Enhancing Lung Cancer Diagnosis: An Optimization-Driven Deep Learning Approach with CT Imaging

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Lung cancer (LC) is a major global health concern, necessitating accurate and timely diagnosis.
• Computed Tomography (CT) is crucial for LC detection, but manual analysis is labor-intensive and prone to errors.
• Existing deep learning models struggle with feature extraction and computational complexity in high-dimensional CT data.

Purpose of the Study:

• To develop an optimized deep learning model for enhanced lung cancer classification from CT images.
• To improve feature extraction efficiency and reduce computational complexity in lung cancer detection.
• To investigate the impact of various optimization algorithms on model performance.

Main Methods:

• An optimized CBAM-EfficientNet model was developed, integrating EfficientNet for reduced complexity and CBAM for feature emphasis.
• Gray Wolf Optimization (GWO), Whale Optimization (WO), and Bat Algorithm (BA) were employed for hyperparameter tuning.
• The model was evaluated on the Lung-PET-CT-Dx and LIDC-IDRI benchmark datasets.

Main Results:

• The GWO-based CBAM-EfficientNet achieved high accuracies: 99.81% on Lung-PET-CT-Dx and 99.25% on LIDC-IDRI.
• The BA-based CBAM-EfficientNet demonstrated strong performance with 99.44% and 98.75% accuracy on the respective datasets.
• The proposed model significantly outperformed existing methods in lung cancer classification.

Conclusions:

• The optimized CBAM-EfficientNet model offers a highly accurate and efficient solution for automated lung cancer diagnosis.
• The integration of optimization algorithms enhances predictive accuracy and model robustness.
• The lightweight architecture facilitates real-time clinical application, aiding radiologists in diagnosis.