The Pharmacogenomics Journal (2006) 6, 375–380
& 2006 Nature Publishing Group All rights reserved 1470-269X/06 $30.00
The drug diagnostic co-development concept
paper
Commentary from the 3rd FDA-DIA-PWG-
PhRMA-BIO Pharmacogenomics Workshop
LM Hinman1, S-M Huang2,
J Hackett3, WH Koch4, PY Love5,
G Pennello6, A Torres-Cabassa5
and C Webster7
1
Hoffmann-La Roche Inc., Nutley, NJ, USA;
2
Office of Clinical Pharmacology and
Biopharmaceutics, Center for Drug Evaluation
and Research (CDER), Food and Drug
Administration (FDA), Silver Spring, MD, USA;
3
Office of In Vitro Device Evaluation and Safety,
Center for Devices and Radiological Health
(CDRH), FDA, Rockville, MD, USA; 4Roche
Molecular Systems, Pleasanton, CA, USA; 5Office
of Combination Products, Office of the
Commissioner, FDA, Rockville, MD, USA;
6
Diagnostics Branch, Division of Biostatisics,
CDRH, FDA, Rockville, MD, USA and 7Millenium
Pharmaceuticals Inc., Cambridge, MA, USA
Correspondence:
Dr LM Hinman, Hoffmann-La Roche Inc., 340
Kingsland Street, Nutley, NJ 07110-1199, USA.
E-mail: lois.hinman@roche.com
Disclaimer: The views expressed here are
those of the authors and do not necessarily
reflect those of FDA, Hoffmann-La Roche or
Millenium.
Received 30 January 2006; revised 7 March
2006; accepted 13 March 2006;
published online 2 May 2006
www.nature.com/tpj
REVIEW
At the Washington DC Pharmacogenomics in Drug Development and
Regulatory Decision-Making: Workshop III – Three Years of Promise,
Proposals and Progress on Optimizing the Benefit/Risk of Medicines (11–13
April 2005), one break-out session (Track 2) focused on co-development of
therapeutic drug and diagnostic products. The Food and Drug Administra-
tion (FDA) released a draft concept paper shortly before the workshop was to
convene. Track 2 was a forum for initial discussion of the content of the
concept paper, and industry’s initial reactions. After the workshop, formal
commentaries on the co-development concept paper were submitted by
several trade associations (e.g., Pharmaceutical Research and Manufacturers
of America (PhRMA), Advanced Medical Technology Association (AdvaMed),
American Association for Clinical Chemistry) and individual companies to
FDA’s Docket No. 2004N-0279. This paper includes a summary of the key
features of the draft concept paper, the discussion in Track 2 of the April,
2005 meeting and highlights of the industry comments submitted to the
FDA docket following the meeting.
The Pharmacogenomics Journal (2006) 6, 375–380. doi:10.1038/sj.tpj.6500392;
published online 2 May 2006
Keywords: pharmacogenetics; drug/diagnostic co-development
Introduction
At the Washington DC Pharmacogenomics in Drug Development and Regulatory
Decision-Making: Workshop III – Three Years of Promise, Proposals and Progress
on Optimizing the Benefit/Risk of Medicines1 (11–13 April 2005), Track 2 focused
on co-development of therapeutic drug and diagnostic products. Many of the
issues were derived from a preceding workshop (July 29, 2004) also held in
Washington DC.2 Both workshops were organized by the Food and Drug
Administration (FDA) in conjunction with the Drug Information Association
(DIA), Pharmaceutical Research and Manufacturers of America (PhRMA), the
Biotechnology Industry Organisation (BIO) and the Pharmacogenetics Working
Group (PWG) to discuss pharmacogenomics in drug development and regulatory
decision making. During the July 2004 workshop, a complex discussion took
place on the co-development of a therapeutic drug with a diagnostic device for
use in selection of a drug in specific patient groups for specific indications.
