To provide objective performance evaluations to surgeons, researchers at Caltech and urologists at Keck Medicine of USC have developed a novel artificial intelligence system called the Surgical AI System (SAIS). Surgeons-in-training typically require the guidance of more seasoned physicians throughout this phase of their education. A new AI-powered SAIS system could soon alter that.
SAIS is analysing videos of surgeries to determine what kind of surgery is being performed and how well it is being conducted. The objective is to inform surgeons of their strengths and areas for growth to serve their patients better.
According to the study’s senior author and Bren Professor of Computing and Mathematical Sciences Anima Anandkumar, AI can only replace human surgeons sometime soon in high-stakes scenarios like robotic surgery.
“Instead we asked how AI can safely improve surgical outcomes for the patients,” he revealed, which is why the team is concentrating on improving human surgeons’ skills and efficiency with the help of AI.
Surgeons’ performances were evaluated down to the level of individual discrete motions, and the SAIS was trained on a significant volume of video data annotated by medical specialists. Some things that surgeons are evaluated on include how well they grip a needle, how well they drive it through tissue, and how well they pull it out of the tissue. After completing the programme, SAIS was entrusted with reviewing and rating the work of surgeons across many institutions utilising footage of surgical procedures.
Dani Kiyasseh, the study’s lead author, said that SAIS could give surgeons reliable, consistent, and scalable feedback. The tool’s value was increased by the development team’s work to teach the AI to provide thorough feedback on its rationale for generating skill assessments, including links to specific video segments. The goal of SAIS is to help surgeons by pointing out areas where they may improve.
When first putting SAIS through its paces, researchers noticed that the AI occasionally gave surgeons higher or lower ratings for proficiency than their actual level of experience would have warranted based only on a study of their movements. To solve this problem, researchers trained the AI system to pay attention only to essential details in the surgical video, which helped to reduce the bias but not eradicate it.
Dr Andrew Hung, an associate professor of urology at the Keck School of Medicine of USC and a urologist at Keck Medicine of USC explained that human-derived surgical feedback is neither unbiased nor scalable at present, so AI-derived feedback like that provided by SAIS represents a significant opportunity to give surgeons access to helpful information.
Andrew Hung, an associate professor of urology at the Keck School of Medicine of USC and a urologist at Keck Medicine of USC mentioned, “Human-derived surgical feedback is not currently objective nor scalable. However, our method offers a significant possibility of giving surgeons actionable feedback from artificial intelligence.”
There have been many attempts to improve healthcare infrastructure in the United States by incorporating technological innovations. Scientists at the United States National Institute of Standards and Technology (NIST) have recently created biosensors for the early diagnosis of diseases like cancer.
The biosensor can identify biomarkers because it can measure the strength of DNA threads’ bonds to the sensor. Its modular design sets it apart from other sensors of its kind and helps keep prices down by making mass production easier and reusing the costliest components.
Expanding the availability of high-quality diagnostics is a revolutionary biosensor chip designed to be accurate and inexpensive. Because of the biomarkers, doctors can make timely, accurate diagnoses and provide customised treatments for the simple reason that conventional methods of screening may be too laborious, too expensive, or too limiting.
