Osteoarthritis and Cartilage 23 (2015) 698e715

Review

OARSI Clinical Trials Recommendations: Knee imaging in clinical trials

in osteoarthritis
D.J. Hunter y z *, R.D. Altman x, F. Cicuttini k, M.D. Crema ¶ #, J. Duryea yy, F. Eckstein zz xx,
A. Guermazi ¶, R. Kijowski kk, T.M. Link ¶¶, J. Martel-Pelletier ##, C.G. Miller yyy,
T.J. Mosher zzz xxx, R.E. Ochoa-Albíztegui kkk, J.-P. Pelletier ##, C. Peterfy ¶¶¶,
J.-P. Raynauld ##, F.W. Roemer ¶ ###, S.M. Totterman yyyy, G.E. Gold zzzz xxxx kkkk
y Institute of Bone and Joint Research, Kolling Institute, University of Sydney, Sydney, NSW, Australia
z Rheumatology Department, Royal North Shore Hospital, University of Sydney, Sydney, NSW, Australia
x Department of Medicine, Division of Rheumatology and Immunology, David Geffen School of Medicine,
University of California at Los Angeles, Los Angeles, CA, USA
k School of Public health and Preventive Medicine, Monash University, Alfred Hospital, Melbourne 3004, Australia
¶ Quantitative Imaging Center, Department of Radiology, Boston University School of Medicine, Boston, MA, USA
# Department of Radiology, Hospital do Coraça ~o (HCor) and Teleimagem, Sa ~o Paulo, SP, Brazil
yy Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, Brazil
zz Institute of Anatomy, Paracelsus Medical University Salzburg & Nuremberg, Salzburg, Austria
xx Chondrometrics GmbH, Ainring, Germany
kk Department of Radiology, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin, USA
¶¶ Department of Radiology and Biomedical Imaging, UCSF, San Francisco, USA
## Osteoarthritis Research Unit, University of Montreal Hospital Research Centre (CRCHUM), Montreal, Quebec, Canada
yyy BioClinica, Newtown, PA, USA
zzz Department of Radiology, Penn State University, Hershey, PA, USA
xxx Department of Orthopaedic Surgery, Penn State University, Hershey, PA, USA
kkk Department of Radiology, The American British Cowdray Medical Center, Mexico City, Mexico
¶¶¶ Spire Sciences, Inc., Boca Raton, Florida, USA
### Department of Radiology, University of Erlangen-Nuremberg, Erlangen, Germany
yyyy Qmetrics Technologies, Rochester, NY, USA
zzzz Department of Radiology, Stanford University, Stanford, CA, USA
xxxx Department of Bioengineering, Stanford University, Stanford, CA, USA
kkkk Department of Orthopaedic Surgery, Stanford University, Stanford, CA, USA

a r t i c l e i n f o s u m m a r y

Article history: Significant advances have occurred in our understanding of the pathogenesis of knee osteoarthritis (OA)
Received 8 December 2014 and some recent trials have demonstrated the potential for modification of the disease course. The
Received in revised form purpose of this expert opinion, consensus driven exercise is to provide detail on how one might use and
9 March 2015
apply knee imaging in knee OA trials. It includes information on acquisition methods/techniques
Accepted 9 March 2015
(including guidance on positioning for radiography, sequence/protocol recommendations/hardware for
magnetic resonance imaging (MRI)); commonly encountered problems (including positioning, hardware
Keywords:
and coil failures, sequences artifacts); quality assurance (QA)/control procedures; measurement
Knee osteoarthritis
Imaging
methods; measurement performance (reliability, responsiveness, validity); recommendations for trials;
Clinical trial recommendations and research recommendations.
© 2015 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.

Introduction

* Address correspondence and reprint requests to: D.J. Hunter, Rheumatology
Department, Royal North Shore Hospital and Kolling Institute, University of Sydney,
Sydney, NSW, Australia. Tel: 61-2-9463-1887; Fax: 61-2-9463-1077.
E-mail address: David.Hunter@sydney.edu.au (D.J. Hunter).
Significant advances have occurred in our understanding of the
pathogenesis of knee osteoarthritis (OA) and some recent trials
have demonstrated the potential for modification of the disease

http://dx.doi.org/10.1016/j.joca.2015.03.012
1063-4584/© 2015 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.
 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
course1,2. Current regulatory requirements for disease modification
require that alongside the assessment of structural effects symp-
tom improvement be demonstrated3,4. The previous guidelines for
clinical trials published in 19965 included recommendations for
imaging with a predominant focus on radiography (commensurate
with the era) with some detail in the appendices on methods of
acquiring radiographs and use of magnetic resonance imaging
(MRI). There was little if any detail on the different imaging
methods available, the pitfalls inherent nor performance metrics of
different imaging markers.
The purpose of this data-based expert opinion, consensus driven
exercise is to provide detail to anyone involved in clinical OA trials,
imaging scientists and their respective teams on how one might use
and apply this knowledge in disease modifying clinical trials uti-
lising knee imaging assessments. It includes information on
acquisition methods/techniques (including guidance on posi-
tioning for radiography, sequence/protocol recommendations/
hardware for MRI); commonly encountered problems (including
positioning, hardware and coil failures, sequences artifacts); quality
assurance (QA)/control procedures; measurement methods; mea-
surement performance (reliability, responsiveness, validity); rec-
ommendations for trials; and research recommendations.
Methods
This paper is the work of a large multidisciplinary group who
were invited to participate in the Osteoarthritis Research Society
(OARSI) Clinical Trials Imaging Working Group. This review began
with a search of Pubmed using terms of OA, knee, and imaging. We
also conducted general searches for manuscripts covering general
imaging and randomized control trial (RCT) methods, covering each
of the sub-topics below; from this literature we identified designs
and methods, as well as other manuscripts of high relevance. It
should be noted that the vast majority of this manuscript is based
upon expert opinion of the diverse multidisciplinary group
involved in this exercise. Authors of this review included multiple
experts familiar with different approaches for imaging of the knee
joint, including radiologists, rheumatologists, and engineers. This
expert opinion was generated via a series of teleconference and
email exchanges followed by generation of a series of recommen-
dations. This correspondence allowed the working group to iden-
tify additional topics and manuscripts for inclusion and to develop
and reach concurrence on a set of recommended principles for
inclusion of knee imaging methods in OA trials. Given the potential
for divergent perspectives for the trial recommendations and
research recommendations a survey was conducted of these
members to determine the strength of recommendation for each
point raised. Final recommendations were obtained by averaging
the responses to a survey among the 14 authors for the strength of
recommendation from 0 to 100. At the commencement of this
exercise all members of the working group were asked about
conflicts of interest and the results from those who were conflicted
were not included in the survey results. The focus of the content is
on radiography and MRI as the preferred imaging techniques with
some content on ultrasound when appropriate.
Image acquisition
Imaging of the knee can be obtained by a variety of modalities.
X-ray, ultrasound, computed tomography (CT) and MRI are most
widely used and well developed at this time. New trials are using
PET and SPECT/scintigraphy to evaluate knee OA6. There continue
to be limitations in the application of arthroscopy, ultrasound and
bone densitometry techniques in clinical OA trials; hence, their
acquisition methods are not discussed below.
699
X-ray of the tibiofemoral joints
Radiography of the tibiofemoral joints is most often evaluated
with anteroposterior (AP) positioning7. Although supine (non-
weight bearing) AP images are easily and frequently obtained, the
joint space width (JSW) cannot be adequately estimated in that
position8. The AP standing view is the most frequently used view
in clinical practice, but not in clinical trials. Both knees can be
included using a 36  43 cm (14  17 in) detector system if the
thighs are not heavy or significant varus is not present. Otherwise,
a single knee can be imaged with a suitable (e.g., 24  30 cm)
detector system. For the standing X-ray, care must be taken that
the knees be fully extended and against the cassette to avoid
distortion. The X-ray beam should be 6 caudad and a 180 cm (72
in) focus to film distance, utilizing 65e72 kVP and a small focal
spot. With feet pointed straight ahead, the X-ray beam should be
centered midway between the knees (bilateral) or horizontally at
mid-patella (single knee) and about 1 cm below the apex of the
patella.
In cross sectional epidemiologic or symptom modifying clinical
trials, the posteroanterior (PA) X-ray is useful for supporting a
diagnosis of knee OA. For longitudinal epidemiologic or structure
(disease) modifying trials, the standing AP radiograph of the knee
does not adequately define joint space narrowing (JSN), as the
most severe changes in condylar articular cartilage are often
somewhat posterior9. To more accurately determine the narrowest
point in the medial compartment, BucklandeWright developed a
method of fluoroscopically positioning the knee10. This and several
other techniques have been developed in order to have repro-
ducible positioning for semi-quantitative OARSI grading of JSN and
quantitative measurements of the medial joint space11,12. The basic
principle of these methods are to fix the alignment of the medial
tibial plateau, while flexing the knee at a reproducible angle in
order to capture the narrowest region of cartilage of the medial
condyle. The most commonly used techniques used at this time
are the fixed-flexion PA view and the fluoroscopically aligned
Lyon-Schuss PA view, with the use of a SynaFlexer positioning
frame (BioClinica (formerly Synarc), San Francisco, CA)13e15. In the
Lyon Schuss view, the patellae and thighs are in contact with the
film cassette and coplanar with the tips of the great toes. In the
(MTP) fixed flexion view the patellae are in contact with the
cassette and aligned vertically with the first MTP joints with the
feet externally rotated 5 . This results in reproducible flexion of
the knee at approximately 20 . The X-ray beam is angulated 10
caudad and centered on a reproducibly identifiable point in the
popliteal fossa of a single knee or midway between the two knees
at the level of the popliteal fossa if both knees are imaged
simultaneously, with the aim of aligning the beam tangential with
the floor of the medial tibial plateau. The Lyon Schuss technique
uses the same positioning and beam centering, but fluoroscopi-
cally aligns the anterior and posterior rims of the medial tibial
plateau. Magnification correction is assisted by a projection
phantom on the SynaFlexer frame or by placing a ball bearing of
known dimension in the same position medial to the knee if im-
aging without the SynaFlexer. Fluoroscopic guidance, if done
correctly, provides high quality and optimally aligned images.
However, this requires more training and is not available at most
sites in a multicenter trial. A modified Lyon-Schuss (mLS) tech-
nique has been used successfully, requiring imaging the knee three
times with a 2 difference in tube head angle to ensure optimum
positioning as observed with the inter-margin distance16,17. It
should be noted that special images (long limb films) were his-
torically recommended to adequately judge varus or valgus de-
formities, but more recently, the knee film alone has been found to
be sufficient18.
700 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715

Table I
Semi-quantitative scoring MRI acquisition techniques30,31,38

Tissue Location Feature MRI technique Imaging plane

Primary Secondary Primary Secondary

Cartilage Tibiofemoral Focal defect PD FSE/T2 TSE with FS 3D T1-GRE with FS or WE Coronal Sagittal
Patella Focal defect PD FSE/T2 TSE with FS 3D T1-GRE with FS or WE Axial Sagittal
Trochlea or posterior femur Focal defect PD FSE/T2 TSE with FS 3D T1-GRE with FS or WE Sagittal
Bone marrow Tibiofemoral Lesions/cyst STIR/PD FSE 3D TSE/T1 w IV cont Coronal Sagittal
TSE with FS or WE
Fracture T2 FSE/TSE with FS or WE or STIR T1 FSE/TSE Coronal Sagittal
Patella Lesions STIR/PD FSE 3D TSE/T1 w IV cont Axial Sagittal
TSE with FS or WE
Fracture T1 without FS or STIR T2 FSE/TSE with FS or WE Axial Sagittal
Bone Tibiofemoral Osteophytes T1 FSE/TSE; GRE/SPGR, FLASH, DESS PD FSE/TSE with or without FS Coronal Axial
Patella Osteophytes T1 FSE/TSE PD FSE/TSE with or without FS Axial Sagittal
Trochlea Osteophytes T1 FSE/TSE PD FSE/TSE with or without FS Sagittal Axial
Meniscus Morphology/Tear PD, T1 FSE/TSE T2 FSE/TSE with FS; 3D FSE Sagittal Coronal
Extrusion PD, T1 FSE/TSE with FS T2 FSE/TSE with FS; 3D FSE Coronal Sagittal
Ligament/tendon ACL/PCL Tear T2 FSE/TSE with or without FS T1 FSE/TSE Sagittal Axial
Collateral Tear T2 FSE/TSE with or without FS T1 FSE/TSE Coronal Axial
Knee extensor Tear T2 FSE/TSE with or without FS T1 FSE/TSE Sagittal Axial
Synovium Peri-patellar fat pads Synovitis T1w þ IV with FS T2 PD FSE/TSE with FS or WE Sagittal Axial
Posterior cruciate recess Synovitis T1w þ IV with FS T2 PD FSE/TSE with FS or WE Sagittal Axial
Para-meniscal recess Synovitis T1w þ IV with FS T2 PD FSE/TSE with FS or WE Coronal Axial
Medial and lateral recesses Synovitis T1w þ IV with FS T2 PD FSE/TSE with FS or WE Axial Sagittal
and suprapatellar bursa
Joint effusion Anterior compartment Fluid volume T1w FS CE, T2 FSE/TSE with or without FS PD FSE/TSE with or without FS Axial Sagittal
Popliteal cyst Fluid volume T1w FS CE, T2 FSE/TSE with or without FS PD FSE/TSE with or without FS Axial Sagittal
Muscle Myositis T2 FSE/TSE with FS or WE STIR Axial Sagittal
Atrophy T1 FSE/TSE Axial Sagittal
Bursae Pes anserine. MCL bursa Inflammation T1w FS CE, T2 FSE/TSE with FS or WE STIR Coronal Axial
Peri-patellar bursae Inflammation T1w FS CE, T2 FSE/TSE with FS or WE STIR Sagittal Axial

Definitions
PD FSE/TSE: Proton density weighted fast or turbo spin echo (TR > 2000 ms, TE in the range of 25 mse35 ms).
T2 FSE/TSE: T2-weighted fast or turbo spin echo (TR > 2000 ms, TE in the range of 40e60 ms).
T1 FSE/TSE: T1-weighted fast or turbo spin echo (TR < 1000 ms, TE in the range of 10e20 ms).
3D GRE: 3D gradient echo sequence 3D GRE 3D gradient echo sequence.
Short tau inversion recovery recommended in post-operative setting where there may be artifact due implanted.
STIR: metal.
FS: fat suppression.
WE: Selective water excitation used at 3 T.
T1w þ IV: T1 with intravenous contrast Gadolinium.
DCE: Dynamic gadolinium contrast enhanced perfusion imaging.

