Artificial intelligence in surgery—a narrative review

Introduction

Artificial intelligence (AI) is poised to redefine surgical practice, offering innovative solutions to enhance patient care and outcomes across the perioperative spectrum. The increasing global demand for surgical procedures highlights the need for AI to address access, precision, and efficiency challenges (1). AI integration into surgical systems opens new possibilities in risk prediction, real-time analytics, and longitudinal monitoring, spanning from preoperative planning to intraoperative decision-making and postoperative care. Beyond patient care, AI can accelerate surgical education through personalized learning experiences and automated skills assessments. However, realizing AI’s full potential requires navigating complex ethical, legal, and practical challenges, including mitigating bias, safeguarding patient privacy, validating AI systems, and effectively integrating AI into surgical workflows and training (2).

This narrative review aims to comprehensively explore AI’s role in surgery by describing its current applications and future trajectory. We delve into the ethical, legal, and practical challenges accompanying AI’s use while highlighting its transformative potential for surgical practice. By synthesizing the current state of knowledge, this review seeks to offer insights and recommendations to guide the responsible and effective implementation of AI in surgical settings. We present this article in accordance with the Narrative Review reporting checklist (available at https://jmai.amegroups.com/article/view/10.21037/jmai-24-111/rc).

Methods

We conducted a comprehensive literature search using PubMed and Google Scholar databases. The search terms included combinations of “artificial intelligence”, “surgery”, “surgical robotics”, “machine learning”, “deep learning”, “ethics”, “surgical education”, “surgical training”, and “clinical trials”. We prioritized articles published within the last 5 years [2020–2024] to capture the most recent advancements and discussions in the field (Table 1).

We included original research articles, systematic reviews, meta-analyses, and expert opinion pieces exploring AI’s applications, implications, and challenges in surgical settings. Articles were selected based on their relevance to the key themes of our review, which included current applications, ethical considerations, impact on surgical education and training, and strategies for safe and effective integration of AI in surgery. Articles were excluded if they were not written in English or if they discussed AI for non-surgical applications. During the preparation of this manuscript, the authors utilized ChatGPT-4 by OpenAI and Claude-3 by Anthropic for writing and editing assistance, including adapting content for Table 2, while carefully reviewing and taking full responsibility for the final manuscript content (details included in the Acknowledgments section).

Applications of AI in surgical settings

AI-enabled technologies augment surgeons’ capabilities in three main areas: preoperative planning, intraoperative assistance, and postoperative monitoring.

Enhanced preoperative planning

The preoperative phase is crucial for assessing patient risks, mapping anatomical structures, and strategizing procedures. AI systems can integrate population-level surgical data to determine patient-specific risks and expected outcomes of different interventions, facilitating evidence-based decisions about the best course of treatment. For example, Li et al. (3) developed a machine learning model to help anesthesiologists assess the risk of complications after hip surgery. Their model showed higher performance in risk assessment than the traditional American Society of Anesthesiologists physical status (ASA-PS) method, and anesthesiologists who used the tool reported high satisfaction. Other validated machine learning algorithms, such as MySurgeryRisk, Trauma Outcomes Predictor (TOP), and Predictive OpTimeal Trees in Emergency Surgery Risk (POTTER), have been developed to predict the risk of significant complications and death after surgery (4-6).

Sophisticated AI algorithms can process high-volume imaging data to construct detailed 3D models of anatomical structures. Deep learning modeling approaches can automate organ and muscle segmentation in medical images, achieving high accuracy while preserving detailed boundaries (7). Unlike traditional imaging, these interactive 3D models allow visualizations from multiple perspectives, enabling surgeons to visualize the surgical site and identify potential intra-operative challenges (8). AI guidance can also optimize surgical parameters such as incision placements or orthopedic implant sizing (9). These advancements in preoperative planning empower surgeons to make more informed decisions and prepare for surgeries with greater precision.

Intraoperative visualization and robotic automation

The adoption of surgical robots has increased over the past decades, with their use rising from 1.8% of all general surgeries in 2012 to 15.1% in 2018 (10). AI systems assist in robotic-assisted surgeries by providing real-time feedback to optimize tool trajectories, perform autonomous tissue retraction, and detect tissue damage (11-13). Whether part of a robotic surgical system or as a separate tool, AI can be used for visualization and navigation during complex surgeries, providing surgeons with real-time feedback and critical insights. AI-enabled intraoperative imaging technologies allow surgeons to clearly distinguish anatomical structures and track instruments (14). For instance, overlaying magnetic resonance imaging (MRI) or computed tomography (CT) visuals directly onto the field of view can outline surgical sites with unprecedented clarity (15). Kok et al. (16) demonstrated the feasibility and safety of real-time tumor tracking during rectal cancer surgery using an AI-enabled tumor sensor placed preoperatively, which could be used to decrease positive margin resections.

The Massachusetts General Hospital Surgical Artificial Intelligence and Innovation Laboratory (SAIIL) has been actively involved in utilizing AI technology to extract knowledge from intraoperative procedure videos. Ward et al. (17) found that AI could accurately predict the operative course based on the Parkland Grading Scale (PGS) of gallbladder inflammation observed in surgical videos. In recent work, they introduced a novel approach for understanding surgical videos using Concept Graph Neural Networks (CGNNs), which incorporate a knowledge graph into the temporal analysis of surgical procedures to enhance AI-augmented surgery. This methodology allows for dynamic learning of surgical concepts and their interrelations within video data, resulting in improved performance on surgical safety view verification, clinical scale estimation, and recognition of complex interactions between surgical instruments and tissues (18). In related work, Yin et al. presented the Hypergraph-Transformer (HGT) for predicting interactive events in laparoscopic and robotic surgery. Compared to traditional methods, their approach achieved superior accuracy in predicting critical surgical scenarios, such as clipping cystic ducts, before attaining the Critical View of Safety (CVS) (19).

