1. www.quintiles.com | 1
Tomorrow’s Path to Improved Early-Phase
Oncology Drug Development:
Maximizing Quality and Efficiency of Go/No-Go Decisions in Early-phase Studies
Philip Breitfeld, M.D., Vice President Therapeutic Strategy, Quintiles
Eric Groves, M.D., Ph.D., Vice President Center for Integrated Drug Development, Quintiles
Chris Learn, Ph.D., PMP, Senior Clinical Program Manager, Oncology, Quintiles
WHITE PAPER
Executive Summary
Due to the size and scope of clinical trials in this therapeutic area, traditional early oncology development has evidenced high
start-up costs and long durations to advance to the Phase II setting. While the website ClinicalTrials.gov reveals that there
are many products and programs in development, soaring costs, long timelines, and high failure rates result in relatively few
investigational drugs progressing all the way to marketing approval. This is unfortunate for patients who may have benefited
from pharmacotherapy earlier, and makes it challenging for biopharmaceutical companies to achieve a return on investment
and hence to be in the financial position to continue with research and development (R&D) programs for other potential
drug candidates.
The high attrition rate occurring between progression to clinical development and marketing approval suggests that initial
candidate selection processes are not optimal. Given the high costs of development and the demands upon patients who
participate in clinical trials, it is essential to select only those molecules from preclinical development programs that are truly
worthy of advancing to Phase I clinical trials and likely to meet the criteria for success in later-phase trials. More focused and
informed decision-making is therefore vital. Fortunately, advances in molecular biology and patient molecular profiling that may
facilitate targeted therapy have ushered in new hope and enthusiasm for better clinical outcomes. Targeted therapy represents
a transition from broader-acting cytotoxic agents with high toxicity levels toward agents with high specificity and hence
therapeutic benefit for a well-defined group of patients with a particular molecular biological profile.

2. www.quintiles.com | 2
Executive Summary 1
Introduction 3
Landscape Review: Early Oncology Development 3
Targeted therapies as a route to R&D success 4
Need for more preclinical information 5
Falling clinical productivity 6
Future Approaches to Improving Phase I Oncology Studies 7
Key steps to success 7
Going beyond the Maximum Tolerated Dose 8
Biomarker-driven approaches 9
Optimizing dose escalation cycles 10
Study conduct mechanics 10
Summary: proposed improvements 11
Opportunities for Molecular Profiling in Oncology Drug Development 12
Improving go/no-go decision-making 13
Molecular-based selection of trial participants 13
Molecular markers as a screening tool 14
Conclusion 15
Acknowledgement 15
References 16
About the Authors 17
table of contents
For such advances in clinical practice and outcomes to be maximized, it is important to
better understand the biological consequence of treating a biological pathway of interest in
the preclinical setting. Identifying candidate biomarkers for mechanism of action (MOA) and
selection of patients to participate in a given clinical trial is of considerable importance, since
drugs without a biomarker-based patient selection strategy are at a profound disadvantage. A
vision of the future, therefore, would be for newly diagnosed patients to have a comprehensive
molecular profile performed and then be matched to participate in the right trial based on that
profile. Leveraging ‘intelligent biomarker selection’ of patients to participate in early phase
clinical trials has potential to make more efficient go/no-go decisions on product candidates
at the earliest possible stage.

3. 3 | www.quintiles.com
Introduction:
Early-phase Oncology Development Realities
This White Paper examines current issues in progressing oncology compounds from the
preclinical arena through early clinical development in humans to predict future best
practices. Given the increasing R&D costs and regulatory hurdles that must be navigated in
getting new drugs to global markets, coupled with the high failure rate of Phase III studies in
oncology, setting the right course early in clinical development is critical. The framework for
deciding whether and when to progress to human studies, and the goals and expectations of
early clinical development, requires critical appraisal.
Landscape Review:
Early Oncology Development
Oncology pipeline pressures have intensified the demand for speed and productivity. Today,
oncology teams need approaches to make better, faster decisions about whether to kill or
progress potential new products. There is a pressing need for better quality Phase I/II data
to help decrease risk and improve decision making. New approaches are needed in oncology
to better indicate an investigational drug’s viability, identify risks and increase compound
knowledge, and facilitate better go/no-go decisions earlier in the development process.
The good news is that across the continuum of indications, the oncology pipeline remains
robust. There are 1,400 oncology drug development programs in progress (Figure 1), with
81% in the Phase I, Phase II or Phase I/II combined space.1
Of note, breast cancer therapeutics
are the most active area of dedicated research (around 155 programs in progress), followed by
leukemia (around 149) and colorectal cancer (around 80). While solid tumors have historically
dominated the pipeline, possibly the greatest strides in the clinic in the last few years have
been in the effective treatment of hematologic malignancies, with the addition of new
products in non-Hodgkin’s and Hodgkin’s lymphoma, myeloma, myelodysplastic disease
and others, offering substantial therapeutic improvements for patients.
Figure 1 Oncology R&D Pipeline
Source: ADIS R&D Pipeline Database, Sept. 2012
250
PHASE I PHASE II PHASE I/II PHASE III PHASE II/III
Number of Development Programs by Phase
Gastric cancer
Head and neck cancer
Bone cancer
Cervical cancer
Renal cancer
Ovarian cancer
Brain cancer
Multiple myeloma
Liver cancer
Pancreatic cancer
Glioblastoma
Bladder cancer
Lung cancer
Hematological malignancies
Malignant melanoma
Prostate cancer
Lymphoma
Colorectal cancer
Leukemia
Breast cancer
Cancer
N/A
Solid tumors
100500 150 200
The framework for
deciding whether
and when to
progress to human
studies requires
critical appraisal.

