Artificial Intelligence in the GI Lab

Artificial Intelligence has made its way into almost every aspect of our lives. We see it in self-driving cars, search engines, and even gaming. We are now seeing it in the GI lab with a computer program that detects polyps. Specialties General Specialties Knowledge

Updated:   Published

Artificial Intelligence in the GI Lab

What is Artificial Intelligence?

We all know the concept of Artificial Intelligence, and understand the use it has in our lives. Whether you appreciate it or not, it is here to stay and will only continue to weave it's way into human life. There are two broad types of A.I., narrow and general1. Heath tells us in his article, "What is AI? Here's Everything You Need to Know About Artificial Intelligence", that narrow is defined by its use with defined tasks (Siri), and general is a more malleable form that can learn and do specific tasks1. A.I. is a system that can learn and modify, and then affix that knowledge to new situations. This ability to learn and adapt to new scenarios can now be used to detect adenomas in the colon during a colonoscopy.

Adenoma Detection Rate

Gastroenterology physicians perform colonoscopies among other procedures. Colonoscopies are meant to screen patients beginning at age 50 for precancer or cancer. The new recommended age is 45 by the American Cancer Society, so hopefully insurance will follow suit soon. If you have ever seen a colonoscopy, you know that there are many folds and corners that polyps can hide. There are many factors to consider in how well an exam results in finding polyps. The prep must be good so  that the doctor can see well. Also, the equipment must be adequate and working, and the physician must  spend the recommended six minutes withdrawing in order to be able to find polyps.

The adenoma detection rate is the percentage of patients in which polyps are found in screening colonoscopies. Since colon cancer is preventable through these screenings, the more polyps that are detected decreases a patient's chances of colon cancer. If the polyps are removed and then studied by a pathologist, a patient can be sure of a diagnosis. This diagnosis helps the doctor to define the path of treatment that best fits the patient.

Artificial Intelligence in GI

Colon cancer has a high incidence in America, the third highest in both men and women according to the American Cancer Society. They estimate that there will be 104,270 new cases of colon cancer this year, and 45,230 of rectal cancer2.   Education combined with screenings is super important in decreasing cases.

Using this new artificial intelligence program, decreased screening age, and increased patient awareness, the number of  colon cancer cases will decrease in our country. Too many cases are found in advanced stages and that increases the patient's mortality. The A.I. program uses a computer program that has a very accurate data set that can recognize polyps of all shapes and sizes3. When an abnormal area of tissue is detected, it puts a circle around it on the screen so that the doctor can go right to it.

Medtronic has developed a program, one that has a 99.7% sensitivity3. The program studied 2684 lesions that were proven with pathology in order to learn what to look for3.

In the GI lab, using this program decreases procedure time by pointing out polyps to the physician. The physician can then remove the polyp and continue on with the exam. Human eyes are good and combined with this program, no polyp will be missed. With a glance away from the screen, a polyp can slip past, not to be seen until years later when the patient returns for another screening or diagnostic colonoscopy. And literally, this could mean the difference between having surgery or not.

I have seen a sample of this program, and I hope to one day get it for the lab that I work in. Eventually, this program should be built into all of the processors so that every patient can have the best outcome possible.


1What is AI? Here's everything you need to know about artificial intelligence

2About Colorectal Cancer

3Use of artificial intelligence in improving adenoma detection rate during colonoscopy: Might both endoscopists and pathologists be further helped

Gastrointestinal Columnist

Brenda F. Johnson, MSN, RN Specialty: 25 years of experience in Gastrointestinal Nursing

61 Articles   326 Posts

Share this post

Share on other sites
Specializes in Ambulatory Care, Community Health, HIV.

Thank you for the article. While I agree with you that AI will inevitably become more pervasive in every aspect of life, it still really creeps me out. I also wonder if physicians using this might be lulled into a false sense of security and skimp on their own visual contribution. Just a thought. 

By using the site, you agree with our Policies. X