Date of Award


Document Type

Open Access Thesis

Degree Name

Medical Doctor (MD)

First Advisor

Cary Gross


Considerable ethical debate surrounding the risks and benefits of Phase I oncology trials is based on older response and toxicity data that does not account for recent changes in the types of agents and trial design. This study aims to not only update these data, but to investigate the impact of novel trial designs on various clinical outcomes. We performed a review of the literature using the Medline database. Part I included nearly all phase I trials published in 2002. Part II identified phase I studies of cytotoxic agents alone, published from 2002 through 2004. 221 Phase I oncology studies, consisting of 6,008 patients, were studied in Part I, while 149 studies, comprising 4,532 patients, were analyzed in Part II. Overall, the response rate for Phase I oncology trials in 2002 was 19%, the mortality rate was 1.1%, and the rates of severe hematologic and non-hematologic toxicities were 19% and 22%, respectively. "Classic" phase I trials of single agent cytotoxic drugs accounted for only 18% of trials, while more than half (55%) included at least one FDA approved therapy. The response and toxicity rates varied with the class of agent (e.g. cytotoxic, biologic, vaccine), and the combinations of agents (e.g. approved, investigational) studied. Only 34% of studies utilized aggressive dose escalation schemes, 22% permitted intra-patient dose escalation, and only 28% enrolled fewer than 3 patients to any dose level before proceeding to the next higher dose level. Studies that allowed intra-patient dose escalation or used fewer than three patients per dose were not associated with rates of response or toxicity that differed from trials using a more "traditional" design, nor did they increase the percentage of patients who received the recommended phase II dose. However, aggressive dose escalations were associated with increased rates of both hematologic (17% vs.10%) and non-hematologic (17% vs. 13%) toxicity for participating patients without increasing response rates. None of these novel design strategies were associated with a smaller patient requirement. Phase I oncology trials represent a spectrum of different classes of agents and design strategies that are often associated with distinct clinical outcomes. Accounting for this variety is critical when evaluating their risk-benefit profiles and ethics. While some innovations in trial design do not appear to be any more helpful or harmful than standard methods in phase I trials of single agent cytotoxic drugs, using aggressive dose escalations may, in fact, be more hazardous for patients. These findings highlight the need for continued effort towards improving trial design and its impact on our patients.


This is an Open Access Thesis.

Open Access

This Article is Open Access