Development of an Ultrasound Joint Inflammation Score for Rheumatoid Arthritis through a Data-Driven Approach

Anna-Birgitte Aga¹, Elisabeth Lie², Inge C Olsen², Hilde Berner Hammer³, Till Uhlig⁴, Désirée van der Heijde^5,6, Tore K. Kvien², Espen A. Haavardsholm² and the ARCTIC study group, ¹Department of Rheumatology, Diakonhjemmet Hospital, Oslo, Norway, ²Dept. of Rheumatology, Diakonhjemmet Hospital, Oslo, Norway, ³Postboks 23 Vinderen, Diakonhjemmet Hospital, Oslo, Norway, ⁴Diakonhjemmet Hospital, Oslo, Norway, ⁵Department of Rheumatology, Leiden University Medical Center, Leiden, the Netherlands, Leiden, Netherlands, ⁶Rheumatology, Diakonhjemmet Hospital, Oslo, Norway

Meeting: 2015 ACR/ARHP Annual Meeting

Date of first publication: September 29, 2015

Keywords: Disease Activity, rheumatoid arthritis (RA) and ultrasonography

Session Information

Date: Sunday, November 8, 2015

Title: Imaging of Rheumatic Diseases I: Ultrasound

Session Type: ACR Concurrent Abstract Session

Session Time: 2:30PM-4:00PM

Background/Purpose:
No consensus exists regarding which joints and
tendons should be systematically assessed by ultrasonography (US) to assess
inflammation in rheumatoid arthritis (RA). Validity and responsiveness must be
weighted against feasibility. Our group has previously developed and validated a
US joint inflammation score (Aga et al, Ann Rheum Dis 2015). The objective of
this study was to optimize the score in terms of sensitivity to change and
feasibility, and to perform comparative analyses versus formerly proposed US
scores in a longitudinal study.

Methods: DMARD-naïve
early RA patients (pts) with <2 yrs symptom duration fulfilling the 2010
ACR/EULAR classification criteria were recruited between Sept 2010 and April
2013 and started on methotrexate. Extensive US examination were performed by
experienced sonographers at baseline (BL) and after 3 and 6 months, using a
validated grey-scale (GSUS) and power Doppler (PDUS) semi-quantitative scoring
system with scores 0-3 for GSUS and PDUS in each of the following 36 joints and
4 tendons: MCP1-5, PIP2-3, radiocarpal, distal radioulnar, intercarpal, elbow,
knee, talocrural, MTP1-5, extensor carpi ulnaris and tibialis posterior (TP)
tendons, bilaterally. An US atlas was used as reference (Hammer et al, Ann
Rheum Dis 2011). We performed principal component factor analyses (PCA) of GSUS
and PDUS changes from BL to 3 months (ΔGSUS and ΔPDUS) to identify
joint groups with high internal correlation. Based on these analyses we identified
several joints/tendons candidate sets to be further tested. Standardised
response means (SRMs) with 95% CI (bootstrapping) were estimated at 3 and 6
months for the candidate sets and formerly proposed scores.

Results: 117 early RA pts were included; 81.2% anti-CCP pos, mean(SD) age 50.3
(13.3) yrs, median(IQR) disease duration (from onset of swollen joint) 5 (3-9)
months, mean(95% CI) 36-joint GSUS score 23(21-25) and PDUS score 11(10-12). Based
on the PCA 12 groups of joint/tendons were identified (Table 1). Elbow,
TP-tendon and MTP1 came out as separate factors, but were omitted from the
scores due to low SRMs. Table 2 shows the 3- and 6-month SRMs for the previously
published candidate sets (A, B), one of the new candidate sets (C) and the
formerly proposed scores.

Conclusion: We used a data-driven approach to further develop an ultrasound
joint inflammation score in RA. We propose a novel candidate set of
joints/tendons with improved sensitivity to change, which is feasible, and performed
numerically better than formerly proposed scores. Our results show that a reduced
US assessment may efficiently contribute to disease assessment in RA.

