AI Model Outperforms Radiology Experts in Diagnosing Crohn’s Disease in Children
Research By: Richard X. Liu, MHS | Jonathan Dillman, MD, MSc
Post Date: August 2, 2023 | Publish Date: Aug. 2, 2023
Radiology | Top Scientific Achievement
Crohn’s disease (CD) affects around 750,000 people in the United States, with cases rising worldwide. This chronic inflammatory disease commonly starts between ages 10 and 25, with about 20% of patients hospitalized annually and 50% requiring surgery within 10 years of diagnosis.
In children, diagnosing CD early is essential for reducing the risks of malnutrition and growth issues. However, diagnosing CD can be challenging. Even expert radiologists can be inconsistent at interpreting MRI scans.
A research team led by first author Richard X. Liu, MHS, a medical student in the University of Cincinnati College of Medicine, and senior author Jonathan Dillman, MD, MSc, hypothesized that an artificial intelligence (AI) model could offer high diagnostic accuracy while eliminating the variability associated with human radiologists.
The study compared the AI model’s performance against three trained radiologists to interpret MRI scans of 135 participants, 70 of whom were diagnosed with ileal CD. While individual radiologists’ accuracy ranged from 83.7% to 88.1%, the AI model achieved 93.5% accuracy.
“These findings suggest that an AI-assisted diagnosis could help standardize readings, reduce errors, and offer a faster path to treatment for children with suspected Crohn’s disease,” Dillman says.
The study also further establishes that ileal-wall radiomic features were strongly predictive of CD. Currently, a single reference standard tool for MRI-based Crohn’s diagnosis is lacking. Meanwhile, existing standards for clinical diagnosis are complex and can include a combination of cross-sectional imaging, ileocolonoscopy with biopsy, histologic testing, laboratory evaluation and clinical assessment.
Cincinnati Children’s co-authors in this study included Hailong Li, PhD, Alexander Towbin, MD, Nadeen Abu Ata, MD, Ethan Smith, MD, Jean Tkach, PhD, and Lili He, PhD, all with the Department of Radiology; and Lee Denson, MD, Division of Gastroenterology, Hepatology and Nutrition.
Explore the 2024 Research Annual Report
| Original title: | Machine Learning Diagnosis of Small-Bowel Crohn Disease Using T2-Weighted MRI Radiomic and Clinical Data |
| Published in: | American Journal of Roentgenology |
| Publish date: | Aug. 2, 2023 |
Research By
Latest Posts
About this blog
The Research Horizons blog features news and insights about the latest discoveries and innovations developed by the scientists of Cincinnati Children's. This blog does not provide medical advice, diagnosis, or treatment.
Subscribe to Our Newsletter