Diagnostic accuracy of magnetic resonance imaging (MRI) for symptomatic knee osteoarthritis: a scoping review

Introduction

Osteoarthritis (OA) is a complex disease that affects the articular cartilage, compromising joint integrity. The estimated incidence of this condition in 2020 was around 595 million people worldwide, affecting 7.6% of the world’s population (1). Knee OA causes an extreme economic burden, with prevalence varying geographically. Developed economies such as North America and Europe, with ageing populations and increasing obesity prevalence, are particularly affected (2). To date, there is no curable treatment, and only symptomatic relief measures can be offered. The characteristic anatomical and histological lesions of OA vary between phenotypes and may involve the articular or subchondral bone, ligaments, synovium, joint capsule, and periarticular muscle structures (3). This highlights the need for early detection of knee OA phenotypes and monitoring of disease progression using more sensitive and accurate diagnostic methods.

Magnetic resonance imaging (MRI), which provides detailed images of soft tissue structures, is increasingly being used in the diagnosis of knee OA. MRI can detect structural changes in the knee without clinical symptoms in asymptomatic individuals, providing valuable information of the knees morphological modifications associated with aging. In contrast to X-rays, which mainly show changes in the bone, MRI can detect early changes in the cartilage and other soft tissues, potentially allowing for earlier diagnosis and treatment intervention (4). The European League Against Rheumatism (EULAR) guidelines provide general management recommendations and considerations for musculoskeletal diseases, usually supported by imaging findings (5). In the context of knee OA, radiography has been used to assess the structural severity of OA using the Kellgren and Lawrence (K-L) scale, which includes the presence of osteophytes and joint space narrowing in its assessment and establishes 5 categories (from absent to severe OA) (6).

Although this scale has been useful for many decades, the increasing use and higher tissue resolution of MRI is opening a debate on how to better approach the diagnosis and classification of knee OA in clinical practice. In fact, the American College of Rheumatology (ACR) currently uses MRI criteria specifically related to the classification of knee OA and provides standardized guidance for identifying people with the condition (7,8). The ACR and the National Institute for Health and Care Excellence (NICE) recommend diagnosing OA based on clinical criteria and emphasize the need for further examinations (9).

There are several scoring systems and development techniques in MRI that need to systematically evaluate overall efficacy and prioritise the accuracy assessments of each technique. It could be essential to provide criteria for assessing various structural features of the knee joint. Such as; Whole-Organ Magnetic Resonance Imaging Score (WORMS) has undergone a systematic development process involving the identification of MRI-based parameters and their integration into clinical practice. Other scoring systems, Boston Leeds Osteoarthritis Knee Score (BLOKS), provide a global view of OA, including aspects like inflammation and ligaments (10). Also, the MRI Osteoarthritis Knee Score (MOAKS) has a broader range of scores, including ligaments, subchondral cysts (11).

Moreover, the role of MRI for certain treatments in OA-related conditions has been emphasized in the literature, reinforcing the usefulness of this imaging modality in clinical practice and warranting the need for further exploring its utility in knee OA (12). One of the most challenging problems that radiologists and clinicians still need to face in the management of knee OA lies in the relatively poor correlation between imaging findings -particularly those from X-rays or even computed tomography (CT) and clinical symptoms. For instance, it has been found that clinical symptoms occur in 40% of patients with moderate radiographic knee OA (X-rays KL-classification stage II–III) and in 60% of patients with severe knee OA. A number of studies have explored how MRI findings can improve our understanding of the connection between imaging findings and clinical symptoms in knee OA.