4. www.quintiles.com | 4
As might be expected from a development process of winnowing the most promising from
the least promising, and one that proceeds stepwise with each phase of development,
the number of molecules in early phase is substantially higher than in late phase. This is
consistent with the oncology field having a strong pipeline of new candidates. Figure 2
illustrates that the great majority of oncology studies are currently in Phase I/II, as
noted earlier and as expected in a productive field. However, it also shows that while a
substantial percentage of molecules advance from Phase I to Phase II, the overall success
rate for molecules progressing from Phase I to approval is low, with the high attrition rate
suggesting that initial candidate selection processes are not optimal. Given the high costs of
development and the demands upon patients, it is essential to select for Phase I only those
molecules that are truly worthy of advancing, and that are likely to meet the criteria
for success in later-phase trials.
Figure 2 Oncology Trials and Success Rates by Phase2,3
Source: cancer.gov, fastcompany.com
In addition, due to the size and scope of trials, early oncology development is plagued by
high start-up costs and long durations to get to the Phase II setting, which can take from
about 2.5 to 8 years. As a result, while metrics from ClinicalTrials.gov show that there are
many products and programs in development, soaring costs, long timelines and high failure
rates make it very challenging to achieve a return on investment.
Targeted therapies as a route to R&D success
What can the biopharma industry do to improve its R&D success in oncology? In the era
of ‘-omics,’ systems biology, and patient molecular profiling, targeted therapy has ushered
in new hope and enthusiasm for better clinical outcomes, moving away from broader-acting
cytotoxic agents with high toxicity levels. To date, targeted therapy development has
generated a significant number of compounds with improved toxicity profiles for Phase I
investigation, with almost all of these having or needing a surrogate biomarker to define the
Phase I
$18.6M
Pre-Phase I
(preIND)
Phase I/II
$23.7M
Phase II
$28.8M
Phase III
$105.8MSuccess Rates by Phase
Modality
Small
Molecule
Large
Molecule
P1 to P2 63% 84%
P2 to P3 38% 53%
P3 to NDA/BLA 61% 74%
Subm. to
Approval
91% 96%
P1 to Approval 13% 32%

5. 5 | www.quintiles.com
mechanism of action and/or efficacy. Unfortunately, however, even with so many targets in
play for therapeutic development, the number of significant clinical advances with targeted
therapies has fallen short of initial hopes, with few good biomarkers effectively utilized in
early phase oncology development (Figure 3).
Figure 3 The Promise of Targeted Therapeutics4
Need for more preclinical information
The reason for this is simple: cancer is complex. There are hundreds, perhaps thousands,
of genetic changes in the over 200 diseases comprising cancer.5
The dysregulated pathways
in a tumor are well elaborated, redundant, and responsive. This means that more drugs are
not necessarily better, and since combinatorial therapy can be too toxic, sequential therapy
is currently necessary. However, sequential therapy carries its own caveats and risks to the
patients and their treatment. These include:
• Exclusion criteria – The patient may not be eligible for future studies/therapies,
because this eligibility is now excluded by receiving current therapy
• Duration of treatment – The study may go on for a period that precludes or excludes
timely treatment with other therapies
• Cumulative toxicity – The patient may experience accumulating sequelae due to
consecutive lines of therapy
• Sequential therapy – Treatment may not allow for the benefits of targeting pathways
with multiple drugs.
Resisting
cell
death
Sustaining
proliferative
signaling
Evading
growth
suppressors
Inducing
angiogenesis
Activating
invasion &
metastasis
Avoiding
immune
destruction
Deregulating
cellular
energetics
Aerobic
glycolysis
inhibitors
Proapoptotic
BH3 mimetics
PARP
inhibitors
Inhibitors of
VEGF signaling
EGFR
inhibitors
Immune
activating
anti-CTLA4
mAb
Telomerase
inhibitors
Selective anti-
inflammatory
drugs
Inhibitors of
HGF/c-Met
Cyclin-dependent
kinase inhibitors
Tumor-
promoting
inflammation
Genome
instability
& mutation
Enabling
replicative
immortality
Therapeutic Targeting of the Hallmarks of Cancer
Drugs that interfere with each of the acquired capabilities necessary for tumor growth and progression have been developed and are in clinical trals or in some cases approved
for clinical use in treating certain forms of human cancer. Additionally, the investigational drugs are being developed to target each of the enabling characteristics and emerging
hallmarks, which also hold promise as cancer therapeutics. The drugs listed are but illustrative examples; there is a deep pipeline of candidate drugs with different molecular
targets and modes of action in development for most of these hallmarks.

