European Journal of Radiology 82 (2013) 112–117

Contents lists available at SciVerse ScienceDirect

European Journal of Radiology

journal homepage: www.elsevier.com/locate/ejrad

A practical MRI grading system for osteoarthritis of the knee:

Association with Kellgren–Lawrence radiographic scores
Hee-Jin Park a,b,1 , Sam Soo Kim b,1 , So-Yeon Lee a,∗ , Noh-Hyuck Park c,1 , Ji-Yeon Park c,1 ,
Yoon-Jung Choi a,1 , Hyun-Jun Jeon d,1
a
Department of Radiology, Sungkyunkwan University School of Medicine, Kangbuk Samsung Hospital, #108 Pyung-dong, Jongno-gu, Seoul 110-746,
Republic of Korea
b
Department of Radiology, Kangwon National University School of Medicine, Baengnyeong-ro 156, Chuncheon-Si, Gangwon-Do Kangwon National
University Hospital 200-722, Republic of Korea
c
Department of Radiology, Myongji Hospital, Kwandong University, College of Medicine, 697-24 Hwajung-dong, Dukyang-ku, Koyang, Kyunggi 412-270,
Republic of Korea
d
Department of Occupational Medicine, Dongsan Medical Center, Keimyung University School of Medicine, 194 Dongsan-Dong, Jung-ku, Taegu, Republic of
Korea

a r t i c l e i n f o a b s t r a c t

Article history: Purpose: To propose a reproducible and constant MR grading system for osteoarthritis of the knee joint
Received 14 December 2011 that provides high interobserver and intraoberver agreement and that does not require complicated
Received in revised form 23 February 2012 calculation procedures.
Accepted 27 February 2012
Materials and methods: This retrospective study sample included 44 men and 65 women who underwent
both MRI and plain radiography of the knee at our institution. All patients were older than 50 years of age
Keywords:
(mean 57.7) and had clinically suspected osteoarthritis of the knee. The standard of 4 grades on the MR
MR
grade scale was based mainly on cartilage injury and additional findings. Kellgren–Lawrence grades were
Knee
Osteoarthritis
assessed for the same patient group. The relationship between the results was determined. Statistical
Cartilage analyses were performed including kappa statistics, categorical regression analysis and nonparametric
correlation analysis.
Results: The interobserver and intraoberver agreements between the two readers in the grading of
osteoarthritis were found to be almost perfect. Interobserver and intraobserver agreements were slightly
lower for the MR grading system than for the Kellgren–Lawrence grading scale. The correlation between
the MR grade and Kellgren–Lawrence grade was very high and did not differ with patient age. The MR
grades were highly correlated with the Kellgren–Lawrence grades and showed excellent interobserver
and intraobserver agreements.
Conclusion: This new MR grading system for osteoarthritis of the knee joint is reproducible and may be
helpful for the grading of osteoarthritis of the knee without requiring reference to plain radiography.
© 2012 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Radiographic grading scales are commonly used to define the

presence of and to estimate the severity of osteoarthritis.
Osteoarthritis is a progressive degenerative disease character- Three radiographic grading scales have been developed, the
ized by a gradual loss of articular cartilage [1]. Bone marrow edema Kellgren–Lawrence (KL), Ahlback, and Brandt scales [3–5]. Among
(BME), osteophytes, subchondral cysts and meniscal tears also these, the KL scale is the most frequently used. The KL scale
can be associated with osteoarthritis [2]. Many clinicians depend was adopted by the World Health Organization as the reference
on the radiography for assessing the degree of osteoarthritis. standard for cross-sectional and longitudinal epidemiologic studies
[6]. However, the KL scale cannot be used to characterize carti-
lage loss and marrow changes of the bony structures associated
with osteoarthritis of the knee. Radiologically-determined joint
∗ Corresponding author. Tel.: +82 2 2001 1035; fax: +82 2 2001 1030. space narrowing is an indirect measure of articular status, because
E-mail addresses: [email protected] (H.-J. Park), [email protected] radiographic evaluation includes only a small percentage of the
(S.S. Kim), [email protected], [email protected] (S.-Y. Lee), total articular surface [2] and radiograph evaluates only the bony
[email protected] (N.-H. Park), [email protected] (J.-Y. Park),
[email protected] (Y.-J. Choi), [email protected] (H.-J. Jeon).
component of the joint. MRI has become a useful tool for the
1
Tel.: +82 2 2001 1035; fax: +82 2 2001 1030. early detection and measurement of cartilage lesions because of

