West Lake staging: A new staging system orchestrated by X-ray and MRI on knee osteoarthritis

Abstract

Purpose: To investigate the differences on X-ray and MRI among each stage of knee osteoarthritis (KOA) and further propose a new staging system called West Lake (WL) staging. Methods: A cross-sectional study was conducted on patients with KOA. Stage I, II, III, and IV were divided based on stepwise treatment strategy of Knee osteoarthritis (KOA). Joint space widths (JSW) were measured on X-rays, whereas cartilage injuries (CI) and bone marrow lesions (BML) were evaluated on MRI. The differences of them across the groups were calculated by T-test. Receiver operating characteristic (ROC) curves were rendered to obtain the areas under the curves (AUC), Youden index and corresponding cut-off points. Results: Eventually, there were significant differences on JSW, CI, and BML between stage II/III and III/IV, while no significant differences between stage I/II. In stage II/III, the AUC of JSW, CI, BML was 0.99, 0.76, 0.71 and the Youden index was 0.94, 0.38, 0.45, meanwhile the cut-off points were ≤5.1 mm, >1, >2. In stage III/IV, the AUC of JSW, CI, BML was 0.96, 0.79, 0.74 and the Youden index was 0.84, 0.58, 0.38, meanwhile the cut-off points were ≤3.2 mm, >3, >4. Conclusion: The WL staging was described as follows: Stage I, X-ray shows no joint space narrow, normal MRI or MRI shows cartilage degeneration and only 1 or 2 sections are involved in BML. Stage II, X-ray shows joint space narrow, MRI shows cartilage defect but no full-thickness cartilage defect, meanwhile 3 or 4 sections are involved in BML. Stage III, X-ray shows serious joint space narrow even JSW disappeared, MRI shows full-thickness cartilage defect, more than 4 sections are involved in BML.

Keywords

osteoarthritis knee staging system X-ray MRI

Introduction

Knee osteoarthritis (KOA) is a common chronic disease manifested by articular cartilage degradation, bone remodeling, osteophyte formation, and eventually the loss of joint space widths (JSW).¹ Indisputably, KOA affected the function of joints and reduced the quality of life, even led to disability, and finally incurred enormous social and economic burdens.^2–4

Although the long and wide implementation since its generation in 1957, Kellgren-Lawrence (KL) staging system of KOA still had some deficiencies.⁵ KL staging based on radiographs has gradually exerted the shortcoming in terms of the ambiguous differentiation due to its non-quantitative depiction on the morphological change especially in patients at the early stage of KOA, leading to the uncertainty in hierarchical remedy. In recent decades, despite numerous attempts based on X-rays on modifying this classification, none has been widely recognized; moreover, the non-quantitative traits of KL staging were hardly radically revolutionized.^6,7

With the popularity of MRI, the perception of KOA has been rapidly upgraded, pathological change including cartilage injuries (CI), and bone marrow lesions (BML) were realized their diagnostic significance.⁸ However, an integrated algorithm encompassing X-rays and MRI considering JSW, CI, BML has never been generated for better KOA grading.

Therefore, in this study, we aimed at investigating the differences on X-ray and MRI among each stage of KOA and further figuring out a new staging system which called West Lake (WL) staging orchestrated with JSW on X-rays and CI, BML on MRI.

Materials and methods

Patients’ enrollment

Patients diagnosed with KOA and treated in our institution from 2013 to 2018 were retrieved on the hospital information system (HIS). The inclusive criteria were (1) knee OA; (2) with both eligible MRI and X-rays document on knee; (3) with entire case record. The exclusive criteria were (1) with at least one musculoskeletal comorbidity, including rheumatoid arthritis, hemophilia arthritis, ankylosing spondylitis, gout, and other painful diseases of the knee joint; (2) with complications affecting knees, including psoriasis, syphilitic neuropathy, ochronotic arthritis, metabolic osteopathy, and acute trauma; (3) the radiographs were unclear.

Eventually, a total of 220 patients met the criteria. Generally, the mean age was 53.68 ± 13.39; the majority was female.