Theator, a company focused on advancing surgical training and quality assurance through the application of AI in surgical video analysis, has made substantial contributions to the field. Their research has demonstrated how AI, trained on a comprehensive dataset of 1,243 laparoscopic cholecystectomy videos, accurately recognized surgical phases and generalized to new settings (20). While the dataset consisted of a single procedural type, it included videos from six different medical centers and more than 50 surgeons, providing a wide range of conditions for the model to learn from. Another study evaluated the ability of Theator’s AI to facilitate the review of laparoscopic cholecystectomy surgical videos, achieving a high agreement rate of over 75% with surgeons on annotated CVS components and enabling surgeons to review up to 50 videos per hour (21).

Beyond instrumentation and visualization, intraoperative predictive analytics have been used to alert the surgical team to potentially harmful physiological changes. Wijnberge et al. (22) conducted a randomized controlled trial of an AI platform to predict and prompt early treatment of intraoperative hypotension. They found that patients in the intervention group were hypotensive for a median of 16.7 minutes less than the control group.

Optimized recovery

AI continues to assist patients beyond the operating room by enabling personalized and proactive post-surgery care. Intelligent algorithms can track vital parameters, analyze data from various sources, and identify early signs of complications such as infections, bleeding, or organ dysfunction (23,24). Domain-specific AI models have been developed to predict a wide range of postoperative complications, including prolonged ventilation, blood loss, cardiac complications, pulmonary complications, renal complications, aspiration, venous thromboembolism, blood transfusion, surgical site infections, harmful opioid use, and mortality (25,26). These predictive capabilities allow healthcare teams to triage postoperative patients appropriately (27,28) and to intervene early to prevent adverse outcomes.

Wearable technology combined with AI offers the potential for more accurate monitoring of postoperative recovery progress. Wearable devices can provide data-driven insights into a patient’s recovery trajectory, enabling healthcare teams to tailor rehabilitation plans to individual needs (29). AI-driven virtual reality (VR) systems are also being explored for pain management and physical therapy, offering immersive experiences that can reduce reliance on opioids and improve recovery outcomes (30). Furthermore, AI can assist in longitudinal patient care by enabling long-term monitoring, identifying patients at risk of chronic complications or readmission, and facilitating early interventions (31). AI conversational agents, capable of simulating human voice, have been effective at following up with post-operative patients and reducing resource utilization compared to manual phone calls (32).

Ethical considerations

Integrating AI in surgical decision-making presents complex ethical challenges that must be carefully navigated to ensure responsible implementation. The ABCD approach (accountability, bias, confidentiality, decision-making) provides a framework to identify and address these ethical issues (33).

Accountability

As AI systems become more integrated into surgical decision-making, questions of accountability and responsibility become increasingly complex. Traditional models that place responsibility solely on individual healthcare providers need to be revised to account for AI’s role. Surgeons must judiciously weigh AI-generated recommendations against their professional expertise, recognizing that while AI tools offer valuable insights, they do not embody a human surgeon’s depth of understanding and experiential judgment (34). To ensure accountability, liability should be clearly defined and allocated based on specific process stages, such as the analysis of X-ray images. This may involve the manufacturer, the operator, or those responsible for maintenance. Culpability should be unequivocally assigned, potentially mandating a secondary human review of any AI-derived decision (35,36). Continuous monitoring of AI systems for accuracy and dependability, along with clear guidelines for human supervision, is crucial for ensuring patient safety and mitigating the risk of errors due to technical issues or misinterpretations of AI data (37).

Robust legislative frameworks involving diverse stakeholders, including legal experts, healthcare policymakers, and judiciary officials, are essential to preemptively address and resolve potential liability issues. Algorithmic accountability requires not just transparency into the AI system itself but second-order transparency into the governance structures around the AI to ensure private sector interests in developing and implementing the technology remain aligned with the public good (38). Regulatory bodies like the Food and Drug Administration must play a pivotal role by establishing rigorous AI system validation and certification standards, including stringent testing for accuracy, reliability, and safety in various clinical scenarios (39). Informed consent also takes on new dimensions in the age of AI. Patients must be thoroughly educated about the role of AI in their treatment, including the risks, benefits, and available alternatives. Ensuring patients clearly understand how AI influences their care is fundamental to ethical medical practice (40).

Bias

The ethical landscape of AI in healthcare is significantly complicated by the inherent biases that can manifest in AI systems. These biases often originate from unbalanced data inputs and have the potential to perpetuate discrimination against specific demographics, including race, gender, or socioeconomic groups (41-43). Such biases fundamentally contravene the ethical principle of justice, potentially leading to unequal healthcare access and quality. Transparency in how AI systems operate and make decisions is crucial (44). However, current algorithms often lack clarity regarding which data is included or excluded and how decisions are made (45,46). This absence of transparency hinders the ability to identify and assess biases and is a crucial factor contributing to patient apprehensions about the use of AI in healthcare (47). Several strategies can be employed to mitigate these biases. These include diversifying the datasets used for training AI systems (48,49), algorithmically identifying and adjusting for biases (50), and implementing ongoing surveillance mechanisms to detect and rectify biased outcomes (41).

Mitigating unfair bias in AI tools necessitates diversity in data collection and involving stakeholders from marginalized groups in the design process. At the same time, AI has the potential to equalize gender disparities in surgical training and education by overcoming barriers related to gender discrimination, enhancing remote mentoring and training, and improving work-life balance for female surgeons (51). Ensuring that AI tools in surgery and healthcare are utilized equitably will aid in achieving the broader objective of creating a just healthcare system (52).