6. www.quintiles.com | 6
It is therefore important to better understand the biological consequence of treating a
pathway in the preclinical setting in order to develop an effective drug for the clinic.
Falling clinical productivity
At the level of clinical trials, this decreased efficiency is clear. The pharma industry has
responded to the decrease in R&D productivity by attempting to control R&D investment.
After a peak in 2008, pharma R&D spend has decreased and flattened, and the total
number of trials in all phases has fallen (Figure 4). Even with this reduction in spend, the
gross-adjusted efficiency of clinical R&D productivity, which historically has not been overly
impressive, continues a downward trend. Against this backdrop, more focused and informed
decision making is vital.
Figure 4 Declining R&D Productivity6
Source: Scannell et al, Nature Reviews, 2012
A key step in informing decisions is to gain a better understanding of the drug during the
preclinical phase, helping to diminish risks at later, more costly, phases of study (Figure 5).
In addition, faster decisions are needed on whether to advance, hold or stop a compound’s
development, such that needless spend is re-appropriated to better development
opportunities. Even idle programs burn substantial money; faster decision making helps
to avoid this. Faster, better informed decisions are also in the best interest of patients
consenting to be subjects in early phase oncology trials. Such decisions help make it
possible to avoid critical pitfalls in Phase I, especially scientifically, operationally, and from an
overall business perspective. Finally, a better understanding is needed of how best to apply
biomarkers and genomic medicine in the future for Phase I studies.
0.1
1.0
10
100
Overall trend in R&D efficiency (inflation-adjusted)
NumberofdrugsperbillionUS$R&Dspending
197019601950 1980 1990 2000 2010
FDA tightens
regulation
post-thalidomide
FDA clears backlog
following PDUFA
regulations plus small
bolus of HIV drugs
First wave of
biotechnology-
derived therapies
The pharma
industry has
responded to
the lack of R&D
productivity by
attempting to
control R&D
investment.

7. 7 | www.quintiles.com
Figure 5 Improving Phase I Oncology Trials
Future Approaches to Improving Phase I Oncology Studies
The main elements of a Phase I program are shown in Figure 6.
Figure 6 Elements of a Phase I Program
Pressures on
Early Oncology
Development
Large #
molecules and
programs
in play
Movement
to targeted
therapeutics
Flat R&D
spend over
5 years
Robust
program
planning
Higher
expectations
from Sr.
Mgmt.
Understand your drug better (and
subsequently decrease risks for
later phases)
Reach faster decisions on whether
to advance, kill, or hold your
compound
Avoid critical pitfalls in Phase 1—
scientifically, operationally, and
overall business perspective
How to incorporate biomarkers
and genomic medicine in the
future of Phase 1 studies
Quintiles Oncology Center of Excellence
Criticality of
biomarkers
TPP
Commercial
Assessment
IP Position
Biomarkers
Pharmacology
CMC
Toxicology
Prococol
PDP
Mgmt Support
Funding
Success Metric
File IND
Site Selection
Enter
Phase 1

