Clinical Trial Imaging Endpoint Process Standards Guidance for Industry

Administrative guidance, not enacted legislation. This captured publisher edition preserves the listed document and attachments; later developments are not independently consolidated.

Clinical Trial Imaging
Endpoint Process
Standards
Guidance for Industry
U.S. Department of Health and Human Services
Food and Drug Administration
Center for Drug Evaluation and Research (CDER)
Center for Biologics Evaluation and Research (CBER)
April 2018
Clinical/Medical

Clinical Trial Imaging
Endpoint Process
Standards
Guidance for Industry
Additional copies are available from:
Office of Communications, Division of Drug Information
Center for Drug Evaluation and Research
Food and Drug Administration
10001 New Hampshire Ave., Hillandale Bldg., 4th Floor
Silver Spring, MD 20993-0002
Phone: 855-543-3784 or 301-796-3400; Fax: 301-431-6353; Email: druginfo@fda.hhs.gov
https://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/default.htm
and/or
Office of Communication, Outreach, and Development
Center for Biologics Evaluation and Research
Food and Drug Administration
10903 New Hampshire Ave., Bldg. 71, rm. 3128
Silver Spring, MD 20993-0002
Phone: 800-835-4709 or 240-402-8010; Email: ocod@fda.hhs.gov
https://www.fda.gov/BiologicsBloodVaccines/GuidanceComplianceRegulatoryInformation/Guidances/default.htm
U.S. Department of Health and Human Services
Food and Drug Administration
Center for Drug Evaluation and Research (CDER)
Center for Biologics Evaluation and Research (CBER)
April 2018
Clinical/Medical

TABLE OF CONTENTS
I. 	INTRODUCTION............................................................................................................. 1
II. 	BACKGROUND ............................................................................................................... 2
III. 	LOGISTICAL AND TECHNICAL CONSIDERATIONS ........................................... 3
A. 	Choice of Imaging Modality .......................................................................................................... 3
B. 	Is Centralized Image Interpretation Important for an Imaging-Based Primary Endpoint? . 3
C. 	Should Image Interpretation Be Blinded to Clinical Data? ....................................................... 4
D. 	How Often Should Imaging Evaluations Be Performed?........................................................... 5
E. 	How Soon After Acquisition Should Images Be Interpreted? ................................................... 5
F. 	What Procedures Should Be Standardized for an Imaging-Based Clinical Trial Primary
Endpoint? ....................................................................................................................................... 5
IV. 	THE EXTENT OF IMAGING PROCESS STANDARDS ........................................... 7
A. 	Are Existing Medical Practice Imaging Process Standards Sufficient for the Trial’s Primary
Endpoint? ....................................................................................................................................... 7
B. 	What Should Be Considered When Augmenting Existing Medical Practice Imaging Process
Standards to Create Trial-Specific Imaging Process Standards? ............................................. 7
REFERENCES .............................................................................................................................. 9
APPENDIX A: BEFORE IMAGING: CHARTER CONSIDERATIONS ......................... 10
APPENDIX B: DURING IMAGING: MONITORING PLANS AND CHARTER
MODIFICATIONS ..................................................................................................................... 27
APPENDIX C: AFTER IMAGING: DATA TRANSFER, ARCHIVING, AND
ANALYSIS OF IMAGING INFORMATION ......................................................................... 28

Contains Nonbinding Recommendations
1
Clinical Trial Imaging Endpoint Process Standards
Guidance for Industry1
This guidance represents the current thinking of the Food and Drug Administration (FDA or Agency) on
this topic. It does not establish any rights for any person and is not binding on FDA or the public. You
can use an alternative approach if it satisfies the requirements of the applicable statutes and regulations.
To discuss an alternative approach, contact the FDA office responsible for this guidance as listed on the
title page.
I. 	INTRODUCTION
The purpose of this guidance is to assist sponsors in optimizing the quality of imaging data
obtained in clinical trials intended to support approval of drugs and biological products.2 This
guidance focuses on imaging acquisition, display, archiving, and interpretation process standards
that we regard as important when imaging is used to assess a trial’s primary endpoint or a
component of that endpoint.
Considerable standardization already exists in clinical imaging. There are a variety of sources,
including picture archiving and communication systems and the Digital Imaging and
Communications in Medicine (DICOM) formats for the handling and transmission of clinical
imaging information that describe the standards generally employed by clinical practitioners.
This guidance recommends additional imaging endpoint process standards that are more specific
to clinical trials. Imaging process standards help sponsors ensure that imaging data are obtained
in a manner that complies with a trial’s protocol, that the quality of imaging data is maintained
within and among clinical sites, and that a verifiable record of the imaging process is created.
Minimization of imaging process variability may importantly enhance a clinical trial’s ability to
detect drug treatment effects.
Standardization, while important for all clinically used measures, becomes essential for an
imaging endpoint used in a clinical trial to reduce variability and to ensure interpretability of the
results. The extent of trial-specific standardization may vary depending upon how standardized
the local imaging procedures are (e.g., routine bone X-rays (relatively standardized) versus bone
mineral density (more variability across sites)). This guidance does not address approaches for
1 This guidance has been prepared by the Division of Medical Imaging Products in the Center for Drug Evaluation
and Research in cooperation with the Center for Biologics Evaluation and Research at the Food and Drug
Administration.
2 For the purposes of this guidance, all references to drugs include both human drugs and biological products unless
otherwise specified.

Contains Nonbinding Recommendations
2
selecting an imaging endpoint for a clinical trial. An imaging endpoint, even if standardized and
considered to be precise and accurate, might not be interpretable if it does not have clinical
meaningfulness (e.g., hip fracture versus changes in local blood flow in the central nervous
system detected by positron emission tomography (PET) imaging).
This guidance does not address whether specific imaging measures would be acceptable in
submissions used to support approval of a drug or biologic. These considerations should be
discussed with the FDA review division responsible for drug development.
Many of the imaging process standard considerations for clinical trials of therapeutic drugs can
also be applied to clinical trials that evaluate the performance of diagnostic drugs. For
considerations involving the development of diagnostic imaging drugs, see the guidance for
industry Developing Medical Imaging Drug and Biological Products (Parts 1, 2, and 3).3
In general, FDA’s guidance documents do not establish legally enforceable responsibilities.
Instead, guidances describe the Agency’s current thinking on a topic and should be viewed only
as recommendations, unless specific regulatory or statutory requirements are cited. The use of
the word should in Agency guidances means that something is suggested or recommended, but
not required.
II. 	BACKGROUND
Almost all medical imaging processes involve some aspects of standardization established by
those within the practice of medicine. In medical practice, images of human anatomy and/or
physiology typically are acquired and interpreted, often with limited or no formal quantification,
by a single facility’s imaging professional staff. The images typically achieve the medical
practice’s diagnostic purposes even though the acquisition, display, and interpretation methods
may vary somewhat among imaging facilities and imaging professionals. This variability may
have little or no impact on the ability to provide a diagnosis in medical practice, yet in a clinical
trial, imaging process variability may result in increased variability in endpoint measurements
and may compromise the ability of the trial to achieve its objectives.
Although the medical practice of diagnostic imaging already follows many standardized
procedures, we recommend that some trials augment these existing standards to create trial-
specific imaging process standards. For purposes of this guidance, we define these trial-specific
imaging process standards as standards that extend beyond those typically performed in the
medical care of a patient (i.e., the process standards are implemented solely for the purposes of
the clinical trial). The extent of trial-specific imaging process standards can range from minimal
processes that are described solely in the clinical protocol, such as obtaining noncontrasted and
contrasted images in all subjects, to more detailed imaging process standards for image
acquisition, display, interpretation, and archiving that are detailed in an imaging charter (see
Appendixes A through C).
3 We update guidances periodically. To make sure you have the most recent version of a guidance, check the FDA
guidance web page at https://www.fda.gov/RegulatoryInformation/Guidances/default.htm.

