The role of artificial intelligence in the minimally invasive thoracic surgery: narrative review of the last 15 years

Introduction

Nowadays, artificial intelligence (AI) is quickly changing every aspect of everyday life. In particular, its application in the medical field is increasing day by day, supporting medical staff in diagnosis and treatment processes, with an active role in patient management and in the prediction of adverse events. AI can also help to optimize the allocation of healthcare resources and provide better resource management by reducing waste.

Machine learning (ML), a subset of AI, utilizes algorithms that iteratively minimize the error between predicted and actual outcomes, thereby enhancing model performance. Deep learning (DL), an advanced branch of ML, employs multilayer artificial neural networks capable of processing complex, high-dimensional data. Fully connected neural networks (FCNNs), based on back-propagation, form one such architecture and are known for their predictive accuracy (Figure 1).

Figure 1 Diagram showing the relationship between AI and its subsets: types of AI implementation and their area of use. Level 1 Machine Learning: autonomous growth of the AI based on the errors it produces. Level 2 Convolutional Neural Networks: growth of AI based on expanded databases. Level 3 Deep Learning: growth of AI based on multi-layered networks. AI, artificial intelligence; ICU, intensive care unit.

The application of AI in surgery is an exciting and rapidly advancing field that extends well beyond robotic surgery (1). In this scenario, many researchers are carefully testing its potential in various areas, such as pre-operative planning, intra-operative assistance and postoperative care (2). Although robotic surgery has been a primary focus, the utilization of AI in minimally invasive thoracic surgery (MITS) is a developing area of interest, with initial reports showing promising results. This technology not only aids surgical trainees in honing their skills on robotic systems—particularly beneficial for those who may face challenges—but also enhances the accuracy of diagnosing non-small-cell lung cancer (NSCLC) and predicting lymph node metastasis. Although AI has a significant role in simplifying the extraction of important data from electronic medical records (EMRs), it is essential to carefully consider other important issues such as data privacy, ethical responsibilities and algorithmic biases (3). Our aim is to give a general overview of the current application of the AI in thoracic surgery, therefore, in this narrative review, we analyzed the state-of-the-art regarding AI application in thoracic surgery, with particular focus on its role in clinical practice of MITS. We focused on general principles of AI in MITS as well as it roles in five specific domains: (I) radiologic and intraoperative evaluation; (II) localization and three-dimensional (3D) reconstruction; (III) high-risk patient evaluation; (IV) surgical training; (V) AI and robotic-assisted thoracic surgery. We didn’t address topics in AI fields such as data bias, cybersecurity, and surgeon autonomy. We present this article in accordance with the Narrative Review reporting checklist (available at https://vats.amegroups.com/article/view/10.21037/vats-25-36/rc).

Methods

Literature research for this scoping narrative review was conducted using the PubMed database (Table 1). We employed the following Medical Subject Headings (MeSH Terms) and keywords: “Artificial Intelligence”, “AI”, “Computing Methodologies”, “Thoracic Surgery”, “Video-Assisted Thoracoscopy”, and “Robotic Surgical Procedures”.

To ensure a rigorous and transparent methodology, the article selection process, although for a narrative review, followed the key principles outlined in the Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines. Specifically, the MOOSE checklist elements pertaining to search strategy, study selection, and data extraction were adapted and applied. This application helped standardize the inclusion process and mitigate potential selection bias.

The following inclusion criteria were applied: original and review studies published in the English language that specifically address the application of AI within the field of thoracic surgery.

The corresponding exclusion criteria were: non-English language papers, papers related to Thoracic Surgery but not directly related to the application of AI. papers published outside the defined date range of 1 January 2010 to 1 April 2025; this specific temporal criterion was set to capture the most recent advancements in the field.

The selection process was performed in three stages: title screening, followed by abstract evaluation, and finally, full-text analyses for definitive inclusion. All relevant studies were independently analyzed by two authors (C.U., G.L.). An interactive peer-review process among the panel was used to summarize the findings. In cases of disagreement between the two primary reviewers regarding a study’s inclusion, a third reviewer (P.S.) was consulted to ensure consistency and resolve discrepancies. A total of 603 articles were initially identified. After applying the selection criteria and following the MOOSE-guided process, 68 articles were assessed for eligibility. Ultimately, 38 articles were selected for the final analysis based on their specificity and direct relevance to the research topic (Table 2).