Confidentiality

The paradox of AI in healthcare lies in its dependence on patient data while simultaneously needing to safeguard patient confidentiality and privacy. The issue of data ownership raises critical ethical questions, especially in areas like robotic surgery, where manufacturers might claim ownership of generated surgical data. While compiling a centralized database of surgical data into a freely accessible format could foster research and innovation, patients may not wish to have their surgical videos freely accessible, and companies are not financially incentivized to share data (53).

The vulnerability of health records to cyberattacks underscores the importance of maintaining the confidentiality of medical records. The progression of AI technology introduces additional privacy concerns, as there is a risk that users might unknowingly consent to extensive data collection under the impression that they are interacting with human operators (54). There have been notable instances where patient data has been used for AI research without explicit approval from the individuals involved. The Google DeepMind and National Health Service patient data controversy, where data was used beyond the publicly disclosed scope to develop a clinical alert app, highlights the importance of utilizing a principled approach to innovation in public health services, creating and following regulatory frameworks, and pursuing meaningful societal engagement with equitable value-sharing in data use (55).

While laws like the European Union’s General Data Protection Regulation (56) attempt to tackle these challenges by limiting personal data usage, these laws can complicate collaborative research efforts, as varying regulations across countries may restrict data availability necessary for training AI systems on a larger scale. Addressing these challenges without stifling innovation requires stricter data security measures. Enhancing client-side data encryption and utilizing federated learning, which allows for model training without extensive data dispersion, are potential solutions that balance data security with advancing AI technology in healthcare (57).

Decision-making

As AI systems take on critical decision-making roles in healthcare, from diagnostics to prescribing treatments and controlling surgical instruments, it is paramount to ensure these AI-driven decisions align with ethical standards and are tailored to individual patient needs. Sarofim (58) emphasized the need for transparency and accountability in AI algorithms used for surgical decision-making, arguing that surgeons and patients are entitled to access evidence regarding these AI systems’ development, validation, and inherent limitations. While AI-assisted clinical decision support systems (CDSS) can provide more accurate and efficient recommendations, there are risks of over-reliance on these systems by clinicians, potential harm to patients if AI recommendations are followed without careful consideration of patient-specific factors, and erosion of public trust if AI is seen as the sole decision-maker (59).

The role of clinicians in AI-augmented decision-making is also critical. Surgeons, nurses, and other healthcare professionals must receive adequate training to understand and effectively utilize AI systems in their practice (60). This training should cover both technical and ethical aspects, ensuring healthcare providers can interpret AI recommendations within a patient’s unique clinical situation (61). As AI-assisted CDSS blurs the line between clinical practice and research by continuously learning from new data, clinicians face potential conflicts in their dual roles as clinical decision-makers and data collectors (59). Ethical frameworks and professional guidelines must account for these challenges, which may be attenuated by co-designing these systems with input from all stakeholders (62).

AI’s impact on surgical practice and training

The AI revolution is poised to transform surgical education by introducing personalized and simulation-based learning experiences that redefine how surgical skills are acquired and honed (63). At the forefront of this change are algorithms designed to meticulously evaluate the strengths and weaknesses of trainees, tailoring the training process to their specific needs.

Transforming surgical education

AI enhances simulation-based training by offering personalized feedback and objective assessment tools. Video-based surgical assessment systems that employ computer vision and machine learning can provide data-driven feedback, enabling a more tailored and practical training experience (64). By analyzing performance metrics such as motion tracking and force measurements, AI-driven tools can pinpoint specific areas for improvement, offering a more customized learning experience for each trainee (65). Julian and Smith (66) created a computer-based intelligent tutoring system to teach robotic suturing to novice robotic surgeons. The process of robotic suturing was documented through videos, instructional sets, and flowcharts; these materials were then verified for accuracy by experienced surgeons and integrated into the tutoring system. Lohre et al. (67) showed that trainees who used an immersive VR shoulder arthroplasty simulation for 60 minutes acquired the same amount of learning that would have been acquired in 47 minutes in an actual operating room. Such AI-augmented simulation tools could bolster traditional procedural simulation tasks to improve procedural skills in low- and middle-income countries with a shortage of surgeons (68,69). Expansion of such programs could bridge geographical and socioeconomic gaps, allowing aspiring surgeons to access high-quality training and resources previously beyond their reach.

AI algorithms can also predict a trainee’s performance and skill development trajectory. This predictive power is beneficial for optimizing training curricula to match individual learning needs and progress. The novel VR Sim-Ortho simulator successfully measured surgical performance in an anterior cervical discectomy scenario. At the same time, a complementary neural network-based AI model identified key procedural safety metrics (70). Baloul et al. (71) showed that AI could use transcribed commentary from trainees watching and commenting on surgical videos to determine training levels more accurately than traditional statistical techniques. The Virtual Operative Assistant, an automated educational feedback platform based on a VR brain tumor resection task, was shown to classify novice vs. skilled participants with 92% accuracy. Similarly, performance metrics from a VR neurosurgical tumor resection were able to classify participants’ level of training with 90% accuracy (72).

Large language models (LLMs), such as OpenAI’s GPT-4, hold significant promise in transforming surgical education. These models can facilitate the transcription, translation, and summarization of feedback, potentially enhancing the communication between trainers and trainees (73). Beyond these capabilities, LLMs are poised to augment surgical curricula, with one study demonstrating their ability to offer suitable and safe procedural advice for common ward-based surgical scenarios (74). Furthermore, research exploring the role of LLMs as teaching assistants in plastic surgery has shown their potential to generate content for interactive case studies, simulations, and ethical dilemmas (75). While these studies point towards the potential of LLMs in advancing surgical education, a notable gap in the literature concerning best practices suggests an imperative area for future research.

Adoption challenges and strategies

Ensuring surgeons and healthcare professionals are proficient with AI technologies is vital to successful integration into clinical practice. However, the adoption of AI tools in surgical education faces challenges. Poor algorithm interpretability lowers trust in AI tools, supporting the need for rigorous validation studies and phased deployment to augment human expertise (76). High costs have also deterred the adoption of AI tools (77), though cost-aware AI models that can function in healthcare environments with limited data infrastructure may help address this barrier (78).

