Early Drug Development in the Era of Precision Medicine...Oct 10, 2016 · drug potentially...
Transcript of Early Drug Development in the Era of Precision Medicine...Oct 10, 2016 · drug potentially...
57
CO
GN
OS
CIE
NT
I
In 1990, an international effort was launched to
coordinate the mapping of the human genome
and with it, the opportunity to understand the
unique self at the most fundamental level. In
the year 2000, on the completion of the first
survey of the entire Human Genome Project,
past US President Bill Clinton declared that
the project would “revolutionize the diagno-
sis, prevention and treatment of most, if not
all, human diseases”2. Furthermore, Clinton
recognized in his remarks that “some forms of
leukemia and breast cancer already are being
treated in clinical trials with sophisticated new
drugs that precisely target the faulty genes and
cancer cells, with little or no risk to healthy cells.”
The Precision Medicine Initiative, announced
by US President Barack Obama in his 2015 State
of the Union Address, was the next major
federally-funded effort to fulfill the potential
of a detailed map of the human genome. 3,4
Advances in bioinformatics and scalable gene
sequencing have paved the way for the current
era of individualized medicine, in which
researchers, providers, and patients work
together to develop a more precise approach
to diagnosis and treatment. In the National
Institutes of Health (NIH) definition, “Precision
medicine is an emerging approach for disease
treatment and prevention that takes into
account individual variability in genes,
environment, and lifestyle for each person.”
The NIH is developing a one million subject
cohort as part of the initiative; in which full
genetic sequence data, together with clin-
ical and biomarker profiles, will allow for
the assessment of the interaction between
genotype and phenotype over time, enabling
the development of tailored disease prevention
and treatment approaches. President Obama
highlighted the treatment of cystic fibrosis
(CF) as an example: “I want the country that
eliminated polio and mapped the human
genome to lead a new era of medicine: one
that delivers the right treatment at the right
time. In some patients with cystic fibrosis, this
approach has reversed a disease once thought
unstoppable.”4
The mapping of the human genome and
an intense effort to characterize the disease
at the genetic level has led to the development
of targeted therapies in patients with cystic
fibrosis. Ivacaftor, a cystic fibrosis transmem-
brane conductance regulator (CFTR) potenti-
ator which mitigates some of the pulmonary
consequences of cystic fibrosis, was approved
by the FDA in 2012. Ivacaftor is indicated for
the treatment of cystic fibrosis in patients
who have a G551D mutation in the CFTR gene;
which represent 4-5% of individuals afflicted
with CF worldwide. 5,6
In 2015, 13 of the 45 drugs approved by FDA
were characterized as precision medicines by
the Personalized Medicine Coalition (PMC). 7,8
These approvals continue a trend that began
in 2014 when nine out of 41 new drugs were
classified as precision medicines. In its evalu-
ation of new drug approvals, the PMC defined
them as those therapeutic products for which
the label included a reference to specific
biomarkers that qualified individual patients for
therapy. In addition to the 13 drugs approved in
2015, FDA approved new indications for
previously existing drugs narrowing their
intended use to subpopulations of patients
who could be identified as more clearly
benefiting from the treatment.