0720-048X/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ejrad.2012.02.023
 H.-J. Park et al. / European Journal of Radiology 82 (2013) 112–117 113

Table 1 the cartilage injury standard but did not satisfy the other additional
Proposed MR grading system of the osteoarthritis of the knee joint.
criteria, we consider the lesion as one grade lower state.
Grade of osteoarthritis Description

0 No cartilage injury with no or minimal 2.1.2. MR parameters

osteophyte (<5 mm) All MR examinations were performed using the same protocol
1 Cartilage injury grade I and at least one of the on a 1.5 T magnet MRI scanner (Intera, Philips) using an extremity
following; Osteophyte >5 mm,a BME >10 mm,
subchondral cyst >10 mm
coil. The sequences and imaging parameters are summarized in
2 Cartilage injury grade II and at least one of the Table 2.
following; Osteophyte >5 mm, BME >10 mm,
subchondral cyst >10 mm
2.1.3. Radiography of the knee
3 Cartilage injury grade III and at least one of the
following; Osteophyte >5 mm, BME >10 mm, In all patients, radiographs of the knee were obtained in lat-
subchondral cyst >10 mm eral and anteroposterior planes. Upright standing 45◦ flexion views
4 Cartilage injury grade III and meniscal injury and sunrise views of the patellofemoral joint were also obtained
grade III (50–70 kV, 8 mAs). The X-ray tube-cassette distance was 100 cm.
Note: If the lesion with each cartilage grade does not have additional findings, the The radiographs were obtained at the same attendance with MRI.
lesion corresponds to one grade lower state. Cartilage grade–Noyes classification.
Meniscal injury grade–Stoller et al. grade.
a
BME = Bone marrow edema. 2.2. Image analysis