Moreover, compliance with evolving standards, particularly concerning patient safety and legal risks, requires external algorithm audits, limiting autonomous functionality, and involving human supervisors in the loop (79). A multi-faceted strategy is essential to overcome these challenges. Medical institutions must invest in high-quality AI tools and ensure they are constantly updated and validated against current surgical standards (80). Training programs must be developed to familiarize surgeons and trainees with these new technologies (81). Collaboration with AI developers can help tailor these tools to specific surgical training needs, ensuring they are practical and relevant (82). Furthermore, continuous research is needed to explore the best practices for implementing these technologies in surgical training, including the effective use of LLMs.

Ensuring safe and effective integration of AI in surgery

Integrating AI tools into surgical practice holds immense potential for enhancing patient care; however, the intrinsic complexity of surgery requires a careful and responsible approach to adopting these rapidly advancing technologies.

Establishing safety and privacy

The complexity of surgical practice necessitates the establishment of robust safety protocols to ensure the safe and effective integration of AI tools. This involves implementing stringent safety measures to ensure accountability, transparency, and the protection of patient privacy. Regulatory frameworks for AI in healthcare must account for AI systems’ ongoing evolution and improvement, favoring periodic, clinically evaluated updates and establishing continuous performance monitoring to address potential drift and maintain model accuracy over time (79). Incorporating AI tools in the healthcare environment requires the development of industry standards to facilitate collaboration among AI developers, clinicians, ethicists, and regulators (83).

The challenges of protecting health information in the era of AI should be addressed through the development of privacy-preserving solutions and the implementation of digital masks to safeguard patient privacy, ensuring the secure exchange of health data while protecting patient confidentiality (84). Secure and private health data exchange platforms, such as the CUREX platform, complement cybersecurity and privacy risk assessment functionality to propose optimal safeguards for mitigating risks in health data exchange (85).

The role of prospective clinical trials

While AI applications have shown high proficiency in controlled lab settings, their real-world performance in more complex and variable environments needs thorough evaluation. Prospective randomized controlled trials are essential in this context. Trials must assess technical accuracy, impact on clinical outcomes, and integration with existing hospital infrastructure and staff. The DECIDE-AI reporting guideline was introduced to bridge the development-to-implementation gap in clinical AI (86). DECIDE-AI provides guidelines for early and small-scale clinical evaluations of AI-based decision-support systems, ensuring the evaluation of human factors and algorithm usability before large-scale trials.

SPIRIT-AI and CONSORT-AI offer specific reporting guidelines for AI interventions at the clinical trial stage (87,88). These guidelines aim to enhance transparency and completeness in reporting clinical trials for AI applications, aiding editors, peer reviewers, and readers in understanding, interpreting, and critically appraising the quality of trial design and potential biases in reported outcomes. The CONSORT-AI extension, for example, includes 14 new items that prompt investigators to provide detailed descriptions of the AI intervention, including the required skills for use, the clinical setting of integration, the management of inputs and outputs, the interaction between humans and AI, and an analysis of error cases (88).

Training surgeons as AI end-users

Surgeons, being at the forefront as end-users and contributors in developing AI tools, play a pivotal role. Their education must encompass an understanding of the strengths and limitations of AI systems. When introduced into clinical practice, AI developers and vendors should provide rich information about a new AI tool’s global properties, strengths, and limitations (89). Educators should develop an easily accessible AI curriculum for medical students and practicing clinicians to enable them to critically appraise, adopt, and use AI tools safely in their practice (79). The curricula should evolve to include training in statistical principles and computing techniques while emphasizing clinical judgment and surgical skills. This will require close collaboration between surgical educators, AI experts, and professional societies to develop comprehensive, up-to-date curricula that balance technical skills with clinical judgment. Finally, surgeons should underline the need for longitudinal mentorship and faculty role modeling to facilitate the understanding and application of AI knowledge in the clinical setting (90,91).

Multidisciplinary development and implementation

Multidisciplinary collaboration is crucial in the development and implementation of AI in surgery. Madaio et al. (92) emphasized the importance of understanding organizational challenges and opportunities around deploying ethical AI systems, highlighting the need for a comprehensive approach that involves experts from various fields. Clinical stakeholders, including nursing staff, should be involved early in the project and implementation planning of any AI-driven intervention designed to take place within a healthcare system (93). Additionally, patient involvement in developing and evaluating AI tools in surgery will ensure that these tools meet patients’ actual needs and address their concerns. For example, Gould et al. (94) found that patients undergoing knee replacement surgery had varying levels of understanding and perceptions of AI, suggesting the need for patient education using simple explanations and practical examples to aid in shared AI decision-making with surgeons. Bringing together diverse perspectives and expertise will help to ensure that AI tools are not only technologically advanced but also clinically relevant, ethically sound, and patient-centered.

Limitations

This narrative review has several limitations. First, the literature search was restricted to English-language publications from 2016 to 2024, potentially excluding relevant studies published outside this timeframe or in other languages. The rapid pace of AI development means that new studies and innovations may have emerged since the literature search was conducted, while historical pioneering research in AI and surgery was not included in this review. Including only English-language papers may have diminished contributions from non-English speaking countries, which should be acknowledged given the global nature of AI research. Additionally, the interpretation of findings is inherently limited by the available data and methodologies employed in the reviewed studies. Most studies were conducted in high-income countries, limiting the generalizability of findings to low- and middle-income settings where surgical resources and infrastructure may differ significantly. Finally, while this review highlighted the importance of multidisciplinary collaboration and ethical considerations in developing and implementing AI in surgery, it did not provide an exhaustive exploration of these topics. The rapid evolution of AI technologies may raise novel ethical questions and challenges for effective teamwork that were not anticipated.