C irca the 5th century BC, Hippocrates established the basis for modern medicine, stating “It’s far more important to know what person the disease has than what disease the person has.” The physician’s appreciation of a patients’ unique identity in health and disease is as old as medicine itself. An example in modern medicine is the discovery of blood groups for which Karl Landsteiner was awarded the Nobel Prize in 1930. Recognition of individual blood types allowed for safe blood transfusions and is one of the major landmarks in modern medicine.1
Early Drug Development in the Era of Precision Medicine by Amparo de la Peña, PhD, Mark A. Deeg, MD, PhD
Eyas Raddad, PhD, MBA, Robert Schott, MD, Joanne Sloan-Lancaster, PhD, John R. Stille, PhD
This year, the Obama Administration is launch-
ing the National Cancer Moonshot, a $1 billion
initiative to provide the funding necessary for
researchers to accelerate the development of
new cancer detection and treatments. These
efforts will include cancer-related research
at the NIH, the FDA, and the Departments of
Defense and Veterans’ Affairs, and is expected
to be another example of scientists, cancer
physicians, advocates, philanthropic organiza-
tions, and representatives of the biotechnology
and pharmaceutical industry working together
to focus on major new innovations in the
understanding of and treatment for cancer.9
From the drug development point of view,
incorporating the concepts of precision medicine
early in the drug development cycle is important,
both to set the stage for success in later phase
drug development, and for delivering optimal
therapeutics to individual patients. The current
paradigm of drug development frequently
includes initial safety and pharmacokinetics
testing in healthy subjects in Phase I studies,
followed by efficacy and dose-ranging studies
in patients with the disease of interest. In some
therapeutic areas, particularly in oncology, this
approach is transitioning to a precision-guided
approach with earlier targeting of patients
who may uniquely benefit from a therapy. We
consider the implications of this paradigm shift
on the drug development cycle and the factors
that influence the transition from traditional
drug development to a precision medicine
approach.
The path from traditional to precision
medicine starts in early clinical development
Historically, most novel therapeutics have been
developed without a full understanding of the
molecular mechanism of the disease and the
treatment. A transition to precision-medicine
based strategies of drug development is made
possible by a greater understanding of the
molecular basis of health and disease, the
proliferation of “-omics”, vast advances in
technology, and innovative methods that
enable to characterize the heterogeneity in
human populations at the molecular level.
To make this transition, there needs to be a
paradigm shift across multiple drug develop-
ment stages, most notably in early drug
development. Early stages of development
comprise clinical and pre-clinical studies and
supporting research activities such as toxicology,
chemistry, manufacturing, and control (CMC),
development of analytics and diagnostic tools,
that culminate with the compound achieving
proof of concept (POC).
Typically, new drugs enter early clinical
development with a strategy that defines a
set of clinical studies, including their high-
level designs, and other supporting research
activities to be performed. Many drug
development programs advance to clinical
testing with at least one defined tailoring
hypothesis. As concepts of precision medicine
become increasingly widespread in the
scientific community, these programs face
key strategic alternatives, each with different
implications and difficult tradeoffs that differ
from the traditional development paradigm.
These alternatives depend on the point in time
when a specific tailoring hypothesis becomes
the main focus of the development program. At
one extreme, a traditional (Broad/Exploratory)
development paradigm starts with the clinical
evaluation of a broad population, to gather
information that may allow for identification
of potential tailoring hypotheses later in
development. On the other extreme, the
development path could commence with a
singular tailoring hypothesis in a narrowly
defined patient population (Narrow/Focused).
In between these two extremes, a myriad of
possibilities for development choices exist
along the development continuum. In particular,
the ends of the spectrum of development
strategies translate into radically different
choices for early clinical development (Table 1).
Broad/Exploratory strategies start with a largely
agnostic view of the ultimate tailoring
approach and intend to prove the therapeutic
concept by enrolling heterogeneous patient
populations in the early clinical trials. Taking
advantage of this diversity, Phase 1 trials may
yield unexpected insights that can be further
explored with flexible or adaptive trial designs.
As the molecules progress to Phase 2, studies
become larger and more complex, due to the
necessarily lower signal-to-noise ratio observ-
able in an early broad population. Failure to
prove the therapeutic concept in the general
population is likely to result in termination of
the drug. Alternatively, the drug response may
subsequently be explored in a further refined
subpopulation. These further explorations can
begin at the POC study, with the potential of
extending to more focused studies in other
subpopulations, which can be either identified
at the POC stage or defined based on a strong
scientific theory.
Narrow/Focused strategies, in contrast, start
with a strong predisposition towards a single
-or a select few- tailoring hypotheses, upon
which the proof of the therapeutic concept
is focused. Typically, the probability of success
is the highest for these tailoring hypotheses,
and positive proof of the hypotheses is a
prerequisite for the continued development
of the program. Once proven successful, the
program can continue on this tailored path,
as well as potentially expand into other
subpopulations and/or broader populations.