X-ray of the patello-femoral joint the patella, anterior to the thigh with the film cassette just distal
to the knee.
A lateral view of the patellofemoral (PF) joint can be used to
supplement the findings on the AP view. In addition, measure-
MRI
ment of the lateral view radiographic tibiofemoral JSW may also
be of value19. The lateral PF view is often obtained with the patient
Due to the limitations of radiography, MRI has been identified by
lying supine on the table, hips and knees flexed with the heels on
the Outcome Measures in Rheumatologic Clinical Trials (OMERACT)
the table enough to flex the knees to 45 . Additionally, determi-
and OARSI as the most appropriate imaging modality to assess joint
nation of PF alignment and specifics of PF JSW (adequate deter-
status in OA research studies20. MRI can detect structural pathology
mination of medial vs lateral PF involvement) are often viewed
associated with pain21,22 and other tissues involved in the disease
with an axial (skyline, sunrise, sunset) image of the knee, obtained
process such as cartilage damage, osteophytes, subchondral cysts,
with the X-ray beam parallel to the table and perpendicular to a
joint effusions, ligament and tendon tears, Baker's cysts, synovitis,
14  30 cm cassette. The X-ray tube is positioned 10e12 cm above
meniscal tears, and subchondral bone marrow lesions23e31. MRI is

Table II
Quantitative cartilage morphometry MRI acquisition techniques

Location MRI technique Imaging plane

Primary Secondary Primary Secondary

Tibiofemoral 3D T1 GRE with FS or WE 3D DESS Sagittal Coronal

Patella 3D T1 GRE with FS or WE 3D DESS Axial Sagittal
Trochlea or Posterior femur 3D T1 GRE with FS or WE 3D DESS Sagittal Axial