Speakers from the FDA and both the drug and diagnostic industries participated
 The drug diagnostic co-development concept paper
LM Hinman et al
376
in the meeting, and took part in both presentations and
panel discussions. At the conclusion of the 2004 meeting,
Dr Larry Lesko (Director of FDA’s Office of Clinical
Pharmacology and Biopharmaceutics and leader of the
Agency’s initiative on pharmacogenomics) confirmed that
the FDA would take industry’s concerns into serious
consideration.2 He further stated that a cross-functional
FDA team, with representatives from the drug and diag-
nostic review divisions, as well as the Office of Combination
Products would work to develop a draft guidance in this
complex field.
In advance of the guidance, the FDA released a draft
concept paper on April 7, 2005,3 shortly before the April
2005 workshop was to convene. Track 2 of the workshop, as
reported here, was a forum for initial discussion of the
concept paper with session presentations from the FDA on
the content of the concept paper4 followed by industry’s
initial reactions.5 Advance copies of the presentation
materials were not available; therefore, audience participa-
tion reflected early review of the concept paper and
reactions to the presentation and perspectives from this
and previous meetings. After the workshop, formal com-
mentaries on the co-development concept paper were
submitted by several trade associations (e.g., PhRMA,
AdvaMed, American Association for Clinical Chemistry)
and individual companies to the FDA’s Docket No. 2004N-
0279. A brief summary of the key features of the draft
concept paper, the discussion in Track 2 of the April, 2005
meeting and highlights of the industry comments sub-
mitted to the docket are provided below.
The draft drug diagnostic co-development concept
paper3,4
The draft concept paper, released on April 7, 2005,3 describes
the Agency’s preliminary thoughts on how to prospectively
co-develop a drug or biological therapy (drugs) and a device
(test) in a ‘scientifically robust and efficient way’. The FDA
Figure 1 The drug–test co-development process describing key steps during development. From the FDA draft concept paper on drug diagnostic
co-development, April, 2005.
The Pharmacogenomics Journal
noted that these concepts do not constitute guidance but
reflect perspectives under consideration. Although a de-
tailed description of its content is beyond the scope of this
article, the concept paper makes the point that while the use
of diagnostic tests in conjunction with drug therapy is well
established (e.g., estrogen receptor hormones, Her 2 neu
DNA and protein, EGFR protein), new technology has
enabled the development of pharmacogenomic in vitro
diagnostic tests to identify inter-individual variability in
drug responses, both in efficacy and safety. Such tests might
be used to guide drug selection and target therapy to
selected patient subsets. Distinctions are made in the paper
between cases in which the use of the test is mandatory (i.e.,
both the drug and diagnostic are necessary) in making a
therapeutic decision for the selection of a drug for a specific
patient and cases in which the test may be informative but
not mandatory. In the latter situation, for example, testing
may be helpful in understanding the mechanisms and
determining how to enrich or select populations to study.
The principles discussed in the concept paper are also
relevant to pharmacogenomic testing for dosing determina-
tion of drugs.
In the concept paper, ‘co-development’ refers to consid-
erations when both the therapeutic drug and diagnostic test
are necessary for the intended clinical indication. These
products may raise development issues that affect both the
drug and the test products. (Co-development products may
or may not meet the regulatory definition of a combination
product under 21 CFR 3.2(e).) The paper presents informa-
tion in five areas: regulatory considerations, analytical test
validation, clinical test validation, clinical test utility and
labeling. Each of these areas addresses prospective co-
development of a single test in conjunction with a single
drug.
