The following medical brief espouses how ordinary clinical trial design is rendered less than perfect for targeted therapies, and suggests how the following 3 interrelated Agile Trial Design principles should be considered:
Our context: The dose for the most commonly used chemotherapies in veterinary medicine were most likely arrived at through trials using a 3 + 3 design. Put simply, this type of trial is a rules based trial that treats the first cohort of patients with a low dose of a drug. Dose is then escalated or de-escalated based upon the observation of dose limiting toxicities.
These phase I clinical trials are familiar to most oncologists and have certain advantages. They are simple, robust, and easily understood by clinicians and investigators. However, this historical approach of selecting the maximum tolerated dose may no longer be appropriate for the new more targeted chemotherapeutic agents being used today. These drugs often don’t need to be given at the maximum tolerated doses, but instead can be given at a biologically relevant dose.
A 2016 article in Clinical Cancer Research discussed the rationale behind advocating for a different method for determining drug dose in the current age of chemotherapy – as the 3+3 approach was designed for evaluating cytotoxic chemotherapy agents (1).
The authors go on to state that when this approach is used for targeted therapies, many non-optimal doses are taken into late stage development or to market. This results in frequent postmarket requirements to study lower doses or alternate regimens for these drugs. These postmarket trials are often quite challenging to complete in a timely manner.
The experience of bringing Palladia® (toceranib phosphate) to the veterinary market illustrates the problem of bringing a targeted therapy drug to market with the optimal dose. Many oncologists now use this drug at a dosage different from that on the label. They use a lower dose (2.4-2.75 mg/kg rather than 3.25 mg/kg) and a longer interval (MWF instead of every other day) based upon their experience with both efficacy and tolerability of the drug.
The authors then make a few recommendations about how to adjust the current trial design for targeted therapies and what methods may be appropriate to find doses for these new anticancer medications. One of their first recommendations is that investigators focus on determining an efficacious dose range rather than a single dose. This efficacious dose range can be based upon biomarkers and/or tumor size. This dose range may be incredibly wide, 5-fold or greater, encompassing both a minimally effective dose and a maximum tolerated dose (MTD).
The authors also recommend an adaptive or Bayesian approach to identify an appropriate dose of drug.
There is a very real tension between speeding up drug development and taking the time needed to determine the most appropriate dose (or dose range). To address this issue, the authors recommend integrating non-clinical, preclinical and clinical data to both increase the speed and accuracy of the trial. The use of clinical data is exactly what the Real World Evidence and Real World Data “movement” spearheaded by FDA is really all about (see our previous blog post).
Importantly, the authors also state that where there is significant unmet medical need, dose optimization could be carried out in the post-approval setting. A situation, I believe after 30 years of practice, exists in veterinary oncology.
Agile dose finding design principles depart from ordinary dose finding trials. These principles may be used to accelerate the finding of efficacious doses of targeted therapies. Certainly the clinical experience with Palladia and post-market research show that lower doses of Palladia are likely still effective while minimizing the adverse event profile (2).