Previous generations of tests, known as chemosensitivity assays, have measured cell death. There are two limitations to measuring cell death:
The wrong cells get measured. When cells are caused to proliferate (divide mother to daughter many times) they are not accurately copied. Both their genomes (genetic information) and phenomes (behaviors) change in unpredictable ways. Therefore, the response of the daughter cells to cancer drugs ex vivo may not match the response of the original tumor cells when the same drugs are administered in vivo.
Cancer cells that are kept alive artificially behave differently. The artificial environment that keeps them alive alters the way the cells respond to a cancer drug. Additionally, many cells will die naturally throughout the duration of the assay. If the cells are heterogeneous (multiple different cancers in the same sample), the different cancers may die at different rates and only the most robust portion of the sample will be tested, providing incomplete or ineffective results.