Staging based on treatment protocols

Referring to the principle of stepwise treatment strategy of KOA, the treatment methods were divided into four steps: basic therapy, drug therapy, restorative therapy, and reconstructive therapy. Stage I, II, III, and IV were divided according to which step was effective. Whether the treatment was effective or not was judged according to whether the symptoms of the affected knee are significantly relieved after treatment⁹ (Figure 1).

Figure 1.

Basic therapy included health education, exercise guidance, losing weight, rehabilitation training, and other non-drug treatment. Drug therapy included topical or oral NSAIDs and traditional Chinese medicine as well as oral glucosamine and intraarticular injection of glucocorticoids and sodium hyaluronate. Restorative therapy included intraarticular injection of mesenchymal stem cell and platelet-rich plasma as well as arthroscopy and osteotomy around the knee joint. Reconstructive therapy included partial and total knee arthroplasty. Patients who received basic therapy effectively were divided into stage I. Patients who received drug and basic therapy effectively were divided into stage II. Patients who received restorative, drug, and basic therapy effectively were divided into stage III. Patients who received reconstructive, restorative, drug, and basic therapy effectively were divided into stage IV.

X-rays

The X-rays were projected in weight-bearing standing position with knees in 20°of flexion and the feet in 5°of extorsion.¹⁰ A line was made parallel to the axis of the tibia through the points on the surface of the femoral condyle, and the distance between the two straight lines was measured and defined as JSW. The narrowest JSW was recorded.

MRI

CI and BML were measured on 3T MRI scanner.

According to the Recht scores, CI were divided into 4 grades: Grade 0, normal articular cartilage; Grade I, articular cartilage degeneration while no cartilage defect; Grade II, cartilage defect and the depth of cartilage defect was less than 50% of the full thickness; Grade III, the depth of cartilage defect was more than 50% while less than 100% of the full thickness; Grade IV, full-thickness cartilage defect and subchondral bone exposed.¹¹

It was difficult to measure the area of BML on MRI and the current WORMs scores were too complex, so referring to WORMs scores, the MRI of knee joint was divided into 15 regions, including lateral patellar section, medial patellar section, lateral–anterior femoral section, lateral–central femoral section, lateral–posterior femoral section, medial–anterior femoral section, medial–central femoral section, medial–posterior femoral section, lateral–anterior tibial section, lateral–central tibial section, lateral–posterior tibial section, medial–anterior tibial section, medial–central tibial section, medial–posterior tibial section and “S” section (the non-articulating portion of the tibial plateau beneath the tibial spines).¹² Meanwhile, the number of regions involved in BML was evaluated.

Statistical analysis

Power analysis was performed for sample size. According to previous studies, the prevalence (p) of KOA was about 30%, and a conservative estimate put it at 29%. The allowable error (d) was set at 6%. A 2-tailed test was required with an alpha level (α) of 0.05. Using the standard formula $n = \frac{Z_{α} \times p \times (1 - p)}{d^{2}}$ , the sample size (n) of our study was 220.

Stata/MP, version 16.0 was used for statistical analysis, and Receiver operating characteristic (ROC) curves were rendered by Medcalc, version 15.2.2. Statistical indicators included age, gender, JSW, the grade of CI, and the number of regions involved in BML. All data were presented as mean ± SD. The difference of those radiological parameters across the groups was calculated by T-test. Chi-square test was used to analyze the difference of gender among four groups. Logistic regression was used to adjust for confounders when there was a statistical difference in age or gender between the two groups. ROC curves were rendered to obtain the areas under the curves (AUC), maximum Youden index, and corresponding cut-off points. Agreement Index (KAPPA) was used to analyze the consistency of WL stages assessed by two qualified orthopedic surgeons. All reported p values are 2-tailed, with an alpha level of 0.05 indicating significance.

Results

Eventually, 25 patients were included in stage I; 62 patients were included in stage II; 109 patients were included in stage III; and 24 patients were included in stage IV.

Stage I/II

There was statistical difference on age (p < 0.05), while there was no statistical difference on gender (p > 0.05) between stage I and II (Table 1).

Table 1.

The results of T-test on age and chi-square test on gender among stage I, II, III, and IV.

