Despite major recent advances in the therapeutic management of Giant cell arteritis (GCA), the diagnosis accuracy of temporal artery ultrasound remains controversial in this disease. We performed a systematic review to determine the sensitivity, specificity, and summary positive (LR+) and negative (LR-) likelihood ratios of temporal artery ultrasound for the diagnosis of GCA. For this, we searched EMBASE, MEDLINE and the Cochrane Database of Systematic Reviews without language restriction. Original articles reporting on diagnostic accuracy of temporal artery ultrasound compared to temporal artery biopsy, for the diagnosis of GCA, were selected. Sensitivity and specificity from each study were used to fit a bivariate diagnosis accuracy model. Of 1280 articles identified, 48 underwent full-text review, and 25 were included. Based on a total of 20 studies, the sensitivity and specificity of hypoechoic halo compared to positive temporal artery biopsy were respectively of 68% (95% CI: 57-78) and 81% (95%CI: 75-86). The summary mean positive and negative likelihood ratios were respectively of 3.64 (95%CI: 2.76-4.73) and 0.40 (0.28-0.52). Taking into account 11 studies reporting on the presence of any abnormal sign on temporal artery ultrasound yielded similar results with largely overlapping 95% confidence interval regions. This study provides the summary estimates of the diagnostic properties of temporal artery ultrasound compared to temporal artery biopsy, for the diagnosis of GCA. Those parameters allow the calculation of the post-test probability of GCA in a given patient, based on the results of temporal artery ultrasound and will help improving the diagnosis strategy for this common disease.