176 - Next-Gen Clinical Trial Designs (S12E11)

getting new medicines from the lab bench to the
bedside, it's a long road, a real marathon. It
absolutely is. And traditional clinical trials,
well, they're crucial obviously for safety, for
efficacy, but they can take ages. And sometimes
it's not even crystal clear who actually benefits
the most. Right, that's a major challenge. But
what's really exciting, I think, is how the field
is adapting. We're seeing these next -gen clinical
trial designs pop up. Next -gen designs, okay.
Yeah, and they're not just small adjustments.
It's more like a, well, a fundamental rethink
aiming for more efficiency. and really trying
to pinpoint which treatments work for which specific
patients. Okay, I'm with you. And that's exactly
what we want to dig into today, isn't it? We're
focusing on two really interesting ones, basket
trials and umbrella trials. That's the plan.
Think of this listener as your sort of inside
look at how drug development is getting smarter.
We'll break down how they work, the principles
behind them, look at how they can adapt as they
go along these adaptive features. Right. And
we'll try to connect it to real world research,
maybe touching on the things we see in places
like OPR and D Literature, which often covers
the foundational science that makes these trials
possible. So the mission for you today is to
get a really clear picture of basket and umbrella
trials and understand how they could potentially
speed things up, getting good treatments out
there faster. Exactly. OK, let's start with basket
trials. What's the basic idea? Lay it out for
us. Right. So basket trials, the core concept
is testing one specific drug, but across several
different diseases, often different types of
cancer, for example. Different diseases. OK,
how does that work? The key link, the thing they
all have in common, is some kind of shared molecular
feature, like maybe a specific gene mutation.
Oh, OK. So I'm thinking like a literal basket.
You've got different kinds of fruit in there,
apples, pears, whatever. What makes them belong
in the same basket? That's a great analogy, maybe
all the fruit is organic, right? Or from the
same farm. In a basket trial, the different diseases
are the fruit. And the shared characteristic
is that common molecular change, that gene mutation,
that shows up across all those different conditions.
OK, this is where it gets really interesting,
I think. So instead of running one trial for
a drug targeting mutation X in lung cancer, and
then a whole separate trial for the same drug
targeting mutation X in, say, skin cancer, you
could potentially put patients with both cancers,
if they have mutation X, into the same single
trial. That's exactly it. That's the efficiency
gain. Wow, okay. That feels like a big leap.
It is. You focus on that specific biomarker,
that mutation relevant across diseases, and you
test the drug in multiple groups, multiple baskets
all at once, each disease type with the marker.
is its own basket within the larger trial. And
what happens if the drug looks really promising
in, I don't know, the skin cancer basket? Well,
that's another potential advantage. You might
see a path to accelerated approval just for that
specific group, for that indication. Yeah, if
the data is strong for that one basket, patients
with skin cancer who have Mutation X regulators
might be able to approve it for that use. even
if it doesn't work as well in the lung cancer
basket, for instance. So it's not just faster.
It's like a whole new way to think about disease
based on the molecules, not just the location
in the body. Exactly. It helps figure out if
the drug's effect is really tied to hitting that
molecular target you see. No matter where that
target happens to be, it's less about the tissue
type. and more about the underlying biology.
OK, let's try and connect this to the literature,
like those OPRND sources you mentioned. How might
that foundational chemistry work support this?
Well, OPRND, being focused on process research
and development, might not explicitly say basket
trial. But you look for clues. Maybe a paper
discusses optimizing the synthesis or formulation
of a compound. Right, making the actual drug
better. Exactly. And maybe that component is
being looked at for several tumor types that
all share a specific marker like, say, a KRAS
mutation. The work described in OPRD on making
that drug stable, making it deliverable, that's
essential groundwork for enabling a basket trial.
So the chemistry enables the clinical strategy.
