Binnen Nederland bestaat er praktijkvariatie wat betreft het diagnostisch proces bij RCA. Van oudsher wordt de diagnose gesteld door middel van een temporale arterie biopsie (TAB), echter non-invasieve beeldvorming in de vorm van echografie, moleculaire beeldvorming (FDG PET/CT-scan) en craniële MRI/MRA-scan (c-MRI/MRA) worden steeds meer toegepast. Deze technieken bieden vele voordelen, maar kunnen alleen veilig worden toegepast wanneer deze goed zijn ingebed in het diagnostische proces en wanneer er expertise aanwezig is voor de uitvoering en interpretatie hiervan. Een uitdaging is dat patiënten zich bij verschillende specialismen kunnen presenteren. Daarnaast is de expertise m.b.t. RCA per praktijk wisselend. Ook bestaat er variatie in beschikbaarheid, standaardisering en kwaliteitswaarborging van diagnostische testen. Deze variatie leidt tot een heterogeen diagnostisch proces dat potentieel kan leiden tot onterecht gestelde of gemiste diagnosen. Echter, een snelle en juiste diagnose is een voorwaarde voor een adequate behandeling en het voorkomen van irreversibele orgaanschade (zoals visusverlies). Dit benadrukt het belang van deze Nederlandse multidisciplinaire richtlijn. Vaatwandbetrokkenheid bij beeldvorming kan ook aanwezig zijn in afwezigheid van lokale klinische verschijnselen en vice versa (Bosch, 2022; Michailidou, 2019).

First, we will describe the results of the BSR guideline (2020)

Then we will describe any new studies results we found in our current update.

BSR guideline

Diagnostic accuracy may be expressed as sensitivity or specificity, or as likelihood ratios; this information can be combined with the pre-test probability (established on clinical grounds) to select and interpret the results of confirmatory diagnostic tests. Compared to temporal artery biopsy, imaging tests such as color Doppler ultrasound have the advantage of access to both superficial temporal arteries in their entirety, and evaluation of other by ultrasound accessible arteries such as the axillary arteries. In the BSR guideline, most diagnostic accuracy studies have focused on the role of ultrasound (n=16) or MRI (n=7). One study addressed the role of FDG PET/CT-scan, and another study examined the role of the FDG PET/CT-scan and CT angiography for GCA diagnosis.

Ultrasound ‘halo sign’ vs. current practice

Seven studies (519 patients with suspected GCA, of whom 169 were diagnosed with GCA) compared the ultrasound ‘halo sign’ with a clinical diagnosis of GCA, giving a pooled sensitivity of 79% (95% CI: 73%-84%) and pooled specificity of 94% (95% CI: 90%-96%) (Aschwanden, 2013; Diamantopoulos, 2014; Habib, 2012; Karahaliou, 2006; Nesher, 2002; Reinhard, 2004; Schmidt, 1997). Quality of evidence (QoE) was +++; downgrading was performed because of risk of bias in 4/7 studies. One of these studies included 12 patients with a final diagnosis of LV-GCA (Diamantopoulos, 2014).

Ultrasound ‘halo sign’ vs. temporal artery biopsy

Five studies (185 patients with suspected GCA, of whom 57 were diagnosed with GCA) compared the ultrasound ‘halo sign’ with temporal artery biopsy, giving a pooled sensitivity of 74% (95% CI: 63%-83%) and pooled specificity of 81% (95% CI: 73%-88%) (Nesher, 2002; Reinhard, 2004; Schmidt, 1997; LeSar, 2002; Murgatroyd, 2003). QoE was +; downgrading was performed because of high risk of bias in all 5 studies, and because of inconsistency. Patients with LV-GCA were not evaluated in these studies.

Ultrasound ‘compression sign’ vs. current practice

Two studies (140 patients with suspected GCA, of whom 67 were diagnosed with GCA) compared the ultrasound ‘compression sign’ of temporal arteries with ACR criteria-based diagnosis of GCA, giving a pooled sensitivity of 79% (95% CI: 67%-88%) and a pooled specificity of 100% (95% CI: 95-100) (Aschwanden, 2013; 2015). QoE ++; downgrading was performed for risk of bias in one of the studies, and for the fact that both studies were performed by the same research group. The ACR criteria for GCA, which are not suitable for clinical diagnosis, served as reference standard in both studies.