Contains Nonbinding Recommendations
3
A clinical trial’s design and clinical context are critical determinants of the extent that imaging
process standards could be important to the trial. In a clinical trial where the imaging process and
imaging primary endpoint measures are not well standardized in medical practice, trial-specific
standards become important.
The following sections outline the imaging process topics that sponsors should address when an
imaging primary endpoint is used within a clinical trial intended to support the approval of a
proposed drug.4
III. 	LOGISTICAL AND TECHNICAL CONSIDERATIONS
The following subsections describe some of the factors a sponsor should assess when
considering the use of an imaging-based primary endpoint in a clinical trial intended to support
approval of a drug. The sponsor should contact the assigned FDA review division to discuss the
clinical meaningfulness of the primary endpoint imaging information to be obtained in the trial.
A. 	Choice of Imaging Modality
If an imaging-based primary endpoint is chosen for a phase 3 trial, the choice of the imaging
modality (such as echocardiography versus single photon emission computerized tomography)
may prove to be an especially important consideration. Imaging modality upgrades and
malfunctions are sometimes unpredictable. Clinical sites may also experience unforeseen
limitations on the use of the modality or modality-specific imaging drugs and processes, such as
the interchange of certain contrast agents that may not affect typical diagnostic imaging but may
alter trial-specific quantitative imaging measures.
B. 	Is Centralized Image Interpretation Important for an Imaging-Based
Primary Endpoint?
In clinical trials, images are interpreted either at the clinical site or at a centralized facility that
receives images from the clinical sites. Sometimes, both site and centralized imaging
interpretation may be performed. Image interpreters are sometimes referred to as readers, and
we use that term in this guidance.5
The usefulness of a centralized image interpretation process is determined by the role,
variability, and susceptibility to bias of imaging within the trial as well as modality-specific
image quality considerations and overall trial design features. Centralized image interpretation is
not always critical, even for a phase 3 trial primary endpoint that uses some aspects of
quantitative imaging, if the quantitative measures are widely performed and reported in clinical
medicine, little imaging acquisition or interpretation variability is anticipated, and potential
4 Evidentiary standards for diagnostic imaging products are addressed in parts 2 and 3 of the guidance for industry
Developing Medical Imaging Drug and Biological Products (Parts 1, 2, and 3).
5 Terms such as image interpretation, image review, or image read are used interchangeably in this guidance, and in
literature and protocols, image readers are sometimes referred to as image reviewers.

Contains Nonbinding Recommendations
4
biases in image interpretation are controlled by the trial design features. However, these
characteristics may not apply to some clinical trials, such as those that might be subject to bias
and cannot be blinded, those that use imaging modalities vulnerable to image quality problems,
and those that use specialized imaging measures.
In open-label clinical trials, availability of clinical information might influence a site-based
image interpretation because the expected relation of clinical features to outcome is known, and
therefore, a site-based image interpretation could raise concern about potential bias. A
centralized image interpretation process, fully blinded, may greatly enhance the credibility of
image assessments and better ensure consistency of image assessments. Some imaging
modalities also may prove vulnerable to site-specific image quality problems, and a centralized
image assessment process may help minimize these problems. For example, the National Lung
Screening Trial’s experience with computed tomography of the chest suggested that centralized
image quality monitoring was important to the reduction of imaging defects (Gierada, Garg, et
al. 2009). Hence, a centralized image interpretation process may be used to help control image
quality as well as to provide the actual imaging-based endpoint measurements.
Some image interpretation methods, such as trial-specific measures of bone or joint disease in
arthritic diseases, rely upon centralized imaging because of the extent of reader training
recommended for the specialized image interpretation.
As compared to site-based image interpretations in multicenter clinical trials, a centralized image
interpretation process may provide more verifiable and uniform reader training as well as
ongoing management of reader performance, helping to ensure quality control of the images and
the interpretations and to decrease variability in image interpretations, leading to a more precise
estimate of treatment effect. Nevertheless, the overarching trial design features and the other
previously described features may justify the use of site-based imaging interpretations even in
large phase 3 multicenter clinical trials, so long as blinding of image interpretation to treatment
can be assured or bias is otherwise controlled.
C. 	Should Image Interpretation Be Blinded to Clinical Data?
The extent of blinding image readers to clinical data depends upon the role of imaging in the
clinical trial, the specific disease and clinical setting, and the potential for unblinding effects
(e.g., toxicity) of the investigational drug. In a randomized controlled trial, we expect that
clinical trial primary endpoint image readers will be blinded to a subject’s treatment assignment
because knowing the assignment would be presumed to create bias. Further, we anticipate that
many, if not most, clinical trials using imaging-based primary endpoints will be conducted with
no reader knowledge of individual-level clinical data because this knowledge also may bias the
reader. In unique situations, a primary endpoint might rely upon integration of clinical data into
an image interpretation (Sargent, Rubinstein, et al. 2009).
To determine whether image readers should be blinded to clinical information, sponsors should
have knowledge of the underlying clinical condition, an understanding of the precedent for the
use of imaging as a trial’s primary endpoint, and detailed insight into the trial’s unique image
interpretation procedures (such as a plan for sequential locked-read image interpretation where

Contains Nonbinding Recommendations
5
an assessment cannot be altered versus an option for modification of prior image interpretations).
In certain disease conditions, readers also should be blinded to the image acquisition date and/or
knowledge of prior imaging observations. Again, we note that even if the image reader is aware
of individual-level clinical information, blinding to treatment assignment is almost always
critical.
D. 	How Often Should Imaging Evaluations Be Performed?
When a medical image serves as a trial’s primary endpoint, its timing and frequency of
ascertainment depends upon the underlying condition being studied, the feasibility of the
imaging schedule, and the overarching trial design features. For a trial using time point-based
imaging measures as a primary endpoint, the frequency of imaging evaluations should be the
same in all trial arms. Asymmetric imaging evaluation time points can introduce bias in the
treatment effect assessment. For a primary endpoint that uses a time-to-event analytical
approach, imaging evaluations should be performed at baseline and at sufficient frequency to
provide a reasonably precise measure of the time to the expected clinical event.
E. 	How Soon After Acquisition Should Images Be Interpreted?
In diagnostic medical imaging practice, images typically are interpreted on site within several
hours following acquisition. In contrast, in clinical trials using centralized imaging
interpretation, the interpretation may require a longer time frame. Therefore, image
interpretation timing typically is more of a consideration when clinical trials use centralized
imaging interpretation. When planning a clinical trial that uses an imaging primary endpoint, the
turnaround time by a central image interpretation facility should be appropriate for the
anticipated trial design. For example, prompt image interpretation may be an important
consideration for trials that use centralized image interpretations as components of interim
analyses, which may occur when imaging-based analyses are important to accommodate
prespecified sample size adjustment plans. Similarly, image interpretation expediency may
prove critical when centralized imaging interpretation is used to help control imaging quality; in
this situation, the centralized imaging readers or an appropriately prespecified centralized
imaging quality control process should promptly identify technical flaws that necessitate repeat
imaging of a subject. In other circumstances, interpretation of batches of randomized images at
specified intervals during a trial may be appropriate. Sponsors should consider the timeliness of
centralized image interpretation when developing a clinical trial protocol that uses an imaging-
based primary endpoint.
F. 	What Procedures Should Be Standardized for an Imaging-Based Clinical
Trial Primary Endpoint?
No single set of detailed imaging process standards is readily applicable to every clinical trial
because the trials differ in design and objectives. When usual medical practice imaging process
standards are acceptable in a trial, the plans for the use of such standards should be stated in the
clinical protocol. Determinations on what to standardize beyond these expectations should be
driven by consideration of the imaging processes that might introduce variability and inaccuracy
to the endpoint and by consideration of the other items outlined below. When determining the

