Background/Purpose

Magnetic resonance imaging (MRI) can detect early joint inflammation with high sensitivity, without use of radiation. The MARVELOUS study (NCT01235598), using the OMERACT RA MRI scoring system (RAMRIS), previously established 16 weeks (wks) as the earliest time point at which certolizumab pegol (CZP) had a statistically significant reduction on synovitis and bone edema.¹ To further evaluate efficacy of CZP on synovitis, dynamic contrast-enhanced MRI (DCE-MRI) parameters were examined on proximal interphalangeal (PIP) joints 2-5 (PIP 2-5), metacarpophalangeal (MCP) joints 2-5 (MCP 2-5), and PIP 2-5 and MCP 2-5 joints combined (PIP+MCP 2-5).

Methods

A total of 41 patients (pts) with active RA despite DMARD therapy and ≤1 biological therapy were randomized 2:1 to CZP, or 2 wks placebo followed by CZP (PBO-CZP). 40 were treated (27 CZP, 13 PBO-CZP), 36 completed Wk16 (24 CZP, 12 PBO-CZP). The approved CZP dose was administered: CZP loading dose 400mg every 2 wks (Q2W) Wks 0-4 for CZP pts or Wks 2-6 for PBO-CZP, followed by CZP 200mg Q2W to Wk16. The DCE-MRI parameters Initial Rate of Enhancement (IRE), Maximum Enhancement (ME; % increase compared to pre-contrast) and Number of voxels with Plateau and Washout Pattern (Nvox) were used to quantify contrast-enhanced MRIs. Data were analyzed using Dynamika software (Image Analysis, UK) by an experienced reader blinded to time, pt identity and treatment group who manually performed precise outlining of regions of interest at each joint for all 6 acquired time points (Wks 0, 1, 2, 4, 8, 16). Each DCE-MRI parameter was analyzed for PIP 2-5, MCP 2-5, and PIP+MCP 2-5. Intrareader variability (intraclass correlation coefficients [ICCs]) was assessed.

Results

There were statistically significant changes from baseline at Wk16 in the CZP group for PIP 2-5 in median IRE (-0.009; p=0.024) and median ME (-0.33; p=0.021), and for MCP 2-5 and PIP+MCP 2-5 in median Nvox (-235.0; p=0.025 and -421.5; p=0.015, respectively). Changes at Wk8 were not statistically significant, precluding earlier timepoint testing (Table 1). For PBO-CZP vs CZP comparisons at Wks 1 and 2, 3 of the 9 (ME for MCP 2-5 and PIP+MCP 2-5, Nvox for MCP 2-5) had nominal p-values <0.05 at Wk1. However, no nominal p-values were <0.05 at Wk2. The complete set of results is displayed in Table 2. All parameters had very good intrareader reliability (ICC>0.90).

Conclusion

Statistically significant improvements in dynamic MRI variables were observed at Wk16 following initiation of CZP therapy, despite small sample size. Dynamic MRI results were consistent with primary analysis (RAMRIS scoring).¹ Thus, both conventional MRI RAMRIS and DCE-MRI criteria confirmed efficacy of CZP in reducing joint inflammation in RA pts.

References

1. Østergaard M. Arthritis Rheum 2013;65(s10):1976