8. www.quintiles.com | 8
Key steps to success
Data requirements for Phase I studies are well established and include: toxicology, animal
pharmacokinetics (PK)/pharmacodynamics (PD); Chemistry, Manufacturing, and Controls
(CMC); and limited pharmacology. When combined, these are usually sufficient to establish
a starting dose and potential multi-dose schedule. However, this information is not enough
for a successful program; key additional steps need to be included:
• Clearly identifying the potential commercial target and its commercial context. It is
never too early to draft a Target Product Profile (TPP).
• Obtaining management approval of the draft product development plan/timeline/
cost, putting a development team in place, and securing adequate funding and
Intellectual Property protection.
• Determining whether the in vitro and in vivo pharmacology data are adequate to
support the proposed TPP, and whether there is sufficient information about the
drug’s mechanism of action (MOA).
• Ensuring that CMC is at sufficient scale and stability to support the proposed
development program, and that the formulation is appropriate.
• Identifying candidate biomarkers for MOA and patient selection, since drugs without
a biomarker-based patient selection strategy are at a profound disadvantage. Ensuring
that the necessary assays are validated and set up for rapid turn-around.
These steps are illustrated in Figure 7 – a complex process further complicated by the fact
that, typically, many team members on a given oncology development program have never
participated in an oncology development program before.
Figure 7 Planning and Program Management in Early Oncology Development
Pre-clinical Phase 1IND
Regulatory meetings,
Mgmt of IND Process, Document preparation
Supervision of Program and Execution of Program
Creation and support of Co-Aligned Incentives
Clinical Planning and Delivery
Asset valuation and Gap Analysis
TPP and Product Development
Plan creation
ongoing
Ongoing re-assessment
Identify target
population, endpoints
Plan for sites,
vendor selection
Study delivery and
reporting of results
Biomarker Driven Development
PK/PD analysis and modeling
Biomarker discovery
and validation
Biomarker development
MoA, pathways, safety,
model systems
Dose/schedule
selection
PK/PD
analysis
Clinical biomarker testing, exploratory
analysis, safety, benefit, MoA

9. 9 | www.quintiles.com
Going beyond the Maximum Tolerated Dose
An efficient development process needs to go beyond the goal of establishing the classic
maximum tolerated dose (MTD). The MTD alone is not sufficient to ensure an expedited
Phase II program. A more appropriate goal may be to establish an optimal biologic dose
where possible, as well as focusing on establishing an appropriate target dosing schedule.
The product’s safety profile must be mapped out, but Phase I safety data are limited, as the
small patient numbers give restricted information about infrequent adverse events (AEs). PK
data for the molecule and its major metabolites (as defined in preclinical studies), including
any food effects, should be documented. The MOA should be confirmed by documenting
receptor or enzyme modulation/blockade. It is also important to confirm at an early stage
that biomarker assays really work, to focus on the most useful ones, and if possible, to
develop new biomarker indicators. Positron Emission Tomography (PET) or other imaging
reagents and techniques should be co-developed and piloted as a baseline for subsequent
development phases.
Early development should include tracking markers of efficacy, for example via imaging or
neo-adjuvant pathology. Looking ahead, it may be possible to track circulating tumor DNA or
circulating tumor cells (CTCs). Biomarkers can also be used to refine the link between MOA
and patient benefit. In order to facilitate patient selection in Phase II, it can be useful to add
a Phase Ib trial at MTD in target diseases to the end of Phase Ia.
Biomarker-driven approaches
A core philosophy towards greater achievement in Phase I studies is to support the effort
with a biomarker-driven approach. At the preclinical stage, this includes biomarker discovery
and technical validation, and establishing an understanding of the biology of disease targets
and pathways. The effect of the compound on cell lines should be analyzed, and up- or
down-regulated genes should be identified. Thus, the MOA is hypothesized and mechanism-
based biomarker candidates are elucidated. At the clinical stage, this information is used
to support planning and design of clinical trials, patient screening, prognosis and disease
monitoring. It also helps with correlation to clinical endpoints, confirming prognostic and
therapeutic utility, and demonstrating cost-effectiveness.
Naturally, biomarkers come at a cost. Pitfalls exist, with each approach having risks and
benefits. These choices need careful consideration, paying attention to factors such as how
the samples will be collected and how frequently, whether it is possible to ship samples
outside the country where the trial is being conducted, and whether the consent form is
written to allow samples to be re-examined at a later date.
Where possible, an optimal biologic dose may
be a more appropriate goal to establish than
maximum tolerated dose.

