Artificial Intelligence Aids Colorectal Polyp Classification Trial

After creating an artificial intelligence tool using retrospective data, Dartmouth-Hitchcock Medical Center and Cheshire Medical Center are implementing the technology in a clinical trial.

The tool was developed by Dartmouth-Hitchcock Norris Cotton Cancer Center researchers and uses artificial intelligence to classify the four major types of colorectal polyps removed during screening colonoscopy. The model produced accurate results comparable to pathologists and withstood evaluation using broad datasets from multiple institutions across the United States.

According to researchers, this proves that the artificial intelligence model is generalizable and can be trained on widespread external data.

The clinical trial was designed to study the performance of the deep learning model as part of an AI-augmented digital system compared to a standard microscope.

“Evaluating this tool through a prospective clinical trial shows that the AI-augmented digital system significantly improves the accuracy of pathologists in the classification of polyps in comparison to the traditional process of using microscopes,” leader of the clinical research team, Saeed Hassanpour, PhD, said in a press release.

Before using the AI-augmented digital system, pathologists watched a five-minute training video, read a brief summary regarding how the model works and how results are generated, and practiced using a set of ten sample slides to become familiar with the new system.

During the clinical trial, the average time of evaluation across all pathologists using the digital system decreased consistently. In compassion, the reading time did not change significantly when using the microscope, a tool that pathologists have relied on for many years.

According to researchers, the average System Usability Scale Score for the digital system indicated positive usability. In addition, pathologists stated that the digital system was easy to operate and ran smoothly.

Additionally, half of the participating pathologists said they would use a version of the artificial intelligence tool in clinical practice. Twelve out of 15 added that their experience either positively changed or supported their positive opinions for the use of artificial intelligence in clinical practice.

Hassanpour and the research team are now working with a digital pathology startup to bring their technology to clinical practice, improving patient outcomes and cancer prevention.

“This AI-augmented digital system shows promise in improving the frequency of surveillance recommendations to prevent cancer development, cutting colorectal cancer surveillance costs, eliminating undue stress to patients, increasing coverage and accuracy of surveillance programs, and ultimately reducing overall colorectal cancer mortality,” the press release concluded.