Abstract
OBJECTIVES: To derive segmental cut-off values and measures of diagnostic accuracy for the intima-media thickness of compressed temporal artery segments for the diagnosis of giant cell arteritis (GCA) on the patient level. To examine the influence of cardiovascular risk. METHODS: Retrospectively, patients evaluated for GCA with an ultrasound of the temporal arteries and an MRI of the head, including a T1-fatsat-black blood (T1-BB) sequence, were identified and classified based on cardiovascular risk and a dual reference standard of T1-BB on the segmental level and the clinical diagnosis on the patient level. Intima-media thickness of the common superficial temporal artery (CSTA), frontal and parietal branches (FB, PB) were measured by compression technique. Statistically and clinically optimal (specificity of approx. 90% for the patient level) cut-offs were derived. Diagnostic accuracy was evaluated on the patient level. RESULTS: The population consisted of 144 patients, 74 (51.4%) with and 70 (48.6%) without GCA. The statistically optimal cut-offs were 0.86 mm, 0.68 mm and 0.67 mm for the CSTA, the FB and PB, respectively. On the patient level sensitivity and specificity were 86.5 and 81.4%. Clinically optimal cut-offs were 1.01 mm, 0.82 mm and 0.69 mm and showed a sensitivity of 79.7% and a specificity of 90.0%. For patients without high cardiovascular risk, statistically optimal cut-offs showed a sensitivity of 89.6% and a specificity of 90.5%. CONCLUSION: Newly derived ultrasound intima-media thickness cut-offs with a dual reference standard show high diagnostic accuracy on the patient level for the diagnosis of GCA, particularly in patients without high cardiovascular risk.