Precisely. They're developing the tool needed
for these targeted investigations. Even if OPR
&D examples aren't explicitly labeled basket
trial, the principle is there. Imagine a new
drug inhibits protein XYZ, and a mutation makes
XYZ overactive in, let's say, some leukemias,
some sarcomas, maybe even some brain tumors.
A basket trial could enroll patients with any
of those cancers, but only if they have that
specific XYZ mutation. And you'd see if the drug
works across the board or just in some basket.
Yeah, all within one study. It's a powerful concept.
OK, basket trials, one drug, multiple diseases
linked by a biomarker. Got it. Now flip side,
umbrella trials. The name alone suggests something
different. It does. If the basket groups different
diseases by commonality, the umbrella covers
patients who all have the same disease. Let's
stick with cancer for a moment. It's a non -small
cell lung cancer. Okay, so everyone under the
umbrella has lung cancer. Right, but here's the
twist. Within that single disease lung cancer,
there are actually many different molecular drivers,
different mutations causing the cancer in different
people. Ah, so it's not monolithic. Not at all.
So an umbrella trial starts by doing detailed
molecular testing on each patient's tumor, finding
out which specific mutation or alteration is
driving their cancer. And then what? Then...
Based on that molecular profile, the patient
is assigned to a specific treatment arm within
the umbrella trial. Each arm tests a different
drug designed to target a specific molecular
alteration found in that disease. OK, wait. So
under the lung cancer umbrella, you might have
one arm testing drug for patients with an EGFR
mutation, another arm testing drug B for patients
with an ALK rearrangement, maybe a third arm
with drug C for a KRAS mutation. All happening
simultaneously. Exactly. That's the essence of
it. Multiple targeted therapies tested under
the umbrella of a single disease type, but matched
to patients based on their tumor -specific molecular
fingerprint. It's highly personalized. That sounds
incredibly efficient, too, in a different way.
Testing multiple drugs for one disease in one
go. It is. Think about the infrastructure. One
trial set up, one protocol, but you're investigating
multiple potential treatments and figuring out
which subgroups benefit from which specific drug.
Big savings in time and resources compared to
running five separate trials. And how might the
OPR &D literature reflect this? Would it be about
developing those different drugs? Precisely.
You might see OPR &D papers detailing the synthesis,
the scale up, the characterization of several
different candidate drugs. And maybe the context
is that these are all potential treatments for
different molecular subtypes of, say, breast
cancer. One targets HGR2, another targets PIK3CA,
and other BRCA mutations. The chemistry work
is creating those different spokes for the potential
umbrella trial. Creating the specific tools needed
for each arm of the trial. Exactly. Even if the
paper focuses purely on the chemistry, the scientific
rationale developing distinct molecules for distinct
targets within one disease directly supports
the umbrella concept. So you might have a trial
where all breast cancer patients get screened.
Right. And based on whether they have ATR2 amplification
or a PIK3CA mutation or something else, they
get assigned to the arm testing the drug specifically
designed for that alteration. It helps find the
right drug for the right subgroup much faster.
Okay, both designs, basket, and umbrella sound
like really smart ways to organize research.
Now, you mentioned adaptive features. What does
that mean in this context? Yeah, that's a really
important layer. Many of these next -gen trials,
both basket and umbrella, aren't static. They
often use adaptive designs. Adaptive, meaning
they can change. Kind of, yeah. But in a pre
-planned way. An adaptive design allows for specific
modifications to the trial based on the data
that's coming in during the trial. It lets the
trial learn as it goes. OK, so it's not just
set in stone from day one. It can react. Give
me an example. Sure. So in that umbrella trial
for lung cancer with multiple arms, let's say
the drug targeting the ALK rearrangement looks
really effective early on. An adaptation might
be to start enrolling more patients into that
ALK arm to confirm the finding faster. OK, focus
the resources. Exactly. Or conversely, if another
arm Say the one for the KRAS mutation just isn't
showing any benefit, or maybe there's an unexpected
side effect. A pre -planned adaptation could
be to stop enrolling patients in that arm. It's
called dropping an arm for futility. Saves patients
from getting an ineffective treatment and frees
up resources. That makes perfect sense. Don't
keep throwing good money after bad, so to speak.
