Imaging in Osteoarthritis 2021 Yjoca

Osteoarthritis and Cartilage 30 (2022) 913e934
Review
Imaging in Osteoarthritis
F.W. Roemer y z *, A. Guermazi y x, S. Demehri k, W. Wirth ¶ # yy, R. Kijowski zz
y Quantitative Imaging Center, Department of Radiology, Boston University School of Medicine, FGH Building, 3rd Floor, 820 Harrison Ave, Boston, MA,
02118, USA
z Department of Radiology, Friedrich-Alexander University Erlangen-Nürnberg (FAU) and Universita €tsklinikum Erlangen, Maximiliansplatz 3, Erlangen,
91054, Germany
x Department of Radiology, VA Boston Healthcare System, 1400 VFW Pkwy, Suite 1B105, West Roxbury, MA, 02132, USA
k Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, 600 N. Wolf Street, Park 311,
Baltimore, MD, 21287, USA
¶ Institute of Anatomy, Paracelsus Medical University Salzburg, Salzburg, Austria, Nüremberg, Germany
# Ludwig Boltzmann Institute for Arthritis and Rehabilitation, Paracelsus Medical University Salzburg, Strubergasse 21, 5020, Salzburg, Austria
yy Chondrometrics, GmbH, Freilassing, Germany
zz Department of Radiology, New York University Grossmann School of Medicine, 550 1st Avenue, 3nd Floor, New York, NY, 10016, USA
a r t i c l e i n f o s u m m a r y
Article history: Osteoarthritis (OA) is the most frequent form of arthritis with major implications on both individual and
Received 31 January 2021 public health care levels. The field of joint imaging, and particularly magnetic resonance imaging (MRI),
Received in revised form has evolved rapidly due to the application of technical advances to the field of clinical research. This
22 April 2021
narrative review will provide an introduction to the different aspects of OA imaging aimed at an audience
Accepted 28 April 2021
of scientists, clinicians, students, industry employees, and others who are interested in OA but who do
not necessarily focus on OA. The current role of radiography and recent advances in measuring joint
Keywords:
space width will be discussed. The status of cartilage morphology assessment and evaluation of cartilage
Osteoarthritis
Radiography
biochemical composition will be presented. Advances in quantitative three-dimensional morphologic
Ultrasound cartilage assessment and semi-quantitative whole-organ assessment of OA will be reviewed. Although
MRI MRI has evolved as the most important imaging method used in OA research, other modalities such as
PET ultrasound, computed tomography, and metabolic imaging play a complementary role and will also be
CT discussed.
© 2022 The Authors. Published by Elsevier Ltd on behalf of Osteoarthritis Research Society International.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Introduction monitor its progression. Magnetic resonance imaging (MRI) is

seldom indicated in the assessment or management of knee or hip
Osteoarthritis (OA) is understood as the clinical and pathologic OA. Because of its high sensitivity, MRI is useful for research studies
outcome of a range of disorders that result in structural damage, to identify early OA, to enroll participants with specific joint con-
functional decline, and eventually failure of synovial joints1. OA ditions (e.g., severe cartilage lesions), and to and document struc-
occurs when the dynamic equilibrium between the breakdown and tural changes over time4. In clinical care, MRI can be useful if there
repair of joint tissues is disturbed and unbalanced, which will result is suspicion of conditions such as displaced meniscus fragment or
in pain, physical disability, and psychologic distress2. intra-articular body causing knee locking, subchondral insuffi-
While imaging is not needed to establish a clinical diagnosis of ciency fracture, tumor, or infection that would be treated differ-
OA, OA structural changes have traditionally been assessed with ently or more urgently than OA5.
radiographs3. Today, radiography is still the modality of choice in Application of advanced imaging methods has led to a rapid
the clinical setting to confirm a structural diagnosis of OA and to increase in our understanding of OA, and large longitudinal
observational studies such as the Osteoarthritis Initiative (OAI,
https://nda.nih.gov/oai/) and the Multicenter Osteoarthritis Study
(MOST, https://most.ucsf.edu/) have added to our knowledge
* Address correspondence and reprint requests to: F.W. Roemer, Department of
Radiology, Boston University Medical Center, FGH Building, 3rd floor, 820 Harrison regarding the natural course of the disease. The field of imaging,
Ave, Boston, MA, 02118, USA. Tel.: 1-617-414-4954; Fax: 1-617-638-6616. and particularly MRI, has evolved markedly over recent years based
E-mail address: froemer@bu.edu (F.W. Roemer).
https://doi.org/10.1016/j.joca.2021.04.018
1063-4584/© 2022 The Authors. Published by Elsevier Ltd on behalf of Osteoarthritis Research Society International. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
914 F.W. Roemer et al. / Osteoarthritis and Cartilage 30 (2022) 913e934
Fig. 1 FUNDAMENTALS OF OA Osteoarthritis and Cartilage
Image examples of the different Kellgren-Lawrence grades. A. Posterior-anterior (p.a.) radiograph of the knee joint shows a normal medial and
lateral joint space and no osteophytes. B. In Kellgren-Lawrence 1, equivocal osteophytes (large arrows) that do not fulfill criteria of a definite
osteophyte are present. In addition there is a small notch osteophyte ((small arrow) not considered in the KellgreneLawrence grading scale. The
medial and lateral joint space is normal. C. Kellgren-Lawrence 2 is characterized by definite osteophytes (arrows) without joint space narrowing. D
Kellgren-Lawrence 3 knees exhibit definite joint space narrowing (arrowheads) and osteophytes (arrows). A knee with joint space narrowing but
no osteophytes would still fulfill criteria for Kellgren-Lawrence 3. E. Kellgren-Lawrence 4 is defined by complete obliteration of the medial or
lateral or both joint spaces (black arrows). In addition there are large medial and lateral osteophytes (white arrows).
on technical advances and application of new and improved im- review is aimed at individuals seeking to acquire basic and clinical
aging technology to the clinical research arena6. Quantitative and knowledge about OA. This includes scientists, graduate students,
semi-quantitative morphological MRI has revealed important educators, and any clinicians involved in OA management including
knowledge on the progression of the disease and its risk factors, Rheumatology, Orthopedics, Primary Care, and Allied Health. The
while compositional MRI has provided a deeper understanding of review will focus on the knee joint since the vast majority of
early and potentially reversible OA pathology7. research in OA and cartilage imaging has focused on this particular
This narrative review of the “Fundamentals of OA”-series of joint. Hip, ankle, shoulder or hand OA is prevalent but has gained
Osteoarthritis and Cartilage is an update of the Osteoarthritis less attention in the imaging community. This is primarily due to
Research Society International (OARSI) online educational tool the challenges in visualizing the thin cartilage on the curved
“Primer of OA - Imaging” published a decade ago and will cover the articular surfaces of these joints, particularly on MRI, which re-
different aspects of OA imaging. The content will focus on the pil- quires thin slices and high spatial resolution that is difficult to
lars and foundations of OA imaging and on recent advances. The obtain in clinically feasible scan times. .
F.W. Roemer et al. / Osteoarthritis and Cartilage 30 (2022) 913e934 915
This narrative review will discuss the current role of radiog- challenges regarding positioning and reproducibility of image
raphy in OA imaging and recent advances in the technology of joint acquisition and its inability to evaluate soft tissues, which may be
space width (JSW) measurement. Largely due to the application of sources of pain in OA10e13.
MRI to large clinical studies, it is now accepted consensus to The severity of radiographic OA can be estimated using semi-
perceive OA as a “whole-organ” disorder involving multiple joint quantitative scoring systems. Published atlases provide example
tissues eventually leading to joint failure. Whole joint MRI assess- images that represent specific grades14e16. The Kellgren and Law-
ment has contributed much to this change in perception, and thus rence grading scheme, which dates back as far as 1957, is still the
the role of the different joint structures in a clinical context and accepted standard to diagnose radiographic OA3. The definition of
their relevance for disease progression will be presented in detail. radiographic OA relies on the presence of a “definite” osteophyte on
In addition, the review will discuss methodologies for evaluating the anterior-posterior (a.p.) or posterior-anterior (p.a.) weight-
cartilage biochemical composition and advances in quantitative bearing radiograph (defined as Kellgren and Lawrence grade 2).
morphologic cartilage assessment based on segmentation ap- Differentiation between grades 2 and 3 is based on absence or
proaches, which have made large and significant contributions to presence of joint space narrowing. Illustrative image examples of
our understanding of disease progression in recent years. Although the Kellgren-Lawrence grading instrument are presented in Fig. 1.
MRI has evolved as the most important imaging method used in OA The Kellgren and Lawrence scoring system suffers from limitations
research, other modalities such as ultrasound do play an important based on the invalid assumption that radiographic changes appear
complementary role and will be discussed. The review will in a consecutive fashion over the course of the disease and that it is
conclude with an overview of metabolic imaging and computed comprised of a composite measure of JSW and osteophyte presence.
tomography (CT) and their current and potential role in the clinical In contrast, the OARSI atlas classification scores joint space nar-
and research setting. rowing and osteophytes separately in each compartment of the
knee14,16. Table I presents a comparative overview of Kellgren and
Radiography Lawrence and OARSI atlas image assessment.
For decades the extended knee radiograph (i.e., a bilateral
For a long time OA structural changes have been assessed with weight-bearing a.p. view of both knees in full extension) has been
radiographs. Radiography is used in clinical practice to establish a the standard radiograph employed to image the tibiofemoral
structural diagnosis of OA and to monitor progression of the dis- joint17. While the diagnostic utility of the extended knee radiograph
ease. Radiographs visualize bony features including marginal is established, this technique has limitations concerning repro-
osteophytes, subchondral sclerosis, and subchondral cysts that are ducibility in longitudinal assessment of JSW18. Several research
associated with OA, and provide an indirect estimate of cartilage groups have developed alternative protocols for standardized
thickness and meniscal integrity by assessment of JSW. Radio- positioning of the knee. Common to all of these is a standard of
graphic assessment of OA mainly relies on the evaluation of knee flexion, that provides contact between the tibia and the
osteophytes and joint space narrowing. Osteophytes, i.e., bony posterior aspect of the femoral condyle19. However, the protocols
outgrowths at the joint margins, generally develop at an earlier differ with respect to the degree of flexion required, the angulation
stage than joint space narrowing. The assessment of OA severity of the x-ray beam, and the parameter that is adjusted to meet the
relies mainly on joint space narrowing and concomitant sub- examination's positioning standards. Some positioning protocols
chondral bone pathology8,9. Shortcomings of radiography include use fluoroscopy to confirm satisfactory anatomical positioning of
its moderate sensitivity to change in longitudinal evaluation, the medial tibial plateau prior to acquisition of the radiograph
Kellgren-Lawrence Grading *,y OARSI Atlas Grading z,x
Grading Scale Definition Grading Scale OA structural feature
Grade 0 No feature of OA Medial femoral condyle (0e3) Marginal Osteophytes