In the Narrow/Focused paradigm, the initial
studies are typically smaller, due to the
expectation of a higher drug effect size.
However, they also present operational
challenges for subject recruitment, given
the demand for a more stringently defined
patient population, rather than a healthy
subject population. Indeed, the first study in
humans may also serve as the POC study in
this paradigm, which differs from the traditional
drug development approach in which the POC
study is usually preceded by single dose and
multiple dose safety/pharmacokinetics studies.
Furthermore, the exploration of alternative
tailoring hypotheses is typically avoided in
the focused strategy approach.
58
Development Strategy Broad/Exploratory Narrow/Focused
Hypothesis testing Broad Specific
Wide exploration of tailoring Pre-defined tailoring hypothesis
hypothesis at the core
Disease Population Extensive Narrow
Heterogeneous population to Homogeneous sub-group focus
built efficacy hypothesis to confirm efficacy hypothesis
Phase I Safety Study Population Typically Healthy Subjects Target Patient Population
Understand safety in Safety in targeted population
heterogeneous population most applicable
Disease Markers Search and Find Refine and Confirm
Evaluate many biomarkers, Select a few, specific prognostic
find those that respond and predictive biomarkers
Strategic Tailoring Decision Post-PoC Early
Decision as late as Phase II/III; Build decision and companion
triggers companion diagnostic diagnostic into early development
Study Designs Adaptive Characteristics Fixed Study Design
Explore breadth and adapt Greater initial information allows
as information is acquired for a more defined design
The challenge, thus, is having enough relevant
information to confidently choose a tailoring
hypothesis early in drug development.
Implications of data and decisions along
the development strategy continuum
Throughout the development of a drug, making
the choices between the two extremes of the
competing early development strategies results
in significant downstream implications (Table 1,
Figure 1). The decision tree in Figure 1 depicts a
simplified view of the tradeoffs a development
team must consider based on the available
information. The primary decision is between
the development of: 1) a broadly applicable
drug potentially benefiting a large population,
with an associated greater market potential or
2) a drug with a narrow patient population,
albeit with a higher effect size and probability
of technical success.
necessary to reach the nodes in the decision
tree, including the probabilities for each
branch, and a probability-weighted overall
outcome based on the tenets of decision
sciences, can be performed to inform and
improve development choices. This informa-
tion can subsequently be used to contrast the
expected cost and duration for competing
strategies and thus facilitate better decision
making. Such assessments are dependent on
the applied designs of the studies, as well as
the knowledge base of the novel therapeutic,
the disease state, the target and the strength
of the tailoring hypothesis.
An example of the Broad/Exploratory approach
is Eli Lilly’s evaluation of a new compound
in a proof of concept trial. An evaluation
of multiple biomarkers -at baseline and over
The reality of drug development is quite
complex, and there is a continuum of options
within the extremes depicted in Figure 1. A
development path that starts broadly can end
up with a narrowly defined subpopulation,
by a midstream shift to a more tailored
approach based on emerging and prior
evidence. Similarly, a path that starts narrow
does not preclude a population expansion
that can end up with broader applicability
and a larger ultimate target population, as
new information is gathered from both
clinical trials and from the extended scientific
community. However, the choice of starting
strategy does indeed have differing implica-
tions, and it is important to properly frame
those implications in order to choose the
most appropriate strategy for the compound.
A formal analysis of the resources and time
59
CO
GN
OS
CIE
NT
I
61
CO
GN
OS
CIE
NT
I
time- was built into the study design, as
an attempt to identify higher responding
subpopulations based on the underlying
pathogenesis. However, the sample size and
randomization algorithm were not designed
to ensure adequate power for any specific
tailoring hypothesis. Although several
biomarker trends were observed, they were
not compelling enough to continue the
program. The limitation of this approach is
that only subpopulations with a substantial
effect size have the potential to be identified.
