Predictive testing in cancer chemotherapy. I. In vivo.

Several in vivo methods have been assessed for their capacity to predict sensitivity for anticancer agents in humans. Standard strategies have been developed for screening purposes. Adjustments of these strategies are frequently suggested in reports in which the correlation between assay results and clinical therapeutic efficacy is analysed. Low predictivity and high costs of these assays are important reasons for changing the screening strategy. In vivo methods which predict the clinical response in the individual patient, are under investigation. Only the results of the subrenal capsule assay (in normal mice) have been correlated with the clinical response in a larger study. The criticism of the method and the low predictivity for sensitivity in a prospective study provide no reason for optimism. Methods which study changes predicting the clinical response in patients are still in a developmental phase.