However, our current knowledge on the topic is still limited, partly because the reference standards used for assessing the reliability of MRI in the diagnosis of knee OA are heterogeneous and because currently available assessment tools present methodological limitations. A comprehensive synthesis of the evidence on the diagnostic test accuracy (DTA) of MRI for OA is essential for developing robust clinical guidelines (13). The current lack of standardized diagnostic criteria presents challenges, including the lack of standardized methodological quality assessment tools specifically designed for scoping reviews. Furthermore, existing guidelines for DTA of MRI in symptomatic knee OA reveal substantial gaps and inconsistencies in the current diagnostic criteria and imaging protocols (14,15). This review aimed to identify, assess and summarise the available systematic reviews on DTA of MRI in patients with knee OA, applying a novel methodological quality assessment tool. We present this article in accordance with the PRISMA-ScR reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-24-1544/rc).

Methods

This review was registered on the Open Registries Network (OSF), https://doi.org/10.17605/OSF.IO/HDNE6 (dated: June 10, 2024).

Information sources, search strategy and study selection

A systematic search of seven electronic databases (MEDLINE/PubMed, Web of Science Core Collection, Scopus, Google Scholar, EMBASE, CINAHL, and Sport Discus) was performed to identify systematic reviews and meta-analyses published from database inception until May 30, 2024. The search strategy combined medical subject headings (MeSH) terms and keywords related to “MRI”, “diagnostic accuracy”, “sensitivity”, “specificity”, and “knee osteoarthritis” (Appendix 1). The inclusion criteria encompassed systematic reviews and meta-analyses focusing on the DTA of MRI in adults aged 18–70 years with symptomatic knee OA. Studies included participants diagnosed with knee OA and evaluated MRI test interventions to diagnose knee OA. In the screening process, all titles and abstracts of interest were examined by two authors and those which did not meet the eligibility criteria were excluded. If consensus could not be reached, a third independent screener was consulted.

The exclusion criteria were other types of reviews, individual primary studies not focusing on MRI diagnosis or knee OA, articles published in languages other than English, and studies with incomplete data. The excluded studies were those that did not provide data suitable for evaluating test accuracy, such as sensitivity, specificity, or likelihood ratios. For instance, articles focusing on treatment effectiveness, prognosis, or other unrelated clinical outcomes were considered irrelevant, as they do not assess the performance of diagnostic tests. Additionally, studies that did not compare test results to a recognized reference standard or failed to provide the essential data needed to calculate sensitivity and specificity were also excluded. The excluded articles based on full-text review are summarized in Appendix 2.

Data extraction and methodological quality assessment

Data were extracted on study characteristics, patient populations, MRI techniques, reference standards, risk of bias methods and DTA metrics. Data extraction was performed independently by two reviewers using a standardized form. The methodological quality of the included reviews evaluated by using a modified version of the Measurement Tool to Assess Systematic Reviews (AMSTAR 2). This new tool was tailored using existing guidance (16) to critically appraise DTA systematic reviews. The 16 items included in the original AMSTAR 2 were retained in the modified version. Each item was rated, and critical weaknesses were identified to determine the overall confidence in the results of the review (17). Further clarifications regarding the modified AMSTAR 2 applied in this study can be consulted in Appendix 3.

Data synthesis

We conducted a data synthesis by systematically compiling and analyzing DTA metrics from included systematic reviews evaluating MRI for knee OA. The key characteristics of each review, including population demographics, reference standards and diagnostic criteria, were meticulously extracted and reported. This synthesis was conducted to identify any methodological weaknesses and sources of heterogeneity. Corrected covered area (CCA) was calculated to evaluate the overlap between reviews (18).

Results

Search and study selection

The search yielded 8,586 records, after which 978 duplicates and 7,537 irrelevant records were removed based on titles and abstracts. This left 73 full-text articles to be assessed for eligibility. One study was excluded for irrelevance, 67 did not meet our criteria for a systematic review and meta-analysis of DTA in knee OA, and two were not available. The three included studies were two systematic reviews and meta-analysis and one systematic review without meta-analysis, which encompassed a total of 99 original studies (7,530 participants) (Figure 1) (19-21).

Figure 1 Flowchart of studies of systematic review and meta-analysis in diagnostic test accuracy of MRI in patients with knee osteoarthritis patients. MRI, magnetic resonance imaging.

