Predicting Malignant Transformation in Oral Epithelial Dysplasia: A Systematic Comparison of Artificial Intelligence-Based Risk Models and Pathologist-Based Microscopy.
Diele Carine Barreto Arantes, Lucas Monteiro Delgado, Marília de Marco Pinto Paiva Dos Santos et al.
✦ AI-curated · Sources linked
In 30 seconds
This systematic review and meta-analysis compared artificial intelligence (AI)-based risk prediction models with traditional human microscopy for predicting malignant transformation in oral epithelial dysplasia. The analysis included 32 studies, revealing that AI models had a pooled odds ratio of 12.64 for predicting malignant transformation, indicating superior discriminative performance compared to human evaluations, which had an odds ratio of 3.42.
Key findings
- AI models demonstrated a pooled odds ratio of 12.64 for predicting malignant transformation.
- Human evaluations showed a significant association with an odds ratio of 3.42.
- AI-based models exhibited higher specificity and consistent performance compared to human assessments.
- Sensitivity of AI models was comparable to that of human pathologists.
Why it matters
The findings highlight the potential of AI-based models to enhance the accuracy of risk stratification in patients with oral potentially malignant disorders, which could lead to more personalized management strategies.