Background Use of ultrasound (US) in rheumatology has increased over the last decade. Until recently, the only US modality used was 2D, which has been shown to be more sensitive for detection of erosions in rheumatoid arthritis (RA) than X-rays.
Within the last years, high frequency volumetric 3D US probes have become available as application for many high-end US machines. With 3D it is possible in the post-processing to inspect two perpendicular planes of the same point. This makes it reasonable to believe 3D will increase the usefulness of US in detection of erosions in patients with RA.
The reliability and validity of 3D for detection of erosions has only been investigated in few pilot studies.
Objectives The aim of this study was to test the reliability, validity and sensitivity and specificity of a 3D US erosion score in RA. MRI was used as gold standard.
Methods Patients with RA were included in the study. A 3D US examination and a 3T MRI scanning over the 2nd and 3rd metacarpophalangeal (MCP) joint were performed on the same day in all patients. The joints were examined by US from both the dorsal and volar aspect.
The 3D blocks were evaluated for erosions by two investigators. One of the investigators evaluated all examinations twice. The erosions were estimated according to both a semi-quantitative score (SQS) with the size of the erosion scored from 0-3 and measured quantitatively (QS) in mm2. The area was estimated by multiplying the transverse and longitudinal measurement of the erosion.
Furthermore, a specialist evaluated the MRI scans by the RAMRIS.
For estimation of the reliability the Intra-class correlation coefficients (ICC) and absolute agreement were used. In order to test validity, the correlation between 3D US and MRI was calculated and the sensitivity and specificity were estimated.
Results Twenty eight patients with RA were included in the study. The ICC for the inter-observer reliability was 0.41 and 0.13 for the metacarpal and phalangeal bone respectively when assessed with QS and 0.86 and 0.16 when assessed according to SQS. The ICC for intra-observer reliability was 0.75 and 0.48 for the metacarpal and phalangeal bone respectively in the QS evaluation and 0.83 and 0.60 in SQS.
The correlation with MRI for the metacarpal was significant for both scores with values of 0.73 (SQ) and 0.74 (SQS). For the phalangeal bone, no significant correlation was found with values of 0.28 (SQ) and 0.26 (SQS). The sensitivity and specificity for the detection of erosions for the metacarpal bone was 86% and 85% respectively. For the phalangeal bone it was 60% and 97%.
Conclusions Both good inter- and intra-observer reliability and correlation with MRI were seen in assessment of erosions with 3D US in the metacarpal bone, while results were low and without significance for the phalangeal bone. The results indicate that 3D US still has room for improvement.
Disclosure of Interest None Declared