946 - Construct Validity and Responsiveness of Dynamic Contrast-Enhanced Magnetic Resonance Imaging in Early Rheumatoid Arthritis – A Comparison with Conventional Magnetic Resonance Imaging and Clinical Measures of Disease Activity

Monday, November 7, 2011: 9:00 AM-6:00 PM
Hall F2 - Poster Hall (McCormick Place West)
Mette Bjørndal Axelsen1, Bo J. Ejbjerg2, Merete L. Hetland1, Kim Hørslev-Petersen3, Mikael Boesen4, Olga Kubassova5, Ulrik B. Lauridsen1, Ole Majgaard2, Henning Bliddal4, Niels Steen Krogh6 and Mikkel Østergaard1, 1Copenhagen University Hospital at Glostrup, Copenhagen, Denmark, 2Copenhagen University Hospital at Slagelse, Slagelse, Denmark, 3University of Southern Denmark, Graasten, Denmark, 4The Parker Institute, Copenhagen University Hospital at Frederiksberg, Frederiksberg, Denmark, 5Image Analysis Ltd., Leeds, United Kingdom, 6ZiteLab ApS, Copenhagen, Denmark
Presentation Number: 946

Background/Purpose: The aim of the study was to assess the responsiveness and construct validity of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) parameters during aggressive treatment of early rheumatoid arthritis (RA), by comparison with conventional MRI and clinical measures of disease activity.

Method: DCE-MRI and conventional contrast-enhanced MRI of the non-dominant hand (wrist and 2.-5. metacarpophalangeal joints (MCP)) and clinical assessment were performed at months 0, 6 and 12 in 14 patients with early RA treated with disease modifying antirheumatic drugs (DMARDs) (methotrexate +/- cyclosporine) and intra-articular glucocorticoid in an investigator initiated clinical trial [1]. Conventional MRI was evaluated using the Rheumatoid Arthritis MRI score (RAMRIS). DCE-MRI (one coronal slice) was analysed using a computer-aided approach by which the temporal contrast uptake is characterized by parameters such as the Initial Rate of Enhancement (IRE), Maximal Enhancement (ME) and number of enhancing voxels (Nvoxel)[2]. The analysis was performed for 2 manually outlined Regions of Interest (ROIs) covering the wrist and 2.–5. MCP, respectively, and for the sum of wrist and MCP ROIs, ('Wrist+MCP').

Results: All DCE-MRI parameters in the 'Wrist+MCP' declined significantly from month 0 to 12 (see table for separate wrist and MCP values and SRMs). The SRMs for RAMRIS synovitis was -0.67 to -1.05 from month 0 - 6 and 0 - 12. DAS28, and tender and swollen joint counts declined similarly (SRMs -0.51 to -1.11). Correlations with RAMRIS synovitis were high for ME (rho 0.56 – 0.87, p<0.04), Nvoxel (rho 0.76 – 0.80, p≤0.002) and IRExNvoxel (rho 0.73 – 0.81, p≤0.003), and low – moderate for IRE (0.09 – 0.56, p ≤0.04 - 0.76).   

Table 1. Responsiveness of DCE-MRI, Conventional MRI and Clinical Parameters

 

Baseline

Median

(range)

6 months

Median

(range)

12 months

Median

(range)

Change

0-6 months

Median

(range)

Change

0-12 months

Median

(range)

SRM

0-6

months

SRM

0-12

months

Dynamic Contrast-Enhanced Magnetic Resonance Imaging Parameters

Wrist+MCP IRE

(%/s)

1.79

(1.37 – 3.54)

1.58

(0.91 – 2.68)

1.55

(0.93 – 3.06)

-0.22

(-2.12 – 1.00)*

-0.42

(-2.22 – 1.38)*

-0.49

-0.53

Wrist+MCP ME

1.53

(1.19 – 2.04)

1.43

(1.15 – 1.80)

1.32

(1.16 – 1.83)

-0.21

(-0.53 – 0.28)

-0.13

(-0.48 – 0.31)*

-0.55

-0.61

Wrist+MCP Nvoxel

(voxels)

1241

(267 – 7964)

642

(12 – 2712)

505

(16 – 2189)

-665

(-6158 - 1867)

-644

(-424 - 1344)**

-0.49

-0.65

Wrist+MCP IRExNvoxel

(%voxels/s)

2193

(365 – 23943)

1168

(11 – 7281)

741

(15 – 6699)

-1192

(-21353 - 5859)

-1337

(-21162 - 5277)**

-0.43

-0.53

Wrist IRE

(%/s)

1.88

(1.36 – 3.74)

1.63

(0.91 – 2.98)

1.55

(1.08 – 3.33)

-0.28

(-2.33 – 1.31)