Stage	Age	Gender
Stage	Age	Male	Female
I (n = 25)	45.08 ± 11.92	14	27
II (n = 62)	51.29 ± 13.37	11	35
III (n = 109)	55.25 ± 12.62	29	80
IV (n = 24)	61.67 ± 12.78	8	16
I/II	t, p/χ², p	−2.020, 0.047	1.109, 0.292
I/II	95% Confidence intervals	−12.323, −0.097	1.109, 0.292
II/III	t, p/χ², p	−1.929, 0.055	5.151, 0.023
II/III	95% Confidence intervals	−8.007, 0.092	5.151, 0.023
III/IV	t, p/χ², p	−2.251, 0.026	0.443, 0.506
III/IV	95% Confidence intervals	−12.059, −0.779	0.443, 0.506

There were no statistical differences on JSW, CI, and BML between stage I and II (p > 0.05) (Table 2).

Table 2.

The results of t-test among stage I, II, III, and IV on JSW, CI, and BML.

Stage		JSW	CI	BML
I (n = 25)		5.89 ± 0.99	0.68 ± 1.03	1.76 ± 1.30
II (n = 62)		6.13 ± 0.80	0.89 ± 0.99	1.65 ± 1.19
III (n = 109)		4.17 ± 0.58	1.99 ± 1.12	3.69 ± 2.87
IV (n = 24)		1.82 ± 1.32	3.33 ± 1.17	6.50 ± 3.23
I/II	t, p	−1.220, 0.226	−0.871, 0.386	0.397, 0.692
I/II	95% Confidence intervals	−0.654, 0.157	−0.680, 0.266	−0.460, 0.690
II/III	t, p	18.443, < 0.001	−6.456, < 0.001	−6.514, < 0.001
II/III	95% Confidence intervals	1.756, 2.177	−1.441, −0.766	−2.662, −1.424
III/IV	t, p	8.523, < 0.001	−5.28, < 0.001	−4.221, < 0.001
III/IV	95% Confidence intervals	1.780, 2.914	−1.845, −0.840	−4.130, −1.494

To rule out the effect of age on results, the logistic regressions included age and each indicator were done and showed there were still no statistical differences on JSW (B = 0.434, Wald = 2.164, p = 0.141, Exp(B) = 1.544, 95%CI (0.866, 2.752)), CI (B = 0.195, Wald = 0.601, p = 0.438, Exp(B) = 1.216, 95%CI (0.742, 1.992)), and BML (B = −0.099, Wald = 0.251, p = 0.617, Exp(B) = 0.905, 95%CI (0.613, 1.336)).

Stage II/III

There was statistical difference on gender (p < 0.05), while there was no statistical difference on age (p > 0.05) between stage II and III (Table 1).

There were statistical differences between stage II and III on JSW, CI as well as BML (p < 0.05). (Table 2)

To rule out the effect of gender on results, the Logistic regressions included gender and each indicator were done and showed there were still statistical differences on JSW (B = −5.440, Wald=26.361, p < 0.001, Exp(B) = 0.004, 95%CI (0.001, 0.035)), CI (B = 0.953, Wald = 27.172, p < 0.001, Exp(B) = 2.594, 95%CI (1.813, 3.712)) and BML (B = 0.451, Wald = 0.104, p < 0.001, Exp(B) = 1.569, 95%CI (1.280, 1.924)).

Respectively, to differentiate stage III from stage II, the cut-off points were JSW≤5.1 mm, CI>1, and BML>2 from ROC curves (Figure 2).

Figure 2.

The results of ROC curve between stage II and III. The ROC curve of JSW showed the AUC was 0.985, Youden index was 0.9424, cut-off point was 5.1 mm, meanwhile the corresponding sensitivity and specificity was 99.08% and 95.16% (a). The ROC curve of CI indicated the AUC was 0.759, Youden index was 0.3838, cut-off point was >1, simultaneously the corresponding sensitivity and specificity was 70.64% and 67.74% (b). The ROC curve of BML pointed the AUC was 0.712, Youden index was 0.4512, cut-off point was >2, simultaneously the corresponding sensitivity and specificity was 59.63% and 85.48% (c).

