Introduction
Materials and method
Search strategy
PUBMED |
("Osteomyelitis"[Mesh] OR "Osteitis"[Mesh] OR ("Surgical Wound Infection"[Mesh] AND bone*[tiab]) OR osteomyelitis[tiab] OR osteitis[tiab]) AND ("Diagnostic Imaging"[Mesh] OR "Magnetic Resonance Imaging"[Mesh] OR "Tomography, X-Ray"[Mesh] OR "Tomography, Emission-Computed"[Mesh] OR "Radionuclide Imaging"[Mesh] OR "Positron-Emission Tomography"[Mesh] OR "Fluorodeoxyglucose F18"[Mesh] OR "Leukocytes/radionuclide imaging"[Mesh] OR "Technetium Tc 99 m Exametazime"[Mesh] OR diagnostic imaging[tiab] OR MRI[tiab] OR "bone scan"[tiab] OR "CT scan"[tiab] OR "computed tomography"[tiab] OR SPECT-CT[tiab] OR SPECT/CT[tiab] OR PET[tiab] OR PET/CT[tiab] OR PET-CT[tiab] OR FDG[tiab] OR fluorodeoxyglucose[tiab] OR scintigraphy[tiab]) NOT Case Reports[ptyp] AND PY: from 2000, added to Pubmed until dec2015 |
EMBASE |
‘osteomyelitis’/mj OR ‘osteitis’/mj OR (‘surgical infection’/exp/mj AND bone*:ab,ti) OR osteomyelitis:ab,ti OR osteitis:ab,ti AND (‘diagnostic imaging’/exp OR ‘nuclear magnetic resonance imaging’/exp OR ‘tomography’/de OR ‘computer assisted tomography’/exp OR ‘emission tomography’/exp OR ‘whole body tomography’/exp OR ‘scintiscanning’/exp OR ‘fluorodeoxyglucose f 18’/exp OR (‘leukocyte’/exp/mj AND imaging) OR (‘technetium 99 m’/exp/mj AND imaging) OR ‘diagnostic imaging’:ab,ti OR mri:ab,ti OR ‘bone scan’:ab,ti OR ‘ct scan’:ab,ti OR ‘computed tomography’:ab,ti OR ‘spect-ct’:ab,ti OR ‘spect/ct’:ab,ti OR pet:ab,ti OR ‘pet/ct’:ab,ti OR ‘pet-ct’:ab,ti OR fdg:ab,ti OR fluorodeoxyglucose:ab,ti OR scintigraphy:ab,ti) NOT ‘case report’/exp AND [2000-2016]/py AND [1-1-1900]/sd NOT [31-12-2015]/sd |
Study selection
Inclusion Criteria |
1) The study must evaluate the accuracy of radiological and nuclear imaging modalities for diagnosing PTO. |
2) The study group must be at least 10 patients of 18 years and older with (suspected) PTO. In case of a mixed population, the data for this subgroup must be available independently. |
3) The studied location must be in the peripheral skeleton. |
4) The study must use a valid reference test (osteomyelitis was proven histologically and/or bacteriologically, and/or there was a clinical follow-up of at least 6 months in which no signs or symptoms of chronic infection were described). |
5) Studies must provide sufficient details to construct a 2 x 2 contingency table expressing the results of the index tests by the disease status. |
6) The study must investigate a commonly used diagnostic imaging test for PTO. These are conventional X-ray, CT, MRI, WBC scintigraphy/AGA scintigraphy (+/- SPECT/CT), bone scintigraphy (+/- SPECT/CT) and FDG-PET (+/- CT). |
Exclusion Criteria |
1) Non-human studies. |
2) Studies that investigate non-trauma-related osteomyelitis (such as osteomyelitis due to spondylodiscitis, diabetic feet, haematogenous dissemination and pressure ulcers). |
3) Studies that investigate a not commonly used diagnostic imaging test [such as 99mTc-ciprofloxacin (Infecton) scintigraphy or 68Ga-citrate PET]. |
Methodological quality assessment
Data extraction
Statistical analysis
Source of funding
Results
Included studies

Study quality
Description of study characteristics
Imaging modality | Author | Year | PTO Patients (n) | Specifics on imaging technique/ tracer | Use of hybrid imaging | Methodology | Timeframe between trauma, first symptoms of infection and diagnostic imaging |
---|---|---|---|---|---|---|---|
Three-phase bone scintigraphy | Ballani et al. [45] | 2007 | 10 | 3-phase 99mTc-MDP bone scan, 740 MBq, γ-camera with 256 x 256 matrix (pixel size ∼ 2 mm). | No | Retrospective. Gold standard: microbiology (n = 5); otherwise, overall clinical assessment, FU unknown. | Unknown. |
Kaim et al. [49] | 2000 | 18/19* | 3-phase 99mTc-DPD bone scan, 740 MBq, γ-camera with 256 x 256 matrix. | No | Retrospective, highly selective patient group without orthopaedic implants or patients whose devices had been removed. Gold standard: microbiology (n = 13); otherwise, overall clinical assessment with minimum of 16 months FU. | Long-standing PTO, time interval between last surgical intervention and present study was 6.5 years (3 months – 39 years). Time interval between last surgery and imaging not clear. | |
Meller et al. [52] | 2002 | 19/21* | 3-phase 99mTc-MDP bone scan, 740 MBq. 150,000 – 400,000 counts for each projection. | No | Prospective. Gold standard: MRI (n = 19); histology/microbiology (n = 12); FU 1–6 months. | Symptoms of infection lasting for more than 6 weeks. | |
WBC (or AGA) scintigraphy | Ballani et al. [45] | 2007 | 10 | WBC scintigraphy with 99mTc-HMPAO labelled autologous WBCs, 740 MBq, γ-camera with a 256 x 256 matrix (pixel size ∼ 2 mm). No late (24 h) phase scan. | No | Retrospective. Gold standard: microbiology (n = 5); otherwise, overall clinical assessment. | Unknown. |
