8. Late-Onset Toxicity

  1. Guosheng Yin

Published Online: 9 JAN 2012

DOI: 10.1002/9781118183335.ch8

Clinical Trial Design: Bayesian and Frequentist Adaptive Methods

Clinical Trial Design: Bayesian and Frequentist Adaptive Methods

How to Cite

Yin, G. (2011) Late-Onset Toxicity, in Clinical Trial Design: Bayesian and Frequentist Adaptive Methods, John Wiley & Sons, Inc., Hoboken, NJ, USA. doi: 10.1002/9781118183335.ch8

Publication History

  1. Published Online: 9 JAN 2012
  2. Published Print: 19 DEC 2011

ISBN Information

Print ISBN: 9780470581711

Online ISBN: 9781118183335



  • 3 + 3 design;
  • dose-finding method;
  • dose-limiting toxicities (DLTs);
  • hypothesis testing;
  • late-onset toxicity;
  • likelihood function;
  • maximum tolerated dose (MTD);
  • statistical methodologies;
  • toxicity outcomes


In a typical phase I oncology trial, the primary objective is to identify the maximum tolerated dose (MTD) of an experimental drug. Most of the dose-finding method requires dose-limiting toxicities (DLTs) to be ascertainable shortly after treatment, so that the toxicity outcomes of all the previously treated patients are available by the time for the next dose assignment. This chapter provides deeper insights into the issues of late-onset toxicity and introduces each dose-finding method that is capable of accommodating delayed outcomes. The statistical methodologies discussed in the chapter can also be used to model other endpoints that are not immediately observable, such as the time-lagged efficacy event or delayed response. The 3 + 3 design requires that toxicity be ascertainable shortly after treatment. Using patients' exposure times as weights, the time-to-event continual reassessment method (TITE-CRM) addresses the issues associated with late-onset toxicity based on a pseudo-binomial likelihood function.

Controlled Vocabulary Terms

hypothesis testing; likelihood