10. www.quintiles.com | 10
There are three main options for Phase I study design:
• Patient-based Phase I: Traditionally, first in man, single agent oncology Phase I trials
are conducted in multi-dose patient trials. These can be quite lengthy, but work reliably.
• Healthy volunteer-based Phase I: Alternatively, single dose and potentially multi-dose
Phase I trials can be conducted in healthy volunteer subjects to save time and money,
provided the agent does not modify DNA, is not likely to be a carcinogen, and does
not carry a long term safety risk. However, when the target patient population is
expected to require different exposure levels than healthy subjects (which would be
unusual in oncology), a bridging study from healthy volunteers to patients will be
required and may cancel out the time savings.
• Mix and match Phase I: In volunteer designs, single dose is usually separated from
multi-dose and single dose information may be used to reduce the number of multi-
dose cohorts. This allows for a switch to patients for a more limited number of levels
of multi-dose escalation.
Optimizing dose escalation cycles
To accelerate Phase I multi-dose trials, a useful approach is to define a functional cycle of
two to six weeks in duration; each dose level cohort receives a cycle of dosing and then is
observed before the next cohort receives the escalated dose. The Phase I target goal needs
to be defined, and the MTD, optimal biological dose and schedule then established. If
possible, the number of dose escalation steps should be reduced, since these impact study
duration and cost. Approaches to reducing the number include:
• Reduce the number per cohort for lowest doses.
• Change from traditional 3+3 design (which does not define the MTD very well) to a
Continuous Reassessment Design (which defines it better).
• Add a PK/PD-guided procedure to define the dose level for the next cohort.
• Avoid focusing on specific disease populations until Phase Ib.
Early clinical modeling and simulation can help translate preclinical data into an accelerated
dose escalation scheme. PK/PD approaches can expand the information from preclinical
studies into Phase I, allowing simulation of exposures and multiple exposures, with the goal
of increasing the efficiency of dose escalation. These are not easy to put in place, but help
make the process work more efficiently, and provide more information about the drug at an
earlier stage.
Early clinical modeling and simulation can help
translate preclinical data into an accelerated dose
escalation scheme.

11. 11 | www.quintiles.com
Study conduct mechanics
For the mechanics of study conduct, smooth processes are important in order to avoid
surprises. The requirements at various stages of early development are illustrated in Figure 8.
Figure 8 Study Conduct Mechanics
Summary: proposed improvements
In summary, key areas for improvement in early phase oncology development, along with
proposed solutions, are as follows:
• Commercial potential is not defined: A gap assessment should be performed and a
TPP and Product Development Plan (PDP) developed at the earliest possible point.
• Criteria for go/no-go are ill-defined and lack tethering to the TPP at key points. This
can be addressed by a TPP and PDP with well-defined criteria for go/no-go decisions.
Management needs to sign on to these criteria.
• Data from the Phase I study are limited and fail to speed program progress to next
stage: Here, more robust and complete preclinical data provide a solution.
• Product program is not adequately resourced to drive quickly and efficiently to key
decision: This can be addressed by real-world objective assessment and
data-driven decisions.
All these elements must be addressed upfront, and followed with ongoing attention to all
details. Total solutions will not be achieved overnight.
Pre-IND
Phase 1
Execution
IND
Acceptance
Governance in place;
Maintenance of good site-sponsor-CRO relationships
Experienced phase I sites
Well written protocol
Electronic CRFs;
How many sites?
• For patient based studies: 3 is the current compromise
between waiting for subjects and sites not being ready
• Experienced high quality monitoring, database mgmt, ….
• Assay turn-around optimized; sample management
under control
• Site concerns (IRB, contracts, internal committees, training)
• Drug supply chain
Regulatory under control
The identified areas
for improvement
must be addressed
upfront, and
followed with
ongoing attention
to all details.

