Statisticians and scientists at Novartis have been at the forefront of developing a new method in early phase oncology trials called the BLRM. Many believe that the BLRM, (short for the Bayesian Logistic Regression Method,) allows for the construction of clinical trials that have the dual benefit of improving treatment for patients participating in the trials, and allowing the trial to complete in a more timely and efficient manner.
The BLRM was built to address two fundamental objectives of a Phase 1 clinical trial: (i) determinig the maximum tolerable dose of a new drug as quickly as possible, and (ii) conducting a trial in a manner that takes into account all of the information available to trial clinicians in order to ensure that patients enrolled in the trial receive the best possible treatment.
The primary objective of nearly any Phase1 clinical trial in oncology is to conduct dose-escalation in a manner that determines the maximum tolerable dose (MTD) of a new drug. Administering a dose which exceeds the MTD can cause serious adverse side-effects for a patient who is already ill. However, a dose which falls below the MTD fails to offer patients all the benefits of a strong therapy. As a result, determining the MTD as quickly as possible, is normally of paramount concern for a Phase 1 trial. Satrajit Roychoudhury of Novartis explains, "while keeping patient safety within acceptable limits, the trials should be small, adaptive, and enable a quick declaration of the maximum tolerable dose (MTD) and/or recommended phase II dose (RP2D)."
Another crucial aspect of Phase 1 trials has been the subject of much discussion in the last few years. When selecting a dose for a patient participating in a Phase 1 trial, can we make use of the information gathered during the course of a trial to ensure improved dosages? For example, data collected at the beginning of a trial might effect how we would adminster a dose later on, if this data were available to clinicians. How can we build trials that allow for such flexibility?
It is generally agreed that the BLRM offers clinicians more flexibility than more traditional designs like the 3+3. The 3+3 outlines a set of rules for administering doses before a trial begins. Due to the fact that these rules are established before data is collected, the 3+3 cannot make use of information gathered during a trial to adjust doses. By contrast the BLRM was designed to make better use of evidence collected duirng a clinical trial. It allows clinicians to make decisions about dosages as a trial proceeds, thereby aiming to ensure that each patient receives the best dose given the data available.
Bayesian statisticians have been advocating for similar trials for many years. However, there has been much discussion and debate on which trial designs are most suitable for this type of procedure given the particular resource constraints of oncology trials. A central question involves the best way to design trials with sufficient flexibility to make use of data acquired during its progress.
In the attached presentation, Satrajit Roychoudhury summarizes a chapter of his work that explains the BLRM model for those unfamiliar with it. In doing so, he provides an eloquent description of the scope and promise of this method. According to Satrajit, benefits to using the BLRM are manifold. It offers clinicians the opportunity to:
(i) Use improved ways to quantify knowledge and assess risk
(ii) Share information collected between doses
(iii) Plan for flexible decision making within the trial
(iv) Improve chances of determining a patient's maximum tolerable dose
In order to see Satrajit's full presentation, please click below:
Interested in applying adaptive approaches in oncology? Click the button below to download a copy of our article 'Are adaptive trials the answer to oncology development success?'