General principles

AI encompasses different technologies and methodologies that can significantly impact surgical practices in general and, specifically, MITS as well. One core principle is the ability of AI to process and interpret vast amounts of data, often exceeding human cognitive capabilities (4). As highlighted by Constable et al., this includes medical images, patient records and real-time physiological data (3).

ML, a key component of AI, enables systems to learn from this data without explicit programming. Supervised learning, for instance, as reported by Teixeira et al., trains models on labelled datasets to predict outcomes, while unsupervised learning identifies patterns in unlabelled data (5). DL, a more advanced kind of ML that use multi-layered neural networks, is particularly adept at handling complex, high-dimensional data such as medical images and videos, crucial in surgical contexts (5,6). These networks can identify subtle patterns and features that could be missed by human observation, leading to more accurate diagnoses and predictions (2).

In addition, AI algorithms can give real-time help during surgery (7). This help includes identifying anatomical structures, pointing out dangers, guiding tools, and checking how well surgery is done (1,2). As more data is gathered, these algorithms get better over time. Using AI in surgery also needs strong data systems and powerful computers (4). Collecting, storing and processing data safely and efficiently is key for AI to work well. Yu et al. noted that special computing power, like cloud services and graphics processing unit (GPUs), are needed to be used for increasingly complex AI (4).

Finally, the interpretability and explainability of AI models are critical considerations in clinical applications. In this setting, understanding the process a model used to arrive at a particular conclusion is crucial for clinicians to trust and effectively utilize these technologies in patient care (2). Specifically, research is ongoing to develop more transparent AI models, particularly in high-stakes fields like surgery (3).

Preoperative assessment in MITS

Preoperative assessment is a critical phase in MITS. In this setting AI helps in improving radiological evaluation, facilitating precise localization and 3D reconstruction and optimizing the assessment of high-risk patients.

Pre- and intra-operative radiological evaluation

AI, particularly DL, plays a significant role in enhancing the interpretation of radiological images for thoracic surgery before and during the operation. Convolutional neural networks (CNNs) and vision transformers (ViTs) can be used on large datasets of medical images to detect and characterize lung nodules, assess tumor invasion, and identify other critical anatomical features. For instance, Shimada et al. developed DL models to predict visceral pleural invasion in clinical stage I lung adenocarcinoma by using thoracoscopic images, showing promising performance (8). These models can identify subtle visual cues, such as white color tones or fibrous changes, that indicate tumor invasion, potentially improving preoperative staging and surgical planning. This information is also valuable during the surgical procedure to better plan some aspects of the operation. The ability of AI to analyze vast numbers of images and learn complex patterns can aid radiologists (and surgeons) in making more accurate and consistent diagnoses, reducing inter-observer variability.

Localization and 3D reconstruction prior surgery

Precise localization of lesions and detailed 3D reconstruction of thoracic anatomy are paramount for safe and effective MITS. AI algorithms can process CT scans and other imaging data to create highly accurate 3D models of the lungs, bronchi, vessels and tumors as well as the segmental planes. This allows surgeons to visualize the spatial relationship between structures, plan resection margins and anticipate potential challenges before entering the operating room. For instance, Ichinose et al. demonstrated that AI-powered surgical support systems can assist in real-time thoracic nerves (phrenic and vagus) recognition during lung cancer surgery, enhancing safety by visualizing important microanatomies (9). Compared to a “classic” MITS monitor, these AI systems demonstrate higher accuracy, with minimal time lag and good image quality, aiding surgeons in navigating complex anatomical regions (9). These advancements facilitate more precise surgical approaches, especially for small or deeply located lesions that are difficult to palpate.

High-risk patients evaluation

Identifying high-risk patients in the preoperative setting is crucial for reducing complications and optimizing outcomes in thoracic surgery. AI models can analyze a wide array of patient data, including clinical history, laboratory results, respiratory function tests and radiological findings, to predict the likelihood of postoperative complications. For example, Zhou et al. utilized ML and DL algorithms to construct early prediction models for post-operative pulmonary complications after thoracoscopic surgery. These models identified factors such as single lung ventilation duration, history of smoking, surgery duration, ASA score, and blood glucose as significant predictors (10). Similarly, Özçıbık İşık et al. showed that AI models, particularly FCNN, demonstrated high accuracy in predicting the need for intensive care unit (ICU) admission after NSCLC surgery, considering clinical, laboratory, respiratory, tumor’s radiological and surgical features (36). While study by Alnajjar et al. on predicting hospital stay for conditions like pleural empyema with AI models has shown mixed results, highlighting the need for further refinement (11), the overall potential of AI to personalize risk assessment and guide preoperative optimization strategies is significant. AI can help stratify patients, identify those who would benefit most from pre-habilitation programs, as suggested by Dennis et al. (37), and allow for more tailored surgical planning, ultimately aiming to improve patient safety and reduce adverse events.