Stage III/IV

There was statistical difference on age (p < 0.05), while there was no statistical difference on gender (p > 0.05) between stage III and IV (Table 1).

There were statistical differences on JSW, CI, and BML between stage III and IV (p < 0.05) (Table 2).

To rule out the effect of age on results, the logistic regressions included age and each indicator were done and showed there were still statistical differences on JSW (B = −3.488, Wald =18.240, p < 0.001, Exp(B)=0.031, 95%CI (0.006, 0.151)), CI (B = 1.030, Wald = 17.021, p < 0.001, Exp(B) = 2.801, 95%CI (1.717, 4.570)) and BML (B = 0.264, Wald =11.685, p = 0.001, Exp(B) = 1.302, 95%CI (1.119, 1.514)).

Respectively, to differentiate stage IV from stage III, the cut-off points were JSW≤3.2 mm, CI>3, and BML>4 from ROC curves (Figure 3).

Figure 3.

The results of ROC curve between stage III and IV. The ROC curve of JSW showed the AUC was 0.956, Youden index was 0.8383, cut-off point was 3.2 mm, meanwhile the corresponding sensitivity and specificity was 87.50% and 96.33% (a). The ROC curve of CI indicated the AUC was 0.793, Youden index was 0.5799, cut-off point was >3, simultaneously the corresponding sensitivity and specificity was 70.83% and 87.16% (b). The ROC curve of BML pointed the AUC was 0.743, Youden index was 0.3781, cut-off point was >4, simultaneously the corresponding sensitivity and specificity was 70.83% and 66.97% (c).

Agreement index (KAPPA)

The results of one surgeon’s WL staging assessment were 53 patients with stage I, 98 patients with stage II, and 69 patients with stage III. The results of another surgeon’s WL staging assessment were 55 patients with stage I, 115 patients with stage II, and 50 patients with stage III. The Kappa coefficient of WL stages assessed by two qualified orthopedic surgeons was 0.76 (Table 3).

Table 3.

The results of Agreement Index (KAPPA) in WL staging system.

	Expected agreement, %	Agreement, %	Kappa	Z	p
WL staging	85.00	36.44	0.76	236.81	<0.05

Discussion

To our knowledge, this was the first study combining MRI and X-ray features to classify KOA grades. Although X-ray was always the most available and convenient tool in the world for screening OA, previous literature uncovered intra-rater reliability ranging from 0.67 to 0.73 and hardship in distinguishing KL II and III by means of KL staging.^13–15 Guermazi¹⁶ proposed that KL staging had little information about the rate of cartilage degeneration through the comparison between same KOA patient’s X-ray and MRI. So, it was necessary to propose a new staging system so that every surgeon could agree on the severity of KOA.

For there was still a lack of an effective quantitative method to evaluate the number of osteophytes, the description of osteophytes in WL staging was removed.¹⁷ On the contrary, JSW was easier to be measured and confirmed that had a significant association with the severity of KOA as well as knee pain.^18,19

The mechanism of KOA could be summarized as follows: Abnormal mechanical loads were applied to the cartilage and transmitted to the subchondral bone, which led to microfractures of the subchondral bone, then an inflammatory response occurred and immune cells were recruited, which could activate bone growth factor and promotes subchondral bone reconstruction which led to the formation of BML finally.^20–22 Current studies suggest that the pathogenesis of KOA is mainly the lesion of the cartilage layer, meanwhile BML could inform on the degenerative state within the osteochondral unit.^23,24 In addition, BML has been known as a phenomenon to assess the severity and prognosis of KOA.^25,26 So, CI and BML were included in the study.

In the study, we found there were no statistical differences on JSW, CI, and BML between stage I and II which let us to make the decision of merging them together. This may be because the JSW difference in early KOA is relatively small. Besides, osteophytes are a main tool to identify KL I or II in KL staging system, neglect of osteophytes may also play a role.¹⁵ The possible reason for the failed correlation of CI between stage I and II was that maybe there was significant difference in the level of cartilage degeneration, it could not be shown by our CI scores.²⁷ It is well known that the formation of BML was closely related to cartilage injuries, with mild cartilage damage, the vicious cycle of BML formation was less severe which caused the mild BML in stage I/II.²⁸ In clinical, surgeons generally decide whether to use drug therapy or not through the patient’s pain, function, and other aspects in early KOA, so it is feasible to combine stage I and II.²⁹ In summary, the imaging findings (JSW, CI, BML) of stage I and stage II are similar, and we can judge whether to use basic therapy alone or combination of basic therapy with drug therapy according to the patient’s pain degree.

