This poster is published under an open license
. Please read the disclaimer
for further details.
Musculoskeletal bone, Musculoskeletal joint, Cone beam CT, MR, Comparative studies, Cost-effectiveness, Arthritides
E. Trajcevska, C. Owen, R. Buchanan, R. B. Zwar; Melbourne/AU
Methods and materials
A single centre retrospective study was performed.
Twenty two patients with a diagnosis of RA were imaged with MRI and CBCT and assessed for evidence of erosions of the carpal bones,
the metacarpal bases,
the distal radius and the distal ulna.
No more than three months between the two imaging modalities was allowed.
The first metacarpal base of one patient was excluded as it was unfortunately not included in the in the CBCT field of view.
Validated-Outcome Measures in Rheumatology Clinical Trials/Rheumatoid Arthritis Magnetic Resonance Imaging Score (OMERACT-RAMRIS) criteria and Sharp/van der Heijde criteria were used as a basis for erosion assessment. Equipment: NewTom Cone Beam CT 5G,
Italy; Siemens Avanto 1.5T MRI,
The number of erosions was scored per bone as 1,
In the event that there were 5 erosions or more,
the number taken was 5.
Bone erosion was scored 1- 10,
according to the proportion (in increments of 10%) of bone involved: 0: 0%,
Two independent readers assessed each MRI and CBCT on standard Hi Resolution radiology PACS workstations (AGFA Health Care NV.
All images were de-identified.
Images were assessed blinded and MRI and CBCT assessed independently.
Scores for the number of erosions and percentage bone involvement were given – 15 sites per patient: carpal bones (8),
base of metacarpals (5),
distal radius and distal ulna.
Diagnostic confidence scores (as a score of 1-5) were recorded per erosion,
were felt non-contributory for the purpose of this study.
Marrow oedema was assessed on MRI and qualitative correlation with trabecular rarefaction seen on CBCT was then performed in consensus.
Mean number and standard deviation of erosions and proportion of bone involvement was calculated per reader per bone across all patients.
CBCT versus MRI assessment results for each reader was compared using the Wilcoxon Signed Rank Test using MedCalc,
A p-value of <0.05 was classified as statistically significant.