Characteristics of the studies

The study characteristics showed differences in populations, range of years covered, number of primary studies, and countries where the studies were conducted. Additionally, there were variations in study designs and assessment methods used. They incorporated various methodological designs, such as diagnostic studies with consecutive patients, prospective and retrospective studies. In addition, they included expert opinions and case-control studies, as well as data from larger clinical trials or cohorts study types. The first study used histology and a variety of other methods as reference standards, the second used arthroscopy or open surgery, and the third also relied on arthroscopy, each yielding different reported sensitivities, specificities, and overall diagnostic performance. The overall CCA was 7.41%, varying from 5.05 to 9.09 across the included reviews. The modified AMSTAR 2 ratings further highlighted the variability in methodological quality across these reviews, with ratings ranging from critically low to low quality (Table 1).

Methodological quality assessment: original AMSTAR 2 checklist

The result of methodological quality assessment based on the original AMSTAR 2 identified deficiencies in research questions and inclusion criteria, Participants, Intervention, Comparison, Outcome (PICO) components (2/3, 66.7%), protocol registration (1/3, 33.3%), heterogeneity exploration (1/3, 33.3%), providing a list of excluded studies (2/3, 66.7%), sources of funding and potential conflict of interest (2/3, 66.7%), and risk of bias assessment (2/3, 66.7%) (Table 2).

Methodological quality assessment: modified AMSTAR 2 checklist

The modified AMSTAR 2 checklist incorporated additional criteria tailored to diagnostic accuracy reviews, leading to the identification of further areas of concern. This assessment highlighted issues in research questions and inclusion criteria components and protocol registration deficiency (1/3, 33.3%), reference standard partial deficiency (2/3, 66.7%), justifying and providing a list of excluded studies deficiencies (2/3, 66.7%), use of appropriate statistical methods (1/3, 33.3%), risk of bias assessment (1/3, 33.3%), heterogeneity exploration (1/3, 33.3%), sources of funding and potential conflict of interest (2/3, 66.7%), risk of bias assessment using QUADAS-2 (2/3, 66.7%), evaluating and discussing the impact of risk of bias in individual studies (1/3, 33.3%).

The reference standards used in the first review were histology, arthroscopy, radiography, CT, clinical evaluation, and direct visual inspection. The second study used arthroscopic or open surgery, while the third study used only arthroscopy. In two reviews with meta-analysis the accuracy of MRI was: sensitivity 61% [95% confidence interval (CI), 53–68%] and specificity 82% (95% CI, 77–87%); and sensitivity 74% (95% CI, 71–77%) and specificity 95% (95% CI, 94–95%). In the third review, without meta-analysis, the sensitivity and specificity of MRI ranged 26% to 96% and from 50% to 100%, respectively (Table 3).

Discussion

The AMSTAR 2 tool was developed to assess the methodological quality of systematic reviews that focus on RCTs and non-randomized studies of interventions. However, it may not be the most appropriate instrument for evaluating reviews of DTA studies, such as those examining the accuracy of MRI for knee OA. The PICO framework in AMSTAR 2 is designed to evaluate interventions, whereas diagnostic studies concentrate on diagnostic tests and reference standards. This assessment tool does not include items designed to evaluate the methodology of DTA studies, being less appropriate to determine their quality. In addition, the instrument includes items that have been specifically designed for the assessment of systematic reviews of other study methodologies, such as RCTs. As a result, it may not be the most suitable tool for the assessment of DTA studies.

In DTA studies, the primary objective is to assess the accuracy of a diagnostic test (e.g., MRI) in identifying a disease. This is achieved by measuring metrics such as sensitivity and specificity (against a reference standard) (22). In this context, AMSTAR 2 is of limited value, as it does not encompass the inclusion of elements pertinent to diagnostic accuracy, such as a comprehensive delineation of the diagnostic index test, the rationale and justification for the selection of reference standards, and the utilization of specific risk of bias tools, such as QUADAS-2.

