Review on the Application of Machine Learning and Deep Learning for the Diagnosis and Survival Prediction in Lung Cancer

Lung cancer continues to rank among the world’s leading causes of cancer-related death. Improvements in early detection, diagnosis, and treatment planning are imperative. Classification, feature extraction, and prognostic modelling of lung cancer have become much more accurate and efficient in recent years due to machine learning (ML) and deep learning (DL) techniques. This paper offers a thorough analysis of current approaches for detecting lung cancer utilising a variety of datasets, imaging techniques, and genomic data. We assess the efficacy of many ML and DL techniques, such as convolutional neural networks (CNN), deep belief networks (DBN), and hybrid models. Along with cutting-edge feature selection algorithms like Sparse Neural Networks and Deep Extreme Machine Learning, important preprocessing methods including Gabor filters, discrete Fourier transformations, and self-supervised learning strategies are covered. Additionally, the possibility of federated learning, generative models, and multi-omics integration to improve classification and prediction accuracy is investigated. The results draw attention to the advantages and disadvantages of various approaches, highlighting the necessity of sizable, superior datasets and strong validation techniques. This review may be a useful resource for researchers and clinicians aiming to leverage ML-driven advances in lung cancer research.