Conclusions

AI is advancing rapidly, enhancing surgeons’ capabilities through refined preoperative planning, nuanced risk assessment, superior intraoperative guidance, and improved postoperative care. However, the safe and effective implementation of AI in surgery requires careful consideration of the ethical, legal, and practical challenges, such as ensuring accountability, mitigating bias, protecting patient privacy, and validating AI systems through rigorous clinical trials (Table 2). The impact of AI on surgical education is also growing, offering opportunities for personalized learning experiences and objective skill assessment. To fully realize the potential of AI in surgery, ongoing research, multidisciplinary collaboration, and the development of robust governance frameworks will be essential. Future studies should focus on evaluating AI interventions’ clinical and economic outcomes in diverse surgical settings and patient populations to guide evidence-based integration into surgical practice.

Acknowledgments

During the preparation of this manuscript, the authors utilized the following AI tools to assist with the writing and editing: (I) ChatGPT-4 by OpenAI (https://openai.com/); dates of use: November 21, 2023, April 10, 2024; purpose: to organize authors’ conceptual ideas into outline format and refine sentence structure and grammar. (II) Claude-3 by Anthropic (https://www.anthropic.com/); dates of use: April 10–16, 2024; purpose: to adapt the manuscript content to create Table 2 (prompt: “Create a detailed summary table showing the key applications of AI in surgery based on the attached manuscript”). The authors carefully reviewed and edited the AI-generated content to ensure accuracy and the authors take full responsibility for the final content of the manuscript. This disclosure adheres to the World Association of Medical Editors (WAME) Recommendations on Chatbots and Generative Artificial Intelligence in Relation to Scholarly Publications.

Funding: This work was supported by the Agency for Healthcare Research and Quality (AHRQ) and the Patient-Centered Outcomes Research Institute (PCORI) (No. P30HS029744) and the Minnesota Learning Health System (MN-LHS) Mentored Career Development Program through the University of Minnesota Office of Academic Clinical Affairs, Clinical Translational Science Institute, and Center for Learning Health System Sciences. The content is solely the responsibility of the authors and does not necessarily represent the official views of AHRQ, PCORI, or MN-LHS.

Footnote

Reporting Checklist: The authors have completed the Narrative Review reporting checklist. Available at https://jmai.amegroups.com/article/view/10.21037/jmai-24-111/rc

Peer Review File: Available at https://jmai.amegroups.com/article/view/10.21037/jmai-24-111/prf

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at https://jmai.amegroups.com/article/view/10.21037/jmai-24-111/coif). The 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 are appropriately investigated and resolved.

