Lung Cancer Classification from CT Scans using Deep Learning Techniques

Lung cancer remains the leading cause of cancer-related mortality globally, with late-stage diagnosis significantly reducing survival rates. This paper presents a robust deep learning framework for classifying lung CT scans into three clinically critical categories: benign, malignant, and normal cases, using the IQ-OTHNCCD dataset. Our convolutional neural network (CNNs) architecture incorporates batch normalization and strategic dropout layers to enhance feature learning while mitigating overfitting, a common challenge in medical image analysis. The model leverages extensive data augmentation to address dataset limitations and class imbalance, achieving an accuracy of 94.3% on the test set. Comparative analysis demonstrates that our approach outperforms existing methods in multi-class classification while maintaining computational efficiency. The proposed system shows strong potential as a decision-support tool for radiologists, particularly in resource-constrained settings where expert availability is limited.