Artificial intelligence (AI) is defined as algorithms that enable machines to perform cognitive functions, such as problem solving and decision-making. AI has been changing the face of healthcare for some time through robot learning (ML) and natural language processing (NLP).
However, its use in surgery has taken longer than in other areas of medicine, mainly because of the lack of information on how to implement the algorithm in actual surgery. Artificial intelligence is advancing so rapidly that it is now seen as complementing, rather than replacing, the skills of human surgeons.
While the potential of the "surgeon-patient-computer" relationship is still a long way from being fully explored, the use of ARTIFICIAL intelligence in surgery is already bringing about significant changes for doctors and patients. For example, continuous improvements in surgical planning and navigation through CT and MRI, while minimally invasive surgery combined with robot-assisted surgery have reduced surgical trauma and improved patient recovery.
How ARTIFICIAL Intelligence shapes "Preoperative planning"
Preoperative planning is the stage in which surgeons plan surgical interventions based on the patient's history and images, using general image analysis techniques and traditional machine learning for classification. Deep learning is enabling pre-operative planning and is used for anatomical classification, detection segmentation and image registration.
Deep learning algorithms are able to identify abnormalities from CT scans, such as skull fractures, intracranial bleeding, and midline displacement. Deep learning enables emergency care for these abnormalities and represents a potential key technology for automating patient identification in the future.
Deep learning recurrent neural networks (RNN) have been used to predict renal failure as well as mortality and postoperative bleeding after cardiac surgery in real time, with better results than standard clinical reference tools. These findings, obtained entirely through the collection of clinical data without manual manipulation, could improve intensive care by focusing more on patients most at risk of such complications.
The role of AI in intraoperative guidance
Computer aided intraoperative guidance has been regarded as the basis of minimally invasive surgery. AI learning strategies have been implemented in several areas of minimally invasive surgery, such as tissue tracking.
Accurate tracking of tissue deformation is very important for intraoperative guidance and navigation of minimally invasive surgery. Unable to accurately shape tissue deformations through "temporary representations," scientists have developed an algorithm-based online learning framework that provides a tracking method for identifying operations in the human body.
Artificial Intelligence AIDS 'surgical robot'
The AI-powered surgical robot is a computer-operated device that assists in the operation and positioning of instruments during surgery, enabling surgeons to focus on complex aspects of surgery.
Their use reduces surgeons' volatility during surgery, helps them improve their skills and perform better during surgery, resulting in superior patient outcomes and lower overall healthcare expenditures.
With the help of machine learning technology, surgical robots identify key information and best practices by browsing millions of data sets. Asensus Surgical's performance-guided laparoscopic AI robot can provide surgeons with critical information, such as tissue size, without the need for a physical tape measure. Meanwhile, robots are programmed by demonstrating human skills; Robots learn by mimicking surgeons.
Learning from Demonstration (LfD) "trains" robots to independently perform new tasks by accumulating information. In the first stage, LfD breaks down complex surgical tasks into several sub-tasks and basic gestures. In the second phase, the surgical robot provides human surgeons with the opportunity to break away from repetitive tasks in the order of identifying, modeling, and performing subtasks.
Expanding the use of autonomous robots in surgery, especially for performing tasks in minimally invasive surgery, is a daunting endeavor. JIGSAWS (Johns Hopkins Institute for Information Security Gesture and Skill Assessment Working Set) is the first public benchmark surgical activity dataset that provides kinematic data and synchronous video of three standard surgical tasks performed by the university's surgeons of varying skill levels.
The three standard surgical tasks analyzed by JIGSAWS were suturing, needle-threading and knotting. Gestures were recognized with about 80 percent accuracy during each task. The results, while promising, suggest there is still room for improvement, especially in predicting gestural activity among different surgeons.
For many surgical tasks, reinforcement learning (RL) is a commonly used machine learning paradigm for solving subtasks that are difficult to model accurately, such as intubation and soft tissue manipulation. The RL algorithm is formed based on strategies learned from demonstrations rather than from scratch, thus reducing the time required for the learning process.
Examples of AI-assisted surgery
Human-robot interaction is an area that allows surgeons to operate surgical robots in a non-contact way. This can be done with head or hand movements, speech recognition, and the surgeon's gaze.
Surgeons can already control laparoscopic robots remotely by moving their heads. FAce MOUSe is a human-machine interface that monitors surgeons' facial movements in real time without any physical contact with the device. Laparoscopic movements are simply and accurately controlled by the surgeon's facial "gestures," providing non-invasive, verbal cooperation between human and robot for a variety of surgical operations.
In 2017, the Maastricht University Medical Center in the Netherlands used an AI-powered robot in a microsurgical intervention. The surgical robot sutured 0.03 to 0.08 mm of blood vessels in a patient affected by lymphedema. The chronic disease is often a side effect of swelling caused by fluid buildup during treatment for breast cancer. The robot used in the surgery was built by Microsure and operated by a surgeon. His hand movements are made smaller and more accurate by a robot hand. Surgical robots are also used to repair tremors in surgeons' movements, ensuring that ai-powered devices perform surgery correctly.
The hair transplant surgery robot enables the robot to collect hair follicles and transplant them to precise areas of the scalp with the help of AI algorithms. The robot can perform minimally invasive surgery without having to surgically remove the donor area, and it eliminates the need for hair transplant surgeons to manually extract hair follicles over several hours of surgery, which can only be done one at a time.
Da Vinci heart surgery is robotic heart surgery, cutting through very small incisions in the chest with robot-operated instruments and very small instruments. Robotic heart surgery has been used for various heart-related procedures, such as coronary artery bypass grafting, valve surgery, heart tissue ablation, tumor removal and repair of heart defects.
Gestonurse is a nurse robot designed specifically for surgeons handling surgical instruments in the operating room, with the goal of reducing errors that can occur that negatively affect surgical outcomes. Gestonurse used fingertip recognition and gesture inference to operate the required instruments during simulated surgery at Purdue University, proving its efficiency and safe use.
Liz Kwo is the Deputy Chief clinical Officer of Anthem, Inc. She is a serial healthcare entrepreneur, physician and lecturer at Harvard Medical School. She received an M.D. from Harvard Medical School, an M.B.A. from Harvard Business School and a Master of Public Health degree from Harvard T.H. Chan School of Public Health.
Manly Battery which located in Shenzhen,China. A leading Lithium battery manufacturer over 12 years ,widly used for Robostic industry,if there is any project need to evluate ,pls feel free tosend email to info@manlybatteries.com
