This transcript has been edited for clarity.
What is artificial intelligence (AI)? This technology term refers to the ability of a computer to perform tasks usually associated with humans. One AI example is the ability for computers to learn, a highly complicated cognitive function that mimics the human mind. In machine learning, a computer can "learn" data, such as medical image pattern recognition, without even being programmed.
Over the past few years, researchers have been working to apply AI to the field of gastroenterology, with great success.
Colorectal cancer is the second leading cause of cancer death in the United States. As gastroenterologists, we are all excited about potentially using the power of AI to assist us with image analysis during colonoscopies to help identify polyps. Computer-aided detection software is being used to improve adenoma polyp detection rates. By removing these polyps during colonoscopy procedures, we can prevent colon cancer in our patients.
The US Food and Drug Administration (FDA) recently approved a really cool technology called GI Genius. This is AI/machine learning program that aids polyp detection in real time, with a tiny square box highlighting the polyp, kind of like those radars locked onto military targets in airplane movies such as Top Gun.
AI can be used to find both malignant lesions and premalignant lesions. A recent meta-analysis by Barua and colleagues, including five randomized controlled trials and 4311 patients, showed that the adenoma detection rate was much better using a combination of computer-aided detection software with colonoscopy (29.6%) compared with colonoscopy alone (19.3%).
Computer-aided detection with colonoscopy can also improve the detection of tiny or diminutive polyps < 5 mm in size.
Using AI in GI and Liver Diseases
We've also shown that AI can help gastroenterologists detect dysplasia, esophageal cancer, and gastric cancers when using esophagogastroduodenoscopy, or the upper scopes we use to look in the patient's esophagus and the stomach.
Liver disease causes 2 million deaths worldwide, and the prevalence of nonalcoholic fatty liver disease is 25%. We've also been using AI to help make liver diagnoses for several years using ultrasound technology, to help analyze the severity of fatty liver disease and to determine the stage of liver fibrosis. AI can help radiologists and gastroenterologists interpret ultrasound images and differentiate between benign liver lesions and liver cancer.
This is important because of the potential element of human error caused by exhaustion from reading a high volume of images or from challenges with interpretation. Currently, we are using AI algorithms such as traditional machine learning and deep learning. These are really important AI applications because ultrasound is so widely available and really inexpensive.
Radiomics is a kind of machine learning that is popular in CT, MRI, and ultrasound. This AI in imaging can help physicians identify complicated patterns that are difficult to recognize with the human eye.
Machine learning has also been used to distinguish ulcerative colitis from Crohn's disease in pediatric patients.
Artificial neural network has also been used to predict how frequently patients with inflammatory bowel disease are going to flare, with a pretty high level of accuracy, by analyzing data from our health records. It's our hope that doctors in the future will be able to use AI to select successful medical therapies for patients with ulcerative colitis and Crohn's disease.
Artificial neural network also has been used to predict 1-year mortality with 90% accuracy in patients with cirrhosis. This might help us with liver transplant evaluations in the future.
We can also use AI to predict prognosis after thermal ablation of liver cancer. With the help of AI in liver cancer, we can predict curative potential by employing transarterial chemoembolization.
AI can also be used to improve the diagnostic capabilities of endoscopic ultrasound, which is a kind of ultrasound probe attached to a scope to detect pancreatic cancer and intraductal papillary mucinous neoplasms. AI has also been used to help gastroenterologists detect pancreatic cancer based on both endoscopic ultrasound and serum markers. It's our hope that in the future, AI may help us detect pancreatic cancer in earlier stages and help us save patients' lives.
Oncologists can now use endoscopic ultrasound with AI technology to provide additional information about rectal cancer before surgery and to improve treatment.
A Technology Ready for Primetime Use
The FDA has developed regulatory plans for closely evaluating and approving updates and modifications to AI software technology. AI software producers will be able to update software for scopes, ultrasound, CT, and MRI remotely and instantaneously via Wi-Fi, just like our smartphone devices constantly update today.
Now, here's something really cool. AI can now help us monitor patients from home, the holy grail of medicine. For example, we have the capability to measure fecal calprotectin, an inflammatory marker in inflammatory bowel disease patients, using smartphones, which could help us better recognize the beginnings of a flare-up before our patient even feels sick, and allow physicians the opportunity to intervene earlier.
This is an incredibly exciting and revolutionary time within gastroenterology. Technological advances in imaging capability and computing power are synergistically allowing us to detect precancerous lesions and cancer more easily while performing endoscopy. Gastroenterologists are ready to embrace these emerging technologies as they become ready for primetime clinical use.
Benjamin H. Levy III, MD, is a board-certified gastroenterologist at the University of Chicago. His clinical projects focus on healthcare disparities and the development of colon cancer screening campaigns. Previously, as division head of gastroenterology at Mount Sinai Hospital, Dr Levy organized a gastroenterology clinic for refugees resettling in Chicago. He is a member of the American College of Gastroenterology (ACG) Public Relations Committee and FDA-Related Matters Committee. During the COVID-19 pandemic, Dr Levy started an international health education campaign named "Concerts & Cocktails" which teamed up musicians with physicians and nurses on the front lines. He was subsequently selected to be a speaker at TEDxWrigleyville for "Humanity: An Inside View of the Pandemic." An avid cellist, Dr Levy recently served as co-chair of the Chicago Symphony Orchestra Soundpost Series. In 2021, he started Tune It Up: A Concert To Raise Colorectal Cancer Awareness with the ACG.
Follow Medscape on Facebook, Twitter, Instagram, and YouTube
Lead image: iStock/Getty Images
Medscape Gastroenterology © 2023
Cite this: Benjamin H. Levy III. AI Revolution in Gastroenterology Is Closer Than We Think - Medscape - Jan 03, 2023.