ACR Meeting Abstracts

ACR Meeting Abstracts

Abstract Number: 385

Validation of a Prognostic Model to Predict Structural Damage Assessed By X-Ray in Patients with RA Using MRI Data from a Clinical Trial

EA Alemao1, S Joo2, S Banerjee1, P Allison3, P Emery4, M Weinblatt5 and Katherine Liao6, 1Bristol-Myers Squibb, Princeton, NJ, 2Bristol-Myers Squibb, Hopewell, NJ, 3University of Pennsylvania, Philadelphia, PA, 4University of Leeds, Leeds, United Kingdom, 5Brigham and Women's Hospital, Boston, MA, 6Division of Rheumatology, Immunology and Allergy, Brigham and Women's Hospital, Boston, MA

Meeting: 2014 ACR/ARHP Annual Meeting

Keywords: ,

Session Information

Title: Rheumatoid Arthritis - Clinical Aspects: Novel Biomarkers and Other Measurements of Disease Activity

Session Type: Abstract Submissions (ACR)

Background/Purpose: We developed and validated a prognostic model for rapid radiographic progression (RRP) using X-ray data to identify RA patients (pts) at risk of structural joint damage. The objective of this study was to determine the RRP model performance using MRI as the outcome for structural joint damage. Methods: We analyzed data from a Phase IIB clinical study with the anti-IL-6 monoclonal antibody clazakizumab. This was a randomized, comparator-controlled, double-blind study of moderate-to-severe RA pts with inadequate response to MTX. Outcome data were available for MRI scans of the hands (12 wks) as well as disease activity and functional status (12 and 24 wks). MRI outcomes of synovitis, erosions and bone edema were evaluated using the RA-MRI scoring system (RAMRIS). We applied our prognostic model for RRP to pts in this study. The model includes seropositivity status, body weight, disease duration, DAS28 (CRP) and Total Sharp Score to determine the probability of RRP for each pt. The external validity of the model was evaluated for discrimination (receiver operating characteristic [ROC]) and calibration (Hosmer–Lemeshow goodness-of-fit chi-square). Based on the calculated probability of RRP, pts were categorized by probability into low (0–0.25), moderate (>0.25–0.75) and high (>0.75) risk of RRP. Analysis of variance was used to study the association between predicted probability of RRP at baseline to MRI outcome at 12 wks. Additionally, we examined the association between RRP prediction and disease activity [SDAI, DAS28 (CRP)] and functional status (HAQ) at 12 and 24 wks.   

Results: There were 418 pts in the clazakizumab Phase IIB study; average age was 50.4 yrs (SD 12.3); 82.1% were female. The RRP model when applied to clazakizumab Phase IIB had an overall ROC of 0.73 (95% CI 0.62, 0.83) and Hosmer–Lemeshow chi-square of 13.5 (df=8; p=0.1). Baseline probability of RRP was evaluated in 387 (92.6%) pts with available data. Of these, the majority (96.1%) were in the moderate- and high-risk (48.1% each) RRP groups. Pts in the moderate-risk group, when compared with those in the high-risk group, tended to be younger (mean age [SD] 48.3 [12.9] vs 52.2 [11.0] yrs). Compared with the moderate-risk RRP group, pts at high risk of RRP had significantly higher MRI measures of erosion, synovitis and joint space narrowing. In addition, pts at baseline in the high–risk group compared with the moderate-risk group had significantly higher disease activity scores and higher physical activity scores as measured by HAQ at 12 and 24 wks (Table).  

Table. Differences between MRI, disease activity, and functional outcomes in pts at high risk versus moderate risk of RRP

Outcomes

12 weeks

24 weeks

D (high versus moderate) risk of RRP

95% CI

D (high versus moderate) risk of RRP

95% CI

MRI scores

Erosion (MRI)

10.84

7.03, 14.66

–

–

Synovitis (MRI)

2.63

1.54, 3.71

–

–

Edema (MRI)

1.06

–0.90, 3.02

–

–

Joint space narrowing (MRI)

6.54

3.59, 9.50

–

–

Disease activity and functional outcomes

DAS28 (CRP) C

0.72

0.36, 1.09

0.78

0.42, 1.14

SDAI

8.22

4.38, 12.05

8.11

4.63, 11.58

HAQ

0.43

0.25, 0.60

0.47

0.29, 0.65

Each line represents a mutivariable linear regression model of outcomes

Conclusion: The RRP model based on X-ray progression had good validity in predicting structural joint damage measured by MRI. In line with clinical thinking, we observed that patients at high baseline risk of RRP compared with those with moderate risk tended to have higher disease activity and worse functional status in the longer term.   

Disclosure:

E. Alemao,

BMS,

1,

BMS,

3;

S. Joo,

BMS,

3,

BMS,

1;

S. Banerjee,

BMS,

1,

BMS,

3;

P. Allison,
None;

P. Emery,

AbbVie, BMS, Merck, Pfizer, Roche, Takeda,

5,

AbbVie, BMS, Merck, Pfizer, Roche,

2;

M. Weinblatt,

BMS, Crescendo Bioscience, UCB, Abbvie, Roche, Janssen,

5,

BMS, Crescendo Bioscience, UCB,

2;

K. Liao,
None.

« Back to 2014 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/validation-of-a-prognostic-model-to-predict-structural-damage-assessed-by-x-ray-in-patients-with-ra-using-mri-data-from-a-clinical-trial/