Ultrasound vs. clinical practice

Three studies (560 patients with suspected GCA, of whom 327 had a clinical diagnosis of GCA) compared the diagnostic performance of ultrasound abnormality (defined as any one of halo sign, stenosis or occlusion) with clinical diagnosis of GCA, giving a pooled sensitivity of 61% (95% CI: 56%-67%) and pooled specificity of 86% (95% CI: 81%-90%) (Luqmani, 2016; Schmidt, 1997; Pfadenhauer, 2003). QoE ++; downgrading was performed for risk of bias in all three studies, and for inconsistency.

Four studies (563 patients with suspected GCA, of whom 180 had a positive temporal artery biopsy) compared the diagnostic performance of ultrasound abnormality (defined as any one of halo sign, stenosis or occlusion) with temporal artery biopsy, giving a pooled sensitivity of 81% (95% CI: 74%-86%) and pooled specificity of 74% (95% CI: 70%-79%) (19, 62, 66, 67). QoE ++; downgrading was performed for risk of bias in three of the four studies, and for imprecision.

A positive temporal artery biopsy, showing features of inflammation characteristic of GCA such as giant cells or panarteriitis (Lie, 2018), confirms the diagnosis of GCA. Although the true sensitivity of temporal artery biopsy is not precisely known, it is accepted that its sensitivity is substantially less than 100%; this is supported by the histological observation of skip lesions in some cases. An imperfect reference standard would result in underestimation of the diagnostic accuracy of ultrasound. When using clinical diagnosis as a reference standard it is important that this is made independently of the index test result in order to avoid bias; this may be done by blinding of the diagnostician to the index test result. Notably a large prospective UK study assessing the diagnostic value of ultrasound addressed this issue by blinding the patient, the treating clinician and the investigator to the ultrasound result (Luqmani, 2016). Ultrasound was found to be more sensitive but less specific than biopsy for diagnosis of GCA, was cost-effective and provided scope for reducing the number of patients who need a temporal artery biopsy (Luqmani, 2016). Overall, the pooled positive and negative likelihood ratios for ultrasound appear to support its use either for ruling out GCA in low-probability cases or for confirming GCA in high-probability cases (Appendix E and Figure 1 of the BSR publication (Mackie, 2020)). Ultrasound of the axillary arteries might add extra diagnostic information to ultrasound of the temporal arteries (Schmidt, 2018).

Cranial MRI vs. current practice

Six studies (500 patients with suspected GCA, of whom 268 were finally diagnosed with GCA) compared cranial artery MRI (vessel wall oedema and contrast enhancement) with clinical diagnosis, giving a pooled sensitivity of 75% (95% CI: 69%-80%) and a pooled specificity of 89% (95% CI: 84%-93%) (Bley, 2007; Bley, 2005; Geiger, 2010; Klink, 2014; Siemonsen, 2015; Rheaume, 2017). QoE ++; downgrading was performed for risk of bias in five of the studies, and for the fact that five of the six studies were performed by the same research group; sensitivity was somewhat lower in the study performed by a separate group (Rheaume, 2017).

Cranial MRI vs. temporal artery biopsy

Five studies (397 patients with suspected GCA, of whom 171 had positive temporal artery biopsy) compared cranial artery MRI (vessel wall oedema and contrast enhancement) with temporal artery biopsy, giving a pooled sensitivity of 94% (95% CI: 90%-97%) and specificity of 79% (95% CI: 73%-84%) (Bley, 2007; Bley, 2005; Geiger, 2010; Klink, 2014; Rheaume, 2017). QoE +; downgrading was performed for risk of bias in five of the studies, for inconsistency, and for likely publication bias.

Overall, MRI of the cranial arteries appears to be potentially useful for ruling out GCA if the result is negative, but false positive test results could occur, such that MRI of the cranial arteries would not be first choice for a confirmatory test in GCA (Rheaume, 2017). Other issues of relevance to cranial vascular MRI are low availability of high-resolution 3T MRI equipment and expertise, higher costs and possible adverse effects of contrast agents.

