A new study has highlighted the value of musculoskeletal ultrasound in diagnosing polymyalgia rheumatica, but Australian rheumatologists say its limitations warrant a rethink of the imaging pathway.
Musculoskeletal ultrasound can help rule out polymyalgia rheumatica, but positive findings are less reliable for distinguishing the condition from other inflammatory rheumatic diseases, new research suggests.
The findings, published in The Lancet Rheumatology, have prompted Australian experts in an accompanying editorial to call for a more targeted approach to imaging, potentially incorporating MRI and PET-CT when ultrasound findings are inconclusive.
The multicentre study involved 185 glucocorticoid-naive patients referred with suspected PMR across two independent cohorts in the Netherlands and Denmark.
“This study provides real-world evidence for the diagnostic performance of musculoskeletal ultrasound in patients with suspected polymyalgia rheumatica,” the researchers wrote.
“Musculoskeletal ultrasound could be particularly useful to exclude non-inflammatory conditions in patients with suspected polymyalgia rheumatica, but positive findings must be interpreted in the context of careful clinical assessment due to overlap with other inflammatory rheumatic diseases.”
The researchers assessed the diagnostic accuracy of protocolised musculoskeletal ultrasound, examining inflammatory lesions around the shoulders and hips and comparing the results with clinical diagnoses established after six or 12 months of follow-up.
Of the 92 patients in the Dutch cohort, 58 (63%) were subsequently diagnosed with PMR. In the Danish cohort, 66 of 93 patients (71%) received a PMR diagnosis.
The researchers found that ultrasound achieved its highest sensitivity when at least one inflammatory lesion was detected, reaching 93% in the Dutch cohort and 92% in the Danish cohort.
The presence of at least one inflammatory lesion in both the shoulder and hip girdles produced the highest specificity, at 85% and 93%, respectively.
However, no individual ultrasound finding was specific to PMR, and the technique performed substantially better at distinguishing PMR from non-inflammatory conditions than from alternative inflammatory rheumatic diseases.
The researchers said the absence of inflammatory lesions on ultrasound made PMR less likely, but positive findings needed to be interpreted alongside a careful clinical assessment.
“Absence of shoulder and hip lesions on musculoskeletal ultrasound makes a polymyalgia rheumatica diagnosis unlikely,” they wrote.
“However, their presence requires careful clinical evaluation, as similar findings occur in other inflammatory rheumatic diseases.”
The findings are particularly relevant given the diagnostic challenges associated with PMR, which lacks a disease-specific biomarker and shares clinical features with conditions including rheumatoid arthritis and osteoarthritis.
Biceps tenosynovitis affecting at least one shoulder was the most sensitive individual ultrasound finding in both cohorts, with sensitivities of 69% and 73%.
Bilateral subacromial-subdeltoid bursitis, bilateral biceps tenosynovitis, and bilateral hip synovitis demonstrated specificities exceeding 80% in both cohorts, although individual lesions generally produced only small changes in diagnostic probability.
Increasing the number of anatomical sites examined improved sensitivity but came at the expense of specificity, the researchers found.
In the Dutch cohort, expanding the assessment from six to 10 sites increased sensitivity from 86% to 93%, while specificity declined from 56% to 53%.
Writing in an accompanying commentary, Melbourne rheumatologist Dr Claire Owen and co-author Dr Bonnia Liu, both from Austin Health and the University of Melbourne, said the findings raised important questions about the role of ultrasound in the diagnostic pathway.
The authors argued that the study findings demonstrated the importance of examining the right anatomical locations rather than simply identifying more inflammatory lesions.
“Specificity in polymyalgia rheumatica diagnosis lies with the anatomical targets selected rather than the number of lesions assessed,” they wrote.
They highlighted evidence from MRI and FDG-PET–CT showing that some of the most distinctive inflammatory changes associated with PMR occurred in deep pelvic and axial structures that ultrasound cannot reliably assess.
These included hamstring peritendonitis and lumbar interspinous inflammation.
The study also identified considerable variation between clinicians interpreting ultrasound abnormalities, reinforcing the need for standardised scanning protocols, objective measurements, and appropriate training, Dr Owen and Dr Liu said.
“The limitations of musculoskeletal ultrasound do not diminish its value; rather, they highlight the importance of defining when and how it should be used within the diagnostic pathway for polymyalgia rheumatica,” they wrote.
They advocated a staged approach, with ultrasound serving as an accessible, low-cost initial investigation and MRI or FDG-PET–CT reserved for patients in whom diagnostic uncertainty persists.
“A negative scan is reassuring, particularly for excluding non-inflammatory mimics, but musculoskeletal ultrasound should not be regarded as the sole arbiter of polymyalgia rheumatica diagnosis,” they wrote.
Related
Dr Owen and Dr Liu concluded that the different imaging modalities should be considered complementary rather than competing diagnostic tools, with selection guided by clinical suspicion, local expertise, and availability.
The Australian authors also highlighted the challenges associated with standardising ultrasound interpretation.
In the new study, agreement between ultrasound assessors ranged from fair for glenohumeral synovitis to moderate for biceps tenosynovitis.
By comparison, the quantitative assessment of hip capsular thickness demonstrated excellent reliability, with an intraclass correlation coefficient of 0.93.
Dr Owen and Dr Liu said developing standardised scanning protocols, training frameworks, and objective definitions of ultrasound abnormalities would be essential to improving reproducibility and supporting wider clinical adoption.
Despite its limitations, they emphasised that ultrasound remained an attractive initial imaging option because of its low cost, portability, lack of ionising radiation, and potential for point-of-care assessment by appropriately trained clinicians.
However, they cautioned against relying on ultrasound as the sole determinant of a PMR diagnosis, particularly when clinical suspicion remained high despite negative or equivocal findings.
The study’s researchers acknowledged that the relatively small study cohorts, differences in scanning protocols, and use of clinical diagnoses incorporating imaging findings as the reference standard were important considerations when interpreting the results.
Nevertheless, the findings provided evidence supporting ultrasound as an adjunct to clinical assessment rather than a definitive diagnostic investigation.
The researchers said positive musculoskeletal ultrasound findings should always be interpreted in the context of a full clinical assessment.
“As polymyalgia rheumatica remains a clinical diagnosis without a definitive test, future diagnostic models that integrate clinical features with musculoskeletal ultrasound findings are needed to improve accuracy and support treatment decisions,” the researchers concluded.
The Lancet Rheumatology, October 2026 (research)
The Lancet Rheumatology, October 2026 (editorial)
