GI Genius Polyp Detection System: Interview with Giovanni Di Napoli, President, Gastrointestinal

We recently reported on the de novo FDA clearance received by Medtronic for its GI Genius AI polyp detection system. The product is hugely flexible as it works with any video colonoscope and provides AI powered assistance for clinicians in identifying colorectal polyps during colonoscopies.

Medtronic reports that the system has significantly improved polyp detection rates, which should translate to better colorectal cancer patient outcomes through prompt diagnosis and early treatment.

As a clinician performs a colonoscopy the system analyzes the video data and highlights suspected polyps using a green box on the video feed in real time.

GI Genius relies on a massive dataset of polyp images that power its identification efforts, and this is constantly growing, which should help to improve the system accuracy and sensitivity further.

Medgadget had the opportunity to speak with Giovanni Di Napoli, President, Gastrointestinal at Medtronic about the GI Genius system.

​

Conn Hastings, Medgadget: Please give us an overview of colon cancer, and how it is currently diagnosed.

Giovanni Di Napoli, Medtronic: Colorectal cancer is the third most common form of cancer globally, with 1.8 million new cases every year. It is estimated that 149,500 people will be diagnosed in 2021, and 52,980 people will die in 2021 of colorectal cancer. These are unacceptable cancer rates, and these deaths are preventable. In fact, 90% of patients can beat colorectal cancer when caught early. Colonoscopies are the gold standard of colorectal cancer diagnosis, and more than 10 million screening colonoscopies are performed in the United States each year. However, according to the CDC, this number should be higher as only about one-third of Americans who meet CRC screening criteria remain unscreened.

And now, with COVID-19, many more have delayed important medical appointments and procedures, and traditional GI screenings and care delivery declined substantially during the pandemic. And while many office visits are being conducted virtually due to COVID-19, virtual platforms dedicated to GI care are also emerging as a standard practice and will benefit patients and healthcare providers in a post-COVID environment as well.

Medgadget: How often do clinicians miss polyps during a colonoscopy and how does this affect patient prognosis?

Giovanni Di Napoli: Some research has shown that 17% to 28% of colorectal polyps are missed during colonoscopies. Detection of all colorectal polyps regardless of size, shape, or morphology can lead to better clinical outcomes for patients and reduce the risk of interval cancers that can occur between colonoscopies. Adding AI can increase the quality of colonoscopies, which can improve diagnosis and outcomes for colon cancer patients, and we are leveraging technology with GI Genius to improve healthcare.

Early diagnosis and treatment are important for achieving the best outcome for GI patients, and patients should follow their GI’s recommendations in a timely manner. While we are all staying home to stay safe, missing or delaying medical care can result in late diagnosis, delayed care, and negatively impact your GI health. Telehealth options exist for patients hesitant to make an appointment amid the pandemic. In fact, Medtronic created PillCam™SB3@HOME to modernize clinical GI services with a telemedicine option. It is an at-home option for accurate, timely, and convenient GI diagnostic tests during COVID and beyond. The program leverages our PillCam technology, which has been approved for 20 years, with Amazon logistics for delivery to ensure timely and accurate results.

Medgadget: Please give us an overview of GI Genius and what it does.

Giovanni Di Napoli: GI Genius™ uses advanced AI to highlight the presence of precancerous lesions with a visual marker in real-time – serving as an ever-vigilant second observer. The GI Genius™ supplements procedures with real-time image analysis of any colonoscope video and alerts clinicians to the presence of lesions — including those with flat (non-polypoid) morphology — with a visual marker on-screen. This can increase diagnostic accuracy, improve the efficacy rate of polyp detection, and reduce the risk of interval cancers that can occur between colonoscopies. GI Genius™ adds to our portfolio of innovative GI solutions to improve patient outcomes through disruptive technologies.

Medgadget: How did you train the AI system to recognize polyps?

Giovanni Di Napoli: The software in the GI Genius™ module has a robust dataset of over 13 million polyp images of various shapes and sizes. And, as the datasets grow, the GI Genius™ intelligent endoscopy module continues to learn. We’re already releasing additional data sets that will lead to tens of millions of more images and videos feeding the system, ultimately leading to better performance.

Medgadget: So, what types of increases in polyp detection have you observed when the system is used?

Giovanni Di Napoli: Use of the GI Genius module has demonstrated a 30% relative increase in ADR compared to colonoscopy alone for both flat (42% increase) and polyploid (36% increase) lesions, thus increasing accuracy and helping reduce the risk of interval cancers which can occur between colonoscopies.

Medgadget: Is the system easy to use? Does it run passively in the background while a colonoscopy is in progress?

Giovanni Di Napoli: The GI Genius intelligent endoscopy module integrates easily and seamlessly with all major brands of endoscopes and has no impact on the current clinic workflow. Different facilities, and clinicians, have different needs. The platform is highly flexible and can be integrated into the existing equipment of each facility.

Source