Finding the next oncology breakthrough starts with a successful phase I clinical trial. While these studies do look at safety or toxicity, they also have another goal – to establish a recommended dose and avoid subtherapeutic participation.

However, not all phase I trials accomplish this in the same way.

Developing Dosing Schedules
Developing appropriate dosing schedules may involve calculating the Maximum Tolerated Dose (MTD) or Recommended Phase 2 Dose (RP2D) of the proposed therapeutic or investigational product. The methods for doing so generally fall into one of two categories: rule-based or model-based.

  • Rule-based Designs: Rule-based dose escalation (also called algorithm-based) includes traditional designs like 3+3. These approaches have very specific rules to guide dose escalation, and they make no assumptions regarding toxicity. Rule-based dosing has been popular for decades because it is very simple. The potential downfall is that this approach might not capture the true MTD, and it is not aggressive. There is a risk that participants could receive subtherapeutic doses.
  • Model-based Designs: Model-based designs are more flexible and often more effective. This approach exposes more participants to the MTD, so there is a somewhat greater risk, but it also tends to be more accurate than rule-based designs.

Traditionally, these two approaches were treated dichotomously, but there is a third opportunity. As medicine in general (and the field of oncology specifically) has evolved to be more personalized, dose-escalation methods are also shifting. Toxicity as an endpoint doesn’t carry the same weight in today’s precision medicine, so there is a greater need for researchers to look past traditional designs.

Port Worthy
American Lung Association
Novasep PharmaZell Group

Looking Beyond Traditional Designs
Novel dose-escalation approaches combine rule-based and model-based designs to create trials that are both flexible and efficient. One such design is the Bayesian Optimal INterval (BOIN) Design. It is model-assisted as opposed to model-based, and the BOIN platform can be used with a single drug or a combination therapy. It also has some of the simplicity of rule-based designs, but the real standout feature is that Bayesian methods include toxicity as well as efficacy.

Advertisement #3

Bayesian Optimal Interval Design
The way BOIN calculates dose liming toxicity (DLT) is unique in that it incorporates the observed DLT rate for all patients at a particular dose as compared to looking at the current cohort of patients separately. This approach is more transparent, but it is also more accessible for non-statisticians. Another advantage is that BOIN establishes a maximum number of patients, which, in turn, informs recruitment timelines and program design overall.

To better understand BOIN in practice, here is a flowchart of the design as used to determine DLT.

BOIN does require several assumptions, such as the target DLT and maximum sample size, but the approach has a variety of benefits. Dose escalation determined by BOIN tends to have a reduced risk of overdosing as compared to other designs. It also has a higher probability of determining the true MTD, and it tends to reduce the risk of sub-therapeutic dosing.

Using BOIN in Oncology Trials
Using a BOIN design in phase I clinical trials can uncover valuable insights, but it is chronically underutilized. While no trial design is without potential pitfalls, the BOIN approach helps minimize uncertainty. It also helps researchers collect the most impactful data, and this improves the odds of being able to propel theory to therapy.

Reference
[1] Le Tourneau C, Lee JJ, Siu LL. Dose escalation methods in phase I cancer clinical trials. J Natl Cancer Inst. 2009;101(10):708-720. doi:10.1093/jnci/djp079
[2] Zhou H, Yuan Y, Nie L. Accuracy, Safety, and Reliability of Novel Phase I Trial Designs. Clin Cancer Res. 2018;24(18):4357-4364. doi:10.1158/1078-0432.CCR-18-0168

Featured image photo by the National Cancer Institute on Unsplash. Used with permission.

Byondis

Advertisement #5