12. www.quintiles.com | 12
Opportunities for Molecular Profiling in Oncology
Drug Development
Speaking at this year’s Economist Global Healthcare Summit, Dr. Stephen Spielberg, Deputy
Commissioner of the U.S. Food and Drug Administration, predicted that diseases would
eventually be classified based on their biological mechanism. “We are dividing up diseases
into ever smaller categories. It has huge implications for those who are discovering and
developing new drugs and huge implications for us as a regulatory agency.”7
Spielberg
referred to the successful drug candidates in this new paradigm as “mini-busters” for their
relatively small piece of the pie compared to the blockbusters of the past. More efficient
and productive early phase oncology development, and specifically genomic or molecular
profiling of patients potentially eligible for such studies, has potential to leverage this new
paradigm and take these new mini-busters to patients.
To achieve the ideal of targeting biologically- and biomarker-defined patient populations for
early phase oncology clinical trials, a better understanding of fundamental cancer biology
and how a drug candidate MOA might counteract that biology is required. This knowledge,
along with understanding of a short list of the biomarkers that influence a drug’s mechanism
of action, can significantly help in improving R&D productivity. The current implementation
of matching biomarker-defined cancer populations to specific trials is inefficient, and future
best practice will depend on unprecedented cooperation between investigators, patients,
biopharma companies and their partner CROs.
An important recent development has been an appreciation of the value to patients of
agents targeting specific functional pathways or circuits in the cancer cell. These can be
fully unlocked only when key nodes in those circuits can be identified. The nodes essentially
dictate how effective the targeted agent might be for a given patient. Figure 9 shows the
epidermal growth factor receptor (EGFR) pathway and highlights the central importance of
k-ras status when predicting clinical benefit in colorectal cancer in the context of monoclonal
antibodies designed to inhibit this circuit.
Figure 9 Patient Selection Using Biomarkers8
Therapeutic Targeting of the Hallmarks of Cancer
Drugs that interfere with each of the acquired capabilities necessary for tumor growth
and progression have been developed and are in clinical trals or in some cases
approved for clinical use in treating certain forms of human cancer. Additionally, the
investigational drugs are being developed to target each of the enabling characteristics
and emerging hallmarks, which also hold promise as cancer therapeutics. The drugs
listed are but illustrative examples; there is a deep pipeline of candidate drugs with different
molecular targets and modes of action in development for most of these hallmarks.
Interacellular Signaling Networks Regulate the Operations of the Cancer Cell
An elaborate integrated circuit operates within normal cells and is reprogrammed to
regulate hallmark capabilities within cancer cells. Separate subcircuits, depicted here
in differently colored fields, are specialized to orchestrate the various capabilities. At
one level, this depiction is simplistic, as there is considerable crosstalk between such
subcircuits. In addition, because each cancer cell is exposed to a complex mixture of
signals from its microenvironment, each of these subcircuits is connected with
signals originating from other cells in the tumor microenvironment.
Hallmark
capabilities
changes
in gene
expression
proteases
adjacent cells b-catenin
extracellular
matrix
growth
factors
hormones
survival
factors
cytokines
receptor
tyrosine
kinases
Apc
TCF4
Myc
abnormality
sensor
E2F p21
p53
DNA-damage
sensor
pRb
cyclin D
p16
Smads
anti-growth
factors
Ras
E-cadherin
integrins
Bcl-2
death
factors
Mobility Circuits
Proliferation
Circuits
Viability
Circuits
Cytostasis and
Differentiation Circuits
Resisting
cell
death
Sustaining
proliferative
signaling
Evading
growth
suppressors
Inducing
angiogenesis
Activating
invasion &
metastasis
Avoiding
immune
destruction
Deregulating
cellular
energetics
Aerobic
glycolysis
inhibitors
Proapoptotic
BH3 mimetics
PARP
inhibitors
Inhibitors of
VEGF signaling
EGFR
inhibitors
Immune
activating
anti-CTLA4
mAb
Telomerase
inhibitors
Selective anti-
inflammatory
drugs
Inhibitors of
HGF/c-Met
Cyclin-dependent
kinase inhibitors
Tumor-
promoting
inflammation
Genome
instability
& mutation
Enabling
replicative
immortality
Sophisticated
patient selection
using smart
biomarkers is
key to unlocking
the potential of
targeted agents.

13. 13 | www.quintiles.com
Improving go/no-go decision-making
Leveraging the intelligent biomarker selection of patients for early phase clinical trials has
potential to make more efficient go/no-go decisions on product candidates at the earliest
possible stage. Figure 10 illustrates the current status on the left-hand side, where a novel
target is discovered, a lead candidate inhibitor identified and moved through standard
toxicological testing and xenograft tumor models, an Investigational New Drug application
(IND) is submitted and Phase I development begins. In this case, the biology of the target
and the MOA of the drug are often not fully understood, so initial testing takes place in an
unselected patient population with enormous biologic heterogeneity.
This issue has been highlighted as a significant issue in the preceding sections. As expected,
most of the time only modest clinical activity is seen in a few patients with this “all-comers”
unselected approach. These inconclusive data, combined with imprecise go/no-go criteria,
can lead to the promotion of drugs to Phase II and even Phase III without a good knowledge
of what biologically-defined patient population is most likely to benefit and without a good
positive clinical signal.
Molecular-based selection of trial participants
On the right-hand side of Figure 10 is an alternative scenario. Here, there is a good biologic
understanding of the target and how this drives the target cancer. Likewise, the way the lead
candidate inhibitor interacts with the target pathway and its biologic consequences are also
understood. This suggests the possibility that patients entering early phase studies of the
drug should be selected based on molecular characteristics believed to be essential for drug
activity. With this set-up, such early phase studies have the potential to be true tests of drug
activity. Such early-phase studies afford the best opportunity to see efficacy if the drug is
truly efficacious, and if little or no activity is detected under these circumstances, it would be
worth considering halting or redirecting development.
Figure 10 Biological Patient Selection
Good
biologic
understanding
of target
Selected
patients in
early phase
development
Good
understanding
of drug MoA
Can make Go/No Go decisions
prior to late phase development
with confidence
Molecular-based
patient selection
Early phase studies
are true test of drug
Biology
of target
poorly
understood
Unselected
patients in
early phase
development
MoA of
drug poorly
understood
Imprecise Go/No Go decisions
prior to late phase development
Heterogeneous
non-biological selection
Early phase studies yield
inconclusive data
Molecular-based
patient selection
can impact the
quality of decisions
to move to late
phase development.

