Background/Purpose: Temporal artery biopsy (TAB) remains the gold standard for the diagnosis of giant cell arteritis (GCA). It is invasive and its sensitivity is limited by segmental arterial involvement. We sought to determine the diagnostic accuracy of high-field MRI compared to TAB in patients with suspected GCA.

Methods: All patients referred for TAB at our center were approached for this study. A high-resolution 3T MRI of the scalp arteries was obtained before the TAB was performed. Arterial abnormalities on MRI were assessed according to a previously published grading scheme, evaluating mural thickness and enhancement based on multiplanar postcontrast T1-weighted spin-echo images. Diagnostic accuracy of MRI was evaluated by comparison with TAB results as a primary analysis. Secondary analyses included comparison to clinician diagnosis and ACR criteria.

Results: 191 patients were screened and 171 were included in the study. Exclusions were based on withdrawal of consent (11), contra-indication to MRI (3), non-diagnostic test (3 MRI, 2 TAB) or failure to obtain MRI (1). ACR criteria were met in 137 patients (80.1%). Physician diagnosis was available for 162 subjects, with 78 (48.2%) considered as GCA. The MRI showed abnormal scalp arteries in 60 patients (35.1%) while biopsy was positive in 31 (18.1%). MRI was positive in 29 of those 31 patients with positive TAB (Sensitivity 93.6%). MRI was normal in 109 of those 140 with negative TAB (Specificity 77.9%). Relative to TAB, the negative predictive value of MRI was 98.2%.

Conclusion: High-resolution MRI detects biopsy-positive GCA with high sensitivity. A negative MRI is highly predictive of a negative TAB, such that patients with a negative MRI could safely be spared TAB. The significance of a positive MRI is not as well defined, and this should be the focus of future research.