Hence, as in this example, when the data
for the primary objective are not sufficient
to support continued development in the
broad population, investment in additional
A
B
C
D
E
F
G
H
Drug
Development
Candidate
Broad:l Explore precision hypothesesl Tailor if data supports
Focused:l Single tailoring hypothesis as the core of development
Hypothesis works broadly;continue development
Inadequate response; precision hypothesis identified
Does not work; tailoring hypotheses not supported
Tailoring hypothesis works
Continue with same tailoring hypothesis
Precision hypothesis works
Expansion successful
Development
Branch
Tailoring hypothesis fails
Explore populationexpansion
Precision hypothesis fails
Expansion not successful
Decision Data/Resolution of Uncertainty
Beneficial drug ina broad populationwith large marketpotential
Beneficial for anarrow population
Expensive late failure
A failed drug withremnant abandonedpotential
A drug with multipletarget populations
A drug with narrowsubpopulation
A drug with narrowsubpopulation
Inexpensive failed drug with remnantabandoned potential
Late PhaseImplication
Figure 1. A decision tree depicting selected broad versus focused drug development strategies. Multiple strategies can be adopted for a new drug candidate
as it enters clinical development, and each of these strategies results in unique late phase implications. Many hybrid strategies can be envisioned.
exploration of subpopulations is also highly
unlikely unless a very strong and compelling
signal is observed. In this case, the scenario
that unraveled is similar to branch D in the
decision tree (Figure 1).
The broad strategy can move a drug with wide
applicability to later stages of development,
yet maintains an opportunity for finding
benefit in a drug that fails the broad hypothesis,
if an alternative tailoring hypothesis is
identified and subsequently proven during
early development. The implied risk is that
of failing to identify a key subpopulation due
to inadequate representation in the early
studies, and the risk of a delayed, and therefore
more expensive, failure (if the drug truly does
not work) due to sequential POC studies in
two or more different subpopulations.
Choosing the narrow strategy for early
development has different implications. In
this focused approach, a drug may show
clinical benefit in the initial targeted patient
population, supporting continued development
in this paradigm. Furthermore, the option to
continue to evaluate in a broader and/or other
defined subpopulations, which will require
additional investment, is maintained. One
drawback to the narrow strategy is that the
burden of proof at the initial stage is higher
than for the traditional approach. Additionally,
62
The abundance of available data leads to tech-
nology and privacy concerns, together with the
difficulty to adequately interpret the informa-
tion. Thus, flexibility is paramount to be able
to adapt the development approach as new data
is received. Because every drug and target is
different, and because information is available
in stages, Bayesian methodologies, where new
information is used to refine or change
the approach in development, should be
considered. This approach can also maximize
the use of real-time data obtained via techno-
logical advances; which will in turn improve
and accelerate the way in which prescribers
and patients are able to modify dosing regi-
mens, adapting them to their specific needs.
The strength of the data to be used in a tailoring
approach ultimately affects the strategy chosen
for the development of a given drug. The
decision tree depicted in Figure 1 merely
outlines the strategic conundrum that is
typical in early development planning;
many permutations or hybrid strategies can
be devised for a novel therapeutic in devel-
opment. In other words, the strategy may
represent a continuum of choices across the
individual decisions depicted in Table 1.
Despite present challenges for progressing to
a more focused or tailored approach, almost
30% of drugs approved in 2015 were precision
medicines. The US Administration’s continued
commitment to accelerating medical research
is exemplified by the Precision Medicine
Initiative in 2015 and the new National
Cancer Moonshot in 2016.
The ultimate goal is to benefit the patients; the
question is how to make it more widespread –
and whether precision medicine is appropriate
in all cases.
on different label indications, such as in
the example of ivacaftor mentioned above.
