Cheaper, Faster, Better: Chasing the Elusive Clinical Trial End Point

Cancer clinical trialists perennially seek to use end points in trials that are objective, easily measured and, if not the definitive end point of interest, are at least highly predictive of a definitive outcome. Re-searchers have been on a persistent quest to identify shorter-term, easier to measure, more frequent event-driven end points such as tumor response, biochemical failure, or othermarkers of disease pro-gression as surrogatemeasures for definitive outcomes such as overall survival (OS). Evidence for this long-term ongoing pursuit can be tracedback to the subject of an article that appeared in thefirst issueof Journal of Clinical Oncology (JCO) by Anderson et al1 The authors illustrated the bias-ridden and then more common reporting of tu-mor response–specific survival. Theypresented the landmarkmethod that corrects for biases by fixing the time point for the determination of tumor status and enumerating cumulative person-time and events from that point forward. An alternative was to use tumor status as