Contains Nonbinding Recommendations
6
extent of imaging process standardization critical for a phase 3 clinical trial that uses an imaging-
based primary endpoint, sponsors should consider the following factors:
• 	Imaging modality availability and the modality’s technical performance variation across
trial sites
• 	Performance features of the imaging modality at the trial sites or any other locations
where subjects may undergo imaging
• 	Qualifications of the imaging technologists and any special technological needs for the
trial
• 	Proposed imaging measures’ reliance on phantoms and/or calibration standards to ensure
consistency and imaging quality control among clinical sites
• 	Any unique image acquisition features of the trial design, including subject positioning,
anatomical coverage of imaging, use of contrast, timing of imaging, importance of
subject sedation, and scanner settings for image acquisition
• 	Image quality control standards, including those specifying the need for repeat imaging to
obtain interpretable images
• 	Procedures for imaging display and interpretation, including technical variations in reader
display stations
• 	Nature of the primary endpoint image measurement, including the importance of training
image readers in trial-specific quantification methods
• 	Extent that image archiving could be important to the trial’s conduct, monitoring, and
data auditing
• 	Potential for imaging modality upgrades or modality failures, as well as the potential
variation in imaging drugs (such as contrast agents) across trial sites
• 	Precedent for use of the imaging-based primary endpoint measure in investigational drug
development, especially previously observed imaging methodological problems
If the existing medical imaging practice standards should be augmented to create trial-specific
imaging process standards and those trial-specific standards are too lengthy to be described in the
clinical protocol, then we encourage sponsors to develop an imaging charter that details the trial-
specific imaging process standards (see Appendix A). An imaging charter can consist of a single
document or a series of technical documents. Overall, the charter should describe how potential
sources of imaging bias and variability are controlled and how imaging process standards are
implemented to a level appropriate to the trial design.

Contains Nonbinding Recommendations
7
IV. 	THE EXTENT OF IMAGING PROCESS STANDARDS
The extent of imaging process standards in a phase 3 clinical trial runs on a continuum from the
standards already in place in imaging medical practice to trial-specific considerations of image
acquisition, display, transmission, interpretation, database development, and archiving.
Similarly, the effort associated with clinical trial imaging standardization may range from
relatively minor to challenging. At a minimum, the sponsor should consider the following
questions when developing primary endpoint imaging process standards for a clinical trial.
A. 	Are Existing Medical Practice Imaging Process Standards Sufficient for the
Trial’s Primary Endpoint?
To rely solely on existing medical practice imaging process standards to support an imaging-
based clinical trial primary endpoint, the sponsor should anticipate the variability inherent in
clinical site-based image acquisition, display, interpretation, and archiving. Consequently, for
sponsors to rely solely on existing imaging process standards, the trial’s imaging endpoint
measure or outcome should be readily apparent on an image that is acquired, displayed, and
interpreted using imaging processes that do not importantly vary among clinical sites. For
example, existing medical practice imaging process standards may be reasonable for an endpoint
focused upon the detection of a long bone fracture or intracranial hemorrhage — an indisputable
outcome that is readily determined with well-accepted and widely implemented imaging
methods that do not importantly vary among clinical sites.
With medical practice imaging process standards, the images might be interpreted and archived
at the clinical site such that the imaging outcome may be documented in the clinical trial records
solely by an investigator’s response on a case report form and/or appended clinical imaging
report. This type of image documentation typically is insufficient for a clinical trial endpoint that
relies upon quantitative imaging assessment unless the quantitative measure is widely accepted
as reliable and consistently reported to clinicians as a component of medical practice. The more
complex the image endpoint quantification process, the more likely the existing medical practice
imaging process standards will not be sufficient and will necessitate trial-specific standards that
minimize variability and document image endpoint measurements.
B. 	What Should Be Considered When Augmenting Existing Medical Practice
Imaging Process Standards to Create Trial-Specific Imaging Process
Standards?
We anticipate that most phase 3 clinical trials using an imaging-based primary endpoint will
benefit from some aspect of trial-specific imaging process standardization. In some situations,
this standardization may be confined to a clinical protocol’s brief statement about the nature of
the imaging to be performed in the trial and the frequency of the imaging evaluations. For
example, the protocol may include a statement that all subjects will be imaged both with and
without contrast enhancement. In other situations, an imaging charter should be developed to
standardize an array of imaging procedures among the clinical sites, such as the timing of
imaging during the trial, details of contrast injection, subject sedation and positioning, image
display and interpretation, and image archiving. Appendix A describes the various components

Contains Nonbinding Recommendations
8
of imaging process standards a sponsor should consider when augmenting medical practice
imaging processes. In comparison to medical practice imaging process standards alone, trial-
specific standards may provide better assurances that the imaging methods for the assessment of
a trial imaging endpoint are well defined and reliable.6
6 See 21 CFR 314.126(b)(6).

Contains Nonbinding Recommendations
9
REFERENCES
Dogan, S, Y Plantinga, J Crouse, et al., 2011, Algorithms to Measure Carotid Intima-Media
Thickness in Trials: A Comparison of Reproducibility, Rate of Progression and
Treatment Effect, J Hypertens, 29(11):2181–93.
Eisenhauer, E, P Therasse, J Bogaerts, et al., 2009, New Response Evaluation Criteria in Solid
Tumors: Revised RECIST Guideline (version 1.1), Eur J Cancer, 45(2):228–47.
Gierada, D, K Garg, H Nath, D Strollo, R Fagerstrom, and M Ford, 2009, CT Quality Assurance
in the Lung Screening Study Component of the National Lung Screening Trial:
Implications for Multicenter Imaging Trials, Am J Roentgenol, 193(2):419–24.
Keen, H, P Mease, C Bingham, J Giles, G Kaeley, and P Conaghan, 2011, Systematic Review of
MRI, Ultrasound, and Scintigraphy as Outcome Measures for Structural Pathology in
Interventional Therapeutic Studies of Knee Arthritis: Focus on Responsiveness, J
Rheumatol, 38(1):142–54.
Krupinski, E, K Berbaum, R Caldwell, K Schartz, M Madsen, and D Kramer, 2012, Do Long
Radiology Workdays Affect Nodule Detection in Dynamic CT Interpretation?, J Am Coll
Radiol, 9(3):191–98.
Sargent, D, L Rubinstein, L Schwartz, J Dancey, C Gatsonis, L Dodd, and L Shankar, 2009,
Validation of Novel Imaging Methodologies for Use as Cancer Clinical Trial End-points,
Eur J Cancer, 45(2):290–99.

