[18F]FDG positron emission tomography in patients presenting with suspicion of giant cell arteritis—lessons from a vasculitis clinic

[18F]FDG positron emission tomography in patients presenting with suspicion of giant cell... Abstract Aims The usefulness of [18F] fluorodeoxyglucose-positron emission tomography/computed tomography ([18F]FDG-PET/CT) for diagnosing giant cell arteritis (GCA) has been previously reported. Yet, the interpretation of PET scans is not clear-cut. The present study aimed at determining the best method to analyse PET/CT in a large, real-life cohort of patients presenting with suspicion of GCA. Methods and results One hundred and three patients with clinical suspicion of GCA undergoing PET/CT between 2006 and 2012 were included. Clinical data were retrieved from patients’ charts. PET/CT was categorized by visual scoring of the uptake and by the artery/liver standardized uptake values (SUV) ratios. Diagnosis of GCA was confirmed in 68 patients and excluded in 35 patients, which served as the controls. GCA patients were older (median age 75 vs. 68 years), and presented more often with ischaemic symptoms. The best discrimination between GCA patients and controls was achieved for PET/CT findings within the supra-aortic arteries (sensitivity 0.71, specificity 0.91 for a SUV/LE cut-off value of 1.0). Specificity of PET/CT for the aorta and the iliofemoral arteries was lower (<0.34). Visual scoring correlated poorly to SUV measurements (Kendall Tau-b 0.13–0.55) and had a lower diagnostic accuracy (sensitivity 0.77, specificity 0.75). Prednisone treatment for ≥10 days significantly reduced PET/CT sensitivity (P = 0.009). Conclusion SUV based analysis of PET/CT enhances diagnostic accuracy with best discrimination in the supra-aortic region, particularly in steroid naïve patients. For discrimination based on the aorta and the iliofemoral region, higher cut-off values have to be applied, resulting in lower sensitivities for diagnosing GCA. large vessel vasculitis, giant cell arteritis, positron emission tomography, standardized diagnostic, prednisone Introduction Giant cell arteritis (GCA) is the most frequent of the primary vasculitidies. GCA manifests with constitutional and/or ischaemic symptoms.1,2 Diagnosis can be challenging, especially if typical ischaemic symptoms are absent and temporal artery biopsy shows no vasculitis. Consequently, several imaging modalities, such as positron emission tomography (PET)3,4 magnetic resonance angiography, and CT angiography have been introduced in clinical practice to aid diagnosis.5 Diagnostic accuracy of these techniques is difficult to assess, resulting in some uncertainty how to interpret imaging findings. Once GCA is suspected, therapy has to be initiated without delay to avoid ischaemic complications (e.g. vision loss), which can occur in around 10–20% of patients.6–9 Steroid treatment may interfere with diagnostic techniques and thus further complicate diagnosis. Positron emission tomography (PET) with [18 F] fluorodeoxyglucose (FDG) is the only imaging technique that allows visualization of metabolic activity by detecting FDG accumulation in cells. Increased tracer uptake is typically seen in inflammatory or neoplastic tissue, and is mainly due to increased expression of glucose transporters.10,11 PET is nowadays almost exclusively combined with CT (PET/CT),12 enabling for the precise correlation of tracer uptake and anatomical structures. PET/CT thereby allows visualizing inflammation in the vessel wall of GCA patients. Visual semiquantitative diagnostic scoring systems, using liver FDG uptake as cut-off for vasculitis have been proposed as a possible diagnostic standard in GCA.13,14 However, reports of significant inter-observer variability as well as a high rate of indeterminate findings in visual scoring systems have called the need for standardization. Consequently, measurement of standardized uptake values (SUV) has been suggested to overcome the subjectivity of the visual scoring systems.15 On the basis of vessel SUV measurements, a significant distinction between GCA patients and control populations could be achieved,15 yet a large overlap between groups remains. Hence, it is still under debate (i) which diagnostic criteria are the most appropriate to apply for the best diagnostic yield in PET/CT examinations, and (ii) how factors such as the vascular region or treatment affect the discriminative capacity of this method in suspected GCA.16,17 We herein aimed to close this gap, by comparing the diagnostic yield of FDG-PET/CT for diagnosis of GCA in a real-life setting of a vasculitis university clinic by analysing consecutive patients having received PET/CT for suspicion of GCA. Methods This study was approved by the local ethics committee. We included patients that presented with suspicion of having GCA and received PET/CT for diagnostic work-up at the University Hospital Basel between December 2006 and August 2012. The final diagnosis of GCA was made either (i) if temporal artery (TA) biopsy was positive, (ii) if patients fulfilled the 1989 ACR criteria, or (iii) if they fulfilled at least 2/5 ACR criteria in combination with typical ‘vasculitic’ ultrasound findings18 or vasculitic findings in other imaging methods.19 Data on clinical presentation, laboratory results, and treatment were retrieved from our prospective local GCA cohort and from retrospective chart review. PET/CT image acquisition All patients underwent FDG PET/CT scanning either at the University Hospital Basel (GE Discovery STE 16, 74 patients) or at the St. Claraspital Basel (Siemens Biograph 40 TruePoint HD, 29 patients). Patients were fasting for at least 6 h before tracer injection. Scans were obtained 1 h after intravenous injection of 5 MBq FDG/kg body weight at glycaemic levels below 10 mmol/L. First, a non-contrast low-dose CT scan was performed for attenuation correction and anatomical correlation. Next, a whole-body PET scan was obtained in 3D mode with iterative ordered subset expectation maximum (OSEM) reconstruction and a 128 × 128 matrix. At the University Hospital, reconstruction was performed using 2 iterations/28 subsets and a 6 mm FWHM Gaussian filter. At the St. Claraspital, the reconstruction parameters were 4 iterations/8 subsets and 5 mm Gaussian filtering. PET/CT analysis All 103 PET/CT scans were read by an experienced nuclear medicine specialist (C.R.), blinded for the diagnosis and treatment regimens of the patients. For interobserver agreement calculation the first 71 of the 103 PET/CT scans were additionally read by a second blinded investigator (E.S.). For the SUV analyses, the following arteries were systematically rated on PET/CT images: vertebral, carotid, subclavian, and axillary arteries (‘supra-aortic region’), thoracic and abdominal aorta, and iliac and common femoral arteries (‘iliofemoral region’). The area with the highest FDG uptake was visually chosen for each vessel on the PET images. The maximum SUVs (SUVmax) were determined in a spherical volume of interest (VOI), containing the artery wall and the lumen of the vessel on axial PET/CT slices within the previously assessed area (Figure 1). For the calculation of SUV ratios, mean SUVs were determined the same way, using a predefined (20 mL) spherical VOI in the liver as individual reference region. The reference region was carefully selected to exclude morphological abnormalities, as e.g. cysts, as well as from adjacent tissue on the CT. The exact localization of the VOIs was confirmed by the co-registered CT images. In case of absent tracer uptake and insufficient direct visualization of the vertebral artery on non-contrast enhanced CT images (see Supplementary data online, Figure S1). VOI measurement of the vertebral arteries was performed in the intervertebral foramen of the third cervical vertebral body. The PET/CT analysis was performed with Siemens SyngoVia software. Figure 1 View largeDownload slide Example of a VOI for SUV measurement, (A) VOI in the thoracic aorta in axial orientation, (B) in sagittal orientation, and (C) localization of the VOI in the maximum intensive projection (MIP) image are displayed. Figure 1 View largeDownload slide Example of a VOI for SUV measurement, (A) VOI in the thoracic aorta in axial orientation, (B) in sagittal orientation, and (C) localization of the VOI in the maximum intensive projection (MIP) image are displayed. Additionally, PET/CTs, were scored on a four-point visual analogue scale (VAS) of vascular tracer uptake, as previously described by Walter et al.:13 Grade 0 = no uptake, Grade I = uptake present but lower than liver uptake, Grade II = similar to liver uptake and Grade III higher than liver uptake. Vessels with Grade II or III were judged as ‘vasculitis’. To describe the influence of the duration of steroid therapy on diagnostic accuracy of PET/CT we divided patients in four different groups reflecting typical clinical scenarios: no steroids before PET/CT, PET/CT after a very short steroid treatment of 1–3 days, PET after short steroid treatment of 4–10 days, and PET/CT after established steroid therapy >10 days. Statistics Continuous variables were analysed using the Mann–Whitney U test and Kruskal–Wallis H test, and are expressed as medians and interquartile ranges. Correlations between quantitative values were performed using Spearman’s rank order test or Kendall Tau-b as appropriate. Categorical variables were compared using the χ2 test. Receiver operating characteristics (ROC) analyses were applied to each vascular territory to calculate sensitivity and specificity, and subsequent area under the curve (AUC) comparison was performed by Hanley and McNeil analysis.20 Cohen’s kappa statistics was used for calculating inter-observer agreement. P-values <0.05 were considered significant, all hypothesis testing was two-tailed for all statistical analyses. Statistical analysis was performed using SPSS (version 22.0, IBM Corp., Armonk, NY, USA) and R (Version 3.3.0, R Development Core Team, Vienna, Austria). Results Patients’ characteristics The final diagnosis of GCA was established in 68 of the 103 included