Medial tibial plateau (0e3)
Lateral femoral condyle (0e3)
Lateral tibial plateau (0e3)
Grade 1 Doubtful JSN and possible osteophytic lipping Medial compartment (0e3) Joint space narrowing
Grade 2 Definite osteophytes and possible JSN Lateral compartment (0e3)
Grade 3 Moderate multiple osteophytes, definite JSN, and some absent/present Other
sclerosis and possible deformity of bone ends absent/present Medial tibial attrition
Grade 4 Large osteophytes, marked JSN, severe sclerosis, and absent/present Medial tibial sclerosis
definite deformity of bone ends Lateral femoral sclerosis
Abbreviations: OARSI eOsteoarthritis Research Society International; OA e osteoarthritis; JSN e joint space narrowing.
*
Note: Radiographic OA is defined as Kellgren-Lawrence grade 2, i.e., presence of a definite osteophyte.
y
Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957; 16:494e502.
z
Altman RD, Gold GE: Atlas of individual radiographic features in osteoarthritis, revised. Osteoarthritis Cartilage 2007; 15 (Supplement A):A1-A56.
x
Note: In contrast to Kellgren-Lawrence grading, which is based on descriptive definitions of each specific grade, OARSI atlas grading is based on image examples. A grade
for a specific compartmental feature is assigned based on a best match with the atlas image.
Table I FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Kellgren and Lawrence and OARSI grading systems for severity assessment of radiographic OA
Direct comparison of MRI and radiography for visualization of knee osteoarthritis. A. Baseline posterior-anterior radiograph shows normal medial
tibiofemoral joint space width (arrows). B. At 3 year follow-up, discrete but definite joint space narrowing is observed (arrows). C. Baseline in-
termediate-weighted fast spin echo (FSE) image of the same knee shows that the medial meniscus is well aligned with the medial joint margins
(white line). No cartilage damage is observed on this image. D. On the corresponding 3 year follow-up intermediate fast spin echo -weighted FSE
image, incident meniscal extrusion of the medial meniscal body is seen, responsible for radiographic joint space narrowing (arrowheads and
white line). No cartilage loss is observed during the follow-up interval.
while others use a non-fluoroscopic positioning standard (semi- mentioning specific imaging modalities27. Traditionally, it is
flexed metatarsophalangeal and fixed flexion views)20e23. required that imaging must provide evidence that the drug can
Standardized measurement of JSW is the prerequisite for improve or even normalize the JSW on X-ray (reverse progression),
assessment of joint space narrowing in clinical trials. Prior to the maintain the JSW on X-ray (halt progression), or slow joint space
development of automated and semi-automated methods, JSW loss by at least a pre-specified amount (slow the rate of progres-
measurement was conducted using manual methods24. Although sion)28. Despite these recommendations regarding the use of X-ray
manual methods offer simplicity of equipment and application and in clinical trials of OA, it is important to acknowledge that JSW as
can be used to measure any linear distance, they are subjective and measured on X-rays does not reflect cartilage damage only29. Fig. 2
labor-intensive. Automated and semi-automated techniques for use illustrates the relevance of the meniscus in the context of JSW
in clinical trials have been developed to provide rapid, objective, measurements. Further, a normal X-ray may not exclude the pres-
and precise measurements of JSW25,26. Different measures of JSW ence of MRI intra-articular OA features30.
have been introduced including minimum JSW, mean JSW or joint
space area, and location-specific JSW. Minimum JSW is defined as MRI
the shortest distance between the tibial and femoral margins of the
joint space within the weight-bearing areas of the medial and Although radiography remains the primary imaging modality
lateral tibio-femoral compartments. for defining inclusion and exclusion criteria as well as the primary
Despite major advances in cross-sectional imaging and partic- structural endpoint in OA clinical trials, known limitations for
ularly MRI, radiographically assessed joint space narrowing is still visualization of OA features significantly limits the utility of radi-
the recommended structural outcome for demonstrating the effi- ography both clinically and in the research setting. Long regarded
cacy of disease-modifying OA drugs in phase-III clinical trials by as a disease of “wear and tear,” the application of MRI has led to a
regulatory agencies, including the U.S. Food and Drug Administra- new paradigm for OA in that OA should be viewed as a whole-organ
tion. These have been amended in a recent draft guidance stating disorder involving multiple joint tissues. Whole-organ MRI
that “… an effect on the candidate structural endpoint will reliably assessment has contributed much to this change in perception. MRI
predict an effect on the clinical outcomes of interest” without has evolved to play a major role in the research setting, with
Direct comparison of non-enhanced and contrast-enhanced imaging for synovitis assessment. A. Axial proton density-weighted fat-suppressed
(FSE) image shows marked hyperintensity within the joint cavity suggesting severe joint effusion (asterisk). In addition, there is bone marrow
lesion at the proximal attachment site of the anterior cruciate ligament (arrow). B. Axial T1-weighted fat-suppressed FSE image after contrast
administration visualizes severe synovial thickening depicted as contrast enhancement (arrows). Asterisk demarcates the true amount of effusion,
which is only discrete and visualized as intra-articular hypointensity. Note that bone marrow lesion is depicted with comparative conspicuity
using T1-weighted contrast-enhanced imaging (arrowhead).
compositional MRI techniques becoming increasingly more that do not fulfill the defined criteria of a full grade change as per a
important due to their ability to assess ‘pre-morphologic’ given scoring system) in a longitudinal study can increase the
biochemical compositional changes of articular and periarticular sensitivity to change of a semi-quantitative outcome measure40.
tissues. There has been an on-going discussion of whether reading blinded
to time points of image acquisition is preferable to reading in an un-
blinded fashion. Reading un-blinded might result in a slight ten-
Semi-quantitative MRI assessment
dency to read more change in comparison to a blinded reading.
However, it has been shown that scoring without knowing the
The knee joint is assumed to represent an organ as it is
chronological sequence substantially decreases sensitivity in the
comprised of multiple tissues whose integrity is essential for the
detection of clinically relevant changes in comparison with scoring
joint to optimally function. The analyses based on semi-quantita-
in chronological order41,42. These studies showed that blinding to
tive (SQ) scoring from MRI have added deeply to the understanding
time point can lead to misclassification of the longitudinal change
of the pathophysiology and natural history of OA as well as the
in a feature and that it may compromise the assessment of the
clinical implications of the structural changes assessed28e36.
relation of that feature and its outcome, which was also translated
The Whole Organ Magnetic Resonance Imaging Score (WORMS)
to OA assessment43.
was the first published scoring system and has been used exten-
The heterogeneous nature of the clinical and structural mani-
sively for almost two decades in a multitude of OA studies world-
festations of the disease have led to recent suggestions that there
wide31. Since then, three more whole-organ knee scoring systems
may be several phenotypes or subpopulations in OA that are
have been developed: The Knee Osteoarthritis Scoring System
characterized by distinct clinical manifestations of disease, by
(KOSS), the Boston Leeds Osteoarthritis Knee Score (BLOKS), and
certain laboratory parameters, biochemical markers, and/or imag-
the MRI Osteoarthritis Knee Score (MOAKS)32e34. All of these
ing findings44. While MRI-based phenotypic characterization of
scoring systems are based on MRI without intravenous or intra-
large cohort data is not yet available, three main structural phe-
articular administration of contrast agents, while other systems
notypes in OA have been proposed, i.e., meniscus/cartilage, sub-
have been developed that are based on contrast-enhanced MRI
chondral bone, and inflammation45,46. These may progress
specifically developed for assessment of synovitis35. Fig. 3 shows an
differently and may represent specific tissue targets for future
example of non-enhanced vs enhanced imaging for synovitis
therapeutic approaches47,48. Atrophic and hypertrophic pheno-
assessment. Literature evidence suggests that semi-quantitative
types have also been defined but are uncommon49. Fig. 4 depicts
assessment of knee OA by expert MRI readers is a valid, reliable, and
image examples of different structural OA phenotypes. Recently, a
responsive tool that helps investigators to understand the natural
semi-quantitative scoring instrument, the rapid osteoarthritis MRI
history of this complex disease and to evaluate potential new drugs
eligibility score (ROAMES), has been presented that enables
in OA clinical trials4,36,37. These approaches have enabled us to
phenotypic characterization based on simplified MRI assessment,
explain associations of structural tissue damage with clinical
which also includes assessment of findings of exclusion such as
manifestations of the disease and with changes in the morphology
subchondral insufficiency fracture, malignant bone marrow infil-
that indicate disease onset and progression38,39. Recording of
tration, and others50. Combined with accelerated image acquisition,
‘within-grade’ changes (i.e., definite visual morphologic changes
Structural knee OA phenotypes as defined by MRI. A. Cartilage/Meniscus phenotype. Coronal intermediate-weighted fast spin echo (FSE) image
shows diffuse full-thickness cartilage loss at the lateral tibia (arrowhead). In addition there is marked meniscal substance loss (i.e., maceration) at
the corresponding lateral meniscal body but also medially (thin arrows) characteristic for the cartilage/meniscus phenotype of knee OA. In
addition, there are osteophytes at the joint margins characteristic of structural OA (thick arrows). B. Inflammatory phenotype. Sagittal inter-
mediate-weighted fat-suppressed FSE image shows diffuse hyperintensity within the joint cavity (asterisk). In addition, there is a small sub-
chondral bone marrow lesion at the medial femur (large arrow) and a discrete horizontal-oblique meniscal tear in the posterior horn of the medial
meniscus (small arrows). Note that structural features of OA rarely are present in isolation. C. Subchondral bone phenotype. The subchondral
bone phenotype is characterized by larger bone marrow lesions as shown in this sagittal intermediate-eighted fat-suppressed FSE image. There
are large bone marrow lesions at the medial femur (arrows) but also tibia (arrowheads). Additional structural OA features are present such as a
minor medial meniscal tear ans small osteophytes. D. Atrophic phenotype. Coronal intermediate-weighted fat-suppressed FSE image shows
marked full thickness cartilage loss at the medial femur (arrow). In addition there is meniscal maceration (arrowhead) and superficial cartilage loss
at the medial tibia. No marginal osteophytes are seen at the medial or lateral joint line defining this knee as exhibiting an atrophic phenotype. E.
Hypertrophic phenotype. Hypertrophic phenotype. Coronal DESS image shows large marginal osteophytes medially and laterally (arrows)
characteristic of the hypertrophic phenotype of knee OA. In addition, there are osteophytes at the femoral notch commonly not included in whole
joint scoring systems (arrowheads). Note that there is only minor superficial cartilage damage in the medial and lateral femur and tibia.
MRI may be useful in screening endeavors for OA clinical trials. A exemplary fashion data extraction from a high-resolution 3D MRI
comparative overview of the main semi-quantitative scoring in- acquisition. Image analysis software can then be used to compute a
struments is presented in Table II. variety of morphological parameters, such as the volume, thickness,
size of the subchondral bone and cartilage surface area, the
denuded and cartilage covered subchondral bone area, and
Quantitative morphologic cartilage assessment by MRI
others51. Quantitative measurement of cartilage morphology re-
quires standardized, high-resolution 3D imaging sequences such as
Quantitative measurement of cartilage morphology exploits the
spoiled gradient-echo (SPGR) images or double-echo steady-state
three-dimensional (3D) nature of MRI data to assess morphologic
(DESS) images that delineate the bone-cartilage interface and
tissue parameters as continuous variables51 based on segmentation
cartilage surface with adequate contrast52,53 and provide a high
of the bone-cartilage-interface and cartilage surface. Fig. 5 shows in
sensitivity to change54. Analysis of total knee cartilage plates and differences in change60. By removing the link between magnitude
compartments has been extended methodologically to reporting and location of change, location-independent analyses such as the
changes in defined subregions as illustrated in Fig. 655. The spatial “ordered values”-approach or thinning/thickening scores circum-
pattern of subregional cartilage change has been reported in vent the challenge of selecting a particular knee compartment or
various samples, including Osteoarthritis Initiative participants subregion as an outcome measure of progression a priori. These
with up to 4-year follow-up56,57. measures have been shown to improve discrimination between
Cartilage morphometry is traditionally used to measure longi- healthy and OA subjects61 and between knees with difference
tudinal change in knees with established radiographic OA that are radiographic OA grades62, to be superior to location-based mea-
likely to show structural progression. The magnitude of progression sures in detecting risk factors of OA progression63, and to be sen-
is strongly linked to the existing cartilage damage, and therefore sitive to treatment interventions64. Location-independent analyses
radiographic scoring systems such as the Kellgren and Lawrence have also been shown to be sensitive to differences in change in
system or compartment-specific joint space narrowing grades are knees without radiographic signs of OA and may therefore allow
commonly used for selecting participants for studies relying on quantitative cartilage morphometry to be used in studies focusing
cartilage morphometry as the outcome measure58. Early OA, in on early OA.
contrast, is not characterized by wide-spread cartilage substance Quantitative measurements of cartilage volume and thickness
loss. Instead, knees without definite radiographic OA are likely to change have been used as outcomes in interventional trials, for
remain stable for longer periods of time before eventually pro- instance, evaluating the effect of non-steroidal anti-inflammatory
gressing to advanced disease. In addition, subregional thinning and drugs (NSAIDs), celecoxib, physical exercise, licofelone, chondroitin
thickening (due to cartilage swelling or hypertrophy) may be sulfate, strontium ranelate, triamcinolone acetonide, and vitamin D
observed simultaneously in early OA cohorts59, which impairs the supplementation. More recently, sprifermin (fibroblast growth
analysis of change in cartilage morphology on a cartilage plate or factor 18) showed a dose-dependent positive effect regarding
joint level. articular cartilage preservation and possibly growth when
Because knees may show both thinning and thickening on a compared to placebo65.
subregional level at the same time and because the location Quantitative joint analyses are not limited to articular cartilage,
affected from change in cartilage thickness varies between OA pa- but have also been applied to other joint tissues such as the bone,
tients, location-independent analyses focusing on the magnitude of meniscus, and infrapatellar fat-pad.
change (cartilage thickness gain and loss) independent of their
location have been suggested to provide a greater sensitivity to
Reference Acronym Features assessed (scale of scoring)
Peterfy (2004) WORMS Cartilage: (0e6)