Contains Nonbinding Recommendations
10
APPENDIX A:
BEFORE IMAGING: CHARTER CONSIDERATIONS
This appendix outlines some of the imaging methodology factors FDA recommends for sponsors
who may be considering using specific imaging standards for a particular clinical trial. We refer
to these augmented imaging process standards as trial-specific standards. Relatively
uncomplicated trial-specific standards can be described in the body of a clinical protocol. If the
trial-specific standards are too extensive for a clinical protocol, a separate imaging charter may
be more appropriate.
An imaging charter (hereafter, charter) can consist of either a single document or a series of
documents. The charter should describe the clinical trial imaging methodology, including
modality-specific technical details, image interpretation, and image archiving procedures.
Sponsors should develop the document or documents with the same rigorous standards typically
applied to the clinical protocol. Indeed, sponsors can choose to develop the charter as an
appended component of a clinical protocol. In this situation, the charter can be attached to a
clinical protocol as an appendix or cited as a supplementary document. The charter can also
consist of a freestanding, overarching summary of the imaging methodology with references to
multiple other imaging-specific documents that form a component of the charter, such as
imaging acquisition protocols, data transfer plans, or image submission guidelines. These
documents should provide detailed information on the methodology for acquiring images and for
transferring and archiving the images and the image interpretation data.
Generally, we do not regard the charter as part of the protocol unless the sponsor specifically
designates it as a component. We encourage sponsors to submit the charter for FDA review and
to do so as soon as possible. In the unusual situation where review of a charter is critical for
completing either a special protocol assessment or review of a trial’s clinical protocol, the review
division can specifically request submission of the charter along with the trial’s clinical protocol.
Submission of a charter for FDA review helps to support the plan for verification of the trial’s
data integrity because compliance with the charter may form an important aspect of the trial
conduct verification process as well as of the data quality assessment procedure.
When imaging forms an important part of a phase 3 clinical trial’s primary endpoint (or some
other important part of a trial), we encourage sponsors to briefly discuss at an end-of-phase 2
meeting whether or not imaging standardization procedures are appropriate.
There is no specific format or content recommended for a charter. When developing a charter,
sponsors should define the requirements for standardization based on the trial’s imaging
objectives and the sponsor-required imaging methods at the participating clinical sites.
Consequently, sponsors should specify key requirements for imaging equipment and image
quality, as well as the processes for image acquisition, display, interpretation, storage, and data
transfer.
Imaging technology evolves rapidly and can be highly technical. The complexity of technical
standardization may markedly limit the use of imaging in a multicenter clinical trial even if the
imaging methods have well-recognized value in clinical medicine (Keen, Mease, et al. 2011).

Contains Nonbinding Recommendations
11
Listed below are the suggested headings and subheadings for the elements within a charter.
Some of these elements may not apply to a particular clinical trial. We encourage sponsors to
address each of these elements within their charters in a manner most appropriate for the trial.
Executive Summary of the Trial Design and the Role of Imaging in the Trial
The charter should summarize the role of imaging within the clinical trial and provide a
description of the imaging database variables (deliverables) to be incorporated into the analysis
of the primary endpoint. It should address important trial design features detailed in the protocol
that might affect the proposed imaging database variables (e.g., procedures to minimize missing
data, plans for the use of off-protocol images).
The executive summary should also provide an overview of the major aspects of the image
acquisition, interpretation, and reader-defined deliverables. Presentation of a flow chart that
identifies the specific steps in the process can be especially useful in summarizing the flow of the
imaging information.
Standards for Image Acquisition
Development of image acquisition standards involves a broad knowledge of imaging modalities,
including knowledge of anticipated imaging equipment upgrades or malfunctions during the
conduct of the clinical trial. In some situations, exploratory clinical trials may be critical to
identify the most important imaging technical details, including those vulnerable to technical
factors and charter noncompliance. For example, an explicit description of the imaging
acquisition time may be critical when rapid dynamic cardiac arteriography is used to assess
coronary artery disease; in this situation, the X-ray energy (kVp) should be standardized and be
considered appropriate for imaging iodinated contrast agent within the heart. Similarly,
optimization of X-ray energy is essential for breast imaging because a high kVp will obscure the
signal intensity differences between adipose, glandular, or cancerous tissue, and variations in
kVp among clinical sites may increase variability in the imaging endpoint. In addition to
equipment settings, other imaging acquisition parameters critical to the trial imaging endpoint
(e.g., the number, angle, and magnification level of radiographic views in assessing arthritis)
should be standardized in the charter. The feasibility of maintaining technical consistency within
and among clinical sites is particularly important when choosing and optimizing the imaging
modality.
Equipment Standardization and Operation
The charter should typically identify the following:
• 	Vendor-specific equipment/platforms (e.g., injectors, scanners, ultrasound probes,
software). The charter should identify the use of any investigational equipment. For
sites in the United States, we recommend the use of imaging equipment that has received
FDA marketing authorization. Use of investigational devices is subject to the
requirements of 21 CFR part 812 governing investigational device exemptions.

Contains Nonbinding Recommendations
12
The charter should specify the important imaging equipment for the trial, including the
imaging drug (contrast) injectors, scanners, ultrasound probe selection/settings, and
software. This is particularly important for certain functional and quantitative imaging
tasks that benefit from the use of specific scanners and models that are able to perform
the imaging and allow the control that ensures the standardization. The importance of the
equipment specifications varies with the role of imaging in the trial and may limit the
number of qualifying clinical sites. For example, imaging scanners may differ in
technical details that can influence image quality, such as image reconstruction software
programs and techniques for respiratory and cardiac gating, subject positioning, scan
times, probe positioning, and technician-dependent procedures.
In trials using quantitative imaging procedures, an ongoing site qualification process may
be important to maintain control of changes in imaging hardware and software. In other
trials (e.g., large multinational trials using standard computed tomography or magnetic
resonance imaging modalities), the ongoing control of hardware and software versions
across all sites may be inapplicable or impractical. The charter should provide a
justification for the approach proposed by the sponsor. The sponsor can provide routine
extensive technical details in companion manuals (e.g., imaging acquisition and
processing manuals) rather than in the charter.
In situations where minimizing variability of imaging data is critical, we encourage the
use of a tabular listing of the acceptable imaging equipment, including the key
characteristics of the acquisition, processing, and display components of each scanner or
review workstation. Another approach could be to identify the physical benchmarks and
testing parameters that should be met by the imaging equipment in accordance with a
prespecified protocol for the acquired images to be used in the trial. When developing
these specifications, sponsors are encouraged to perform exploratory analyses of imaging
outcomes grouped by imaging platforms.
Most three-dimensional imaging technology relies upon raw data processing using
proprietary software algorithms. Software upgrades within the trial period may affect
how images are generated. Changes in an image may be caused by these software
changes and be incorrectly attributed to actual clinical changes. The charter should
specify minimum requirements for important software and also identify any situations
when alternatives are acceptable.
Occasionally, requisite imaging equipment may become unavailable at a qualified site
because of equipment malfunction or unavailability of technical support. In these
situations, a clinical site might choose to substitute one imaging modality for another
(such as magnetic resonance for computed tomography). The charter should identify the
situations when these changes are acceptable. We anticipate that, in many situations,
modalities will not prove interchangeable (such as arteriography for ultrasound) when the
endpoint assessment involves a quantitative imaging measurement. Ad hoc, unplanned
interchange of modalities (including substitution of film for digitized imaging data) may

Contains Nonbinding Recommendations
13
compromise trial objectives if changes occur for a given subject during the course of the
clinical trial.
• 	Equipment technical settings to be used at each site. The charter should summarize
the technical settings for image acquisition for each type of important imaging equipment
and should identify any acceptable deviations from these settings. We encourage
sponsors to identify these settings based upon the findings from exploratory clinical trials
or other trials that attempted to standardize the technology among multiple clinical sites.
Typically, such information is included in site imaging manuals. Details critical to
quantitative imaging, such as tomographic slice thickness, pulse sequence, and contrast
agent injection time (especially for dynamic imaging), should be specified in the charter.
• 	Role of site imaging technicians in equipment operation, including the identification
of faulty or unacceptable images and the importance of repeating imaging. The
charter should describe the role of the imaging technician in the image acquisition
process, including the recommended qualifications and the role of the technician, if any,
in the initial assessment of image quality. Situations should be identified when repeat
imaging is critical (and exposure to additional radiation dose is justified) because
technical failure has made the images uninterpretable. In some situations, such as
ultrasound imaging, detailed procedures should describe the technician’s role in
manipulation of the imaging probe. Depending upon the imaging modality and the
technical demands, the charter can describe or reference the trial documents describing a
technician training process that will help ensure consistency in image acquisition. In this
situation, the charter should specify that administrators of trial sites will document
modality-specific training standards and maintain records that show the technologists
participating in the trial have met all training requirements specified by the charter.
• 	Phantoms to be used for site qualification and image quality monitoring. In many
situations, the use of phantoms (i.e., prespecified objects for scanning) is a critical part of
site qualification and image quality monitoring during the conduct of a clinical trial. In
other situations, phantoms may not be important if the equipment and imaging
acquisition parameters are well standardized. Phantoms can simulate a variety of
conditions and have been developed for a range of imaging modalities (e.g., magnetic
resonance, nuclear medicine, radiography). The choice of the specific phantom type
depends upon the imaging objectives as well as upon the specific imaging modality.
Standardization of image acquisition using imaging and dosimetry phantoms often
enhances the consistent performance of the imaging equipment during the course of the
trial.
• 	Subject preparation, positioning, and comfort measures. Many imaging modalities
require specific subject preparation (e.g., fasting, special dietary limitations), positioning
(e.g., supine, right lateral decubitus, weight-bearing status for lower extremity
radiographs), preparation (e.g., removal of jewelry and eyeglasses), and comfort
measures (e.g., ear plugs, sedation). These common aspects of imaging could vary
markedly among clinical sites. The charter’s specifications for these items may prove
especially useful because significant site-to-site variations in subject preparation can

