State-of-the-art artificial intelligence methods for pre-operative planning of cardiothoracic surgery and interventions: a narrative review

Abstract

Background and Objective: Artificial intelligence (AI) has been increasingly explored as a tool to enhance clinical decision-making and optimize and speed up preoperative planning in cardiothoracic surgery. By improving precision and efficiency, AI has the potential to streamline workflows and improve outcomes. This study aimed to examine the current applications of AI in preoperative planning for cardiothoracic procedures. Methods: We systematically reviewed the literature in PubMed. Two search strings were employed to identify research articles related to AI applications in preoperative cardiothoracic surgery planning published up to August 2024. Studies were screened, and articles were included based on predefined criteria. Key Content and Findings: A total of 525 articles were extracted from the PubMed database. After applying exclusion criteria and analyzing the articles, 32 articles were included. These articles were categorized into and described according to their application: aortic (valve) surgery/intervention, mitral valve surgery/intervention, other cardiac surgeries, and lung, thoracic wall, and mediastinal surgeries. Key AI applications included segmentation of anatomical structures, tumor detection, prosthesis sizing for transcatheter aortic valve implantation (TAVI), and automated measurement of surgical parameters. The reviewed studies demonstrated that AI could increase segmentation accuracy, reduce preoperative planning time, and automate critical steps in surgical preparation. Conclusions: AI has been introduced in preoperative planning for cardiothoracic procedures to support clinicians by increasing segmentation accuracy, reducing preoperative planning time, and automating several preoperative planning steps such as tumor detection, TAVI prosthesis sizing and other planning measurements. However, the widespread adoption faces several challenges, including the need for robust validation, regulatory approval, and integration into clinical workflows. Additionally, the implementation of AI involves substantial costs, including investments in software development, computational infrastructure, and training of clinical staff. Future research should focus not only on advancing AI technology but also on evaluating the cost-effectiveness to ensure it delivers measurable benefits while remaining accessible and sustainable for healthcare systems. Addressing these issues is essential to realize the full potential of AI in cardiothoracic surgery.

Ähnliche Arbeiten