Main findings

This scoping review synthesizes the available evidence in the MRI assessment of knee OA, summarizing the potential and limitations of this imaging technique. We included 3 systematic reviews encompassing 99 original studies with over 7,530 participants and found a high sensitivity and specificity of MRI for the assessment of knee OA. However, the methodological quality of the reviews was variable, which weakens the conclusions that can be drawn from our findings. To address this issue, we developed a modified version of the AMSTAR 2 checklist adapted to the particular needs of DTA studies.

The quality of the included reviews was assessed using the AMSTAR 2 tool, and further evaluated using a modified AMSTAR 2 checklist specific to DTA studies. The first included review was rated as “critically low” by the original AMSTAR 2 and as “low quality” by the modified AMSTAR 2 checklist, i.e., after improving the assessment tool to be relevant to diagnostic test components. This discrepancy indicates that while the general quality is poor, the diagnostic-specific criteria show some strengths not captured by AMSTAR 2. The second study was rated as “low quality” by the original AMSTAR 2 and as “high quality” by the modified AMSTAR 2 checklist, indicating that the modified checklist’s focus on diagnostic accuracy components revealed a higher methodological quality review. Finally, the third study was consistently rated “critically low” in both assessments, indicating pervasive methodological issues.

There is considerable variation in the reported sensitivities and specificities for MRI, reflecting differences in the reference standards. The heatmap showing the CCA for structural missingness provided a visual representation of which primary studies were included in which systematic reviews. Systematic review 1 showed a slightly above-the-low overlap threshold with a CCA of 5.05%, while systematic reviews 2 and 3 showed moderate overlap (9.09% and 8.08, respectively) (Figure 2).

Figure 2 This diagram shows the CCA adjusted for structural missingness in the studies included in this review. CCA, corrected covered area.

Strengths and limitations

The primary limitation of this review is the critically low quality of the included studies, as assessed by the modified AMSTAR 2. This limitation underscores the need for more robust and methodologically sound research in this area. The persistent deficiencies in some reviews, including inadequate details and descriptions of the diagnostic index test and reference standards which may be not fully improved using the modified tool. In sum, our findings stress ongoing challenges in methodology in the diagnostic accuracy studies, including the need to improve risk of bias assessment of the included articles, and standardizing criteria for evaluating diagnostic methods, ultimately to obtain higher-quality reviews that contribute effectively to evidence-based medicine.

Comparison with other studies

Understanding the etiology and pathology of OA remains not well understood. It is usually classified into primary (idiopathic) and secondary (post-traumatic) forms, with idiopathic risk factors (23). Many institutions, including ACR, NICE, and EULAR, support guidelines primarily based on clinical assessment, with minimal contribution of radiographic findings (24,25).

Nevertheless, the use of MRI in knee OA is growing due to its ability of depicting undetectable features by other imaging modalities, potentiating the identification of lesions even in asymptomatic patients or in the presence of normal radiographs (26). Therefore, the inclusion of MRI as a diagnostic tool in knee OA requires validated criteria. One of the first attempts for this purpose used a Delphi method, which involves experts engaging in a structured process to establish a consensus or standard regarding specific aspects of diagnosis. As a result of this method, major and minor criteria to diagnose knee OA were proposed. For instance, the major criteria were the presence of osteophytes and cartilage loss, strengthening the importance of bone and cartilage in the diagnosis of OA (27).

The current guidelines for the diagnostic test of MRI in knee OA cases exhibit substantial gaps and inconsistencies, which highlight the need for improved diagnostic criteria and imaging protocols (14,15). It is still needed to clarify and validate the quantifiable changes detected on MR to make it the standard index tests for knee OA.

Challenges and methodological issues

Despite the advanced imaging capabilities of MRI, there is insufficient conclusive evidence to prioritize and accurately define these index tests. The systematic reviews examined were predominantly of low quality, revealing significant methodological weaknesses and gaps in the literature. There is a notable unclear methodology for conducting and evaluating systematic reviews and meta-analyses within the DTA study design field, as shown by the study characteristics (Table 1). In the available evidence of DTA studies, the reference standard for the target disease is either incomplete or not available.

