Related Experiment Video
Updated: Jan 16, 2026
07:15
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
7.4K
Automated classification and explainable AI analysis of lung cancer stages using EfficientNet and gradient-weighted
Abdulmajeed Alqhatni1, T K S Rathish Babu2, T R Mahesh3
1Department of Information Systems, College of Computer Science and Information Systems, Najran University, Najran, Saudi Arabia.
Frontiers in Medicine
|September 26, 2025
Summary
A new deep learning model using EfficientNet-B0 achieves 99% accuracy in classifying lung cancer from CT scans. Explainable AI with Grad-CAM enhances transparency and reliability in automated cancer diagnosis.
Area of Science:
- Oncology
- Medical Imaging Analysis
- Artificial Intelligence in Healthcare
Background:
- Accurate lung cancer staging from CT scans is crucial for patient prognosis and treatment planning.
- Traditional manual interpretation of CT scans can be subjective and inconsistent.
- Automated methods are needed to improve the accuracy and reliability of lung cancer classification.
Purpose of the Study:
- To develop and evaluate an automated deep learning model for classifying lung cancer stages from CT images.
- To enhance the interpretability and transparency of the deep learning model using Explainable AI (XAI) techniques.
- To improve diagnostic accuracy and reliability in lung cancer detection.
Main Methods:
- An EfficientNet-B0 based deep learning architecture was employed for image classification.
- Gradient-weighted Class Activation Mapping (Grad-CAM) was utilized to provide visual explanations for model predictions.
- The model was trained and validated using 1,190 CT scans from the IQ-OTH/NCCD dataset, categorized as benign, malignant, or normal.
Main Results:
- The proposed model achieved high performance metrics: 99% accuracy, 99% precision.
- Recall rates were 96% for benign, 99% for malignant, and 100% for normal cases.
- Grad-CAM successfully identified critical regions in CT scans influencing classification, enhancing model interpretability.
Conclusions:
- The EfficientNet-B0 based deep learning model demonstrates exceptional performance in classifying lung cancer from CT scans.
- The integration of Grad-CAM significantly improves the interpretability and trustworthiness of the automated diagnostic system.
- This approach offers a reliable and transparent tool for medical image analysis in oncology, potentially aiding clinical decision-making.
