Prediction of individual patient response to chemotherapy by the fluorometric microculture cytotoxicity assay (FMCA) using drug specific cut-off limits and a Bayesian model.

The semiautomated fluorometric microculture cytotoxicity assay (FMCA) based on the measurement of fluorescence generated from cellular hydrolysis of fluorescein diacetate (FDA) to fluorescein in microtiter plates, has been used for determination of cytotoxic drug resistance of tumor cells from patients with hematological and solid tumors. In the present study we describe a calibration procedure based on statistically derived cut-off limits and assay-predicted response probabilities using Bayes' theorem. Test results at a specified drug concentration were divided into three categories: low, intermediate or extreme drug resistance (LDR, IDR and EDR, respectively) using the median and median +1 standard deviation as the cut-off limits. When correlated with clinical outcome, LDR samples showed a higher response rate than expected, IDR a lower and EDR samples no response at all. The sensitivity and specificity of the test, using the median as cut-off limit, were 0.92 and 0.69 respectively. By fitting these test characteristics to a statistical model based on Bayes' theorem it is possible to calculate response probabilities for each individual patient taking into consideration not only the test characteristics and the particular assay result, but also the clinical and patient specific characteristics influencing the pre-test probability of response. EDR predicts clinical drug resistance with high specificity and is also observed in tumor types with high response rate.