When evaluating diagnostic accuracy, it is crucial to recognize that the quality of existing systematic reviews is often suboptimal. The exhibit reveals methodological inadequacies, including the inconsistent application of reference standards and insufficient reporting of results. Furthermore, the diagnostic criteria for knee OA are currently insufficient, primarily due to the lack of a universal consensus on the criteria that should be included.

In the absence of universal agreement to create a criterion to evaluate the interventions of MRI in knee OA case, due to lack of consensus and methodology assessment of the index test diagnosis. This leads to unclear guidance in the medical field and a lack of medical based evidence, which we faced in our case to evaluate the sensitivity and specificity of MRI in patients with knee OA. Although, efforts exist to support the reporting of systematic reviews of DTA using best practice methodological assessment, focus on the challenges for MRI imaging (28).

The sensitivity and specificity values range highlight the variability in the diagnostic performance of MRI for knee OA. High specificity indicates that MRI can correctly identify those cases without the disease, which is crucial to avoid unnecessary treatment. However, the variable sensitivity suggests inconsistency in detecting all true cases of OA. Factors influencing this variability include different magnet strength, coils and sequences employed, with higher performance for cartilage in 3 Tesla magnets (29). These findings underline the importance of using adequate reference standards for a comprehensive evaluation of diagnostic tests.

The modified AMSTAR 2 for evaluating diagnostic tests includes specific changes which are intended to address the unique challenges associated with diagnostic accuracy studies. While the traditional AMSTAR is focused on systematic reviews of interventions, the modified version incorporates elements that are crucial for assessing the quality of DTA studies. These efforts aim to provide a standardized approach to evaluating DTA studies, enhancing the consistency and reliability of systematic reviews. The modified AMSTAR contributes to improve diagnostic accuracy and, consequently, better patient management and outcomes. However, the methodological quality of the reviews is low, which undermines the conclusions that can be drawn from our findings. To further address this issue, we developed a modified version of the AMSTAR 2 checklist adapted to the particular needs derived from DTA studies.

The small sample size (i.e., 3 systematic reviews) in this scoping review reflects and limits the generalizability of findings, although the use of a modified AMSTAR 2 checklist still represents a significant methodological improvement. AMSTAR 2 originally does not address key aspects specific to diagnostic accuracy studies, and when used in the case of DTA, it would likely be under-assessed by not having value of some questions components. The modified AMSTAR 2 checklist improves the evaluation of diagnostic test studies by including specific components in the checklist’s questions relevant to diagnostics, such as detailed descriptions of the index test and reference standards. It also uses the QUADAS-2 tool to assess bias, providing a more accurate analysis of study quality. These help to ensure systematic reviews of diagnostic tests are assessed more effectively and reliably.

Implications for research

Enhancing the quality assessment framework, it is crucial to improve key aspects like research questions and inclusion criteria, ensure the relevant diagnostic components. This means accurately describing the diagnostic index test, justify the reference standard with adequate details, assess the research questions and inclusion criteria providing enough information on included studies, and correctly incorporating QUADAS-2 risk of bias assessment tool for this type of study. Specifically, the studies showed weaknesses in consistently registering their protocols and adequately assessing the risk of bias. For example, the first and third studies were rated as critically low quality due to these deficiencies, while the second was rated as a low-quality review.

One significant improvement observed with the modified checklist was the ability to detect deficiencies related to the description of the diagnostic index test, specification of the comparator or reference standard, and definition of accuracy measures. This resulted in a more nuanced assessment of methodological quality. For instance, the second study was upgraded to a high-quality review under the modified checklist due to better adherence to these diagnostic-specific criteria, while the first was rated as a low-quality review and the third study remained as of critically low quality. By addressing these key aspects, it is possible to significantly enhance the quality of the evaluation framework for diagnostic tests. Accurate descriptions and justifications are crucial for ensuring that the findings of systematic reviews are both relevant and applicable to clinical practice.

