Related Experiment Video
Updated: Jun 10, 2025
05:56
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
Published on: April 14, 2023
2.4K
DETECTION OF ORAL SQUAMOUS CELL CARCINOMA USING PRE-TRAINED DEEP LEARNING MODELS
K Dhanya1, D Venkata Vara Prasad1, Y Venkataramana Lokeswari1
1Department of Computer Science Engineering, Sri Sivasubramaniya Nadar College of Engineering, Rajiv Gandhi Salai, Kalavakkam, India.
Experimental Oncology
|October 13, 2024
Summary
This study explored using transfer learning and convolutional neural networks (CNNs) to classify oral squamous cell carcinoma (OSCC) from histopathological images. The developed models show promise for early cancer diagnosis, improving patient outcomes.
Area of Science:
- Oncology
- Medical Imaging
- Artificial Intelligence
Background:
- Oral squamous cell carcinoma (OSCC) is a significant global health concern, with early detection crucial for improved patient prognosis.
- Traditional tissue biopsy for OSCC diagnosis is costly and time-consuming.
- Advancements in artificial intelligence, particularly transfer learning, offer potential for more efficient diagnostic tools.
Purpose of the Study:
- To evaluate the efficacy of transfer learning using pre-trained models and CNNs for binary classification of OSCC from histopathological images.
- To assess the potential of these AI methodologies as a tool for early cancer detection.
Main Methods:
- Utilized a dataset of 5192 histopathological images.
- Employed pre-trained deep learning models (ResNet-50, VGG16, InceptionV3) for feature extraction.
- Developed and tuned a convolutional neural network (CNN) for classification.
Main Results:
- The best-performing model achieved an accuracy of 0.90, sensitivity of 0.97, and AUC of 0.94.
- Evaluated performance against state-of-the-art methods, highlighting high sensitivity.
- Results visualized using ROC curves and confusion matrices.
Conclusions:
- Transfer learning methodologies demonstrate practical viability for medical applications.
- The findings support the use of AI in improving diagnostic precision for early cancer detection.
- Suggests future research directions for refining these AI tools for clinical use.
