In a recent review published in Nature Medicine, researchers investigated the use of artificial intelligence (AI) in surgery, focusing on improving pre-, intra- and post-operative care.
study: Artificial intelligence in surgery. Image credit: LALAKA/Shutterstock.com
background
AI is making significant advances in healthcare, including fundamental model architectures, wearable technology, and surgical data infrastructure. Research shows that artificial intelligence can exceed or complement human skills, especially in radiology.
However, surgery remains a slow-growing specialty, with global differences in access, complications, and post-complication mortality.
A comprehensive approach to strengthening surgical systems is needed, including enhancing access, education, detecting and treating problems, and improving system efficiency.
About reviews
In this review, researchers highlight the potential of AI technologies to improve patient outcomes, surgical education, and care optimization, highlight existing deep learning applications, and explore future possibilities using multimodal foundational models. expected development.
Preoperative surgical application of AI
Artificial intelligence (AI) is being utilized in surgical planning and patient selection, particularly in preoperative imaging, to support early diagnostic assessment and planning. A model-free method for reinforcement learning shows promise for detecting and removing tumor tissue while limiting the impact on functioning living tissue during neurosurgical procedures.
Innovative techniques for preoperative planning use AI and virtual reality-based segmentation algorithms to significantly improve surgical techniques.
Accurate preoperative clinical diagnosis is essential for treatment decisions and planning. RadioLOGIC AI algorithms extract unstructured data from radiology reports in medical records to improve radiology diagnoses.
Advances in large-scale language models (LLMs) and their interaction with electronic data systems have the potential to enable early detection and treatment of diseases before they progress.
Diagnosis is a highly developed aspect of surgical artificial intelligence, and the versatility and accuracy of the models will enable early clinical use. Task-specific breakthroughs may enable more accurate tumor staging and improve surgical planning.
Intraoperative surgical application of AI
In surgical practice, the intraoperative phase is a data-rich environment in which physiological indicators and complex injuries are continuously monitored.
Advances in intraoperative computer vision have enabled anatomical analysis, tissue characterization, cut surface evaluation, pathology detection, and reliable instrument identification. However, the data that can be recorded, evaluated, and collected in the operating room is limited.
Researchers must use valuable intraoperative data streams to aid surgical automation and artificial intelligence in the operating room.
AI in surgical decision-making can help enhance surgical resection margins, reduce operative time, and improve efficiency. Recent patient-independent transfer learning neural networks use rapid nanopore sequencing to provide accurate intraoperative diagnosis within 40 minutes.
Multimodal AI interrogation could help determine relevant anatomy, enhance the surgeon's visual assessment, provide biopsy information, and quantify cancer risk.
Digitized surgical platforms have the potential to pave the way to an AI-enhanced future, and investing in them is essential to take advantage of digital advances. Intraoperative apps are essential in the future of surgery as they improve non-technical activities such as communication, collaboration, and performance assessment.
Postoperative surgical application of AI
Home health care aims to increase access and equity in health care by allowing patients to recuperate in a familiar environment and resume normal functioning within society.
Although progress has been made in shortening postoperative hospital stays, achieving early discharge, promoting functional recovery through minimally invasive surgical techniques, allowing early return to daily life, improving postoperative evaluation, and providing early warning systems, the data is insufficient. Innovation based on this is lacking. During the postoperative period.
Wearables provide continuous monitoring by allowing input of multimodal physiological indicators, aiding in data-driven discharge planning. A systematic review identified 31 wearable devices that monitor physiological data, vital signs, and physical activity.
Postoperative complications are difficult to predict because there are many variables that influence care and outcome. Early diagnosis of complications, especially life-threatening complications such as anastomotic leaks and pancreatic fistulas, is important for healthcare systems to reduce mortality.
MySurgeryRisk is one of the few advances in predicting complications using machine learning algorithms. However, knowledge regarding the scalability of these algorithms to other health systems is limited.
Researchers developed an AI-driven home rehabilitation model that incorporates real-time data collection, advanced assessments of daily living measures, and novel assessments of ADLs.
conclusion
Based on the review results, AI technology can improve surgical treatment by optimizing patient selection, intraoperative performance, and operating room procedures. Transformers, a breakthrough in neural network design, enable multimodal AI models for critical surgical applications.
These include clinical risk prediction, automation, computer vision in robotic surgery, intraoperative diagnostics, enhanced training, sensor-based postoperative monitoring, resource management, and discharge planning.
However, thorough scrutiny and regulatory oversight are required, and stakeholder engagement is essential to providing improved surgical care. Research must guide AI systems by providing benefits such as accurate diagnosis and increased system efficiency.