Bone marrow lesions; subchondral cysts; bone attrition; effusion/synovitis; periarticular cysts/bursitis; loose bodies: (0
e3)
Osteophytes: (0e7)
Meniscal pathology: (0e4)
Cruciate and collateral ligaments: (0e1)
Kornaat (2005) KOSS Cartilage size and depth (i.e., % that is full thickness loss): (0e3)
Bone marrow lesions; subchondral cysts; osteophytes; effusion; meniscal tear; meniscal extrusion; popliteal cysts: (0
e3)
Synovial thickening (0e1)
Hunter (2008) BLOKS Cartilage size and depth: (0e3), plus extent of any cartilage loss at specified point
Bone marrow lesions (each lesion is evaluated on a 0e3 scale for: size by volume; size by % surface area adjacent to
subchondral plate; % of BML that is non-cystic)
Osteophytes; effusion; meniscal extrusion: (0e3)
Synovitis (size of signal changes in Hoffa's fat pad 0-3: five additional sites 0e1)
Meniscal status (evaluated on a presence/absence scale for: intrameniscal signal; tear; maceration; meniscal cyst)
Ligaments; periarticular cysts/bursitis; loose bodies: (0e1)
Hunter (2011) MOAKS Cartilage size and depth: (0e3); includes % of subregion that is full thickness loss
Bone marrow lesions (each subregion is evaluated on a 0e3 scale for: size by volume; % of BML that is non-cystic,
number of BMLs in subregion counted)
Osteophytes, effusion-synovitis; Hoffa-synovitis; meniscal extrusion: (0e3)
Meniscal status (evaluated on a 0e1 scale for: intrameniscal signal; tear; maceration; meniscal cyst; hypertrophy)
Ligaments; periarticular cysts/bursitis; loose bodies: (0e1)
Roemer (2020 ROAMES For cross-sectional phenotypic stratification and eligibility assessment in clinical trials:
Only maximum subregional WORMS/MOAKS grade per compartment is coded for cartilage, bone marrow lesions,
inflammation, osteophytes and meniscus
Diagnoses of exclusion are recorded in addition (0e1): meniscal root tears, osteonecrosis, subchondral insufficiency
fracture, tumors, malignant marrow infiltration and acute traumatic changes
Abbreviations: BML ebone marrow lesion; WORMS - Whole-Organ Magnetic Resonance Imaging Score; KOSS - Knee Osteoarthritis Scoring System; BLOKS - Boston Leeds
Osteoarthritis Knee Score; MOAKS - MRI Osteoarthritis Knee Score; ROAMES - Rapid OsteoArthritis MRI Eligibility Score.
Table II FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Overview of different semi-quantitative whole-organ MRI scoring systems for assessment of knee osteoarthritis
3D morphometry illustrated using a Kellgren and Lawrence grade 4 knee with areas of denuded, full-thickness cartilage loss in the medial
compartment from the OAI. The knee was imaged using 3T coronal SPGR MRI (top row) and sagittal DESS MRI (bottom row). Figures A. and D.
show the knee without segmentation and without the lines marking the femoral region of interest (sagittal orientation only). Figures B. and E.
illustrate the cartilage plates. Figures C. and F. illustrate the segmentation of the cartilage contours (green: subchondral bone area, magenta:
cartilage surface).
MRI pulse sequence protocols Compositional MRI is best suited for evaluating subjects at risk for
OA who do not have evidence of pre-existing joint degeneration
Concerning specific pulse sequence protocols for imaging the on radiographs or conventional MRI. A wide variety of composi-
knee joint to investigate OA, the minimum requirement would tional MRI methods have been used to evaluate patients with OA
consist of proton density- or intermediate-weighted fat-suppressed including T2 mapping, sodium imaging, T2* mapping, gagCEST
sequences in three planes. An additional non-fat suppressed T1- imaging, T1-rho mapping, diffusion-weighted imaging, ultra-short
weighted sequence for superior visualization of bony contours, echo-time imaging, and delayed gadolinium enhanced MRI of
loose bodies, and subchondral sclerosis is desirable. If 3D volu- cartilage (dGEMRIC)68,69. These techniques provide information
metric analysis is to be included, an additional high resolution 3D regarding different constituents of cartilage and have different
gradient-echo (GRE) sequence providing high intrinsic cartilage advantages and limitations which are summarized in Table III.
signal is necessary52. Although developed for dedicated cartilage However, the vast majority of studies have used T2 mapping, T1-
imaging, the relatively poorer contrast between cartilage and joint rho mapping, and dGEMRIC to evaluate subjects with OA, and thus
fluid of 3D GRE-based techniques renders them inferior in depict- these compositional MRI techniques will be described in more
ing focal cartilage defects when compared to standard two- detail.
dimensional (2D) fast spin-echo (FSE) techniques as show in T2 relaxation time can provide information regarding the
exemplary fashion in Fig. 766. An overview on protocol suggestions collagen component of cartilage as it is influenced by the organi-
for semi-quantitative assessment of knee OA was published several zation of the collagen fiber network, which is disrupted during the
years ago67. earliest stages of cartilage degeneration70. However, the T2 relax-
ation time of cartilage is influenced by multiple additional factors
including water and macromolecular concentration71,72, cartilage
Compositional MRI loading73, and magic angle effect due to the orientation of cartilage
relative to the main magnetic field74. Thus, changes in cartilage T2
Compositional MRI can detect changes in cartilage composition relaxation time may be difficult to interpret and often challenging
and ultra-structure during the earliest stages of cartilage degen-
eration prior to the onset of morphologic cartilage loss.
Femorotibial subregions of the medial and lateral compartment as commonly applied to 3D quantitative cartilage assessment. A. The central,
weight-bearing part of the femoral cartilage (bottom left) is divided into external (green), central (red), and internal (blue) subregions. The tibia
cartilage is divided into external, central, internal, anterior (turquoise) and posterior subregions (yellow).
to detect due to the multiple competing biological and mechanical a T2 mapping sequence in the imaging protocol of the longitudinal
factors which influence the measurement. OAI study. Higher cartilage T2 relaxation time in patients at risk for
T1-rho relaxation time can provide information regarding the knee OA can predict the incidence and progression of cartilage and
proteoglycan component of cartilage as it is influenced by the ex- bone marrow edema lesions88,89 and the onset of radiographic
change of protons between water molecules and the hydroxyl and OA90. However, cartilage T2 relaxation time measurements are of
amine groups on the glycosaminoglycan side chains of proteogly- limited value for predicting structural and symptomatic progres-
can. T1-rho relaxation time of both native and enzymatically sion in patients with established knee OA91,92. Furthermore, there is
degraded ex-vivo cartilage samples is strongly correlated with limited information on changes in cartilage T2 relaxation time or
proteoglycan content75. However, T1-rho relaxation time is not a other compositional MRI parameters in healthy knees over time,
specific measure of proteoglycan within cartilage and is also which makes it difficult to interpret small longitudinal changes in
influenced by multiple additional factors including other biological cartilage T2 and T1-rho relaxation time in knees with or at risk for
changes that occur during cartilage degeneration76, the orientation OA.
of cartilage relative to the main magnetic field77, and the degree of Both T2 mapping and T1-rho mapping have been used to eval-
cartilage loading78. uate post-traumatic and post-surgical changes in the cartilage of
dGEMRIC measures the T1 relaxation time of cartilage following the knee joint. Elevated cartilage T2 and T1-rho relaxation time
intra-venous administration of gadolinium contrast. dGEMRIC within the posterior lateral tibia plateau in areas of acute cartilage
provides a sensitive and specific measure of the proteoglycan injury is present during the early phase after anterior cruciate lig-
content of cartilage as negatively charged gadolinium contrast ament (ACL) tear93,94 and persists at both one year94,95 and two
distributes within cartilage indirectly proportional to the concen- year96 follow-up despite resolution of adjacent bone marrow le-
tration of negatively charged glycosaminoglycan side chains of sions. Multiple studies with follow-up periods ranging between six
proteoglycan79. However, dGEMRIC has several disadvantages months and two years show elevated cartilage T2 and T1-rho
including the long waiting time between contrast administration relaxation time within the medial compartment of the knee joint
and MRI which is inconvenient for patients and the risk of allergic following ACL injury96,97. Fig. 8 shows an example of elevated
reactions and nephrogenic systemic sclerosis with use of ionic cartilage T2 and T1-rho relaxation time in areas of morphologically
gadolinium contrast80. normal cartilage in a subject with ACL injury. Increased cartilage T2
T2 and T1-rho mapping have been used to evaluate subjects relaxation time is also present in subjects with isolated meniscus
with OA. The T2 and T1-rho relaxation time of knee cartilage in- tears98.
creases with age81,82 and with the severity of cartilage degenera- dGEMRIC has also been used to evaluate subjects with or at risk
tion83,84. T1-rho relaxation time is more sensitive than T2 for OA. Decreased T1 relaxation time of cartilage following gado-
relaxation time for detecting early cartilage degeneration in both linium contrast administration in animal studies correspond to
ex-vivo specimens85 and within the knee joint of human sub- areas of proteoglycan loss on histological analysis99. dGEMRIC has
jects86,87. However, cartilage T2 relaxation time has been more high sensitivity for detecting surgically and histologically
extensively studied in subjects with knee OA due to the inclusion of confirmed early cartilage degeneration within the knee joint in
Focal defects on MRI. A. Coronal intermediate-weighted fast spin echo (FSE) image shows a focal superficial thickness cartilage defect at the
medial femur (arrow). This finding cannot be depicted by radiography as joint space width will not be affected by such subtle changes. B.
Corresponding coronal fast low angle shot (FLASH) image at the same slice position shows only minor hypointensity in this area (ellipse) but no
definite defect. This example illustrates that not all MRI sequences are suited to visualize cartilage changes in the same manner.
human subjects100. Lower T1 relaxation times of cartilage following patellofemoral joint including the anterior femoral surface and
gadolinium contrast administration indicating reduced proteogly- Hoffa's fat pad, and the medial and lateral joint line including
can content is also present in subjects with higher Kellgren and osteophytes and the body of the meniscus. Fig. 9 shows an example
Lawrence grades of radiographic knee OA and in subjects with ACL of meniscal extrusion as shown by ultrasound. Furthermore, peri-
injury101,102 and meniscectomy103 who are at high risk for knee OA. articular cystic lesions are well-depicted. Ultrasonography has
DGEMRIC can also be used to monitor changes in cartilage pro- proven to be a useful adjuvant to routine clinical assessment to aid
teoglycan content following cartilage repair procedures104e108. management of disease, rather than a stand-alone diagnostic
test113.
Ultrasonography In knee OA, ultrasound adds diagnostic information over con-
ventional radiography primarily by its ability to directly visualize
Ultrasound provides real time, multi-planar imaging at rela- structural changes in cartilage and meniscus, as well as to sensi-
tively low cost. It offers reliable assessment of OA-associated fea- tively detect effusion and synovitis112. A cross sectional, multi-
tures, including inflammatory and structural abnormalities, center European study analyzed 600 patients with painful knee OA
without contrast administration or exposure to radiation109. A and found that ultrasound-detected synovitis correlated with
comprehensive overview of the topic was presented some years advanced radiographic OA and clinical symptoms and signs sug-
ago by Keen et al.110. Ultrasound allows sensitive and specific gestive of an acute inflammatory episode109.
identification of soft tissue and superficial bone changes with the It has been shown that ultrasound is more sensitive than con-
advantage of non-invasive assessment of vascularity111. However, ventional radiography for detecting tibiofemoral osteophytes114.
acquisition of ultrasonography skills takes time and practice and The benefits of ultrasonography over radiography include the
ongoing maintenance of competency112. The main limitation is the ability to image soft tissue structures, and the potential to detect
inability to pass through bone interfaces or into deeper articular small or early structural lesions115. Whether there will be a role for
structures. However, ultrasound is able to visualize the ultrasound in the assessment of pre-radiographic OA remains to be
MRI Technique Cartilage Constituent Advantages Limitations