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. Varghese C, Harrison EM, O’Grady G, et al. Artificial intelligence in surgery. Nat Med 2024;30:1257-68. [Crossref] [PubMed]
2. 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]
3. Li YY, Wang JJ, Huang SH, et al. Implementation of a machine learning application in preoperative risk assessment for hip repair surgery. BMC Anesthesiol 2022;22:116. [Crossref] [PubMed]
4. Bertsimas D, Dunn J, Velmahos GC, et al. Surgical Risk Is Not Linear: Derivation and Validation of a Novel, User-friendly, and Machine-learning-based Predictive OpTimal Trees in Emergency Surgery Risk (POTTER) Calculator. Ann Surg 2018;268:574-83. [Crossref] [PubMed]
5. Bihorac A, Ozrazgat-Baslanti T, Ebadi A, et al. MySurgeryRisk: Development and Validation of a Machine-learning Risk Algorithm for Major Complications and Death After Surgery. Ann Surg 2019;269:652-62. [Crossref] [PubMed]
6. Maurer LR, Bertsimas D, Bouardi HT, et al. Trauma outcome predictor: An artificial intelligence interactive smartphone tool to predict outcomes in trauma patients. J Trauma Acute Care Surg 2021;91:93-9. [Crossref] [PubMed]
7. Villarini B, Asaturyan H, Kurugol S, et al. 3D Deep Learning for Anatomical Structure Segmentation in Multiple Imaging Modalities. Proc IEEE Int Symp Comput Based Med Syst 2021;2021:166-71. [Crossref] [PubMed]
8. Huettl F, Saalfeld P, Hansen C, et al. Virtual reality and 3D printing improve preoperative visualization of 3D liver reconstructions-results from a preclinical comparison of presentation modalities and user’s preference. Ann Transl Med 2021;9:1074. [Crossref] [PubMed]
9. Lambrechts A, Wirix-Speetjens R, Maes F, et al. Artificial Intelligence Based Patient-Specific Preoperative Planning Algorithm for Total Knee Arthroplasty. Front Robot AI 2022;9:840282. [Crossref] [PubMed]
10. Sheetz KH, Claflin J, Dimick JB. Trends in the Adoption of Robotic Surgery for Common Surgical Procedures. JAMA Netw Open 2020;3:e1918911. [Crossref] [PubMed]
11. Afebu KO, Tian J, Liu Y, et al. AI-Assisted Dynamic Tissue Evaluation for Early Bowel Cancer Diagnosis Using a Vibrational Capsule. IEEE Robotics and Automation Letters 2023;8:2341-8. [Crossref]
12. Attanasio A, Scaglioni B, Leonetti M, et al. Autonomous Tissue Retraction in Robotic Assisted Minimally Invasive Surgery – A Feasibility Study. IEEE Robotics and Automation Letters 2020;5:6528-35. [Crossref]
13. Shahkoo AA, Abin AA. Autonomous Tissue Manipulation via Surgical Robot Using Deep Reinforcement Learning and Evolutionary Algorithm. IEEE Transactions on Medical Robotics and Bionics 2023;5:30-41. [Crossref]
14. Mascagni P, Alapatt D, Sestini L, et al. Computer vision in surgery: from potential to clinical value. NPJ Digit Med 2022;5:163. [Crossref] [PubMed]
15. Ivanov VM, Krivtsov AM, Strelkov SV, et al. Practical Application of Augmented/Mixed Reality Technologies in Surgery of Abdominal Cancer Patients. J Imaging 2022;8:183. [Crossref] [PubMed]
16. Kok END, Eppenga R, Kuhlmann KFD, et al. Accurate surgical navigation with real-time tumor tracking in cancer surgery. NPJ Precis Oncol 2020;4:8. [Crossref] [PubMed]
17. Ward TM, Hashimoto DA, Ban Y, et al. Artificial intelligence prediction of cholecystectomy operative course from automated identification of gallbladder inflammation. Surg Endosc 2022;36:6832-40. [Crossref] [PubMed]
18. Ban Y, Eckhoff JA, Ward TM, et al. Concept Graph Neural Networks for Surgical Video Understanding. IEEE Trans Med Imaging 2024;43:264-74. [Crossref] [PubMed]
19. Yin L, Ban Y, Eckhoff J, et al. Hypergraph-Transformer (HGT) for Interactive Event Prediction in Laparoscopic and Robotic Surgery. rXiv:2402.01974 [Preprint]. 2024. Available online: https://doi.org/10.48550/arXiv.2402.01974
20. Bar O, Neimark D, Zohar M, et al. Impact of data on generalization of AI for surgical intelligence applications. Sci Rep 2020;10:22208. [Crossref] [PubMed]
21. Korndorffer JR Jr, Hawn MT, Spain DA, et al. Situating Artificial Intelligence in Surgery: A Focus on Disease Severity. Ann Surg 2020;272:523-8. [Crossref] [PubMed]
22. Wijnberge M, Geerts BF, Hol L, et al. Effect of a Machine Learning-Derived Early Warning System for Intraoperative Hypotension vs Standard Care on Depth and Duration of Intraoperative Hypotension During Elective Noncardiac Surgery: The HYPE Randomized Clinical Trial. JAMA 2020;323:1052-60. [Crossref] [PubMed]
23. Aasvang EK, Meyhoff CS. The future of postoperative vital sign monitoring in general wards: improving patient safety through continuous artificial intelligence-enabled alert formation and reduction. Curr Opin Anaesthesiol 2023;36:683-90. [Crossref] [PubMed]
24. Alsareii SA, Raza M, Alamri AM, et al. Machine Learning and Internet of Things Enabled Monitoring of Post-Surgery Patients: A Pilot Study. Sensors (Basel) 2022;22:1420. [Crossref] [PubMed]
25. Bates DW, Levine D, Syrowatka A, et al. The potential of artificial intelligence to improve patient safety: a scoping review. NPJ Digit Med 2021;4:54. [Crossref] [PubMed]
26. Lopez CD, Gazgalis A, Boddapati V, et al. Artificial Learning and Machine Learning Decision Guidance Applications in Total Hip and Knee Arthroplasty: A Systematic Review. Arthroplast Today 2021;11:103-12. [Crossref] [PubMed]
27. Byrd TF 4th, Ho J, Tignanelli CJ. Invited Commentary: Postoperative Artificial Intelligence Model for ICU Triage. J Am Coll Surg 2023;236:292-3. [Crossref] [PubMed]
28. Loftus TJ, Ruppert MM, Ozrazgat-Baslanti T, et al. Postoperative Overtriage to an Intensive Care Unit Is Associated With Low Value of Care. Ann Surg 2023;277:179-85. [Crossref] [PubMed]
29. Ianculescu M, Andrei B, Alexandru A. A Smart Assistance Solution for Remotely Monitoring the Orthopaedic Rehabilitation Process Using Wearable Technology: re.flex System. Studies in Informatics and Control 2019;28:317-26. [Crossref]
