Precision lung cancer screening from CT scans using a VGG16-based convolutional neural network.

Area of Science:

• Medical Imaging Analysis
• Artificial Intelligence in Healthcare
• Oncology Diagnostics

Background:

• Lung cancer remains a leading cause of cancer-related mortality worldwide.
• Early detection is crucial for improving patient outcomes and survival rates.
• Traditional screening methods have limitations in sensitivity and specificity.

Purpose of the Study:

• To develop and validate an advanced Convolutional Neural Network (CNN) model for precise lung cancer screening.
• To leverage medical big data and Computed Tomography (CT) scans for enhanced diagnostic accuracy.
• To establish a reliable tool for early lung tumor identification in clinical settings.

Main Methods:

• A VGG16-Based CNN model was developed using CT scans from a large patient cohort.
• Data augmentation techniques were employed to enhance model generalization and prevent overfitting.
• The model was trained and fine-tuned using SGD optimizer, learning rate scheduler, dropout layers, and early stopping.
• Performance was rigorously assessed using five-fold cross-validation and external validation datasets.

Main Results:

• The VGG16-Based CNN model achieved a high Area Under the Curve (AUC) of 0.963 ± 0.004.
• Classification accuracy reached 0.917 ± 0.004, with a Specificity of 0.962 ± 0.005.
• External validation confirmed the model's robustness and effectiveness across diverse patient data.

Conclusions:

• The developed VGG16-Based CNN lung screening model demonstrates significant reliability and effectiveness.
• The model accurately identifies lung tumors, supporting its application in clinical lung cancer screening.
• This research provides strong evidence for the use of AI-powered tools in early cancer detection.