Evaluated
T2 Mapping Collagen network, Well validated, widely Not specific to any

water available, baseline cartilage constituent,
measurements provide unable to evaluate
prognostic information, osteochondral junction,
does not require influenced by magic
contrast angle effect
Sodium Imaging Proteoglycan Specific for Requires high field
proteoglycan in strength 3.0T or 7.0T
cartilage, does not scanner and specialized
require contrast coils, low signal results
in long scan time and
low spatial resolution
T2* Mapping Collagen network, Faster acquisition time Not well validated,
water than T2 mapping, able susceptible to magnetic
to evaluate field inhomogeneity,
osteochondral junction, influenced by magic
does not require angle effect
contrast
gagCest Imaging Proteoglycan Specific for Not well validated,
proteoglycan in requires high field
cartilage, does not strength 7.0T scanner,
require contrast pulse sequence not
widely available
T1-Rho Mapping Proteoglycan Well validated, more Not specific to any
sensitive to early cartilage constituent,
cartilage degeneration unable to evaluate
than T2 mapping, less osteochondral junction,
influenced by magic pulse sequence not
effect than T2 and T2* widely available, high
mapping, does not specific absorption rate
require contrast
Diffusion-Imaging Proteoglycan, collagen Simultaneously Not well validated,
network provides information requires high field
regarding proteoglycan strength 3.0T or 7.0T
and collagen network, scanner, pulse
does not require sequence not widely
contrast available
UTE Imaging Proteoglycan, water, Able to evaluate Not well validated,
collagen network osteochondral junction, superiority over
can be combined with conventional T2, T2*,
T2, T2*, and T1-rho and T1-rho mapping
mapping unclear
dGEMRIC Proteoglycan Specific for Requires gadolinium
proteoglycan in contrast, long wait
cartilage, well validated prior to MRI, potential
toxicity of gadolinium
Table III FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Characteristics, strengths and drawbacks of compositional MRI techniques to assess knee cartilage
seen. Possible examples of its clinical use include evaluation of joint approach for measuring joint space parameters from weight-
effusion and synovitis, which has been shown to be a pre-cursor of bearing CT (WBCT) has been introduced120 CT arthrography can
cartilage damage, and evaluation of meniscal extrusion under improve tissue contrast through use of intra-articular iodinated
weight-bearing conditions116,117. contrast. CT arthrography is considered the in vivo reference
standard for measuring cartilage thickness121,122 and has high
Computed tomography diagnostic performance for detecting cartilage defects, meniscus
tears, and ACL injuries within the knee joint as exemplified in
CT has been applied to study OA-associated changes including Fig. 10123,124. The imaging technique can also be used to simulta-
bone trabecular remodeling, subchondral cysts, and subchondral neously measure cartilage thickness and subchondral bone min-
bone sclerosis118. In addition to subchondral bone changes, CT is eral density, which allows investigation of the regional
useful for detecting and quantifying associated tissue mineraliza- interactions between cartilage and bone in the onset and pro-
tion such as chondrocalcinosis, which is thought to play a role in gression of OA125. Furthermore, CT arthrography can provide in
disease initiation and progression119 and recently a quantitative vivo assessment of the proteoglycan content of cartilage as anionic
Early cartilage degeneration after ACL injury. (A) Sagittal fat-suppressed SPGR image in a subject with prior ACL injury shows a small intra-
chondral osteophyte (arrowhead) but otherwise morphologically normal cartilage on the posterior lateral tibial plateau without surface defects. (B)
Corresponding sagittal T2 mapping image and (C) sagittal fat-suppressed T1-rho mapping image show elevated cartilage T2 and T1-rho
relaxation time on the posterior lateral tibial plateau (arrows), indicating changes in cartilage composition and ultra-structure due to early cartilage
degeneration.
intra-articular iodinated contrast diffuses into cartilage in inverse DECT can also create bone subtraction images that can identify
proportion to the concentration of negatively charged glycosami- attenuation changes within bone marrow due to post-traumatic
noglycan in the macromolecular matrix126. However, CT arthrog- and degenerative bone marrow edema lesions131. Newly developed
raphy is invasive and is unable to detect all structural features of extremity cone beam CT provides high resolution imaging of the
joint degeneration including joint effusion, synovitis, and bone knee and ankle in the weight-bearing position with low radiation
marrow edema lesions, which has limited its widespread use in doses. WBCT can assess structural features of knee joint degener-
OA research studies. ation132 and measure tibiofemoral JSW with high scan-rescan
Recent technologic advances have provided new CT applications reliability133. Tibiofemoral JSW measurements obtained using
for evaluating subjects with OA. Dual-energy CT (DECT) can WBCT correlate strongly with measurements obtained using fixed
distinguish between structures with similar densities but different flexion radiography134 and are associated with clinical symptoms
elemental compositions based upon attenuation differences on and physical function in subjects with knee OA135. WBCT may have
image datasets acquired at different energy levels127. DECT can a role in assessing relevance of articular structures that show
improve characterization of crystal deposition diseases, which is change under weight-bearing conditions as exemplified in Fig. 11.
thought to play an important role in the onset and progression of In addition to WBCT, four-dimensional CT (4D-CT) has been
knee OA128, through detection of monosodium urate deposits in developed, which is not only able to detect patellofemoral mal-
early gout129 and distinguishing between calcium pyrophosphate tracking136, which is an important risk factor for patellofemoral OA,
and calcium hydroxyapatite deposits in meniscus and bone130. but can also provide an accurate and reliable visual assessment of
pain138, the Kellgren and Lawrence grade of radiographic OA139,

and the severity of cartilage degeneration on MRI139,140. 18F-FDG
uptake on PET imaging is correlated with the severity of synovitis
in subjects with knee OA141. Furthermore, 18F-NaF uptake but not
18
F-FDG uptake is seen in bone marrow edema lesions, osteo-
phytes, and subchondral sclerosis in subjects with knee OA on
PET-MRI with the extent of radiopharmaceutical uptake directly
correlating with the severity of bone abnormalities142. However,
more than a third of regions of 18F-NaF uptake on PET show no
correlative abnormality on MRI indicating that bone remodeling
in knee OA may precede morphologic joint degeneration142. 18F-
NaF uptake in bone in subjects with or at risk for knee OA is also
associated with proteoglycan loss and collagen matrix degrada-
tion detected on compositional MRI, which provides evidence of
the interaction between bone and cartilage in joint
degeneration143.
Artificial Intelligence (AI) in OA imaging
An abundance of literature has been published in recent years

on AI-based approaches in the context of OA imaging. It will go far
beyond the scope of this narrative review to cover these in
comprehensive fashion, although thorough review articles on the
applications of AI in OA imaging have been published144e146.
However, some recent developments that may be relevant for the
non-expert will be described.
Several deep-learning (DL) approaches have been used to
identify structural joint pathology on MRI associated with knee OA
including cartilage lesions147,148, meniscus tears148,149, and ACL in-
Fig. 9 Osteoarthritis and Cartilage juries149,150. Most of these studies first isolated the structure of
interest prior to disease classification using a convolutional neural
network (CNN) and showed similar diagnostic performance as
FUNDAMENTALS OF OA human readers for detecting joint pathology. One study described
the use of a single neural network that detected meniscus tears and
Sonographic image of the medial tibiofemoral joint space in mod-
ACL injuries on MRI without first isolating the structure of interest.
erate structural osteoarthritis. Image shows marked extrusion of the
However, the DL approach had significantly lower diagnostic per-
body of the medial meniscus (arrows). In addition, a small femoral
formance than human readers despite using large training data-
osteophyte is depicted (arrowhead). Note that there is sound
sets150. This suggests that isolation of individual joint structures or
extinction toward the more central parts of the joint (asterisk), which
use of more sophisticated neural networks may be needed to
does not allow for assessment of the cartilage surface and the
maximize diagnostic performance of DL approaches for detecting
ligaments in these deeper areas of the joint.
joint pathology.
Many studies have described DL methods for fully-automated
segmentation of articular cartilage using a wide variety of ap-
proaches including statistical shape modelling151, 3D active
appearance models152, 2D and 3D CNNs153,154, and sophisticated
patellar tracking137. Fig. 12 shows an example of time-resolved combinations or variants of generative adversarial networks
patellofemoral tracking during knee flexion and extension using (GANs)155. Although progress in OA image segmentation has led to
4D-CT. a point where algorithm errors lie within the intra-reader vari-
ability range, deployment in clinical practice still requires some
Molecular imaging form of supervision, which may include visual inspection of seg-
mentation outputs as well as error maps when a ground truth is
Molecular imaging techniques have been used to evaluate available146.
subjects with OA including single-photon emission computed DL methods have been used to assign Kellgren-Lawrence
tomography (SPECT), which utilizes Tc-99 hydroxymethane grades to assess the severity of OA on knee radiographs. In
diphosphonate (HDP) to assess bone turnover, and positive one study, the neural network had an average multi-class
emission tomography (PET) imaging, which utilizes 18F-fluo- accuracy of 67% for assigning a Kellgren-Lawrence grade with an
rodeoxyglucose (FDG) to assess metabolic activity as exemplified area under the curve of 0.93 for determining the presence or
in Fig. 13 and 18F-SodiumFluoride (NaF) to assess bone turnover. absence of radiographic OA156. In another study, the neural
Both SPECT and PET are typically combined with CT or MRI to network had a sensitivity ranging between 69% and 86% and a
provide improved anatomic localization of radiopharmaceutical specificity ranging between 84% and 99% for assigning
uptake and to correlate findings of metabolic activity and bone individual Kellgren-Lawrence grades157. Other applications of AI
turn-over with morphologic and compositional findings of joint in OA imaging include automated prescription of imaging pro-
degeneration. Quantitative uptake of Tc-99 HDP on SPECT-CT in tocols, image acquisition acceleration, reconstruction, and
subjects with knee OA is associated with the severity of knee prognosis158.
CT arthrography is considered the gold standard for the visualization of cartilage surface defects. A. Coronal reformatted CT image after intra-
articular administration of contrast material shows a subtle but definite focal superficial defect at the medial tibia (arrow). B. The same defect is
confirmed in a sagittal reformat (arrow).
AI-based imaging of OA has several advantages and proven improving interpretability and estimation of uncertainty of DL
potential. However, studies on the use DL methods for evaluating models159.
patients with OA on imaging studies are in their preliminary phase,
and additional validation is needed before these techniques could Conclusions and outlook
be of widespread use in clinical practice and research settings.
Furthermore, availability of datasets involving large cohorts of Conventional radiography remains the initial and most widely
participants, as well as multi-institutional collaborations are used imaging technique for evaluation of a patient with a known or
fundamental to improve generalizability of AI's technological suspected diagnosis of OA. In research and clinical trials, it is still an
promise. Future direction of research will likely be devoted to important tool for stratifying OA patients into different categories
for inclusion criteria and eligibility. However, due to inherent
Weight-bearing CT. Novel weight-bearing CT devices allow for imaging in an upright, standing position. A. Standard non-weight bearing coronal
intermediate-weighted fat-suppressed fast spin echo (FSE) image show a regular alignment of the medial meniscus with the joint margin (arrows).
B. Corresponding weight-bearing CT acquired on the same day shows definite extrusion of the medial meniscus beyond the tibial margin
(arrows).
Four-dimensional CT (4D-CT). Novel 4D-CT devices allow for imaging of the patellofemoral joint throughout extension-flexion knee motion to
elucidate the patellar tracking pattern. A. 3D volume rendering CT image of bilateral patellofemoral joint in a 19-year old athlete with history of left
patellar instability demonstrates symmetric lateral subluxation of patella at full extension (A, arrows) which resolves during knee flexion (B,
arrowheads).
limitations of plain radiography (i.e., inability to visualize most OA in clinical practice. The role of CT, scintigraphy, and FDG-PET in the
features, poor sensitivity and specificity to change, and difficulties diagnosis and follow-up of OA is still limited but is rapidly
of reproducibility in longitudinal trials), MRI plays a crucial role in advancing. A comparative summary of the strengths and limita-
understanding the natural history of the disease and in guiding tions of currently available imaging methods for the evaluation of
future therapies due to its ability to image the knee as a whole knee OA is shown in Table IV.
organ and to directly and three-dimensionally assess cartilage The use of high-resolution techniques to assess cartilage
morphology and composition. Ultrasound plays an important role morphology combined with available compositional techniques
in the diagnosis and follow-up of treatment of OA-related synovitis will improve the sensitivity of MRI in the detection of early joint
2-18F-fluoro-2-deoxyeD-glucose (FDG) positron emission tomography (PET) enables visualization of hypermetabolic joint tissues such as sy-
novial activation in osteoarthritis. A. Standard coronal reformatted CT image shows features of OA such as joint space narrowing in the lateral
compartment (arrows) and intra- and peri-meniscal calcification at the medial joint margin consistent with calcium pyrophosphate deposition
disease (CPPD) (large arrowheads). In addition, there is a definite osteophyte at the lateral femoral joint margin (small arrowhead). B. The cor-
responding fusion image of CT with FDG-PET shows marked peri-meniscal synovitis medially and laterally (short arrows) as well as around the
cruciate ligaments (arrowhead). In addition, there is synovitis at the posterior joint recess medially (long arrow).
928
Radiography US CT CT arthrography MRI SPECT and PET
Strengths - Low cost - Quickly performed - High anatomical [In addition to advantages of CT]: - All tissues of the joint - Provides metabolic
- Wide availability - Readily available resolution and excellent - Accurate evaluation of focal visualized information on bone
- Minimal radiation exposure - Enables real-time dynamic depiction of bony structures surface cartilage defects - Enables assessment (SPECT) and synovial
- Quick and inexpensive to imaging - 3D imaging with - Evaluation of meniscal tears of the tissue (FDG-PET)
perform - No radiation volumetric reconstruction and other intrinsic joint joint as a whole organ - Sensitive
- Readily available - Suitable for imaging of - Weight-bearing and structures such as ligaments using established
F.W. Roemer et al. / Osteoarthritis and Cartilage 30 (2022) 913e934

