Background/Purpose: While ultrasound assessment of patients with RA using the 7-joint score (US7) demonstrates reliability in the hands of experienced readers it is unclear to what degree this has external validity, and whether satisfactory calibration can be attained over the course of a single calibration exercise conducted over one day at a single location. We aimed to assess the feasibility and impact on inter-reader reliability of a one-day structured program of training based on an eCRF designed to provide real-time feedback on reader reliability during the process of scanning.
Methods: Six patients with rheumatoid arthritis were examined by 12 sonologists from 6 countries and 12 centers in 6 rater pairs who performed the US7 score (1). The US7 score includes the clinically dominant wrist, the second and third metacarpophalangeal (MCP) and proximal interphalangeal joints, and the second and fifth metatarsophalangeal (MTP) joints, which were evaluated for synovitis (SYN), tenosynovitis/paratenonitis (TS), and erosions (ER) from the dorsal side and palmar/plantar aspects by gray-scale (GS) and power Doppler (PD) ultrasound. Additional lateral scans were performed at the MCP2 and MTP5 joints. Scores were entered into an eCRF custom designed to provide immediate calculation of reliability data (eCaRe-US reliability). Training of readers focused on the most discrepant features observed in exercise 1. All reader pairs repeated the examination in different patient order. Mean (SD) weighted kappa values, mean (SD) per cent agreement rates, inter-observer intra-class correlation (ICC) for summed scores (SYN, TS, ER) were calculated.
Results: Improvement in reliability was mainly observed in assessment of SYN-GS and PD. Primary regions of improvement in SYN-GS were dorso-median and ulnar wrist (k 0.31 to 0.54 and 0.17 to 0.34), PIP2 and 3 dorsal (k 0.03 to 1 and 0.13 to 1), MCP2 and 3 palmar (k 0.04 to 0.37 and 0.06 to 1.0), and MTP5 dorsal (k 0.06 to 1.0). Improvement in detection of SYN-PD was mainly observed in the wrist. Regions where reliability was not improved were MCP2 dorsal, PIP2 and 3 palmar, and MTP2 dorsal. Erosion assessment was lengthy and considered challenging for routine practice.
|
|
Ex |
SYN-GS |
SYN-PD |
TS-GS |
TS-PD |
ER |
|
% agree |
1 |
39.8 (25.9) |
75.6 (25.9) |
60.0(19.0) |
80.0(13.9) |
67.8(24.0) |
|
|
2 |
48.7 (41.6) |
73.1 (26.0) |
53.3(32.1) |
80.0(18.3) |
nd |
|
kappa |
1 |
0.12 (0.10) |
0.53 (0.47) |
0.40 (0.22) |
0.40 (0.42) |
0.29 (0.35) |
|
|
2 |
0.47 (0.40) |
0.45 (0.43) |
0.31(0.31) |
0.30 (0.41) |
nd |
|
ICC |
1 |
0.63 |
0.53 |
0.60 |
0.45 |
0.55 |
|
|
2 |
0.84 |
0.77 |
0.56 |
0.37 |
nd |
Conclusion: Substantial enhancement of reliability for detection of synovitis by ultrasound may be observed with limited calibration of inexperienced readers. We have identified specific regions that require more intensive calibration, specifically, MCP2 dorsal, PIP2 and 3 palmar, and MTP2 dorsal.
1. Backhaus M, et al. Arthritis Rheum 2009;61:1194
Disclosure:
M. Backhaus,
None;
S. G. Werner,
None;
S. Ohrndorf,
None;
S. Alraqi,
None;
S. Crowther,
None;
S. Dhillon,
None;
N. Dhindsa,
None;
A. J. Fernandes,
None;
O. FitzGerald,
None;
H. B. Hammer,
None;
C. Hudry,
None;
S. Jousse-Joulin,
None;
R. G. Lambert,
None;
M. Larche,
None;
A. Remy-Moulard,
None;
L. Terslev,
None;
R. Yazdani,
None;
R. Dadashova,
None;
J. Paschke,
None;
W. P. Maksymowych,
None.
« Back to 2013 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/can-satisfactory-reliability-of-the-7-joint-ultrasound-score-be-attained-by-inexperienced-readers-in-a-single-calibration-exercise-results-from-the-biodam-program/