Multilevel Deep Learning Technique Based Classification of Lung Cancer as Per Tumor Node Metastasis Coding

Aim/Background: Accurate staging of lung cancer is critical for treatment planning and prognosis. Traditional methods typically rely on binary classification and manual analysis of computed tomography (CT) scans, limiting diagnostic precision and scalability in clinical settings. Objectives: This study introduces a deep learning-based system to automatically classify lung CT scans into three tumor– node–metastasis (TNM) categories. Materials and Methods: The proposed method starts with automatic segmentation of various lung abnormalities, including internal and juxtapleural nodules, using an enhanced conditional generative adversarial network (c-GAN). Subsequently, a hybrid approach combining support vector machines (SVM) and deep neural networks is used to assign TNM stages. The model leverages a multilevel architecture to extract both local and global image features, integrates patient-independent data separation, and incorporates explainable artificial intelligence (AI) components. Results: The automated system eliminates the need for manual region of interest (ROI) selection and delivers staging performance that rivals current high-end TNM classification tools. Compared to conventional binary classifiers, the proposed approach provides a comprehensive diagnostic model with improved accuracy and clinical relevance. Conclusion: This work presents a robust and scalable AI-driven framework for lung cancer staging that aligns with established clinical protocols. By minimizing manual intervention and enhancing model interpretability, the approach supports more efficient and accurate analysis of CT scans in real-world healthcare environments.