Table 1: Joint/tendon groups identified by principal component factor analyses of ΔGSUS and ΔPDUS scores after 3 months with methotrexate treatment in early RA
Group 1	Group 2	Group 3	Group 4	Group 5	Group 6	Group 7	Group 8	Group 9	Group 10	Group 11	Group 12
Radiocarpal* Intercarpal DRUJ	MCP3* MCP4 MCP5	MTP2*	Elbow	MCP2*	MCP1*	MTP3* MTP4 MTP5	TP-t	ECU-t*	PIP2*	PIP3*	MTP1
DRUJ = distal radioulnar joint; MTP = metatarsophalangeal; MCP = metacarpophalangeal; PIP = proximal interphalangeal joints; TP-t = tibialis posterior tendon; ECU-t = extensor carpi ulnaris tendon. *Joints/tendon included in the new candidate set C

Table 2: 3- and 6- month standardised response means (SRMs) for candidate set A, B, C and formerly proposed ultrasound joint scores for RA

GSUS	*SRM(95%CI) 3 months**	*SRM(95%CI) 6 months**
12-joint score Naredo¹	– 1.01 ( – 1.17 to – 0.84)	– 0.92 ( – 1.08 to – 0.76)
7-joint score Backhaus²	– 1.10 ( – 1.30 to – 0.90)	– 1.15 ( – 1.36 to – 0.94)
6-joint score Perricone³	– 0.99 ( – 1.19 to – 0.78)	– 0.97 ( – 1.15 to – 0.79)
Candidate set A score⁴	– 1.04 ( – 1.23 to – 0.84)	– 1.06 ( – 1.29 to – 0.84)
Candidate set B score⁵	– 1.06 ( – 1.27 to – 0.87)	– 1.14 ( – 1.35 to – 0.94)
Candidate set C score⁶	– 1.25 ( – 1.48 to – 1.02)	– 1.25 ( – 1.47 to – 1.02)

PDUS	*SRM(95%CI) 3 months**	*SRM(95%CI) 6 months**
12-joint score Naredo¹	– 0.74 ( – 0.89 to – 0.58)	– 0.74 ( – 0.89 to – 0.60)
7-joint score Backhaus²	– 0.90 ( -1.07 to – 0.74)	– 1.01 ( – 1.16 to – 0.87)
6-joint score Perricone³	– 0.70 ( – 0.87 to – 0.52)	– 0.78 ( – 0.93 to – 0.62)
Candidate set A score⁴	– 0.90 ( – 1.12 to – 0.68)	– 1.06 ( – 1.27 to – 0.87)
Candidate set B score⁵	– 0.90 ( – 1.11 to – 0.69)	– 1.08 ( – 1.30 to – 0.88)
Candidate set C score⁶	– 0.97 ( – 1.18 to – 0.77)	– 1.08 ( – 1.26 to – 0.92)
SRM (standardised response mean) = mean change divided by the standard deviation of the change. CI = confidence interval; GSUS = grey-scale ultrasound; PDUS = power Doppler ultrasound. Patients in the early RA cohort with data on 3 months (n=113) and 6 months (n=109). ¹12-joint score Naredo et al:* MCP2, MCP3, wrist, elbow, knee, ankle and TP tendon bilaterally (Naredo et al, Arthritis Rheum 2008). ²7-joint score Backhaus: MCP2, MCP3, PIP2, PIP3, wrist, MTP2, MTP5 and ECU tendon dominant side (Backhaus et al, Arthritis Rheum 2009). ³6-joint score Perricone: MCP2, wrist, knee bilaterally (Perricone et al, Rheumatology 2012). ⁴Candidate set A score: MCP1, MCP2, PIP3, radiocarpal, elbow, MTP1, MTP2, tibialis posterior (TP) tendon group and extensor carpi ulnaris (ECU) tendon bilaterally (Aga et al, Ann Rheum Dis 2015). ⁵Candidate set B score: same as candidate set A with addition of MCP5 and MTP5 (Aga et al, Ann Rheum Dis 2015). ⁶Candidate set C score: MCP1, MCP2, MCP3, PIP2, PIP3, radiocarpal, MTP2, MTP3 and ECU tendon bilaterally.

Disclosure: A. B. Aga, None; E. Lie, None; I. C. Olsen, None; H. B. Hammer, None; T. Uhlig, None; D. van der Heijde, None; T. K. Kvien, None; E. A. Haavardsholm, AbbVie, 2,Pfizer Inc, 2,MSD, 2,Roche Pharmaceuticals, 2,UCB, 2.

To cite this abstract in AMA style:

Aga AB, Lie E, Olsen IC, Hammer HB, Uhlig T, van der Heijde D, Kvien TK, Haavardsholm EA. Development of an Ultrasound Joint Inflammation Score for Rheumatoid Arthritis through a Data-Driven Approach [abstract]. Arthritis Rheumatol. 2015; 67 (suppl 10). https://acrabstracts.org/abstract/development-of-an-ultrasound-joint-inflammation-score-for-rheumatoid-arthritis-through-a-data-driven-approach/. Accessed .

« Back to 2015 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/development-of-an-ultrasound-joint-inflammation-score-for-rheumatoid-arthritis-through-a-data-driven-approach/