Arthritis & Rheumatism, Volume 60,
October 2009 Abstract Supplement

The 2009 ACR/ARHP Annual Scientific Meeting
Philadelphia October 16-21, 2009.


Rheumatologist Training to Recognize Lesions On T1-Weighted MRI Enhances Diagnostic Utility of MRI in Patients with Ankylosing Spondylitis

Maksymowych1,  Walter P., Weber2,  Ulrich, Hodler2,  Juerg, Ostergaard3,  Mikkel, Pedersen3,  Susanne J., Lambert1,  Robert GW

University of Alberta, Edmonton, AB,
University Clinic Balgrist, Zurich, Switzerland,
Copenhagen University Hospitals at Hvidovre and Gentofte. Copenhagen, Denmark

Purpose:

In using MRI to diagnose AS, we have shown that rheumatologists rely primarily on diagnostic information from fat-suppressed sequences where inflammation is readily discernable and de-emphasize the contribution of abnormalities on T1-weighted sequences (T1W) which may reflect difficulty in interpreting complex structural abnormalities such as erosions and sclerosis and the uncertain diagnostic significance of fat infiltration. We aimed to test the hypothesis that systematic calibration of readers directed at recognition of abnormalities on T1W MRI would enhance the diagnostic utility of MRI in early AS.

Method:

Six experienced readers (4 rheumatologists, 2 radiologists) from 3 international centres, blinded to patient and diagnosis, independently assessed MRI scans (T1W and STIR) from the following subjects all aged < 45 years: 77 patients with AS; 85 controls (26 with mechanical low back pain (mLBP) and 59 healthy individuals). MRI scans of the SIJ were read systematically as described in a standardized online training module. We recorded bone edema, fat infiltration, erosions, and ankylosis according to standardized definitions using an online data entry system. In addition to deciding on the presence/absence of SpA, readers were asked to record the following: 1. Whether their diagnostic conclusion was primarily based on abnormalities observed on the STIR, T1W, or both. 2. Which type of lesion was the primary basis for the diagnostic conclusion. The exercise was repeated 6 months later on a random selection of 30 AS patients (symptom duration <= 5 years) and 34 controls (26 with mLBP and 8 healthy) from the original cohort after 2 training exercises directed at abnormalities on T1W MRI and development of a reference image set developed by group consensus.

Results:

Diagnostic utility of MRI improved for both concordant as well as individual readers following calibration. Rheumatologist readers also based their diagnostic conclusion on the T1W MRI and recorded erosions as the principal diagnostic feature significantly more frequently after calibration (p = 0.013 and <0.0001, respectively). However, this was not accompanied by significant improvement in the inter-reader agreement of detection of erosions as recorded in individual SIJ quadrants.

 Pre-CalibrationPost-calibration
 SensitivitySpecificityLR+LR-SensitivitySpecificityLR+LR-
Individual reader (mean (range))92.2% (76.7–100)93.6% (79.4–100)14.40.0897.9% (96.7–100)95.6% (88.2–100)18.10.02
All Rheumatologists (mean (range)91.7% (76.7–100)91.2% (79.4–100)10.40.0997.8% (96.7–100)94.1% (88.2–97.1)16.60.02
Concordant data (all 6 readers)66.7%76.5%2.80.4490.0%82.4%5.10.12

Conclusion:

Rheumatologist training directed specifically at detection of MRI abnormalities on T1W scans enhances diagnostic utility by increasing overall interpretation of the T1W scan.

To cite this abstract, please use the following information:
Maksymowych, Walter P., Weber, Ulrich, Hodler, Juerg, Ostergaard, Mikkel, Pedersen, Susanne J., Lambert, Robert GW; Rheumatologist Training to Recognize Lesions On T1-Weighted MRI Enhances Diagnostic Utility of MRI in Patients with Ankylosing Spondylitis [abstract]. Arthritis Rheum 2009;60 Suppl 10 :540
DOI: 10.1002/art.25622

Abstract Supplement

Meeting Menu

2009 ACR/ARHP