Surgical training

AI is upgrading the training of young surgeons by providing personalized feedback and enhancing the acquisition of “standard” as well as complex surgical skills. Traditional surgical training often relies on subjective evaluations by experienced surgeons, whereas AI-powered systems offer a more standardized and data-driven approach.

One key application is the automated assessment of surgical performance. As demonstrated by Cavallo et al. (29) and Hashemi et al. (12), AI can analyze video recordings of surgical procedures to identify and quantify various metrics related to technical skill, such as instrument movements, task completion time, and efficiency. DL, particularly CNNs combined with long short-term memory (LSTM) layers, can be used to extract spatial and temporal features from video recordings, facilitating precise action recognition and skill assessment in robot-assisted thoracic surgery (RATS) (12). This allows for objective evaluation and identification of areas where trainees need improvement, even for subtle nuances that might be missed by human observers (30).

AI-based video feedback systems have shown promise in improving novice performance on robotic suturing skills. Ma et al. conducted a pilot study demonstrating that trainees who receive AI-generated feedback improve significantly compared to those without such feedback (31). This automated feedback can include real-time alerts or post-procedure summaries highlighting errors, inefficient movements or deviations from optimal technique (31).

Furthermore, AI can be used to obtain accreditation through the standardization of training in robotic surgery (30). This includes evaluating technical skills across various procedures and settings. The development of intelligent tutors and virtual reality (VR) simulations augmented with AI can create highly realistic and interactive training environments, allowing trainees to practice complex procedures repeatedly without patient risk. As a result, these systems can check trainee progress, adapt difficulty levels and provide individualized learning paths, ultimately leading to more competent and confident surgeons.

AI and robotic-assisted thoracic surgery

The association between AI and robotic surgery represents a transformative frontier in MITS, promising to elevate precision, enhance surgeon capabilities and optimize patient outcomes (13). Robotic surgical systems, such as the widely known da Vinci platform, already offer superior dexterity, enhanced 3D visualization and tremor filtration compared to conventional laparoscopic or thoracoscopic approaches (14). The integration of AI during surgery improves these capabilities by assisting surgeons in real-time and automating certain surgical steps as well (15,16).

Learning curve

One of the most significant applications of AI in RATS is related to improvement in surgical training and skills evaluation. As a rule, traditional methods to analyze surgical proficiency are often subjective and resource-intensive, relying on expert supervision (17,18).

Two studies by Hashemi et al. and Vasey et al. have demonstrated that AI-driven tools can analyze surgical gestures, recognize specific actions and provide quantitative feedback on performance, distinguishing between novice and experienced surgeons (17,19). Specifically, these authors developed a DL-based automated skills assessment tool for robotic surgery using CNNs combined with LSTM layers, showing promising results in analyzing spatial and temporal features from video recordings in porcine models (31). Similarly, the potential of AI to generate and deliver highly specialized intra-operative surgical feedback holds immense promise for accelerating the learning curve for new surgeons and maintaining the proficiency of experienced ones (14,20). In this scenario, Moglia et al. conducted a systematic review highlighting the significant potential of AI in RATS, particularly in improving patient safety, though emphasizing the need for more evidence on its direct efficacy (21).

Intraoperative decision making

Beyond training, AI plays a crucial role in enhancing intra-operative guidance and decision-making. By processing vast amounts of preoperative imaging data (e.g., CT, MRI) and combining it with real-time intra-operative video, AI algorithms can create augmented reality overlays, providing surgeons with crucial anatomical information, tumor boundaries and safe dissection planes (13,14). This can lead to more precise resections, especially in complex thoracic procedures where critical structures are in proximity (nerves, etc.). For example, Furube et al. demonstrated that AI can assist in identifying the recurrent laryngeal nerve during RATS esophagectomy, potentially reducing the risk of vocal cord paralysis (22). The ability of AI to detect positive surgical margins in real-time is another area of active research, which could significantly impact oncological outcomes by minimizing the need for re-surgery (14).

