Cancer cell line panels have proved useful disease models to, among others, identify genomic markers of drug sensitivity and to develop new anticancer drugs. The increasing availability of in vitro sensitivity and cell line profiling data sets raises the question of whether this information could be used, and to which extent, to predict the activity of drugs in cancer cell lines and, ultimately, in patients tumors. Drug sensitivity prediction embraces those approaches aiming at predicting in vitro drug activity on cancer cell lines by integrating genomic and/or chemical information using machine learning models. In this review, we summarize the cytotoxicity assays generally used to determine in vitro activity on cultured cell lines, and revisit the drug sensitivity prediction studies that have leveraged chemical and cell line profiling data from the NCI60, Cancer Cell Line Encyclopedia (CCLE) and Genomics of Drug Sensitivity in Cancer (GDSC) projects. A section outlining current limitations and future perspectives in the field closes the review.