Measurement of drug activity as an oncolytic effect, use of the control only to monitor the quality of tissue implanted, and the rapid clearance of necrotic tissue from the subcapsular site, as significant factors incorporated into the design of the assay, have permitted use of a simple tumor size parameter for evaluating drug activity. The simplicity and economy of such a parameter, the predictability and reproducibility of the 6-day assay observed thus far, and evidence that the assay does measure a biological property of the tumor apart from host response, have warranted the continued use of the 6-day time frame and the normal immunocompetent CDF1 mouse as xenograft host. These studies have demonstrated the feasibility of using human tumor explants obtained from a variety of solid human malignancies in a straightforward, short term, in vivo predictive assay system. Preliminary correlations between in vivo (assay) tumor sensitivity and clinical response have given reasonable concurrence. This crucial point will require further study, with larger numbers of patients, under more rigid conditions. Final validation of this, and other, predictive assays will require a prospective, randomized study in large numbers of patients. Our present prospective study is being continued, therefore, with expansion to a multi-institutional design over a broader geographic area.