- Availability of established superficial structures with kinematic CT systems now - Potential to assess cartilage grading schemes
semiquantitative grading high spatial resolution available biochemical composition - Contrast-enhanced
schemes for osteoarthritis - Color Doppler adds imaging
information on vascularity provides additional
information,
especially for
synovitis
- No radiation
- Excellent anatomical
resolution
- 3D and compositional
assessment of
cartilage and other
tissues possible
Shortcomings - Projectional technique - Poor anatomical resolution - Higher cost relative to [In addition to disadvantages - Relatively long - Long imaging time
- 2D imaging only - Limited visualization of radiographs of CT] imaging - Non-specific
- Inability to visualize intraarticular structures in - Radiation exposure - High cost due to the use of time - Low spatial
important OA-related the weight-bearing aspects - Inability to delineate soft contrast and involved - May be resolution
pathologies including of the joint tissue structures in detail logistics contraindicated - Ionizing radiation
cartilage, bone marrow - Subchondral bone not - Invasive e risk of pain, in some patients - Logistical issues
lesions, ligaments, effusion depicted due to sound infection and hemorrhage - Costs regarding handling
and synovitis extinction at the subchondral post-procedure of radioactive tracers
- Technical limitations plate - Bone marrow lesions not - High costs
including positioning of the adequately visualized
joint and reproducibility in
longitudinal studies
Abbreviations: US e ultrasound, CT-computed tomography, MRI - magnetic resonance imaging, SPECT - single photon emission computed tomography; FDG-PET - fluorodeoxyglucose positron emission tomography.
Table IV FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Strengths and shortcomings of radiography, US, CT, CT arthrography, MRI and metabolic imaging techniques
degeneration, which most probably will strengthen the role of 6. Roemer FW, Demehri S, Omoumi P, Link TM, Kijowski R,
imaging in the clinical setting in the future. This will be of particular Saarakkala S, et al. State of the art: imaging of osteoarthritis-
importance once structure- or disease-modifying drugs become revisited 2020. Radiology 2020;296(1):5e21.
available, which will require monitoring the treatment response in 7. Guermazi A, Alizai H, Crema MD, Trattnig S, Regatte RR,
large sets of OA patients. With the availability of disease-modifying Roemer FW. Compositional MRI techniques for evaluation of
interventions, imaging will play an ever more important role in cartilage degeneration in osteoarthritis. Osteoarthritis Carti-
clinical decision-making and practice. lage 2015;23(10):1639e53.
8. Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K,
et al. Development of criteria for the classification and
Authors contributions
reporting of osteoarthritis. Classification of osteoarthritis of
the knee. Diagnostic and Therapeutic Criteria Committee of
(1) All authors were involved in the conception and design of the
the American Rheumatism Association. Arthritis Rheum
study, or acquisition of data, or analysis and interpretation of
1986;29(8):1039e49.
data.
9. Spector TD, Hart DJ, Byrne J, Harris PA, Dacre JE, Doyle DV.
(2) All authors contributed to drafting the article or revising it
Definition of osteoarthritis of the knee for epidemiological
critically for important intellectual content.
studies. Ann Rheum Dis 1993;52(11):790e4.
(3) All authors gave their final approval of the manuscript to be
10. Neogi T. Structural correlates of pain in osteoarthritis. Clin
submitted.
Exp Rheumatol 2017;35(5):75e8. Suppl 107.
11. Hunter DJ, Guermazi A, Roemer F, Zhang Y, Neogi T. Struc-
Additional contributions:
tural correlates of pain in joints with osteoarthritis. Osteo-
arthritis Cartilage 2013;21(9):1170e8.
Analysis and interpretation of the data: FWR, AG, SD, WW, RK.
12. Felson DT, Nevitt MC, Yang M, Clancy M, Niu J, Torner JC, et al.
Drafting of the article: FWR, AG, SD, WW, RK.
A new approach yields high rates of radiographic progression
Provision of study materials or patients: FWR, AG, SD, WW, RK.
in knee osteoarthritis. J Rheumatol 2008;35(10):2047e54.
Collection and assembly of data: FWR, AG, SD, WW, RK.
13. Kinds MB, Vincken KL, Hoppinga TN, Bleys RL, Viergever MA,
Marijnissen AC, et al. Influence of variation in semiflexed
Responsibility for the integrity of the work as a whole, from
knee positioning during image acquisition on separate
inception to finished article, is taken by F. Roemer, MD (first author;
quantitative radiographic parameters of osteoarthritis,
froemer@bu.edu).
measured by Knee Images Digital Analysis. Osteoarthritis
Cartilage 2012;20(9):997e1003.
Conflict of interest 14. Altman RD, Gold GE. Atlas of individual radiographic features
F.W.R. is Chief Medical Officer and shareholder of Boston Imaging in osteoarthritis, revised. Osteoarthritis Cartilage
Core Lab (BICL), a company providing image assessment services. 2007;15(Suppl A):A1eA56.
He received consultancies from Calibr e California Institute of 15. Scott Jr WW, Lethbridge-Cejku M, Reichle R, Wigley FM,
Biomedical Research. Tobin JD, Hochberg MC. Reliability of grading scales for in-
A.G. has received consultancies, speaking fees, and/or honoraria dividual radiographic features of osteoarthritis of the knee.
from Novartis, Pfizer, AstraZeneca, Regeneron,-, Merck Serono, and The Baltimore longitudinal study of aging atlas of knee
TissuGene and is President and shareholder of BICL. osteoarthritis. Invest Radiol 1993;28(6):497e501.
WW is shareholder and part-time employee with Chondro- 16. Altman RD, Hochberg M, Murphy Jr WA, Wolfe F,
metrics GmbH. He received consultancies from Galapagos. Lequesne M. Atlas of individual radiographic features in
None of the other authors have declared any competing osteoarthritis. Osteoarthritis Cartilage 1995;3(Suppl A):3e70.
interests. 17. Leach RE, Gregg T, Siber FJ. Weight-bearing radiography in
osteoarthritis of the knee. Radiology 1970;97(2):265e8.
18. Mazzuca SA, Brandt KD, Katz BP. Is conventional radiography
Funding and role of the funding source suitable for evaluation of a disease-modifying drug in patients
No funding was received for this study. with knee osteoarthritis? Osteoarthritis Cartilage 1997;5(4):
217e26.
References 19. Messieh SS, Fowler PJ, Munro T. Anteroposterior radiographs
of the osteoarthritic knee. J Bone Joint Surg Br 1990;72(4):
1. Bijlsma JW, Berenbaum F, Lafeber FP. Osteoarthritis: an up- 639e40.
date with relevance for clinical practice. Lancet 20. Buckland-Wright JC, Wolfe F, Ward RJ, Flowers N, Hayne C.
2011;377(9783):2115e26. Substantial superiority of semiflexed (MTP) views in knee
2. Eyre DR. Collagens and cartilage matrix homeostasis. Clin osteoarthritis: a comparative radiographic study, without
Orthop Relat Res 2004;(427 Suppl):S118e22. fluoroscopy, of standing extended, semiflexed (MTP), and
3. Kellgren JH, Lawrence JS. Radiological assessment of osteo- schuss views. J Rheumatol 1999;26(12):2664e74.
arthrosis. Ann Rheum Dis 1957;16(4):494e502. 21. Mazzuca SA, Brandt KD, Buckland-Wright JC, Buckwalter KA,
4. Hunter DJ, Zhang W, Conaghan PG, Hirko K, Menashe L, Li L, Katz BP, Lynch JA, et al. Field test of the reproducibility of
et al. Systematic review of the concurrent and predictive automated measurements of medial tibiofemoral joint space
validity of MRI biomarkers in OA. Osteoarthritis Cartilage width derived from standardized knee radiographs.
2011;19(5):557e88. J Rheumatol 1999;26(6):1359e65.
5. Katz JN, Arant KR, Loeser RF. Diagnosis and treatment of hip 22. Peterfy C, Li J, Zaim S, Duryea J, Lynch J, Miaux Y, et al.
and knee osteoarthritis: a review. J Am Med Assoc Comparison of fixed-flexion positioning with fluoroscopic
2021;325(6):568e78. semi-flexed positioning for quantifying radiographic joint-
space width in the knee: test-retest reproducibility. Skeletal 36. Guermazi A, Roemer FW, Haugen IK, Crema MD, Hayashi D.
Radiol 2003;32(3):128e32. MRI-based semiquantitative scoring of joint pathology in
23. Kan H, Arai Y, Kobayashi M, Nakagawa S, Inoue H, Hino M, osteoarthritis. Nat Rev Rheumatol 2013;9(4):236e51.
et al. Radiographic measurement of joint space width using 37. Collins JE, Losina E, Nevitt MC, Roemer FW, Guermazi A,
the fixed flexion view in 1,102 knees of Japanese patients Lynch JA, et al. Semiquantitative imaging biomarkers of knee
with osteoarthritis in comparison with the standing osteoarthritis progression: data from the foundation for the
extended view. Knee Surg Relat Res 2017;29(1):63e8. national institutes of health osteoarthritis biomarkers con-
24. Ravaud P, Chastang C, Auleley GR, Giraudeau B, Royant V, sortium. Arthritis Rheumatol 2016;68(10):2422e31.
Amor B, et al. Assessment of joint space width in patients 38. Roemer FW, Guermazi A, Collins JE, Losina E, Nevitt MC,
with osteoarthritis of the knee: a comparison of 4 measuring Lynch JA, et al. Semi-quantitative MRI biomarkers of knee
instruments. J Rheumatol 1996;23(10):1749e55. osteoarthritis progression in the FNIH biomarkers con-
25. Duryea J, Zaim S, Genant HK. New radiographic-based sur- sortium cohort - methodologic aspects and definition of
rogate outcome measures for osteoarthritis of the knee. change. BMC Musculoskelet Disord 2016;17(1):466.
Osteoarthritis Cartilage 2003;11(2):102e10. 39. Runhaar J, Schiphof D, van Meer B, Reijman M, Bierma-
26. Bruyere O, Henrotin YE, Honore A, Rovati LC, Seidel L, Zeinstra SM, Oei EH. How to define subregional osteoarthritis
Dardenne C, et al. Impact of the joint space width measure- progression using semi-quantitative MRI osteoarthritis knee
ment method on the design of knee osteoarthritis studies. score (MOAKS). Osteoarthritis Cartilage 2014;22(10):1533e6.
Aging Clin Exp Res 2003;15(2):136e41. 40. Roemer FW, Nevitt MC, Felson DT, Niu J, Lynch JA,
27. U.S. Department of Health and Human Services Food and Crema MD, et al. Predictive validity of within-grade scoring
Drug Administration Center for Drug Evaluation and of longitudinal changes of MRI-based cartilage morphology
Research (CDER) Center for Biologics Evaluation and Research and bone marrow lesion assessment in the tibio-femoral
(CBER) Center for Devices and Radiological Health (CDRH). joint–the MOST study. Osteoarthritis Cartilage 2012;20(11):
Osteoarthritis: Structural Endpoints for the Development of 1391e8.
Drugs, Devices, and Biological Products for Treatment Guid- 41. Bruynesteyn K, Van Der Heijde D, Boers M, Saudan A,
ance for Industry. 07/18/2018. URL: https://www.fda.gov/ Peloso P, Paulus H, et al. Detecting radiological changes in
media/71132/download accessed 01/20/2021. rheumatoid arthritis that are considered important by clinical
28. Hunter DJ, Altman RD, Cicuttini F, Crema MD, Duryea J, experts: influence of reading with or without known
Eckstein F, et al. OARSI Clinical Trials Recommendations: sequence. J Rheumatol 2002;29(11):2306e12.
knee imaging in clinical trials in osteoarthritis. Osteoarthritis 42. Ross PD, Huang C, Karpf D, Lydick E, Coel M, Hirsch L, et al.
Cartilage 2015;23(5):698e715. Blinded reading of radiographs increases the frequency of
29. Crema MD, Nevitt MC, Guermazi A, Felson DT, Wang K, errors in vertebral fracture detection. J Bone Miner Res
Lynch JA, et al. Progression of cartilage damage and meniscal 1996;11(11):1793e800.
pathology over 30 months is associated with an increase in 43. Felson DT, Nevitt MC. Blinding images to sequence in oste-
radiographic tibiofemoral joint space narrowing in persons oarthritis: evidence from other diseases. Osteoarthritis
with knee OA–the MOST study. Osteoarthritis Cartilage Cartilage 2009;17(3):281e3.
2014;22(10):1743e7. 44. Deveza LA, Melo L, Yamato TP, Mills K, Ravi V, Hunter DJ.
30. Guermazi A, Niu J, Hayashi D, Roemer FW, Englund M, Knee osteoarthritis phenotypes and their relevance for out-
Neogi T, et al. Prevalence of abnormalities in knees detected comes: a systematic review. Osteoarthritis Cartilage
by MRI in adults without knee osteoarthritis: population 2017;25(12):1926e41.
based observational study (Framingham Osteoarthritis 45. Oo WM, Yu SP, Daniel MS, Hunter DJ. Disease-modifying
Study). BMJ 2012;345:e5339. drugs in osteoarthritis: current understanding and future
31. Peterfy CG, Guermazi A, Zaim S, Tirman PF, Miaux Y, White D, therapeutics. Expert Opin Emerg Drugs 2018;23(4):331e47.
et al. Whole-organ magnetic resonance imaging score 46. Roemer FW, Kwoh CK, Hayashi D, Felson DT, Guermazi A. The
(WORMS) of the knee in osteoarthritis. Osteoarthritis Carti- role of radiography and MRI for eligibility assessment in
lage 2004;12(3):177e90. DMOAD trials of knee OA. Nat Rev Rheumatol 2018;14(6):
32. Kornaat PR, Ceulemans RY, Kroon HM, Riyazi N, 372e80.
Kloppenburg M, Carter WO, et al. MRI assessment of knee 47. Karsdal MA, Christiansen C, Ladel C, Henriksen K, Kraus VB.
osteoarthritis: knee Osteoarthritis Scoring System (KOSS)– Bay-Jensen AC. Osteoarthritis–a case for personalized health
inter-observer and intra-observer reproducibility of a care? Osteoarthritis Cartilage 2014;22(1):7e16.
