A Comprehensive Survey of Computational Techniques for Lung Cancer Diagnosis and Prediction

Background and Objective: Lung cancer continues to be a major global health issue, with a pressing need for improved diagnostic and prognostic methods to enhance early detection and patient outcomes. The objective of this survey is to review and evaluate the current methods and models used in lung cancer diagnosis and prognosis, focusing on their strengths, limitations, and potential for future advancements. Methods: A systematic review of the literature was conducted across key databases, focusing on studies that utilize deep learning architectures, such as CNN, GoogleNet, VGG-16, U-Net, and machine learning algorithms, including XGBoost, SVM, KNN, ANN, and Random Forest. The review synthesized findings from these studies to assess the effectiveness and limitations of these computational models in the context of lung cancer detection. Results: The review identified several strengths in current models, including high accuracy in controlled environments and potential for early detection. However, significant limitations were also highlighted, such as issues with model interpretability, a lack of real-world validation, and challenges in integrating diverse diagnostic techniques. These gaps indicate the need for further research to enhance the applicability and reliability of AI-driven models in clinical settings. Conclusions: Advanced computational methods, particularly those utilizing deep learning and machine learning, hold transformative potential for lung cancer diagnosis and prognosis. However, to fully realize this potential, future research must address current challenges, such as improving model interpretability and ensuring robust validation in real-world scenarios. By overcoming these obstacles, AI-driven approaches can significantly improve patient care and outcomes in lung cancer treatment.