patients. In 50 of the 62 patients having received temporal artery biopsies, the histology was diagnostic. Of the remaining, 13 full-filled at least 3/5 ACR criteria, and 5 patients fulfilled 2/5 ACR criteria, but had clear confirmation of vasculitis in cranial or extracranial vascular ultrasound (n = 4).18,21 or in magnetic resonance imaging and angio-CT (n = 1). In 35 patients, the diagnosis of GCA was refuted. Final diagnoses in this group were: polymyalgia (n = 14), non-vasculitic (i.e. arteriosclerotic) anterior ischaemic optic neuropathy (n = 4), myelodysplastic syndrome (n = 3), infection (n = 2), oligoarthritis (n = 2), interstitial lung disease (n = 2), coronary heart disease (n = 2), amaurosis fugax due to arteriosclerotic disease (n = 2), limb restricted vasculitis (n = 1), periodic fever syndrome (n = 1), systemic inflammation secondary to aortic dissection (n = 1) and trigeminal neuralgia (n = 1). These patients served as controls and are thereafter termed ‘non-GCA’. The GCA patients had a higher median age than non-GCA patients (75 vs. 68 years, P = 0.02), but were not significantly different in respect to gender distribution (P = 0.63) or systemic inflammation markers (C-reactive protein, P = 0.72; erythrocyte sedimentation rate, P = 0.67). Cranial ischaemic symptoms were more often present in the GCA group (Table 1). Table 1 Patients characteristics: data are expressed as number (%) or median (interquartile range) GCA (n = 68) Non-GCA (n = 35) P-value Gender (female) 44 (65%) 21(60%) 0.671 Median age (years) at PET scan 75 (65.5–80) 68 (61–76) 0.021 Amaurosis fugax 5 (7%) 1 (2%) 0.661 Loss of vision 10 (14%) 4 (11%) 0.768 New onset headache 29 (42%) 2 (6%) 0.001 Jaw claudication 26 (38%) 1 (2%) <0.001 Scalp tenderness 26 (38%) 3 (8%) 0.001 Pathological temporal artery 18 (26%) 1 (2%) 0.003 Proximal muscle pain 26 (38%) 16 (46%) 0.528 Fever 9 (13%) 4 (11%) 0.794 Erythrocyte sedimentation rate 68 mm/h (34–82) 59 mm/h (28–94) 0.722 C-reactive-protein 39.9 mg/L (10–71) 40.2 mg/L (8.3–100.7) 0.668 GCA (n = 68) Non-GCA (n = 35) P-value Gender (female) 44 (65%) 21(60%) 0.671 Median age (years) at PET scan 75 (65.5–80) 68 (61–76) 0.021 Amaurosis fugax 5 (7%) 1 (2%) 0.661 Loss of vision 10 (14%) 4 (11%) 0.768 New onset headache 29 (42%) 2 (6%) 0.001 Jaw claudication 26 (38%) 1 (2%) <0.001 Scalp tenderness 26 (38%) 3 (8%) 0.001 Pathological temporal artery 18 (26%) 1 (2%) 0.003 Proximal muscle pain 26 (38%) 16 (46%) 0.528 Fever 9 (13%) 4 (11%) 0.794 Erythrocyte sedimentation rate 68 mm/h (34–82) 59 mm/h (28–94) 0.722 C-reactive-protein 39.9 mg/L (10–71) 40.2 mg/L (8.3–100.7) 0.668 Table 1 Patients characteristics: data are expressed as number (%) or median (interquartile range) GCA (n = 68) Non-GCA (n = 35) P-value Gender (female) 44 (65%) 21(60%) 0.671 Median age (years) at PET scan 75 (65.5–80) 68 (61–76) 0.021 Amaurosis fugax 5 (7%) 1 (2%) 0.661 Loss of vision 10 (14%) 4 (11%) 0.768 New onset headache 29 (42%) 2 (6%) 0.001 Jaw claudication 26 (38%) 1 (2%) <0.001 Scalp tenderness 26 (38%) 3 (8%) 0.001 Pathological temporal artery 18 (26%) 1 (2%) 0.003 Proximal muscle pain 26 (38%) 16 (46%) 0.528 Fever 9 (13%) 4 (11%) 0.794 Erythrocyte sedimentation rate 68 mm/h (34–82) 59 mm/h (28–94) 0.722 C-reactive-protein 39.9 mg/L (10–71) 40.2 mg/L (8.3–100.7) 0.668 GCA (n = 68) Non-GCA (n = 35) P-value Gender (female) 44 (65%) 21(60%) 0.671 Median age (years) at PET scan 75 (65.5–80) 68 (61–76) 0.021 Amaurosis fugax 5 (7%) 1 (2%) 0.661 Loss of vision 10 (14%) 4 (11%) 0.768 New onset headache 29 (42%) 2 (6%) 0.001 Jaw claudication 26 (38%) 1 (2%) <0.001 Scalp tenderness 26 (38%) 3 (8%) 0.001 Pathological temporal artery 18 (26%) 1 (2%) 0.003 Proximal muscle pain 26 (38%) 16 (46%) 0.528 Fever 9 (13%) 4 (11%) 0.794 Erythrocyte sedimentation rate 68 mm/h (34–82) 59 mm/h (28–94) 0.722 C-reactive-protein 39.9 mg/L (10–71) 40.2 mg/L (8.3–100.7) 0.668 63 of the 68 GCA patients underwent PET/CT at initial disease manifestation. In the other five cases, PET/CT was performed at a later time point for suspected relapse during follow-up. The latter were on a stable median steroid dose of 5 mg/day (interquartile range 5–15 mg) at the time of PET/CT scanning. Of the 63 patients who underwent PET/CT at initial diagnosis, 29 patients were steroid-naïve at the time of PET/CT scanning. The remaining 34 (54%) patients were on a median dose of 60 mg (IQR 50–60 mg) prednisone for a median of 6 (IQR 3–11) days before PET/CT. Amongst the non-GCA patients 15 of 35 (43%) took a median of 27.5 mg (IQR 5–60 mg) prednisone for a median of 13 (IQR 5–96) days before PET/CT, whilst the other 20 were steroid-naïve. SUV/LE ratio in the supra-aortic arteries best discriminates GCA from non-GCA The ratios of SUV of arteries divided by mean liver values (SUV/LE) was calculated for the four defined vessel regions (supra-aortic, thoracic aorta, abdominal aorta, and iliofemoral arteries) and compared between GCA-patients and non-GCA patients. In the supra-aortic region, SUV/LE clearly distinguished GCA from non-GCA patients with only three of the 35 patients of the non-GCA group having a SUV/LE ratio > 1. In contrast, there was a considerable overlap between the two groups in the thoracic aorta, the abdominal aorta as well as in the iliofemoral region, with the SUV/LE ratio frequently exceeding 1 also in the non-GCA group (Table 2). Table 2 Mean ± SD of raw SUV measurements and ratios adjusted to the mean liver uptake GCA Non-GCA P-value SUVmean Liver 2.10 ± 0.34 2.17 ± 0.41 0.295 SUVmax Liver 2.57 ± 0.39 2.73 ± 0.47 0.057 SUVmax supra-aortic region 2.73 ± 0.93 1.92 ± 0.34 <0.001 SUVmax thoracic aorta 2.51 ± 0.71 2.22 ± 0.45 0.048 SUVmax abdominal aorta 2.58 ± 0.75 2.11 ± 0.47 0.001 SUVmax iliofemoral region 2.22 ± 0.70 1.89 ± 0.40 0.010 SUVmax supra-aortic region/SUVmean liver 1.33 ± 0.50 0.90 ± 0.14 <0.001 SUVmax thoracic aorta/SUVmean liver 1.22 ± 0.38 1.04 ± 0.16 0.007 SUVmax abdominal aorta/SUVmean liver 1.25 ± 0.41 0.98 ± 0.19 <0.001 SUVmax iliofemoral region/SUVmean liver 1.08 ± 0.37 0.89 ± 0.18 0.001 GCA Non-GCA P-value SUVmean Liver 2.10 ± 0.34 2.17 ± 0.41 0.295 SUVmax Liver 2.57 ± 0.39 2.73 ± 0.47 0.057 SUVmax supra-aortic region 2.73 ± 0.93 1.92 ± 0.34 <0.001 SUVmax thoracic aorta 2.51 ± 0.71 2.22 ± 0.45 0.048 SUVmax abdominal aorta 2.58 ± 0.75 2.11 ± 0.47 0.001 SUVmax iliofemoral region 2.22 ± 0.70 1.89 ± 0.40 0.010 SUVmax supra-aortic region/SUVmean liver 1.33 ± 0.50 0.90 ± 0.14 <0.001 SUVmax thoracic aorta/SUVmean liver 1.22 ± 0.38 1.04 ± 0.16 0.007 SUVmax abdominal aorta/SUVmean liver 1.25 ± 0.41 0.98 ± 0.19 <0.001 SUVmax iliofemoral region/SUVmean liver 1.08 ± 0.37 0.89 ± 0.18 0.001 Table 2 Mean ± SD of raw SUV measurements and ratios adjusted to the mean liver uptake GCA Non-GCA P-value SUVmean Liver 2.10 ± 0.34 2.17 ± 0.41 0.295 SUVmax Liver 2.57 ± 0.39 2.73 ± 0.47 0.057 SUVmax supra-aortic region 2.73 ± 0.93 1.92 ± 0.34 <0.001 SUVmax thoracic aorta 2.51 ± 0.71 2.22 ± 0.45 0.048 SUVmax abdominal aorta 2.58 ± 0.75 2.11 ± 0.47 0.001 SUVmax iliofemoral region 2.22 ± 0.70 1.89 ± 0.40 0.010 SUVmax supra-aortic region/SUVmean liver 1.33 ± 0.50 0.90 ± 0.14 <0.001 SUVmax thoracic aorta/SUVmean liver 1.22 ± 0.38 1.04 ± 0.16 0.007 SUVmax abdominal aorta/SUVmean liver 1.25 ± 0.41 0.98 ± 0.19 <0.001 SUVmax iliofemoral region/SUVmean liver 1.08 ± 0.37 0.89 ± 0.18 0.001 GCA Non-GCA P-value SUVmean Liver 2.10 ± 0.34 2.17 ± 0.41 0.295 SUVmax Liver 2.57 ± 0.39 2.73 ± 0.47 0.057 SUVmax supra-aortic region 2.73 ± 0.93 1.92 ± 0.34 <0.001 SUVmax thoracic aorta 2.51 ± 0.71 2.22 ± 0.45 0.048 SUVmax abdominal aorta 2.58 ± 0.75 2.11 ± 0.47 0.001 SUVmax iliofemoral region 2.22 ± 0.70 1.89 ± 0.40 0.010 SUVmax supra-aortic region/SUVmean liver 1.33 ± 0.50 0.90 ± 0.14 <0.001 SUVmax thoracic aorta/SUVmean liver 1.22 ± 0.38 1.04 ± 0.16 0.007 SUVmax abdominal aorta/SUVmean liver 1.25 ± 0.41 0.98 ± 0.19 <0.001 SUVmax iliofemoral region/SUVmean liver 1.08 ± 0.37 0.89 ± 0.18 0.001 ROC analysis shows a maximum AUC of 0.83 in the supra-aortic region, followed by significantly lower AUCs in the abdominal aorta (0.73, P = 0.04), the iliofemoral region (0.71, P = 0.02), and the thoracic aorta (0.66, P = 0.002) (Figure 2). For the supra-aortic region, a cut-off for the SUV/LE ratio of 1.0 yielded a diagnostic sensitivity of 0.71 and a specificity of 0.91. To achieve the same specificity of 0.91 for diagnosing GCA in the other less distinctive regions, a higher cut-off of 1.3 was required, resulting in a considerably lower sensitivity of 0.25 for the thoracic aorta and 0.34 for the abdominal aorta. For the iliofemoral region, a specificity of 0.91 is achieved with a cut-off of 1.14 resulting in a sensitivity of 0.26 (Figure 3). The combination of all vessel regions (maximal SUV/LE per patient) did not result in an improvement of the AUC (0.81) compared to the supra-aortic region alone (0.83). Figure 2 View largeDownload slide Distinct diagnostic performance of SUV/LE ratios in different vascular regions. ROC curves of the four different vessel regions. Figure 2 View largeDownload slide Distinct diagnostic performance of SUV/LE ratios in different vascular regions. ROC curves of the four different vessel regions. Figure 3 View largeDownload slide Distribution of SUV/LE for the four vessel regions; (A) supra-aortic, (B) thoracic aorta, (C) abdominal aorta, and (D) iliofemoral. The grey line indicates the three proposed cut-off values (solid = 1, dotted = 1.14, and dashed = 1.3, respectively). Blue bars = non-GCA patients, red bars = GCA patients, n = number of patients. Figure 3 View largeDownload slide Distribution of SUV/LE for the four vessel regions; (A) supra-aortic, (B) thoracic aorta, (C) abdominal aorta, and (D) iliofemoral. The grey line indicates the three proposed cut-off values (solid = 1, dotted = 1.14, and dashed = 1.3, respectively). Blue bars = non-GCA patients, red bars = GCA patients, n = number of patients. Visual grading correlates poorly with standardized FDG uptake ratio based grading The visual score was assessed and compared to the SUV/LE based analysis in all patients. The two scores correlated poorly for the individual vessel regions (Kendall Tau-b 0.13–0.55), with the exception of the vertebral artery showing a moderate correlation (Kendall Tau-b 0.56–0.61) (Figure 4). Accordingly, the overall diagnostic accuracy of visual scoring for all 4 regions combined showed a lower specificity of 0.77 compared to SUV/LE based analysis (0.86) with similar sensitivity (0.75 vs.0.72). This can be attributed to difficulties in visual discrimination of VAS scores 1–3, showing a considerable overlap of these scores in all vessel regions. SUV/LE ratios higher than 1.0 (expected to be classified as score 3, i.e. uptake higher than liver) were potentially underestimated using VAS interpretation, with 8.5% classified as score 0 (no uptake), 11.4% as score 1 (uptake lower than liver) and 72% of the measurements as score 2 (equal uptake as liver), although SUV/LE values in this group were as high as 1.66. Figure 4 View largeDownload slide Comparison of SUV/LE ratio and visual scoring of the four different vessel regions for all patients. The y-axis shows the visual scores (VAS) 0–3, the x-axis the SUV/LE values. (A) The supra-aortic region, (B) the thoracic aorta, (C) the abdominal aorta, and (D) the iliofemoral region. Gray shading indicates areas of overt disagreement between the scores. VAS, visual analogue scale (Basel score), n = number of patients. Figure 4 View largeDownload slide Comparison of SUV/LE ratio and visual scoring of the four different vessel regions for all patients. The y-axis shows the visual scores (VAS) 0–3, the x-axis the SUV/LE values. (A) The supra-aortic region, (B) the thoracic aorta, (C) the abdominal aorta, and (D) the iliofemoral region. Gray shading indicates areas of overt disagreement between the scores. VAS, visual analogue scale (Basel score), n = number of patients. Overall interobserver agreement of visual scores in between the two blinded readers was only moderate (Cohen’s kappa 0.49) and varied within the separate regions analysed from moderate (supra-aortic region 0.42, thoracic aorta 0.49, and iliofemoral region 0.44) to poor (abdominal aorta 0.17). Duration of steroid treatment affects the diagnostic accuracy of PET/CT Steroid treatment was already initiated in 52% of the cases [39/68 GCA patients (57%) and in 15/35 (43%) non-GCA patients, P = 0.21] before PET/CT imaging was available. To test the impact of treatment on the FDG uptake, we categorized patients according to the duration of steroid treatment before PET/CT study into four groups: (i) steroid-naïve patients (0 days) n = 49, (ii) 1–3 days, n = 16, (iii) 4–10 days, n = 15, or (iv) more than 10 days on prednisone treatment, n = 23 patients. In the patients without GCA, no difference in the SUV/LE ratios in between patients having received prednisone before PET/CT or steroid naïve could be detected. However, patients with GCA that received >10 days of steroids had significantly lower SUV/LE ratios for all four regions, compared to those who had no steroids before PET/CT (P < 0.01). For the abdominal aorta, the difference was already significant for the group having received 1–3 days of steroids compared to no steroid treatment (P = 0.04) (Figure 5). In the supra-aortic region, SUV/LE ratios performed very well in discriminating GCA from non-GCA if the treatment duration was less than 10 days. The AUC of the ROC analysis for patients not having received prednisone before PET/CT was significantly higher than if those on treatment were also included (0.94 vs. 0.86, P = 0.022, see Supplementary data online, Figure S2). Figure 5 View largeDownload slide Duration of steroid treatment before PET impacts SUV/LE values in GCA patients but not controls. Boxplots for SUV/LE in relation to steroid treatment duration for the four defined vascular regions. (0) = without steroid treatment, n = 49 patients, after 1–3 days of steroid treatment, n = 16 patients, after 4–10 days of steroid treatment, n = 15 patients, and >10 days of steroid treatment, n = 23 patients. Key statistics are indicated. Figure 5 View largeDownload slide Duration of steroid treatment before PET impacts SUV/LE values in GCA patients but not controls. Boxplots for SUV/LE in relation to steroid treatment duration for the four defined vascular regions. (0) = without steroid treatment, n = 49 patients, after 1–3 days of steroid treatment, n = 16 patients, after 4–10 days of steroid treatment, n = 15 patients, and >10 days of steroid treatment, n = 23 patients. Key statistics are indicated. Discussion Visualization of metabolically active lesions in arteries using PET/CT is a promising technique; yet, its validity to diagnose GCA is still under debate.22 The aim of this study was to assess diagnostic usefulness of PET/CT in patients presenting at a vasculitis clinic with suspicion of GCA and to establish criteria for interpreting imaging. The study population reflects clinical practice and current understanding of GCA as a disease entity beyond temporal arteritis. Consequently 5 out of 68 included GCA patients fulfilled only 2/5 ACR criteria, but all 5 had clear vasculitic findings in at least one (additional to PET/CT) imaging technique. This may be a limitation when comparing our data to previous studies with more restrictive inclusion criteria. Studying a large patient collective in a single vasculitis centre, we found that PET/CT has an excellent diagnostic accuracy, for diagnosing vasculitis in the supra-aortic branches. However, the specificity of PET/CT in the diagnosis of GCA was significantly lower for both, the aorta and the iliofemoral vessel region. Moreover, we found that the duration of glucocorticoid treatment prior to PET/CT reduces its diagnostic power. Different performance of PET/CT in distinct vascular regions has also been described by others.15,23 Specifically, supra-aortic vessels were more often judged as being PET positive than the aorta,23 which might reflect, in part, a pattern of GCA vessel involvement. Indeed, we and others found that by ultrasound GCA patients, more often showed typical vasculitic vessel wall thickening in the supra-aortic arteries compared to the iliofemoral region.18,24 A similar distribution of vasculitic findings has also been described by CT angiography with a higher incidence of supra-aortic than iliac and femoral artery involvement25 although the aorta was the most commonly involved blood vessel in CT. The aorta has also been reported to be frequently involved by visual qualitative scoring in 10 patients with GCA.26 In line with our findings, a reduced specificity of PET/CT for the diagnosis of GCA of the lower extremities has been reported in a small patient population.27 The reason for the overall higher FDG uptake in the femoral arteries is not known. Arteriosclerosis, which is more prevalent in the aorta and iliofemoral region, may play a role through increased FDG uptake in macrophages within the plaques.28 Although we have avoided measurements within visible calcified plaques in CT, soft plaques might have been missed in non-contrast enhanced CT examinations and thus contribute to higher SUV values in this vascular region. Discrepancies of the findings between the different imaging techniques remain unsolved, because histology—the gold-standard for vasculitis cannot be obtained for the involved vessels. However, our finding of a high number of non-GCA patients showing SUV/LE >1 within the aorta and the iliofemoral arteries indicates that we need to be aware of the risk of false positive GCA diagnoses and argues for adapted cut-off values for different vascular regions. Thus, based on our findings, we propose a SUV/LE cut-off for diagnosis of GCA of 1.0 for the supra-aortic region, but a higher ratio of 1.3 for the aorta and the iliofemoral region. Steroid treatment must be initiated upon suspicion of GCA with the effect that the diagnostic accuracy of PET/CT may be reduced. A decrease of FDG uptake in patients with GCA was reported after 3 months of treatment in longitudinal PET studies.23,29 Yet, others recently found that in a majority of patients FDG uptake may persist beyond 6 months. However, quantitative analyses were not performed.30 About half of the patients in our study received prednisone therapy before PET/CT, giving us the opportunity to study the impact of glucocorticoid treatment on PET findings in the different vascular regions. In general, SUV/LE ratios of patients with GCA were lower when patients received steroids before the PET/CT. This significantly impacted the performance of the PET/CT as early as 3 days after treatment initiation in the abdominal aorta, and 10 days in the remaining examined arteries. However, these data have to be interpreted with caution, because our study was not designed as a longitudinal PET/CT study and the decision to start prednisone before PET/CT may result in a bias through selection of specific patients (e.g. patients with clear ischaemic symptoms). Yet, our data strongly argue that PET/CT should be performed immediately, and—if the clinical setting permits—before initiation of the glucocorticoid therapy. When comparing SUV/LE ratios and visual scoring to diagnose GCA on PET/CT, we found that SUV analysis outperforms visual semi-quantitative scoring. The visual score correlated poorly with objective, ratio based SUV analysis, and showed poor interobserver correlation. Thus the risk of misclassification and subjective interpretation by visual scoring seems to be higher than previously assumed. In addition, visual interpretation is prone to perception phenomena such as simultaneous contrast and assimilation. Taken together, our data suggests that visual grading should not be used as sole read-out when interpreting PET imaging in the setting of suspected GCA. The overall specificity of PET/CT for the diagnosis of GCA in our cohort was comparable to other studies with smaller patient numbers that used visual or SUV based scoring.16 However, sensitivity of PET/CT for diagnosing GCA in our study was lower than reported in previous studies with sensitivities ranging from 0.77 to 1.0.31,32 This may in part be due to an already established steroid treatment in about half of our patients at the time of PET/CT. In the majority of the previous studies, patients were excluded when having received steroids for longer than 3 days prior to PET/CT.15 Moreover, our ‘control group’ which consisted of patients with inflammatory syndromes, and thereby may differ from non-inflammatory controls like cancer patients or healthy controls which have been used in other studies. 