compartment-based scoring system. Skeletal Radiol 48. Hunter DJ. Pharmacologic therapy for osteoarthritis–the era
2005;34(2):95e102. of disease modification. Nat Rev Rheumatol 2011;7(1):
33. Hunter DJ, Lo GH, Gale D, Grainger AJ, Guermazi A, 13e22.
Conaghan PG. The reliability of a new scoring system for knee 49. Roemer FW, Guermazi A, Niu J, Zhang Y, Mohr A, Felson DT.
osteoarthritis MRI and the validity of bone marrow lesion Prevalence of magnetic resonance imaging-defined atrophic
assessment: BLOKS (Boston Leeds Osteoarthritis Knee Score). and hypertrophic phenotypes of knee osteoarthritis in a
Ann Rheum Dis 2008;67(2):206e11. population-based cohort. Arthritis Rheum 2012;64(2):
34. Hunter DJ, Guermazi A, Lo GH, Grainger AJ, Conaghan PG, 429e37.
Boudreau RM, et al. Evolution of semi-quantitative whole 50. Roemer FW, Collins J, Kwoh CK, Hannon MJ, Neogi T,
joint assessment of knee OA: MOAKS (MRI Osteoarthritis Felson DT, et al. MRI-based screening for structural definition
Knee Score). Osteoarthritis Cartilage 2011;19(8):990e1002. of eligibility in clinical DMOAD trials: rapid OsteoArthritis
35. Guermazi A, Roemer FW, Hayashi D, Crema MD, Niu J, MRI Eligibility Score (ROAMES). Osteoarthritis Cartilage
Zhang Y, et al. Assessment of synovitis with contrast- 2020;28(1):71e81.
enhanced MRI using a whole-joint semiquantitative scoring 51. Eckstein F, Ateshian G, Burgkart R, Burstein D, Cicuttini F,
system in people with, or at high risk of, knee osteoarthritis: Dardzinski B, et al. Proposal for a nomenclature for
the MOST study. Ann Rheum Dis 2011;70(5):805e11. magnetic resonance imaging based measures of articular
cartilage in osteoarthritis. Osteoarthritis Cartilage 65. Hochberg MC, Guermazi A, Guehring H, Aydemir A, Wax S,
2006;14(10):974e83. Fleuranceau-Morel P, et al. Effect of intra-articular sprifermin
52. Crema MD, Roemer FW, Marra MD, Burstein D, Gold GE, vs placebo on femorotibial joint cartilage thickness in pa-
Eckstein F, et al. Articular cartilage in the knee: current MR tients with osteoarthritis: the FORWARD randomized clinical
imaging techniques and applications in clinical practice and trial. J Am Med Assoc 2019;322(14):1360e70.
research. Radiographics 2011;31(1):37e61. 66. Kijowski R, Blankenbaker DG, Davis KW, Shinki K, Kaplan LD,
53. Peterfy CG, Schneider E, Nevitt M. The osteoarthritis initia- De Smet AA. Comparison of 1.5- and 3.0-T MR imaging for
tive: report on the design rationale for the magnetic reso- evaluating the articular cartilage of the knee joint. Radiology
nance imaging protocol for the knee. Osteoarthritis Cartilage 2009;250(3):839e48.
2008;16(12):1433e41. 67. Peterfy CG, Gold G, Eckstein F, Cicuttini F, Dardzinski B,
54. Wirth W, Nevitt M, Hellio Le Graverand MP, Benichou O, Stevens R. MRI protocols for whole-organ assessment of the
Dreher D, Davies RY, et al. Sensitivity to change of cartilage knee in osteoarthritis. Osteoarthritis Cartilage 2006;14(Suppl
morphometry using coronal FLASH, sagittal DESS, and coro- A):A95eA111.
nal MPR DESS protocols–comparative data from the Osteo- 68. MacKay JW, Low SBL, Smith TO, Toms AP, McCaskie AW,
arthritis Initiative (OAI). Osteoarthritis Cartilage 2010;18(4): Gilbert FJ. Systematic review and meta-analysis of the reli-
547e54. ability and discriminative validity of cartilage compositional
55. Wirth W, Hellio Le Graverand MP, Wyman BT, Maschek S, MRI in knee osteoarthritis. Osteoarthritis Cartilage
Hudelmaier M, Hitzl W, et al. Regional analysis of femoroti- 2018;26(9):1140e52.
bial cartilage loss in a subsample from the Osteoarthritis 69. Atkinson HF, Birmingham TB, Moyer RF, Yacoub D, Kanko LE,
Initiative progression subcohort. Osteoarthritis Cartilage Bryant DM, et al. MRI T2 and T1rho relaxation in patients at
2009;17(3):291e7. risk for knee osteoarthritis: a systematic review and meta-
56. Eckstein F, Wirth W, Nevitt MC. Recent advances in osteo- analysis. BMC Musculoskelet Disord 2019;20(1):182.
arthritis imaging–the osteoarthritis initiative. Nat Rev Rheu- 70. Nieminen MT, Rieppo J, Toyras J, Hakumaki JM,
matol 2012;8(10):622e30. Silvennoinen J, Hyttinen MM, et al. T2 relaxation reveals
57. Buck RJ, Wirth W, Dreher D, Nevitt M, Eckstein F. Frequency spatial collagen architecture in articular cartilage: a
and spatial distribution of cartilage thickness change in knee comparative quantitative MRI and polarized light micro-
osteoarthritis and its relation to clinical and radiographic scopic study. Magn Reson Med 2001;46(3):487e93.
covariates - data from the osteoarthritis initiative. Osteoar- 71. Liess C, Lusse S, Karger N, Heller M, Gluer CC. Detection of
thritis Cartilage 2013;21(1):102e9. changes in cartilage water content using MRI T2-mapping in
58. Wirth W, Nevitt M, Hellio Le Graverand MP, Lynch J, vivo. Osteoarthritis Cartilage 2002;10(12):907e13.
Maschek S, Hudelmaier M, et al. Lateral and medial joint 72. Nishioka H, Hirose J, Nakamura E, Oniki Y, Takada K,
space narrowing predict subsequent cartilage loss in the Yamashita Y, et al. T1rho and T2 mapping reveal the in vivo
narrowed, but not in the non-narrowed femorotibial extracellular matrix of articular cartilage. J Magn Reson Im-
compartment - data from the Osteoarthritis Initiative. Oste- aging 2012;35(1):147e55.
oarthritis Cartilage 2014;22(1):63e70. 73. Apprich S, Mamisch TC, Welsch GH, Stelzeneder D, Albers C,
59. Buck RJ, Wyman BT, Le Graverand MP, Hudelmaier M, Totzke U, et al. Quantitative T2 mapping of the patella at 3.0T
Wirth W, Eckstein F, et al. Osteoarthritis may not be a one- is sensitive to early cartilage degeneration, but also to loading
way-road of cartilage loss–comparison of spatial patterns of of the knee. Eur J Radiol 2012;81(4):e438e43.
cartilage change between osteoarthritic and healthy knees. 74. Shao HP S, Li S, Ma Y, Tadros A, Chang E, Tang G, Du J. Magic
Osteoarthritis Cartilage 2010;18(3):329e35. angle effect plays a major role in both T1rho and T2 relaxa-
60. Eckstein F, Buck R, Wirth W. Location-independent analysis tion in articular cartilage. Osteoarthritis Cartilage
of structural progression of osteoarthritis-Taking it all apart, 2017;25(12):2022e30.
and putting the puzzle back together makes the difference. 75. Wheaton AJ, Dodge GR, Elliott DM, Nicoll SB, Reddy R.
Semin Arthritis Rheum 2017;46(4):404e10. Quantification of cartilage biomechanical and biochemical
61. Buck RJ, Wyman BT, Le Graverand MP, Hudelmaier M, properties via T1rho magnetic resonance imaging. Magn
Wirth W, Eckstein F. Does the use of ordered values of sub- Reson Med 2005;54(5):1087e93.
regional change in cartilage thickness improve the detection 76. Menezes NM, Gray ML, Hartke JR, Burstein D. T2 and T1rho
of disease progression in longitudinal studies of osteoar- MRI in articular cartilage systems. Magn Reson Med
thritis? Arthritis Rheum 2009;61(7):917e24. 2004;51(3):503e9.
62. Wirth W, Buck R, Nevitt M, Le Graverand MP, Benichou O, 77. Li X, Cheng J, Lin K, Saadat E, Bolbos RI, Jobke B, et al.
Dreher D, et al. MRI-based extended ordered values more Quantitative MRI using T1rho and T2 in human osteoarthritic
efficiently differentiate cartilage loss in knees with and cartilage specimens: correlation with biochemical measure-
without joint space narrowing than region-specific ap- ments and histology. Magn Reson Imaging 2011;29(3):
proaches using MRI or radiography–data from the OA 324e34.
initiative. Osteoarthritis Cartilage 2011;19(6):689e99. 78. Souza RB, Stehling C, Wyman BT, Hellio Le Graverand MP,
63. Eckstein F, Le Graverand MP, Charles HC, Hunter DJ, Kraus VB, Li X, Link TM, et al. The effects of acute loading on T1rho and
Sunyer T, et al. Clinical, radiographic, molecular and MRI- T2 relaxation times of tibiofemoral articular cartilage. Oste-
based predictors of cartilage loss in knee osteoarthritis. Ann oarthritis Cartilage 2010;18(12):1557e63.
Rheum Dis 2011;70(7):1223e30. 79. Bashir A, Gray ML, Boutin RD, Burstein D. Glycosaminoglycan
64. Eckstein F, Kraines JL, Aydemir A, Wirth W, Maschek S, in articular cartilage: in vivo assessment with delayed
Hochberg MC. Intra-articular sprifermin reduces cartilage Gd(DTPA)(2-)-enhanced MR imaging. Radiology
loss in addition to increasing cartilage gain independent of 1997;205(2):551e8.
location in the femorotibial joint: post-hoc analysis of a 80. Bellin MF, Van Der Molen AJ. Extracellular gadolinium-based
randomised, placebo-controlled phase II clinical trial. Ann contrast media: an overview. Eur J Radiol 2008;66(2):160e7.
Rheum Dis 2020;79(4):525e8.
81. Mosher TJ, Dardzinski BJ, Smith MB. Human articular carti- 95. Li X, Kuo D, Theologis A, Carballido-Gamio J, Stehling C,
lage: influence of aging and early symptomatic degeneration Link TM, et al. Cartilage in anterior cruciate ligament-recon-
on the spatial variation of T2–preliminary findings at 3 T. structed knees: MR imaging T1{rho} and T2–initial experi-
Radiology 2000;214(1):259e66. ence with 1-year follow-up. Radiology 2011;258(2):505e14.
82. Goto H, Yuki I, Fujii M, Aoyama N, Kubo S, Kuroda R, et al. The 96. Su F, Hilton JF, Nardo L, Wu S, Liang F, Link TM, et al. Cartilage
natural degenerative course of T1rho values of normal knee morphology and T1rho and T2 quantification in ACL-recon-
cartilage. Kobe J Med Sci 2011;57(4):155e70. structed knees: a 2-year follow-up. Osteoarthritis Cartilage
83. Dunn TC, Lu Y, Jin H, Ries MD, Majumdar S. T2 relaxation time 2013;21(8):1058e67.
of cartilage at MR imaging: comparison with severity of knee 97. Van Ginckel A, Verdonk P, Victor J, Witvrouw E. Cartilage
osteoarthritis. Radiology 2004;232(2):592e8. status in relation to return to sports after anterior cruciate
84. Witschey WR, Borthakur A, Fenty M, Kneeland BJ, Lonner JH, ligament reconstruction. Am J Sports Med 2013;41(3):550e9.
McArdle EL, et al. T1rho MRI quantification of arthroscopi- 98. Kai B, Mann SA, King C, Forster BB. Integrity of articular
cally confirmed cartilage degeneration. Magn Reson Med cartilage on T2 mapping associated with meniscal signal
2010;63(5):1376e82. change. Eur J Radiol 2011;79(3):421e7.
85. Regatte RR, Akella SV, Lonner JH, Kneeland JB, Reddy R. T1rho 99. Laurent D, O'Byrne E, Wasvary J, Pellas TC. In vivo MRI of
relaxation mapping in human osteoarthritis (OA) cartilage: cartilage pathogenesis in surgical models of osteoarthritis.
comparison of T1rho with T2. J Magn Reson Imaging Skeletal Radiol 2006;35(8):555e64.
2006;23(4):547e53. 100. Woertler K, Buerger H, Moeller J, Rummeny EJ. Patellar
86. Li X, Benjamin Ma C, Link TM, Castillo DD, Blumenkrantz G, articular cartilage lesions: in vitro MR imaging evaluation
Lozano J, et al. In vivo T(1rho) and T(2) mapping of articular after placement in gadopentetate dimeglumine solution.
cartilage in osteoarthritis of the knee using 3 T MRI. Osteo- Radiology 2004;230(3):768e73.
arthritis Cartilage 2007;15(7):789e97. 101. Tiderius CJ, Olsson LE, Nyquist F, Dahlberg L. Cartilage
87. Stahl R, Luke A, Li X, Carballido-Gamio J, Ma CB, Majumdar S, glycosaminoglycan loss in the acute phase after an anterior
et al. T1rho, T2 and focal knee cartilage abnormalities in cruciate ligament injury: delayed gadolinium-enhanced
physically active and sedentary healthy subjects versus early magnetic resonance imaging of cartilage and synovial fluid
OA patients–a 3.0-Tesla MRI study. Eur Radiol 2009;19(1): analysis. Arthritis Rheum 2005;52(1):120e7.
132e43. 102. Fleming BC, Oksendahl HL, Mehan WA, Portnoy R, Fadale PD,
88. Joseph GB, Baum T, Alizai H, Carballido-Gamio J, Nardo L, Hulstyn MJ, et al. Delayed gadolinium-enhanced MR imaging
Virayavanich W, et al. Baseline mean and heterogeneity of of cartilage (dGEMRIC) following ACL injury. Osteoarthritis
MR cartilage T2 are associated with morphologic degenera- Cartilage 2010;18(5):662e7.
tion of cartilage, meniscus, and bone marrow over 3 years– 103. Ericsson YB, Tjornstrand J, Tiderius CJ, Dahlberg LE. Rela-
data from the Osteoarthritis Initiative. Osteoarthritis Carti- tionship between cartilage glycosaminoglycan content
lage 2012;20(7):727e35. (assessed with dGEMRIC) and OA risk factors in meniscec-
89. Roemer FW, Eckstein F, Duda G, Guermazi A, Maschek S, tomized patients. Osteoarthritis Cartilage 2009;17(5):
Sharma L, et al. Is laminar cartilage composition as deter- 565e70.
mined by T2 relaxometry associated with incident and 104. Domayer SE, Welsch GH, Nehrer S, Chiari C, Dorotka R,
worsening of cartilage or bone marrow abnormalities? Szomolanyi P, et al. T2 mapping and dGEMRIC after autolo-
Cartilage 2020, https://doi.org/10.1177/1947603520932197. gous chondrocyte implantation with a fibrin-based scaffold
1947603520932197. Online ahead of print. in the knee: preliminary results. Eur J Radiol 2010;73(3):
90. Liebl H, Joseph G, Nevitt MC, Singh N, Heilmeier U, 636e42.
Subburaj K, et al. Early T2 changes predict onset of radio- 105. Gillis A, Bashir A, McKeon B, Scheller A, Gray ML, Burstein D.
graphic knee osteoarthritis: data from the osteoarthritis Magnetic resonance imaging of relative glycosaminoglycan