Other considerations when making drug
development choices are the impact on
operational efficiency at the trial level/program
level and the impact on program costs and
length, when companion diagnostics are
required. At the enterprise level, portfolios
need to balance their cost, quality, and
operational efficiency. Including a large
proportion (>20%) of Phase 1-3 trials with
a focus on precision medicines in the
portfolio, significantly reduced the overall
operational efficiency. 10 The same effect was
observed when a higher percentage of a
portfolio consisted of patients with rare versus
“traditional” diseases. However, there may
be justification for these choices through
benefits in value or attrition. Thus, these factors,
together with the regulatory environment,
need to be taken carefully into account when
development options are considered.
Complexities along the journey towards
a Precision Medicine approach
Advances in technology have yielded enormous
repositories of information, as the sequence
of entire individual genomes can be catalogued
and queried, and have been instrumental in
improving the understanding of the molecular
mechanisms defining health and disease.
Adding more precision to medicine requires
a continued collaborative effort between
the discovery engines of academics, national
institutions, industry, and health care
organizations, in order to better systematize
and organize this plethora of information for
optimal application in drug development.
This is crucial for growing the understanding
of the heterogeneity and complexity of drug
responses, diseases, and patient populations.
The objective of Precision Medicine is to link
information, such as the phenotypic and
genotypic data available from the Human
Genome project, to therapeutic information
which will benefit the patient.
if a broader population is subsequently
evaluated (population expansion), there is
a risk that development will remain more
focused than if an agnostic broad approach
had been initially used. Moreover, even if the
chosen focused population failed to respond,
a patient profile which could benefit from
the drug (and was missed) could still exist,
particularly if the choice of focused population
was based on limited early data. A benefit of
the narrow strategy, however, is that a failure
to prove the core hypothesis would likely
result in a quick and relatively inexpensive
stop of the development program.
The probability of following each of the
branches in the decision tree in Figure 1
depends on the quality and quantity of
information known at the time of making
each decision. When knowledge about the
heterogeneity of disease and response is
sparse, and there are many competing tailoring
hypotheses, the ultimately successful drug is
likely to journey through different branches
of the decision tree (e.g., branches A,B, E,F
or G) than the drug that ultimately fails (e.g.,
branches C, D, or H). In both cases, the broad
strategy would be a better fit for the drug, as
it has a shorter path to success and lower risk
of unjustified failure (branch H). On the other
hand, information in support of a narrow and
specific subpopulation with clearly defined
biomarkers (e.g., genetic polymorphisms)
favors the narrow and focused strategy
(branches F and G) and would ultimately
result in a precision medicine.
The choice of the most appropriate population
to include in the POC studies is extremely
relevant; stronger “tailoring” evidence allows
to study a more focused patient population,
and to influence the study design to directly
test the precision medicine hypothesis. At
the start of each development program, it
is important to define the most relevant
information to identify the right population,
and the tools to be used for that purpose.
These choices can have a significant impact
35
CO
GN
OS
CIE
NT
I
References
1. Landsteiner K, Levine P. on individual differences in human blood. The Journal of Experimental Medicine. 1928;47(5):757-775.2. June 2000 White House Event_files. 2000; https://www.genome.gov/10001356/june-2000-white-house-event/.3. President Obama’s Precision Medicine Initiative. 2015; 5. Available at: https://www.whitehouse.gov/the-press-office/2015/01/30/fact-sheet-president-obama-s-precision-medicine-initiative.4. Remarks by the President in State of the Union Address. 2015; https://www.whitehouse.gov/the-press-office/2015/01/20/remarks-president-state-union-address-january-20-2015#content-start. Accessed January 20, 2015, 2015.5. Brodlie M, Haq IJ, Roberts K, Elborn JS. Targeted therapies to improve CFTR function in cystic fibrosis. Genome Med. 2015;7:101.6. FDA. KALYDECOTM (ivacaftor) Tablets Label. 2012:13.7. Pritchard D. Personalized Medicine at FDA: 2015 Progress Report. Journal of Precision Medicine. 2016:2(1):40-42.8. Personalized Medicine Coalition (PMC). 2016; http://www.personalizedmedicinecoalition.org/.9. FACT SHEET: Investing in the National Cancer Moonshot. 2016; https://www.whitehouse.gov/the-press-office/2016/02/01/fact-sheet-investing-national-cancer-moonshot.10. Ringel M, Martin L, Hawkins C, et al. What drives operational performance in clinical R&D? Nat Rev Drug Discov. 2016;15(3):155-156.
