Medicine used to be extremely hierarchal. Decisions tend to be made based on someone’s experiences rather than objective measurement. The emergence of evidence-based medicine changed it all. The change believed would continue as data from various sources are becoming available. At the same time, new trends are also being introduced. For example, Surgical Data Science (SDS) was in the limelight a few years ago. Basically, SDS work out meaning from Big Data to facilitate technologies like decision making systems, surgical robotics, speech recognition and so on.
The promising SDS applications
Traditionally, clinical guidelines and textbooks narrating high level concepts are what drive surgical procedure planning. The surgeon will use these knowledge as well as his or her own experiences to execute surgical plans tailored for the patients. SDS could help to develop a system for preoperative planning and execution of procedures. The system will digest patients’ data and other granular details and allows surgical plans to be carried out with greater precision.
During the surgery, preoperative planning will be used as a model for Computer Integrated Surgery system (CIS) to execute various tasks on robotic platforms. SDS could create a framework for recording, storing, and annotating data generated by the CIS so that the information can be used for a wider variety of clinical applications.
Besides, CIS often faces the challenge of integrating automated procedures into the surgical workflow. Most surgeons are also not comfortable of high level automation due to legal reasons. SDS may render insights and make possible the development applications that observe surgical scene, detect surgical phases and access when it will be appropriate for surgeons to automate procedures under their close supervision.
Machine learning models process large amount of data to provide operating surgeons with decision support. SDS also sift out correlations and trends which highlight important surgical consequences to ascertain patients with more cohesive care. In terms of training, SDS may contribute to the creation of surgery simulations, providing trainees with more realistic scenarios or experienced surgeons who wished to be trained up in certain workflow.
The development of SDS remains slow
In spite of the promises given by SDS, its development and application remain slow. In fact, as Dr. Anthony Chang, AIMed founder and Chief Artificial Intelligence Officer at Children’s Hospital of Orange County (CHOC) noted at the recent AIMed webinar: Key trends in Surgical AI, although most surgeons have a strong mathematics or engineering background, yet when it comes to artificial intelligence (AI) discussions and even adoptions, surgery does not have the kind of robustness or velocity as witnessed in other medical sub-specialities like radiology or cardiology.
Dr. Chang wondered if it could be an illusion but the two speakers – Dr. Max R. Langham Jr., Professor of Surgery and Pediatrics at the University of Tennessee Health Science Center and Dr. Thomas Ward, Artificial Intelligence and Innovation Fellow at the Massachusetts General Hospital’s Surgical Artificial Intelligence and Innovation Laboratory (SAIIL) believed it’s not. This is probably due to the fact that surgeons need to spend years of their career on rigorous clinical training.
AI, thus far, is still consider fairly novel in medicine, hence, time is needed to bridge the gap and get more people to acquire it. Moreover, AI has a relatively high entry barrier for complete novice. One will need to put in some effort in learning before he or she can start to use related terminologies, so that’s why surgeons are still stuck in the early days.
Nevertheless, the AI in medicine community is looking forward to a change and on 23 September, AIMed will be hosting AIMed Surgery virtual event to connect the likeminded. Feel free to join us in this revolution by registering your interest or obtain a copy of the event agenda here.