initiative. Ann Rheum Dis 2015;74(7):1353e9. distribution in patients with autologous chondrocyte trans-
91. Jungmann PM, Kraus MS, Nardo L, Liebl H, Alizai H, plants. Invest Radiol 2001;36(12):743e8.
Joseph GB, et al. T(2) relaxation time measurements are 106. Bekkers JE, Bartels LW, Benink RJ, Tsuchida AI, Vincken KL,
limited in monitoring progression, once advanced cartilage Dhert WJ, et al. Delayed gadolinium enhanced MRI of carti-
defects at the knee occur: longitudinal data from the osteo- lage (dGEMRIC) can be effectively applied for longitudinal
arthritis initiative. J Magn Reson Imaging 2013;38(6): cohort evaluation of articular cartilage regeneration. Osteo-
1415e24. arthritis Cartilage 2013;21(7):943e9.
92. Wirth W, Maschek S, Beringer P, Eckstein F. Subregional 107. Watanabe A, Wada Y, Obata T, Ueda T, Tamura M, Ikehira H,
laminar cartilage MR spin-spin relaxation times (T2) in et al. Delayed gadolinium-enhanced MR to determine
osteoarthritic knees with and without medial femorotibial glycosaminoglycan concentration in reparative cartilage after
cartilage loss - data from the Osteoarthritis Initiative (OAI). autologous chondrocyte implantation: preliminary results.
Osteoarthritis Cartilage 2017;25(8):1313e23. Radiology 2006;239(1):201e8.
93. Bolbos RI, Ma CB, Link TM, Majumdar S, Li X. In vivo T1rho 108. Chang G, Madelin G, Sherman OH, Strauss EJ, Xia D, Recht MP,
quantitative assessment of knee cartilage after anterior cru- et al. Improved assessment of cartilage repair tissue using
ciate ligament injury using 3 Tesla magnetic resonance im- fluid-suppressed (2)(3)Na inversion recovery MRI at 7 Tesla:
aging. Invest Radiol 2008;43(11):782e8. preliminary results. Eur Radiol 2012;22(6):1341e9.
94. Theologis AA, Kuo D, Cheng J, Bolbos RI, Carballido-Gamio J, 109. Conaghan PG, D'Agostino MA, Le Bars M, Baron G,
Ma CB, et al. Evaluation of bone bruises and associated Schmidely N, Wakefield R, et al. Clinical and ultrasonographic
cartilage in anterior cruciate ligament-injured and -recon- predictors of joint replacement for knee osteoarthritis: re-
structed knees using quantitative t(1rho) magnetic resonance sults from a large, 3-year, prospective EULAR study. Ann
imaging: 1-year cohort study. Arthroscopy 2011;27(1): Rheum Dis 2010;69(4):644e7.
65e76.
110. Keen HI, Wakefield RJ, Conaghan PG. A systematic review of non-osteoarthritic and severely osteoarthritic knees: in vivo
ultrasonography in osteoarthritis. Ann Rheum Dis concomitant 3D analysis using CT arthrography. Osteoar-
2009;68(5):611e9. thritis Cartilage 2019;27(4):621e9.
111. Schmidt WA, Volker L, Zacher J, Schlafke M, Ruhnke M, 126. Kokkonen HT, Suomalainen JS, Joukainen A, Kroger H, Sirola J,
Gromnica-Ihle E. Colour Doppler ultrasonography to detect Jurvelin JS, et al. In vivo diagnostics of human knee cartilage
pannus in knee joint synovitis. Clin Exp Rheumatol lesions using delayed CBCT arthrography. J Orthop Res
2000;18(4):439e44. 2014;32(3):403e12.
112. Oo WM, Bo MT. Role of ultrasonography in knee osteoar- 127. Johnson TR, Krauss B, Sedlmair M, Grasruck M, Bruder H,
thritis. J Clin Rheumatol 2016;22(6):324e9. Morhard D, et al. Material differentiation by dual energy CT:
113. Qvistgaard E, Kristoffersen H, Terslev L, Danneskiold- initial experience. Eur Radiol 2007;17(6):1510e7.
Samsoe B, Torp-Pedersen S, Bliddal H. Guidance by ultra- 128. Ea HK, Nguyen C, Bazin D, Bianchi A, Guicheux J, Reboul P,
sound of intra-articular injections in the knee and hip joints. et al. Articular cartilage calcification in osteoarthritis: insights
Osteoarthritis Cartilage 2001;9(6):512e7. into crystal-induced stress. Arthritis Rheum 2011;63(1):
114. Podlipska J, Guermazi A, Lehenkari P, Niinimaki J, 10e8.
Roemer FW, Arokoski JP, et al. Comparison of diagnostic 129. Bongartz T, Glazebrook KN, Kavros SJ, Murthy NS, Merry SP,
performance of semi-quantitative knee ultrasound and knee Franz 3rd WB, et al. Dual-energy CT for the diagnosis of gout:
radiography with MRI: oulu knee osteoarthritis study. Sci Rep an accuracy and diagnostic yield study. Ann Rheum Dis
2016;6:22365. 2015;74(6):1072e7.
115. Monteforte P, Rovetta G. Sonographic assessment of soft 130. Pascart T, Norberciak L, Legrand J, Becce F, Budzik JF. Dual-
tissue alterations in osteoarthritis of the knee. Int J Tissue energy computed tomography in calcium pyrophosphate
React 1999;21(1):19e23. deposition: initial clinical experience. Osteoarthritis Cartilage
116. Roemer FW, Guermazi A, Felson DT, Niu J, Nevitt MC, 2019;27(9):1309e14.
Crema MD, et al. Presence of MRI-detected joint effusion and 131. Li M, Qu Y, Song B. Meta-analysis of dual-energy computed
synovitis increases the risk of cartilage loss in knees without tomography virtual non-calcium imaging to detect bone
osteoarthritis at 30-month follow-up: the MOST study. Ann marrow edema. Eur J Radiol 2017;95:124e9.
Rheum Dis 2011;70(10):1804e9. 132. Segal NA, Nevitt MC, Lynch JA, Niu J, Torner JC, Guermazi A.
117. Nogueira-Barbosa MH, Gregio-Junior E, Lorenzato MM, Diagnostic performance of 3D standing CT imaging for
Guermazi A, Roemer FW, Chagas-Neto FA, et al. Ultrasound detection of knee osteoarthritis features. Phys Sportsmed
assessment of medial meniscal extrusion: a validation study 2015;43(3):213e20.
using MRI as reference standard. AJR Am J Roentgenol 133. Segal NA, Bergin J, Kern A, Findlay C, Anderson DD. Test-
2015;204(3):584e8. retest reliability of tibiofemoral joint space width measure-
118. Johnston JD, Masri BA, Wilson DR. Computed tomography ments made using a low-dose standing CT scanner. Skeletal
topographic mapping of subchondral density (CT-TOMASD) Radiol 2017;46(2):217e22.
in osteoarthritic and normal knees: methodological devel- 134. Segal NA, Frick E, Duryea J, Nevitt MC, Niu J, Torner JC, et al.
opment and preliminary findings. Osteoarthritis Cartilage Comparison of tibiofemoral joint space width measurements
2009;17(10):1319e26. from standing CT and fixed flexion radiography. J Orthop Res
119. Misra D, Guermazi A, Sieren JP, Lynch J, Torner J, Neogi T, 2017;35(7):1388e95.
et al. CT imaging for evaluation of calcium crystal deposition 135. Kothari MD, Rabe KG, Anderson DD, Nevitt MC, Lynch JA,
in the knee: initial experience from the Multicenter Osteo- Segal NA, et al. The relationship of three-dimensional joint
arthritis (MOST) study. Osteoarthritis Cartilage 2015;23(2): space width on weight bearing CT with pain and physical
244e8. function. J Orthop Res 2019 Dec 16, https://doi.org/10.1002/
120. Turmezei TD, S BL, Rupret S, Treece GM, Gee AH, MacKay JW, jor.24566. Online ahead of print.
et al. Quantitative three-dimensional assessment of knee 136. Demehri S, Thawait GK, Williams AA, Kompel A, Elias JJ,
joint space width from weight-bearing CT. Radiology 2021 Carrino JA, et al. Imaging characteristics of contralateral
Apr 13:203928, https://doi.org/10.1148/radiol.2021203928. asymptomatic patellofemoral joints in patients with unilat-
Online ahead of print. eral instability. Radiology 2014;273(3):821e30.
121. Wyler A, Bousson V, Bergot C, Polivka M, Leveque E, Vicaut E, 137. Best MJ, Tanaka MJ, Demehri S, Cosgarea AJ. Accuracy and
et al. Hyaline cartilage thickness in radiographically normal reliability of the visual assessment of patellar tracking. Am J
cadaveric hips: comparison of spiral CT arthrographic and Sports Med 2020;48(2):370e5.
macroscopic measurements. Radiology 2007;242(2):441e9. 138. Kim HR, So Y, Moon SG, Lee IS, Lee SH. Clinical value of (99m)
122. Haubner M, Eckstein F, Schnier M, Losch A, Sittek H, Becker C, Tc-methylene diphosphonate (MDP) bone single photon
et al. A non-invasive technique for 3-dimensional assessment emission computed tomography (SPECT) in patients with
of articular cartilage thickness based on MRI. Part 2: valida- knee osteoarthritis. Osteoarthritis Cartilage 2008;16(2):
tion using CT arthrography. Magn Reson Imaging 1997;15(7): 212e8.
805e13. 139. Kim J, Lee HH, Kang Y, Kim TK, Lee SW, So Y, et al. Maximum
123. Vande Berg BC, Lecouvet FE, Poilvache P, Dubuc JE, standardised uptake value of quantitative bone SPECT/CT in
Maldague B, Malghem J. Anterior cruciate ligament tears and patients with medial compartment osteoarthritis of the knee.
associated meniscal lesions: assessment at dual-detector Clin Radiol 2017;72(7):580e9.
spiral CT arthrography. Radiology 2002;223(2):403e9. 140. Dordevic M, Hirschmann MT, Rechsteiner J, Falkowski A,
124. Vande Berg BC, Lecouvet FE, Poilvache P, Jamart J, Materne R, Testa E, Hirschmann A. Do chondral lesions of the knee
Lengele B, et al. Assessment of knee cartilage in cadavers with correlate with bone tracer uptake by using SPECT/CT? Radi-
dual-detector spiral CT arthrography and MR imaging. Radi- ology 2016;278(1):223e31.
ology 2002;222(2):430e6. 141. Nakamura H, Masuko K, Yudoh K, Kato T, Nishioka K,
125. Omoumi P, Babel H, Jolles BM, Favre J. Relationships between Sugihara T, et al. Positron emission tomography with 18F-
cartilage thickness and subchondral bone mineral density in
FDG in osteoarthritic knee. Osteoarthritis Cartilage 152. Bowes MA, Guillard GA, Vincent GR, Brett AD,
2007;15(6):673e81. Wolstenholme CBH, Conaghan PG. Precision, reliability, and
142. Kogan F, Fan AP, McWalter EJ, Oei EHG, Quon A, Gold GE. PET/ responsiveness of a novel automated quantification tool for
MRI of metabolic activity in osteoarthritis: a feasibility study. cartilage thickness: data from the osteoarthritis initiative.
J Magn Reson Imaging 2017;45(6):1736e45. J Rheumatol 2020;47(2):282e9.
143. Savic D, Pedoia V, Seo Y, Yang J, Bucknor M, Franc BL, et al. 153. Wirth W, Eckstein F, Kemnitz J, Baumgartner CF,
Imaging bone-cartilage interactions in osteoarthritis using Konukoglu E, Fuerst D, et al. Accuracy and longitudinal
[(18)F]-NaF PET-MRI. Mol Imaging 2016;15:1e12. reproducibility of quantitative femorotibial cartilage mea-
144. Pedoia V, Majumdar S, Link TM. Segmentation of joint and sures derived from automated U-Net-based segmentation of
musculoskeletal tissue in the study of arthritis. Magma two different MRI contrasts: data from the osteoarthritis
2016;29(2):207e21. initiative healthy reference cohort. Magma 2020 Oct 6,
145. Hayashi D, Roemer FW, Guermazi A. Magnetic resonance https://doi.org/10.1007/s10334-020-00889-7. Online ahead
imaging assessment of knee osteoarthritis: current and of print.
developing new concepts and techniques. Clin Exp Rheu- 154. Cheng R, Alexandridi NA, Smith RM, Shen A, Gandler W,
matol 2019;37(5):88e95. Suppl 120. McCreedy E, et al. Fully automated patellofemoral MRI seg-
146. Recht MP, Dewey M, Dreyer K, Langlotz C, Niessen W, mentation using holistically nested networks: implications
Prainsack B, et al. Integrating artificial intelligence into the for evaluating patellofemoral osteoarthritis, pain, injury, pa-
clinical practice of radiology: challenges and recommenda- thology, and adolescent development. Magn Reson Med
tions. Eur Radiol 2020;30(6):3576e84. 2020;83(1):139e53.
147. Liu F, Zhou Z, Samsonov A, Blankenbaker D, Larison W, 155. Ambellan F, Tack A, Ehlke M, Zachow S. Automated seg-
Kanarek A, et al. Deep learning approach for evaluating knee mentation of knee bone and cartilage combining statistical
MR images: achieving high diagnostic performance for shape knowledge and convolutional neural networks: data
cartilage lesion detection. Radiology 2018;289(1):160e9. from the Osteoarthritis Initiative. Med Image Anal 2019;52:
148. Pedoia V, Norman B, Mehany SN, Bucknor MD, Link TM, 109e18.
Majumdar S. 3D convolutional neural networks for detection 156. Tiulpin A, Thevenot J, Rahtu E, Lehenkari P, Saarakkala S.
and severity staging of meniscus and PFJ cartilage morpho- Automatic knee osteoarthritis diagnosis from plain radio-
logical degenerative changes in osteoarthritis and anterior graphs: a deep learning-based approach. Sci Rep 2018;8(1):
cruciate ligament subjects. J Magn Reson Imaging 1727.
2019;49(2):400e10. 157. Norman B, Pedoia V, Noworolski A, Link TM, Majumdar S.
149. Bien N, Rajpurkar P, Ball RL, Irvin J, Park A, Jones E, et al. Applying densely connected convolutional neural networks
Deep-learning-assisted diagnosis for knee magnetic reso- for staging osteoarthritis severity from plain radiographs.
nance imaging: development and retrospective validation of J Digit Imaging 2019;32(3):471e7.
MRNet. PLoS Med 2018;15(11), e1002699. 158. Bowes MA, Kacena K, Alabas OA, Brett AD, Dube B, Bodick N,
150. Chang PD, Wong TT, Rasiej MJ. Deep learning for detection of et al. Machine-learning, MRI bone shape and important
complete anterior cruciate ligament tear. J Digit Imaging clinical outcomes in osteoarthritis: data from the Osteoar-
2019;32(6):980e6. thritis Initiative. Ann Rheum Dis 2020;80(4):502e8.
151. Van Dijck C, Wirix-Speetjens R, Jonkers I, Vander Sloten J. 159. Reyes M, Meier R, Pereira S, Silva CA, Dahlweid FM, von
Statistical shape model-based prediction of tibiofemoral Tengg-Kobligk H, et al. On the interpretability of artificial
cartilage. Comput Methods Biomech Biomed Engin intelligence in radiology: challenges and opportunities.
2018;21(9):568e78. Radiol Artif Intell 2020;2(3), e190043.