Radiography and MR examinations were interpreted by two

its ability to noninvasively visualize hyaline cartilage in vivo [1]. fellowship-trained academic musculoskeletal radiologists with 12
To our knowledge, no standard MR grading scale for osteoarthritis and 10 years of experience who were blinded to clinical infor-
of the knee joint currently exists. In order to evaluate the sever- mation and radiologic reports. After grading the KL scores of the
ity of arthritis, we rely on plain radiography, which is an indirect entire study sample, each radiologist evaluated the MRI results
measure. This study was performed to propose reproducible and without referring to information about the radiographic grades of
constant MR grading system for osteoarthritis of the knee joint that the lesions.
provides high interobserver agreement and that does not require
complicated calculation procedures.
2.2.1. Radiographic assessment
Anteroposterior, lateral and axial views of the radiographs
2. Materials and methods were evaluated without reference to clinical information using
the KL scale [2,6]. KL scores were based on osteophyte formation,
2.1. Case selection joint space narrowing, sclerosis, and joint deformity characteris-
tics according to the five-level scale defined as follow: grade 0,
Our retrospective study sample included 44 (40%) men and 65 normal; grade 1, minute osteophytes of doubtful clinical signifi-
(60%) women who underwent both MRI and plain radiography of cance; grade 2, definite osteophytes with unimpaired joint space;
the knee at our institution between January 2010 and May 2011. All grade 3, definite osteophytes with moderate joint space narrowing;
patients were over age 50 (mean age 57.7 ± 7.0) and had clinically grade 4, definite osteophytes with severe joint space narrowing and
suspected osteoarthritis of the knee. The exclusion criteria were: subchondral sclerosis.
trauma, tumor, major ligament structure tears and previous his-
tory of knee surgery. This study was approved by our Institutional 2.2.2. MR imaging assessment
Ethics Review Board, and the requirement for informed consent During assessment of MR images, the readers were blinded to
was waived because of its retrospective nature. Thirty-nine (36%) KL scores and clinical information. MR images were used to assess
patients underwent arthroscopy and 13 (12%) patients underwent the medial and femoral condyle and patella for cartilage lesions,
microfracture surgery. osteophytes, BME and subchondral cysts. Cartilage lesions were
graded according to the Noyes classification [7]. The cartilages of
2.1.1. Establishment of MR grading system for osteoarthritis the three compartments (medial femorotibial, lateral femorotib-
We determined the criteria for osteoarthritis of the knee on ial and patellofemoral) and seven surfaces (medial tibial, medial
sagittal, coronal and axial MR imaging. Coronal T2-weighted fat femoral, lateral tibial, lateral femoral, trochlea, medial and lat-
saturation image were used as main sequences and sagittal and eral patellar facets) were evaluated and grading was determined
axial T2-weighted fat saturation images were used as complemen- according to the most severe lesions at each location. Osteophytes
tary sequences. The main standard of grading was cartilage injury were defined as any abnormal bone growth arising from the mar-
and other accessory findings such as osteophytes, BME and sub- gins of the involved compartment. The lengths of osteophytes from
chondral cysts were combined [2]. Cartilage injuries were classified base to the tip were evaluated by proton density imaging. BME
according to the Noyes classification system as follows [7]; grade 0, was defined as a non-circumscribed area of abnormally high signal
normal; grade 1, internal signal intensity alteration; grade 2, car- intensity on fat saturation T2-weighted fast spin echo images (FS
tilage defect less than 99%; grade 3, 100% cartilage defect with or T2FSE) in a subchondral location and was verified in at least two
without BME and bony ulcerations. Four grades were developed imaging planes for each patient. The largest diameter of each lesion
(Table 1). Grade 0 refers to the absence of cartilage injury with no was evaluated. Subchondral cysts were defined as well marginated
or minimal (<5 mm) osteophytes; grade 1 refers to cartilage injury circular or oval areas of homogeneously high signal intensity on
grade 1 and showing at least one of the findings below; grade 2 fat saturation T2-weighted fast spin echo images (FS T2FSE) in sub-
refers to cartilage injury grade 2 and showing at least one of the chondral locations and were verified in at least two imaging planes.
findings below; grade 3 refers to cartilage injury grade 3 and show- The greatest dimension of each lesion was evaluated.
ing at least one of the findings below; grade 4 refers to cartilage Meniscal tears were evaluated using the grading system of
injury grade 3 and meniscal injury grade 3 [8]. Additional crite- Stoller [8]. The grades were as follows: grade 0, normal; grade 1,
ria included: osteophytes > 5 mm, BME > 10 mm in maximal length, nonarticular focal or globular intrasubstance tear; grade 2, hori-
subchondral cyst > 10 mm in maximal diameter. If a lesion satisfied zontal, linear intrasubstance tear; grade 3, displaced, complex and
114 H.-J. Park et al. / European Journal of Radiology 82 (2013) 112–117

Table 2
Imaging parameters for MR sequences.

Imaging parameter Coro FS T2 FSE Coro PD FSE Sag PD FSE Axial FS T2 FSE Sag FS T2 FSE

TR (msec) 3600 1500 1500 2000 4200

TE (msec) 70 15 15 70 70
Flip angle (◦ ) 90 90 90 90 90
Matrix size 304 × 243 304 × 243 304 × 243 304 × 243 304 × 243
Field of view (cm) 16 16 16 16 16
Slice thickness (mm) 4 4 4 4 4
Bandwidth (kHz) 260 250 250 190 190
Echo train length 14 10 12 15 15
Signal average 2 2 2 2 2
Scan time (min:s) 2:07 2:27 2:07 2:13 2:02

Note: Coro = coronal, FS = fat saturation, FSE = fast spine-echo, Sag = sagittal.

extended into the articular surface. Among these categories, grade

3 was included in the MR grading system.