30. Dreesmann NJ, Su H, Thompson HJ. A Systematic Review of Virtual Reality Therapeutics for Acute Pain Management. Pain Manag Nurs 2022;23:672-81. [Crossref] [PubMed]
31. Eigner I, Reischl D, Bodendorf F. Development and evaluation of ensemble-based classification models for predicting unplanned hospital readmissions after hysterectomy. Australasian Conference on Information Systems 2018, Sydney. 2018. [Epub ahead of print]. doi: 10.5130/acis2018.cg.10.5130/acis2018.cg
32. Bian Y, Xiang Y, Tong B, et al. Artificial Intelligence-Assisted System in Postoperative Follow-up of Orthopedic Patients: Exploratory Quantitative and Qualitative Study. J Med Internet Res 2020;22:e16896. [Crossref] [PubMed]
33. AlHasan A. Ethical Concerns Grow as AI Takes on Greater Decision-Making Role. 2023. [accessed November 22, 2023]. Available online: https://www.facs.org/for-medical-professionals/news-publications/news-and-articles/bulletin/2023/february-2023-volume-108-issue-2/ethical-concerns-grow-as-ai-takes-on-greater-decision-making-role/
34. 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]
35. 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]
36. Wolff J, Pauling J, Keck A, et al. Success Factors of Artificial Intelligence Implementation in Healthcare. Front Digit Health 2021;3:594971. [Crossref] [PubMed]
37. Bartlett B. The possibility of AI-induced medical manslaughter: Unexplainable decisions, epistemic vices, and a new dimension of moral luck. Medical Law International 2023;23:241-70. [Crossref]
38. Kaminski ME. Understanding Transparency in Algorithmic Accountability. In: Barfield W. editor. The Cambridge Handbook of the Law of Algorithms. Cambridge: Cambridge University Press; 2020:121-38.
39. Amann J, Blasimme A, Vayena E, et al. Explainability for artificial intelligence in healthcare: a multidisciplinary perspective. BMC Med Inform Decis Mak 2020;20:310. [Crossref] [PubMed]
40. Loftus TJ, Filiberto AC, Li Y, et al. Decision analysis and reinforcement learning in surgical decision-making. Surgery 2020;168:253-66. [Crossref] [PubMed]
41. Char DS, Shah NH, Magnus D. Implementing Machine Learning in Health Care - Addressing Ethical Challenges. N Engl J Med 2018;378:981-3. [Crossref] [PubMed]
42. Zou J, Schiebinger L. AI can be sexist and racist - it’s time to make it fair. Nature 2018;559:324-6. [Crossref] [PubMed]
43. Byrd TF 4th, Southwell B, Ravishankar A, et al. Validation of a Proprietary Deterioration Index Model and Performance in Hospitalized Adults. JAMA Netw Open 2023;6:e2324176. [Crossref] [PubMed]
44. Jobin A, Ienca M, Vayena E. The global landscape of AI ethics guidelines. Nat Mach Intell 2019;1:389-99. [Crossref]
45. Racine E, Boehlen W, Sample M. Healthcare uses of artificial intelligence: Challenges and opportunities for growth. Healthc Manage Forum 2019;32:272-5. [Crossref] [PubMed]
46. Lu JH, Callahan A, Patel BS, et al. Assessment of Adherence to Reporting Guidelines by Commonly Used Clinical Prediction Models From a Single Vendor: A Systematic Review. JAMA Netw Open 2022;5:e2227779. [Crossref] [PubMed]
47. Richardson JP, Smith C, Curtis S, et al. Patient apprehensions about the use of artificial intelligence in healthcare. NPJ Digit Med 2021;4:140. [Crossref] [PubMed]
48. Buolamwini J, Gebru T. Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification. In: Proceedings of the 1st Conference on Fairness, Accountability and Transparency. PMLR 2018;81:77-91.
49. Obermeyer Z, Powers B, Vogeli C, et al. Dissecting racial bias in an algorithm used to manage the health of populations. Science 2019;366:447-53. [Crossref] [PubMed]
50. Mehrabi N, Morstatter F, Saxena N, et al. A Survey on Bias and Fairness in Machine Learning. ACM Comput Surv. 2021;54:1-35. [Crossref]
51. Mari V, Spolverato G, Ferrari L. The potential of artificial intelligence as an equalizer of gender disparity in surgical training and education. Art Int Surg 2022;2:122-31. [Crossref]
52. Grote T, Berens P. On the ethics of algorithmic decision-making in healthcare. J Med Ethics 2020;46:205-11. [Crossref] [PubMed]
53. August AT, Sheth K, Brandt A, et al. The Value of Surgical Data-Impact on the Future of the Surgical Field. Surg Innov 2022;29:98-102. [Crossref] [PubMed]
54. Huriye AZ. The Ethics of Artificial Intelligence: Examining the Ethical Considerations Surrounding the Development and Use of AI. American Journal of Technology 2023;2:37-44. [Crossref]
55. Powles J, Hodson H. Google DeepMind and healthcare in an age of algorithms. Health Technol (Berl) 2017;7:351-67. [Crossref] [PubMed]
56. What is GDPR, the EU’s new data protection law? GDPR.eu. 2018. [accessed April 13, 2024]. Available online: https://gdpr.eu/what-is-gdpr/
57. Zhang A, Xing L, Zou J, et al. Shifting machine learning for healthcare from development to deployment and from models to data. Nat Biomed Eng 2022;6:1330-45. [Crossref] [PubMed]
58. Sarofim M. Devil’s advocate: exploring the potential negative impacts of artificial intelligence on the field of surgery. J Med Artif Intell 2024;7:7. [Crossref]
59. Lysaght T, Lim HY, Xafis V, et al. AI-Assisted Decision-making in Healthcare. Asian Bioeth Rev 2019;11:299-314. [Crossref] [PubMed]
60. Guerrero DT, Asaad M, Rajesh A, et al. Advancing Surgical Education: The Use of Artificial Intelligence in Surgical Training. Am Surg 2023;89:49-54. [Crossref] [PubMed]
61. 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]
62. Higgins D, Madai VI. From Bit to Bedside: A Practical Framework for Artificial Intelligence Product Development in Healthcare. Advanced Intelligent Systems 2020;2:2000052. [Crossref]
63. Vedula SS, Ghazi A, Collins JW, et al. Artificial Intelligence Methods and Artificial Intelligence-Enabled Metrics for Surgical Education: A Multidisciplinary Consensus. J Am Coll Surg 2022;234:1181-92. [Crossref] [PubMed]
64. Sarofim M. Navigating the inevitable convergence of artificial intelligence and surgical training programs. Surgeon 2024;22:e155-6. [Crossref] [PubMed]