In contrast to the 2010 guideline, where the authors outlined that imaging techniques are promising for diagnosis and monitoring of GCA (Dasgupta, 2010), in this guideline there is now sufficient evidence, taken together, to state that all patients with GCA should have at least one confirmatory diagnostic test, which could be either temporal artery biopsy, or temporal and axillary artery ultrasound. However, temporal artery biopsy and ultrasound differ in their positive and negative likelihood ratios for GCA, with biopsy having relatively greater “rule-in” value and ultrasound having relatively greater “rule-out” value (Appendix E). Selection of the most appropriate confirmatory diagnostic test(s) therefore requires an assessment of the pre-test probability as outlined elsewhere (Dejaco, 2018); if both ultrasound and biopsy are possible, an approach to this is suggested in Figure 1 of the BSR guideline by Mackie (2020).

The ultrasound ‘halo sign’ diminishes in size during the first week of glucocorticoid therapy, indicating that sensitivity of the test is likely to depend on the delay between initiation of glucocorticoid therapy and the ultrasound test (Luqmani, 2016). Ultrasound is operator-dependent and requires adequate training. Ultrasound performs best in the “fast-track” setting, assuming rapid access, good technical equipment and high expertise with this method. With ultrasound, the non-compressible ’halo’ sign is the most important finding suggesting GCA (Chrysidis, 2018). Temporal artery biopsy should be performed by a surgeon experienced in this procedure, and samples should be at least 1cm in length post-fixation. The pathologist evaluating the biopsy should be experienced in diagnosing GCA. Data from the TABUL study (Luqmani, 2016) suggested significant variation between pathologists in the interpretation of temporal artery biopsy histology, so where biopsy findings are ambiguous (eg low-level inflammation restricted to the adventitia), discussion between the requesting clinician and the pathologist is desirable. In the absence of inflammatory infiltrate, a report of healed arteritis is not sufficient to diagnose GCA. Isolated vasa vasorum vasculitis is not diagnostic of GCA. Contralateral biopsy may slightly increase the yield of temporal artery biopsy, but is usually unnecessary. Biopsy may remain positive for several weeks after initiation of glucocorticoid therapy (Jakobsson, 2016).

Evaluation of involvement of the aorta and its proximal branches (LV-RCA)

FDG PET/CT-scan vs. current practice

One study (24 patients with suspected GCA, of whom 15 were diagnosed with GCA) compared FDG PET/CT-scan with clinical diagnosis of GCA, giving a sensitivity of 67% (95% CI: 38%-88%) and a specificity of 100% (95% CI: 66% to 100%) (Lariviere, 2016). QoE ++; downgraded because of indirectness and publication bias.

FDG PET/CT-scan vs. temporal artery biopsy

One study (69 patients with suspected GCA/PMR, of whom 13 had biopsy evidence of GCA) compared vascular 18F-glucose uptake in thorax and legs on FDG PET/CT-scan with temporal artery biopsy, giving a sensitivity of 77% (95% CI: 46% to 95%) and specificity of 66% (95% CI: 52% to 78%). Comparing vascular 18F-glucose uptake in thorax on FDG PET/CT-scan with temporal artery biopsy gave a sensitivity of 54% (95% CI: 25% to 81%) and specificity of 86% (95% CI: 74% to 94%). QoE +; downgraded because of risk of bias, indirectness and imprecision (Blockmans, 2000).

CT angiography vs. current practice

One study (24 patients with suspected GCA, of which 15 were diagnosed with GCA) compared CT angiography (CTA) with clinical diagnosis of GCA, giving a sensitivity of 73% (95% CI: 45%-92%) and specificity of 78% (95% CI: 40%-97%) (Lariviere, 2016). QoE++; downgraded for indirectness and publication bias. CTA can reveal wall thickening with contrast enhancement in biopsy-proven GCA (Prieto-González, 2012). There is also experience with CTA for accurate assessment of luminal diameter for large vessel stenosis in Takayasu arteritis (Yamada, 1998).

No studies of MR angiography for the diagnosis of LV-GCA were found meeting our criteria, but there is experience with MRI for detection of vessel wall oedema reflective of inflammation and accurate assessment of luminal diameter for large vessel dilatation and stenosis in diseases of the major arteries, such as Takayasu arteritis. Gadolinium-enhanced MR angiography may help identify aortitis in the large-vessel vasculitides but appears to be very sensitive to glucocorticoid therapy (Adler, 2017).

