Taxanes are used for the treatment of many human cancers, as first- and second-line chemotherapeutics. In the course of treatment many patients develop resistance or hypersensitivity to one form of taxane and require a different taxane to rescue the therapeutic benefit of the drug. There is currently no method to reliably predict tumor responses to taxanes prior to therapy or when resistance or hypersensitivity develops. We adapted the quartz crystal microbalance (QCM) biosensor technique to study responses of human mammary epithelial tumor cells to taxanes. Studies indicate that stable frequency and resistance levels are reached at 24 h. Cells in the QCM can then be treated with taxanes and responses monitored in real time via frequency and resistance changes reflecting alterations of cell mass distribution and viscoelastic properties. Distinct shifts in frequency and resistance accurately predicted apoptosis or resistance to treatment, as determined in parallel convention assays. QCM analysis accurately predicted docetaxel was more effective than paclitaxel and MCF-7 cells were more resistant to taxanes compared to MDA-MB-231 cells. These studies suggest "signature" patterns for taxane responsivity could be compared to those of patient biopsy samples to predict therapy outcome prior to treatment for initial therapy or to rescue therapy efficacy.