Resolution

Field strength Section thickness In-plane

3T 0.7 mme1.0 mm 0.3 mme0.5 mm

1.5 T 1.0 mme1.5 mm 0.3 mme0.6 mm
 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
Table III
Compositional cartilage imaging techniques142
Biomarker Underlying biological feature the biomarker
is believed to be a proxy for
T2 Collagen anisotropy and hydration
T1rho Glycosaminoglycan content and meniscus
dGEMRIC Glycosaminoglycan content
Contrast enhance CT Cartilage glycosaminoglycan
Sodium MRI Cartilage glycosaminoglycan
gagCEST Cartilage glycosaminoglycan
Diffusion Cartilage water mobility
Ultra-short TE Osteochondral junction
Definition
T2: spin spin relaxation time.
T1rho: Spin lattice relaxation in the rotating frame.
becoming more widely available and requires no exposure to
ionizing radiation32.
The mechanisms of pain and mechanical dysfunction in OA are
not completely understood but are believed to involve multiple
interrelated pathways involving all joint structures23,26,33e37. MRI
protocols in OA research studies should be tailored for optimized
visualization of the tissue of interest and take into account the
image analysis method such as semi-quantitative scoring, quanti-
tative assessment of tissue dimensions (volume, thickness) or
biochemical composition20,30. For 3D volumetric cartilage analysis
dedicated high resolution sequences with good contrast between
cartilage and surrounding structures will be needed. “Whole-joint”
MRI protocols need to be tailored to evaluate all joint structures in
three imaging planes including ligaments, tendons, menisci, sub-
chondral bone, and synovium (see Table I)33,38.
MRI protocols for semi-quantitative evaluation of joint struc-
tures include two-dimensional fast or turbo spin-echo (FSE/TSE)
sequences, which have highly versatile tissue contrast that can be
used for comprehensive joint assessment. Two-dimensional FSE
images should be acquired in multiple planes to provide detailed
evaluation of all articular surfaces and all joint structures. Two-
dimensional fluid sensitive sequences such as fat-suppressed in-
termediate-weighted or T2-weighted FSE or short tau inversion
recovery should be performed in at least one plane to allow
improved detection of ligament tears, bone marrow lesions, and
joint effusions. Protocols are optimally performed on a 3.0 T scan-
ner using multi-channel phased-array extremity coils to optimize
signal-to-noise ratio (SNR)33. However, current 1.5 T scanners
provide image quality sufficient for reproducible analysis for most
protocols.
MRI protocols should include 3D sequences if quantitative,
volumetric evaluation of articular cartilage is to be performed (see
Table II). Fat suppression is typically added to these sequences to
reduce chemical shift artifact and to optimize the overall dynamic
contrast range of the image. 3D sequences are acquired with high in-
plane spatial resolution and thin continuous slices which reduce
partial volume averaging. Many 3D sequences have near-isotropic
resolution which allows articular cartilage and other joint struc-
tures to be evaluated in any orientation through the creation of high
quality multi-planar reformat images. Various three-dimensional
sequences have been used for quantitative cartilage assessment
within the knee joint including T1-weighted, fat-suppressed
gradient-echo39e41 and dual-echo in the steady-state (DESS)42.
Although the three-dimensional FSE/TSE techniques are
improving, there are still numerous drawbacks compared to the
more widely available 2D techniques: (1) image blurring when
using 3D FSE sequences secondary to T2 modulation of the point
spread function, (2) decreased contrast for evaluating soft tissue
structures including bone marrow lesions, ligaments, tendons, and
menisci (3) greater variation between manufacturers, (4) longer
701
Contributing tissues
Collagen and content, water diffusivity
Changes in water content predominantly meniscus
Water diffusivity
Novel CT contrast agents rodent models of OA
Suppresses signal from synovial fluid
Preliminary validation in the intervertebral disc and cartilage repair tissue
Calculates apparent diffusion coefficient and T2
Isolate signal from the zone of calcified cartilage
image acquisition time which increases likelihood of motion, (5)
increased susceptibility to metallic artifacts. Although, 3D gradient-
echo sequences have been validated and applied for volumetric
cartilage assessment for over a decade43, there are few studies
available using 3D FSE/TSE sequences for semi-quantitative
assessment44.
Gradient-echo sequences were the first 3D sequences used to
evaluate articular cartilage. Gradient-echo sequences can be
divided into dark fluid sequences and bright fluid sequences based
upon the signal intensity of synovial fluid. Dark fluid sequences
include T1-weighted spoiled gradient recalled-echo (SPGR, GE
Healthcare), fast low angle shot (FLASH, Siemens Medical Systems),
or steady-state free precession (SSFP, GE or 3D-FISP, Siemens) and
T1-fast field echo (T1-FFE, Philips Healthcare). Bright fluid se-
quences include T2*-weighted gradient recalled-echo acquired in
the steady-state (GRASS, GE Healthcare), gradient recalled-echo
(GRE, Siemens Medical Systems), and fast field-echo (FFE, Philips
Healthcare). These sequences have been combined with a variety of
fat-suppression techniques including frequency selective fat-satu-
ration41,45,46, water excitation47,48, and iterative decomposition of
water and fat with echo asymmetry and least squares estimates
(IDEAL)49e51. Bright fluid gradient-echo sequences have been found
to be the most useful for detecting cartilage lesions in clinical
practice and OA to perform cartilage thickness measurements in OA
research39e41,45,46,52. The main disadvantage of using dark fluid
sequences for clinical cartilage imaging is that the low signal in-
tensity of synovial fluid may decrease the conspicuity of superficial
cartilage lesions51,53.
Driven equilibrium fourier transform (DEFT) and DESS are
additional three-dimensional sequence used to evaluate articular
cartilage. DEFT uses a 90 pulse to return transverse magnetization
to the z-axis which increases the signal intensity of tissues such as
synovial fluid with long T1 relaxation times. Cartilage signal is
relatively preserved on DEFT images due the use of short echo
times48,54,55. DESS acquires two gradient echoes separated by a
refocusing pulse which are combined into a single image. Adding
the two echoes enhances the T2*-weighting of the image and in-
creases the signal intensity of both cartilage and synovial fluid56.
DEFT and DESS utilize frequency selective fat-saturation54,55 or
water excitation57,58 to suppress signal from adipose tissue. Both
sequences produce multi-planar images of the knee joint with
bright synovial fluid which creates an arthrogram-like effect that
increases the conspicuity of superficial cartilage lesions. A water
excitation DESS sequence with 0.4 mm  0.5 mm in-plane spatial
resolution and 0.7 mm slice thickness is currently being used in the
Osteoarthritis Initiative to provide detailed cartilage assessment42.
Three-dimensional FSE sequences such as fast spin-echo Cube
(FSE-Cube, GE Healthcare) and sampling perfection with applica-
tion optimized contrasts using different flip angle evolutions
(SPACE, Siemens Medical Systems) have also been used to evaluate
702 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
articular cartilage. These sequences utilize flip angle modulation to
constrain T2 decay over an extended echo train which allows
intermediate-weighted images of the knee joint with isotropic
resolution to be acquired with minimal blurring. Three-
dimensional FSE sequences typically use frequency selective fat-
saturation to suppress signal from adipose tissue59,60. FSE-Cube
and SPACE have lower in-plane spatial resolution and greater im-
age blurring when compared to other three-dimensional cartilage
imaging sequences with similar acquisition times which may
decrease the conspicuity of superficial cartilage lesions61,62. How-
ever, these sequences have highly versatile intermediate-weighted
contrast which can be used to evaluate all joint structures. In fact,
three-dimensional FSE sequences with multi-planar reformats has
been shown to provide near identical “whole-organ” knee joint
assessment in OA research studies as axial, sagittal, and coronal
two-dimensional FSE sequences in shorter periods of time by
eliminating the need to acquire images with identical tissue
contrast in multiple planes44. It should be noted, however, that the
image quality of the reformations is not comparable with the ac-
quired source images and that reformations cannot fully substitute
2D images obtained in the same orientation63.
For cartilage quality studies quantitative mapping techniques
have been used to measure MRI parameters that are sensitive to
early changes in the composition and structural organization of the
extra-cellular cartilage matrix that precede loss of tissue detected
with standard MRI techniques (Table III). Techniques that have been
applied in clinical trials include cartilage T2 mapping (which has
been included in the OAI protocol), T1rho mapping, and delayed
gadolinium enhanced MRI of cartilage (dGEMRIC) and have been
the focus of several review articles64e66. dGEMRIC can provide
sensitive and specific information regarding the proteoglycan
content of cartilage67e70, but the technique requires a long waiting
period between contrast administration and image acquisition and
carries the risk of nephrogenic systemic sclerosis71,72. T2 and T1-
rho mapping have been shown to be sensitive for detecting early
cartilage degeneration in human subjects without the inconve-
nience and risks associated with the use of exogenous contrast
agents73,74. However, both techniques are not chemically specific,
but are influenced by concurrent changes in water content,
macromolecule content, and anisotropic organization that occur
during early cartilage degeneration. As such it is important when
interpreting the results of these techniques to realize that the
measured response to tissue degeneration is nonlinear and that
biochemical specificity of the response will generally decrease with
more advanced degeneration.
CT arthrography
CT arthrography can also be used to provide knee joint assess-
ment in OA research studies. CT is a widely available and relatively
Fig. 1. (a). Positioning of the conventional extended knee AP or PA radiograph with horizontal beam. (b). Skewed radio-anatomic alignment of the medial tibial plateau apparent in
the displacement of anterior and posterior margins (arrows) and difficulty identifying the cortex of the floor of the plateau.
inexpensive imaging modality which can rapidly acquire high
resolution volumetric source data that can be used to create
reformatted images in multiple planes. CT imaging has high
sensitivity for detecting various features of joint degeneration
including meniscus and ligament tears, cartilage loss (these first
three features following the injection of iodinated contrast into the
knee joint), subchondral cysts and sclerosis, and osteophyte for-
mation75e78. The technique can also be used to measure the
thickness79,80 and proteoglycan concentration81,82 of cartilage.
However, CT arthrography is an invasive technique requiring radi-
ation exposure and intra-articular contrast administration which
has limited its use in longitudinal OA research studies. Also eval-
uation of bone marrow and ligaments is limited with CT compared
to MRI. A more recent study suggested CT may also have promise
for evaluation of calcium crystal deposition in the knee83.
Commonly encountered problems: positioning, hardware and
coil failures, sequences artifacts
Commonly-encountered problems in radiography
Projection
Radiography is a projectional technique in which 3D anatomy is
projected onto a 2D receptor, and therefore subject to morpholog-
ical distortion, magnification and superimposition of overlaying
structures. A reproducible radiographic image of the tibiofemoral
joint space thus requires adherence to exacting standards of
centering and angulation of the X-ray beam, and positioning of the
knee, which include specifications for flexion and rotation of the
joint11. In a majority of patients, the anatomy of the knee is such that
full extension of the joint (as is required for a conventional weight-
bearing extended-knee radiograph) tilts the tibial plateau to an
angle that is skew (i.e., not parallel) to a horizontally directed X-ray
beam84 [Fig. 1(a)]. Skewed radio-anatomic alignment of the tibial
plateau in an AP or PA knee radiograph is apparent in the excessive
displacement of the anterior and posterior margins of the medial
plateau [Fig. 1(b)]. When the alignment of the plateau is notably
skewed, the floor of the joint space can become indistinct, and a key
reference point for the measurement of minimum JSW (minJSW) is
less reliably ascertained. Typically it is recommended that the inter-
marginal distance (IMD) should be 1.5 mm85 or even 1 mm86.
Although, other studies suggest that while stability of the IMD be-
tween visits matters, as a metric of reproducible projection, small
differences in absolute IMD may not empirically affect JSN87.
1 Flexion of the knee, rotation of the knee
If the articulating surfaces of tibia and femur overlap or the
anterior and posterior tibial rims are grossly malaligned, unclear
visualization or artificial variation of the JSW may result [Fig. 1(b)].
 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
Different degrees of knee flexion between time points cause
problems in longitudinal studies88, Fig. 2).
Further, articular cartilage loss in the knee most commonly
begins along a region of the femoral condyle that is posterior to that
which articulates with the tibial surface when the knee is in full
extension. Since radiographic joint-space width reflects cartilage
thickness only where the femoral and tibial articular surfaces are in
direct contact, positioning the knee in full extension can be
insensitive to early cartilage loss. Mild flexion is thus needed to
articulate this active portion of the femur with the central tibia and
detect early cartilage loss. Moreover, since cartilage loss progresses
heterogeneously in the knee, knee flexion as well as beam
centering and angulation must be exactly the same on serially ac-
quired radiographs for longitudinal assessments of JSN to be valid.
Different degrees of rotation of the knee may cause different
degrees of overlap between osteophytes and the margin of femur/
tibia, leading to non-visualization of previously seen osteophytes,
exposure of previously obscured osteophytes, or a discrepancy in
their apparent size between timepoints when in fact there has been
no real change in size.
2 Weight bearing
As noted above, radiographic joint-space width can be valid as a
measure of cartilage thickness only when the two opposing carti-
lage surfaces of interest are in direct contact with each other. The
knee must therefore be loaded in order to ensure this. However,
weight bearing during radiography poses a number of additional
challenges89. For example, longitudinal changes in weight bearing
(for example, due to weight gain or loss) may affect the extent of
voluntary knee extension. Accordingly, effort must be taken in se-
rial examinations to fix the degree of flexion reproducibly. Changes
in the distance between the knee and radiographic cassette may
alter the degree of radiographic magnification in the image. These
sources of error seriously limit the utility of the knee radiograph to
Fig. 2. Radiographs at baseline and 2-year follow-up of a 61-year-old woman with OA. At two time points, radiographs were taken with 5 , 10 , and 15 angulation of the knee. (a)
AP radiograph taken at 5 angulation shows medial JSN (OARSI grade 2 and KLG 3). (b) However, AP radiograph taken at 10 angulation shows OARSI grade 3 JSN (KLG 4). (c)
Similarly, at follow-up, AP radiograph taken at 5 angulation shows OARSI grade 2 JSN (KLG 3) whereas that taken at 15 angulation (d) shows OARSI grade 3 JSN (KLG 4). If images
(b) and (c) are compared longitudinally, one observes a paradoxical ‘joint space widening’.
703
detect true reduction in JSW, the cardinal indicator of progression
of knee OA. While these challenges are recognized, recent studies
have demonstrated that with careful site technologist training,
using a weight bearing acquisition technique, careful unilateral
imaging of each knee (not a bilateral knee radiography), proficient
quality control (QC) of the radiography by a core lab experienced in
the technique, either the fixed flexion or mLS technique can provide
reproducible images to allow evaluation of a disease modifying
osteoarthritis drugs (DMOAD) in a large multi-center clinical trial.
3 X-ray beam angle
Variation in X-ray beam angle can result in apparent difference
in the JSW (Fig. 3). This can also cause artificial variation in JSW in
longitudinal studies.
Commonly-encountered problems in MRI
In clinical trials of knee OA, the images need to cover the whole
knee joint. For optimal images, the joint should be centered at the
midpoint of the coil and the iso-center aligned caudal to the distal
end of the patella. That guarantees full coverage of the knee. However
if the synovial fluid is to be measured, the field of view (FOV) must be
shifted proximally to ensure that all parts of the joint capsule are
covered in both the proximal-distal and medial-lateral directions.
Imaging plane alignment
The curved shape of the femoral condyles and the patella re-
quires selection of an imaging plane perpendicular to the curvature
to provide clear views of the cartilage surface and to minimize
partial volume effects. Imaging planes should provide the best
views of articulating cartilage, bones and soft tissue structures both
for scoring and quantitative assessment of cartilage, bone marrow
lesions, and meniscus30,90. Selection of an oblique imaging plane
aligned with the anterior cruciate ligament may be helpful for
704 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
Fig. 3. In this example using fixed flexion PA protocol, the joint space will appear
wider when a beam that is parallel to the tibial plateau is used (thick dotted line)
compared to when a beam is slightly angulated with respect to the tibial plateau (thin
dotted line). Therefore, tube angulation needs to be kept constant between time
points, as well as the angle of flexion of the knee. For the mLS technique with varying
angles at baseline, the technologist at acquisition has to reference the baseline
radiograph and tube angle used.
assessing ligament integrity30,90. However, for quantitative analysis
of the whole femoral cartilage including the trochlea and posterior
femoral condyles and patella, sagittal images are preferred. If
analysis is limited to cartilage of the weight bearing regions, cor-
onal images can be used.
In order to avoid oblique knee images, the feet should be posi-
tioned with the big toe up91. This eliminates knee rotation, and
standardizes knee positioning across study subjects and time-
points. Further, MRI sections should be oriented with careful
biplanar alignment based on reproducible anatomical landmarks.
Image resolution
Quantitative cartilage imaging is demanding for resolution, with
slice thickness from 0.7 to 1.5 mm and in plane resolution from
0.3  0.3 mm to 0.5  0.5 mm necessary to reliably detect and
follow small defects less than 5 mm2. This is driven by sampling
theory, which recommends that twice the sampling frequency of
the signal bandwidth is required for reliable detection of lesions;
this case means twice the geometric diagonal of the voxel size for
reliable feature detection. In knee OA the pattern of cartilage loss
varies from an increase in size of a single defect to thinning of the
cartilage over a large area, therefore a slice thickness of 1e1.5 mm
and in-plane resolution of 0.4  0.4 mm or greater are recom-
mended30,90. In this context it is important to note that quantitative
analysis of cartilage volume and thickness does not require
isotropic resolution; a pixel size of 0.3  0.5 mm is acceptable
allowing for faster anatomical coverage along the longer dimension.
Standardization of slice thickness across longitudinal timepoints
is important in scoring change in both cartilage and bone marrow
lesions as the grade is determined by the number of slices the
feature is observed in. Therefore all sites must comply with the
specified standardized imaging protocol. In addition, incoming data
QC needs to include verification of imaging parameters, including
slice thickness and in-plane resolution92. To avoid future errors, and
maximize the chance of recalling subjects when necessary for re-
imaging and retaining them in the study, feedback to imaging
sites needs to happen quickly and on an ongoing basis.
Imaging sequences
The goals of imaging sequences for morphological cartilage
analysis are high in-plane resolution with high contrast to noise ratio
between cartilage and all adjacent tissues and fluid. Many of the
techniques are manufacturer-dependent, but most include some
type of fat suppression or selective water excitation and 3D FSE or
gradient echo sequence, although 2D FSE is sometimes used for
semi-quantitative scoring. All of those sequences have been vali-
dated for imaging of cartilage62,93e97. Their ability to show the
boundaries between bone-cartilage, cartilageecartilage, cartilage-
degenerated meniscus, cartilage-inflamed synovium, and cartilage-
inflamed synovial fluid varies. While selecting a sequence it is
important to take care that the imaging parameters are also chosen
in such a way that the visualization of those interfaces are optimized,
particularly those with the greatest impact on study endpoints.
MRI is used for compositional assessment including tech-
niques such as T2 imaging, dGEMRIC and T1 rho imaging. This
imaging has similar resolution dependencies as quantitative
morphologic imaging. Thick slices and low in-plane resolution will