Regulatory considerations focus on the need for industry
and the FDA to include both drug and device firms and
Center for Drug Evaluation and Research (CDER) and Center
for Devices and Radiological Research (CDRH) in the

Figure 2 Drug–test co-development process: formal Industry–FDA interactions. From the FDA draft concept paper on drug diagnostic co-
development, April, 2005.
development discussions. The paper contains two diagrams
(Figures 1 and 2), which are included in this report for
reference. Figure 1 highlights key steps in the co-develop-
ment process that relate to various sections of the concept
paper and the type of issues that may be considered during
development. For example, the process begins with basic
research, target selection and validation and ends with
clinical validation for the use of the drug and the diagnostic
test. Figure 2 illustrates the parallel formal Industry–FDA
interactions that would occur with both the drug and
diagnostic review processes when two investigational appli-
cations are used. For example, the figure identifies formal
meetings that would occur in the investigational develop-
ment for the IDE(investigational device exemption) applica-
tions and the drug (IND – investigational new drug). These
interactions would be with one or more of the following
FDA centers: CDER, Center for Biologic Evaluation and
Research (CBER) and CDRH. The section on regulatory
considerations also contains a list of questions that a
sponsor might want to consider in planning early FDA
discussions: for example, what analytical and clinical data
are needed to support retrospective development and
validation of the diagnostic test, or what is the most
appropriate regulatory pathway for the proposed project.
Analytical test validation
The paper provides general recommendations and prelimin-
ary details on the type of analytical and clinical validation
information that may be expected for the co-developed
diagnostic test. The paper recognizes that a key hurdle for
prospective co-development relates to the availability of
pharmacogenomic samples. There are recommendations for
the device description, analytical studies, software and
instrumentation, feasibility studies and considerations for
test changes during the late stages of co-development. The
paper includes examples of the type of information that
may be necessary to submit; for example, estimates of
The drug diagnostic co-development concept paper
LM Hinman et al
377
clinical test performance and test quality measurements,
primarily from the diagnostic perspective.
Clinical test validation
The section on clinical test validation and statistical
considerations focuses mainly on evaluation of clinical
performance of the test and setting test parameters when
the test will be used to select patients to receive or avoid a
drug, or to stratify patients in some manner; for example, in
the context of the intended clinical outcome. It provides
information on statistical considerations in co-develop-
ment. Also, there is a related section on pre-clinical pilot
feasibility studies.
Clinical utility
The section on clinical utility discusses various clinical trial
design considerations to help minimize bias and to assure
that results of the trials address the primary study hypoth-
esis. The amount of clinical trial data required to verify
utility will differ depending on prior knowledge of patho-
physiology involved, the magnitude of differences in
outcome, and the extent of previous relevant clinical
experience. Clinical development considerations were simi-
lar to expectations to support other clinical hypotheses (i.e.,
the end points that are used to establish safety and
effectiveness of the drug and diagnostic test should reflect
the end points used to establish the test population
selection). Also, this section addresses statistical issues in
the context of the therapeutic and the diagnostic. The FDA
also notes that pre-specified retrospective clinical utility
studies utilizing previously collected pharmacogenomic
samples may be possible. In addition, the paper discusses
data collection, data standards, statistical issues and infor-
mation that would be submitted to allow an FDA audit of
the clinical and other material.
The recommendations in this section are based on the
premise of a prospective co-development model; for example,
The Pharmacogenomics Journal
 The drug diagnostic co-development concept paper
LM Hinman et al
378
the test and drug go through the early stage of development
in parallel, emerging from early development together to be
used in conjunction in pivotal trials. This proposed pathway
indicates that clinical utility should be established for the test
with the specific drug. As such, the FDA would expect
concurrent reviews and approval of the therapeutic drug and
diagnostic test to achieve the intended indication of both
products. The co-review process could be facilitated in some
cases by the Office of Combination products (OCP).
The drug diagnostic co-development concept paper con-
cludes with a glossary of terms, examples of device
descriptors and examples of study design issues. Labeling
of the drug and diagnostic (e.g., what information should be
in each label or both) was not addressed in the concept
paper. At a recent advisory committee for pharmaceutical
science–clinical pharmacology subcommittee meeting,7
however, the FDA presented general recommendations for
when and where the pharmacogenomic information will be
included in the drug/biologics or device labeling.