In addition to showing inflammation of the large vessels, whole body FDG PET/CT-scan may detect malignancy or infection so can be of use in the differential diagnosis of GCA. Contrast-enhanced CT of the chest and abdomen is also often used in clinical practice to screen for deep infection or occult malignancy. Moreover, aortic wall thickening on a contrast CT might help to identify GCA, albeit with lower sensitivity than FDG PET/CT-scan and could also potentially have uses in settings where FDG PET/CT-scan is unavailable (Lariviere, 2016; Vaidyanathan, 2018; Muto, 2014). Additional advantages of FDG PET/CT-scan and CT therefore include potential value in the workup of alternative diagnoses such as malignancy and infection. 

As well as detecting axillary artery involvement for diagnosis of large-vessel involvement in GCA, vascular ultrasound may also be able to visualize the carotid arteries and obtain more limited views of the subclavian arteries, vertebral arteries, and parts of the aorta, but a higher level of operator expertise is required for these studies.

Overall, there is indirect evidence for the use of imaging tests to evaluate involvement of the aorta and its proximal branches in GCA, but the published evidence is extrapolated from other diseases such as Takayasu arteritis (Dejaco, 2018) and there is currently insufficient evidence from prospective studies of suspected GCA to yield precise estimates of diagnostic accuracy for these tests.

Update

An overview of the results is shown in Table 1.

Description of studies

Halo-sign ultrasound vs clinical practice and/or TAB:

Molina Collada (2021) performed a retrospective cohort study to investigate the diagnostic values of ultrasound (US) ‘halo sign’ compared with the gold standard for GCA diagnosis (i.e., clinical confirmation after 6-month follow-up). Patients with suspected GCA were referred to the fast-track pathway and underwent US examination within 24 h. temporal artery biopsy (TAB) and FDG PET/CT-scan were performed according to the treating clinician criteria. The gold standard for GCA diagnosis was clinical confirmation after 6-month follow-up. Patients underwent bilateral US examination of the three temporal artery (TA) segments (common superficial TA, its parietal and frontal branches) and extracranial (carotid, subclavian, and axillary) arteries. This was performed by the same assessor, using EsaoteMyLab8 with a high-frequency (12-18MHz) transducer. Ethical approval for this study was obtained. Importantly, the ultrasonographer was not blinded to clinical data.

In total 64 patients who were referred to the fast-track pathway were included in the analysis. In total 65% of them were female, and the mean age was 75.3 years. The prevalence of GCA was 26.5%, meaning that 17 patients were diagnosed with GCA. Limitation of the current study were the retrospective design, assessors were not blinded regarding clinical data, and steroids were used at baseline by 40-52% of the included patients.

Mukhtyar (2020) performed a cohort study to investigate the diagnostic values of US ‘halo sign’ compared with the gold standard for GCA diagnosis and TAB. Patients were eligible for inclusion if they had an US within 7 days and TAB within 28 days of commencing high-dose prednisolone. The first author performed US on a Toshiba Viamo US machine with a linear transducer (4–14 MHz) using tissue harmonic imaging mode. Under local anesthesia by an ophthalmic surgeon, TAB was performed. The following definitions of results were used: the US was defined as positive in the presence of non-compressible vessel wall oedema (the ‘halo sign’) in longitudinal and transverse views, stenosis or obstruction. TAB was defined as positive in the presence of intramural inflammatory infiltrate. Clinical decisions were recorded as GCA if clinicians chose to treat patients with the hospital-approved Norwich regimen for prednisolone. In total 25 patients were included in the analysis. Detailed information about patient characteristics is missing. Twenty of them were initial treated as GCA, after 100 days 16 of them were treated as GCA (i.e., prevalence 64%). Limitations of the current study were the relatively low sample size, and the fact that patients underwent first US and TAB (possible that if US was negative, TAB was not performed).

González Porto (2020) performed a prospective cohort study to investigate the diagnostic values of US ‘halo sign’ compared with the gold standard for GCA diagnosis (i.e., clinical confirmation after 6-month follow-up). Patients with suspected temporal arteritis were eligible for inclusion. At the second visit, patients underwent diagnostic tests; 1. Colour Doppler ultrasound on both temporal arteries using Mindray Z6 ultrasound machine with 7L4P linear probe. All US examinations were performed by the same assessor; 2. 36h after US, temporal artery biopsy was performed. Ethical approval for this study was obtained.

