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 ddoherty@thejournalofprecisionmedicine.com