limit detection to generalized changes and prevent detection of
local abnormalities and their progression. For T2 measurements the
ideal slice thickness is 2 mm, but for practical purposes 3 mm
resolution may be needed98,99. There are varying opinions of the
number of echoes needed in the multi-echo spin echo series for
repeatable T2 calculations, varying from two echoes to 11
echoes98,99. The practical issue is that the larger the number of
echoes the smaller the number of slices, subsequently limiting
coverage of the knee. Regardless of the number of slices, however,
assessments are only useful for comparison if the area covered and
the imaging plane used are the same from time point to time point.
Standardization (consistency of image acquisition parameters
within the study) of image acquisition cannot be overemphasized.
The variability of T2 measurements between different manu-
facturers and different imaging units is a well-known confounding
factor. The effect of the variability can be minimized by using small,
standardized T2 phantoms, which are positioned in the imaging
field of view of the T2 series acquired for each subject at each time
point. This will enable pooling of T2 data across multiple sites and
equipment and verify possible change between time points100. At
this time, T1 rho sequences have not been validated for use in OA
trials, therefore their use is currently limited to experimental ter-
tiary endpoints.
Since both quantitative and qualitative assessments may use fat
suppressed/saturated sequences, heterogeneous fat suppression
due to field heterogeneity will affect analysis results73,101. Since the
cartilage thickness changes over time are generally relatively small,
geometric distortion can affect the reliability of the results. The
stability of field homogeneity should be monitored by scanning
special uniformity and linearity phantoms at the start of the study
and periodically throughout92. Transmit/receive as well as receive
only coils, which are used for knee imaging have a tendency to start
failing slowly102. Therefore, phantom scanning also should monitor
their performance throughout the study. Although a relatively rare
occurrence, changes in imaging performance can be caused by
software or hardware upgrades to the scanner, or problems
requiring a re-shim to correct. Periodic phantom scanning can
detect these issues before too much study data has been acquired.
In addition to the difficulties outlined above there can be diffi-
culties in safely putting patients with pacemakers or imbedded
metal foreign objects inside the magnet. A small proportion of pa-
tients have problems with claustrophobic reactions. The size of the
joint being imageddfor example a large knee in an obese patient-
dmay be too large for the cylindrical RF coil typically used, leading
to potential impaired image quality. Researchers need to be aware
of commonly encountered MR artifacts. Screening for artifacts that
potentially hinder adequate and appropriate image assessment
should be part of standardized QA procedures that are discussed in
detail in the next section. The challenges in conjunction with MRI
 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
artifacts have been reviewed extensively. A detailed discussion of
these would be beyond the scope of this manuscript103.
QA/control procedures
Clinical trials are becoming more complex, and from a regula-
tory perspective, they are underpinned by an over-arching quality
process or Good Clinical Practice (GCP)104. The understanding of the
GCP framework is critical to ensure the management of all aspects
of clinical trials, not least of which is the imaging. Essentially, this is
twofold: data integrity and subject protection105. From an imaging
perspective, more recently the requirements specifically for medi-
cal imaging in clinical trials, including the necessary documenta-
tion, such as an Imaging Charter have been detailed by the Food and
Drug Administration (FDA)92. Potentially more practical details for
the charter and review of read systems have been described in
detail recently106. A systematic approach is required to ensuring
quality as detailed below107:
 Personnel roles and responsibilities
 Training
 Policies and procedures
 QA and auditing
 Document management, record retention, and reporting
 Corrective and preventive action (CAPA)
There is a major difference between QC and QA:
Quality control, as stated by the ISO definition, is the summari-
zing term covering the operational techniques that are used to
fulfill requirements for quality. In the context of medical imaging in
clinical trials, this includes the framework of the documents
described and the checking of the medical imaging data by a third
party such as an imaging core lab. A diagrammatic representation is
shown in Fig. 4.
Quality Assurance comprises the overarching procedures that
are put in place to ensure quality, which includes the checking and
audit of the process and QC systems that have been instigated.
Medical imaging QC in clinical trials
When evaluating the QC procedures one needs to understand
the key aspects of the use of medical imaging in medicine in general:
(1) Diagnosis; (2) Prognosis or screening; (3) Monitoring of therapy
and (4) Monitoring of natural history in order to initiate, refrain from
or change therapy. In clinical trials one usually considers efficacy
measures, which are items 3 & 4 (monitoring therapy, and moni-
toring natural history for placebo controlled studies). For eligibility,
items 1 and 2 are relevant. There are different metrics that have to
be considered for each of these uses. KLG108 is used for eligibility but
it does not have the sensitivity or precision for efficacy and hence
the radiographic QC may differ when applying radiographic KLG for
Fig. 4. QC, retrieved from http://www.transition-support.com/Quality_control.htm accessed 05 February 2014.
705
eligibility compared to reduction in radiographic JSW being used as
an outcome in longitudinal studies looking at efficacy.
The challenge at screening in regard to required image quality is
to account for both, the requirements of eligibility and efficacy: e.g.,
if JSW loss is the primary end point in a clinical trial, then the
acquisition of the knee X-ray has to be of sufficient quality for the
JSW measurement, but for a pure screening KLG acquisition, a
standard AP film may be adequate, as when an analgesic is being
developed without DMOAD properties.
Cross-calibration
Depending on the end-point, cross-calibration may be an
optional requirement. For many outcomes in clinical trials the key
end point is percentage change from baseline. If each subject is used
as their own control (i.e., compare follow up with baseline or pre-
vious follow-up visits), then cross calibration is not required. This is
certainly the case with all radiographic and MRI endpoints in
DMOAD studies. However, it has to be ensured that image acquisi-
tion at serial visits is performed in identical fashion including tech-
nical aspects such as MRI or radiography system, sequence protocols
used, positioning and others. This leads to a set of QC features that are
required regardless of the anatomical area being evaluated:
1. Radiographs. If a grading scale is being used, then no calibration
is required, but the above mentioned items in regard to repro-
ducibility of image acquisition apply also for studies using
ordinal outcomes. If JSW or another quantitative end point is
being used, then a calibration marker or markers have to be
placed and be visible in the FOV at about the same distance to
the detector as the joint is located, to minimize the effects of
beam divergence and magnification.
2. MRI. If a scoring or grading system is being used, no calibration
is required (other than the routine standard ones conducted at
the site for correct instrument use). However, full anatomical
coverage and comparable sequence protocols are a requirement
for longitudinal comparison. If cartilage volume, thickness or
other quantitative parameters are being measured, very specific
calibration phantoms should be used, which are anatomic and
end-point specific. For continuous biomarkers such as these
quantitative measures then multivendor trials should demon-
strate and document that the biomarker values are comparable
across vendors and across centres. Subjects must always be
scanned on the same instrument on which they had their
baseline measurement, and using identical imaging parameters,
regardless of methodology being applied. If change in quanti-
tative cartilage is to be measured, then 3 T with phased-arrays is
recommended, while if whole-organ scoring is the primary aim,
then any 1.5 T or 3 T system which passes QA can be used.
3. Ultrasound. Ideally all the equipment should be standardized
across all sites. If not, then each subject must be scanned using
706 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
the same equipment at each visit. The technologist acquiring the
data has to be very conscientious in marking up the scans so the
technologist providing QC can follow the joint and positioning.
Instructions manual
For all imaging, instruction manuals or “cheat sheets” must be
provided to all sites detailing the specific acquisition requirements.
Without this set of details, QC cannot be conducted.
Reader “training” or calibration
For all the imaging evaluations performed in OA clinical trials,
radiologists or radiologist-trained readers may be involved in the
clinical study as readers. If only one reader is used, a regular cali-
bration process is needed during the course of a longitudinal study
to avoid shift/drift in readings. If more than one reader is to read the
images (highly recommended, except for small studies), then it is
important to bring the readers together for a calibration meeting
for the following specific reasons:
1. Training on the reading protocol
2. Training on the reading interface (unless they have used the same
specific software and hardware configuration in prior work).
3. To gain concordance in understanding the grading requirements
4. To undertake intra- and inter-reader agreement evaluation. It is
important that the reliability assessment is performed on an
adequate number of cases that takes into account the charac-
teristics of a specific cohort. It is important to assess reliability in
a cross-sectional as well as longitudinal fashion as detection of
change is the most important parameter in clinical trials.
Further an adequate number needs to be included in reliability
exercises. During the course of the study, inter-reader concor-
dance exercises should be conducted in an on-going manner. In
general the concepts provided here are outlines in the FDA draft
Guidance entitled “Guidance for Industry on Standards for
Clinical Trial Imaging Endpoints92”.
Knee X-ray specific QC
All knee acquisitions need to use a weight bearing fixed flexion
technique, such as fixed flexion or mLS using a positioning device,
such as the Synaflexer equipped with a calibration standard.
There are a number of factors that need to be considered when
evaluating QC for the knee for both eligibility and efficacy:
1. Is there a single knee imaged per radiograph?
2. Are there calibration beads visible?
a. If not JSW cannot be measured
3. Is the knee in the center of the image?
4. Is there minimal rotation (as evaluated by the patellar position)
5. Is the cortex of the medial tibial plateau sharply delineated in its
entirety
a. Some studies suggest inter-margin distance should be 
1.5 mm
6. Does the inter-margin distance vary between visits?
7. Is anatomical coverage 10 cm above and below the tibial spines?
a. Is there correct alignment (defined by anatomical angle axis)
and is it within protocol limits? (if needed for eligibility)
Knee MRI specific QC
After checking the identification of the participant, sequences
relative to the study protocol should be verified. Then, for each
acquisition, image quality should be validated according to the
following criteria:
- Integrity of all sequences (legible and complete sequences)
- Position of the knee in the FOV
- Contrast of tissues
- Motion artifact
- Other artifacts
- At follow-up visit, it is important to check that it is the same
knee as the knee that was imaged at baseline and that the im-
aging parameters are identical.
- QC results should be documented and archived in the partici-
pant's file.
- According to the study, the rejection of a QC may necessitate the
new submission of images, resumption of the MRI or dismissal
of the participant.
Measurement methods
Semi-quantitative radiographic scores
Radiography can be used to visualize osteophytes, subchondral
cysts and sclerosis, alignment, and cortical irregularities. Radiog-
raphy does not directly visualize articular cartilage, menisci, cruciate
or collateral ligaments, synovium, bursae, or the periarticular mus-
cles and ligaments, all of which can be better assessed by MRI.
Moreover, JSW is nonspecific, as it includes not only articular carti-
lage but meniscal tissue. Radiography is also limited by providing
only a 2D projection of the knee, which has a complex, 3D structure.
In clinical trials, the PA radiograph is useful for supporting a
diagnosis of knee OA. Although the KLG system is over 60 years old, it
continues to be a useful method for eligibility screening for clinical
trials108,109. According to the KLG the presence of a definite osteo-
phyte is consistent with the diagnosis of OA (KLG 2). The presence of
definite JSN and osteophytes is consistent with moderate OA (KLG 3).
There are more detailed methods for defining individual
radiographic features (including the OARSI atlas) that are useful in
more specifically defining the study population and ascertaining
progression of JSN110e112. Other radiographic scoring methods are
available (including the Spector Atlas113 and Doherty's line drawing
atlas114), however, they are less widely used.
Quantitative radiographic measures
Traditionally the progression of knee OA has been assessed by
measuring changes in JSW (typically in millimeters), i.e., the dis-
tance between the medial femoral condyle and medial tibial
plateau on plain radiographs, as the medial compartment is the
most common site of involvement in knee OA. Among the various
radiographic features of OA, JSW is considered the surrogate for
articular cartilage thickness although it also reflects the integrity of
the meniscus. Methods of measurement of JSW can be either
manual using callipers or a simple graduated ruler and a micro-
metric eyepiece, or semi-automated using computer software115.
MRI semi-quantitative scores
Semi-quantitative assessment of the joints by expert in-
terpreters of MRI data has increased our understanding of the
natural history of this complex disease. These are usually in ordinal
scales although for some features they are dichotomous/binary
(e.g., ACL disruption). Several reliable and validated semi-
quantitative scoring systems now exist and have been applied to
large-scale, multicenter, cross-sectional and longitudinal observa-
tional epidemiological studies. Semi-quantitative MRI outcome
 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
measures have also been applied in several clinical trials in OA.
These include WORMS30, BLOKS116 and MOAKS31, which have been
extensively reviewed elsewhere recently38.
MRI e quantitative measurements
There is a wide variety of quantitative MRI measurement
methods including33 compositional measures such as (T1rho, T2
(sec) and T2* relaxation time measurements, GAG CEST, Sodium,
Diffusion and DTI, and dGEMRIC)64, as well as quantitative mea-
surements of cartilage (such as cartilage thickness, denuded area117
and morphological (CALS) score118) or other joint tissues for
quantitative morphological assessment (including 3D statistical
shape models and bone shape and dynamic contrast-enhanced and
non-contrast MRI synovitis). There are a range of methods used for
quantitative assessment of cartilage morphometry (volume (typi-
cally measured in mm3), thickness (typically measured in mm),
denuded area (typically measured in mm2)) ranging from fully
manual to fully automatic119e126. There is no clear data demon-
strating measurement performance superiority of one method over
another after direct comparison127. Similarly, quantification of the
knee bone marrow lesions, osteophytes, meniscus, muscle volume,
bone area and shape128,129, synovitis and synovial fluid are possible
and are demonstrating some important findings130,131.
Measurement performance (reliability, responsiveness,
validity)
Reliability of radiography in the knee joint
Reliability assessments for observer based measures include both
inter and intra-observer measures as provided in ICCs or kappa
scores. There are different review articles available which provide
relatively consistent numbers132 with intra-reader variability of
about 0.95e0.97 (ICC) and only slightly lower inter-reader vari-
ability of about 0.93e0.95 (ICC) for JSW. However, these values are
highly influenced by the quality and grades of the radiographs being
evaluated, so should not be used as a gold standard, unless the same
set of images are being used to compare another set of radiologists.
The radiographic method did not affect the reliability132, however
the positioning of the knee joint plays a very important role.
Reliability of MRI in the knee joint
Concerning MRI the literature gets much more inconsistent and
the reliability is in part very much dependent on what one is
measuring and how one is doing this. To this end we would assert
that reliability is study specific and should be tested and reported
for each endeavour. The inter-reader and intra-reader ICCs are high
for quantitative as well as semi-quantitative morphological MRI
(range 0.80e0.94)133. However, the results are very much depen-
dent on the specific anatomical area within the knee joint which is
analysed with an increased range of ICCs e.g., for the quantitative
assessment of the synovium (range 0.61e1.00) or the semi-
quantitative assessment of the meniscus (range 0.49e1.00)133. For
quantitative cartilage morphometry the ICC ranges in various
studies for intra- and inter-rater reliability in between 0.86 and
0.95133. In compositional MR techniques, the range of the reliability
is higher and even more heterogeneous134. So far there is no sys-
tematic review on reliability issues in compositional MRI available.
Existing studies are reporting on ICC values for T2 mapping in be-
tween 0.61 and 0.9899,135. The reproducibility of T1-r values for
example was higher in the thicker patellar cartilage (ICC range,
0.86e0.93) than in the femoro-tibial joint with a large range of ICC
values in between 0.20 and 0.84.
707
Responsiveness of radiography in the knee joint
The responsiveness of radiographic JSW can be assessed by
calculating the standardized response mean (SRM)132. The SRM is
one of several available effect size indices used to gauge the
responsiveness of scales to clinical change. The overall pooled SRM
was 0.33. Responsiveness of change in JSW measurement was
improved substantially in studies of greater than 2 years duration
(0.57). Further stratifying this result in studies of greater than
2 years duration, radiographs obtained with the knee in a flexed
position yielded an SRM of 0.71132 although, like MRI this is likely
study specific.
Responsiveness of cartilage morphometry: thickness, area, volume
in the knee joint
The pooled SRM for quantitative measures of cartilage
morphometry for the medial tibiofemoral joint was reported
as 0.86133. For the quantitative analysis, SRMs are negative
because the quantitative value, indicating a loss of cartilage, goes
down. The pooled SRM for the semi-quantitative measurement of
cartilage morphology for the medial tibiofemoral joint was 0.55133.
For the semi-quantitative analysis, SRMs indicating a loss in carti-
lage are positive (increase in score).
Responsiveness of compositional measures: (T2, T1-rho, dGEMRIC,
Na) in the knee joint
The responsiveness as an instrument's ability to detect change
over time has so far not been assessed in the literature for
compositional MR techniques. Nevertheless for quantitative T2
mapping, recent reports from the incident and normal cohort of the
OAI support the use of cartilage T2 as an early marker of cartilage
degeneration, demonstrating responsiveness to cartilage change
over time136. Comparably also in other studies T2 mapping is seen
to provide stable and subtle longitudinal data137. A comparable
promising detection of longitudinal results can be found for
dGEMRIC138. For T1rho a number of studies also showed
responsiveness139.
For all other compositional sequences, the availability of
possible responsiveness to change is not yet available in literature.
Additionally the possibility of a real quantification of most of the
compositional MR techniques enables them to reliably detect
changes of cartilage micro-structure.
Validity of radiography in the knee joint
The validity or accuracy is the degree to which an instrument
measures what it is supposed to measure is assessed here. In OA
progression can be assessed validly. For radiography, there is some
evidence for construct and predictive validity, with good evidence
for reliability of metric measurement of JSW140. It is important to
note that radiographic assessment concurrent and predictive val-
idity for clinical outcomes is weak and that JSW reflects a number of
structures not just hyaline articular cartilage3.
Validity of MRI in the knee joint
Concurrent validity of MRI in OA has been examined compared
to different parameters131. The relation of bone marrow lesions,
synovitis and effusion to pain was moderate to strong. There was a
weak or no relation of cartilage damage or meniscal tears to pain.
The relation of cartilage morphology to radiographic OA and
radiographic joint space was inconsistent. There was a higher fre-
quency of meniscal tears and extrusion, synovitis and other
 708

