'No such thing as a free lunch': Spine surgeons on the promise, pitfalls of AI

Spine

Artificial intelligence in spine surgery is in its infancy, and physicians are discussing the exciting and nerve-wracking aspects of it.

Ask Spine Surgeons is a weekly series of questions posed to spine surgeons around the country about clinical, business and policy issues affecting spine care. Becker's invites all spine surgeon and specialist responses.

Next question: Halfway through the year, what's one accomplishment you're proud of and what's a goal you have for the remainder of 2024?

Please send responses to Carly Behm at cbehm@beckershealthcare.com by 5 p.m. CST Wednesday, June 26.

Editor's note: Responses were lightly edited for clarity and length.

Question: When it comes to AI in spine surgery, what’s one thing you're excited about and one thing you're nervous about?

Neel Anand, MD. Cedars-Sinai Spine Center (Los Angeles): AI is paving the way for the type of innovation in spine surgery that will help our patients get the exact care they need, delivered with incredible precision, to the specific area of the spine where they need it, all while allowing for minimal invasion and faster recovery. It remains in the infancy stage, but the power and promise of AI in spine surgery are right here in front of us.

In the diagnosis of spinal conditions – AI shows great promise in helping healthcare providers achieve more accurate and timely diagnoses of spinal conditions. Using predictive-modeling type AI applications, the ability to amass loads of clinical data and predict which patients may end up developing spine conditions is advantageous. As with most medical issues, the earlier the diagnosis, the earlier the intervention – before problems become catastrophic.

In the surgical treatment of spinal conditions – When coupled with robot-assisted spine surgery, which is already in use, AI has the potential to predict which patients might benefit most from surgical intervention and can also help the precision and skill of surgeons when performing spine procedures. The predictive accuracy of AI in the operating room also offers the potential to help reduce surgeon fatigue. In many cases, most spine surgeries are relatively long – four hours or more. 

In the prediction of spine surgery outcomes – AI offers the potential for scientists and statisticians to predict spine surgery outcomes based on a patient's unique health metrics and circumstances. For example, the result of the same spine procedure on a 46-year-old sedentary female may not be the same or may be accompanied by additional considerations as the one performed on a 75-year-old active male. AI can help spine surgeons analyze the data available on both populations. In turn, spine surgeons can have honest conversations with their patients, and patients can make well-informed decisions about their spine surgery and health care decisions.

In the recovery and return to daily living for spine surgery patients – With addiction to opioid pain relief medication at an all-time high, one of the timeliest promises of AI is in the prediction of which patients will need post-operative pain management, how much they will need, and which medications are best. As a spine surgeon, I must ensure my patients can manage their pain well after surgery. Yet I don't want them to become addicted to the medication tools I prescribe during their healing. There is a delicate balance between just right and too much when it comes to pain medication. I am excited about the potential of AI to help pinpoint the right spot on the pain management continuum so that my patients can recover well without future risks of harm from pain medication.

While we might be a way off from harnessing the most potent potential of AI in all of the above areas of spine surgery, I am confident we will get there. Spine surgery research as it relates to AI is more proliferated than it was even five years ago, and if I know one thing about the history of spine surgery and the tenacious researchers behind it – where there is a will, there is a way.

Choll Kim, MD, PhD. Excel Spine Center (San Diego): Patients considering spine surgery are hungry for information. I imagine a day soon when there is an AI version of myself that has studied everything I have ever said and written, along with a selective body of information pertinent to my practice, that my patients, and their loved ones, have access to 24 hours a day, seven days a week, 365 days a year. They can ask me any question they want, for as long as they want, and the same question in as many ways as they want, without me getting tired or impatient. I hope and pray that this will free up my time to focus on higher level activities that I enjoy more. Just like the convenience of the internet and the apps on our mobile devices, AI will make life easier in many ways. But of course, there is no such thing as a free lunch. The tradeoff is that my office will be a lot lonelier, as the need for ancillary staff will be greatly diminished, and there will be many people that might prefer interacting with my AI more than me. For an extrovert, not having real people to spend time with is terrifying.  

Philip Schneider, MD. The Centers for Advanced Orthopaedics (Bethesda, Md.): AI and machine learning have demonstrated the ability to assist in surgical decision-making, patient selection, and predicting functional outcomes – which can all result in improved patient care. However, I am concerned that surgeons may become overly dependent on AI over time. All technology is prone to failure at some point, and my trust in AI is not yet fully established.

Christian Zimmerman, MD. St. Alphonsus Medical Group and SAHS Neuroscience Institute (Boise, Idaho): There certainly seems to be a societal press in medicine (like every other facet of our professional lives) to meld AI beyond our perspectives to a point of newfound dependence, should elevate nervousness amongst actual and proper practicing clinicians. Some aspects of discernable AI techniques such as risk stratification, image analysis and intraoperative navigation may be useful to some, the variability of patient symptoms and diversity of anatomic presentation requires more individualistic prescience. My fears of further dehumanizing the art and science of medicine will only be furthered and justified with added dilution of separation as this reliance and dependence grows.

The goals of mitigating doubt and errors in medicine and its delivery is paramount. Methods to do so are supported.

Copyright © 2024 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.