2.3. Statistical analysis

Interobserver and intraoberver agreements between the two

radiologists were analyzed using kappa statistics. The kappa
value interpretation was: poor (k < 0.1), slight (0.1 ≤ k ≤ 0.2), fair
(0.2 < k ≤ 0.4), moderate (0.4 < k ≤ 0.6), substantial (0.6 < k ≤ 0.8),
and almost perfect (0.8 < k ≤ 1). Correlation coefficients (Rs)
between MR grade and KL grade were calculated with categor-
ical regression analysis and nonparametric correlation analysis
(Spearman’s correlation). In analyses of the relationships between
findings and patient characteristics, the relationship between the
MR grade and KL grade was evaluated by age (<58, ≥58, 58 = mean
age of the sample). Statistical analyses were performed using SPSS
software version 10.1 (SPSS Inc., Chicago IL, USA). P-values less
than or equal to 0.05 were considered statistically significant. Cor-
relation coefficients (R) between 0.7 and 0.9 were considered to
represent relatively high positive correlations, and R greater than
0.9 were considered to represent very high positive correlations
(Figs. 1–4).

3. Results

According to MR grade, reader 1 detected grade 0 in 44(43)

patients, grade 1 in 21(23) patients, grade 2 in 19(18) patients,
grade 3 in 13(14) patients and grade 4 in 12(11) patients. Reader
2 detected grade 0 in 41(43) patients, grade 1 in 22(24) patients,
grade 2 in 22(20) patients, grade 3 in 10(8) patients, and grade 4
in 14 patients (Table 3). According to KL grade, reader 1 detected
grade 0 in 42(41) patients, grade 1 in 27(28) patients, grade 2 in 17
patients, grade 3 in 15 patients, and grade 4 in 8 patients. Reader
2 detected grade 0 in 43(42) patients, grade 1 in 27(28) patients, Fig. 1. (A and B) A case of MR grade 3 and KL grade 3. T2-weighted fat suppression
grade 2 in 17 patients, grade 3 in 15 patients, and grade 4 in 7(8) fast spine echo sagittal image of a 58-year-old man showing a full thickness tear of
patients. Interobserver and intraobserver agreements between the the cartilage of the medial femoral condyle and combined BME greater than 10 mm.
two readers for the grading of osteoarthritis were found to be Meniscal injury is not combined. An anteroposterior view of the knee reveals definite
osteophyte and moderate joint space narrowing.
almost perfect (Table 4). Interobserver and intraobserver agree-
ments were slightly lower for the MR grading system than for
the Kellgren–Lawrence grading scale. The R for reader 1 between

Table 3
Incidence of grades for MR grade and Kellgren–Lawrence grade.

Grade 0 Grade 1 Grade 2 Grade 3 Grade 4 Total

a
MR grade (first) 44/41 21/22 19/22 13/10 12/14 109/109
MR grade (second) 43/43 23/24 18/20 14/8 11/14 109/109
KL grade (first)b 42/43 27/27 17/17 15/15 8/7 109/109
KL grade (second) 41/42 28/27 17/17 15/15 8/8 109/109
a
Numbers of the patients detected by reader 1/reader 2.
b
KL grade = Kellgren–Lawrence grade.
 H.-J. Park et al. / European Journal of Radiology 82 (2013) 112–117 115

Fig. 2. (A and B) A case of MR grade 4 and KL grade 4. T2-weighted fat suppression Fig. 3. (A and B) Case of MR grade 2 and KL grade 1. T2-weighted fat suppression fast
fast spine echo coronal image of a 58-year-old man showing a full thickness tear spin echo coronal image of a 53-year-old man showing a partial thickness tear of the
of the cartilage of the lateral femoral condyle and combined marrow edema larger cartilage of the medial femoral condyle (large arrow) and combined BME larger than
than 10 mm. Meniscal injury is combined. A standing 45◦ flexion anteroposterior 10 mm (small arrow). Anteroposterior view of the knee reveals minute osteophytes
view of the knee reveals definite osteophyte and severe joint space narrowing. and no joint space narrowing.