recommend.

i. Site training

terminology141.
visit studies
iv. Adjudication
i. Central reading
content of cartilage134.

Experience with radiography

draft guidance is as follows:

when deciding on study end-point.

groups or different practice settings).

Research recommendations (Table V)

2. Improve the X-ray (halt progression)
Recommendations for trials (Table IV)

ii. Multiple readers of the same images

1. Normalise the X-ray (reverse progression)

mining the complications of disease or treatment.

ii. Image data-quality checks (completeness, correctness)

of progression) accompanied by symptom improvement.

Regarding image analysis process we would recommend:

vices and Biological Products Intended for the Treatment of OA

defined by the FDA Clinical Development Programs for Drugs, De-
and detecting changes e.g., in the collagen matrix or the GAG

communication we would advocate careful and correct use of

For image acquisition (whether radiography or MRI) we would

We support the continued use of JSW as one option for assessing

3. Slow JSW loss by at least a pre-specified amount (slow the rate
ability to predict total knee replacement (TKR), change in symp-

that authors describe in sufficient detail the methods of

in clinical trials. For example imaging may play a role in: subject
Trial recommendations are contingent on the goals of imaging

The current hierarchy of claims for structural outcome as

others ability to replicate it. To facilitate clear and transparent
In reporting imaging studies in OA we strongly advocate
iii. Individual patient imagesgrouped for processing for multi-
selection; or monitoring progression/improvement; or deter-

design considerations for phase III trials or phase IV (patient sub-

decision making studies may include proof of concept studies,
internal decision making or for definitive proof of efficacy. Internal
logical samples with very promising results in exactly quantifying
toms, radiographic progression as well as MRI progression131.

acquisition, image processing/analytic methods, QA and QC

Further to this the role of imaging in the clinical trial maybe for

structural OA change, taking all the previous points into account

stronger. Predictive validity of MRI in OA has been examined for
to other constructs including histology and arthroscopy was

enhance both the readers ability to appraise the work and

steps in the methods section of the manuscript. This will
Compositional MR techniques have been validated against histo-
features in persons with radiographic OA. The relation of cartilage

Table IV
Summary of recommendations for clinical trials

Process Strength of recommendation

(range 0e100)*

Radiography3
The use of fluoroscopic positioning and semi-flexed views improves responsiveness, although it is acknowledged that access to fluoroscopic facilities is restricted. 74.6
Studies will generally need to be at least 12 and more likely 24 months duration.
The IMD of the tibial plateau should not change between visits, and ideally should be no more than 1.5 mm but preferably below 1 mm. 58.4
50.5
D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715

It is advisable to “enrich” a knee OA study population to increase the rate of JSW loss.
Automated methods for assessing parameters of JSW offer promise of improved precision and therefore improved responsiveness. 65.5
MRI3
For assessing MRI cartilage morphometry in knee OA, there is some evidence for construct and predictive validity, with good evidence for reliability and responsiveness. 83.5
Using MRI it is possible to accurately and feasibly measure change in cartilage morphometry over 12 months for knee OA and we recommend the use of MRI for
assessing cartilage morphometry in trials of OA structure modification.
It is possible to “enrich” a knee OA study population with MRI outcomes in order to increase the rate of cartilage loss, for example, by including higher KL grade, 73.4
maligned knees, BMLs or meniscal extrusion at baseline.
In terms of correlations with concurrent symptoms, there is a weak association between progression of cartilage loss and increasing symptoms. There is little information 79.5
on how change in cartilage parameters during the course of a study reflects post-study change in symptoms. There is some predictive validity with progression of
cartilage loss predicting subsequent total joint replacement.
More information is required on the performance metrics of MRI semi-quantitative and compositional measures of cartilage. There may be a role for semi-quantitative 76.9
assessments for assessing focal cartilage defects.
Structure modification should be considered in a broader context than that of cartilage alone. Since MRI has the capacity to image the other tissues, further work is 87.4
needed on the quantification and predictive validity of non-cartilage MRI pathologies The performance metrics of non-cartilage MRI features have not been extensively
studied but there is a rapidly emerging literature in this field.
Final recommendations
Radiographic JSW is still a recommended option for trials of structure modification, with the understanding that the construct represents a number of pathologies and 60.9
trial duration may be long.
MRI is now recommended for clinical trials in terms of cartilage morphometry assessment. 81.7
*
Mean response from the 13 persons on the working group who responded to the survey. Scale ranged from 0 (don't Recommend) to 100 (strongly recommend).
Table V
Summary of research recommendations

Process Strength of recommendation

(range 0e100)*

Conventional radiography
With conventional radiography a special focus and further work should be on:
JSW: Measurement performance (reliability and responsiveness) dependency on projection technique, dependency on axial loading 72.2
JSW: Measurement performance (reliability and responsiveness) in the lateral femorotibial compartment 60.3
Bony changes (osteophytes, erosions, cysts, attrition): dependency of reproducibility on projection technique 51.8
Further investigation of the correlation of MRI cartilage thickness change, MRI meniscus change (position and size), and radiographic JSW change in different 74.1
stages of radiographic disease (i.e., healthy, pre-radiographic, early radiographic, advanced radiographic should be analysed). This knowledge is required to
achieve a paradigm shift from radiographic JSW to MR measurements.
What is their reliability (test-retest precision) and sensitivity to change across different (radiographic) stages of knee OA (i.e., from healthy, to pre-radiographic, 71.3

D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715

early radiographic, advanced radiographic).
What is their ability to predict the onset of pain in non-symptomatic subjects with and without radiographic change. 62.7
What is their ability to predict the onset of functional limitations in non-symptomatic subjects with and without radiographic change. 72
What is their ability to predict the onset of radiographic or other structural changes in subjects without definite radiographic change (KLG 0 or 1, with and 70.2
without symptoms).
What is their ability to predict a worsening in pain (progression) in subjects with mild to moderate pain levels (with and without radiographic change). 69.5
What is their ability to predict an increase in functional limitations (progression) in subjects with mild to moderate limitations (with and without 69.8
radiographic change).
What is their ability to predict the progression of radiographic (or other structural) changes in subjects with definite radiographic change (KLG  2) with 73.5
and without symptoms
MRI
Although there exists a comparison between quantitative cartilage volume/thickness with semi-quantitative scoring further comparison with scoring systems 55.4
such as e WORMS, BLOKS, KOSS and MOAKS should be performed.
Further investigation should also be done for compositional MR methods such as T2 mapping and gagCEST to find correlations between macro-and microstructure 69.6
changes in the knee joint.
A correlation with clinical scores, clinical importance of individual variables in the semi-quantitative scoring systems as well as with quantitative values of the 78.8
compositional MR techniques should be performed.
Develop and appraise methods that will allow qualitative (semi-quantitative) measures to be quantitatively assessed as well as semi-automated to fully automated: 77.8
cartilage; bone (osteophytes, erosions, cysts, attrition, fractures, BML); menisci (surface, morphology, thickness, volume).
Evaluation of age dependence of structural changes (cartilage thickness and others) due to mechanical stress in different age groups (20e30, 30e40, 40e50 etc) 63.5
Improved detection, quantification and measurement performance of structural, compositional changes of hyaline and fibrocartilage (GAG, fibre orientation, water content) 73.8
with the use of: T2 relaxation time measurements, T2* relaxation time measurements, gagCEST, T1rho, DWI, Sodium imaging, dGEMRIC
Development of semi-quantitative and quantitative as well as fully automated evaluation of inflammatory changes of the synovium, with the use of synovitis score, 81.8
non-contrast-enhanced sequences, dynamic contrast-enhanced perfusion studies, development of reproducible, standardized quantification of perfusion
Evaluation of “inflammatory” changes in the subchondral bone marrow with the use of: DWI, T2*, dynamic contrast-enhanced perfusion studies. 72.8
Evaluation of muscle cross sectional areas (thigh), inter-muscular adipose tissue, intramuscular adipose tissue, and other imaging measures of muscle quality (in context of 75.2
muscle strength measurements) which are thought to play a potential role in the onset and progression of knee OA, but in particular for intra-muscular adipose tissue and
muscle quality, it is currently unclear which imaging technique is most accurate and precise for its evaluation.
Clinical validation of MRI methods and techniques in general: 91.2
What is their ability to predict virtual knee replacement endpoints?
What is their ability to predict real knee replacement?
How sensitive are different imaging measures to drug treatment (with different modes of action: i.e., affecting cartilage resorption, cartilage anabolism, bone resorption, other)
once DMOADs with these modes of action become available?
*
Mean response from the 13 persons on the working group who responded to the survey. Results of participants who were conflicted are not included. Strength of recommendation scale ranged from 0 (don't Recommend) to
100 (strongly recommend).

