The in vivo hollow fibre model was developed by the National Cancer Institute (NCI) in the United States of America (USA) at a time when the number of potential anti-cancer drugs arising from in vitro screening efforts exceeded the available capacity for testing in traditional xenograft models. Updated analysis of the predictive value of the hollow fibre model continues to indicate that the greater the response in the hollow fibre assay, the more likely it is that activity will be seen in subsequent xenograft models. The original 12 cell line hollow fibre panel has been supplemented with histology-specific panels, and we begin here to analyse their utility in predicting activity in subsequent in vivo models. The key goal of using the hollow fibre model as a way to decrease the cost, both financial and in the number of animals used, to evaluate initial evidence of a compound's capacity to act across physiological barriers continues to be reinforced with our enlarging experience.