14. www.quintiles.com | 14
While this approach is promising, a short list of critical success factors is needed to leverage
it. These are: (1) a demonstrated link between a biomarker and efficacy in a preclinical model;
(2) a robust assay for the marker; (3) an “all-comers” strategy not being suitable; and finally, (4)
a scientific and clinical development team that has experience navigating these issues.
Clearly, many development programs are not in a good position to select patients for early
phase studies using molecular markers. The fundamental hurdle can be the limits of our
biologic understanding of a target and its inhibitor. There are many examples of targeted
therapies approved for use where selection markers were simply not available at the time
of approval, despite great efforts to understand the drug in these terms. In some cases,
selection markers become known post-approval, improving the benefit-risk profile. However,
even when there is a good biologic understanding of a target and its inhibitor, there are
practical barriers to executing molecularly-based patient selection early clinical development.
Commonly, patients do not come to trial screening with relevant molecular profiling already
having been performed, especially for novel biomarkers. The resulting high screen failure
rates in certain situations may limit investigator and patient enthusiasm. Molecular profiling of
patients prior to consideration of trial participation could overcome some of these barriers.
Molecular markers as a screening tool
Figure 11 outlines the current approach for using molecular markers as a screening tool for
clinical trials. This example involves a patient with non-small cell lung cancer (NSCLC) who
is being screened for a trial where a specific EGFR mutation is an entry criterion. If positive,
the patient has the potential for participation. If negative, then another trial, perhaps one
requiring an Alk mutation, is considered and the screening testing proceeds. If this is
negative, additional options can be considered. The downside of this process is clear: The
iterative/sequential testing of a series of markers costs time and money.
A vision of the future would be for newly diagnosed patients to have a comprehensive
molecular profile performed and then be matched to the right trial based on that profile. The
time and cost advantages are clear. This approach is now starting to be used at some sites
and organizations. Examples include recently published efforts by Daniel Von Hoff of the
Translational Genomics Research Institute (Arizona, USA) and colleagues,9
and by Fabrice
Andre and colleagues at the Institute Gustave-Roussy (France).10
In each, the key theme
is to gain a better understanding of the molecular profile of a patient’s tumor, followed by
the intelligent matching of this profile to known targeted therapies or clinical trials relevant
for the biology of the tumor. To properly leverage such a system, having access to a large
number of studies to match to patients is an important success factor.
The fundamental hurdle to using molecular
markers can be the limits of our biologic
understanding of a target and its inhibitor.
A vision of the
future would be for
newly diagnosed
patients to have
a comprehensive
molecular profile
performed and then
be matched to the
right trial based on
that profile.

15. www.quintiles.com | 15
Figure 11 Approaches for Patient Profiling for Trial Enrollment
Conclusion:
Novel approaches hold potential to improve efficiency
To summarize, drug development today is both expensive and inefficient, and there is
a pressing need to improve productivity if we are to continue to succeed in developing
oncology therapies in the future. This is especially relevant as our understanding of the
biology of cancer is becoming more sophisticated and generating more opportunities, while
also revealing fundamental challenges due to the complexities of this group of diseases.
High quality is essential in early phase oncology development planning, with a particular
need for experienced teams who pay attention to detail, planning and goals. During Phase I
design, all options should be considered thoroughly, including novel approaches to PK/PD
guided escalation and patient/subject selection to ensure we address the unmet medical
needs of tomorrow. In the future, Biomarkers will be a crucial element in getting the
maximum information from a Phase I program. Molecular profiling and leveraging molecular
selection of patients has the potential to significantly improve early decisions in oncology
drug development.
Current approach: Inefficient molecular screening
to determine eligibility for enrollment
Future approach: Efficient molecular screening of cancer patients
to determine clinical trial eligibility
20%
+ Enrolled
80%
Other
Clin. Trials?
5%
+ Enrolled
95%
SOC/Other
Options
Newly
diagnosed
NSCLC
Patient –
Non-
smoker
EGFR
Mutation
Test
Protocol
Requiring
EGFR
Mutation
ALK
Mutation
Test
Protocol
Requiring
ALK
Mutation
Protocol
Requiring
EGFR
Mutation
Protocol
Requiring
ALK
Mutation
Newly
diagnosed
NSCLC
Patient –
Non-
smoker
Genetic
profiling
If EGFR mut+, go to
EGFR mut protocol
If ALK mut+, go to
ALK mut protocol
If both negative, search for another
trial with remaining molecular data