The novelty of our study lies in the adaptation of the AMSTAR 2 checklist to better structure the evaluation of DTA systematic review. The modified AMSTAR 2 checklist adapted for DTA study designs provided a more comprehensive assessment by identifying additional methodological quality aspects related to diagnostic test reviews.

To improve the quality assessment, we have identified the critical components that justify the index test, reference standard, classification, research questions and inclusion criteria, heterogeneity exploration, and risk of bias assessment.

Future directions

Future research should focus on developing standardized criteria for MRI-based diagnosis of knee OA. The modified checklists used in this review provided valuable insights, although they may not perfectly capture all aspects necessary for assessing DTA studies and could benefit from further refinement. There is also a need for high-quality, large-scale studies to confirm the diagnostic accuracy of MRI and to ensure that findings are generalizable across diverse populations and clinical settings. Investigating the potential of advanced and novel MRI techniques, such as quantitative MRI, could provide more detailed and accurate assessments of knee OA, mainly in early phases with no or mild morphological changes. Finally, exploring the cost-effectiveness of these MRI techniques in routine clinical practice remains necessary.

Conclusions

This scoping review offers a comprehensive synthesis of the evidence pertaining to DTA of MRI for knee OA. Our findings highlight the need to clarify the role of MRI in knee OA diagnosis and establish its priority over traditional methods as a gold standard. The current literature indicates a pressing need for improved quality and standardization in future studies to enhance the reliability of MRI as a diagnostic tool. In particular, there is an urgent need for more robust research methodologies that incorporate detailed diagnostic criteria, address sources of heterogeneity, and ensure thorough risk of bias assessments.

Our adapted AMSTAR 2 checklist offers a more nuanced assessment, emphasizing detailed reporting and tailored criteria. By enhancing the evaluation framework through such methodological improvements, future studies will achieve higher reliability and clinical utility of MRI in diagnosing knee OA. Consequently, these advancements will lead to more rigorous systematic reviews and meta-analyses, thereby strengthening the evidence base for the use of MRI as a diagnostic tool for knee OA. These studies should define the role of early and more advanced pathological changes in bone, cartilage of soft tissues detected by MRI in ruling out and ruling in knee OA, classifying OA, and choosing the appropriate therapy. This role will probably coexist with clinical and radiography criteria.

Acknowledgments

I would like to express my sincere gratitude to the Radiology department professors at the University of Granada.

Funding: None.

Footnote

Provenance and Peer Review: With the arrangement by the Guest Editors and the editorial office, this article has been reviewed by external peers.

Reporting Checklist: The authors have completed the PRISMA-ScR reporting checklist. Available at: https://qims.amegroups.com/article/view/10.21037/qims-24-1544/rc

