Artificial intelligence is becoming a powerful tool within the medical community. From surgeries to diagnoses, professionals are turning to AI for treatment. Now doctors are taking advantage of the technology to help those with inflammatory bowel disease (IBD).
Researchers from Japan have developed an AI system to classify inflammatory bowel disease neoplasia accurately. It’s hard for endoscopists to identify the severity and grade of the neoplasia due to inflammation in the colorectal region. Because of this, it leaves biopsy as the only option, which is associated with high risks and often leads to inaccurate diagnoses.
To this end, researchers from Okayama University Graduate School of Medicine conducted a pilot study to develop this AI system.
“Our AI system prototype proved successful in determining the degree of malignancy of IBD-tumors and is valuable enough to contribute to clinical practice in the coming years,” says Hideaki Kinugasa, an assistant professor at the Okayama University Graduate School of Medicine, in a statement.
To develop the AI system’s prototype, researchers used a conventional neural network — a type of neural network used for the analysis of visual imagery. The AI system was trained using 862 endoscopic images of 99 IBDN lesions from IBD patients from two hospitals between 2003 and 2021, and validated it using a deep-learning framework. Researchers then asked highly-trained endoscopists to analyze the images and classify the lesions into two types based on the need for proctocolectomy, and compared their classification to that of the AI system.
The AI model generated around 6 million images from the original data set, and were used to analyze clinicopathological characteristics and the lesions.
After examining the data, researchers found that most patients had ulcerative colitis, with more than 95% of them presenting pancolitis and left-sided colitis. The AI system was fairly accurate with its diagnosis rate. The model displayed an image-based diagnostic ability with 64.5% sensitivity, 89.5% specificity, and 80.6% accuracy, and a lesion-based diagnostic ability with 74.4% sensitivity, 85% specificity, and 80.8% accuracy. The correct diagnosis of the AI model was 79%, compared to that of 77.8% from endoscopists.
“Using this AI system can ensure that endoscopists do not misdiagnose IBD neoplastic lesions, patients receive prompt treatment, and more appropriate treatment strategies are developed and applied for early as well as advanced stages of IBD,” explains Kinugasa.
There are around 1.6 million Americans who suffer from IBD and as many as 70,000 new cases are diagnosed each year.
This study is published in the Journal of Gastroenterology and Hepatology.