.

AI in cardiac imaging and disease classification, emphasizing machine learning techniques to enhance echocardiographic analysis and improve disease classification accuracy. “Machine learning” as a potent tool for data analysis, is being widely applied in the classification, diagnosis, quantitative analysis, and quantification of the impact of heart diseases. We observed that “machine learning” shares a cluster with keywords such as “calcification”, “classification”, “diagnosis”, “disease”, “dysfunction”, “echocardiography”, “heart failure”, “impact”, “quantification” and “recommendations” in the keyword co-occurrence network. This clustering indicates that the application of machine learning technology in the fields of heart calcification, disease classification, and diagnosis has garnered widespread attention (23, 24). Moreover, the co-occurrence of terms like “echocardiography”, “heart failure” and “dysfunction” further underscores the significance of machine learning in evaluating cardiac function and researching heart failure (25). For instance, in the detection and quantification of cardiac calcification, machine learning is capable of processing complex imaging data to reveal pathological features and their association with aortic stenosis. Additionally, “echocardiography” an essential diagnostic tool in cardiology, benefits from the application of machine learning, enhancing diagnostic accuracy, particularly in evaluating cardiac dysfunction and heart failure. Studies have shown that the data-driven insights provided by machine learning have a significant impact on forming treatment recommendations and clinical practice guidelines (19). In summary, “machine learning” serves as a nexus connecting several key research areas in cardiology, heralding a new era in cardiac research and practice. Through in-depth analysis of this cluster, researchers can better understand how machine learning optimizes the diagnostic and treatment pathways for heart diseases and how it leads the future trends in medical research.

TAVI and imaging-based clinical outcomes, focusing on the role of imaging technologies like computed tomography in evaluating outcomes of transcatheter valve implantation. Analysis of the keyword co-occurrence network reveals a distinct cluster closely associating “Implantation” with a series of related technologies and clinical outcomes in the domain of aortic stenosis, indicating that utilizing computed tomography (CT) for cardiac structural imaging segmentation, assessing ejection fraction, and predicting mortality and other clinical outcomes are current hotspots in research and clinical practice, especially when employing transcatheter aortic valve implantation (TAVI) strategies (9). The formation of this cluster not only emphasizes the central role of imaging in the pre and post-assessment of AVS treatment (26) but also highlights the importance of multidisciplinary approaches in improving patient prognosis (22). Thus, focusing on the literature within this domain is crucial for enhancing the success rate of AVS treatments and improving long-term patient outcomes. Further analysis shows that within this cluster, not only do the keywords co-occur frequently, but their mutual citations in the literature also prove their technological and conceptual interconnections and dependencies, together constituting a comprehensive and multidimensional research domain in AVS studies.

Management and screening of aortic stenosis, addressing guidelines, prevalence, and early detection strategies.The keyword co-occurrence analysis of literature on aortic stenosis management strategies has revealed a tight network of associations between the word “Management” and other keywords such as “accuracy”, “aortic-valve-replacement”, “association”, “prevalence”, “screening”, “society”, “stenosis” and “valvular heart disease”. This cluster reflects the comprehensive management needs of AVS patients in medical practice, including ensuring the precision of treatments and interventions (27), understanding the prevalence of the condition (28), enhancing early screening (29), and assessing the outcomes of cardiac valve replacement surgeries (3). Therefore, the research suggests that utilizing the keywords within this cluster can promote a multifaceted management strategy for AVS, which is crucial for improving clinical outcomes.

AI for risk prediction and personalized treatment, highlighting the use of predictive modeling to assess risks and validate AI frameworks in AVS management. In the modern treatment and research of aortic stenosis, artificial intelligence, as an innovative technology, is forming a tight interactive network with traditional clinical keywords such as “accuracy”, “screening” and “disease prevalence trends.” The growing application of artificial intelligence in medical image recognition, pathological prediction, and patient management, is demonstrating immense potential in the diagnosis and treatment of cardiac valve diseases, especially aortic stenosis. Moreover, the integration of artificial intelligence technology not only enhances the accuracy and efficiency of disease management (20) but also promotes data-driven decision-making in developing screening guidelines and treatment recommendations by public health organizations and professional associations. Consequently, artificial intelligence plays an increasingly important role in optimizing disease detection (30), therapeutic interventions (31, 32), and improving patient quality of life (33), signalling a significant transformation in the research and treatment methodologies of cardiac valve diseases and opening a new chapter in personalized medicine and precision treatment (34).

Deep learning for disease severity assessment and prognosis, exploring how deep learning is applied to evaluate cardiac function and AS severity. The application of artificial intelligence and deep learning technologies in the assessment of the severity of aortic stenosis pathology and treatment decisions is increasingly prevalent. This cluster reveals how researchers employ complex algorithms to deepen their understanding of aortic stenosis and other cardiac diseases (13), and develop tools capable of accurately assessing disease severity and predicting patient prognosis (35). This not only demonstrates the potential of deep learning in disease classification and prognosis prediction but also highlights the significant role of artificial intelligence in the diagnosis and treatment of cardiovascular diseases, especially against the backdrop of the rapid development of non-invasive diagnostic technologies (36–38). These interdisciplinary technological advancements offer new perspectives for personalized medicine in cardiac diseases and could drive the management of cardiac diseases towards more precise and efficient directions.