Exploring the potential of AI for gynecologic surgery

The world of artificial intelligence (AI) is ever-evolving and diverse. With its multitude of uses, AI can be a useful tool to implement into your practice. Gaby Moawad, MD, FACOG, minimally invasive gynecologic surgeon (MIGS), The Center for Endometriosis and Advanced Pelvic Surgery, McLean, Virginia, spoke about the benefits of using AI for minimally invasive gynecologic surgery and its future potential in the specialty at AAGL’s 50th Global Congress on MIGS in Austin, Texas.¹

Moawad’s presentation was the first MED TALK of AAGL, a newly introduced TED Talk-style lecture at the congress.

“We have been using virtual reality for quite some time to help surgeons, whether in training or proacting surgeons, to be accustomed to operating room scenarios in a safer way. Virtual reality can also be incorporated into preoperative planning of surgery when used in unison with AI, it helps map the surgery beforehand,” said Moawad.

With AI, there is no creative scenario, and instead, there is an actual event that is being altered in real-time, according to Moawad, helping surgeons optimize their visual field or surgery. These scenarios are created through preoperative construction that uses any available data and disruptive technology based on computer vision and AI. “The surgeon in their OR can now have an enhanced vision of an anatomy they could not see with their bare eyes,” he added.

Moawad said an example of this is using augmented reality to visualize adenomyomas. He added that being able to incorporate this tech into surgery will help MIGS to do more precise surgery and improve the quality of surgical care for patients.

He also explained that AI is the theories, techniques, and information systems capable of completing a task that would otherwise require human intelligence. Notably, AI requires 2 things—data and an algorithm.

Under the umbrella of AI, there is machine learning and deep learning. In machine learning, the computer is fed a data set as a curriculum and uses an algorithm to complete a task. Deep learning is a subset of machine learning that consists of neuron networks that will process complex, layered regression models and algorithms to come up with a decision.

Moawad also explained the significance of artificial neurons. “An artificial neuron is basically a mathematical function…We have multiple variables fed into a system with each different algorithm to analyze different tasks to come up with a decision.”

In surgery, deep learning is used to teach the machine the location of the uterus. From this point, other images can be added in to help the machine paint a better picture of the anatomy. By overlaying these images, can help identify the optimal way to do the incision with minimal bleeding and less damage to the fibers of the uterus, according to Moawad.

“An important aspect to recognize is the huge mountain we need to climb to reach safer care. And I believe safer care comes with the evolution of newer surgical techniques and technology.

Reference:

1. Moawad G. AI for Dummies: From Machine Learning to Automation Surgery. Presented at: MIGS in Austin, Texas ; November 15, 2021; Austin, Texas.