Analysis of the clinical utility of a predictive chemosensitivity assay

We present an elementary statistical analysis for quantifying the clinical utility of a predictive chemosensitivity assay based on retrospective data. An assay predictive of patient response could possibly improve response rates and reduce morbidity by individualizing chemotherapy for each patient. An analysis of retrospective data can suggest appropriate patient populations and sample sizes for study in prospective trials using the chemosensitivity assay. We provide an example with data from a clonogenic assay collected on melanoma patients.