Industry feedback/discussion on the draft concept
paper – April, 2005 meeting5,6
During the 45-min workshop session, the audience provided
its feedback on the presentation on the co-development
concept paper. Industry’s critical concerns on the concepts
reflected regulatory and procedural issues previously identi-
fied in the July, 2004 workshop and in other informal
discussions between the FDA and industry. These included
the following points and recommendations for information
to include in a subsequent FDA draft guidance:
A. Practical concerns
Overall, industry expressed concern that the scope of the
concept paper (the single drug/test concept) may have
limited applicability. The more common industry practice is
for non-parallel drug and test development. Diagnostic test
development programs often do not begin until a validated
biomarker has been identified and its clinical potential has
been examined in early exploratory studies. As such,
industry indicated that more information was needed on
the FDA’s expectations on the co-development process to
assist firms in developing partnering strategies for the
products. For example, what timing and procedures would
be useful for interaction of the sponsor (or sponsors if the
diagnostic and drug sponsors are not the same) with CDER/
CBER and CDRH. Is it expected that the diagnostic
manufacturer will begin to participate in the FDA discus-
sions as early as the time of a voluntary genomic data
submissions (VGDS) meeting. It is not clear how the co-
development process will affect anticipated FDA/sponsor
interactions (with both CDER/CBER and CDRH) meetings,
agreements, labeling discussions and post-marketing
changes. Clarification is also needed about how joint review
and joint meetings will be managed throughout the co-
development process.
Industry also noted that the life cycle of a drug and
diagnostic test are different. Specifically, the life cycle of a
The Pharmacogenomics Journal
diagnostic test involves frequent improvements in the test
and platform, post-approval. Also, early development of a
partner diagnostic test (validated pre-phase 3) is considered
an expensive and high-risk endeavor in those cases when
the approvability of the partner drug is uncertain before
study in controlled pivotal trials. Industry indicated that it
would be useful if the future guidance clarified how life cycle
changes will impact the co-developed product’s label. In
addition, given the clinical utility expectations, industry
questioned whether it is possible to add general use
indications to the diagnostic test label. Similarly, the
comments suggested that the required steps in the review
process for follow-on tests should be defined. During the
discussion, it was noted that in general diagnostics are not
required to show clinical utility in relation to a drug to
support regulatory approval. It is not clear to industry why
the proposed analytical performance and validation expec-
tations for use with a drug would be different from those in
current CDRH guidance.
Some recommended consideration should be given as to
developing a decision tree to clarify when co-development is
necessary (when a test is required for co-approval) based on
efficacy and safety considerations.
B. Regulatory issues
Guidance is needed on how to bridge across technologies in
co-developing a diagnostic test assuming a qualified bio-
marker has been identified, and how this would affect
labeling. Interaction of the drug-test review process with
other FDA initiatives, including discussions under the
provisions of the guidance on Pharmacogenomic Data
Submissions, should be clarified. For example, what is the
role of advice obtained from the Interdisciplinary Pharma-
cogenomic Review Group and how will it be integrated into
co-development. Clarification is needed for which center or
review area would be responsible for the review and
approval of the pivotal study for a partner product.
From a regulatory perspective, industry inquired about
the possibility of developmental incentives (e.g., exclusivity
extensions based on co-development approval) Also, to
facilitate development, industry recommend more regula-
tory acceptance of retrospective data.
There was a general concern that the draft concept paper
may represent a higher hurdle for diagnostic approval than
current CDRH requirements. It was recommended that
future guidance clarifies the usefulness and distinctions
between other possible approval pathways for a test that
might be used in conjunction with a drug in a co-
development scenario (i.e., PMA vs CLIA vs IUO vs 510K
or de novo 510K). Clarification was requested on the
proposed pre-set ‘cutoff’ or ‘indeterminate zone’ before
pivotal phase 3 studies because in some cases these values
may represent a high hurdle for a diagnostic test. Addition-
ally, industry urged inclusion of labeling and cross-labeling
considerations including post-marketing changes in future
guidance.

