The Prince of Wales Hospital Woollahra, New South Wales, Australia
Disclosure information not submitted.
Keren Port1, Ivan Ho Shon1, Sally Ayesa1, Rachel Langford1, Olivia Bennett1, Andrew Csillag1, Stacey Fredericks1, Luz Palacios-Derflingher2, Eva Wegner1 and Anthony Sammel1, 1The Prince of Wales Hospital, Sydney, Australia, 2University of New South Wales, Sydney, Australia
Background/Purpose: GCA is a systemic medium-large vessel vasculitis (M-LVV) involving the cranial arteries, aorta and its major branches. Recent studies have shown that PET/CT including the cranial and large vessels offers high diagnostic accuracy for GCA in research settings. At our institution in Sydney, Australia, PET/CT is the protocolized first-line investigation for GCA, and here we report its diagnostic performance in this "all comer" real world setting.
Methods: We audited all inpatients and outpatients investigated with PET/CT for suspected new onset GCA between July 2019 and March 2022. Patients were scanned from vertex to thighs with arms by their side, 90 minutes post-fluorodeoxyglucose (FDG) administration. Patients completed a self-reported questionnaire at the time of scan. Scans were reported by the experienced nuclear medicine physician of the day as overall positive, equivocal or negative for M-LVV. FDG uptake by vascular region was also reported. The gold standard comparator was the treating clinician's diagnosis of active M-LVV at a minimum of 6 months post the time of the scan. Active M-LVV included diagnoses of GCA, large vessel vasculitis and aortitis. Temporal artery biopsy (TAB) and vascular US results conducted within 4 weeks of scan were recorded.
Results: 135 patients underwent PET/CT for suspected new onset GCA, with a median age of 73 years (range 48-98), 70% were female. 86 (63.7%) patients were taking glucocorticoids at the time of scan. 44 (32.6%) patients had a final clinical diagnosis of active M-LVV. 35 (25.9%) scans were reported as positive, 27 (20%) equivocal and 73 (54.1%) negative for M-LVV. The likelihood ratio for a diagnosis of active M-LVV with a positive scan was 70.3, equivocal scan 0.59, and a negative scan 0.12. Sensitivity (Sn) and specificity (Sp) of PET/CT was dependent on treatment of equivocal scans in the binary analysis. If equivocal scans were excluded, Sn was 89.5% and Sp 98.6%. If equivocal scans were considered positive, Sn was 90.9% and Sp 75.8%. If equivocal scans were considered negative, Sn was 77.3% and Sp 98.9%. 55/135 (40.7%) patients underwent TAB and/or US (27 TAB, 33 US, 5 both TAB and US) as an additional investigation to PET/CT. 13/55 (23.6%) in this cohort had equivocal PET/CT scans, and 11/13 of the additional test (TAB/US) results matched the clinical diagnosis. In the total cohort, TAB had a Sn of 42.9% and Sp of 100%, and US had a Sn range of 53.3% to 66.7% (dependent on treatment of the 2 equivocal US) and Sp of 88.9%. For the 34 patients with positive PET/CT scans and a clinical diagnosis of M-LVV, significant FDG uptake was most commonly reported in the supra-aortic (88.2%), followed by cranial (55.9%), aortic (52.9%) and infra-aortic (23.5%) regions.
Conclusion: This is the largest real-world study of combined cranial and large vessel PET/CT as a first-line test for GCA. PET/CT had good diagnostic performance with sensitivity ranging from 77.3 to 90.9%, and specificity 75.8 to 98.9% for a clinician diagnosis of active M-LVV, dependent on treatment of equivocal scans. TAB and/or US can be used as ancillary tests to help confirm a diagnosis when PET/CT is indeterminate or discordant with the clinical presentation.
Table 1: Patient baseline characteristics Values are number (%) of patients, except where indicated otherwise PMR = polymyalgia rheumatica TAB = temporal artery biopsy US = vascular ultrasound temporal +/- axillary arteries High dose glucocorticoid use = prednisone ≥40mg oral daily or methylprednisolone ≥500mg IV daily M-LVV = medium-large vessel vasculitis
Table 2: Overall interpretation of PET/CT compared to clinician final diagnosis
Table 3: PET/CT performance analysis for clinician final diagnosis of M-LVV, M-LVV = medium-large vessel vasculitis PPV = positive predictive value NPV = negative predictive value CI = confidence interval AUC = area under the receiver operating curve
K. Port: None; I. Ho Shon: None; S. Ayesa: None; R. Langford: None; O. Bennett: None; A. Csillag: None; S. Fredericks: None; L. Palacios-Derflingher: None; E. Wegner: None; A. Sammel: None.
.jpg)
.jpg)
.jpg)