Glaudemans et al. [47] | 2013 | 49 | WBC scintigraphy with 99mTc-HMPAO labelled autologous WBCs, 500 MBq. Late phase scan included. | Yes | Retrospective. Gold standard: microbiology (n = 13), otherwise, overall clinical assessment at 6 months follow-up. | Unknown. | |
Horger et al. [21] | 2003 | 27/29* | AGA scintigraphy with 99mTc labelled murine monoclonal antibodies. 750 MBq. Late phase scan included. γ-camera with 128 x 128 matrix. | Yes | Prospective. Gold standard: microbiology (n = 18); otherwise, overall clinical assessment with minimum of 6 months FU. 25 patients with suspected PTO (of which 1 non peripheral) and 2 (suspected) PJI. | Reactivation of chronic PTO suspected because of clinical inflammatory symptoms or elevated laboratory markers. Timeframe not specified. | |
Kaim et al. [49] | 2000 | 18/19* | AGA scintigraphy with 99mTc labelled murine IgG antibodies, 555 MBq, γ-camera with a matrix of 256 x 256. Only one imaging time point (17 h). | No | Retrospective, highly selective patient group without orthopaedic implants or patients whose devices had been removed. Gold standard: microbiology (n = 13); otherwise, overall clinical assessment with minimum of 16 months FU. | Longstanding PTO, time interval between last surgical intervention and present study was 6.5 years (3 months – 39 years). Time interval between last surgery and imaging not clear. | |
Meller et al. [52] | 2002 | 19/21* | WBC scintigraphy autologous labelled with 111In, 18–37 MBq. Late phase scan included. 128 x 128 matrix, 250,000 – 500,000 counts for each projection. | No | Prospective. Gold standard: MRI (n = 19); histology/microbiology (n = 12); FU 1–6 months. | Symptoms of infection lasting for more than 6 weeks. | |
FDG-PET | Goebel et al. [46] | 2007 | 50 |
18F-FDG-PET, 200 MBq. 128 x 128 matrix. | No | Prospective. Gold standard: microbiology (n = 50). 2 suspected trauma-related PJI included. | Symptoms of infection lasting for more than 6 weeks. |
Hartmann et al. [48] | 2006 | 23 |
18F-FDG-PET, 300–400 MBq. | Yes | Retrospective, 15 patients with osteosynthesis, 3 prosthesis, 5 no material in situ. All trauma-related. Gold standard: microbiology (n = 23). | Symptoms of infection lasting for more than 6 weeks or presence of recurrent osteomyelitis. | |
Meller et al. [52] | 2002 | 19/21* |
18–F-FDG-PET, 296 MBq. | No | Prospective. Gold standard: MRI (n = 19); histology/microbiology (n = 12); FU 1–6 months. | Symptoms of infection lasting for more than 6 weeks. | |
Schiesser et al. [53] | 2003 | 17/20* |
18F-FDG-PET, 300–400 MBq. | No | Prospective. Gold standard: microbiology (n = 20), clinical FU 6 months. | Pain at motion or rest for at least 6 weeks, interval between last surgical intervention and FDG-PET scan 6 weeks – 14 months. | |
Shemesh et al. [54] | 2015 | 10 |
18F -FDG-PET, 296–555 MBq. | Yes | Retrospective. Gold standard: microbiology (n = 9, five cultures minimum); otherwise, overall clinical assessment with minimum of 1 year FU. | Time from initial surgery to PET/CT 2 months – 20 years. | |
Wenter et al. [55] | 2016 | 84 |
18F-FDG-PET, weight adapted dose (mean 252 +/- 76 MBq). 131 patients underwent PET/CT with mostly a full dose CT (n = 130) with iv contrast (n = 106). | No | Retrospective. Gold standard: microbiology (n = 143); otherwise, overall clinical assessment with minimum of 1 year FU. 215 patients, 192 suspected PTO (of which 12 non peripheral), 11 (suspected) PJI. | Causative event prior to PET scan dated back 12 ± 13 year in the clinically infected group and 10 ± 12 years in the clinically uninfected group. | |
131 | Yes | ||||||
MRI | Goebel et al. [46] | 2007 | 18 | No details on MRI technique provided | n/a | Prospective. Gold standard: microbiology (n = 18). | Symptoms of infection lasting for more than 6 weeks. |
Kaim et al. [49] | 2000 | 18/19* | 1.5-T MRI, slice thickness 3–10 mm. Both body coil (n = 10) and extremity coil (n = 8) used. All scans with iv gadolinium contrast. | n/a | Retrospective, highly selective patient group without orthopaedic implants or patients whose devices had been removed. Gold standard: microbiology (n = 13); otherwise, overall clinical assessment with minimum of 16 months FU. | Longstanding PTO, time interval between last surgical intervention and present study was 6,5 years (3 months – 39 years). Time interval between last surgery and imaging not clear. | |
CT scan | Goebel et al. [46] | 2007 | 22 | No details on CT technique provided. | n/a | Prospective. Gold standard: microbiology (n = 22). | Symptoms of infection lasting for more than 6 weeks. |