15 Conclusion In conclusion, we found that a cut-off value of 1.0 for SUV/LE ratios for the supra-aortic region performed very well in the diagnostic work-up for suspected GCA. In the aorta and iliofemoral regions, a higher cut-off value of 1.3 needs to be applied to minimize false positive diagnoses. PET/CT presumably performs best before or within the first few days of starting steroid treatment. Analysis of SUV/LE ratios should be included in the interpretation of the PET/CT studies, whereas exclusive visual analysis should be avoided. Supplementary data Supplementary data are available at European Heart Journal - Cardiovascular Imaging online. Acknowledgements We thank the physicians of the rheumatology and medical outpatient clinic for their help in the patient care and the clinical trial unit of the Department of clinical research of the university hospital Basel for supporting the local cohort of GCA patients. Funding C.T.B. is supported by the Swiss National Science Foundation (SNSF) [PZ00P3-148000]. Conflict of interest: None declared. References 1 Berger CT , Wolbers M , Meyer P , Daikeler T , Hess C. High incidence of severe ischaemic complications in patients with giant cell arteritis irrespective of platelet count and size, and platelet inhibition . Rheumatol (Oxford) 2009 ; 48 : 258 – 61 . Google Scholar CrossRef Search ADS 2 Haering M , Holbro A , Todorova MG , Aschwanden M , Kesten F , Berger CT et al. Incidence and prognostic implications of diplopia in patients with giant cell arteritis . J Rheumatol 2014 ; 41 : 1562 – 4 . Google Scholar CrossRef Search ADS PubMed 3 Meller J , Strutz F , Siefker U , Scheel A , Sahlmann CO , Lehmann K et al. Early diagnosis and follow-up of aortitis with [(18)F]FDG PET and MRI . Eur J Nucl Med Mol Imaging 2003 ; 30 : 730 – 6 . Google Scholar CrossRef Search ADS PubMed 4 Moosig F , Czech N , Mehl C , Henze E , Zeuner RA , Kneba M,. U A. Correlation between 18-fluorodeoxyglucose accumulation in large vessels and serological markers of inflammation in polymyalgia rheumatica: a quantitative PET study . Ann Rheum Dis 2004 ; 63 : 870 – 3 . Google Scholar CrossRef Search ADS PubMed 5 Lariviere D , Benali K , Coustet B , Pasi N , Hyafil F , Klein I et al. Positron emission tomography and computed tomography angiography for the diagnosis of giant cell arteritis: a real-life prospective study . Medicine (Baltimore) 2016 ; 95 : e4146 . Google Scholar CrossRef Search ADS PubMed 6 Liozon E , Ly K-H , Robert P-Y. [Ocular complications of giant cell arteritis] . Rev Med Interne 2013 ; 34 : 421 – 30 . Google Scholar CrossRef Search ADS PubMed 7 Aiello PD , Trautmann JC , McPhee TJ , Kunselman AR , Hunder GG. Visual prognosis in giant cell arteritis . Ophthalmology 1993 ; 100 : 550 – 5 . Google Scholar CrossRef Search ADS PubMed 8 González-Gay MA , García-Porrúa C , Llorca J , Hajeer AH , Brañas F , Dababneh A et al. Visual manifestations of giant cell arteritis. Trends and clinical spectrum in 161 patients . Medicine (Baltimore) 2000 ; 79 : 283 – 92 . Google Scholar CrossRef Search ADS PubMed 9 Salvarani C , Cimino L , Macchioni P , Consonni D , Cantini F , Bajocchi G et al. Risk factors for visual loss in an Italian population-based cohort of patients with giant cell arteritis . Arthritis Rheum 2005 ; 53 : 293 – 7 . Google Scholar CrossRef Search ADS PubMed 10 Younes M , Lechago LV , Somoano JR , Mosharaf M , Lechago J. Wide expression of the human erythrocyte glucose transporter Glut1 in human cancers . Cancer Res 1996 ; 56 : 1164 – 7 . Google Scholar PubMed 11 Zhao S , Kuge Y , Tsukamoto E , Mochizuki T , Kato T , Hikosaka K et al. Fluorodeoxyglucose uptake and glucose transporter expression in experimental inflammatory lesions and malignant tumours: effects of insulin and glucose loading . Nucl Med Commun 2002 ; 23 : 545 – 50 . Google Scholar CrossRef Search ADS PubMed 12 Slomka PJ , Pan T , Germano G. Recent advances and future progress in PET instrumentation . Semin Nucl Med 2016 ; 46 : 5 – 19 . Google Scholar CrossRef Search ADS PubMed 13 Walter MA , Melzer RA , Schindler C , Müller-Brand J , Tyndall A , Nitzsche EU. The value of [18F]FDG-PET in the diagnosis of large-vessel vasculitis and the assessment of activity and extent of disease . Eur J Nucl Med Mol Imaging 2005 ; 32 : 674 – 81 . Google Scholar CrossRef Search ADS PubMed 14 Besson FL , Parienti J-J , Bienvenu B , Prior JO , Costo S , Bouvard G et al. Diagnostic performance of 18F-fluorodeoxyglucose positron emission tomography in giant cell arteritis: a systematic review and meta-analysis . Eur J Nucl Med Mol Imaging 2011 ; 38 : 1764 – 72 . Google Scholar CrossRef Search ADS PubMed 15 Prieto-González S , Depetris M , García-Martínez A , Espígol-Frigolé G , Tavera-Bahillo I , Corbera-Bellata M et al. Positron emission tomography assessment of large vessel inflammation in patients with newly diagnosed, biopsy-proven giant cell arteritis: a prospective, case–control study . Ann Rheum Dis 2014 ; 73 : 1388 – 92 . Google Scholar CrossRef Search ADS PubMed 16 Soussan M , Nicolas P , Schramm C , Katsahian S , Pop G , Fain O et al. Management of large-vessel vasculitis with FDG-PET: a systematic literature review and meta-analysis . Medicine (Baltimore) 2015 ; 94 : e622. Google Scholar CrossRef Search ADS PubMed 17 Stellingwerff MD , Brouwer E , Lensen K-JDF , Rutgers A , Arends S , van der Geest KSM et al. Different scoring methods of FDG PET/CT in giant cell arteritis: need for standardization . Medicine (Baltimore) 2015 ; 94 : e1542 . Google Scholar CrossRef Search ADS PubMed 18 Aschwanden M , Kesten F , Stern M , Thalhammer C , Walker UA , Tyndall A et al. Vascular involvement in patients with giant cell arteritis determined by duplex sonography of 2x11 arterial regions . Ann Rheum Dis 2010 ; 69 : 1356 – 9 . Google Scholar CrossRef Search ADS PubMed 19 Stone JH , Tuckwell K , Dimonaco S , Klearman M , Aringer M , Blockmans D et al. Trial of tocilizumab in giant-cell arteritis . N Engl J Med 2017 ; 377 : 317 – 28 . Google Scholar CrossRef Search ADS PubMed 20 Hanley JA , McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve . Radiology 1982 ; 143 : 29 – 36 . Google Scholar CrossRef Search ADS PubMed 21 Diamantopoulos AP , Haugeberg G , Hetland H , Soldal DM , Bie R , Myklebust G. Diagnostic value of color doppler ultrasonography of temporal arteries and large vessels in giant cell arteritis: a consecutive case series . Arthritis Care Res 2014 ; 66 : 113 – 9 . Google Scholar CrossRef Search ADS 22 Weyand CM , Goronzy JJ. Clinical practice. Giant-cell arteritis and polymyalgia rheumatica . N Engl J Med 2014 ; 371 : 50 – 7 . Google Scholar CrossRef Search ADS PubMed 23 Blockmans D , Ceuninck L. D , Vanderschueren S , Knockaert D , Mortelmans L , Bobbaers H. Repetitive 18F-fluorodeoxyglucose positron emission tomography in giant cell arteritis: a prospective study of 35 patients . Arthritis Care Res 2006 ; 55 : 131 – 7 . Google Scholar CrossRef Search ADS 24 Assie C , Janvresse A , Plissonnier D , Levesque H , Marie I. Long-term follow-up of upper and lower extremity vasculitis related to giant cell arteritis: a series of 36 patients . Medicine (Baltimore) 2011 ; 90 : 40 – 51 . Google Scholar CrossRef Search ADS PubMed 25 Prieto-González S , Arguis P , García-Martínez A , Espígol-Frigolé G , Tavera-Bahillo I , Butjosa M et al. Large vessel involvement in biopsy-proven giant cell arteritis: prospective study in 40 newly diagnosed patients using CT angiography . Ann Rheum Dis 2012 ; 71 : 1170 – 6 . Google Scholar CrossRef Search ADS PubMed 26 Henes JC , Müller M , Krieger J , Balletshofer B , Pfannenberg AC , Kanz L et al. [18F] FDG-PET/CT as a new and sensitive imaging method for the diagnosis of large vessel vasculitis . Clin Exp Rheumatol 2008 ; 26 (3 Suppl 49): S47 – 52 . [WorldCat] Google Scholar PubMed 27 Förster S , Tato F , Weiss M , Czihal M , Rominger A , Bartenstein P et al. Patterns of extracranial involvement in newly diagnosed giant cell arteritis assessed by physical examination, colour coded duplex sonography and FDG-PET . Vasa 2011 ; 40 : 219 – 27 . Google Scholar CrossRef Search ADS PubMed 28 Yun M , Jang S , Cucchiara A , Newberg AB , Alavi A. 18F FDG uptake in the large arteries: a correlation study with the atherogenic risk factors . Semin Nucl Med 2002 ; 32 : 70 – 6 . Google Scholar CrossRef Search ADS PubMed 29 Papathanasiou ND , Du Y , Menezes LJ , Almuhaideb A , Shastry M , Beynon H et al. 18F-Fludeoxyglucose PET/CT in the evaluation of large-vessel vasculitis: diagnostic performance and correlation with clinical and laboratory parameters . Br J Radiol 2012 ; 85 : e188 – 94 . Google Scholar CrossRef Search ADS PubMed 30 de Boysson H , Aide N , Liozon E , Lambert M , Parienti J-J , Monteil J et al. Repetitive (18)F-FDG-PET/CT in patients with large-vessel giant-cell arteritis and controlled disease . Eur J Intern Med 2017 ; doi:10.1016/j.ejim.2017.08.013. 31 Meller J , Sahlmann CO , Gürocak O , Liersch T , Meller B. FDG-PET in patients with fever of unknown origin: the importance of diagnosing large vessel vasculitis . Q J Nucl Med Mol Imaging 2009 ; 53 : 51 – 63 . Google Scholar PubMed 32 Hautzel H , Sander O , Heinzel A , Schneider M , Müller H-W. Assessment of large-vessel involvement in giant cell arteritis with 18F-FDG PET: introducing an ROC-analysis–based cutoff ratio . J Nucl Med 2008 ; 49 : 1107 – 13 . Google Scholar CrossRef Search ADS PubMed Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2017. For permissions, please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png European Heart Journal – Cardiovascular Imaging Oxford University Press

[18F]FDG positron emission tomography in patients presenting with suspicion of giant cell arteritis—lessons from a vasculitis clinic

Loading next page...