Table 4
(≥58), the R for reader 1 between the MR grade and KL grade was
The interobserver and intraobserver agreements between readers.
0.934–0.960. The R for reader 2 between the MR grade and KL grade
MR grade KL gradea was 0.859–0.918 (Table 5). The correlations between the MR grade
Interobserver agreement 0.829/0.840b 0.975/0.963 and KL grade were very high and showed no differences according
Intraobserver agreement 0.938/0.914c 0.963/0.975 to patient age. The MR grading system showed a high correlation
a
KL grade = Kellgren–Lawrence grade. with the KL grading system as well as excellent interobserver and
b
Kappa value of first evaluation/second evaluation. intraobserver agreements.
c
Kappa value of reader 1/reader 2.

4. Discussion
the MR grade and KL grade was 0.938–0.965. The R for reader 2
between the MR grade and KL grade was 0.896–0.922. The corre- The widely-used KL grading scale has some limitations. A
lation between the MR grade and KL grade was very high. In the major criticism of the KL grading scale is its emphasis on osteo-
younger patients (<58 years), the R for reader 1 between the MR phytes in the diagnosis of osteoarthritis [9]. KL grades 1 and 2 are
grade and KL grade was 0.928–0.937. The R for reader 2 between defined exclusively by the presence of osteophytes. As a result, a
the MR grade and KL grade was 0.925–0.936. In the older patients patient with joint space narrowing or some cartilage injury but no

Table 5
Correlation coefficients for the MR grading system and Kellgren–Lawrence grading system.

Observer MR grade vsa . KL gradeb MR grade vs. KL grade (<58) MR grade vs. KL grade (≥58)

Reader 1 (first) 0.938 (0.883–0.982) 0.928 (0.830–0.995) 0.934 (0.823–0.976)

Reader 1 (second) 0.965 (0.950–0.976) 0.937 (0.895–0.962) 0.960 (0.930–0.977)
Reader 2 (first) 0.896 (0.806–0.948) 0.936 (0.837–0.996) 0.859 (0.733–0.929)
Reader 1 (second) 0.922 (0.888–0.946) 0.925 (0.877–0.955) 0.918 (0.860–0.952)

The level of correlation significance was 0.01. Numbers in parentheses are 95% confidence intervals of each correlation coefficient.
a
MR grade = MR grading system.
b
KL grade = Kellgren–Lawrence grading system.
116 H.-J. Park et al. / European Journal of Radiology 82 (2013) 112–117

diameter were increasingly common with increasing KL grade.