Technical accuracy

The progression towards increased autonomy in robotic surgery is heavily reliant on AI. While fully autonomous surgical robots are still in their birth stages and raise significant ethical and regulatory questions (15,23,24), AI is already enabling various levels of autonomy, from intelligent instrument control to semi-autonomous task execution. Liu et al. introduced an innovative method for gesture recognition in medical robotics using CNNs and a Modified Spring Search Algorithm (MSSA), demonstrating enhanced precision and efficiency in robot control (24). These advancements lay the groundwork for robots that can perform repetitive or technically challenging sub-tasks with high accuracy, reducing surgeon fatigue and potentially improving consistency (14,16). The development of systems capable of analyzing surgical scenes and providing intelligent assistance requires robust data annotation and robust ML models, as explored by De Backer et al. (25).

Collaborative and continuous efforts between clinicians, engineers and ethicists will be of paramount importance to unlock the full potential of AI in RATS safely and effectively (such as technical skills, accuracy of procedure, reducing of complication, time of operation, planning of operation, patient´s selection).

Ethical considerations

Integrating AI into surgery presents significant ethical, legal, and financial considerations (26,27,38). A core concern is liability for errors, particularly with “black box” AI systems whose decision-making processes are opaque (28,39). Determining accountability (liability, culpability) for autonomous robotic surgical errors is a major challenge for current legal frameworks, though a “doctor-in-the-loop” model is often advocated to ensure patient safety (32).

Specifically, critical concerns in this setting are data privacy, cybersecurity as well as algorithmic biases (33). Errors and inaccuracies within datasets can led to significant consequences, such as the creation of biased algorithms that, in turn, risk exacerbating existing health disparities; thus, it essential to thoroughly assess data quality.

In the setting of cybersecurity, an important issue is the possible hacking of surgical robots that require critical attention and future considerations for improving healthcare defense (34).

Furthermore, it is essential to thoughtfully evaluate the future role of surgeons that use AI and human judgment in general (34,35). Since AI lacks ethical judgment, clear ethical guidelines are needed right now to foster trust among healthcare professionals. By collaborating effectively, we can leverage the advantages of AI while preserving the “humanity” that makes healthcare uniquely meaningful.

Moreover, each hospital center and each nation have its own regulations and there is no international standardization in term of cyber security, data manager and privacy. Obviously, sooner or later we hope for the advent of international regulation on AI in the medical field, especially regarding sensitive data and patient privacy.

Strengths and limitations

The purpose of this paper is to provide the reader with a current review of the literature regarding the role of AI in thoracic surgery, with a focus on minimally invasive techniques. Our review was not intended to be a systematic review of the literature; therefore, a limitation of our study is the subjective nature of the evaluation of clinical studies for inclusion. Furthermore, only certain aspects of the role of AI in thoracic surgery were assessed, and important aspects such as cancer prevention were not considered.

A further limitation of the study is the total inhomogeneity of the literature on this topic. Therefore, in the next future, better to understand the effective role of AI in thoracic surgery, we need to focus on creating more transparent AI models, establishing clear regulatory frameworks and ensuring AI integrates smoothly into existing surgical workflows. This will require strong collaboration among clinicians, engineers, and ethicists to ensure AI is used safely and effectively.

Conclusions

It is clear to the entire scientific community that AI has clear negative and positive aspects, as summarized in Table 3.

Further conclusions could be identified after our overview:

• AI models are significantly improving thoracic surgery (particularly MITS) by enhancing preoperative assessment, surgical training and RATS. Specifically, AI improves radiological evaluation, helps in localizing lung lesions and reconstructs 3D anatomy, aiding accurate diagnosis and planning. It also assesses high-risk patients, predicting complications and optimizing outcomes.
• AI is continuously improving surgical training of young surgeons by providing objective and personalized feedback. Furthermore, when combined with robotic systems, it takes surgical precision and dexterity to an improved level, offering advanced features like augmented reality overlays and intelligent instrument control.
• We need to carefully address ethical concerns such as data privacy, who is liable if AI makes a mistake, algorithmic bias and how AI might affect a surgeon’s autonomy.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the Narrative Review reporting checklist. Available at https://vats.amegroups.com/article/view/10.21037/vats-25-36/rc

