Early Detection of Lung Cancer Using Predictive Modeling Incorporating CTGAN Features and Tree-Based Learning

Artificial intelligence (AI) has revolutionized several domains and medical science has significantly benefited from AI-based approaches. Machine learning models enable the robust and accurate analysis of large streams of medical data, assisting medical experts in creating specialized and targeted treatment plans. Lung cancer also requires further automated detection tools as it remains one of the most lethal cancers worldwide, necessitating advancements in early and accurate detection methods. This study presents a novel approach to lung cancer classification by leveraging synthetic data generated using conditional tabular generative adversarial networks (CTGAN) and applying a random forest (RF) classifier. The proposed model has shown exceptional performance with an accuracy of 98.93%, and precision, F1 score, and recall of 99%. To validate the efficacy of the approach, we conducted extensive experiments using nine other classifiers, including support vector machines, k-nearest neighbors, decision trees, and others. These classifiers were evaluated with different data balancing techniques such as synthetic minority oversampling technique (SMOTE), borderline-SMOTE, and SMOTE-ENN, alongside original, unbalanced dataset. Comparative analysis revealed that the CTGAN-RF model outperforms these traditional classifiers, particularly in handling class imbalance and improving predictive accuracy. Additionally, the robustness of the proposed model was confirmed through 5-fold cross-validation, further emphasizing its reliability. The approach was benchmarked against state-of-the-art (SOTA) methods in lung cancer detection, highlighting significant improvements in classification metrics. This comprehensive evaluation highlights the potential of synthetic data augmentation combined with machine learning techniques in enhancing the early detection and diagnosis of lung cancer, opening the way for improved patient outcomes and personalized treatment strategies.