65. Varas J, Coronel BV, Villagrán I, et al. Innovations in surgical training: exploring the role of artificial intelligence and large language models (LLM). Rev Col Bras Cir 2023;50:e20233605. [Crossref] [PubMed]
66. Julian D, Smith R. Developing an intelligent tutoring system for robotic-assisted surgery instruction. Int J Med Robot 2019;15:e2037. [Crossref] [PubMed]
67. Lohre R, Bois AJ, Pollock JW, et al. Effectiveness of Immersive Virtual Reality on Orthopedic Surgical Skills and Knowledge Acquisition Among Senior Surgical Residents: A Randomized Clinical Trial. JAMA Netw Open 2020;3:e2031217. [Crossref] [PubMed]
68. Dean WH, Buchan J, Gichuhi S, et al. Simulation-based surgical education for glaucoma versus conventional training alone: the GLAucoma Simulated Surgery (GLASS) trial. A multicentre, multicountry, randomised controlled, investigator-masked educational intervention efficacy trial in Kenya, South Africa, Tanzania, Uganda and Zimbabwe. Br J Ophthalmol 2022;106:863-9. [Crossref] [PubMed]
69. Tansley G, Bailey JG, Gu Y, et al. Efficacy of Surgical Simulation Training in a Low-Income Country. World J Surg 2016;40:2643-9. [Crossref] [PubMed]
70. Mirchi N, Bissonnette V, Ledwos N, et al. Artificial Neural Networks to Assess Virtual Reality Anterior Cervical Discectomy Performance. Oper Neurosurg (Hagerstown) 2020;19:65-75. [Crossref] [PubMed]
71. Baloul MS, Yeh VJH, Mukhtar F, et al. Video Commentary & Machine Learning: Tell Me What You See, I Tell You Who You Are. Journal of Surgical Education 2022;79:e263-72. [Crossref] [PubMed]
72. Winkler-Schwartz A, Yilmaz R, Mirchi N, et al. Machine Learning Identification of Surgical and Operative Factors Associated With Surgical Expertise in Virtual Reality Simulation. JAMA Netw Open 2019;2:e198363. [Crossref] [PubMed]
73. Abd-Alrazaq A, AlSaad R, Alhuwail D, et al. Large Language Models in Medical Education: Opportunities, Challenges, and Future Directions. JMIR Med Educ 2023;9:e48291. [Crossref] [PubMed]
74. Siu AHY, Gibson D, Mu X, et al. Employing Large Language Models for Surgical Education: An In-depth Analysis of ChatGPT-4. J Med Edu 2023;22:e137753. [Crossref]
75. Mohapatra DP, Thiruvoth FM, Tripathy S, et al. Leveraging Large Language Models (LLM) for the Plastic Surgery Resident Training: Do They Have a Role? Indian J Plast Surg 2023;56:413-20. [Crossref] [PubMed]
76. Price WN. Big data and black-box medical algorithms. Sci Transl Med 2018;10:eaao5333. [Crossref] [PubMed]
77. Panch T, Duralde E, Mattie H, et al. A distributed approach to the regulation of clinical AI. PLOS Digit Health 2022;1:e0000040. [Crossref] [PubMed]
78. Erion G, Janizek JD, Hudelson C, et al. CoAI: Cost-Aware Artificial Intelligence for Health Care. Nat Biomed Eng 2022;6:1384-98. [Crossref] [PubMed]
79. Kelly CJ, Karthikesalingam A, Suleyman M, et al. Key challenges for delivering clinical impact with artificial intelligence. BMC Med 2019;17:195. [Crossref] [PubMed]
80. Hashimoto DA, Rosman G, Rus D, et al. Artificial Intelligence in Surgery: Promises and Perils. Ann Surg 2018;268:70-6. [Crossref] [PubMed]
81. Pakkasjärvi N, Luthra T, Anand S. Artificial Intelligence in Surgical Learning. Surgeries 2023;4:86-97. [Crossref]
82. Park JJ, Tiefenbach J, Demetriades AK. The role of artificial intelligence in surgical simulation. Front Med Technol 2022;4:1076755. [Crossref] [PubMed]
83. He J, Baxter SL, Xu J, et al. The practical implementation of artificial intelligence technologies in medicine. Nat Med 2019;25:30-6. [Crossref] [PubMed]
84. Murdoch B. Privacy and artificial intelligence: challenges for protecting health information in a new era. BMC Med Ethics 2021;22:122. [Crossref] [PubMed]
85. Mohammadi F, Panou A, Ntantogian C, et al. CUREX: seCUre and pRivate hEalth data eXchange. In: IEEE/WIC/ACM International Conference on Web Intelligence - Companion Volume. New York, NY, USA: Association for Computing Machiner; 2019:263-8.
86. Vasey B, Nagendran M, Campbell B, et al. Reporting guideline for the early-stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AI. Nat Med 2022;28:924-33. [Crossref] [PubMed]
87. Cruz Rivera S, Liu X, Chan AW, et al. Guidelines for clinical trial protocols for interventions involving artificial intelligence: the SPIRIT-AI extension. Nat Med 2020;26:1351-63. [Crossref] [PubMed]
88. Liu X, Cruz Rivera S, Moher D, et al. Reporting guidelines for clinical trial reports for interventions involving artificial intelligence: the CONSORT-AI extension. Nat Med 2020;26:1364-74. [Crossref] [PubMed]
89. Cai CJ, Winter S, Steiner D, et al. “Hello AI”: Uncovering the Onboarding Needs of Medical Practitioners for Human-AI Collaborative Decision-Making. Proceedings of the ACM on Human-Computer Interaction 2019;3:1-24. [Crossref]
90. Kolachalama VB, Garg PS. Machine learning and medical education. NPJ Digit Med 2018;1:54. [Crossref] [PubMed]
91. Wood EA, Ange BL, Miller DD. Are We Ready to Integrate Artificial Intelligence Literacy into Medical School Curriculum: Students and Faculty Survey. J Med Educ Curric Dev 2021;8:23821205211024078. [Crossref] [PubMed]
92. Madaio MA, Stark L, Wortman Vaughan J, et al. Co-Designing Checklists to Understand Organizational Challenges and Opportunities around Fairness in AI. In: Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems. New York, NY, USA: Association for Computing Machinery; 2020:1-14.
93. O’ConnorSPeltonenLMRonquilloCA Framework for Embedding Artificial Intelligence in Nursing: Education, Innovation, Collaboration, and Implementation - the ALIGN Model.OSF Preprints; 2023. Available online: 10.31219/osf.io/3b8pc
94. Gould DJ, Dowsey MM, Glanville-Hearst M, et al. Patients’ Views on AI for Risk Prediction in Shared Decision-Making for Knee Replacement Surgery: Qualitative Interview Study. J Med Internet Res 2023;25:e43632. [Crossref] [PubMed]