the opinions of the investigators prior to the onset of the trial. It isthrough the prior that information from previous trials, clinical andpreclinical experience, and medical theory are incorporated into theanalysis. The prior distribution should be concentrated in some mean-ingful way around a prior guess ˆN0(provided by the clinicians), yet itshould also be suﬃciently diffuse as to allow for dose escalation in theabsence of dose limiting toxicity (Gasparini and Eisele, 2000). We notethat several authors (e.g., Tsutakawa, 1972, 1975) have be taken to be auniform distribution on a suitably defined interval (Babb et al., 1998) or anormal distribution with appropriately large variance (Gatsonis andGreenhouse, 1992).Example: 5-FU Trial (continued)The statistical goal of the trial was to determine the MTD of 5-FUwhen administered in conjunction with 20 mg/m2leucovorin and 0.5 mg/m2topotecan. The dose-toxicity model used to design the trial was thatgiven by Eq. (3), reparameterized in terms ofr= [gU0]. Preliminary
This is the end of the preview.
access the rest of the document.