Imaging Musculoskeletal Infection

MRI Assessment of Osteomyelitis

MRI with gadolinium contrast is the imaging modality of choice for osteomyelitis — but quote its accuracy with the setting attached, because MRI is not one test with one number.

• Acute osteomyelitis and the diabetic foot: MRI performs well. Pooled across 16 studies of foot osteomyelitis the diagnostic odds ratio was 42.1, with specificity 82.5% at a 90% sensitivity cut point — enough to rule the diagnosis both in and out, and far ahead of bone scanning, radiography or white-cell studies.
• Chronic osteomyelitis: MRI degrades badly. Pooled across 23 studies its sensitivity was 84% but its specificity only 60% (95% CI 38-78%). Previous surgery, healed infection, metalwork and remodelled marrow all produce the same signal changes, so a positive MRI in a chronically infected limb means much less than the same scan in a previously untouched bone.

That collapse in specificity is the whole reason FDG-PET exists in this algorithm — in the same meta-analysis it achieved 96% sensitivity and 91% specificity for chronic osteomyelitis, the best of any single modality. Evidence level 1 (meta-analyses of diagnostic accuracy studies).

The key MRI findings (PACES mnemonic):

Bone marrow oedema: The most sensitive but least specific finding. On T1: the normal bright fatty marrow is replaced by dark signal. On STIR/T2 fat-sat: marrow shows bright signal indicating oedema. This pattern is also seen in tumour, stress fracture, and contusion — clinical correlation is essential.

Abscess formation: On contrast-enhanced T1 fat-sat images, an abscess appears as a collection with ring enhancement (wall enhances, centre does not). This distinguishes it from phlegmon, which shows diffuse, homogeneous enhancement. Abscess identification is critical for surgical planning — abscesses typically require drainage, while phlegmon may respond to antibiotics alone.

Penumbra sign: A thin rim of T1 hyperintensity (1-2mm) immediately surrounding an intraosseous abscess (Brodie abscess). This finding is SPECIFIC for infection and helps distinguish a Brodie abscess from a bone tumour. The hyperintense rim represents granulation tissue at the interface between abscess and viable bone.

Cortical erosion and sinus tract: Cortical destruction is best assessed on contrast-enhanced images. A sinus tract appears as a linear enhancing track extending from the infected bone through soft tissues to the skin surface — this is specific for chronic osteomyelitis.

Paediatric considerations: In children under 18 months, the growth plate has transphyseal blood vessels that allow infection to cross from the metaphysis to the epiphysis, potentially causing concurrent osteomyelitis and septic arthritis. After 18 months, the growth plate acts as a barrier to infection spread.

Imaging of Prosthetic Joint Infection

PJI imaging is challenging because metalwork creates significant artefact on MRI and the distinction between aseptic loosening and infection is often unclear on radiographs. The imaging approach:

Radiographs: May show progressive osteolysis, periosteal reaction, endosteal scalloping, or hardware loosening — but these signs are NON-SPECIFIC and may be identical in aseptic loosening. Rapid progressive osteolysis within the first 2 years raises suspicion for infection.

Nuclear medicine: (1) Triple-phase bone scan: sensitive but NOT specific — positive for ANY cause of increased bone turnover (including normal post-operative remodelling for up to 12-18 months). Its pooled specificity in chronic osteomyelitis is 25%, which is why it is a screening test at best. (2) Combined labelled WBC scan + sulphur colloid scan: the most accurate nuclear medicine combination for PJI. WBC accumulation at the prosthesis WITHOUT corresponding sulphur colloid uptake (normally taken up by displaced marrow) indicates infection rather than normal marrow distribution. In 59 painful failed hip and knee replacements this achieved sensitivity 100%, specificity 91% and accuracy 95% — figures from a single small series, so treat the 100% as "very high" rather than absolute. (3) FDG-PET.

Do not weld the FDG-PET figures from two different settings into one claim, which is an easy way to lose a viva. The often-quoted 96% sensitivity with 91% specificity belongs to CHRONIC OSTEOMYELITIS, where FDG-PET is the single most accurate modality. Around a prosthesis it is a different and much weaker test: in the head-to-head comparison against combined leucocyte/marrow imaging, FDG-PET's best accuracy by any interpretive criterion was 71%, against 95% for the leucocyte/marrow study. So FDG-PET is the right answer for chronic osteomyelitis and for the axial skeleton, and the wrong answer for confirming or excluding a prosthetic joint infection.

MRI with metal artefact reduction: Newer sequences (MARS, SEMAC, VAT) can partially overcome metalwork artefact. MRI may show periosteal oedema, fluid collections, and synovitis, but interpretation requires expertise.

Joint aspiration: REMAINS THE GOLD STANDARD for PJI diagnosis — imaging supports but does not replace microbiological assessment. Synovial fluid analysis: WBC count, differential, culture, and increasingly alpha-defensin and leukocyte esterase strip testing.

Imaging Features of Septic Arthritis

Septic arthritis is a clinical and microbiological diagnosis confirmed by joint aspiration — imaging supports it and excludes adjacent osteomyelitis but never replaces the aspirate. The role of imaging is to detect the effusion, guide aspiration, and assess for bone involvement.

MRI evidence supports its role as the gold standard for osteomyelitis diagnosis.

Evidence guides the complex imaging approach to musculoskeletal infection.

Society guidelines anchor imaging selection to microbiological confirmation.