Contains Nonbinding Recommendations
14
result in unacceptable levels of image variability. Subject preparation also might be
based on subject-specific factors, such as age, weight, and physical condition; the
importance of standardization of these aspects may widely vary. For example, a trial
conducted among pediatric subjects may call for some form of sedation. A description of
the acceptable sedatives (including doses, route of administration, and potential for repeat
dosing) may prove essential to quality imaging as well as to the avoidance of missing
images.
• 	Schedule for imaging and alternatives. Typically, a trial’s clinical protocol would
outline the schedule for imaging evaluations; hence, the charter might not include the
schedule. In other situations, the clinical trial protocol might include only a superficial
description of the imaging schedule, so the charter should include more details about
imaging times. For example, in certain circumstances, subjects should be imaged at a
specific time of day or night or following the development of certain clinical features
(such as pain in a joint) that prompt imaging-specific procedures (such as region-of-
interest imaging). The charter should describe these expectations and also identify date
and time windows that represent acceptable alternatives to the planned imaging
evaluations.
• 	Off-protocol imaging. Subjects in a clinical trial lasting many months are likely to
undergo imaging examinations in addition to the ones intended to assess the response to
therapy or to detect disease progression. These off-protocol images may or may not be
made available to image readers and considered as part of the trial’s imaging-based
outcomes. The clinical protocol may address some of the off-protocol imaging plans, but
we recommend that the charter elaborate upon these plans to clarify whether and in what
way the off-protocol image interpretation would contribute to the trial’s endpoint
assessment.
• 	Imaging risks. Imaging may involve risks to subjects, such as exposure to radiation and
contrast agents. These risks should be described in the trial’s clinical protocol so that
they can be considered by institutional review boards and appropriately described within
consent documents (see 21 CFR parts 50 and 56). Therefore, we anticipate that most
charters will not contain a section that describes imaging risks.
Occasionally, clinical trial imaging detects incidental findings that may be important for
further clinical evaluation. Incidental image findings may result in health benefits as well
as risks for subjects. For example, some incidental findings may needlessly prompt
invasive diagnostic evaluations. If the detection of incidental findings is likely with
clinical trial imaging, the clinical protocol and consent documents should describe the
procedures for handling the incidental findings. The charter also should summarize how
these incidental findings will be handled based upon the description within the clinical
protocol. A description of these plans within the clinical protocol is important to ensure
that the process is reviewed by institutional review boards and, as necessary, summarized
within trial consent documents (see 21 CFR parts 50 and 56).

Contains Nonbinding Recommendations
15
• 	Site qualification process. The charter should address the process used to qualify
clinical sites for trial participation, specifically describing and/or referencing the tests to
be performed to verify equipment performance, technical support, and capability for
compliance with charter expectations. We anticipate that phantom imaging, on-site
inspection, and/or training will provide sufficient site qualification for many trials. In
some situations, the site qualification process should build upon these expectations by
imaging subjects as part of a qualifying clinical trial. These types of site qualifications
can be particularly important for highly technical imaging modalities or international
trials that include countries where the imaging technology might be uncommon in clinical
practice.
• 	Acquisition quality control monitoring process. The charter should describe the plan
for periodic, quality control monitoring of imaging acquisition, storage, and transfer,
including the plan for repetitive phantom imaging and the correction of deviations from
the quality expectations. The importance and nature of this type of monitoring vary,
depending upon the nature of the imaging technology, but at a minimum, should involve
some form of episodic imaging quality reporting from clinical sites. In instances where
imaging is particularly complex or novel, we recommend periodic on-site inspection by
the trial’s imaging-specific monitors to assess the imaging technical compliance of each
clinical site or a subset of all the sites. Situations should be identified in which sites will
be requalified or terminated because of failure to comply with image quality
expectations. All requalification procedures should be described.
• 	Data storage and transfer. The charter should describe the expectations for imaging
data storage and transfer to any separate facility from the imaging site (e.g., centralized
laboratory, the sponsor). In general, the charter should do the following:
‒ 	Specify the storage of imaging data at the clinical site
‒ 	Describe any and all plans for transfer and storage of imaging data outside the clinical
site
‒ 	Describe any image alteration procedures to be performed at the site (such as removal
of all subject-identifying information (i.e., anonymization))
‒ 	Specify the time period for storage of images at clinical sites and the format for data
storage
Imaging Drug Standardization
Drugs are commonly used as a component of imaging and often require administration
procedures related to the scanning of a subject. Most notable drugs used as components of
imaging include the following:
• 	Preparative drugs
• 	Contrast agents

Contains Nonbinding Recommendations
16
• 	Radiopharmaceutical agents
Depending upon the nature of the imaging evaluation, the charter should identify the important
aspects of drug selection, dosage, and administration for each of these drugs, as exemplified
below.
• 	Preparative drugs. In situations where preparative (or other) drugs are important for the
planned imaging evaluations, the charter should identify acceptable and/or requisite pre-
imaging drugs, including sedatives, stimulants, beta-blockers, vasodilators, intravenous
fluids, or contrast agents. The drugs should be identified by their names and by dosages
and routes of administration. These specifications can be particularly important for trials
in which pediatric subjects are enrolled and for the imaging of subjects following
administration of drugs that will affect images (such as drugs essential for cardiac stress
testing).
• 	Contrast agents. Many modality-specific contrast agents are not interchangeable and
differ importantly in doses, techniques for administration, and risks. If critical to the
imaging evaluation, the charter should identify acceptable and/or requisite contrast
agents, including specific names. The charter should also identify the doses, routes of
administration, rates of administration, and any special administration procedures (such
as automatic injectors or administration times that may trigger scanning).
Some contrast agents can be safely administered only to subjects with acceptable renal
function or other characteristics. The charter should identify any laboratory tests and
outcomes critical for supporting the administration of contrast agents. Risks associated
with imaging, including those associated with contrast agents, are best described in the
clinical protocol.
• 	Radiopharmaceutical agents. In addition to specification of the administered activity,
mass, and route of administration, the charter should briefly identify the major drug
quality features for any clinical trial radiopharmaceutical manufactured at clinical sites.
Unlike preparative drugs and contrast agents, some radiopharmaceuticals (e.g., PET
agents) are commonly produced at clinical sites, and the quality of these drugs may vary
from site to site. Standardization of these drug attributes may be important in achieving
the trial’s imaging objectives. The charter should identify any site-specific production
considerations for site qualification.
Standards for Image Interpretation
Image interpretation generally is carried out by trained readers, such as qualified radiology,
nuclear medicine, and/or clinical specialists, who review and interpret, or read, images obtained
in the course of a clinical trial.
The following elements pertain predominantly to the use of a centralized facility for image
interpretation in a clinical trial. Whether images are interpreted solely at the clinical site or at