Dramatically, there were significant differences on JSW, CI, and BML between stage II and III as well as III and IV. ROC curves of JSW, CI, and BML showed the AUC was 0.99, 0.76, 0.71 in stage II/III and 0.96, 0.79, 0.74 in stage III/IV respectively which suggested high diagnostic value and JSW was the best indicator among them. Theoretically, a cut-off point with 100% sensitivity and 100% specificity is the most appropriate diagnostic value. While in practice, sensitivity decreases as specificity increases, and vice versa. So, a balance must be struck between sensitivity and specificity to obtain an optimal cut-off point. The Youden index is equal to (sensitivity + specificity −1) and when picked the maximum Youden index which meant optimal sensitivity and specificity, JSW ≤5.1 mm, CI >1, and BML >2 were the cut-off points to differentiate stage II and III, and JSW ≤3.2 mm, CI >3 and BML>4 were the cut-off points to distinguish stage IV from III. Although there was error in joint clearance measurements and people in different regions maid have different JSW, our study confirmed that compared with stage I and II, the JSW of stage III were narrow and JSW of stage IV were the most serious one, which has also been verified by previous studies.^30,31 At present, no studies have explored the difference of CI between stage II and III, while it was known that narrowing of the joint space could reflect the degree of cartilage wear, so cartilage damage was gradually exacerbated with the increase of KL staging and full-thickness wear of the cartilage was often present in stage IV for the reason of extreme narrowing or loss of JSW.³² Besides, it has been confirmed that partial and total knee arthroplasty is always used on patients with full-thickness cartilage defect.¹ Since we used the number of regions involved in BML as an indicator, no data were found in the current literature to support our results. However, we have known that the area of BML were associated with the progression of KOA, meanwhile BML size was associated with longitudinal cartilage loss.^33,34

Agreement Index (KAPPA) is a tool to verify the consistency of WL staging between two qualified orthopedic surgeons. In further research, the KAPPA value was 0.76 which indicated their conclusions were highly consistent. This result showed that WL staging can be well unified among surgeons, which will be of great help to clinical treatment and academic discussion.

Our study still had several limitations. First, we did not take into consideration of synovium, joint effusion, ligaments, meniscus, and other imaging on MRI since they had no quantitative or semi-quantitative evaluative methods. Second, given the inadequate following period, we could not verify the validity of the new staging system; however, we will publish our short-middle term results of this grading later. Third, patellofemoral osteoarthritis was not included in this study, for patellofemoral osteoarthritis often exists as an independent concept while tibiofemoral was still the main body of KOA. For example, the KL staging system also only includes the tibiofemoral JSW. Certainly, we hope to conduct further statistical analysis of the patellofemoral JSW to make the staging system more complete in the following series of studies.

Conclusion

In summary, The WL staging was described as follows: Stage I: X-ray shows no joint space narrow, normal MRI or MRI shows cartilage degeneration, and only 1 or 2 sections are involved in BML; Stage II: X-ray shows joint space narrow, MRI shows cartilage defect but no full-thickness cartilage defect, meanwhile 3 or 4 sections are involved in BML; Stage III: X-ray shows serious joint space narrow even JSW disappeared, MRI shows full-thickness cartilage defect, more than 4 sections are involved in BML (Table 4).

Table 4.

The WL staging system.

Stage	X-ray (JSW)	MRI
Stage	X-ray (JSW)	CI	BML area
I	No joint space narrow	None or cartilage degeneration	No more than 2 sections
II	Joint space narrow	Cartilage defect but no full-thickness cartilage defect	3 or 4 sections
III	Serious joint space narrow even JSW disappeared	Full-thickness cartilage defect	More than 4 sections

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was in part supported by the National Natural Science Foundation of China (grant nos. 81873324) and the project of Administration of Traditional Chinese Medicine of Zhejiang Province of China (grant no. 2018ZZ011).