Comments on the draft drug diagnostic co-develop-
ment concept paper (Docket No. 2004N-0279)
Following the April 2005 meeting, the FDA docket on this
topic remained open through the summer of 2005. Exam-
ples of docket comments from several individual companies
and trade associations are summarized below. Details on
these comments can be found by searching for the Docket
no. 2004N-0279 at the FDA Dockets Management Website:
http://www.fda.gov/ohrms/dockets/default/htm.
1. The presented model for co-development is generally unrepre-
sentative of reality and its scope is too limited: Several
industry organizations and firms noted that the model
for simultaneous co-development, whereas a useful
illustration, did not represent most development strate-
gies. Currently, most diagnostic development occurs in
late phase 2 to late phase 3. The proposed model may not
accommodate ongoing scientific advances that occur in
diagnostic development after the biomarker is established
and early co-development may present a significant risk
hurdle for device firms. Additionally, the differences in
development strategies between drugs and diagnostics
make the ‘single drug/single test’ concept a fleeting reality.
It should not be the case that this co-development
pathway obstructs adding an improved diagnostic to a
drug label. Industry requests for future guidance include
(1) information on the pathway when a diagnostic is
added late in the development of a drug, (2) the pathway
to add second-generation diagnostics to drug label devel-
oped for a co-development product and whether this will
vary from current regulatory procedures, (three) clarifica-
tion/consideration of appropriate least burdensome
approaches in these scenarios, (4) consideration of
pre-analytical assessments before committing to a com-
prehensive analytical validation study and (5) clarification
on whether a single test may include one or more analyte.
2. Requirements for establishing clinical utility are a high hurdle
for diagnostics: Several comments noted the dichotomy
between general diagnostic and drug diagnostic develop-
ment approaches. Currently, many diagnostics can be
approved for general use without establishing proof of
clinical utility with a particular drug product. Clinical
utility is not explicitly defined in the Act. It is seen as
critical that the proposed guidance does not impose
unnecessary regulatory burdens on the partner diagnostic
product. Additionally, the concept paper suggests that
two confirmatory trials might be needed to support the
co-development application. Although this is a normal
expectation for a therapeutic product, it is considered an
excessive request for the diagnostic partner. Industry
recommendations include (1) further clarification of
when co-development is considered necessary for mar-
keting, and (2) clarification of when the 510K, PMA or
analyte-specific reagent approaches may be appropriate.
3. Interaction with the VGDS process should be clarified:
Industry noted that the FDA has published guidance on
the voluntary genomic data submission process.8 The
The drug diagnostic co-development concept paper
LM Hinman et al
379
VGDS process is distinct from the regulatory submission
and review process. Industry recommendations include
clarification of the operational expectations for the
interplay of pre-IDE meetings with CDRH with early
interactions under VGDS or meetings with the FDA
review divisions need to be clarified.
4. Intercenter review process: Several comments noted that the
draft concept paper appears to be heavily directed toward
technical concerns from CDRH details for the diagnostic
test. However, the development strategy concerns (proof
of clinical utility, possibly needing two clinical trials for
the partner products, etc.) focus on CDER/CBER drug
development paradigms and may represent more strin-
gent requirements than normally expected for a diag-
nostic. Industry requests for future guidance include
(1) further clarification of the relationship of the diagnostic
test information, the drug clinical trial information, and
(2) how the center review issues will be considered during
the development, review and approval process. Also,
industry requests more practical guidance on the current
development strategies and not just the optimal process.