We established our diagnostic criteria following these size crite-
ria. Many clinicians want that radiologist characterize the severity
of osteoarthritis based on the grade of osteoarthritis obtained from
MRI, similar to the way that plain radiographic films are used to
diagnose osteoarthritis according to the KL grading scale. Medical
insurance systems such as the National Health Insurance Corpo-
ration (NHIC) of Korea require formal reports detailing the exact
radiological grading of osteoarthritis of the knee. This information
is used to decide whether the costs of specific medical practices
are appropriate given the radiologic severity of the osteoarthri-
tis. The KL grading scale is preferred for this purpose because it
is certified by the WHO. However, when MRI is performed, it is
very unreasonable to rely on plain radiography again which con-
tain uncertainty for the grading of osteoarthritis, because MRI is a
more expensive and advanced study and MR imaging procedures
are time-consuming tasks. A standardized grading system that does
not involve difficult procedures or complicated calculations such
as volume measurements is required. Previous studies relying on
cartilage volume measurements were very accurate, but were com-
plicated and time consuming, and such methods are therefore not
practical for use in real clinical practice [1,10]. Practical MR grading
methods require easy and quick measurements and should be as
concordant with the KL grading system as possible. If a unified MR
grading method were established and adopted for widespread use,
its objectivity and consistency would be improved, leading to the
approval of MR grading of osteoarthritis by the medical insurance
systems such as the NHIC. We therefore suggest a new MR grading
method; however, we note that this method is not perfect. Previous
studies have reported high interobserver reliability for KL grading
among readers (intraclass correlation coefficient, 076–0.84) [13]. In
our study, the new MR grading system also showed high interob-
server and intraobserver agreements. However, the agreement was
lower than that of the KL system for the same patients and observers
Fig. 4. (A and B) A case of MR grade 3 and KL grade 1. T2-weighted fat suppression (Table 4). This discrepancy may result from the uncertainty of carti-
fast spine echo coronal image of a 53-year-old man showing a full thickness tear lage injury evaluation, because the MR evaluation of cartilage is still
of the cartilage of the medial femoral condyle (small arrow) and combined marrow
limited. However, with advances in cartilage evaluation techniques,
edema larger than 10 mm (not shown). Meniscal injury is not combined. A standing
45◦ flexion anteroposterior view of the knee reveals minute osteophyte and no joint this diagnostic uncertainty will be decreased. The new grading sys-
space narrowing. tem mainly evaluates cartilage injury (Noyes class) from grade 1 to
grade 3 with some additional findings, and grade 4 includes ipsilat-
eral meniscal injury grade 3 [8]. Although many previous studies
associated osteophytes cannot be classified as having osteoarthri- tried quantitative analysis of the osteoarthritis of the knee, they
tis according to the KL system. Recently, many studies evaluating were focused on the scoring concerning disease severity and eval-
the grading of osteoarthritis using cartilage volume on MRI have uated too many features [14–16]. In the study of Hunter et al. [14],
been performed [1,2,10]. Cicuttini et al. [10] studied the relation- they evaluated 13 items for scoring and Peterfy et al. [15] evalu-
ship between tibial cartilage volume and radiographic grading of ated 14 items. Compared to the previous MR evaluation of knee
osteoarthritis during different stages of disease. They found a sig- joint osteoarthritis, which requires complex and time-consuming
nificant negative linear association between medial and lateral calculation, our grading system is more practical and shows good
tibial cartilage volumes and increasing grades of joint space nar- reproducibility. Furthermore, the findings using our grading system
rowing. They also compared joint space width to cartilage volume were also well correlated with those of the KL system in our patient
as measured from standing knee radiographs [11] and concluded sample (correlation coefficients; 0.859–0.960). The KL grade is not
that there was a significant correlation between joint space width a gold standard for evaluating osteoarthritis but is simply a plain
and cartilage volume adjusted by the medial tibial bone area in radiologic diagnosis. However, the KL grading scale is commonly
the medial tibiofemoral compartment. The relationship between accepted by radiologists and clinicians as a standard grading system
meniscal tears and osteoarthritis is also interesting. The purpose for knee joint osteoarthritis.
of the meniscus is to reduce contact stress by enlarging the con- Some of the limitations of this study were that the MR hardware
tact surface, distributing load and increasing stability. Meniscal used was 1.5 T, and variable cartilage specific imaging sequences
tears increase focal pressure and can lead to articular cartilage such as spoiled gradient–recalled echo (SPGR) or iterative decom-
damage [1]. Berthiaume et al. [12] suggested that there was a rela- position of water and fat with echo asymmetry and least–squares
tionship between meniscal damage and cartilage volume change estimation (IDEAL) were not used. We evaluated MR imaging with
in osteoarthritis. Hayes et al. [2] reported correlations between routine fast spin echo sequences only. The use of higher magnetic
the MR imaging parameters of synovial proliferation, menis- field strengths such as 3.0 T improves cartilage imaging because
cal abnormalities, cartilage defects, effusion, subchondral lesions, such field strengths increase signal to noise ratio and spatial res-
and osteophytes and radiographically-determined KL scores. They olution [16]. Therefore, we will initiate further studies using 3.0 T
classified osteophyte and subchondral cysts into size categories MR in the near future. Another limitation was that only some of
of 5 mm and 10 mm, respectively. BME greater than 1 cm in the cases (35%) were confirmed through arthroscopic examination
 H.-J. Park et al. / European Journal of Radiology 82 (2013) 112–117 117

