Elsevier

European Journal of Cancer

Volume 46, Issue 16, November 2010, Pages 2870-2878
European Journal of Cancer

Current perspective
Dose-escalation models for combination phase I trials in oncology

https://doi.org/10.1016/j.ejca.2010.07.002Get rights and content

Abstract

Designing combination drug phase I trials has become increasingly complex, due to the increasing diversity in classes of agents, mechanisms of action, safety profiles and drug-administration schedules. With approximately 850 agents currently in development for cancer treatment, it is evident that combination development must be prioritised, as based on a specific hypothesis, as well as a projected development path for the involved combination.

In this manuscript the most relevant issues and pitfalls for combination drug phase I trial design are discussed. Several phase I study designs that incorporate controls to circumvent bias due to imbalances in observed background toxicity are discussed.

Introduction

Designs for phase I trials in general and combination drug phase I trials in particular are facing several challenges with the increasing diversity of classes of agents and mechanisms of action. In addition, these new classes of anticancer agents are often administered continuously rather than intermittently, and they manifest safety profiles that are completely different from those of conventional cytotoxic agents. Determining the best dose for new combination drug treatments will have to be balanced against these aspects. Longer observation periods than the usual 3–4 weeks are required, given that continuously administered drugs may show relevant toxicity only after a prolonged exposure.1 There are currently approximately 850 agents in the development for cancer treatment,2 which could potentially be combined into approximately 400,000 two-drug combinations, not to mention potential combinations involving already marketed agents. Computational modelling of new drug combinations may be a way forward, but as long as this has not been developed, we will have to work with realistic phase I study modelling and planning.

Section snippets

Rationale

A phase I study of a new combination of drugs is only the first clinical step in the development of that specific combination and should be considered a means, not an end in itself. Before commencing a combination drug phase I trial, a plan for (tumour-type specific) subsequent developmental phase II/III studies must clearly be defined. The choice for tumour type should be based on a scientific rationale, ideally including data in an appropriate preclinical model or on the basis of clinical

It all starts with an hypothesis

A single ideal template to perform a combination drug phase I trial will likely never exist. Every trial has to be designed based on prior knowledge of the preclinical and clinical pharmacology of the individual agents. Given the enormous number of potential combinations, in addition to a proper rationale and further development plan as outlined above, an appropriate hypothesis to guide the protocol design is also crucial. We have identified three potential general hypotheses for such studies.

Defining the MTD in Types 1 and 2 studies

In theory the number of MTDs of two drugs is infinite and does describe a curve, which we can call an envelope of tolerability. Obviously, the full envelope of tolerability of a drug combination will never be described, but defining multiple MTDs as derivatives of the envelope will generate knowledge on the dosing-range of the combination (Fig. 2).

At first glance, the MTD often seems to be determined by the data generated during the trial. However, some choices made with regard to the

MTD/DLT-incidence in perspective of known toxicity profiles

Combination phase I studies should preferably be initiated if knowledge is available on the single agent toxicity profile of the involved agents. Combination phase I trials should aim to model the toxicity as function of dose and PK. While defining the limits of ‘tolerability’ based on an incidence of unacceptable side-effects is relatively straightforward in single agent phase I trials, this turns out to be more difficult for combination phase I studies. And yet defining MTD and DLT is

Avoiding background noise on toxicity

One of the challenges in conducting combination drug phase I trials is to carefully weigh whether or not the frequency of observed toxicity is representative for the novel combination or that, due to chance, the toxicity attributable to just one of the drugs occurs in the trial in a higher than usual incidence.

Two instruments may be helpful by valuing the toxicity data generated more carefully: (1) the 3 + 3 + 3 design and (2) introducing control groups into phase I studies.

Determining feasibility in Type 3 studies

If no PK or PD interaction is anticipated, the feasibility of a combination can be proven in a kind of phase I/II trial. If we assume that drug A is already regarded as standard treatment for a certain tumour type, then patients with this tumour type can be enroled in a trial in which all patients start with a combination of drug A and drug B, both at full dose. Such an approach is feasible in, for example, the non-small lung cancer patient population in which the epidermal growth factor

Conclusion

Dose-finding studies are the first step in the clinical development of new combination treatment strategies. Combination phase I studies are extremely complex. Designing phase I studies is as important as the conduct itself and should be done by dedicated phase I researchers since a standard template cannot be made.

If there is a rationale to develop a specific drug combination, a hypothesis should be generated. This hypothesis, based on the anticipated levels of interaction, will be guiding for

Conflict of interest statement

None declared.

References (4)

There are more references available in the full text version of this article.

Cited by (41)

  • Phase I study of the checkpoint kinase 1 inhibitor GDC-0575 in combination with gemcitabine in patients with refractory solid tumors

    2018, Annals of Oncology
    Citation Excerpt :

    In the dose escalation stage (stage I), patients were enrolled in a 3 + 3 (arm 1) or 3 + 3 + 3 (arm 2) design to determine the maximum tolerated dose (MTD). A 3 + 3 + 3 design was used for the GDC-0575 and gemcitabine arm to limit the chance of falsely halting dose escalation due to gemcitabine toxicities [4]. Patients in arm 1 (monotherapy) received oral GDC-0575 15-90 mg (supplied by Genentech Inc., South San Francisco, CA) alone daily for three consecutive days starting on days 1, 8, and 15 of a 21-day cycle.

  • The challenges of implementing multiarmed early phase oncology clinical trials

    2018, Novel Designs of Early Phase Trials for Cancer Therapeutics
  • Clinical Trial Methodology in Lung Cancer

    2018, IASLC Thoracic Oncology
  • Designs of drug-combination phase I trials in oncology: A systematic review of the literature

    2015, Annals of Oncology
    Citation Excerpt :

    Therefore, several toxicity orderings between combinations are possible (two examples are given in Figure 1B and C). In practice, drug-combination phase I trials raise several challenging points to be defined before the trial onset [9–17]: (i) starting dose of each agent; (ii) choice of the dose range of each agent and the number of combinations to be evaluated; and (iii) total sample size that is strongly related to the number of possible combinations. In this study, we aimed at evaluating how drug-combination phase I trials in oncology have been designed in the last 3 years and what the principal investigator's choices were with regard to the dose range, number of combinations, and statistical design.

View all citing articles on Scopus
e

Present address: Department of Internal Medicine, Sint Franciscus Gasthuis, Rotterdam, The Netherlands.

View full text