5. Some guidance should be given on labeling principles: During
the meeting, the FDA noted that labeling was not part of
the concept paper but would be addressed in future
guidance. Many industry comments continued to note
that labeling needs to be addressed and labeling should
not be a case-by-case matter. Some comments requested
clarification on how the FDA intercenter agreements on
labeling would apply. (Post Meeting Note – general
labeling recommendations for when and where the
pharmacogenomic information may be included in the
drug/biologics or device labeling were discussed at a
recent FDA advisory committee for pharmaceutical
science–clinical pharmacology subcommittee meeting.)7
6. Clinical study design and analysis considerations: Although
the concept paper indicates that retrospective pharma-
cogenomic analyses may be acceptable, several industry
comments request further clarification on the role and
acceptability of retrospective analysis. Also, more clar-
ification is requested on when and how a test negative
population should be considered. For example, it may be
more appropriately addressed on a case-by-case basis, as
there are too many variables to set a general standard.
7. Statistical issues: Several industry comments requested
additional clarification on coordinating the drug and
diagnostic studies, considerations on the study popula-
tions and the relationship of the statistical considerations
for the test vis-à-vis the statistical considerations
for the therapeutic drug. Also, clarification was requested
for the results that indicate a test is informative vs the
results that indicate it is predictive of a clinical drug
response.
8. Special protocol assessment (SPA) should be available for co-
developed products: Some comments noted that the SPA is
a highly successful method of establishing agreement
between the FDA and sponsor on the specifics of phase 3
trial design. If the phase 3 program for co-development
drug diagnostic partner products is to be reviewed by
The Pharmacogenomics Journal
 The drug diagnostic co-development concept paper
LM Hinman et al
380
both CDER/CBER and CDRH, a pathway to get binding
agreement from all parties would be highly desirable.
Next steps
Biomarkers and pharmacogenomic tests may represent
critical path milestones for innovative products Conceiva-
bly, co-development may lead to enhanced safety or
effectiveness of targeted therapies. As the FDA prepares its
draft guidance on drug/test co-development, it is antici-
pated that it will incorporate and consider the stakeholder
comments. Before the release of the draft guidance, the FDA
encourages industry to take advantage of the voluntary
genomic data submission process in order to gain access to
the current thinking and regulatory experience pertaining
to drug and test co-development. Also, the FDA encourages
the therapeutic and diagnostic developers to maintain
active intercenter dialogue with both the therapeutic and
diagnostics review divisions.
Acknowledgments
We thank Dr Larry Lesko for his thoughtful review and helpful
comments on this manuscript.
The Pharmacogenomics Journal
References
1 Salerno RA, Lesko LJ. Three years of promise, proposals and progress on
optimizing the medicines: a Commentary on the 3rd FDA-DIA-PWG-
PhRMA-BIO Pharmacogenomics Workshop. Pharmacogenomics J 2006; 1:
1–4.
2 FDA/DIA Workshop: Co development of drug, biologic, and device
products, July 2004, Arlington, VA, http://www.diahome.org/Content/
Events/04040.pdf.
3 Drug-diagnostic co-development – preliminary concept paper, FDA
publication, not for implementation: http://www.fda.gov/cder/geno-
mics/pharmacoconceptfn.pdf.
4 Huang S-M. Track 2: introduction: concept paper: drug-diagnostic co-
development – introduction: April 12, 2005, Bethesda, MD, http://
www.fda.gov/cder/offices/ocpb/workshops.htm.
5 Hinman L. Track 2 industry comments, April 12, 2005, Bethesda, MD,
http://www.fda.gov/cder/offices/ocpb/workshops.htm.
6 Huang S-M, Hinman LM, Webster C, Hackett J, Kock W, Love P et al.
Track 2 summary, April 13, 2005, Bethesda, MD, http://www.fda.gov/
cder/offices/ocpb/workshops.htm.
7 Labeling background http://www.fda.gov/ohrms/dockets/ac/05/brief-
ing/2005-4194B1_03_Topic-1A.pdf; background information for an
advisory committee for pharmaceutical science, clinical pharmacology
subcommittee meeting on November 14, 2005; http://www.fda.gov/
ohrms/dockets/ac/05/slides/2005-4194S1_Slide-Index.htm.
8 FDA Guidance for Industry: Pharmacogenomic Data Submissions, issued
March, 2005. http://www.fda.gov/cder/guidance/6400fnl.htm.