Early rheumatoid arthritis, two distinctive structural damage patterns revealed by MRI: an 8-year longitudinal study

Abstract

Objective: To determine how inflammatory and structural parameters change long-term on standard treatment in rheumatoid arthritis patients and which baseline parameter best predicts long-term structural damage.

Material and methods: Prospective study of early rheumatoid arthritis (ERA) patients (symptom duration ≤ 24 months) who underwent identical clinical, serological, radiographic, and dynamic contrast-enhanced MRI of the wrist assessments at baseline, year-1, and year-8. MR images were analyzed semi-quantitatively (Rheumatoid Arthritis Magnetic Resonance Imaging Score [RAMRIS]) and quantitatively (synovial volume (cm³); synovial perfusion; bone marrow edema (BME) proportion [%]). Multivariate analyses and receiver operating curves were applied to find the best predictor of long-term structural damage.

Results: 81 patients (61 ± 12 years, F/M:67/14) were studied. MRI-detected inflammatory parameters markedly improved from baseline to year-1 and slightly deteriorated from year-1 to year-8 (synovial volume:6.7 ± 5.0→2.6 ± 2.9→3.6 ± 3.3 cm³ (p < 0.01); BME proportion:13.1 ± 9.3→7.4 ± 5.0→9.2 ± 9.7% [p < 0.01]). Structural damage progressively deteriorated from baseline to year-8. Two long-term structural damage pattern groups were apparent, namely a "non-progressive structural damage pattern" (62%, 50/81) and a "progressive structural damage pattern" (38%, 31/81). Functional impairment was more frequent and more severe at year-8 in patients with progressive structural damage. MRI-detected bone erosion score better predicted (AUC = 0.81, CI: 0.71-0.91) year-8 structural damage than clinical (SDAI AUC = 0.61, CI: 0.48-0.74), serological (CRP AUC = 0.60, CI: 0.47-0.73), or radiographic (AUC = 0.59, CI: 0.45-0.72) assessment.

Conclusion: In ERA patients, two distinct structural damage patterns are evident. Baseline bone erosion score is better than clinical, serological, or radiographic assessment at predicting long-term structural damage.

Key points: Questions The value of MRI in predicting long-term structural damage in ERA patients is not clear. Findings This study identified two distinct long-term structural damage progression patterns of ERA patients. MRI can better differentiate between these two groups at baseline than clinical, serological, or radiographic assessment. Clinical relevance MRI examination should be performed in all ERA patients at baseline to determine their structural damage pattern. This will allow a better prediction of patient outcomes in the long-term.

Keywords: Follow-up study; MRI; Synovitis; Wrist, Rheumatoid arthritis.

Fig. 1

Flowchart of patient enrolment at baseline, year-1, and year-8. Sixty-four patients completed the baseline, year-1, and year-8 assessments. 17 patients did not undergo year-1 follow-up. #Excluded at year-1 due to diagnosis of systemic lupus erythematosus (n = 3) and peripheral spondylarthritis (n = 1). * Contrast enhanced MRI was contradicted because of implant (n = 3), claustrophobia (n = 1), malignant tumor (n = 4), kidney failure (n = 1), and age more than 90 years (n = 1). There was no significant difference in clinical, serological, radiographic, or MRI parameters at baseline between the 50 patients who did not complete and the 81 patients who did complete the year-8 follow-up (p > 0.05 for all)

Fig. 2

Two structural damage progression patterns at baseline, year-1, and year-8. a Non-progressive structural damage group: 50-year-old female T2FS coronal MR images of wrist. Structural damage score (bone erosion+ JSN) was 1 at baseline, 1 at year-1, and 2 at year-8. b Progressive structural damage group; 51-year-old female T2FS coronal MR images of wrist. Structural damage score was 22 at baseline, 25 at year-1, and 69 at year-8

Fig. 3

Case-by-case representation of structural damage score from baseline to year-8 in 81 ERA patients. Red lines represent patients (n = 31, 38%) patients with a “progressive structural damage pattern”. Blue lines represent patients (n = 50, 62%) with a “non-progressive structural damage pattern”. A structural damage score of 15 (actual value 14.5) was the optimal Youden Index value for differentiating between non-progressive and progressive structural damage groups at baseline

Fig. 4

Inflammatory and structural damage parameters in the non-progressive structural damage group (blue line) and progressive structural damage group (red line) at baseline, year-1 and year-8. Mean and 95% confidence interval are depicted. a Synovial volume (cm³). b Bone marrow edema proportion (%). c Bone erosion score, and d Joint space narrowing score

Fig. 5

Receiver operating characteristics (ROC) curve in predicting structural damage progression at year-8. a Predictive capacity of baseline clinical parameters. b Predictive capacity of baseline MRI parameters and SHS radiographic score. c Predictive capacity of year-1 MRI parameters and SHS radiographic score. AUC, area under the curve; SDAI, Simple Disease Activity Index; DAS, disease activity score; BME, bone marrow edema; CRP, C-reactive protein; Syn. Vol, synovitis volume; SHS, Sharp-van der Heijde score