Advances in AI-based genomic data analysis for cancer survival prediction

Cancer is one of the deadliest diseases prevalent in the world. Survivability, early diagnosis, and accurate prognosis are of utmost importance for the therapeutics and clinical management of cancer patients. To achieve accurate and timely prediction of the survival of cancer patients, several machine-learning models based on genomic data have been proposed but a comprehensive review of recent applications in cancer survival prediction is lacking.

This paper represents a review of the most recent application of machine learning and deep learning on cancer survival prediction, with a particular focus on the use of genomic data. It specifically targets the most prominent cancer types such as breast cancer, glioblastoma, lung cancer, renal cell cancer, and oral cancer.

A systematic review approach is employed to analyze recent studies on machine learning techniques applied to cancer survival prediction. Emphasis is placed on methodologies utilizing genomic data due to its effectiveness in predicting survival outcomes.

This review highlights the efficacy of different machine/deep learning-based techniques in predicting survival outcomes for different cancer types with genomic data. It also provides a summary of the contributions made by different research groups, critically examines the associated challenges, and suggests potential areas for further investigation.

Machine learning and deep learning techniques, especially those utilizing genomic data, hold significant promise for accurate cancer survival prediction across diverse cancer types. Despite advancements, challenges such as data heterogeneity and model interpretability still persist. Further research is warranted to address these challenges and develop a comprehensive framework for cancer survival prediction applicable to various cancer types. This review lays the foundation for future investigations in the area of cancer research.