AGE AND SEX BIAS OF ULTRASOUND SCORES FOR DIAGNOSIS OF GIANT CELL ARTERITIS: INSIGHTS FROM THE PROSPECTIVE, LONGITUDINAL HAS-GCA STUDY

K. Van der Geest, A. Sebastian, E. Conticini, S. Inness, J. Jackson, A. Kayani, M. Khurshid, G. Klinowski, P. Macchioni, D. Prieto-Peña, C. Salvarani, M. Tariq, A. Tomelleri, B. DasguptaUniversity of Groningen, University Medical Center Groningen, Rheumatology and Clinical Immunology, Groningen, Netherlands Mid and South Essex University Hospital Groups, Southend University Hospital, Rheumatology, Westcliff-on-sea, United Kingdom University of Essex, School of Sport, Rehabilitation and Exercise science, Colchester, United Kingdom University Hospital Limerick, Rheumatology, Limerick, Ireland University of Siena, Rheumatology Unit, Department of Medicine, Surgery and Neurosciences, Siena, Italy University Hospitals Dorset NHS Foundation Trust, Rheumatology, Dorset, United Kingdom Azienda USL - IRCCS - di Reggio Emilia, Rheumatology, Reggio Emilia, Italy University of Modena and Reggio Emilia, Rheumatology, Modena, Italy Marqués de Valdecilla University Hospital, Rheumatology, Santander, Spain San Raffaele Hospital, Unit of Immunology, Rheumatology, Allergy and Rare Diseases, Milan, Italy  Background Giant cell arteritis (GCA) is an age-linked, critically ischaemic vasculitis where ultrasound (US) serves as an important tool enabling urgent diagnosis and therapy. US scores have been developed to quantify the arterial inflammation in GCA [1,2], but optimal diagnostic cut-off points and potential bias of age and sex on these values remain un-elucidated. Objectives To compare the diagnostic accuracy of previously reported US scores for GCA, and to determine whether these scores require correction for age and sex. Methods The HAS-GCA study, with a prospective, multicentre, cohort design, recruited 229 consecutive patients with suspected, new-onset GCA from 7 sites located in the UK, Italy, Spain and the Netherlands. All cases underwent bilateral US evaluation of the three temporal artery segments (common superficial, parietal, frontal) and axillary arteries. Intimal-medial thickness (IMT) was measured in each segment. Three ultrasound scores were calculated based on the IMT: halo count (HCount) [1], Halo Score (HScore) [2] and the OMERACT GCA US Score (OGUS) [2]. The definition of halo was based on previously reported IMT cut-off values [3]. The reference standard was the final clinical diagnosis after 6 months follow-up. ROC with AUC analysis was performed. Optimal cut-off points were determined by Youden Index. Group comparisons were made by Mann Whitney U test and correlations by Spearman’s rank correlation coefficient. P values < 0.05 were considered statistically significant. Results A final diagnosis of GCA was made in 84 patients, whereas 145 cases received an alternative diagnosis (i.e. non-GCA patients). HCount, HScore and OGUS were significantly higher in GCA patients than non-GCA patients. In non-GCA patients, all three scores were significantly higher in men than women, and showed a statistically significant, positive correlation with age. In GCA patients, the US scores showed a positive correlation (HCount and OGUS) or strong trend for a positive correlation (HScore) with age; and the HCount was significantly higher in men. All scores effectively discriminated GCA from non-GCA patients as indicated by ROC analysis: AUC for halo count 0.936 (0.899-0.974), Halo Score 0.894 (0.848-0.939) and OGUS 0.937 (0.902-0.973). The optimal diagnostic cut-off points were: 2 for HCount (sens 86%, spec 92%), 17 for HScore (sens 76%, spec 91%) and 0.80 for OGUS (sens 87%, spec 93%). These diagnostic cut-off points differed in sensitivity and specificity within different age- and sex-defined groups with the highest sensitivity and lowest specificity observed among old (>81yr) male patients. Alternatively, age- and sex-stratified cut-off points could be established for the US scores (Table 1). Conclusion All previously reported US scores show excellent diagnostic accuracy for GCA, but are prone to over-diagnose GCA in older males. Our data highlight the need for age- and sex-specific diagnostic cut-off points. The excellent comparative performance of the simple HCount in GCA diagnosis suggests an exciting role in future clinical practice. References van der Geest KSM, et al. Annals of the Rheumatic Diseases 2020 Mar;79(3):393-399. Dejaco C, et al. Annals of the Rheumatic Diseases 2022 doi: 10.1136/ard-2022-223367 Schafer VS, et al. Rheumatology 2017 Sep 1;56(9):1632. Table 1. Ultrasound scores in different age and sex groups. Age cut-offs were based on 25th percentile and 75th percentile of age in female and male patients with GCA. Category Score AUC in ROC analysis Optimal cut-off point Sensitivity (%) Specificity (%) WomenAge <67 Halo count 0.897 2 75 98 Halo Score 0.883 12 83 90 OGUS 0.925 0.66 83 85 Womenage 67-79 Halo count 0.927 1 88 93 Halo Score 0.873 16 72 93 OGUS 0.933 0.80 80 98 WomenAge >79 Halo count 0.938 2 92 94 Halo Score 0.889 16 92 88 OGUS 0.928 0.83 92 94 MenAge <74 Halo count 0.919 1 100 68 Halo Score 0.877 15 88 74 OGUS 0.911 0.87 75 97 MenAge 74-81 Halo count 0.972 3 94 88 Halo Score 0.837 20 72 100 OGUS 0.972 0.76 100 88 MenAge >81 Halo count 0.958 4 100 83 Halo Score 0.885 19 88 83 OGUS 0.896 0.90 100 83 Acknowledgements K. van der Geest and A. Sebastian share first authorship. Disclosure of Interests Kornelis van der Geest Speakers bureau: Roche, Grant/research support from: AbbVie, Alwin Sebastian: None declared, Edoardo Conticini: None declared, Sue Inness: None declared, Jo Jackson: None declared, Abdul Kayani: None declared, Muhammad Khurshid: None declared, Giulia Klinowski: None declared, Pierluigi Macchioni: None declared, Diana Prieto-Peña: None declared, Carlo Salvarani: None declared, Mohammad Tariq: None declared, Alessandro Tomelleri: None declared, Bhaskar Dasgupta Consultant of: Roche, Chugai, Sanofi, Grant/research support from: Roche, Sanofi, AbbVie, and GlaxoSmithKline. Keywords: Ultrasound, Imaging, Vasculitis DOI: 10.1136/annrheumdis-2023-eular.5139Citation: , volume 82, supplement 1, year 2023, page 657Session: Vasculitis - large vessel vasculitis (Poster View)