RESUMEN
BACKGROUND: Extracapsular inflammation at entheseal sites in the pelvic girdle as demonstrated by magnetic resonance imaging (MRI) was shown to be useful as an additional tool for diagnosing polymyalgia rheumatica (PMR). However, it is unclear whether MRI needs to be performed with contrast enhancement or whether oedema-sensitive sequences are sufficient. OBJECTIVE: To evaluate the performance of T2w TIRM (turbo inversion recovery magnitude) imaging compared to fat-saturated contrast-enhanced (ce) T1w at predefined pelvic sites to detect extracapsular inflammation in patients with PMR. METHODS: A total of 120 pelvic MRIs of patients with pelvic girdle pain, 40 with clinically diagnosed PMR and 80 controls, were retrospectively scored by three blinded radiologists separately evaluating the MRI with and without contrast enhancement at 19 previously defined pelvic structures. The intra- and interrater reliability and the diagnostic performance of both techniques were statistically analysed and evaluated. RESULTS: The detection of inflammatory MRI signals correlated moderately between both techniques (Cohen's κ 0.583). With ceT1w imaging 20.7% more sites were detected as inflamed compared to T2w TIRM in PMR patients. Inter- and intrareader reliability was superior with ceT1w imaging. If the inflammatory signal was detected at three sites bilaterally including the origin of the rectus femoris muscle or adductor longus muscle, the sensitivity and specificity was 100% and 97.1% by ceT1w imaging vs. 80.8% and 93.3% by T2w TIRM, respectively. CONCLUSION: Contrast enhancement is superior to oedema-sensitive MRI in the detection of extracapsular inflammation in PMR. However, using T2w TIRM also detects many but not all PMR cases.
RESUMEN
OBJECTIVE: There is currently no diagnostic test for PMR. A characteristic pattern of extracapsular inflammation as assessed by contrast-enhanced MRI (ceMRI) has recently been described in the pelvis of patients with PMR. We aimed to evaluate the performance of inflammatory ceMRI signals at predefined pelvic sites as a diagnostic test for PMR. METHODS: Pelvic MRI scans of patients with pelvic girdle pain (n = 120), including 40 patients with an expert diagnosis of PMR and 80 controls with other reasons for pelvic pain were scored by three blinded radiologists, who evaluated the degree of contrast enhancement at 19 predefined tendinous and capsular pelvic structures. Different patterns of involvement were analysed statistically. RESULTS: The frequency of bilateral peritendinitis and pericapsulitis including less common sites, such as the proximal origins of the m. rectus femoris and m. adductor longus, differed significantly between PMR cases and controls: 13.4 ± 2.7 vs 4.0 ± 2.3. A cut-off of ≥10 inflamed sites discriminated well between groups (sensitivity 95.8%, specificity 97.1%). Bilateral inflammation of the insertion of the proximal m. rectus femoris or adductor longus tendons together with ≥3 other bilaterally inflamed sites performed even better (sensitivity 100%, specificity 97.5%). CONCLUSION: This study confirms that a distinctive MRI pattern of pelvic inflammation (bilateral peritendinitis and pericapsulitis and the proximal origins of the m. rectus femoris and m. adductor longus) is characteristic for PMR. The high sensitivity and specificity of the set of anatomical sites evaluated suggests their clinical usefulness as a confirmatory diagnostic test.