ORCID iD

Peijian Tong

Appendix

The MRI of knee joint was divided into 15 regions, including lateral patellar section, medial patellar section, lateral–anterior femoral section, lateral–central femoral section, lateral–posterior femoral section, medial–anterior femoral section, medial–central femoral section, medial–posterior femoral section, lateral–anterior tibial section, lateral–central tibial section, lateral–posterior tibial section, medial–anterior tibial section, medial–central tibial section, medial–posterior tibial section and “S” section (the non-articulating portion of the tibial plateau beneath the tibial spines). Meanwhile, the number of regions involved in BML was evaluated as BML area.

References

Kolasinski

Neogi

Hochberg

, et al. 2019 American college of rheumatology/arthritis foundation guideline for the management of osteoarthritis of the hand, hip, and knee. Arthritis Care Res 2019; 72(2): 149–162.

Sharma

Kwoh

Lee

, et al. Development and validation of risk stratification trees for incident slow gait speed in persons at high risk for knee osteoarthritis. Ann Rheum Dis 2019; 78(10): 1412–1419.

Tang

Wang

Zhan

, et al. The prevalence of symptomatic knee osteoarthritis in China: results from the China health and retirement longitudinal study. Arthritis Rheumatol 2016; 68(3): 648–653.

Hunter

Bierma-Zeinstra

. Osteoarthritis. Lancet 2019; 393(10182): 1745–1759.

Kellgren

Lawrence

. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957; 16(4): 494–502.

Schiphof

Boers

Bierma-Zeinstra

SMA

. Differences in descriptions of Kellgren and Lawrence grades of knee osteoarthritis. Ann Rheum Dis 2008; 67(7): 1034–1036.

Schiphof

De Klerk

Kerkhof

, et al. Impact of different descriptions of the Kellgren and Lawrence classification criteria on the diagnosis of knee osteoarthritis. Ann Rheum Dis 2011; 70(8): 1422–1427.

Linklater

Hunter

. Imaging in knee osteoarthritis. Curr Opin Rheumatol 2017; 29(1): 86–95.

Bruyère

Honvo

Veronese

, et al. An updated algorithm recommendation for the management of knee osteoarthritis from the European society for clinical and economic aspects of osteoporosis, osteoarthritis and musculoskeletal diseases (ESCEO). Semin Arthritis Rheum 2019; 49(3): 337–350.

10.

Luyten

Denti

Filardo

, et al. Definition and classification of early osteoarthritis of the knee. Knee Surg Sports Traumatol Arthrosc 2012; 20(3): 401–406.

11.

Recht

Kramer

Marcelis

, et al. Abnormalities of articular cartilage in the knee: analysis of available MR techniques. Radiology 1993; 187(2): 473–478.

12.

Peterfy

Guermazi

Zaim

, et al. Whole-Organ Magnetic Resonance Imaging Score (WORMS) of the knee in osteoarthritis. Osteoarthr Cartil 2004; 12(3): 177–190.

13.

Duryea

Neumann

Niu

, et al. Comparison of radiographic joint space width with magnetic resonance imaging cartilage morphometry: analysis of longitudinal data from the osteoarthritis Initiative. Arthritis Care Res 2010; 62(7): 932–937.

14.

Kohn

Sassoon

Fernando

N D

. Classifications in brief: Kellgren-Lawrence classification of osteoarthritis. Clin Orthopaedics Relat Res 2016; 474(8): 1886–1893.

15.

Culvenor

Engen

Øiestad

, et al. Defining the presence of radiographic knee osteoarthritis: a comparison between the Kellgren and Lawrence system and OARSI atlas criteria. Knee Surg Sports Traumatol Arthrosc 2015; 23(12): 3532–3539.

16.

Guermazi

Roemer

Burstein

, et al. Why radiography should no longer be considered a surrogate outcome measure for longitudinal assessment of cartilage in knee osteoarthritis. Arthritis Res Ther 2011; 13(6): 247.

17.