Contains Nonbinding Recommendations
17
both the clinical site and a centralized facility, we regard these elements as important aspects to
address within the charter when a centralized facility is used.
Image Transfer, Receipt Documentation, and Initial Quality Assessment
The charter should identify the process for transfer of imaging data from each clinical site to the
centralized image interpretation facility, including the plans for the following:
• 	Verification of the image technical adequacy
• 	Transfer of images and supportive information to the centralized facility
• 	The centralized facility process for querying sites for missing images, data, or imaging
technical problems
• 	Obtaining repeat images of subjects
• 	Logging of images received at the centralized facility, including the subject-specific
tracking system
• 	Formatting for image data transfer
• 	Digitization of received images or data
• 	Any technical evaluation (or pre-interpretation) or alteration of images, including de-
identification of subject information, biasing marks, or other undesired image signals
• 	Monitoring compliance with the transfer, receipt, and initial image assessment process
• 	Correction of deficiencies and failures in the transfer, receipt, or initial image assessment
process
The process should be highlighted for removal of all subject-identifying information from
images relayed through electronic communication (e.g., Internet, laptop computers) or other
pathways that are vulnerable to a security breach (e.g., courier or postal transfer of hard copy
images or digital images on disk).
Image Display and Interpretation
The paradigm shift from film-based to filmless imaging has redefined clinicians’ processes of
image display, and interpretation of images within a clinical trial may critically depend on the
quality of the displayed image. Image display in many digital systems is a flexible and dynamic
process whereby radiologists directly interact with the soft-copy image, which is displayed on a
computer workstation. The hardware component of a display system usually is composed of a
display device and a display driver or graphics card. The specifications given for a system are
valid only for that particular combination of devices. Another important aspect of the display

Contains Nonbinding Recommendations
18
system is the hardware and software components used for maintaining the display presentation
mapping between image values and luminance levels under a desired calibration model.
Information regarding the calibration hardware, software, and procedures, including frequency
and nature of the performed tests, should be identified in the charter and referenced as
appropriate to a standard operation manual.
• 	Selection of images for interpretation, display sequence, and randomization. The
charter should identify the nature and extent of images to be interpreted (e.g., all
scheduled images, off-protocol images) and any important sequence aspects (e.g.,
baseline images followed by subsequent time point images). The appropriateness of
excluding images or portions of images from the interpretation (read) process should be
emphasized and justified. The charter should prespecify the following:
‒ 	Criteria for classifying an image as uninterpretable based on a technical failure or
other considerations that lead to the exclusion of an image from the interpretation
process
‒ 	Qualifications of an individual or individuals who are to make the determination of
whether an image is included or excluded in the reading queue
‒ 	Determination of and criteria needed to determine whether image readers can also
decide if an image is uninterpretable when an individual or individuals other than the
actual image readers have the responsibility of excluding certain images from the
interpretation process
‒ 	Criteria for how uninterpretable or missing imaging data will be accounted for
(imputation scheme) in the data analyses
‒ 	Potential for reader interpretation drift (i.e., deviation from study-specific image
interpretation criteria on which readers have been trained) if images are assessed on
an ongoing basis for a trial that includes multiple images obtained over time
The process of randomized image presentation is often a key component of the overall
image presentation plan. If images (or image sets for a subject at any specific time point)
are to be randomized for display to readers, the charter should describe the randomization
process. For each type of image presentation, the charter should describe the data locks
(i.e., procedures that prevent modification of a reader’s final interpretation) to be used at
the subject image set level. These locks generally are critical to evaluating the
contribution of each image examination to the overall assessment of a subject’s image
set. The following are some examples of a trial’s image presentation process:
‒ 	In a time-sequential presentation, a subject’s complete image set (from baseline
through the follow-up evaluations) is shown in the order in which the images were
obtained. In this process (unless prespecified and justified in the charter), the reader
does not initially know the total number of time points in each subject’s image set.

Contains Nonbinding Recommendations
19
‒ 	In the simultaneous image presentation, a subject’s complete image set is displayed
(there is no blinding of date, sequence, or total number of images).
‒ 	In the simultaneous, randomized temporal image presentation, a subject’s complete
image set is shown at the same time in a random order with respect to the date (there
is no blinding to total number of images).
‒ 	In the simultaneous time point presentation, a subject’s single time point image set is
randomized among many other subjects’ image sets.
‒ 	In a hybrid, randomized image presentation, a subject’s complete image set (or only
the postbaseline images) is shown fully randomized. After the read results have been
locked for each time point, the images are shown again in known chronological order
for reread. Changes in any of the randomized assessments are tracked and
highlighted in the final assessment. In within-subject-control trials (e.g., comparative
imaging), images obtained before and after use of the investigational drug should be
presented in fully randomized unpaired fashion and in randomized paired fashion in
two separate image evaluations. The minimum number of images in each
randomized block necessary to minimize recall should be considered.
• 	Readers and their background qualifications. When developing the charter, sponsors
should identify the number of image readers and their requisite background
qualifications. The charter developers should consider the following:
‒ 	The extent of technical knowledge essential to image interpretation.
‒ 	The avoidance of any other reader involvement in the clinical trial (e.g., participation
as an investigator) that might bias the interpretation of the images.1
‒ 	The importance of confidentiality of image reads and/or the reading process.
‒ 	The potential for reader fatigue and the use of substitute readers. Decline in the
quality of image interpretation caused by reader fatigue has been reported in clinical
medicine (Krupinski, Berbaum, et al. 2012).2
1 Sponsors should also determine whether image readers should be considered to be clinical investigators. Under
the applicable regulations (21 CFR parts 54, 312, 314, 320, 330, 601, 807, 812, 814, and 860), a sponsor is required
to submit to the FDA a list of clinical investigators who conducted covered clinical trials and to certify and/or
disclose certain financial arrangements. Additional information is available in the guidance for clinical
investigators, industry, and FDA staff Financial Disclosure by Clinical Investigators. We update guidances
periodically. To make sure you have the most recent version of a guidance, check the FDA guidance web page at
https://www.fda.gov/RegulatoryInformation/Guidances/default.htm.
2 Similarly, in clinical trials, reader performance may be compromised if the readers are assigned an excessive
number of images to examine within a short period of time. The charter developers should consider whether to
specify a maximum number of images to be read in a session if reader fatigue is an important consideration for
image interpretation.

Contains Nonbinding Recommendations
20
‒ 	The time commitment of readers and reader availability.
‒ 	Any unique considerations for identification of a reader that adjudicates discordance
in image interpretation between readers and that determines the final assessment.
‒ 	The importance, if any, of clinical readers (i.e., image interpretation by clinicians
aware of nonimaging clinical trial information).
• 	Reader training and qualification. The reader training process should be described,
emphasizing the training documentation process and the use of any specific training
materials (e.g., a training manual, training images), image display training sessions, and
image read testing process. The origin (e.g., other clinical trials) of training images
should be described. In addition, the charter should prespecify whether any performance
criteria will be used to qualify readers after training and during the course of the trial.
Reader training manuals are key documents that contain more details of the reader
training procedures and should be provided for FDA review.
Sponsors should consider the importance of the following items in the development of
the reader training process:
‒ 	An overview of the major goals of the image interpretation. In general, reader
training should emphasize only the image-specific aspects of the image interpretation
process unless the process also integrates the use of clinical information. The process
should also minimize the potential for introduction of bias into image interpretation
through knowledge of any potential image signatures that may break the desired
blind-to-treatment assignment (e.g., if a PET ligand uptake is more common among
the elderly, the coregistration of PET-computed tomography may bias the PET
assessment because of recognition of aging-related cerebral atrophy on the
tomogram).
‒ 	An overview of the major expectations for image manipulation, lesion measurement,
and other image evaluations. Readers may benefit from special training in computer-
assisted interpretation, measurement, or other analysis tools as well as in the process
for performing and recording measurements, especially if this process involves
unique software data lock features and password-protected features. The reading
process may assume knowledge of unique assessment tools, such as Response
Evaluation Criteria in Solid Tumors (RECIST) outcome expectations (Eisenhauer,
Therasse, et al. 2009). The charter should describe these expectations in detail and
address situations when images may not be conducive to the requisite lesion
measurement or other tool expectations.
‒ 	Identification of any unique read definitions and/or criteria, including the use of
image case report forms. Some clinical trials may benefit from predefined criteria for
reads (e.g., identification of the specific basis for an unreadable image), and these
criteria may differ from commonly used clinical criteria. Training and verification of

