An Efficient Combination of Convolutional Neural Network and LightGBM Algorithm for Lung Cancer Histopathology...

Area of Science:

• Oncology
• Bioinformatics
• Medical Imaging

Background:

• Lung cancer is a leading cause of mortality.
• Histopathological image analysis via biopsy is crucial for accurate diagnosis.
• Deep learning shows promise in medical image analysis.

Purpose of the Study:

• To develop an efficient method for identifying and classifying lung tissue histopathology images.
• To combine a novel Convolutional Neural Networks (CNN) model with an enhanced Light Gradient Boosting Model (LightGBM) classifier.

Main Methods:

• Image pre-processing followed by feature extraction using a proposed CNN model with minimal parameters.
• Classification of lung tissues using a multi-threaded LightGBM model.
• Evaluation on the LC25000 dataset.

Main Results:

• Achieved 99.6% accuracy and sensitivity in lung tissue classification.
• The proposed CNN model utilized only one million parameters.
• Feature extraction completed in one second, demonstrating rapid processing.

Conclusions:

• The novel hybrid deep learning approach significantly enhances lung cancer diagnosis accuracy.
• The method offers a faster and more effective alternative to existing state-of-the-art techniques.
• This technique shows potential for improving clinical workflows in lung cancer detection.