709
710 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
Summary and conclusion
The goals of imaging the knee in clinical trials can include
subject selection, monitoring disease progression and treatment
effect, and/or identifying complications of the disease or the
treatment. For acquisition we have provided guidance with
regards both plain radiographic and MRI protocols. MRI protocols
must be optimized around the scientific objectives and unique
practical constraints of the specific study in question, particularly
with respect to the study centers. This manuscript includes a
number of recommendations for clinical trials that we would
advise anyone planning on using imaging in knee OA trial to
follow.
Author contributions
All authors were involved in collecting data, reviewing the
literature and drafting the article or revising it critically for
important intellectual content, and all authors approved the final
version to be published.
Disclosure
The comments and editorial expressed herein represent those of
the author/s and do not reflect those of any official scientific role
or institution that the author/s may be hold or be affiliated with.
The corresponding author had full access to all the data in the
study and had final responsibility for the decision to submit for
publication.
Conflict of interest
1. David Hunter has royalties from DJO for a patent granted for a
patellofemoral buttress. Has grants from the NIH, Australian
Research Council and NHMRC. Associate Editor for Osteoar-
thritis and Cartilage.
2. Roy D. Altman is a consultant to Cytori, Ferring, Iroko,
Johnson and Johnson, Novartis, Oletec, Pfizer, Q Med, Rotta-
pharma, Strategic Sciences and Technologies, and Teva.
3. Flavia Cicuttini e Receives honoraria from Jansen. Has
grants from the NHMRC for Statin RCT, Zolendronic Acid
RCT and ACL study. Holds a volunteered paid position in
Repatriation Medical Authority as a Medical Officer and as
an Associate Editor in the BMC Musculoskeletal Journal,
ART Journal.
4. Michel D. Crema e has shares in Boston Imaging Core Lab,
LLC a company providing image assessment services to
academia and the pharmaceutical and medical device
industry.
5. Jeffrey Duryea e has a grant from NIH/NIAMS for a Quanti-
tative MRI analysis method for longitudinal assessment of
knee OA.
6. Felix Eckstein receives honoraria from Mariel Therpeutics,
MerckSerono, and Medtronic. Is CEO and co-owner of
Chondrometrics Gmbh, a quantitative image analysis
company and is involved in, clinical trials including FGF 18
from MerckSerono. Holds a voluntary unpaid position in
Annals of Anatomy, Cell Tissue Organs as an associate ed-
itor and in Osteoarthritis & Cartilage as part of the editorial
board.
7. Ali Guermazi receives honoraria from Genzyme, TissueGene,
OrthoTrophix and Merck Serono. He has shares in Boston
Imaging Core Lab, LLC. Deputy Editor of Radiology.
8. Richard Kijowski has grants from the National Institute of
Health for NIAMS/Rapid MRI for Assessing Joint
Degeneration, for the NAIMS/Cartilage Contact and for Early
degeneration after ACL Injury.
9. Thomas M. Link receives honoraria from Springer. He has
grants from the National Institutes of Health (NIAMS and
NIBIB) and General Electric for the MAVRIC protocol and
Insightec Exablate protocols. Holds a paid position in Current
Radiology Reports as Editor-in-Chief. He is on the editorial
boards of Osteoarthritis and Cartilage, Radiology, AJR and
Skeletal Radiology. He is also section editor for European
Radiology.
10. Johanne Martel-Pelletier & Jean-Pierre Pelletier e receives
honoraria from ArthroLab/ArthroVision, Bioiberica, Elanco,
Ferring, Merck, Pfizer, Servier, TRB Chemedica. Has grants
from The Arthritis Society (as co-investigators), Canadian
Institutes of Health Research (CIHR) (as co-investigator). Has
investment in non-medical industry at ArthroLab/ArthroVi-
sion and is a Shareholder. Holds a volunteered lecturer po-
sition at the ESCEO13-IOF European Congress on
Osteoporosis and Osteoarthritis (travel grant).
11. Colin G. Miller- Senior vice president at Bioclinica.
12. Timothy J. Mosher receives honoraria from Medical Met-
rics, Depuy Orthopaedics, eImage, Pirima Healthcare, has
shares in Johnson & Johnson, holds a voluntary paid po-
sition in Radiological Society of North America (RSNA),
Academy for Radiology Research and the Society for Skel-
etal Radiology.
13. R. Elena Ochoa Albíztegui e no conflict of interest to disclose.
14. Charles Peterfy e has shares in Spire Sciences, Inc., which
provides Image analysis for clinical trials to multiple phar-
maceutical and medical device companies. He holds a vol-
unteered paid position in The International Society for
Musculoskeletal Imaging in Rheumatology as treasurer.
15. Jean-Pierre Raynauld receives honoraria as a consultant for
ArthroLab/ArthroVision.
16. Frank Roemer e has shares in Boston Imaging Core Lab, LLC, a
company providing image assessment services to academia
and the pharmaceutical and medical device industry. Asso-
ciate Editor for Osteoarthritis and Cartilage.
17. Saara M. Totterman-employee and shareholder of Qmetrics
Technologies.
18. Garry E. Gold receives honoraria as a consultant from Zim-
mer Inc., Isto, Inc. and Boston Scientific. Has a grant from GE
Healthcare for improved MR Systems and one from the NIH
for advanced MRI of Osteoarthritis.
Acknowledgments
Dr Hunter is supported by an NHMRC Health Practitioner
Fellowship. Dr Gold is supported by an NIH K24 Grant (AR062068)
and RO1 (EB002524).
We acknowledge the support and contribution of the entire
OARSI Imaging Clinical Trials Recommendations working group
which includes: Ali Guermazi, Bernard Dardzinski, Charles Peterfy,
Colin Miller, Diann Stern, Emmanuel Maheu, Felix Eckstein, Flavia
Cicuttini, Frank Roemer, Garry Gold, John Carino, Hollis Potter, Ida
Haugen, Jean-Pierre Pelletier, Jean-Pierre Raynauld, Jeffrey Duryea,
Joanne Jordan, Johanne Martel-Pelletier, Kim Bennell, Lee Simon,
Marc Hochberg, Margreet Kloppenburg, Michael Nevitt, Michel
Crema, Mike Bowes, Nancy Lane, Philip Conaghan, Richard Kijow-
ski, Roy Altman, Saara Totterman, Siegfried Trattnig, Souhil Zaim,
Thomas Link, Thomas Schnitzer, Tim McAlindon, Tim Mosher, Yves
Henrotin, Frank Roemer, Goetz Welsch, Graeme Jones, Jordan
Renner, Marie-Pierre Hellio Le Graverand and Nigel Arden. We also
acknowledge the administrative support of Diann Stern.
 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
OARSI gratefully acknowledges support to defer in part the cost
of printing of the Clinical Trial Recommendations from Abbvie,
BioClinica, Boston Imaging Core Lab, and Flexion. The funding
sources for printing had no role in the outcome of this manuscript.
References
1. Reginster JY, Badurski J, Bellamy N, Bensen W, Chapurlat R,
Chevalier X, et al. Efficacy and safety of strontium ranelate in
the treatment of knee osteoarthritis: results of a double-
blind, randomised placebo-controlled trial. Ann Rheum Dis
2013 Feb;72(2):179e86.
2. Matthews GL, Hunter DJ. Emerging drugs for osteoarthritis.
Expert Opin Emerg Drugs 2011;16(3):479e91.
3. Conaghan PG, Hunter DJ, Maillefert JF, Reichmann WM,
Losina E. Summary and recommendations of the OARSI FDA
osteoarthritis Assessment of Structural Change Working
Group. Osteoarthritis Cartilage 2011;19(5):606e10.
4. Abramson SB, Berenbaum F, Hochberg MC, Moskowitz RW.
Introduction to OARSI FDA initiative OAC special edition.
Osteoarthritis Cartilage 2011 May;19(5):475e7.
5. Altman R, Brandt K, Hochberg M, Moskowitz R, Bellamy N,
Bloch DA, et al. Design and conduct of clinical trials in pa-
tients with osteoarthritis: recommendations from a task
force of the Osteoarthritis Research Society. Results from a
workshop. Osteoarthritis Cartilage 1996 Dec;4(4):217e43.
6. Kobayashi N, Inaba Y, Tateishi U, Yukizawa Y, Ike H, Inoue T,
et al. New application of 18F-fluoride PET for the detection of
bone remodeling in early-stage osteoarthritis of the hip. Clin
Nucl Med 2013 Oct;38(10):e379e83.
7. Dieppe P, Altman RD, Buckwalter JA, Felson DT, Hascall V,
Lomander LS, et al. Standardization of methods used to assess
the progression of osteoarthritis of the hip or knee joints. In:
Keuttner KE, Goldberg VM, Eds. Osteoarthritic Disorders.
Rosemont, IL: American Academy of Orthopaedic Surgeons;
1995:481e96.
8. Ravaud P, Chastang C, Auleley GR, Giraudeau B, Royant V,
Amor B, et al. Assessment of joint space width in patients
with osteoarthritis of the knee: a comparison of 4 measuring
instruments. J Rheumatol 1996 Oct;23(10):1749e55
[Comparative Study].
9. Niinimaki T, Ojala R, Niinimaki J, Leppilahti J. The standing
fixed flexion view detects narrowing of the joint space better
than the standing extended view in patients with moderate
osteoarthritis of the knee. Acta Orthop 2010 Jun;81(3):
344e6.
10. Buckland-Wright JC, Macfarlane DG, Williams SA, Ward RJ.
Accuracy and precision of joint space width measurements in
standard and macroradiographs of osteoarthritic knees. Ann
Rheum Dis 1995 Nov;54(11):872e80.
11. Brandt KD, Mazzuca SA, Conrozier T, Dacre JE, Peterfy CG,
Provvedini D, et al. Which is the best radiographic protocol
for a clinical trial of a structure modifying drug in patients
with knee osteoarthritis? J Rheumatol 2002 Jun;29(6):
1308e20 [Research Support, Non-U.S. Gov't Review].
12. Gossec L, Jordan JM, Mazzuca SA, Lam MA, Suarez-
Almazor ME, Renner JB, et al. Comparative evaluation of three
semi-quantitative radiographic grading techniques for knee
osteoarthritis in terms of validity and reproducibility in 1759
X-rays: report of the OARSI-OMERACT task force. Osteoar-
thritis Cartilage 2008 Jul;16(7):742e8.
13. Vignon E, Piperno M, Le Graverand MP, Mazzuca SA,
Brandt KD, Mathieu P, et al. Measurement of radiographic
joint space width in the tibiofemoral compartment of the
osteoarthritic knee: comparison of standing anteroposterior
711
and Lyon schuss views. Arthritis Rheum 2003 Feb;48(2):
378e84.
14. Kothari M, Guermazi A, von Ingersleben G, Miaux Y,
Sieffert M, Block JE, et al. Fixed-flexion radiography of the
knee provides reproducible joint space width measurements
in osteoarthritis. Eur Radiol 2004 Sep;14(9):1568e73.
15. Peterfy C, Li J, Zaim S, Duryea J, Lynch J, Miaux Y, et al.
Comparison of fixed-flexion positioning with fluoroscopic
semi-flexed positioning for quantifying radiographic joint-
space width in the knee: test-retest reproducibility. Skelet
Radiol 2003 Mar;32(3):128e32.
16. Hellio le Graverand MP, Clemmer RS, Redifer P, Brunell RM,
Hayes CW, Brandt KD, et al. A 2-year randomised, double-
blind, placebo-controlled, multicentre study of oral selective
iNOS inhibitor, cindunistat (SD-6010), in patients with
symptomatic osteoarthritis of the knee. Ann Rheum Dis 2013
Feb;72(2):187e95.
17. Hellio Le Graverand MP, Clemmer RS, Brunell RM, Hayes CW,
Miller CG, Vignon E. Considerations when designing a
disease-modifying osteoarthritis drug (DMOAD) trial using
radiography. Semin Arthritis Rheum 2013 Aug;43(1):1e8.
18. Kraus VB, Vail TP, Worrell T, McDaniel G. A comparative
assessment of alignment angle of the knee by radiographic
and physical examination methods. Arthritis Rheum 2005
Jun;52(6):1730e5.
19. Szebenyi B, Dieppe PA, Buckland-Wright JC. Radio-anatomic
position for the lateral radiographic view of the human
patello-femoral joint. Surg Radiol Anat: SRA 1995;17(1):
79e81.
20. Peterfy C, Woodworth T, Altman R. Workshop for consensus
on Osteoarthritis Imaging: MRI of the knee. Osteoarthritis
Cartilage 2006;14(Suppl 1):44e5.
21. Sofat N, Ejindu V, Kiely P. What makes osteoarthritis painful?
The evidence for local and central pain processing. Rheu-
matology (Oxford) 2011 Dec;50(12):2157e65.
22. Hunter DJ, Guermazi A, Roemer F, Zhang Y, Neogi T. Struc-
tural correlates of pain in joints with osteoarthritis. Osteo-
arthritis Cartilage 2013 Sep;21(9):1170e8.
23. Hill CL, Gale DG, Chaisson CE, Skinner K, Kazis L, Gale ME,
et al. Knee effusions, popliteal cysts, and synovial thickening:
association with knee pain in osteoarthritis. J Rheumatol
2001 Jun;28(6):1330e7.
24. Guermazi A, Zaim S, Taouli B, Miaux Y, Peterfy CG,
Genant HG. MR findings in knee osteoarthritis. Eur Radiol
2003 Jun;13(6):1370e86.
25. Hayes CW, Jamadar DA, Welch GW, Jannausch ML,
Lachance LL, Capul DC, et al. Osteoarthritis of the knee:
comparison of MR imaging findings with radiographic
severity measurements and pain in middle-aged women.
Radiology 2005 Dec;237(3):998e1007.
26. Sowers MF, Hayes C, Jamadar D, Capul D, Lachance L,
Jannausch M, et al. Magnetic resonance-detected sub-
chondral bone marrow and cartilage defect characteristics
associated with pain and X-ray-defined knee osteoarthritis.
Osteoarthritis Cartilage 2003 Jun;11(6):387e93.
27. Link TM, Steinbach LS, Ghosh S, Ries M, Lu Y, Lane N, et al.
Osteoarthritis: MR imaging findings in different stages of