Contains Nonbinding Recommendations
21
training (with mock image reads) may be important in documenting reader
proficiency.
‒ 	Description of any reader retraining procedures. Some image interpretation
processes may include the use of test images intermixed among the clinical trial
images such that readers are intermittently tested as to the proficiency and/or
consistency in their reads. Failure to sustain proficiency may result in replacement of
a reader with another trained and qualified reader. The charter should describe the
reader testing and retraining or replacement procedure.
• 	Timing of image reads and the read process. The charter should describe the timing of
image reads with respect to the clinical trial conduct. In some situations, prompt
interpretation of images is important (e.g., for determining trial eligibility, confirming
disease progression in trial subjects). In other situations, images are interpreted only
following completion of all subject evaluations. Perhaps most commonly, readers can
interpret images in batches periodically during the trial. If readers interpret images in
batches, the size of the batches should be specified, and the batch size should be justified
to minimize recall bias. The allowable time interval between the batch sessions also
should be predefined.
The charter should provide a detailed description of the image review process. We
recommend that the following be identified:
‒ 	The review setting.
‒ 	Whether readers interpret images independent of any other individuals. Other
individuals who may be present during the read should be specified and their roles in
image interpretation described. Any consensus read process should be detailed.
‒ 	A description of any image adjudication process.
‒ 	A detailed description of the use of any clinical information in the read.
‒ 	A description of the read outcome information, to be included in case report forms,
and any special procedures in this process (e.g., an initial read followed by a redisplay
of images to form a global reassessment).
‒ 	The assessment tools to be used and qualitative and/or quantitative measurements to
be performed during the image read (e.g., modified RECIST criteria assessment of
each image set).
‒ 	A description of any computer software or other electronic processes involved in
image interpretation, such as an automatic calculation of progression.

Contains Nonbinding Recommendations
22
‒ 	Any lesion tracking system (e.g., certain requisite target lesions), particularly any
nuances related to the appearance of new lesions for tracking, or inability to identify
any previously tracked lesions (e.g., imaging problems, lesion resolution).
‒ 	The options and/or requirements for image manipulation, including application of
calipers, zoom, pan, adjustment of window/level, contrast inversion, and application
of image enhancement features.
‒ 	A description of any process for reread of images. For example, a reader may
experience a sudden illness that results in an incomplete image interpretation. A
prespecified plan for the rereading of incompletely read images helps to verify the
integrity of the read process.
‒ 	The reader’s role in citation of missing images or technical deficiencies within the
images.
‒ 	A description of the plan to ensure that all original read outcome information is
locked and available for subsequent verification and comparison to any reread
outcomes.
When developing the image display process, sponsors should consider, as appropriate for
the chosen modality, the key performance characteristics of medical displays such as
luminance range; contrast ratio; reflection coefficients; grayscale; spatial, temporal (for
image stacks), and color resolution; and spatial and temporal noise. The charter should
specify these details as well as other modality-specific items, such as the process for
displaying dynamic images in relation to static images and any software manipulation of
images for the minimization of degradations that may occur along the imaging process or
transfer chain.
Computer-assisted image interpretation may form an important component of the read
process. In general, the extent of computer assistance should be described explicitly
within the charter, including a plan for quality control checks upon any critical software
functions. For example, the image interpretation may be driven primarily by a reader
who uses a computer-assisted analysis tool to complement the reader’s initial assessment.
Such reliance on computer assistance can be algorithmic (with prespecified parameters
for the use of a tool) or can be elective. In either case, such use should be defined within
the charter in a manner that results in a sufficient audit trail and assessment of the roles of
reader and reading tool. To evaluate for systematic errors, we suggest that a subset of
computer-generated analyses be verified by blinded external readers.
If interpretation tools are to be used, the charter should specify the use of computer-
assisted interpretation tools that FDA has authorized for marketing. Alternatively, an
investigational interpretation tool justified for use with a given imaging modality can be
used in some situations if the tool is compliant with all applicable FDA regulations,
including the investigational device exemption requirements under 21 CFR part 812.
Charter developers should review information regarding the software development

Contains Nonbinding Recommendations
23
process and testing for investigational interpretation tools. The same computer-assisted
interpretation tool should be available to all readers at a site or centralized facility.
• 	Imaging case report forms. We anticipate that specific imaging interpretation case
report forms could be important for many clinical trials, particularly trials that involve
quantitative imaging within endpoint construction. The charter should briefly describe
the content of the case report form and emphasize the specific data content or notations
that will be subsequently transferred to the sponsor to form the imaging database for the
trial’s endpoint analyses. We encourage the attachment of a case report form example to
the charter. On this case report form, sponsors should denote the specific items to be
transferred to the sponsor to form the imaging analytical database. In some situations, the
case report form may consist of a tabular display of numbers (such as lesion
measurements) or categories (such as predefined categories of bone erosion). An
example of the tabular display within the charter may help lessen the potential for errors
during the imaging flow process.
• 	Imaging data lock process. At a predetermined point during the image review process,
the image interpretation data (case report form information and any other important
reader notations, including notations on images) generated by the readers should be
locked. Locking data means that no further modification of image assessment is allowed.
The data locking process and timing should be closely linked with the image read
process. Data can be automatically locked by the imaging display equipment or triggered
in response to reader notations. In some situations, the reading process may include a
reread of previously interpreted images, including access to locked data. In all situations,
the charter should describe the locking of data and any potential rereads.
We encourage the use of an image lock approach to the read process whereby readers
interpret the assigned image (or image set) and lock their reads (e.g., lesion
measurements, response category, lesion severity) such that the contribution of each
image read to the read outcome in each image set is documented and not altered.
• 	Quality control of the image display and interpretation process. The charter should
describe the process for monitoring compliance with the image display and interpretation
process. This monitoring should include technical assessment of equipment, such as
display systems and data locking software, as well as the reader interpretation process.
Digital test patterns for quality control purposes can be used on a daily basis to ensure
consistent performance and to detect changes in the hardware or software that can
degrade image quality. In some instances, automatic luminance corrections might
compensate for the reduction in luminance that is expected over time. Some of these
quality control approaches offer the convenience of centralized reporting that facilitates
the comparison of different display systems used in a given trial. In some circumstances,
these automatic adjustment features may actually complicate measurements if the
features are unaccounted for. In either case, information about such automatic
compensation should be known and accounted for.