Peer Review File: Available at https://vats.amegroups.com/article/view/10.21037/vats-25-36/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://vats.amegroups.com/article/view/10.21037/vats-25-36/coif). A.P. declares serving as advisory boards, receiving consultancies, travel accommodations, speaker fees, writing fees, and PI role in profit trials: Janssen, Bayer, Merck Sharp & Dohme (MSD), Eli Lilly& Co., Bristol Myers Squibb (BMS), AstraZeneca. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was approved in accordance with the Vorarlberg Ethical Commission rules, Austria.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Nwoye E, Woo WL, Gao B, et al. Artificial Intelligence for Emerging Technology in Surgery: Systematic Review and Validation. IEEE Rev Biomed Eng 2023;16:241-59. [Crossref] [PubMed]
2. Aleem MU, Khan JA, Younes A, et al. Enhancing Thoracic Surgery with AI: A Review of Current Practices and Emerging Trends. Curr Oncol 2024;31:6232-44. [Crossref] [PubMed]
3. Constable MD, Shum HPH, Clark S. Enhancing surgical performance in cardiothoracic surgery with innovations from computer vision and artificial intelligence: a narrative review. J Cardiothorac Surg 2024;19:94. [Crossref] [PubMed]
4. Yu KH, Beam AL, Kohane IS. Artificial intelligence in healthcare. Nat Biomed Eng 2018;2:719-31. [Crossref] [PubMed]
5. Teixeira J. One Hundred Years of Evolution in Surgery: From Asepsis to Artificial Intelligence. Surg Clin North Am 2020;100:xv-xvi. [Crossref] [PubMed]
6. Chand M, Ramachandran N, Stoyanov D, et al. Robotics, artificial intelligence and distributed ledgers in surgery: data is key! Tech Coloproctol 2018;22:645-8. [Crossref] [PubMed]
7. Sone K, Tanimoto S, Toyohara Y, et al. Evolution of a surgical system using deep learning in minimally invasive surgery Biomed Rep 2023;19:45. (Review). [Crossref] [PubMed]
8. Shimada Y, Ojima T, Takaoka Y, et al. Prediction of visceral pleural invasion of clinical stage I lung adenocarcinoma using thoracoscopic images and deep learning. Surg Today 2024;54:540-50. [Crossref] [PubMed]
9. Ichinose J, Kobayashi N, Fukata K, et al. Accuracy of thoracic nerves recognition for surgical support system using artificial intelligence. Sci Rep 2024;14:18329. [Crossref] [PubMed]
10. Zhou CM, Xue Q, Li H, et al. A predictive model for post-thoracoscopic surgery pulmonary complications based on the PBNN algorithm. Sci Rep 2024;14:7035. [Crossref] [PubMed]
11. Alnajjar I, Alshakarnah B, AbuShaikha T, et al. Assessing artificial intelligence ability in predicting hospitalization duration for pleural empyema patients managed with uniportal video-assisted thoracoscopic surgery: a retrospective observational study. BMC Surg 2025;25:218. [Crossref] [PubMed]
12. Hashemi N, Mose M, Østergaard LR, et al. Video-based robotic surgical action recognition and skills assessment on porcine models using deep learning. Surg Endosc 2025;39:1709-19. [Crossref] [PubMed]
13. Zhang C, Hallbeck MS, Salehinejad H, et al. The integration of artificial intelligence in robotic surgery: A narrative review. Surgery 2024;176:552-7. [Crossref] [PubMed]
14. Knudsen JE, Ghaffar U, Ma R, et al. Clinical applications of artificial intelligence in robotic surgery. J Robot Surg 2024;18:102. [Crossref] [PubMed]
15. Panesar S, Cagle Y, Chander D, et al. Artificial Intelligence and the Future of Surgical Robotics. Ann Surg 2019;270:223-6. [Crossref] [PubMed]
16. Iftikhar M, Saqib M, Zareen M, et al. Artificial intelligence: revolutionizing robotic surgery Ann Med Surg (Lond) 2024;86:5401-9. review. [Crossref] [PubMed]
17. Hashemi N, Svendsen MBS, Bjerrum F, et al. Acquisition and usage of robotic surgical data for machine learning analysis. Surg Endosc 2023;37:6588-601. [Crossref] [PubMed]
18. Chatterjee S, Das S, Ganguly K, et al. Advancements in robotic surgery: innovations, challenges and future prospects. J Robot Surg 2024;18:28. [Crossref] [PubMed]
19. Vasey B, Lippert KAN, Khan DZ, et al. Intraoperative Applications of Artificial Intelligence in Robotic Surgery: A Scoping Review of Current Development Stages and Levels of Autonomy. Ann Surg 2023;278:896-903. [Crossref] [PubMed]
