HHIT Series Episode 9: Using artificial intelligence and allied technologies to deliver precision endoscopy and imaging to gastroenterology, with Dr. Michael Byrne
On Tuesday, November 21st 2023, Dr. Michael Byrne, Clinical Professor of Medicine in the Division of Gastroenterology at Vancouver General Hospital and the University of British Columbia, and CEO/founder of (opens in a new window)Satisfai Health, spoke to host Prof. Patricia Maguire about using artificial intelligence and allied technologies to deliver precision endoscopy and imaging to gastroenterology on our 15-minute Human Health Impact Technology webinar series (#HHITseries).
Human stagnation
Satisfai Health was inspired by Byrne’s dual passion for entrepreneurship and addressing a critical gap in medical technology in his field of gastrointestinal (GI) medicine. Though advancements have been made in endoscopy technology (whereby a camera is taken through the bowel and intestine), the ability to extract valuable information from the data on screens has not kept pace.
“Human performance has stagnated,” he said, which motivated Byrne to explore digital technologies as a solution in medical diagnostics. The goal of Satisfai Health is to use AI to extract more information during live endoscopy.
Optical Biopsy
Byrne envisions a ground-breaking shift in medical diagnostics through the concept of ‘optical biopsy’ or virtual pathology. This innovative approach involves real-time analysis during medical procedures, such as colonoscopy to look at the colon or gastroscopy to look at the oesophagus, eliminating the need for physical biopsies and the associated waiting time. By leveraging machine capabilities to identify pathology types within “milliseconds”, this approach not only enhances diagnostic accuracy but also offers significant cost savings. It enables the physician to tell the patient “there and then what’s going on”. In the case of human optical biopsy of a polyp, physicians are trained to look for three features: colour, pit pattern and vascular pattern. By contrast, the machine is “looking at 1000 or more features per polyp--things that we can’t even perceive as humans”.
“I have no doubt in the old adage that I've seen on many slides in the last couple of years, which is that AI won't replace physicians, but physicians who use AI will replace physicians who don't. I truly believe in that.”
Perfect harmony
With the rising prevalence of colon cancer globally, especially among younger people, Byrne underscores the critical importance of colon screening. Most colon cancers stem from polyps or adenomas which are often undetected.
“Many of us have colon polyps we don’t even know about,” he added, explaining that the Adenoma Detection Rate (ADR) is a key metric in their detection.
In an “ideal scenario”, ADR rates should be up to 70% but for “many people it’s 20/30/40% and machines have already been shown to improve that by at least 8-10%”. Just a 1% ADR increase can decrease colon cancer mortality by 3%. Byrne advocates for a synergistic approach, combining the skills of physicians with the capabilities of machines, “in a kind of perfect harmony”, to achieve optimal screening outcomes.
Targeted biopsies
He highlights the rising incidence of oesophageal cancer, particularly in women, and the challenges doctors face in accurately assessing the upper GI tract. Unlike traditional non-targeted biopsies, he envisions a paradigm shift towards targeted biopsies using machines. By adopting a more precise approach that targets specific areas in the oesophagus based on machine insights, there is potential for earlier and more accurate cancer detection, ultimately reducing the significant uncertainty associated with early-stage oesophageal cancer.
Pushback
While acknowledging the potential reservations from patients about machines diagnosing health conditions, Byrne argues that the significant miss rates in gastrointestinal procedures and the demonstrated ability of machines to improve detection rates should outweigh concerns. As for getting physicians on board with the technology, he says if their ADR is low, even a modest increase in performance with the help of AI can lead to substantial improvement, especially when this is applied right across the medical community.
Who pays?
A critical barrier to the widespread adoption of AI in medicine are the substantial cost implications. Byrne says while the adoption of AI is progressing in radiology, there is a lack of clarity regarding reimbursement for AI tools in the GI space. He calls for action from financial and health organisations, like the NHS in the UK and HSE in Ireland, to help to resolve this.
Exciting journey
Byrne says the AI in healthcare space is at “the beginning of a very exciting journey”, adding: “I have no doubt in the old adage that I've seen on many slides in the last couple of years, which is that AI won't replace physicians, but physicians who use AI will replace physicians who don't. I truly believe in that.”