disease and correlation with clinical findings. Radiology 2003
Feb;226(2):373e81.
28. Wluka AE, Ding C, Jones G, Cicuttini FM. The clinical corre-
lates of articular cartilage defects in symptomatic knee
osteoarthritis: a prospective study. Rheumatology 2005
Oct;44(10):1311e6.
29. Wildi LM, Raynauld JP, Martel-Pelletier J, Abram F, Dorais M,
Pelletier JP. Relationship between bone marrow lesions,
712 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
cartilage loss and pain in knee osteoarthritis: results from a
randomised controlled clinical trial using MRI. Ann Rheum
Dis 2010 Dec;69(12):2118e24.
30. Peterfy CG, Guermazi A, Zaim S, Tirman PF, Miaux Y, White D,
et al. Whole-Organ Magnetic Resonance Imaging Score
(WORMS) of the knee in osteoarthritis. Osteoarthritis Carti-
lage 2004 Mar;12(3):177e90.
31. Hunter DJ, Guermazi A, Lo GH, Grainger AJ, Conaghan PG,
Boudreau RM, et al. Evolution of semi-quantitative whole
joint assessment of knee OA: MOAKS (MRI Osteoarthritis
Knee Score). Osteoarthritis Cartilage 2011;19(8):990e1002.
32. Eckstein F, Cicuttini F, Raynauld JP, Waterton JC, Peterfy C.
Magnetic resonance imaging (MRI) of articular cartilage in
knee osteoarthritis (OA): morphological assessment. Osteo-
arthritis Cartilage 2006;14(Suppl A):A46e75.
33. Peterfy CG, Gold G, Eckstein F, Cicuttini F, Dardzinski B,
Stevens R. MRI protocols for whole-organ assessment of the
knee in osteoarthritis. Osteoarthritis Cartilage 2006;14(Suppl
A):A95eA111.
34. Hochberg MC, Lawrence RC, Everett DF, Cornoni-Huntley J.
Epidemiologic associations of pain in osteoarthritis of the
knee: data from the National Health and Nutrition Exami-
nation Survey and the National Health and Nutrition
Examination-I Epidemiologic Follow-up Survey. Semin
Arthritis Rheum 1989 May;18(4 Suppl 2):4e9.
35. Hirasawa Y, Okajima S, Ohta M, Tokioka T. Nerve distribution
to the human knee joint: anatomical and immunohisto-
chemical study. Int Orthop 2000;24(1):1e4.
36. Felson DT, Chaisson CE, Hill CL, Totterman SM, Gale ME,
Skinner KM, et al. The association of bone marrow lesions
with pain in knee osteoarthritis. Ann Intern Med 2001 Apr
3;134(7):541e9.
37. Conaghan PG, Felson D, Gold G, Lohmander S, Totterman S,
Altman R. MRI and non-cartilaginous structures in knee
osteoarthritis. Osteoarthritis Cartilage 2006;14(Suppl A):
A87e94.
38. Guermazi A, Roemer FW, Haugen IK, Crema MD,
Hayashi D. MRI-based semiquantitative scoring of joint
pathology in osteoarthritis. Nat Rev Rheumatol 2013
Apr;9(4):236e51.
39. Marshall KW, Mikulis DJ, Guthrie BM. Quantitation of artic-
ular cartilage using magnetic resonance imaging and three-
dimensional reconstruction. J Orthop Res 1995 Nov;13(6):
814e23.
40. Peterfy CG, van Dijke CF, Janzen DL, Gluer CC, Namba R,
Majumdar S, et al. Quantification of articular cartilage in the
knee with pulsed saturation transfer subtraction and fat-
suppressed MR imaging: optimization and validation. Radi-
ology 1994 Aug;192(2):485e91.
41. Recht MP, Kramer J, Marcelis S, Pathria MN, Trudell D,
Haghighi P, et al. Abnormalities of articular cartilage in the
knee: analysis of available MR techniques. Radiology 1993
May;187(2):473e8.
42. Peterfy CG, Schneider E, Nevitt M. The osteoarthritis initia-
tive: report on the design rationale for the magnetic reso-
nance imaging protocol for the knee. Osteoarthritis Cartilage
2008 Dec;16(12):1433e41.
43. Crema MD, Roemer FW, Marra MD, Burstein D, Gold GE,
Eckstein F, et al. Articular cartilage in the knee: current MR
imaging techniques and applications in clinical practice and
research. Radiographics 2011 JaneFeb;31(1):37e61.
44. Crema MD, Nogueira-Barbosa MH, Roemer FW, Marra MD,
Niu J, Chagas-Neto FA, et al. Three-dimensional turbo spin-
echo magnetic resonance imaging (MRI) and semi-
quantitative assessment of knee osteoarthritis: comparison
with two-dimensional routine MRI. Osteoarthritis Cartilage
2013 Mar;21(3):428e33.
45. Recht MP, Piraino DW, Paletta GA, Schils JP, Belhobek GH.
Accuracy of fat-suppressed three-dimensional spoiled
gradient-echo FLASH MR imaging in the detection of patel-
lofemoral articular cartilage abnormalities. Radiology 1996
Jan;198(1):209e12.
46. Disler DG, McCauley TR, Wirth CR, Fuchs MD. Detection of
knee hyaline cartilage defects using fat-suppressed three-
dimensional spoiled gradient-echo MR imaging: comparison
with standard MR imaging and correlation with arthroscopy.
AJR Am J Roentgenol 1995 Aug;165(2):377e82.
47. Hardy PA, Recht MP, Piraino DW. Fat suppressed MRI of
articular cartilage with a spatial-spectral excitation pulse.
J Magn Reson Imaging 1998 NoveDec;8(6):1279e87.
48. Yoshioka H, Alley M, Steines D, Stevens K, Rubesova E,
Genovese M, et al. Imaging of the articular cartilage in oste-
oarthritis of the knee joint: 3D spatial-spectral spoiled
gradient-echo vs. fat-suppressed 3D spoiled gradient-echo
MR imaging. J Magn Reson Imaging 2003 Jul;18(1):66e71.
49. Siepmann DB, McGovern J, Brittain JH, Reeder SB. High-res-
olution 3D cartilage imaging with IDEAL SPGR at 3 T. AJR Am J
Roentgenol 2007 Dec;189(6):1510e5.
50. Chen CA, Lu W, John CT, Hargreaves BA, Reeder SB, Delp SL,
et al. Multiecho IDEAL gradient-echo water-fat separation for
rapid assessment of cartilage volume at 1.5 T: initial experi-
ence. Radiology 2009 Aug;252(2):561e7.
51. Kijowski R, Tuite M, Passov L, Shimakawa A, Yu H, Reeder SB.
Cartilage imaging at 3.0T with gradient refocused acquisition
in the steady-state (GRASS) and IDEAL fat-water separation.
J Magn Reson Imaging: JMRI 2008 Jul;28(1):167e74
[Research Support, N.I.H., Extramural Research Support, Non-
U.S. Gov't Research Support, U.S. Gov't, Non-P.H.S.].
52. Disler DG, McCauley TR, Kelman CG, Fuchs MD, Ratner LM,
Wirth CR, et al. Fat-suppressed three-dimensional spoiled
gradient-echo MR imaging of hyaline cartilage defects in the
knee: comparison with standard MR imaging and arthros-
copy. AJR Am J Roentgenol 1996 Jul;167(1):127e32.
53. Mosher TJ, Pruett SW. Magnetic resonance imaging of su-
perficial cartilage lesions: role of contrast in lesion detection.
J Magn Reson Imaging: JMRI 1999 Aug;10(2):178e82.
54. Gold GE, Fuller SE, Hargreaves BA, Stevens KJ, Beaulieu CF.
Driven equilibrium magnetic resonance imaging of articular
cartilage: initial clinical experience. J Magn Reson Imaging:
JMRI 2005 Apr;21(4):476e81.
55. Hargreaves BA, Gold GE, Lang PK, Conolly SM, Pauly JM,
Bergman G, et al. MR imaging of articular cartilage using
driven equilibrium. Magn Reson Med 1999 Oct;42(4):
695e703.
56. Hardy PA, Recht MP, Piraino D, Thomasson D. Optimization of
a dual echo in the steady state (DESS) free-precession
sequence for imaging cartilage. J Magn Reson Imaging: JMRI
1996 MareApr;6(2):329e35.
57. Self weighted combination of SSFP echoes with a double-
echo steady state (DESS) sequence. In: Deimling M, Jellus V,
Horger W, Haider S, Eds. Proceedings of the 13th Annual
Scientific Meeting of International Society of Magnetic
Resonance in Medicine May, 2005. South Beach, Miami,
Florida.
58. Moriya S, Miki Y, Yokobayashi T, Ishikawa M. Three-dimen-
sional double-echo steady-state (3D-DESS) magnetic reso-
nance imaging of the knee: contrast optimization by
adjusting flip angle. Acta Radiol 2009 Jun;50(5):507e11.
59. Gold GE, Busse RF, Beehler C, Han E, Brau AC, Beatty PJ, et al.
Isotropic MRI of the knee with 3D fast spin-echo extended
 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
echo-train acquisition (XETA): initial experience. AJR Am J
Roentgenol 2007 May;188(5):1287e93 [Comparative Study].
60. Notohamiprodjo M, Horng A, Pietschmann MF, Muller PE,
Horger W, Park J, et al. MRI of the knee at 3T: first clinical
results with an isotropic PDfs-weighted 3D-TSE-sequence.
Invest Radiol 2009 Sep;44(9):585e97.
61. Friedrich KM, Reiter G, Kaiser B, Mayerhofer M, Deimling M,
Jellus V, et al. High-resolution cartilage imaging of the knee at
3T: basic evaluation of modern isotropic 3D MR-sequences.
Eur J Radiol 2011 Jun;78(3):398e405.
62. Chen CA, Kijowski R, Shapiro LM, Tuite MJ, Davis KW,
Klaers JL, et al. Cartilage morphology at 3.0T: assessment of
three-dimensional magnetic resonance imaging techniques.
J Magn Reson Imaging: JMRI 2010 Jul;32(1):173e83
[Research Support, N.I.H., Extramural].
63. Ristow O, Steinbach L, Sabo G, Krug R, Huber M, Rauscher I,
et al. Isotropic 3D fast spin-echo imaging versus standard 2D
imaging at 3.0 T of the kneeeimage quality and diagnostic
performance. Eur Radiol 2009 May;19(5):1263e72.
64. Burstein D, Gray M, Mosher T, Dardzinski B. Measures of
molecular composition and structure in osteoarthritis. Radiol
Clin North Am 2009 Jul;47(4):675e86.
65. Link TM, Stahl R, Woertler K. Cartilage imaging: motivation,
techniques, current and future significance. Eur Radiol 2007
May;17(5):1135e46.
66. Mosher TJ, Dardzinski BJ. Cartilage MRI T2 relaxation time
mapping: overview and applications. Semin Musculoskelet
Radiol 2004 Dec;8(4):355e68.
67. Bashir A, Gray ML, Boutin RD, Burstein D. Glycosaminoglycan
in articular cartilage: in vivo assessment with delayed
Gd(DTPA)(2-)-enhanced MR imaging. Radiology 1997
Nov;205(2):551e8.
68. Bashir A, Gray ML, Burstein D. Gd-DTPA2- as a measure of
cartilage degradation. Magn Reson Med 1996 Nov;36(5):
665e73.
69. Bashir A, Gray ML, Hartke J, Burstein D. Nondestructive im-
aging of human cartilage glycosaminoglycan concentration
by MRI. Magn Reson Med 1999 May;41(5):857e65.
70. Tiderius CJ, Olsson LE, Leander P, Ekberg O, Dahlberg L.
Delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) in
early knee osteoarthritis. Magn Reson Med 2003 Mar;49(3):
488e92.
71. Kaewlai R, Abujudeh H. Nephrogenic systemic fibrosis. AJR
Am J Roentgenol 2012 Jul;199(1):W17e23 [Review].
72. Chrysochou C, Buckley DL, Dark P, Cowie A, Kalra PA. Gado-
linium-enhanced magnetic resonance imaging for renovas-
cular disease and nephrogenic systemic fibrosis: critical
review of the literature and UK experience. J Magn Reson
Imaging: JMRI 2009 Apr;29(4):887e94 [Review].
73. Li X, Benjamin Ma C, Link TM, Castillo DD, Blumenkrantz G,
Lozano J, et al. In vivo T(1rho) and T(2) mapping of articular
cartilage in osteoarthritis of the knee using 3 T MRI. Osteo-
arthritis Cartilage 2007 Jul;15(7):789e97 [Research Support,
N.I.H., Extramural].
74. Stahl R, Luke A, Li X, Carballido-Gamio J, Ma CB, Majumdar S,
et al. T1rho, T2 and focal knee cartilage abnormalities in
physically active and sedentary healthy subjects versus early
OA patientsea 3.0-Tesla MRI study. Eur Radiol 2009
Jan;19(1):132e43 [Comparative Study Evaluation Studies
Research Support, Non-U.S. Gov't].
75. Omoumi P, Mercier GA, Lecouvet F, Simoni P, Vande Berg BC.
CT arthrography, MR arthrography, PET, and scintigraphy in
osteoarthritis. Radiol Clin North Am 2009 Jul;47(4):595e615
[Review].
713
76. Vande Berg BC, Lecouvet FE, Poilvache P, Maldague B,
Malghem J. Spiral CT arthrography of the knee: technique
and value in the assessment of internal derangement of the
knee. Eur Radiol 2002 Jul;12(7):1800e10.
77. Vande Berg BC, Lecouvet FE, Poilvache P, Jamart J, Materne R,
Lengele B, et al. Assessment of knee cartilage in cadavers with
dual-detector spiral CT arthrography and MR imaging. Radi-
ology 2002 Feb;222(2):430e6.
78. Vande Berg BC, Lecouvet FE, Poilvache P, Dubuc JE, Bedat B,
Maldague B, et al. Dual-detector spiral CT arthrography of the
knee: accuracy for detection of meniscal abnormalities and
unstable meniscal tears. Radiology 2000 Sep;216(3):851e7
[Comparative Study].
79. Anderson AE, Ellis BJ, Peters CL, Weiss JA. Cartilage thickness:
factors influencing multidetector CT measurements in a
phantom study. Radiology 2008 Jan;246(1):133e41
[Research Support, Non-U.S. Gov't].
80. Wyler A, Bousson V, Bergot C, Polivka M, Leveque E, Vicaut E,
et al. Hyaline cartilage thickness in radiographically normal
cadaveric hips: comparison of spiral CT arthrographic and
macroscopic measurements. Radiology 2007 Feb;242(2):
441e9 [Comparative Study].
81. Bansal PN, Joshi NS, Entezari V, Grinstaff MW, Snyder BD.
Contrast enhanced computed tomography can predict the
glycosaminoglycan content and biomechanical properties of
articular cartilage. Osteoarthritis Cartilage 2010 Feb;18(2):
184e91.
82. Silvast TS, Jurvelin JS, Lammi MJ, Toyras J. pQCT study on
diffusion and equilibrium distribution of iodinated anionic
contrast agent in human articular cartilageeassociations to
matrix composition and integrity. Osteoarthritis Cartilage
2009 Jan;17(1):26e32 [In Vitro Research Support, Non-U.S.
Gov't].
83. Misra D, Guermazi A, Sieren JP, Lynch J, Torner J, Neogi T,
et al. CT imaging for evaluation of calcium crystal deposition
in the knee: initial experience from the Multicenter Osteo-
arthritis (MOST) study. Osteoarthritis Cartilage 2015
Feb;23(2):244e8.
84. Mazzuca SA, Brandt KD, Dieppe PA, Doherty M, Katz BP,
Lane KA. Effect of alignment of the medial tibial plateau and
X-ray beam on apparent progression of osteoarthritis in the