20. O’Sullivan S, Leonard S, Holzinger A, et al. Operational framework and training standard requirements for AI-empowered robotic surgery. Int J Med Robot 2020;16:1-13. [Crossref] [PubMed]
21. Moglia A, Georgiou K, Georgiou E, et al. A systematic review on artificial intelligence in robot-assisted surgery. Int J Surg 2021;95:106151. [Crossref] [PubMed]
22. Furube T, Takeuchi M, Kawakubo H, et al. Usefulness of an Artificial Intelligence Model in Recognizing Recurrent Laryngeal Nerves During Robot-Assisted Minimally Invasive Esophagectomy. Ann Surg Oncol 2024;31:9344-51. [Crossref] [PubMed]
23. Chappell AG, Teven CM. How Should Surgeons Consider Emerging Innovations in Artificial Intelligence and Robotics? AMA J Ethics 2023;25:E589-597. [Crossref] [PubMed]
24. Liu Z, Jian X, Sadiq T, et al. Efficient control of spider-like medical robots with capsule neural networks and modified spring search algorithm. Sci Rep 2025;15:13828. [Crossref] [PubMed]
25. De Backer P, Eckhoff JA, Simoens J, et al. Multicentric exploration of tool annotation in robotic surgery: lessons learned when starting a surgical artificial intelligence project. Surg Endosc 2022;36:8533-48. [Crossref] [PubMed]
26. Collins JW, Marcus HJ, Ghazi A, et al. Ethical implications of AI in robotic surgical training: A Delphi consensus statement. Eur Urol Focus 2022;8:613-22. [Crossref] [PubMed]
27. Bhandari M, Zeffiro T, Reddiboina M. Artificial intelligence and robotic surgery: current perspective and future directions. Curr Opin Urol 2020;30:48-54. [Crossref] [PubMed]
28. Andras I, Mazzone E, van Leeuwen FWB, et al. Artificial intelligence and robotics: a combination that is changing the operating room. World J Urol 2020;38:2359-66. [Crossref] [PubMed]
29. Cavallo F, Sinigaglia S, Megali G, et al. Biomechanics-machine learning system for surgical gesture analysis and development of technologies for minimal access surgery. Surg Innov 2014;21:504-12. [Crossref] [PubMed]
30. Boal MWE, Anastasiou D, Tesfai F, et al. Evaluation of objective tools and artificial intelligence in robotic surgery technical skills assessment: a systematic review. Br J Surg 2024;111:znad331. [Crossref] [PubMed]
31. Ma R, Kiyasseh D, Laca JA, et al. Artificial Intelligence-Based Video Feedback to Improve Novice Performance on Robotic Suturing Skills: A Pilot Study. J Endourol 2024;38:884-91. [Crossref] [PubMed]
32. O’Sullivan S, Nevejans N, Allen C, et al. Legal, regulatory, and ethical frameworks for development of standards in artificial intelligence (AI) and autonomous robotic surgery. Int J Med Robot 2019;15:e1968. [Crossref] [PubMed]
33. Morris MX, Song EY, Rajesh A, et al. Ethical, Legal, and Financial Considerations of Artificial Intelligence in Surgery. Am Surg 2023;89:55-60. [Crossref] [PubMed]
34. Arjomandi Rad A, Vardanyan R, Athanasiou T, et al. The ethical considerations of integrating artificial intelligence into surgery: a review. Interdiscip Cardiovasc Thorac Surg 2025;40:ivae192. [Crossref] [PubMed]
35. Eppler MB, Sayegh AS, Maas M, et al. Automated Capture of Intraoperative Adverse Events Using Artificial Intelligence: A Systematic Review and Meta-Analysis. J Clin Med 2023;12:1687. [Crossref] [PubMed]
36. Özçıbık Işık G, Kılıç B, Erşen E, et al. Prediction of postoperative intensive care unit admission with artificial intelligence models in non-small cell lung carcinoma. Eur J Med Res 2025;30:293. [Crossref] [PubMed]
37. Dennis BM, Gondek SP, Guyer RA, et al. Use of an evidence-based algorithm for patients with traumatic hemothorax reduces need for additional interventions. J Trauma Acute Care Surg 2017;82:728-32. [Crossref] [PubMed]
38. Balch JA, Abbott KL, Loftus TJ. Critical appraisal of artificial intelligence in robotic surgery. Surgery 2024;176:558-9. [Crossref] [PubMed]
39. Sullivan HR, Schweikart SJ. Are Current Tort Liability Doctrines Adequate for Addressing Injury Caused by AI? AMA J Ethics 2019;21:E160-6. [Crossref] [PubMed]