In total 57 patients were included, of them 4 dropped out due to refusing biopsy (n=1) or died (n=3). Of the included patients 21/57 were diagnosed with GCA (prevalence of 37%). Of the patients with GCA, 62% were female, and their mean age was 77 years. A limitation of the current study was the fact that assessors were not blinded regarding clinical data.

Conway (2019) performed a study using prospectively collected data from a registry. The aim of the study was to evaluate the performance of temporal artery ultrasound (TA US) compared to expert clinical judgement in patients presenting with suspected GCA. If patients were referred with a new presentation of GCA and had both a TA US and TAB, they were eligible for inclusion. TA US were performed on a Philips iU22 scanner using a high-frequency linear array 12-MHz transducer. The examination was considered positive for temporal arteritis if any segment of the TA demonstrated circumferential hypoechoic mural thickening

(the ‘halo’ sign). The clinical diagnosis was made by a rheumatologist at 6 months. Importantly, the rheumatologist was not blinded to the other results. Ethical approval for this study was obtained.
In total 291 patients were recruited from the registry. Of them 256 were followed up for at least 6 months. However, only 169 of them had both a TA US and TAB. In addition, seven of them had non-arterial TAB specialism. Therefore only 162 patients were included in the analysis. Of the included patients 123/162 were diagnosed with GCA (prevalence of 76%). Of the patients with GCA, 59% were females, and their mean age was 71 years. A limitation of the current study is the fact that the reference standard (i.e., clinical diagnosis) is not described in detail, and that the assessors (i.e., rheumatologists) were not blinded to the outcome of TA US.

Cranial MRI vs clinical practice

Rodriguez-Régent (2020) performed a prospective cohort study to investigate the diagnostic values of cranial MRI (CUBE T1, 3D) compared with the gold standard for GCA diagnosis. Patients with suspected CGA who underwent a cranial MRI as part of the diagnosis process were included. Exclusion criteria were a previous TAB or glucocorticoid (GC) therapy >48 before the MRI scan. All patients underwent the ‘usual’ diagnostic process as well. Based on this work-up (i.e., usual), the medical doctor established the final diagnosis. MRI was performed in 1 center on a 3T unit (MR 750, GE Healthcare). The images were reviewed independently by two experienced neuroradiologists, blinded to all other data. Ethical approval for this study was obtained.
In total 32 patients were included, of them 78% were female and the mean age was 70.2 years. Of the included patients 10/32 were diagnosed with GCA (prevalence of 31%). A limitation of the current study is the fact that the reference standard (i.e., clinical diagnosis) is not described in detail. In addition, the 3 T MRI is not widely available.

Poillon (2020) performed a prospective cohort study to investigate the diagnostic values of a cranial MRI (2D and 3D) compared with the clinical diagnosis according to the ACR criteria in patients with suspected GCA. If patients met the in- and exclusion criteria, they were eligible for inclusion. Cranial MRI was performed on a 3 T Philips Ingenia or a 3 T General Electric Discovery MRI with a dedicated 32-channel head coil. The clinical diagnosis of GCA was made in the case of a positive TAB. In patients with a negative TAB, a review of the clinical and biological charts including response to corticosteroid therapy (but excluding MRI findings) was performed by an interdisciplinary panel of rheumatologists and internists not involved in the management of the patient. By consensus, they established a final diagnosis of GCA based on ACR criteria. Importantly, the rheumatologist was blinded to the other results. Ethical approval for this study was obtained.
In total 79 patients were included, of them 51 were diagnosed with GCA (prevalence of 64%). Of all patients, 47% were female, and the median was 75 years. A limitation of the current study is that no 95% confidence intervals were presented by the diagnostic values, and it was not possible to calculate these. In addition, the 3 T MRI is not widely available.

Ultrasound and FDG PET/CT vs. clinical practice or TAB

Imfeld (2020) performed a (prospective) cohort study to investigate the diagnostic values of ultrasound (US) and FDG PET/CT-scan compared with the clinical diagnosis according to the ACR criteria in patients with suspected GCA. If patients were referred with a new presentation of GCA and had both a US and FDG PET/CT-scan, they were eligible for inclusion. The FDG PET/CT-scan was performed with a GE Discovery STE 16 or Siemens Biograph. A cut-off value of 1.3 was used. US was performed with Philips iU22 with 17-5 MHz and 9-3 MHz broadband transducers. The final diagnosis of GCA was made either if TAB was positive if patients fulfilled the 1990 ACR criteria, or if they fulfilled at least two out of five ACR criteria in combination with typical ‘vasculitic’ US findings in accordance to the OMERACT definition or vasculitic findings in FDG PET/CT-scan or MRI.

