Lung cancer detection based on computed tomography image using convolutional neural networks

Area of Science:

• Medical imaging analysis
• Artificial intelligence in oncology

Background:

• Lung cancer is the leading global cancer, often diagnosed late due to non-specific symptoms.
• Early detection of lung cancer is critical for effective treatment and improved patient outcomes.

Purpose of the Study:

• To classify benign versus malignant lung lesions using deep learning and CNNs.
• To develop a decision support system for lung cancer diagnosis.

Main Methods:

• Utilized a dataset of 4459 CT scans (2242 benign, 2217 malignant).
• Employed a GoogLeNet-based deep learning architecture for image analysis.
• Implemented a retrospective case-control study design.

Main Results:

• The CNN model achieved a high accuracy of 0.98 during training.
• The model demonstrated a positive predictive value of 0.984 in the testing phase.
• Sensitivity and specificity metrics also indicated strong classification performance.

Conclusions:

• Deep learning techniques significantly enhance the accuracy of lung cancer diagnosis from CT images.
• CNN-based approaches offer a promising tool for classifying lung lesions, supporting clinical decision-making.