Current Advances and Limitations of Deep Learning in Anticancer Drug Sensitivity Prediction.

Anticancer drug screening can accelerate drug discovery to save the lives of cancer patients, but cancer heterogeneity makes it challenging. Prediction of anticancer drug sensitivity is useful for anticancer drug development and biomarker discovery. Deep learning, as a branch of machine learning, is an important part of in silico studies. Its outstanding computational performance means that deep learning has been applied to solving many biomedical problems, such as medical image recognition, biological sequence analysis, and drug discovery. There have been some studies of anticancer drug sensitivity prediction based on deep learning algorithms. Deep learning has made some progress in model performance and multi-omics data fusion. However, deep learning is limited by the number of studies performed and data sources available so it is not perfect as a pre-clinical model for screening anticancer drugs. How to improve the performance of deep learning models is a pressing problem for researchers. In this review, we introduce the research history of anticancer drug sensitivity prediction and the applications of deep learning in anticancer drug prediction. To provide reference for future research, we also review some common data sources and previous machine learning methods. Lastly, we discuss the advantages and disadvantages of deep learning, as well as the limitations and future perspectives of this approach.