Related Experiment Video
Updated: Sep 14, 2025
09:10
Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures
Published on: August 5, 2021
1.9K
Predicting the Recurrence of Sporadic Odontogenic Keratocyst Using Whole-Slide Histopathology Images With the Hybrid
Samahit Mohanty1, Divya Biligere Shivanna1, Roopa S Rao2
1Department of Computer Science and Engineering, Faculty of Engineering and Technology, M.S. Ramaiah University of Applied Sciences, Bangalore, India.
Clinical and Experimental Dental Research
|July 23, 2025
Summary
A new AI model, Hybrid Encoder Iterative Attention Convolution (HEIAC), accurately distinguishes recurrent from non-recurrent Odontogenic Keratocysts (OKCs). This advances diagnostic capabilities for challenging OKC cases.
Area of Science:
- Oral Pathology
- Artificial Intelligence in Medicine
- Computational Pathology
Background:
- Odontogenic keratocysts (OKCs) present significant clinical challenges due to their aggressive nature and high recurrence rates.
- Accurate prediction of OKC recurrence is crucial for effective patient management but remains difficult.
Purpose of the Study:
- To develop a reliable artificial intelligence (AI) model for distinguishing between recurrent and non-recurrent Odontogenic Keratocysts (OKCs) using whole-slide images (WSIs).
- To enhance the predictive accuracy of OKC recurrence through advanced image analysis.
Main Methods:
- A dataset of 84 OKC cases, including whole-slide images (WSIs), was utilized for model development and evaluation.
- The Hybrid Encoder Iterative Attention Convolution (HEIAC) model, integrating an encoder, attention mechanism, and convolutional layers, was employed for image classification.
- The HEIAC model was trained on 64 WSIs and validated on 20 WSIs (14 non-recurrent, 6 recurrent).
Main Results:
- The HEIAC model achieved a testing accuracy of 0.98.
- Exceptional performance metrics included 96% recall, 100% precision, a 97% F1-score, and an AUC of 1.0.
- The model demonstrated superior efficiency, using 96% fewer trainable parameters compared to standard vision transformers.
Conclusions:
- The developed HEIAC model significantly improves the ability to predict OKC recurrence from WSIs.
- This AI-driven approach offers a promising tool for clinicians and pathologists in managing OKC cases.
