APPLICATION OF DEEP LEARNING IN CANCER PROGNOSIS: PREDICTING TUMOR PROGRESSION, RECURRENCE, AND PATIENT OUTCOMES USING MULTI-OMICS DATA

Abstract

From multi-omics data, deep learning determined the prediction of tumor development, recurrence, and patient outcomes and, thus, revolutionized cancer prognosis. With next-generation sequencing and advanced imaging technologies, multi-omics data now provide a unique opportunity to increase predictive accuracy. This review evaluates the impact of deep-learning models on cancer prognosis and their application with several architectures, including convolutional neural networks (CNNs) and recurrent neural networks (RNNs), in genomics, transcriptomics, and clinical data. It ends with an overview of the developments, challenges, and prospects in the AI-driven context for precision oncology.