 
/lp/ou_press/18f-fdg-positron-emission-tomography-in-patients-presenting-with-2FQ4IPCi08
Publisher
Oxford University Press
Copyright
Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2017. For permissions, please email: journals.permissions@oup.com.
ISSN
2047-2404
D.O.I.
10.1093/ehjci/jex259
Publisher site
See Article on Publisher Site

Abstract

Abstract Aims The usefulness of [18F] fluorodeoxyglucose-positron emission tomography/computed tomography ([18F]FDG-PET/CT) for diagnosing giant cell arteritis (GCA) has been previously reported. Yet, the interpretation of PET scans is not clear-cut. The present study aimed at determining the best method to analyse PET/CT in a large, real-life cohort of patients presenting with suspicion of GCA. Methods and results One hundred and three patients with clinical suspicion of GCA undergoing PET/CT between 2006 and 2012 were included. Clinical data were retrieved from patients’ charts. PET/CT was categorized by visual scoring of the uptake and by the artery/liver standardized uptake values (SUV) ratios. Diagnosis of GCA was confirmed in 68 patients and excluded in 35 patients, which served as the controls. GCA patients were older (median age 75 vs. 68 years), and presented more often with ischaemic symptoms. The best discrimination between GCA patients and controls was achieved for PET/CT findings within the supra-aortic arteries (sensitivity 0.71, specificity 0.91 for a SUV/LE cut-off value of 1.0). Specificity of PET/CT for the aorta and the iliofemoral arteries was lower (<0.34). Visual scoring correlated poorly to SUV measurements (Kendall Tau-b 0.13–0.55) and had a lower diagnostic accuracy (sensitivity 0.77, specificity 0.75). Prednisone treatment for ≥10 days significantly reduced PET/CT sensitivity (P = 0.009). Conclusion SUV based analysis of PET/CT enhances diagnostic accuracy with best discrimination in the supra-aortic region, particularly in steroid naïve patients. For discrimination based on the aorta and the iliofemoral region, higher cut-off values have to be applied, resulting in lower sensitivities for diagnosing GCA. large vessel vasculitis, giant cell arteritis, positron emission tomography, standardized diagnostic, prednisone Introduction Giant cell arteritis (GCA) is the most frequent of the primary vasculitidies. GCA manifests with constitutional and/or ischaemic symptoms.1,2 Diagnosis can be challenging, especially if typical ischaemic symptoms are absent and temporal artery biopsy shows no vasculitis. Consequently, several imaging modalities, such as positron emission tomography (PET)3,4 magnetic resonance angiography, and CT angiography have been introduced in clinical practice to aid diagnosis.5 Diagnostic accuracy of these techniques is difficult to assess, resulting in some uncertainty how to interpret imaging findings. Once GCA is suspected, therapy has to be initiated without delay to avoid ischaemic complications (e.g. vision loss), which can occur in around 10–20% of patients.6–9 Steroid treatment may interfere with diagnostic techniques and thus further complicate diagnosis. Positron emission tomography (PET) with [18 F] fluorodeoxyglucose (FDG) is the only imaging technique that allows visualization of metabolic activity by detecting FDG accumulation in cells. Increased tracer uptake is typically seen in inflammatory or neoplastic tissue, and is mainly due to increased expression of glucose transporters.10,11 PET is nowadays almost exclusively combined with CT (PET/CT),12 enabling for the precise correlation of tracer uptake and anatomical structures. PET/CT thereby allows visualizing inflammation in the vessel wall of GCA patients. Visual semiquantitative diagnostic scoring systems, using liver FDG uptake as cut-off for vasculitis have been proposed as a possible diagnostic standard in GCA.13,14 However, reports of significant inter-observer variability as well as a high rate of indeterminate findings in visual scoring systems have called the need for standardization. Consequently, measurement of standardized uptake values (SUV) has been suggested to overcome the subjectivity of the visual scoring systems.15 On the basis of vessel SUV measurements, a significant distinction between GCA patients and control populations could be achieved,15 yet a large overlap between groups remains. Hence, it is still under debate (i) which diagnostic criteria are the most appropriate to apply for the best diagnostic yield in PET/CT examinations, and (ii) how factors such as the vascular region or treatment affect the discriminative capacity of this method in suspected GCA.16,17 We herein aimed to close this gap, by comparing the diagnostic yield of FDG-PET/CT for diagnosis of GCA in a real-life setting of a vasculitis university clinic by analysing consecutive patients having received PET/CT for suspicion of GCA. Methods This study was approved by the local ethics committee. We included patients that presented with suspicion of having GCA and received PET/CT for diagnostic work-up at the University Hospital Basel between December 2006 and August 2012. The final diagnosis of GCA was made either (i) if temporal artery (TA) biopsy was positive, (ii) if patients fulfilled the 1989 ACR criteria, or (iii) if they fulfilled at least 2/5 ACR criteria in combination with typical ‘vasculitic’ ultrasound findings18 or vasculitic findings in other imaging methods.19 Data on clinical presentation, laboratory results, and treatment were retrieved from our prospective local GCA cohort and from retrospective chart review. PET/CT image acquisition All patients underwent FDG PET/CT scanning either at the University Hospital Basel (GE Discovery STE 16, 74 patients) or at the St. Claraspital Basel (Siemens Biograph 40 TruePoint HD, 29 patients). Patients were fasting for at least 6 h before tracer injection. Scans were obtained 1 h after intravenous injection of 5 MBq FDG/kg body weight at glycaemic levels below 10 mmol/L. First, a non-contrast low-dose CT scan was performed for attenuation correction and anatomical correlation. Next, a whole-body PET scan was obtained in 3D mode with iterative ordered subset expectation maximum (OSEM) reconstruction and a 128 × 128 matrix. At the University Hospital, reconstruction was performed using 2 iterations/28 subsets and a 6 mm FWHM Gaussian filter. At the St. Claraspital, the reconstruction parameters were 4 iterations/8 subsets and 5 mm Gaussian filtering. PET/CT analysis All 103 PET/CT scans were read by an experienced nuclear medicine specialist (C.R.), blinded for the diagnosis and treatment regimens of the patients. For interobserver agreement calculation the first 71 of the 103 PET/CT scans were additionally read by a second blinded investigator (E.S.). For the SUV analyses, the following arteries were systematically rated on PET/CT images: vertebral, carotid, subclavian, and axillary arteries (‘supra-aortic region’), thoracic and abdominal aorta, and iliac and common femoral arteries (‘iliofemoral region’). The area with the highest FDG uptake was visually chosen for each vessel on the PET images. The maximum SUVs (SUVmax) were determined in a spherical volume of interest (VOI), containing the artery wall and the lumen of the vessel on axial PET/CT slices within the previously assessed area (Figure 1). For the calculation of SUV ratios, mean SUVs were determined the same way, using a predefined (20 mL) spherical VOI in the liver as individual reference region. The reference region was carefully selected to exclude morphological abnormalities, as e.g. cysts, as well as from adjacent tissue on the CT. The exact localization of the VOIs was confirmed by the co-registered CT images. In case of absent tracer uptake and insufficient direct visualization of the vertebral artery on non-contrast enhanced CT images (see Supplementary data online, Figure S1). VOI measurement of the vertebral arteries was performed in the intervertebral foramen of the third cervical vertebral body. The PET/CT analysis was performed with Siemens SyngoVia software. Figure 1 View largeDownload slide Example of a VOI for SUV measurement, (A) VOI in the thoracic aorta in axial orientation, (B) in sagittal orientation, and (C) localization of the VOI in the maximum intensive projection (MIP) image are displayed. Figure 1 View largeDownload slide Example of a VOI for SUV measurement, (A) VOI in the thoracic aorta in axial orientation, (B) in sagittal orientation, and (C) localization of the VOI in the maximum intensive projection (MIP) image are displayed. Additionally, PET/CTs, were scored on a four-point visual analogue scale (VAS) of vascular tracer uptake, as previously described by Walter et al.:13 Grade 0 = no uptake, Grade I = uptake present but lower than liver uptake, Grade II = similar to liver uptake and Grade III higher than liver uptake. Vessels with Grade II or III were judged as ‘vasculitis’. To describe the influence of the duration of steroid therapy on diagnostic accuracy of PET/CT we divided patients in four different groups reflecting typical clinical scenarios: no steroids before PET/CT, PET/CT after a very short steroid treatment of 1–3 days, PET after short steroid treatment of 4–10 days, and PET/CT after established steroid therapy >10 days. Statistics Continuous variables were analysed using the Mann–Whitney U test and Kruskal–Wallis H test, and are expressed as medians and interquartile ranges. Correlations between quantitative values were performed using Spearman’s rank order test or Kendall Tau-b as appropriate. Categorical variables were compared using the χ2 test. Receiver operating characteristics (ROC) analyses were applied to each vascular territory to calculate sensitivity and specificity, and subsequent area under the curve (AUC) comparison was performed by Hanley and McNeil analysis.20 Cohen’s kappa statistics was used for calculating inter-observer agreement. P-values <0.05 were considered significant, all hypothesis testing was two-tailed for all statistical analyses. Statistical analysis was performed using SPSS (version 22.0, IBM Corp., Armonk, NY, USA) and R (Version 3.3.0, R Development Core