standing anteroposterior knee radiograph. Arthritis Rheum
2001 Aug;44(8):1786e94.
85. Vignon E, Brandt KD, Mercier C, Hochberg M, Hunter D,
Mazzuca S, et al. Alignment of the medial tibial plateau
affects the rate of joint space narrowing in the osteoar-
thritic knee. Osteoarthritis Cartilage 2010 Nov;18(11):
1436e40.
86. Buckland-Wright JC, Ward RJ, Peterfy C, Mojcik CF, Leff RL.
Reproducibility of the semiflexed (metatarsophalangeal)
radiographic knee position and automated measurements of
medial tibiofemoral joint space width in a multicenter clin-
ical trial of knee osteoarthritis. J Rheumatol 2004 Aug;31(8):
1588e97.
87. Nevitt MC, Peterfy C, Guermazi A, Felson DT, Duryea J,
Woodworth T, et al. Longitudinal performance evaluation and
validation of fixed-flexion radiography of the knee for
detection of joint space loss. Arthritis Rheum 2007
May;56(5):1512e20.
88. Guermazi A, Roemer FW, Burstein D, Hayashi D. Why
radiography should no longer be considered a surrogate
outcome measure for longitudinal assessment of cartilage
in knee osteoarthritis. Arthritis Res Ther 2011;13(6):247
[Review].
714 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
89. Mazzuca SA, Brandt KD, Katz BP. Is conventional radiography
suitable for evaluation of a disease-modifying drug in pa-
tients with knee osteoarthritis? Osteoarthritis Cartilage 1997
Jul;5(4):217e26.
90. Eckstein F, Wirth W, Hunter DJ, Guermazi A, Kwoh CK,
Nelson DR, et al. Magnitude and regional distribution of cartilage
loss associated with grades of joint space narrowing in radio-
graphic osteoarthritisedata from the Osteoarthritis Initiative
(OAI). Osteoarthritis Cartilage 2010 Jun;18(6):760e8.
91. The Osteoarthritis Initiative: A Knee Health Study. MRI procedure
manual for examinations of the Knee and Thigh2006. Available
from: http://oai.epi-ucsf.org/datarelease/operationsManuals/
MRI_ManualRev.pdf.
92. Food and Drug Administration. In: Services DoHaH, Ed. Draft
Guidance for Industry on Standards for Clinical Trial Imaging
Endpoints 2011.
93. Balamoody S, Williams TG, Waterton JC, Bowes M,
Hodgson R, Taylor CJ, et al. Comparison of 3T MR scanners in
regional cartilage-thickness analysis in osteoarthritis: a
cross-sectional multicenter, multivendor study. Arthritis Res
Ther 2010;12(5):R202.
94. Jaremko JL, Cheng RW, Lambert RG, Habib AF, Ronsky JL.
Reliability of an efficient MRI-based method for estimation of
knee cartilage volume using surface registration. Osteoar-
thritis Cartilage 2006 Sep;14(9):914e22.
95. Kijowski R, Gold GE. Routine 3D magnetic resonance im-
aging of joints. J Magn Reson Imaging 2011 Apr;33(4):
758e71.
96. McMahon CJ, Madhuranthakam AJ, Wu JS, Yablon CM, Wei JL,
Rofsky NM, et al. High-resolution proton density weighted
three-dimensional fast spin echo (3D-FSE) of the knee with
IDEAL at 1.5 Tesla: comparison with 3D-FSE and 2D-FSE-
einitial experience. J Magn Reson Imaging 2012 Feb;35(2):
361e9.
97. Raynauld JP, Kauffmann C, Beaudoin G, Berthiaume MJ, de
Guise JA, Bloch DA, et al. Reliability of a quantification im-
aging system using magnetic resonance images to measure
cartilage thickness and volume in human normal and oste-
oarthritic knees. Osteoarthritis Cartilage 2003 May;11(5):
351e60.
98. Anderson SW, Sakai O, Soto JA, Jara H. Improved T2 mapping
accuracy with dual-echo turbo spin echo: effect of phase
encoding profile orders. Magn Reson Med 2013 Jan;69(1):
137e43.
99. Mosher TJ, Zhang Z, Reddy R, Boudhar S, Milestone BN,
Morrison WB, et al. Knee articular cartilage damage in oste-
oarthritis: analysis of MR image biomarker reproducibility in
ACRIN-PA 4001 multicenter trial. Radiology 2011
Mar;258(3):832e42.
100. Balamoody S, Williams TG, Wolstenholme C, Waterton JC,
Bowes M, Hodgson R, et al. Magnetic resonance transverse
relaxation time T2 of knee cartilage in osteoarthritis at 3-T: a
cross-sectional multicentre, multivendor reproducibility
study. Skelet Radiol 2013 Apr;42(4):511e20 [Multicenter
Study Research Support, Non-U.S. Gov't].
101. Deoni SC. Correction of main and transmit magnetic field (B0
and B1) inhomogeneity effects in multicomponent-driven
equilibrium single-pulse observation of T1 and T2. Magn
Reson Med 2011 Apr;65(4):1021e35 [Research Support,
Non-U.S. Gov't].
102. Eckstein F, Kunz M, Hudelmaier M, Jackson R, Yu J, Eaton CB,
et al. Impact of coil design on the contrast-to-noise ratio,
precision, and consistency of quantitative cartilage
morphometry at 3 Tesla: a pilot study for the osteoarthritis
initiative. Magn Reson Med 2007 Feb;57(2):448e54.
103. Shapiro L, Harish M, Hargreaves B, Staroswiecki E, Gold G.
Advances in musculoskeletal MRI: technical considerations.
J Magn Reson Imaging 2012 Oct;36(4):775e87.
104. Getz K. Protocol design trends and their effect on clinical trial
performance. RAJ Pharma 2008:315e6.
105. Bhatt A. Quality of clinical trials: a moving target. Perspect
Clin Res 2011 Oct;2(4):124e8.
106. Medical Imaging in Clinical Trials. London: Springer; 2014.
107. Kleppinger CF, Ball LK. Building quality in clinical trials with
use of a quality systems approach. Clin Infect Dis 2010 Aug
1;51(Suppl 1):S111e6.
108. Kellgren JH, Lawrence JS. Radiological assessment of osteo-
arthrosis. Ann Rheum Dis 1957 Dec;16(4):494e502.
109. The atlas of standard radiographs of arthritis. Rheumatology
December 1, 2005;44(Suppl 4):iv43e72.
110. Scott Jr WW, Lethbridge-Cejku M, Reichle R, Wigley FM,
Tobin JD, Hochberg MC. Reliability of grading scales for in-
dividual radiographic features of osteoarthritis of the knee.
The Baltimore longitudinal study of aging atlas of knee
osteoarthritis. Invest Radiol 1993 Jun;28(6):497e501.
111. Altman RD, Gold GE. Atlas of individual radiographic features
in osteoarthritis, revised. Osteoarthritis Cartilage
2007;15(Suppl A):A1eA56 [Research Support, Non-U.S.
Gov't].
112. Burnett S, Hart DJ, Cooper C, Spector TD. A Radiographic Atlas
of Osteoarthritis. London: Springer Verlag; 1994.
113. Spector T, Cooper C, Cushnighan J, Hart D, Dieppe PA.
A Radiographic Atlas of Knee Osteoarthritis. London: Springer
Verlag; 1992.
114. Nagaosa Y, Mateus M, Hassan B, Lanyon P, Doherty M.
Development of a logically devised line drawing atlas for
grading of knee osteoarthritis. Ann Rheum Dis 2000
Aug;59(8):587e95.
115. Neumann G, Hunter D, Nevitt M, Chibnik LB, Kwoh K, Chen H,
et al. Location specific radiographic joint space width for
osteoarthritis progression. Osteoarthritis Cartilage 2009
Jun;17(6):761e5 [Article].
116. Hunter DJ, Lo GH, Gale D, Grainger AJ, Guermazi A,
Conaghan PG. The reliability of a new scoring system for knee
osteoarthritis MRI and the validity of bone marrow lesion
assessment: BLOKS (Boston-Leeds Osteoarthritis Knee Score).
Ann Rheum Dis 2008 Feb;67(2):206e11.
117. Eckstein F, Ateshian G, Burgkart R, Burstein D, Cicuttini F,
Dardzinski B, et al. Proposal for a nomenclature for magnetic
resonance imaging based measures of articular cartilage in
osteoarthritis. Osteoarthritis Cartilage 2006 Oct;14(10):
974e83.
118. Alizai H, Virayavanich W, Joseph GB, Nardo L, Liu F, Liebl H,
et al. Cartilage lesion score: comparison of a quantitative
assessment score with established semiquantitative MR
scoring systems. Radiology 2014 May;271(2):479e87.
119. Martel-Pelletier J, Roubille C, Abram F, Hochberg MC,
Dorais M, Delorme P, et al. First-line analysis of the effects of
treatment on progression of structural changes in knee
osteoarthritis over 24 months: data from the osteoarthritis
initiative progression cohort. Ann Rheum Dis 2015
Mar;74(3):547e56.
120. Prasoon A, Igel C, Loog M, Lauze F, Dam EB, Nielsen M.
Femoral cartilage segmentation in knee MRI scans using two
stage voxel classification. Conf Proc IEEE Eng Med Biol Soc
2013:5469e72.
121. Duryea J, Neumann G, Brem MH, Koh W, Noorbakhsh F,
Jackson RD, et al. Novel fast semi-automated software to
segment cartilage for knee MR acquisitions. Osteoarthritis
Cartilage 2007 May;15(5):487e92.
 D.J. Hunter et al. / Osteoarthritis and Cartilage 23 (2015) 698e715
122. Eckstein F, Buck RJ, Burstein D, Charles HC, Crim J,
Hudelmaier M, et al. Precision of 3.0 Tesla quantitative
magnetic resonance imaging of cartilage morphology in a
multicentre clinical trial. Ann Rheum Dis 2008 Dec;67(12):
1683e8.
123. Hunter DJ, Bowes MA, Eaton CB, Holmes AP, Mann H,
Kwoh CK, et al. Can cartilage loss be detected in knee oste-
oarthritis (OA) patients with 3-6 months' observation using
advanced image analysis of 3T MRI? Osteoarthritis Cartilage
2010;18(5):677e83.
124. Hunter DJ, Niu J, Zhang Y, Totterman S, Tamez J, Dabrowski C,
et al. Change in cartilage morphometry: a sample of the
progression cohort of the Osteoarthritis Initiative. Ann
Rheum Dis 2009 Mar;68(3):349e56.
125. Zhang M, Driban JB, Price LL, Harper D, Lo GH, Miller E, et al.
Development of a rapid knee cartilage damage quantification
method using magnetic resonance images. BMC Muscu-
loskelet Disord 2014;15:264.
126. Dodin P, Pelletier JP, Martel-Pelletier J, Abram F. Automatic
human knee cartilage segmentation from 3D magnetic
resonance images. IEEE Trans Biomed Eng Nov 2010;57(11):
2699e711.
127. Schneider E, Nevitt M, McCulloch C, Cicuttini FM, Duryea J,
Eckstein F, et al. Equivalence and precision of knee cartilage
morphometry between different segmentation teams, carti-
lage regions, and MR acquisitions. Osteoarthritis Cartilage
2012 Aug;20(8):869e79.
128. Bowes MA, Vincent GR, Wolstenholme CB, Conaghan PG.
A novel method for bone area measurement provides new
insights into osteoarthritis and its progression. Ann Rheum
Dis 2015 Mar;74(3):519e25.
129. Dodin P, Martel-Pelletier J, Pelletier JP, Abram F. A fully
automated human knee 3D MRI bone segmentation using the
ray casting technique. Med Biol Eng Comput 2011
Dec;49(12):1413e24.
130. Guermazi A, Hayashi D, Eckstein F, Hunter DJ, Duryea J,
Roemer FW. Imaging of osteoarthritis. Rheum Dis Clin North
Am 2013 Feb;39(1):67e105.
131. Hunter DJ, Zhang W, Conaghan PG, Hirko K, Menashe L, Li L,
et al. Systematic review of the concurrent and predictive
validity of MRI biomarkers in OA. Osteoarthritis Cartilage
2011;19(5):557e88.
132. Reichmann WM, Maillefert JF, Hunter DJ, Katz JN,
Conaghan PG, Losina E. Responsiveness to change and reli-
ability of measurement of radiographic joint space width in
osteoarthritis of the knee: a systematic review. Osteoarthritis
Cartilage 2011 May;19(5):550e6.
715
133. Hunter DJ, Zhang W, Conaghan PG, Hirko K, Menashe L,
Reichmann WM, et al. Responsiveness and reliability of MRI
in knee osteoarthritis: a meta-analysis of published evidence.
Osteoarthritis Cartilage 2011;19(5):589e605.
134. Surowiec RK, Lucas EP, Ho CP. Quantitative MRI in the eval-
uation of articular cartilage health: reproducibility and vari-
ability with a focus on T2 mapping. Knee Surg Sports
Traumatol Arthrosc 2014 Jun;22(6):1385e95.
135. Pachowsky ML, Trattnig S, Apprich S, Mauerer A, Zbyn S,
Welsch GH. Impact of different coils on biochemical T2 and
T2* relaxation time mapping of articular patella cartilage.
Skeletal Radiol 2013 Nov;42(11):1565e72.
136. Pan J, Pialat JB, Joseph T, Kuo D, Joseph GB, Nevitt MC,
et al. Knee cartilage T2 characteristics and evolution in
relation to morphologic abnormalities detected at 3-T MR
imaging: a longitudinal study of the normal control cohort
from the Osteoarthritis Initiative. Radiology 2011
Nov;261(2):507e15.
137. Welsch GH, Mamisch TC, Marlovits S, Glaser C, Friedrich K,
Hennig FF, et al. Quantitative T2 mapping during follow-up
after matrix-associated autologous chondrocyte trans-
plantation (MACT): full-thickness and zonal evaluation to
visualize the maturation of cartilage repair tissue. J Orthop
Res 2009 Jul;27(7):957e63.
138. Crema MD, Hunter DJ, Burstein D, Roemer FW, Li L,
Eckstein F, et al. Association of changes in delayed
gadolinium-enhanced MRI of cartilage (dGEMRIC) with
changes in cartilage thickness in the medial tibiofemoral
compartment of the knee: a 2 year follow-up study using
3.0 T MRI. Ann Rheum Dis 2014 Nov. Jul 19.
139. Schooler J, Kumar D, Nardo L, McCulloch C, Li X, Link TM, et al.
Longitudinal evaluation of T1rho and T2 spatial distribution
in osteoarthritic and healthy medial knee cartilage. Osteo-
arthritis Cartilage 2015 Jan;22(1):51e62.
140. Emrani PS, Katz JN, Kessler CL, Reichmann WM, Wright EA,
McAlindon TE, et al. Joint space narrowing and Kellgren-
Lawrence progression in knee osteoarthritis: an analytic
literature synthesis. Osteoarthritis Cartilage 2008 Aug;16(8):
873e82.
141. Kessler LG, Barnhart HX, Buckler AJ, Choudhury KR,
Kondratovich MV, Toledano A, et al. The emerging science of
quantitative imaging biomarkers terminology and definitions
for scientific studies and regulatory submissions. Stat
Methods Med Res 2014 Jun 11.
142. Mosher TJ, Walker EA, Petscavage-Thomas J, Guermazi A.
Osteoarthritis year 2013 in review: imaging. Osteoarthritis
Cartilage 2013 Oct;21(10):1425e35.