Imaging in Osteoarthritis 2021 Yjoca

Uploaded by

Copyright:

Available Formats

Imaging in Osteoarthritis 2021 Yjoca

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Imaging in Osteoarthritis 2021 Yjoca

Uploaded by

Copyright:

Available Formats

Osteoarthritis and Cartilage 30 (2022) 913e934

Introduction monitor its progression. Magnetic resonance imaging (MRI) is

Fig. 1 FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Kellgren-Lawrence Grading *,y OARSI Atlas Grading z,x

Grading Scale Deﬁnition Grading Scale OA structural feature

Grade 0 No feature of OA Medial femoral condyle (0e3) Marginal Osteophytes

Table I FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Fig. 2 FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Fig. 3 FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Fig. 4 FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Reference Acronym Features assessed (scale of scoring)

Peterfy (2004) WORMS Cartilage: (0e6)

Table II FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Fig. 5 FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Fig. 6 FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Fig. 7 FUNDAMENTALS OF OA Osteoarthritis and Cartilage

MRI Technique Cartilage Constituent Advantages Limitations

T2 Mapping Collagen network, Well validated, widely Not speciﬁc to any

Table III FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Fig. 8 FUNDAMENTALS OF OA Osteoarthritis and Cartilage

pain138, the Kellgren and Lawrence grade of radiographic OA139,

Artiﬁcial Intelligence (AI) in OA imaging

An abundance of literature has been published in recent years

Fig. 10 FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Fig. 11 FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Fig. 12 FUNDAMENTALS OF OA Osteoarthritis and Cartilage

Fig. 13 FUNDAMENTALS OF OA Osteoarthritis and Cartilage

F.W. Roemer et al. / Osteoarthritis and Cartilage 30 (2022) 913e934

Table IV FUNDAMENTALS OF OA Osteoarthritis and Cartilage

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.