16. www.quintiles.com | 16
References
1. Pipeline information based on Adis R&D Insights data (Sept 2012).
2. Westly E. The Price of Winning FDA Approval. Fast Company, December 1, 2009.
http://www.fastcompany.com/1460583/price-winning-fda-approval
3. Center for Cancer Research Clinical Program web site: http://home.ccr.cancer.gov/
ccrataglance/clinicalProg.asp
4. Hanahan D, Weinberg RA. Hallmarks of Cancer: The Next Generation. Cell.
4 March 2011 (Vol. 144, Issue 5, pp. 646-674) http://download.cell.com/pdf/
PIIS0092867411001279.pdf?intermediate=true
5. National Cancer Institute Office of Clinical Sciences – Oncology web site:
http://physics.cancer.gov/ps1/
6. Scannell JW, Blanckley A, Boldon H, Warrington B. Diagnosing the decline in
pharmaceutical R&D efficiency. Nature Reviews Drug Discovery. March 2012 (11, 191-
200) http://www.nature.com/nrd/journal/v11/n3/fig_tab/nrd3681_F1.html#figure-title
7. Spielberg S. Presentation at the 2012 Economist Global Healthcare Summit.
Addressing tomorrow’s healthcare challenges: a governmental perspective. November 29,
2012 (London, UK). Quote provided at: http://www.epvantage.com/Universal/View.
aspx?type=Story&id=362970&isEPVantage=yes
8. Hanahan D, Weinberg RA. Hallmarks of Cancer: The Next Generation. Cell.
4 March 2011 (Vol. 144, Issue 5, pp. 646-674) http://download.cell.com/pdf/
PIIS0092867411001279.pdf?intermediate=true
9. Von Hoff DD et al. Pilot Study Using Molecular Profiling of Patients’ Tumors to
Find Potential Targets and Select Treatments for Their Refractory Cancer. Journal
of Clinical Oncology. Published online before print October 4, 2010, doi: 10.1200/
JCO.2009.26.5983. Available at: http://jco.ascopubs.org/content/early/2010/10/01/
JCO.2009.26.5983
10. Andre F, Delaloge S, Soria J-C. Biology-Driven Phase II Trials: What Is the Optimal
Model for Molecular Selection? Journal of Clinical Oncology. Published online before
print February 22, 2011, doi: 10.1200/JCO.2010.31.6877. Available at: http://jco.
ascopubs.org/content/29/10/1236.long

17. www.quintiles.com | 17
About the authors
Philip Breitfeld, M.D.
Vice President and Therapeutic Strategy Head,
Oncology Therapeutic Area, Quintiles
Dr. Breitfeld has over 25 years of work experience in oncology, including 20
years of experience in academic medical institutions in the US, and 7 years
of experience in the pharma industry focused exclusively on oncology drug
development and execution of clinical programs. Prior to joining Quintiles he held
senior oncology clinical development positions at BioCryst and Merck Serono.
He has around 50 peer-reviewed publications in the scientific literature, and was a
Visiting Scientist at the Whitehead Institute at MIT.
Eric Groves, M.D., Ph. D.
Vice President, Center for Integrated Drug Development, Quintiles
Dr. Groves has over 25 years of experience in oncology drug developmentas senior
executive or corporate officer, clinician and researcher. During this time, he has
held various senior positions on projects for clinical and pre-clinical development
of oxaliplatin, rasburicase, IL-2, tirapazamine, immunotoxins, bexarotine, ONTAK,
AVINZA, and thrombopoietin. Prior to joining Quintiles, he held senior oncology
development positions at Ligand Pharmaceuticals and Sanofi.
Chris Learn, Ph.D., PMP
Senior Clinical Program Manager, Oncology, Quintiles
Dr. Learn has over 10 years of experience leading investigator led oncology trials
in academic settings and in industry. His expertise includes the development
of molecular immunotherapies for malignant glioma. Prior to joining Quintiles,
he held senior positions in clinical research at Surgical Review Corporation, The
Hamner Institutes for Health Sciences and Duke University Medical Center.
Acknowledgement
The authors would like to acknowledge Jill Dawson, Ph.D. for her assistance in
crafting and editing this document.

20. Copyright ©2013 Quintiles 02.15.16-042013
Contact Us:
US Toll Free: 1 866 267 4479
Direct: +1 973 850 7571
On the web: www.quintiles.com
Email: clinical@quintiles.com