Importantly, the rheumatologist was blinded to the other results. Ethical approval for this study was obtained.

In total 102 patients were included, of them 68 were diagnosed with GCA (prevalence of 66%). Of these patients, 65% were female, and the median age was 75 years. One limitation of the current study is that no 95% confidence intervals were presented by the diagnostic values, and it was not possible to calculate these.

Results

Results of the included studies are descripted per comparison. An overview of the results is shown in Table 1.

Ultrasound ‘halo sign’ vs. current practice

Molina Collada (2021) described the diagnostic values of ultrasound (US) with current practice as reference. The study showed that 15/17 GCA patients with a positive US were true positives, and that 45/47 patients without GCA and with a negative US were true negatives. 2/17 GCA patients with a negative US were false negative, and 2/47 patients without GCA were classified as false positive. This results in a sensitivity of 88% (95%CI 64% to 99%), and a specificity of 96% (95%CI 85% to 98%).

The study of Mukhtyar (2020) described the diagnostic values of US with current practice as reference. The study showed that 14/14 patients with a positive US were true positives, and that 9/11 patients with a negative US were true negatives. 2/11 patients with a negative US were false negative, and none of them were classified as false positive. This results in a sensitivity of 88% (95%CI 62% to 98%), and a specificity of 100% (95%CI 66% to 100%).

The study González Porto (2020) described the sensitivity and specificity values of US (i.e., halo sign, temporal stenosis, arterial occlusion) with current practice as reference. The sensitivity (95%CI) for US halo sign, temporal stenosis, arterial occlusion was 33% (15% to 57%), 14% (3% to 36%) and 10% (1% to 30%), respectively. The specificity (95%) for US halo sign, temporal stenosis, arterial occlusion was 69% (51% to 83%), 94% (81% to 99%), 97% (85% to 100%), respectively.

Conway (2019) described the sensitivity and specificity values of temporal artery US (i.e., halo sign) with current practice as reference. 65/123 patients with GCA had a positive US, resulting in a sensitivity of 53% (95%CI 44% to 62%). 28/39 patients without GCA had a negative US, resulting in a specificity of 72% (95%CI 55% to 85%).

Imfeld (2020) described the sensitivity and specificity of US compared with current clinical practice as reference standard. This resulted in a sensitivity of 57%, and a specificity of 97%.

Overall, based on 5 studies including 449 patients, the sensitivity ranges from 33% to 98%, and the specificity ranges from 55% to 100%.

Ultrasound ‘halo sign’ vs. TAB

The study of Mukhtyar (2020) described the diagnostic values of ultrasound with temporal artery biopsy (TAB) as reference. The study showed that 7/8 patients with a positive US were true positives, and that 10/17 patients with a negative US were true negatives. 1/8 patient with a negative US was false negative, and 7/17 patients with a positive US were false positives. This results in a sensitivity of 88% (95%CI 47% to 100%), and a specificity of 59% (33% to 82%).

Cranial MRI vs. current practice

Rodriguez-Régent (2020) described the diagnostic values of cranial MRI (i.e., mural enhancement in 3D post-contrast CUBE T1) compared with the clinical diagnose as reference. 8/10 patients with GCA had a positive cranial MRI, and none of the patients without GCA. This resulted in a sensitivity of 80% and a specificity of 100%.

Poillon (2020) described the diagnostic values of cranial MRI (i.e., 2D contrast-enhanced vessel-wall, axial only 3D contrast-enhanced vessel-wall reformatted 3D contrast-enhanced vessel-wall) compared with the clinical diagnose as reference. This resulted in a sensitivity of 70%, 73%, 80%, and a specificity of 85%, 93%, 100% for 2D, axial only 3D, and reformatted 3D, respectively.

Overall, based on 2 studies including 111 patients, the sensitivity ranges from 73% to 80%, and the specificity ranges from 93% to 100% using a 3D MRI.

