Colonoscopy can Miss Out Polyps but Artificial Intelligence Will Not

The introduction of artificial intelligence (AI) into healthcare has been revolutionary, transforming how we diagnose, treat, and monitor patients. This technology is significantly improving results by allowing for more accurate diagnoses and more targeted therapies. Mayo Clinic researchers are working on ways to use artificial intelligence to detect possibly malignant polyps during colonoscopy. According to some studies, the technique, which is still the standard of care for screening for and preventing colorectal cancer, may miss lesions that subsequently contribute to more than half of post-colonoscopy cancer cases.

Success of Colonoscopy

Colonoscopy's success is based on its capacity to detect and remove pre-cancerous polyps, which aids in the prevention of colorectal cancer. Polyps, on the other hand, are often the most difficult to detect among the people most in need of screening—those with inflammatory bowel diseases (IBD) such as Crohn's disease or ulcerative colitis, who are at the highest risk of getting colorectal cancer (1^✔ ✔Trusted Source
The role of artificial intelligence in colon polyps detection

Go to source).

Pre-cancerous lesions in this patient population are flat or only slightly elevated, making them easy to miss throughout the operation.

The Mayo Clinic team, on the other hand, is building an AI tool trained to recognize these difficult-to-see traits and to work alongside physicians in real-time to alert them oftheir presence by drawing red boxes around polyps that could have been missed.

"We’re all familiar with facial recognition software," said James East, M.D., a gastroenterologist at Mayo Clinic Healthcare in London. "Instead of training AI to recognize faces, we train it to recognize polyps."

Mayo Clinic used its enormous database of surveillance colonoscopies performed in the past to assist train the AI model. According to Nayantara Coelho-Prabhu, M.B.B.S., a gastroenterologist at the Mayo Clinic in Rochester, Minnesota, the institution is uniquely prepared to build this type of AI because it performs between 800 and 900 colonoscopies every year.

She emphasized that the Mayo database contains ‘ground truth’—real-world observations and measurements that may be utilized to train and test AI algorithms.

Advertisement

Coelho-Prabhu and his colleagues have chosen a group of 1,000 patients and are currently methodically annotating data from them by watching each colonoscopy video to label lesions in every frame from every angle, which will be used to train the algorithm to distinguish IBD-specific polyps.

Furthermore, Coelho-Prabhu and her colleague gastroenterologist Cadman Leggett, M.D., are working on a new digital endoscopic platform that will allow for the filming of all in-house procedures, correlation with patient medical information, and integration of AI back into normal procedures as needed. "Once we develop algorithms, we can run them in our procedure videos to test their performance," Coelho-Prabhu said.

Artificial Intelligence System

While the creation of an AI system to aid in colorectal cancer detection and treatment is the most advanced initiative at the Mayo Clinic, it is not the only one trying to use its automated learning skills to aid in illness detection and treatment.

Advertisement

The academic medical institution also has an AI-focused study that aims to employ natural-language processing (NLP) to scan medical notes in patients' electronic health records to identify risk factors for pancreatic cancer. Shounak Majumder, M.D., the program's director, expects that the NLP screening to identify individuals most at risk might be used as a tool to identify patients who should have extra screening.

Reference:

1. The role of artificial intelligence in colon polyps detection - (https://pubmed.ncbi.nlm.nih.gov/32821348/)

Source-Medindia