We have found the short-term MTT assay to be a simple, reproducible chemosensitivity technique, suitable for use throughout the time course of disease. We now have 12 years' experience of using this method in a variety of tumour types, both haematological and solid malignancies. Tumour cells are isolated from bone marrow, malignant effusions or solid biopsies and subjected to drug exposure for 48-96 h. Cell survival is measured by re-incubation in MTT for 4 h. We have found a significant correlation of in vitro results with in vivo outcome for acute myeloid leukaemia (AML) and for ovarian cancer (both p < 0.0001) with an assay sensitivity of 98% for AML and 81% for ovarian cancer. Furthermore, the 5-year survival of ovarian cancer patients treated with a drug found sensitive in vitro is significantly higher than that for patients treated with a drug found resistant in vitro (p = 0.033). We have correlated assay results with drug resistance markers. For example, expression of the newly described half transporter BCRP is related to daunorubicin resistance (p < 0.05). The MTT assay is also suitable for screening for modulation of drug resistance. We have found that the DNA polymerase inhibitor aphidicolin markedly increases in vitro sensitivity to the platinum drugs in ovarian cancer and cytosine arabinoside in AML in the majority of patients. The greatest effect was seen for patients deemed resistant in vitro to these agents. We have identified novel drug combinations which demonstrate significant synergism using this methodology and have also used it to study the emergence of drug resistance in cell line models with a view to its prevention. In conclusion, we have found the MTT assay to be a simple, repeatable, adaptable technique which produces accurate information to help the clinician select suitable treatment for individual cases.