Hakky

Jarraya

Ratzlaff

, et al. Validity and responsiveness of a new measure of knee osteophytes for osteoarthritis studies: data from the osteoarthritis initiative. Osteoarthr Cartil 2015; 23(12): 2199–2205.

18.

Fang

Wang

, et al. New variables for measuring joint space width to evaluate knee osteoarthritis. Chin Med J 2011; 124(23): 3886–3890.

19.

Muraki

Akune

En-Yo

, et al. Joint space narrowing, body mass index, and knee pain: the ROAD study (OAC1839R1). Osteoarthr Cartil 2015; 23(6): 874–881.

20.

Chen

, et al. Subchondral trabecular rod loss and plate thickening in the development of osteoarthritis. J Bone Mineral Res 2018; 33(2): 316–327.

21.

Vaysbrot

Osani

Musetti

M-C

, et al. Are bisphosphonates efficacious in knee osteoarthritis? A meta-analysis of randomized controlled trials. Osteoarthr Cartil 2018; 26(2): 154–164.

22.

Weber

Chan

PMB

Wen

. Do immune cells lead the way in subchondral bone disturbance in osteoarthritis? Prog Biophys Mol Biol 2019; 148: 21–31.

23.

Showiheen

SAA

Sun

, et al. Application of metabolomics to osteoarthritis: from basic science to the clinical approach. Curr Rheumatol Rep 2019; 21(6): 26.

24.

Muratovic

Findlay

Cicuttini

, et al. Bone marrow lesions in knee osteoarthritis: regional differences in tibial subchondral bone microstructure and their association with cartilage degeneration. Osteoarthr Cartil 2019; 27(11): 1653–1662.

25.

Lynch

Roemer

Nevitt

, et al. Comparison of BLOKS and WORMS scoring systems part I. Cross sectional comparison of methods to assess cartilage morphology, meniscal damage and bone marrow lesions on knee MRI: data from the osteoarthritis initiative. Osteoarthr Cartil 2010; 18(11): 1393–1401.

26.

Felson

Lynch

Guermazi

, et al. Comparison of BLOKS and WORMS scoring systems part II. Longitudinal assessment of knee MRIs for osteoarthritis and suggested approach based on their performance: data from the Osteoarthritis Initiative. Osteoarthr Cartil 2010; 18(11): 1402–1407.

27.

Amin

LaValley

Guermazi

, et al. The relationship between cartilage loss on magnetic resonance imaging and radiographic progression in men and women with knee osteoarthritis. Arthritis Rheum 2005; 52(10): 3152–3159.

28.

Tanamas

Wluka

Pelletier

J-P

, et al. The association between subchondral bone cysts and tibial cartilage volume and risk of joint replacement in people with knee osteoarthritis: a longitudinal study. Arthritis Res Ther 2010; 12(2): R58.

29.

Madry

Kon

Condello

, et al. Early osteoarthritis of the knee. Knee Surg Sports Traumatol Arthrosc 2016; 24(6): 1753–1762.

30.

Ratzlaff

Ashbeck

Guermazi

, et al. A quantitative metric for knee osteoarthritis: reference values of joint space loss. Osteoarthr Cartil 2018; 26(9): 1215–1224. DOI: 10.1016/j.joca.2018.05.014.

31.

Hellio Le Graverand

M-P

Mazzuca

Duryea

, et al. Radiographic grading and measurement of joint space width in osteoarthritis. Rheum Dis Clin North America 2009; 35(3): 485–502.

32.

Simoni

Jamali

Albert

, et al. Minimum joint space width (mJSW) of patellofemoral joint on standing “skyline” radiographs: test-retest reproducibility and comparison with quantitative magnetic resonance imaging (qMRI). Skeletal Radiol 2013; 42(11): 1573–1582.

33.

Felson

McLaughlin

Goggins

, et al. Bone marrow edema and its relation to progression of knee osteoarthritis. Ann Intern Med 2003; 139(5 Pt 1): 330–336.

34.

Pang

Driban

Destenaves

, et al. Quantification of bone marrow lesion volume and volume change using semi-automated segmentation: data from the osteoarthritis initiative. BMC Musculoskelet Disord 2013; 14: 3.