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-24-1544/coif). The special issue “Advances in Diagnostic Musculoskeletal Imaging and Image-guided Therapy” was commissioned by the editorial office without any funding or sponsorship. F.R.S. served as the unpaid Guest Editor of the issue and serves as an unpaid editorial board member of Quantitative Imaging in Medicine and Surgery. The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. GBD 2021 Osteoarthritis Collaborators. Global, regional, and national burden of osteoarthritis, 1990-2020 and projections to 2050: a systematic analysis for the Global Burden of Disease Study 2021. Lancet Rheumatol 2023;5:e508-22. [Crossref] [PubMed]
2. Leifer VP, Katz JN, Losina E. The burden of OA-health services and economics. Osteoarthritis Cartilage 2022;30:10-6. [Crossref] [PubMed]
3. Hayashi D, Roemer FW, Guermazi A. Imaging for osteoarthritis. Ann Phys Rehabil Med 2016;59:161-9. [Crossref] [PubMed]
4. Jackson DW, Jennings LD, Maywood RM, Berger PE. Magnetic resonance imaging of the knee. Am J Sports Med 1988;16:29-38. [Crossref] [PubMed]
5. Zhang W, Doherty M, Peat G, Bierma-Zeinstra MA, Arden NK, Bresnihan B, Herrero-Beaumont G, Kirschner S, Leeb BF, Lohmander LS, Mazières B, Pavelka K, Punzi L, So AK, Tuncer T, Watt I, Bijlsma JW. EULAR evidence-based recommendations for the diagnosis of knee osteoarthritis. Ann Rheum Dis 2010;69:483-9. [Crossref] [PubMed]
6. KELLGREN JH. LAWRENCE JS. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957;16:494-502. [Crossref] [PubMed]
7. Schiphof D, Runhaar J, Waarsing EJ, van Spil EW, van Middelkoop M, Bierma-Zeinstra SM. The 10-year course of the clinical American college of rheumatology (acr) criteria for hip and knee osteoarthritis in an early symptomatic cohort, data from check. Osteoarthritis Cartilage 2018;26:S347-8.
8. England BR, Smith BJ, Baker NA, Barton JL, Oatis CA, Guyatt G, et al. 2022 American College of Rheumatology Guideline for Exercise, Rehabilitation, Diet, and Additional Integrative Interventions for Rheumatoid Arthritis. Arthritis Care Res (Hoboken) 2023;75:1603-15. [Crossref] [PubMed]
9. Oo WM, Linklater J. Prevalence of osteoarthritis-related imaging abnormalities in asymptomatic healthy adults. Osteoarthritis Cartilage 2024;32:1181-3. [Crossref] [PubMed]
10. Hunter DJ, Lo GH, Gale D, Grainger AJ, Guermazi A, Conaghan PG. The reliability of a new scoring system for knee osteoarthritis MRI and the validity of bone marrow lesion assessment: BLOKS (Boston Leeds Osteoarthritis Knee Score). Ann Rheum Dis 2008;67:206-11. [Crossref] [PubMed]
11. Roemer FW, Demehri S, Omoumi P, Link TM, Kijowski R, Saarakkala S, Crema MD, Guermazi A. State of the Art: Imaging of Osteoarthritis-Revisited 2020. Radiology 2020;296:5-21. [Crossref] [PubMed]
12. Ehmig J, Engel G, Lotz J, Lehmann W, Taheri S, Schilling AF, Seif Amir Hosseini A, Panahi B. MR-Imaging in Osteoarthritis: Current Standard of Practice and Future Outlook. Diagnostics (Basel) 2023;13:2586. [Crossref] [PubMed]
13. Culvenor AG, Øiestad BE, Hart HF, Stefanik JJ, Guermazi A, Crossley KM. Prevalence of knee osteoarthritis features on magnetic resonance imaging in asymptomatic uninjured adults: a systematic review and meta-analysis. Br J Sports Med 2019;53:1268-78. [Crossref] [PubMed]
14. Hunter DJ, Bierma-Zeinstra S. Osteoarthritis. Lancet 2019;393:1745-59. [Crossref] [PubMed]
15. Roemer FW, Guermazi A, Felson DT, Niu J, Nevitt MC, Crema MD, Lynch JA, Lewis CE, Torner J, Zhang Y. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study. Ann Rheum Dis 2011;70:1804-9. [Crossref] [PubMed]