Team, Vienna, Austria). Results Patients’ characteristics The final diagnosis of GCA was established in 68 of the 103 included patients. In 50 of the 62 patients having received temporal artery biopsies, the histology was diagnostic. Of the remaining, 13 full-filled at least 3/5 ACR criteria, and 5 patients fulfilled 2/5 ACR criteria, but had clear confirmation of vasculitis in cranial or extracranial vascular ultrasound (n = 4).18,21 or in magnetic resonance imaging and angio-CT (n = 1). In 35 patients, the diagnosis of GCA was refuted. Final diagnoses in this group were: polymyalgia (n = 14), non-vasculitic (i.e. arteriosclerotic) anterior ischaemic optic neuropathy (n = 4), myelodysplastic syndrome (n = 3), infection (n = 2), oligoarthritis (n = 2), interstitial lung disease (n = 2), coronary heart disease (n = 2), amaurosis fugax due to arteriosclerotic disease (n = 2), limb restricted vasculitis (n = 1), periodic fever syndrome (n = 1), systemic inflammation secondary to aortic dissection (n = 1) and trigeminal neuralgia (n = 1). These patients served as controls and are thereafter termed ‘non-GCA’. The GCA patients had a higher median age than non-GCA patients (75 vs. 68 years, P = 0.02), but were not significantly different in respect to gender distribution (P = 0.63) or systemic inflammation markers (C-reactive protein, P = 0.72; erythrocyte sedimentation rate, P = 0.67). Cranial ischaemic symptoms were more often present in the GCA group (Table 1). Table 1 Patients characteristics: data are expressed as number (%) or median (interquartile range) GCA (n = 68) Non-GCA (n = 35) P-value Gender (female) 44 (65%) 21(60%) 0.671 Median age (years) at PET scan 75 (65.5–80) 68 (61–76) 0.021 Amaurosis fugax 5 (7%) 1 (2%) 0.661 Loss of vision 10 (14%) 4 (11%) 0.768 New onset headache 29 (42%) 2 (6%) 0.001 Jaw claudication 26 (38%) 1 (2%) <0.001 Scalp tenderness 26 (38%) 3 (8%) 0.001 Pathological temporal artery 18 (26%) 1 (2%) 0.003 Proximal muscle pain 26 (38%) 16 (46%) 0.528 Fever 9 (13%) 4 (11%) 0.794 Erythrocyte sedimentation rate 68 mm/h (34–82) 59 mm/h (28–94) 0.722 C-reactive-protein 39.9 mg/L (10–71) 40.2 mg/L (8.3–100.7) 0.668 GCA (n = 68) Non-GCA (n = 35) P-value Gender (female) 44 (65%) 21(60%) 0.671 Median age (years) at PET scan 75 (65.5–80) 68 (61–76) 0.021 Amaurosis fugax 5 (7%) 1 (2%) 0.661 Loss of vision 10 (14%) 4 (11%) 0.768 New onset headache 29 (42%) 2 (6%) 0.001 Jaw claudication 26 (38%) 1 (2%) <0.001 Scalp tenderness 26 (38%) 3 (8%) 0.001 Pathological temporal artery 18 (26%) 1 (2%) 0.003 Proximal muscle pain 26 (38%) 16 (46%) 0.528 Fever 9 (13%) 4 (11%) 0.794 Erythrocyte sedimentation rate 68 mm/h (34–82) 59 mm/h (28–94) 0.722 C-reactive-protein 39.9 mg/L (10–71) 40.2 mg/L (8.3–100.7) 0.668 Table 1 Patients characteristics: data are expressed as number (%) or median (interquartile range) GCA (n = 68) Non-GCA (n = 35) P-value Gender (female) 44 (65%) 21(60%) 0.671 Median age (years) at PET scan 75 (65.5–80) 68 (61–76) 0.021 Amaurosis fugax 5 (7%) 1 (2%) 0.661 Loss of vision 10 (14%) 4 (11%) 0.768 New onset headache 29 (42%) 2 (6%) 0.001 Jaw claudication 26 (38%) 1 (2%) <0.001 Scalp tenderness 26 (38%) 3 (8%) 0.001 Pathological temporal artery 18 (26%) 1 (2%) 0.003 Proximal muscle pain 26 (38%) 16 (46%) 0.528 Fever 9 (13%) 4 (11%) 0.794 Erythrocyte sedimentation rate 68 mm/h (34–82) 59 mm/h (28–94) 0.722 C-reactive-protein 39.9 mg/L (10–71) 40.2 mg/L (8.3–100.7) 0.668 GCA (n = 68) Non-GCA (n = 35) P-value Gender (female) 44 (65%) 21(60%) 0.671 Median age (years) at PET scan 75 (65.5–80) 68 (61–76) 0.021 Amaurosis fugax 5 (7%) 1 (2%) 0.661 Loss of vision 10 (14%) 4 (11%) 0.768 New onset headache 29 (42%) 2 (6%) 0.001 Jaw claudication 26 (38%) 1 (2%) <0.001 Scalp tenderness 26 (38%) 3 (8%) 0.001 Pathological temporal artery 18 (26%) 1 (2%) 0.003 Proximal muscle pain 26 (38%) 16 (46%) 0.528 Fever 9 (13%) 4 (11%) 0.794 Erythrocyte sedimentation rate 68 mm/h (34–82) 59 mm/h (28–94) 0.722 C-reactive-protein 39.9 mg/L (10–71) 40.2 mg/L (8.3–100.7) 0.668 63 of the 68 GCA patients underwent PET/CT at initial disease manifestation. In the other five cases, PET/CT was performed at a later time point for suspected relapse during follow-up. The latter were on a stable median steroid dose of 5 mg/day (interquartile range 5–15 mg) at the time of PET/CT scanning. Of the 63 patients who underwent PET/CT at initial diagnosis, 29 patients were steroid-naïve at the time of PET/CT scanning. The remaining 34 (54%) patients were on a median dose of 60 mg (IQR 50–60 mg) prednisone for a median of 6 (IQR 3–11) days before PET/CT. Amongst the non-GCA patients 15 of 35 (43%) took a median of 27.5 mg (IQR 5–60 mg) prednisone for a median of 13 (IQR 5–96) days before PET/CT, whilst the other 20 were steroid-naïve. SUV/LE ratio in the supra-aortic arteries best discriminates GCA from non-GCA The ratios of SUV of arteries divided by mean liver values (SUV/LE) was calculated for the four defined vessel regions (supra-aortic, thoracic aorta, abdominal aorta, and iliofemoral arteries) and compared between GCA-patients and non-GCA patients. In the supra-aortic region, SUV/LE clearly distinguished GCA from non-GCA patients with only three of the 35 patients of the non-GCA group having a SUV/LE ratio > 1. In contrast, there was a considerable overlap between the two groups in the thoracic aorta, the abdominal aorta as well as in the iliofemoral region, with the SUV/LE ratio frequently exceeding 1 also in the non-GCA group (Table 2). Table 2 Mean ± SD of raw SUV measurements and ratios adjusted to the mean liver uptake GCA Non-GCA P-value SUVmean Liver 2.10 ± 0.34 2.17 ± 0.41 0.295 SUVmax Liver 2.57 ± 0.39 2.73 ± 0.47 0.057 SUVmax supra-aortic region 2.73 ± 0.93 1.92 ± 0.34 <0.001 SUVmax thoracic aorta 2.51 ± 0.71 2.22 ± 0.45 0.048 SUVmax abdominal aorta 2.58 ± 0.75 2.11 ± 0.47 0.001 SUVmax iliofemoral region 2.22 ± 0.70 1.89 ± 0.40 0.010 SUVmax supra-aortic region/SUVmean liver 1.33 ± 0.50 0.90 ± 0.14 <0.001 SUVmax thoracic aorta/SUVmean liver 1.22 ± 0.38 1.04 ± 0.16 0.007 SUVmax abdominal aorta/SUVmean liver 1.25 ± 0.41 0.98 ± 0.19 <0.001 SUVmax iliofemoral region/SUVmean liver 1.08 ± 0.37 0.89 ± 0.18 0.001 GCA Non-GCA P-value SUVmean Liver 2.10 ± 0.34 2.17 ± 0.41 0.295 SUVmax Liver 2.57 ± 0.39 2.73 ± 0.47 0.057 SUVmax supra-aortic region 2.73 ± 0.93 1.92 ± 0.34 <0.001 SUVmax thoracic aorta 2.51 ± 0.71 2.22 ± 0.45 0.048 SUVmax abdominal aorta 2.58 ± 0.75 2.11 ± 0.47 0.001 SUVmax iliofemoral region 2.22 ± 0.70 1.89 ± 0.40 0.010 SUVmax supra-aortic region/SUVmean liver 1.33 ± 0.50 0.90 ± 0.14 <0.001 SUVmax thoracic aorta/SUVmean liver 1.22 ± 0.38 1.04 ± 0.16 0.007 SUVmax abdominal aorta/SUVmean liver 1.25 ± 0.41 0.98 ± 0.19 <0.001 SUVmax iliofemoral region/SUVmean liver 1.08 ± 0.37 0.89 ± 0.18 0.001 Table 2 Mean ± SD of raw SUV measurements and ratios adjusted to the mean liver uptake GCA Non-GCA P-value SUVmean Liver 2.10 ± 0.34 2.17 ± 0.41 0.295 SUVmax Liver 2.57 ± 0.39 2.73 ± 0.47 0.057 SUVmax supra-aortic region 2.73 ± 0.93 1.92 ± 0.34 <0.001 SUVmax thoracic aorta 2.51 ± 0.71 2.22 ± 0.45 0.048 SUVmax abdominal aorta 2.58 ± 0.75 2.11 ± 0.47 0.001 SUVmax iliofemoral region 2.22 ± 0.70 1.89 ± 0.40 0.010 SUVmax supra-aortic region/SUVmean liver 1.33 ± 0.50 0.90 ± 0.14 <0.001 SUVmax thoracic aorta/SUVmean liver 1.22 ± 0.38 1.04 ± 0.16 0.007 SUVmax abdominal aorta/SUVmean liver 1.25 ± 0.41 0.98 ± 0.19 <0.001 SUVmax iliofemoral region/SUVmean liver 1.08 ± 0.37 0.89 ± 0.18 0.001 GCA Non-GCA P-value SUVmean Liver 2.10 ± 0.34 2.17 ± 0.41 0.295 SUVmax Liver 2.57 ± 0.39 2.73 ± 0.47 0.057 SUVmax supra-aortic region 2.73 ± 0.93 1.92 ± 0.34 <0.001 SUVmax thoracic aorta 2.51 ± 0.71 2.22 ± 0.45 0.048 SUVmax abdominal aorta 2.58 ± 0.75 2.11 ± 0.47 0.001 SUVmax iliofemoral region 2.22 ± 0.70 1.89 ± 0.40 0.010 SUVmax supra-aortic region/SUVmean liver 1.33 ± 0.50 0.90 ± 0.14 <0.001 SUVmax thoracic aorta/SUVmean liver 1.22 ± 0.38 1.04 ± 0.16 0.007 SUVmax abdominal aorta/SUVmean liver 1.25 ± 0.41 0.98 ± 0.19 <0.001 SUVmax iliofemoral region/SUVmean liver 1.08 ± 0.37 0.89 ± 0.18 0.001 ROC analysis shows a maximum AUC of 0.83 in the supra-aortic region, followed by significantly lower AUCs in the abdominal aorta (0.73, P = 0.04), the iliofemoral region (0.71, P = 0.02), and the thoracic aorta (0.66, P = 0.002) (Figure 2). For the supra-aortic region, a cut-off for the SUV/LE ratio of 1.0 yielded a diagnostic sensitivity of 0.71 and a specificity of 0.91. To achieve the same specificity of 0.91 for diagnosing GCA in the other less distinctive regions, a higher cut-off of 1.3 was required, resulting in a considerably lower sensitivity of 0.25 for the thoracic aorta and 0.34 for the abdominal aorta. For the iliofemoral region, a specificity of 0.91 is achieved with a cut-off of 1.14 resulting in a sensitivity of 0.26 (Figure 3). The combination of all vessel regions (maximal SUV/LE per patient) did not result in an improvement of the AUC (0.81) compared to the supra-aortic region alone (0.83). Figure 2 View largeDownload slide Distinct diagnostic performance of SUV/LE ratios in different vascular regions. ROC curves of the four different vessel regions. Figure 2 View largeDownload slide Distinct diagnostic performance of SUV/LE ratios in different vascular regions. ROC curves of the four different vessel regions. Figure 3 View largeDownload slide Distribution of SUV/LE for the four vessel regions; (A) supra-aortic, (B) thoracic aorta, (C) abdominal aorta, and (D) iliofemoral. The grey line indicates the three proposed cut-off values (solid = 1, dotted = 1.14, and dashed = 1.3, respectively). Blue bars = non-GCA patients, red bars = GCA patients, n = number of patients. Figure 3 View largeDownload slide Distribution of SUV/LE for the four vessel regions; (A) supra-aortic, (B) thoracic aorta, (C) abdominal aorta, and (D) iliofemoral. The grey line indicates the three proposed cut-off values (solid = 1, dotted = 1.14, and dashed = 1.3, respectively). Blue bars = non-GCA patients, red bars = GCA patients, n = number of patients. Visual grading correlates poorly with standardized FDG uptake ratio based grading The visual score was assessed and compared to the SUV/LE