-.54

(-2.42 – 1.66)*

-0.49

-0.54

Wrist ME

1.58

(1.19 – 2.16)

1.44

(1.14 – 1.92)

1.33

(1.15 – 1.93)

-0.16

(-0.64 - 0.37)

-0.13

(-0.60 – 0-38)*

-0.49

-0.57

Wrist Nvoxel

(voxels)

759

(193 – 5473)

552

(12 – 1732)

356

(9 – 1547)

-238

(-4687 – 958)

260

(-773 – 4511)*

-0.37

-0.56

Wrist IRExNvoxel

(%voxels/s)

1394

(269 – 17951)

936

(11 – 5161)

522

(10 – 5151)

-467

(-16756 – 3869)

-679

(-16287 - 3858)*

-0.36

-0.46

MCP IRE

(%/s)

1.61

(1.21 – 2.66)

1.51

(0.00 – 1.84)

1.56

(0.50 – 1.95)

-0.07

(-1.59 – 0.44)

-0.05

(-1.3 – 0.28)

-0.41

-0.37

MCP ME

1.48

(1.16 – 1.75)

1.26

(0.00 – 1.60)

1.25

(1.08 – 1.83)

-0.12

(-1.27 – 0.29)

-0.064

(-0.38 – 0.29)

-0.54

-0.38

MCP Nvoxel

(voxels)

389

(69 – 2491)

131

(0.00 – 1020)

90

(6 – 900)

-278

(-1676 - 909)*

-215

(-1913 - 571)*

-0.61

-0.70

MCP IRExNvoxel

(%voxels/s)

536

(96 – 5082)

196

(0 – 1803)

135

(4 – 1647)

-484

(-3694 - 1675)*

-342

(-3972 – 1123)*

-0.58

-0.65

Rheumatoid Arthritis Magnetic Resonance Imaging Scores (RAMRIS)

Wrist+MCP RAMRIS Synovitis

13

(3 – 18)

5.5

(3 – 19)

5

(3 – 21)

-4.5

(-14 – 1)*

-3.5

(-15 – 4)*

-1.05

-0.89

Wrist
RAMRIS synovitis

7

(3 - 9)

3

(3 - 6)

4

(2 - 8)

-3

(-6 – 2)**

-2

(-6 – 2)**

-1.04

-1.02

MCP
RAMRIS synovitis

6

(1 - 9)

3

(0 - 7)

2

(0 - 6)

-2

(-9 – 2)**

-1.5

(-9 – 4)*

-0.88

-0.67

Clinical Parameters

DAS28(crp)

3.99

(2.76 – 6.85))

2.78

(1.66 – 3.46)

2.23

(1.67 – 4.45)

-0.63

(-3.68 – 1.83)*

-0.78

(-4.54 – 1.81)**

-0.52

- 0.53

Swollen Joint Count

(28 joints)

3

(1 – 14)

0

(0 – 2)

0

(0 – 1)

-3.5

(-12 – 0)***

-3.0

(-14 – 0)**

-1.11

-1.10

Tender Joint Count
(28 joints)

5

(0 – 19)

0

(0 – 5)

0

(0 – 8)

-2

(-17 – 4)*

-2.5

(-19 – 4)*

-0.69

-0.51

Wilcoxon Signed Rank test: p≤0.05:*, p≤0.01:**, p≤0.001:***
Correlations between parameters (Spearman’s Correlation Coefficient, rho), changes over time (Wilcoxon Signed Rank test) and responsiveness (Standardized Response Mean (SRM)) were calculated. 

Conclusion: 

All DCE-MRI parameters declined significantly during treatment. With regard to responsiveness, DCE-MRI was comparable to DAS28 but inferior to RAMRIS synovitis. DCE-MRI demonstrated construct validity and moderate responsiveness.

References:
1 Hetland ML et al. Arthritis Rheum 2006;54:1401-1409
2 Kubassova O et al. Med Image Comput Comput Assist Interv Int Conf Med Image Comput Comput Assist Interv 2007;10:261-269.


Keywords: magnetic resonance imaging (MRI) and rheumatoid arthritis (RA)

Disclosure: M. B. Axelsen, Abbott Laboratories, 2 ; B. J. Ejbjerg, None; M. L. Hetland, None; K. Hørslev-Petersen, Abbott Immunology Pharmaceuticals, 6, UCB, Roche, 8, Danish Association agains Rheumatism, 2 ; M. Boesen, Image Analysis Ltd., 9 ; O. Kubassova, Image Analysis Ltd., 4 ; U. B. Lauridsen, None; O. Majgaard, None; H. Bliddal, None; N. S. Krogh, None; M. Østergaard, Abbott Laboratories, 2 .