Page 21 - TD-2-2

Tumor Discovery Practice and consideration of master protocol design

master protocols in these particular areas, such as 7.1.2. Subjects with multiple biomarkers
multiple treatments investigated in a single rare disease In master protocols designed to investigate therapies
using a common control arm, the TCM treatment of targeting multiple biomarkers, a challenge may arise if
different diseases with the same syndrome, to promote patients have positive results for multiple biomarkers
the development of master protocols and therapies for of interest. The multi-biomarkers patients are eligible
patients. for more than one subgroup but can only be allocated
7. The practical considerations in master to one of them. FDA recommends, in this case, that the
protocol trial protocol should contain a prespecified plan for allocation
of these multi-biomarkers subjects , such as the pragmatic
[1]
7.1. Design considerations allocation to the eligible sub-study with the currently
7.1.1. A common control arm fewest included patients, the random allocation to one
of the eligible sub-studies . In practice, Lung-MAP
[64]
When a sponsor tends to evaluate multiple therapies trial assigned patients using a prespecified plan that the
simultaneously in a single disease, such as umbrella trials, groups for biomarkers with lower prevalence receive more
FDA recommends that the sponsor use a common control patients . Notably, some researchers argue that patients
[23]
arm of SOC to improve efficiency in randomized trials . with multiple biomarkers might be underrepresented in
[1]
The SOC may change over time as newer drugs develop, some sub-studies, and the variety of schemes for allocation
possibly affecting a long-running master protocol. This could hamper the interpretation, thus proposing some
occurred in Lung-MAP when nivolumab was approved, methods to handle this problem, such as using random or
which is superior to the SOC (docetaxel) selected for pragmatic sub-study allocation .
[64]
the biomarker-matched sub-studies. When comparing
several therapies to a common control group, the test 7.2. Ethical consideration
statistics will be correlated, which will lead to a lower Patients tend to give their informed consent early in the
probability of at least one false positive rejection than clinical trial, which is usually based on scarce protocol
independent trials with a dedicated control group for each information. With more complicated informed consent
comparison [62,63] . Therefore, consistent with the standard caused by the increasing complexity and length of
setting of a series of separate studies, the multiplicity the master protocol trial, patients cannot thoroughly
adjustment as regards the Type I error rate would not be understand these documents and the trial concept to
a need for sub-studies comparing several active therapies give truly informed consent. As the trial progresses, the
to the same control group. This would be reinforced by settings not yet fully established might change because
the fact that different therapies would support separate of the characteristics of master protocol designs . The
[65]
and specific claims of efficacy even investigated globally Declaration of Helsinki considers it crucial and ethically
within the same trial . obligatory that each informed consent document guarantee
[62]
However, another type of error rate led using a common that if new information having an impact on benefits or
control group should be recognized. In master protocol risks becomes available, this information is shared with
[66]
sub-studies comparing multiple therapies to a common the patients, and patients are asked to re-consent .
control group, corresponding regulatory decisions are One approach is that patients could be informed about
correlated using a common control group, which may recruitment status during the informed consent process
inflate the chance of simultaneous multiple false positive and choose another sub-study with an already established
regulatory decisions . For example, if the control group setting. Such an approach would involve the patients more
[62]
obtained weak results by chance, the statistical tests may in the decision-making process, which will challenge both
[65]
then be significant for many of the therapy comparisons, patients and investigators .
even with no treatment effect for these therapies; if the 7.3. Statistics considerations
control group obtained overoptimistic results by chance,
few of the therapies could exhibit a statistically significant 7.3.1. The multiplicity issue
effect, even in the case of true efficacy . In summary, the Multiplicity is the Type I error rate inflation from a
[19]
chance of multiple simultaneous false positive regulatory multiple-testing problem. From the regulatory perspective,
decisions might increase in the case of a common control it has to be adjusted at an acceptable level when several tests
group. To reduce it, there are some strategies such as are performed simultaneously for decision-making about
lower individual values of the Type I error rate for each efficacy. In master protocol trials, the master protocol-
comparison, and more randomized patients to the wise error (MPWE) rate is referred to as the probability
control arm . of declaring at least one of the sub-studies of the master
[62]

Volume 2 Issue 2 (2023) 15 https://doi.org/10.36922/td.342