based analysis in all patients. The two scores correlated poorly for the individual vessel regions (Kendall Tau-b 0.13–0.55), with the exception of the vertebral artery showing a moderate correlation (Kendall Tau-b 0.56–0.61) (Figure 4). Accordingly, the overall diagnostic accuracy of visual scoring for all 4 regions combined showed a lower specificity of 0.77 compared to SUV/LE based analysis (0.86) with similar sensitivity (0.75 vs.0.72). This can be attributed to difficulties in visual discrimination of VAS scores 1–3, showing a considerable overlap of these scores in all vessel regions. SUV/LE ratios higher than 1.0 (expected to be classified as score 3, i.e. uptake higher than liver) were potentially underestimated using VAS interpretation, with 8.5% classified as score 0 (no uptake), 11.4% as score 1 (uptake lower than liver) and 72% of the measurements as score 2 (equal uptake as liver), although SUV/LE values in this group were as high as 1.66. Figure 4 View largeDownload slide Comparison of SUV/LE ratio and visual scoring of the four different vessel regions for all patients. The y-axis shows the visual scores (VAS) 0–3, the x-axis the SUV/LE values. (A) The supra-aortic region, (B) the thoracic aorta, (C) the abdominal aorta, and (D) the iliofemoral region. Gray shading indicates areas of overt disagreement between the scores. VAS, visual analogue scale (Basel score), n = number of patients. Figure 4 View largeDownload slide Comparison of SUV/LE ratio and visual scoring of the four different vessel regions for all patients. The y-axis shows the visual scores (VAS) 0–3, the x-axis the SUV/LE values. (A) The supra-aortic region, (B) the thoracic aorta, (C) the abdominal aorta, and (D) the iliofemoral region. Gray shading indicates areas of overt disagreement between the scores. VAS, visual analogue scale (Basel score), n = number of patients. Overall interobserver agreement of visual scores in between the two blinded readers was only moderate (Cohen’s kappa 0.49) and varied within the separate regions analysed from moderate (supra-aortic region 0.42, thoracic aorta 0.49, and iliofemoral region 0.44) to poor (abdominal aorta 0.17). Duration of steroid treatment affects the diagnostic accuracy of PET/CT Steroid treatment was already initiated in 52% of the cases [39/68 GCA patients (57%) and in 15/35 (43%) non-GCA patients, P = 0.21] before PET/CT imaging was available. To test the impact of treatment on the FDG uptake, we categorized patients according to the duration of steroid treatment before PET/CT study into four groups: (i) steroid-naïve patients (0 days) n = 49, (ii) 1–3 days, n = 16, (iii) 4–10 days, n = 15, or (iv) more than 10 days on prednisone treatment, n = 23 patients. In the patients without GCA, no difference in the SUV/LE ratios in between patients having received prednisone before PET/CT or steroid naïve could be detected. However, patients with GCA that received >10 days of steroids had significantly lower SUV/LE ratios for all four regions, compared to those who had no steroids before PET/CT (P < 0.01). For the abdominal aorta, the difference was already significant for the group having received 1–3 days of steroids compared to no steroid treatment (P = 0.04) (Figure 5). In the supra-aortic region, SUV/LE ratios performed very well in discriminating GCA from non-GCA if the treatment duration was less than 10 days. The AUC of the ROC analysis for patients not having received prednisone before PET/CT was significantly higher than if those on treatment were also included (0.94 vs. 0.86, P = 0.022, see Supplementary data online, Figure S2). Figure 5 View largeDownload slide Duration of steroid treatment before PET impacts SUV/LE values in GCA patients but not controls. Boxplots for SUV/LE in relation to steroid treatment duration for the four defined vascular regions. (0) = without steroid treatment, n = 49 patients, after 1–3 days of steroid treatment, n = 16 patients, after 4–10 days of steroid treatment, n = 15 patients, and >10 days of steroid treatment, n = 23 patients. Key statistics are indicated. Figure 5 View largeDownload slide Duration of steroid treatment before PET impacts SUV/LE values in GCA patients but not controls. Boxplots for SUV/LE in relation to steroid treatment duration for the four defined vascular regions. (0) = without steroid treatment, n = 49 patients, after 1–3 days of steroid treatment, n = 16 patients, after 4–10 days of steroid treatment, n = 15 patients, and >10 days of steroid treatment, n = 23 patients. Key statistics are indicated. Discussion Visualization of metabolically active lesions in arteries using PET/CT is a promising technique; yet, its validity to diagnose GCA is still under debate.22 The aim of this study was to assess diagnostic usefulness of PET/CT in patients presenting at a vasculitis clinic with suspicion of GCA and to establish criteria for interpreting imaging. The study population reflects clinical practice and current understanding of GCA as a disease entity beyond temporal arteritis. Consequently 5 out of 68 included GCA patients fulfilled only 2/5 ACR criteria, but all 5 had clear vasculitic findings in at least one (additional to PET/CT) imaging technique. This may be a limitation when comparing our data to previous studies with more restrictive inclusion criteria. Studying a large patient collective in a single vasculitis centre, we found that PET/CT has an excellent diagnostic accuracy, for diagnosing vasculitis in the supra-aortic branches. However, the specificity of PET/CT in the diagnosis of GCA was significantly lower for both, the aorta and the iliofemoral vessel region. Moreover, we found that the duration of glucocorticoid treatment prior to PET/CT reduces its diagnostic power. Different performance of PET/CT in distinct vascular regions has also been described by others.15,23 Specifically, supra-aortic vessels were more often judged as being PET positive than the aorta,23 which might reflect, in part, a pattern of GCA vessel involvement. Indeed, we and others found that by ultrasound GCA patients, more often showed typical vasculitic vessel wall thickening in the supra-aortic arteries compared to the iliofemoral region.18,24 A similar distribution of vasculitic findings has also been described by CT angiography with a higher incidence of supra-aortic than iliac and femoral artery involvement25 although the aorta was the most commonly involved blood vessel in CT. The aorta has also been reported to be frequently involved by visual qualitative scoring in 10 patients with GCA.26 In line with our findings, a reduced specificity of PET/CT for the diagnosis of GCA of the lower extremities has been reported in a small patient population.27 The reason for the overall higher FDG uptake in the femoral arteries is not known. Arteriosclerosis, which is more prevalent in the aorta and iliofemoral region, may play a role through increased FDG uptake in macrophages within the plaques.28 Although we have avoided measurements within visible calcified plaques in CT, soft plaques might have been missed in non-contrast enhanced CT examinations and thus contribute to higher SUV values in this vascular region. Discrepancies of the findings between the different imaging techniques remain unsolved, because histology—the gold-standard for vasculitis cannot be obtained for the involved vessels. However, our finding of a high number of non-GCA patients showing SUV/LE >1 within the aorta and the iliofemoral arteries indicates that we need to be aware of the risk of false positive GCA diagnoses and argues for adapted cut-off values for different vascular regions. Thus, based on our findings, we propose a SUV/LE cut-off for diagnosis of GCA of 1.0 for the supra-aortic region, but a higher ratio of 1.3 for the aorta and the iliofemoral region. Steroid treatment must be initiated upon suspicion of GCA with the effect that the diagnostic accuracy of PET/CT may be reduced. A decrease of FDG uptake in patients with GCA was reported after 3 months of treatment in longitudinal PET studies.23,29 Yet, others recently found that in a majority of patients FDG uptake may persist beyond 6 months. However, quantitative analyses were not performed.30 About half of the patients in our study received prednisone therapy before PET/CT, giving us the opportunity to study the impact of glucocorticoid treatment on PET findings in the different vascular regions. In general, SUV/LE ratios of patients with GCA were lower when patients received steroids before the PET/CT. This significantly impacted the performance of the PET/CT as early as 3 days after treatment initiation in the abdominal aorta, and 10 days in the remaining examined arteries. However, these data have to be interpreted with caution, because our study was not designed as a longitudinal PET/CT study and the decision to start prednisone before PET/CT may result in a bias through selection of specific patients (e.g. patients with clear ischaemic symptoms). Yet, our data strongly argue that PET/CT should be performed immediately, and—if the clinical setting permits—before initiation of the glucocorticoid therapy. When comparing SUV/LE ratios and visual scoring to diagnose GCA on PET/CT, we found that SUV analysis outperforms visual semi-quantitative scoring. The visual score correlated poorly with objective, ratio based SUV analysis, and showed poor interobserver correlation. Thus the risk of misclassification and subjective interpretation by visual scoring seems to be higher than previously assumed. In addition, visual interpretation is prone to perception phenomena such as simultaneous contrast and assimilation. Taken together, our data suggests that visual grading should not be used as sole read-out when interpreting PET imaging in the setting of suspected GCA. The overall specificity of PET/CT for the diagnosis of GCA in our cohort was comparable to other studies with smaller patient numbers that used visual or SUV based scoring.16 However, sensitivity of PET/CT for diagnosing GCA in our study was lower than reported in previous studies with sensitivities ranging from 0.77 to 1.0.31,32 This may in part be due to an already established steroid treatment in about half of our patients at the time of PET/CT. In the majority of the previous studies, patients were excluded when having received steroids for longer than 3 days prior to PET/CT.15 Moreover, our ‘control group’ which consisted of patients with inflammatory syndromes, and thereby may differ from non-inflammatory controls like cancer patients or healthy controls which have been used in other studies. 