and imaging findings were not correlated with arthroscopic find- [6] Link TM, Steinbach LS, Ghosh S, Ries M, Lu Y, Lane N, et al. Osteoarthritis:
ings. But this study did not focus on the accuracy of the MR in the MR imaging findings in different stage of disease and correlation with clinical
findings. Radiology 2003;226(2):373–81.
diagnosis of the cartilage injury itself. We intended to propose new [7] Noyes F, Stabler C. A system for grading articular cartilage lesions at
grading scale in the osteoarthritis of the knee joint on the bases of arthroscopy. American Journal of Sports Medicine 1989;17(4):505–13.
prominent MR findings. [8] Stoller DW. MR imaging in orthopaedic and sports medicine. 3rd ed. Baltimore,
MD: Lippincott Williams & Wilkins; 2007. p. 382.
In conclusion, these results suggest that our new MR grading [9] Spector TD, Cooper C. Radiographic assessment of osteoarthritis in popula-
system for osteoarthritis of the knee joint is reproducible and may tion studies: whither Kellgren and Lawrence? Osteoarthritis and Cartilage
be helpful for the grading of osteoarthritis of the knee without 1993;1(4):203–6.
[10] Cicuttini FM, Wluka AE, Forbes A, Wolfe R. Comparison of tibial cartilage vol-
requiring reference to plain radiography. Further studies concern-
ume and radiologic grade of the tibiofemoral joint. Arthritis and Rheumatism
ing the clinical utility and sensitivity of our new MR grading system 2003;48(3):682–8.
are needed. [11] Cicuttini F, Hankin J, Jones G, Wluka A. Comparison of conventional
standing knee radiographs and magnetic resonance imaging in assessing
progression of tibiofemoral joint osteoarthritis. Osteoarthritis and Cartilage
References 2005;13(8):722–7.
[12] Berthiaume MJ, Raynauld JP, Martel-Pelletier J, Labonte F, Beaudoin G, Blouch
[1] Harada Y, Tokuda O, Fukuda K, Shiraichi G, Motomura T, Kimura M, et al. DA, et al. Meniscal tear and extrusion are strongly associated with progres-
Relationship between cartilage volume using MRI and Kellgren–Lawrence sion of symptomatic knee osteoarthritis as assessed by quantitative magnetic
radiographic score in knee osteoarthritis with and without meniscal tears. resonance imaging. Annals of the Rheumatic Diseases 2005;64(4):556–63.
American Journal of Roentgenology 2011;196(3):W298–304. [13] Gunther KP, Sun Y. Reliability of radiographic assessment in hip and knee
[2] Hayes CW, Jamadar DA, Welch GW, Jannausch ML, Lachance LL, Capul DC, et al. osteoarthritis. Osteoarthritis and Cartilage 1999;7(2):239–46.
Osteoarthritis of the knee: comparison of MR imaging findings with radio- [14] Hunter DJ, Lo GH, Gale D, Grainger AJ, Guermazi A, Conaghan PG. The reliability
graphic severity measurements and pain in middle-aged women. Radiology of a new scoring system for knee osteoarthritis MRI and the validity of bone
2005;237(3):998–1007. marrow lesion assessment: BLOKS (Boston-Leeds Osteoarthritis Knee Score).
[3] Kellgren J, Lawrence J. Radiologic assessment of osteoarthritis. Annals of the Annals of the Rheumatic Diseases 2008;67(2):206–11.
Rheumatic Diseases 1957;16:494–501. [15] Peterfy CG, Guermazi A, Zaim S, Tirman PFJ, Miaux Y, White D, et al.
[4] Brandt K, Fife R, Raunstein E, Katz B. Radiographic grading of the severity of knee Whole-organ magnetic resonance imaging score (WORMS) of the knee in
osteoarthritis: relation of the Kellgren and Lawrence grade to a grade based on osteoarthritis. Osteoarthritis and Cartilage 2004;12(3):177–90.
joint space narrowing and correlation with arthroscopic evidence of articular [16] Kornaat PR, Ceulemans RT, Kroon HM, Riyazi N, Kloppenburg M, Carter WO,
cartilage degeneration. Arthritis and Rheumatism 1989;32(11):1584–91. et al. MRI assessment of knee osteoarthritis: knee osteoarthritis scoring system
[5] Ahlback S. Osteoarthritis of the knee: a radiographic investigation. Acta Radi- (KOSS)–inter-observer and intra-obsever reproducibility of a compartment-
ologica Diagnosis (Stockholm) 1968;(Suppl.):7–72. based scoring system. Skeletal Radiology 2005;34(2):95–102.