FDG PET/CT-scan vs. current practice

Imfeld (2020) described the sensitivity and specificity of FDG PET/CT-scan compared with current clinical practice as reference standard. This resulted in a sensitivity of 72%, and a specificity of 85%. By using the US and FDG PET/CT-scan both, the sensitivity increased to 88%, and the specificity decreased to 82%.

Table 1: Summary of results.

FU= follow-up, TAB= temporal artery biopsy

Level of evidence of the literature

The level of evidence of the literature from the BSR-guideline (Mackie, 2020) was not assessed in the current summary of literature. We refer to this publication for detailed information.

The level of evidence (GRADE method) of literature included in the update was determined per comparison and diagnostic outcome measure and was based on results from diagnostic accuracy studies and therefore starts at level “high”. Subsequently, the level of evidence was downgraded if there were relevant shortcomings in one of the several GRADE domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias.

Ultrasound ‘halo sign’ vs. current practice

The level of evidence regarding the outcome measure sensitivity and specificity was downgraded by 2 levels because of risk of bias (1 level, assessors were not blinded in 3/4 studies), and imprecision (1 level, wide ranges).

Ultrasound ‘halo sign’ vs. TAB

The level of evidence regarding the outcome measure sensitivity and specificity was downgraded by 3 levels because of risk of bias (assessors were aware of result of the index test) and, imprecision (2 levels, only one study of 25 patients).

Cranial MRI vs. current practice

The level of evidence regarding the outcome measure sensitivity and specificity was downgraded by 2 levels because of risk of bias (potential selection bias in one of the included studies) and, imprecision (1 levels, only two studies including in 111 patients in total).

FDG PET/CT-scan vs. current practice

The level of evidence regarding the outcome measure sensitivity and specificity was downgraded by 3 levels because of risk of bias (potential selection bias in the included study) and, imprecision (2 levels, only one study of 102 patients).

Een update van het literatuuronderzoek van de BSR- richtlijn is uitgevoerd voor de huidige uitgangsvraag. Het betreft hier een search naar nieuwe literatuur tot en met 2021. Medio 2022 gepubliceerde artikelen zijn niet meer meegenomen in deze search.

Search and select

What is the value of diagnostic tests compared to current practice in the diagnosis of giant cell arteritis (GCA)?

P:           patients with suspected GCA

I:            additional diagnostic tests (advanced imaging, MRI, FDG-PET, duplex and B-mode ultrasound of cranial and/or extracranial arteries, halo sign, halo score, halo-to-Doppler ratio, compression sign, stenosis, occlusion, intima-media thickness, intima-media

thickness, adventitia thickness and slope sign, very-high frequency ultrasound, very-high resolution ultrasound)

C:           current practice ((clinical diagnosis (without formal criteria), ACR classification criteria and temporal artery biopsy result; GCA with/without extra-cranial)

O:          diagnostic value, diagnostic accuracy, diagnosis, reliability, utility, visual loss, patient reported outcomes

Relevant outcome measures

The guideline development group considered diagnostic values, diagnostic accuracy as a critical outcome measure for decision making; and diagnosis, reliability, utility, visual loss, patient reported outcomes as an important outcome measure for decision making.

The working group defined the outcome measures as follows: Diagnostic studies included direct outcomes (true positives, true negatives, false positives, false negatives, sensitivities and specificities; complications of the index test and of the reference standard; resource use), the number of studies and quality assessment related to each of these outcomes and the effect estimate (i.e. number of individuals classified per 1000 people) according to different pretest probabilities [low (<20%), intermediate (20–50%) and high (>50%) pretest

probability].

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched (updated) with relevant search terms from 2018 until May, 2021. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 860 hits. Studies were selected based on the following criteria;

- patients with suspected GCA,

- comparison between intervention (i.e., laboratory test or additional diagnostic tests) and control (i.e., current practice ((clinical diagnosis (without formal criteria), ACR classification criteria and temporal artery biopsy result; GCA with/without extra-cranial))

- At least one of the outcome measures of interest, as described in the PICO, is reported.

Thirty-six studies were initially selected based on title and abstract screening. After reading the full text, twenty-nine studies were excluded (see the table with reasons for exclusion under the tab Methods), and seven studies were included.

Results

Seven studies were included in the current update of the analysis of the literature. Important study characteristics and results are summarized in the evidence tables. The assessment of the risk of bias is summarized in the risk of bias tables.