16. Rogozińska E, Khan K. Grading evidence from test accuracy studies: what makes it challenging compared with the grading of effectiveness studies? Evid Based Med 2017;22:81-4. [Crossref] [PubMed]
17. Shea BJ, Reeves BC, Wells G, Thuku M, Hamel C, Moran J, Moher D, Tugwell P, Welch V, Kristjansson E, Henry DA. AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both. BMJ 2017;358:j4008.
18. Kirvalidze M, Abbadi A, Dahlberg L, Sacco LB, Calderón-Larrañaga A, Morin L. Estimating pairwise overlap in umbrella reviews: Considerations for using the corrected covered area (CCA) index methodology. Res Synth Methods 2023;14:764-7. [Crossref] [PubMed]
19. Menashe L, Hirko K, Losina E, Kloppenburg M, Zhang W, Li L, Hunter DJ. The diagnostic performance of MRI in osteoarthritis: a systematic review and meta-analysis. Osteoarthritis Cartilage 2012;20:13-21. [Crossref] [PubMed]
20. Smith TO, Drew BT, Toms AP, Donell ST, Hing CB. Accuracy of magnetic resonance imaging, magnetic resonance arthrography and computed tomography for the detection of chondral lesions of the knee. Knee Surg Sports Traumatol Arthrosc 2012;20:2367-79. [Crossref] [PubMed]
21. Quatman CE, Hettrich CM, Schmitt LC, Spindler KP. The clinical utility and diagnostic performance of magnetic resonance imaging for identification of early and advanced knee osteoarthritis: a systematic review. Am J Sports Med 2011;39:1557-68. [Crossref] [PubMed]
22. Gates M, Gates A, Pieper D, Fernandes RM, Tricco AC, Moher D, Brennan SE, Li T, Pollock M, Lunny C, Sepúlveda D, McKenzie JE, Scott SD, Robinson KA, Matthias K, Bougioukas KI, Fusar-Poli P, Whiting P, Moss SJ, Hartling L. Reporting guideline for overviews of reviews of healthcare interventions: development of the PRIOR statement. BMJ 2022;378:e070849. [Crossref] [PubMed]
23. Andriacchi TP, Favre J, Erhart-Hledik JC, Chu CR. A systems view of risk factors for knee osteoarthritis reveals insights into the pathogenesis of the disease. Ann Biomed Eng 2015;43:376-87. [Crossref] [PubMed]
24. Wang Q, Runhaar J, Kloppenburg M, Boers M, Bijlsma JWJ, Bierma-Zeinstra SMA. CREDO expert group. Evaluation of the Diagnostic Performance of American College of Rheumatology, EULAR, and National Institute for Health and Clinical Excellence Criteria Against Clinically Relevant Knee Osteoarthritis: Data From the CHECK Cohort. Arthritis Care Res (Hoboken) 2024;76:511-6. [Crossref] [PubMed]
25. Birch S, Lee MS, Robinson N, Alraek T. The U.K. NICE 2014 Guidelines for Osteoarthritis of the Knee: Lessons Learned in a Narrative Review Addressing Inadvertent Limitations and Bias. J Altern Complement Med 2017;23:242-6. [Crossref] [PubMed]
26. Hunter DJ, Zhang YQ, Tu X, Lavalley M, Niu JB, Amin S, Guermazi A, Genant H, Gale D, Felson DT. Change in joint space width: hyaline articular cartilage loss or alteration in meniscus? Arthritis Rheum 2006;54:2488-95. [Crossref] [PubMed]
27. Hunter DJ, Arden N, Conaghan PG, Eckstein F, Gold G, Grainger A, Guermazi A, Harvey W, Jones G, Hellio Le Graverand MP, Laredo JD, Lo G, Losina E, Mosher TJ, Roemer F, Zhang W. OARSI OA Imaging Working Group. Definition of osteoarthritis on MRI: results of a Delphi exercise. Osteoarthritis Cartilage 2011;19:963-9. [Crossref] [PubMed]
28. McGrath TA, Bossuyt PM, Cronin P, Salameh JP, Kraaijpoel N, Schieda N, McInnes MDF. Best practices for MRI systematic reviews and meta-analyses. J Magn Reson Imaging 2019;49:e51-e64. [Crossref] [PubMed]
29. Cheng Q, Zhao FC. Comparison of 1.5- and 3.0-T magnetic resonance imaging for evaluating lesions of the knee: A systematic review and meta-analysis (PRISMA-compliant article). Medicine (Baltimore) 2018;97:e12401. [Crossref] [PubMed]