Amparo de la Peña, Ph.D, is a PK/PD Research
Advisor in The Chorus Group, Lilly Research
Laboratories, Eli Lilly & Co. A native of Uruguay
and University of Florida graduate, she combines
her multicultural and scientific backgrounds to
facilitate drug development and support scientific
and community organizations.
Mark Deeg, MD, Ph.D, FNLA. is Sr. Director and
Chief Medical Officer of The Chorus Group. His
interests include the use of pharmacogenomics in
drug development.
Eyas Raddad, Ph.D, MBA, is a Sr. Research Advisor in
The Chorus Group. He helped advance more than 50
novel therapeutics in discovery and development stages,
where he utilized various modeling techniques to aid
decision making. His current research focus includes
translational therapeutic research and the application
of quantitative and decision analytic methods in drug
research.
Robert J. Schott, MD, MPH has been a Senior
Research Fellow at Eli Lilly and is currently working
with The Chorus Group as a Research Physician. He
brings a cardiovascular clinician’s perspective to drug
development, with an interest in clinical trial design
and conduct across many therapeutic areas.”
Joanne Sloan-Lancaster, Ph.D, is a Senior Clinical
Research Advisor in The Chorus Group. She combines
her discovery biology, analytical, immunology and
pharmacology experience to clinical study design and
drug development.
John R. Stille, Ph.D, is a Senior Clinical Research
Advisor in The Chorus Group. He combines
biochemical, adaptive, and operational strategies
to clinical study design in drug development.
P E R S O N A L I Z E
Connecting the World of Precision Medicine
May/June 2015
thejournalofprecisionmedicine.com
Building an ‘open source’ecosystem for
precision medicine
Blood Work – liquid biopsies
come of age
Too much precision,
not enough medicine?
Why precision medicine
needs precision design
Venter on life and longevity
P E R S O N A L I Z E
October/Novem
ber 2015
CDx/Rx - Complicated companions?
Personalized Health Care - A new model
for improved clinical delivery
NGS-Movingtowards clinical
actionability?
Soon-Shiongdances with
proteins
Waves of precisionroll into the bay
RThe Evolution of “P–Medicine”
thejournalofprecisionmedicine.comP E R S O N A L I Z E
January/February 2016
NGS- a path to next gen CDx tests?
Precision Medicine - a new paradigm in antimicrobial resistance?
FDA seal of approval -13 stars of 2015
Rothberg on the virtuous circle of life.
Cambridge UK -the original DNA hub
Precision Medicine-the iceberg of hopeand complexity
thejournalofprecisionmedicine.com
P E R S O N A L I Z E
March/April 2016
Imaginary Worlds -reproducibility and modeled claims in precision medicine.
Guardian Angels - delivering precision medicine to the community.
Digital Biomarkers -a new frontier in precision oncology.
Precision Promise - prostate tumor genes tell the story.
A Silent Epidemic -a precision - systems approach to Alzheimers.
Pharma sets its sights - the continued journey towards targeted therapeutics.
thejournalofprecisionmedicine.comP E R S O N A L I Z E
June/July 2016
Call to Action - Breaking downBarriers for Patient Participation in Oncology Clinical TrialsLiquid Biopsies bleed into the clinic
Precision Medicine - a new paradigm in drug development
DTC & CD products – a move towards clinical integration?
Dementia – Immunology holds clinical promise
Glimcher glimpses the future
thejournalofprecisionmedicine.com
P E R S O N A L I Z Ethejournalofprecisionmedicine.com
Our next issue will be published in December, so to ensure
regular copies please subscribe online at the link below and choose
your preference for either print or online versions or both!
If you have any suggestions for contributed articles we
would be delighted to hear from you.
Please email [email protected]