What else can be adapted? You can also adapt
patient allocation more broadly. In a basket
trial, If the drug seems to work wonders in the
sarcoma basket, but not the leukemia basket,
you might adapt to focus enrollment more on sarcoma
patients with that biomarker. Or you might adjust
the total number of patients needed the sample
size. If the effect is really strong and clear
early on, you might need fewer patients than
originally planned to prove the point. So the
trial itself becomes more efficient as it learns.
For us trying to follow this, it means the research
is potentially getting to the answer faster and
more reliably. That's the goal. More efficient,
higher chance of success, maybe lower overall
cost. These adaptive features really boost the
power of the basket and umbrella structures.
It sounds almost too good to be true. I mean,
there must be challenges, right? It can't be
that simple. Oh, definitely not simple. You're
right to ask. There are complexities. For one
thing, the statistics get pretty sophisticated.
How so? Well, analyzing data. When you've got
multiple subgroups within baskets or when you've
dropped arms or changed enrollment numbers mid
-trial, it's not your standard analysis. You
need specialized statistical methods to make
sure the conclusions are sound. Right, because
the groups aren't fixed and independent like
in older designs. Exactly. And then there are
the operational hurdles. Just think about the
logistics. Like what? Like screening potentially
hundreds or thousands of patients with advanced
molecular tests to find the ones with the right
biomarker for a basket trial, or the right mutation
for a specific arm of an umbrella trial. Yeah,
that sounds complex. And then managing all those
different arms, making sure the right patients
get the right drug, especially across multiple
hospitals or clinics, it requires a lot of coordination.
And I bet the regulators, like the FDA, have
to look at these differently too. Absolutely.
That's a key area. Regulatory agencies are actively
working on guidance for these novel designs.
How do you review data from a trial that changed
shape part way through? How do you ensure the
results are robust enough to approve a drug?
It's an ongoing conversation. So the challenges
are real, but they're part of the process of
innovating, of pushing forward to find better
ways. That's a good way to put it. The complexity
is kind of the price of progress towards more
efficient, more patient -centered research. OK,
so we've explored basket trials, umbrella trials,
adaptive designs. Let's pull it all together.
What's the big takeaway here for the future of
drug development? I think the main message is
that these next -gen designs are really about
efficiency and precision. They help us learn
faster, use resources smarter, and, crucially,
get closer to the goal of precision medicine.
Matching the right drug to the right patient.
Exactly. Basket trials do it by finding a common
molecular thread across different diseases. Umbrella
trials do it by dissecting a single disease into
its molecular subtypes and targeting each one
specifically. Both get us away from the one -size
-fits -all model. Which ultimately means better
treatments getting to the patients who will actually
benefit and hopefully faster. That's the potential,
yes. By focusing on the underlying molecular
cause, not just the symptoms or the location
of the disease, these designs could really accelerate
how we develop and approve new medicines. Which
leads to a fascinating thought for you, the listener.
As our understanding of the molecular details
of disease gets even deeper, what other kinds
of innovative trial designs might we see emerge?
It really highlights how fundamental science
drives clinical progress. It really does. The
more we learn about the biology, the smarter
we can be in designing trials. It's a continuous
cycle of discovery and refinement. Well, this
has been incredibly insightful. Thanks for walking
us through these next generation approaches.
It's genuinely exciting to see how clinical trial
science is evolving. My pleasure. It's a dynamic
field, and understanding these designs gives
you a real glimpse into where pharmaceutical
R &D is heading. And for you listening, if this
sparked your interest, you might want to explore
maybe the specific statistical methods used in
adaptive trials or perhaps dive deeper into the
role of biomarkers across different diseases.
There's always more to learn. Absolutely. The
landscape of drug development is always shifting,
always innovating.