15 Conclusion In conclusion, we found that a cut-off value of 1.0 for SUV/LE ratios for the supra-aortic region performed very well in the diagnostic work-up for suspected GCA. In the aorta and iliofemoral regions, a higher cut-off value of 1.3 needs to be applied to minimize false positive diagnoses. PET/CT presumably performs best before or within the first few days of starting steroid treatment. Analysis of SUV/LE ratios should be included in the interpretation of the PET/CT studies, whereas exclusive visual analysis should be avoided. Supplementary data Supplementary data are available at European Heart Journal - Cardiovascular Imaging online. Acknowledgements We thank the physicians of the rheumatology and medical outpatient clinic for their help in the patient care and the clinical trial unit of the Department of clinical research of the university hospital Basel for supporting the local cohort of GCA patients. Funding C.T.B. is supported by the Swiss National Science Foundation (SNSF) [PZ00P3-148000]. Conflict of interest: None declared. References 1 Berger CT , Wolbers M , Meyer P , Daikeler T , Hess C. High incidence of severe ischaemic complications in patients with giant cell arteritis irrespective of platelet count and size, and platelet inhibition . Rheumatol (Oxford) 2009 ; 48 : 258 – 61 . Google Scholar CrossRef Search ADS 2 Haering M , Holbro A , Todorova MG , Aschwanden M , Kesten F , Berger CT et al. Incidence and prognostic implications of diplopia in patients with giant cell arteritis . J Rheumatol 2014 ; 41 : 1562 – 4 . Google Scholar CrossRef Search ADS PubMed 3 Meller J , Strutz F , Siefker U , Scheel A , Sahlmann CO , Lehmann K et al. Early diagnosis and follow-up of aortitis with [(18)F]FDG PET and MRI . Eur J Nucl Med Mol Imaging 2003 ; 30 : 730 – 6 . Google Scholar CrossRef Search ADS PubMed 4 Moosig F , Czech N , Mehl C , Henze E , Zeuner RA , Kneba M,. U A. Correlation between 18-fluorodeoxyglucose accumulation in large vessels and serological markers of inflammation in polymyalgia rheumatica: a quantitative PET study . Ann Rheum Dis 2004 ; 63 : 870 – 3 . Google Scholar CrossRef Search ADS PubMed 5 Lariviere D , Benali K , Coustet B , Pasi N , Hyafil F , Klein I et al. Positron emission tomography and computed tomography angiography for the diagnosis of giant cell arteritis: a real-life prospective study . Medicine (Baltimore) 2016 ; 95 : e4146 . Google Scholar CrossRef Search ADS PubMed 6 Liozon E , Ly K-H , Robert P-Y. [Ocular complications of giant cell arteritis] . Rev Med Interne 2013 ; 34 : 421 – 30 . Google Scholar CrossRef Search ADS PubMed 7 Aiello PD , Trautmann JC , McPhee TJ , Kunselman AR , Hunder GG. Visual prognosis in giant cell arteritis . Ophthalmology 1993 ; 100 : 550 – 5 . Google Scholar CrossRef Search ADS PubMed 8 González-Gay MA , García-Porrúa C , Llorca J , Hajeer AH , Brañas F , Dababneh A et al. Visual manifestations of giant cell arteritis. Trends and clinical spectrum in 161 patients . Medicine (Baltimore) 2000 ; 79 : 283 – 92 . Google Scholar CrossRef Search ADS PubMed 9 Salvarani C , Cimino L , Macchioni P , Consonni D , Cantini F , Bajocchi G et al. Risk factors for visual loss in an Italian population-based cohort of patients with giant cell arteritis . Arthritis Rheum 2005 ; 53 : 293 – 7 . Google Scholar CrossRef Search ADS PubMed 10 Younes M , Lechago LV , Somoano JR , Mosharaf M , Lechago J. Wide expression of the human erythrocyte glucose transporter Glut1 in human cancers . Cancer Res 1996 ; 56 : 1164 – 7 . Google Scholar PubMed 11 Zhao S , Kuge Y , Tsukamoto E , Mochizuki T , Kato T , Hikosaka K et al. Fluorodeoxyglucose uptake and glucose transporter expression in experimental inflammatory lesions and malignant tumours: effects of insulin and glucose loading . Nucl Med Commun 2002 ; 23 : 545 – 50 . Google Scholar CrossRef Search ADS PubMed 12 Slomka PJ , Pan T , Germano G. Recent advances and future progress in PET instrumentation . Semin Nucl Med 2016 ; 46 : 5 – 19 . Google Scholar CrossRef Search ADS PubMed 13 Walter MA , Melzer RA , Schindler C , Müller-Brand J , Tyndall A , Nitzsche EU. The value of [18F]FDG-PET in the diagnosis of large-vessel vasculitis and the assessment of activity and extent of disease . Eur J Nucl Med Mol Imaging 2005 ; 32 : 674 – 81 . Google Scholar CrossRef Search ADS PubMed 14 Besson FL , Parienti J-J , Bienvenu B , Prior JO , Costo S , Bouvard G et al. Diagnostic performance of 18F-fluorodeoxyglucose positron emission tomography in giant cell arteritis: a systematic review and meta-analysis . Eur J Nucl Med Mol Imaging 2011 ; 38 : 1764 – 72 . Google Scholar CrossRef Search ADS PubMed 15 Prieto-González S , Depetris M , García-Martínez A , Espígol-Frigolé G , Tavera-Bahillo I , Corbera-Bellata M et al. Positron emission tomography assessment of large vessel inflammation in patients with newly diagnosed, biopsy-proven giant cell arteritis: a prospective, case–control study . Ann Rheum Dis 2014 ; 73 : 1388 – 92 . Google Scholar CrossRef Search ADS PubMed 16 Soussan M , Nicolas P , Schramm C , Katsahian S , Pop G , Fain O et al. Management of large-vessel vasculitis with FDG-PET: a systematic literature review and meta-analysis . Medicine (Baltimore) 2015 ; 94 : e622. Google Scholar CrossRef Search ADS PubMed 17 Stellingwerff MD , Brouwer E , Lensen K-JDF , Rutgers A , Arends S , van der Geest KSM et al. Different scoring methods of FDG PET/CT in giant cell arteritis: need for standardization . Medicine (Baltimore) 2015 ; 94 : e1542 . Google Scholar CrossRef Search ADS PubMed 18 Aschwanden M , Kesten F , Stern M , Thalhammer C , Walker UA , Tyndall A et al. Vascular involvement in patients with giant cell arteritis determined by duplex sonography of 2x11 arterial regions . Ann Rheum Dis 2010 ; 69 : 1356 – 9 . Google Scholar CrossRef Search ADS PubMed 19 Stone JH , Tuckwell K , Dimonaco S , Klearman M , Aringer M , Blockmans D et al. Trial of tocilizumab in giant-cell arteritis . N Engl J Med 2017 ; 377 : 317 – 28 . Google Scholar CrossRef Search ADS PubMed 20 Hanley JA , McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve . Radiology 1982 ; 143 : 29 – 36 . Google Scholar CrossRef Search ADS PubMed 21 Diamantopoulos AP , Haugeberg G , Hetland H , Soldal DM , Bie R , Myklebust G. Diagnostic value of color doppler ultrasonography of temporal arteries and large vessels in giant cell arteritis: a consecutive case series . Arthritis Care Res 2014 ; 66 : 113 – 9 . Google Scholar CrossRef Search ADS 22 Weyand CM , Goronzy JJ. Clinical practice. Giant-cell arteritis and polymyalgia rheumatica . N Engl J Med 2014 ; 371 : 50 – 7 . Google Scholar CrossRef Search ADS PubMed 23 Blockmans D , Ceuninck L. D , Vanderschueren S , Knockaert D , Mortelmans L , Bobbaers H. Repetitive 18F-fluorodeoxyglucose positron emission tomography in giant cell arteritis: a prospective study of 35 patients . Arthritis Care Res 2006 ; 55 : 131 – 7 . Google Scholar CrossRef Search ADS 24 Assie C , Janvresse A , Plissonnier D , Levesque H , Marie I. Long-term follow-up of upper and lower extremity vasculitis related to giant cell arteritis: a series of 36 patients . Medicine (Baltimore) 2011 ; 90 : 40 – 51 . Google Scholar CrossRef Search ADS PubMed 25 Prieto-González S , Arguis P , García-Martínez A , Espígol-Frigolé G , Tavera-Bahillo I , Butjosa M et al. Large vessel involvement in biopsy-proven giant cell arteritis: prospective study in 40 newly diagnosed patients using CT angiography . Ann Rheum Dis 2012 ; 71 : 1170 – 6 . Google Scholar CrossRef Search ADS PubMed 26 Henes JC , Müller M , Krieger J , Balletshofer B , Pfannenberg AC , Kanz L et al. [18F] FDG-PET/CT as a new and sensitive imaging method for the diagnosis of large vessel vasculitis . Clin Exp Rheumatol 2008 ; 26 (3 Suppl 49): S47 – 52 . [WorldCat] Google Scholar PubMed 27 Förster S , Tato F , Weiss M , Czihal M , Rominger A , Bartenstein P et al. Patterns of extracranial involvement in newly diagnosed giant cell arteritis assessed by physical examination, colour coded duplex sonography and FDG-PET . Vasa 2011 ; 40 : 219 – 27 . Google Scholar CrossRef Search ADS PubMed 28 Yun M , Jang S , Cucchiara A , Newberg AB , Alavi A. 18F FDG uptake in the large arteries: a correlation study with the atherogenic risk factors . Semin Nucl Med 2002 ; 32 : 70 – 6 . Google Scholar CrossRef Search ADS PubMed 29 Papathanasiou ND , Du Y , Menezes LJ , Almuhaideb A , Shastry M , Beynon H et al. 18F-Fludeoxyglucose PET/CT in the evaluation of large-vessel vasculitis: diagnostic performance and correlation with clinical and laboratory parameters . Br J Radiol 2012 ; 85 : e188 – 94 . Google Scholar CrossRef Search ADS PubMed 30 de Boysson H , Aide N , Liozon E , Lambert M , Parienti J-J , Monteil J et al. Repetitive (18)F-FDG-PET/CT in patients with large-vessel giant-cell arteritis and controlled disease . Eur J Intern Med 2017 ; doi:10.1016/j.ejim.2017.08.013. 31 Meller J , Sahlmann CO , Gürocak O , Liersch T , Meller B. FDG-PET in patients with fever of unknown origin: the importance of diagnosing large vessel vasculitis . Q J Nucl Med Mol Imaging 2009 ; 53 : 51 – 63 . Google Scholar PubMed 32 Hautzel H , Sander O , Heinzel A , Schneider M , Müller H-W. Assessment of large-vessel involvement in giant cell arteritis with 18F-FDG PET: introducing an ROC-analysis–based cutoff ratio . J Nucl Med 2008 ; 49 : 1107 – 13 . Google Scholar CrossRef Search ADS PubMed Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2017. For permissions, please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

European Heart Journal – Cardiovascular ImagingOxford University Press

Published: Nov 8, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off