Article Text

SP0093 Role of imaging in osteoarthritis: Biomarker, prognostic factor, outcome measure in RCT?
  1. P.G. Conaghan
  1. Division of Musculoskeletal Disease, Nihr Musculoskeletal Biomedical Research Unit, University of Leeds, Leeds, United Kingdom


A biomarker of structural integrity seems essential in understanding the natural and post-therapy progression of osteoarthritis (OA), perhaps even more so at a time when a new wave of pharmacological agents promises structure modification. So how well do our current imaging modalities fare as useful biomarkers?

Conventional radiography has provided an essential “common language” for defining OA in epidemiological cohorts, especially using categorical classification systems such as Kellgren-Lawrence. However for clinical trial use, assessment of joint space width (JSW) remains the standard outcome for regulatory approval studies of large joints. Although there are a range of acquisition techniques and measures of JSW, it has some predictive validity for joint replacement and good reliability. In large trials of knee OA of over 2 years duration and with flexed-knee acquisition, a SRM of 0.71 has been reported1. The responsiveness of JSW in hip OA is also very good, although based on fewer studies, with cohort studies showing better responsiveness than clinical trials2.

Magnetic resonance imaging (MRI) of course provides the ability to visualise all the structures involved in the whole-organ OA process and large cohort studies have demonstrated frequent abnormalities of cartilage, menisci, bone (bone marrow lesions –BML- osteophytes and attrition). In terms of concurrent validity, BML, synovitis and effusion have moderate to strong relationship with pain, with strong construct validity for MR cartilage compared with histology and arthroscopy3. A number of MR pathologies including quantitative cartilage volume, cartilage defects and BML are predictors of later joint replacement, a meaningful measure for patients with OA. Quantitative cartilage volume has been the most studied MR outcome, predominantly in knee OA studies, with a SRM of 0.86 for the medial tibiofemoral joint4. The semi-quantitative scores used for cartilage morphology have demonstrated good responsiveness and data is accruing on the responsiveness of scoring other pathological features. Automated methods of measuring bone area will improve the assessment of non-cartilage OA joint progression, and semi-quantitative scoring systems of multiple pathologies for both hip and hand OA have been reported recently, further advancing the range of MR biomarkers.

While ultrasonography (US) provides an excellent and feasible tool to assist with OA diagnosis in the clinic, there is as yet only a small amount of data on how the different US-detected pathologies will perform as biomarkers in clinical trials. An US hand OA score has been developed with good reliability.

Overall, both conventional radiography and MRI can currently be recommended for use as biomarkers in clinical trials of OA, but the performance metrics vary with respect to the individual pathologies measured and the specific modality employed, and with respect to the anatomical site of interest. This all bodes well for evaluation of putative structure-modifiers.

  1. Reichmann et al. Osteoarthritis Cartilage 2011, 550-6.

  2. Chu Miow Lin et al. Osteoarthritis Cartilage 2011, 543-9.

  3. Hunter et al. Osteoarthritis Cartilage 2011, 557-88.

  4. Hunter et al. Osteoarthritis Cartilage 2011, 589-605.

Disclosure of Interest None Declared

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.