Contains Nonbinding Recommendations
24
In some clinical trials, evaluation of reader interpretation performance will be of value to
help assess the extent to which readers are consistent in image interpretation and comply
with the trial specifications. We recommend evaluating intra- and inter-reader
performance with defined and prespecified metrics based upon evaluations that are
ongoing during the image interpretation process.
In many situations, intra-reader variability as a measure of reader performance should be
assessed by periodic blinded testing of the reader with a preselected or predefined set of
images interspersed with the clinical trial images. These reader testing images can come
from a source external to the trial or, with proper data locking methods, from
reinterpretation of selected clinical trial images. If clinical trial images will be used as
reader testing images, then prespecified methodology should ensure that the original
image interpretation is considered final and, for intra-reader variability, that a sufficient
period occurred between the reading and rereading of the images to minimize recall bias.
A change in reader performance is not infrequently observed in clinical trials, and
depending upon the role of imaging in the trial, periodic reader retraining and
requalification may be critical. All details of reader testing, retraining, requalification,
and possible replacement should be prespecified within the charter and/or supporting trial
documents.
Image interpretation is inherently subjective, and readers of the same image may
justifiably disagree in their interpretations of the image. In these situations, a strategy for
reconciliation of discordant image interpretation to provide a final image outcome should
be detailed in the charter.
In some situations, the frequency of the reader adjudication may provide a sense of the
inherent subjectivity within the image outcome. Knowledge of the adjudication rate may
be important for interpreting the results. When developing clinical trial protocols and the
charter/supporting documents, the trial developers should consider the potential effect of
reader interpretation variability upon the clinical trial outcomes, and if the image
adjudication rate is regarded as an important consideration, then the trial documents
(protocol and statistical analysis plan) should prespecify this rate determination process.
Charter Modifications Before Imaging
The charter should briefly describe the process for modifying the charter in response to potential
deficiencies within the imaging process or in response to a need to improve the process.
Sponsors should describe the plan for submitting charter modifications to the FDA and other
regulatory authorities. In general, we anticipate charter revisions to be uncommon, particularly if
imaging has been used in exploratory clinical trials and the imaging processes follow precedents.
To assess the sufficiency of the imaging plans for a phase 3 trial, sponsors can conduct pilot
reading studies to test the reading and image lock process, the report form, and other important
aspects of imaging. If so, these pilot studies should not involve images or data that will be used
in the phase 3 trial.

Contains Nonbinding Recommendations
25
Imaging Data Transfer Process to the Sponsor
Image interpretation should result in the completion of a case report form and/or tabular display
of numbers, measures, or categories of responses. The charter should describe the process for
transfer of this information to the sponsor and the time point or points for transmission of this
information. The charter should describe how the sponsor will use the transferred information to
establish the variables used in the analysis of the primary endpoint.
Archiving of Images and Image Interpretations
Images should be archived as a usual component of subject care as well as for use as the source
documentation in clinical trials. Electronic source data should meet the same elements of data
quality that are expected of paper records and should comply with all applicable statutory and
regulatory requirements. The FDA’s acceptance of data from clinical trials for decision-making
purposes relies upon verification of the quality and integrity of data, generally based upon the
findings from audits and inspections.3 In addition to images themselves, the image
interpretations (case report forms or assessment tabulations) represent source data and should be
retained for potential inspection and auditing. All required source records, whether electronic or
paper, must be retained for a period of no less than 2 years following approval of a marketing
application or discontinuation of shipment and delivery of the drug for investigational use, as
described in 21 CFR 312.57(c) and 21 CFR 312.62(c).
The charter should describe the process for archiving imaging information by the site
investigator as well as by the sponsor. In some situations, the sponsor may choose to archive the
imaging at a centralized contractual facility or institution. Regardless of the physical storage
route, the archiving process should address the following items:
• 	Limiting access to ensure images and data are retained in their original forms
• 	Providing back-up storage
• 	Archiving in a manner conducive to a clear audit trail, including date and time stamps
Additional information regarding systems and personnel controls for computerized source data is
described in the guidance for industry Computerized Systems Used in Clinical Investigations.
Verification of Fidelity of Charter Documents With the Clinical Protocol and Statistical
Analysis Plan
Because the charter may consist of an ensemble of technical documents, the developers of the
charter should include a final step in which all the documents are reviewed to ensure that the
charter’s technical specifications do not contradict or modify the protocol-specified imaging
endpoints. Many individuals and organizations may be involved in the development of a charter.
These entities may offer perspectives and proposals that may be thought to enhance or
functionally adapt technical specifications in response to clinical protocol imaging expectations.
These alterations and interpretations may, in fact, redefine important trial endpoints via the
3 See the guidance for industry Computerized Systems Used in Clinical Investigations.

Contains Nonbinding Recommendations
26
proposed imaging technical details. We encourage sponsors to include in their charters a brief
section that states all imaging technical documents will be reviewed to ensure that the imaging-
specific details produce imaging outcomes consistent with the trial’s clinical protocol and
statistical analysis plan.

Contains Nonbinding Recommendations
27
APPENDIX B:
DURING IMAGING: MONITORING PLANS AND CHARTER MODIFICATIONS
A plan for ongoing monitoring of the imaging process is critical to ensure the quality of the
acquired images. Revisions to the imaging procedures might be recommended if unanticipated
technical issues arise.
Monitoring Plans
The charter should outline the complete plan for monitoring the imaging process. The extent of
monitoring is anticipated to vary widely, dependent upon the use of imaging within a trial. In
some situations, monitoring will be minimal; in other trials, intense monitoring (to include
requalification of equipment with phantoms and periodic retesting of readers) will be critical.
Sponsors should comply with the monitoring plan described within the charter. Verification of
this compliance may prove an important component of the assessment of imaging data integrity.
Charter Modifications
During the clinical trial, circumstances may prompt modification of the imaging procedures. For
example, unanticipated technical features may obscure a portion of an image or preclude one of
the expected quantitative assessments. In these situations, we recommend that the sponsor revise
the charter to correct the problem and maintain a record of the modification. The revision should
identify any potential effect of the modification upon the trial’s important endpoint analyses. In
some situations, modification of the charter may affect the definition of the primary endpoint
(e.g., alteration of the method for lesion measurement may call into question the clinical
meaningfulness of any size changes) and result in reconsideration of the role of imaging in the
trial or premature termination of the trial. To avoid these difficulties, we encourage sponsors to
thoroughly consider the role of imaging (including the technical aspects) in a clinical trial,
especially if the imaging is highly technical and/or relies upon quantitative assessments that
require vigilant subject and site cooperation with the imaging process. The use of imaging in
early phases of drug development may help lessen the challenges associated with wider use of
the technology in trials intended to support a drug’s approval.

Contains Nonbinding Recommendations
28
APPENDIX C:
AFTER IMAGING: DATA TRANSFER, ARCHIVING, AND
ANALYSIS OF IMAGING INFORMATION
A description of the procedures for the transfer of imaging data (e.g., from a trial site to a
centralized facility for image interpretation, from a centralized facility to a sponsor for analysis)
is critical for verification of data quality and integrity.
Data Transfer
It is important for sponsors to document fidelity to the charter-specified process of imaging
information transfer from a site to a centralized facility and from the centralized facility to the
sponsor throughout the clinical trial. Many clinical trials are likely to involve transfer of imaging
data to the sponsor only following completion of all image assessments and interpretations, and
some may involve image annotations before this transfer (such as a documented caliper-based
lesion measurement). In these situations, documentation of image and imaging information
transfer may greatly facilitate the completion of audits of the clinical trial’s imaging procedures.
Archiving
Sponsors and investigators should comply with the charter-specified plan for imaging source
data archiving. Deviations from this plan and/or loss of imaging information may compromise
the ability of the FDA to verify data quality and/or prompt reassessment of images. We do not
accept images as a component of new drug applications or biologics license applications.
However, we may request that sponsors display images, as part of records and reports relating to
a clinical investigation; upon our request, sponsors must permit the FDA to have access to these
records (see 21 CFR 312.58(a)).
Analysis of Image Information
We anticipate that most analyses of imaging information will be performed by the sponsor (or
the sponsor’s designated analysts) in accordance with the clinical protocol specifications. In
some situations, clinical sites or a centralized facility may analyze certain aspects of imaging as a
quality control measure (such as the determination of reader interpretation consistency).
Sponsors should specify these site and centralized facility roles in the charter. Clinical trial
imaging data should not be analyzed in an ad hoc, unplanned manner.
Imaging processes that took place during the conduct of the trial (such as image acquisition,
image interpretation, data transfer, and other processes described in the guidance) should be
thoroughly presented in the final report submitted for FDA review.