Biopsy vs imaging for the diagnosis of giant cell arteritis. Viewpoint 2: in favour of biopsy

Abstract

Objectives

Both temporal artery biopsy (TAB) and imaging are widely used to support the diagnosis of giant cell arteritis (GCA). The objective of this study was to compare the use of TAB and imaging.

Methods

This article was based on a debate presented at the 21st Vasculitis Meeting, discussing the advantages and disadvantages of using TAB with histology vs imaging for the diagnosis of suspected GCA.

Results

TAB is the diagnostic procedure with the highest specificity. Its sensitivity may be improved by removing an appropriate artery length, practice, examining multiple sections at various levels, and by recognizing incomplete histological findings (which may lead to a more definitive diagnosis by further sectioning or imaging or be related to other inflammatory diseases). TAB may provide histopathological clues useful for diagnosing GCA mimics that may produce similar imaging abnormalities. TAB is a useful research resource, and our current understanding of GCA physiopathology mostly relies on tissue immunopathology studies.

Conclusion

A suspected diagnosis of GCA should be supported by an objective test. TAB is the procedure with the highest specificity, and its sensitivity may be improved by training. Histopathologic examination provides data for an alternative diagnosis, when diseases other than GCA involve the temporal artery. Imaging is essential for the assessment of large-vessel involvement and allows follow-up studies.

Introduction

The diagnosis of giant cell arteritis (GCA) requires a multistep process starting from the clinical suspicion of GCA, followed by physical examination and laboratory tests (most commonly CRP, ESR, haemoglobin concentration and platelet counts). This initial assessment configures a clinical pre-test probability estimation, which is highly influenced by the physician’s experience [1].

Since clinical diagnosis is of limited accuracy, a supporting test needs to be performed. Paradoxically, clinical diagnosis is the reference standard in many studies assessing the sensitivity and specificity of a particular test. In this setting, pre-test probability influences the interpretation of the test, particularly in the case of negative results, and this presents a circularity that is difficult to overcome [2].

The first temporal artery biopsy (TAB) was performed in 1931 and revealed a granulomatous, chronic arteritis. Since then, TAB has been widely used worldwide to support the diagnosis of GCA. A TAB directly reveals the nature of the disease and has the highest specificity for diagnosis compared with imaging, which shows surrogate changes that may have variable accuracy [2, 3].

As for any test, TAB has limitations, since ∼70% of TABs performed in expert referral centres are negative [4–6]. However, a negative TAB is helpful to the clinician, since it decreases the likelihood of GCA; in fact, the majority of patients with a negative biopsy will not be diagnosed with GCA [6–8]. When a TAB is negative, the clinician should widen the differential diagnosis and pursue an alternative diagnosis explaining the patient’s symptoms.

However, only ∼20–30% of TABs performed in expert centres are truly negative, and GCA diagnosis is then based on expert opinion and typical course, and/or on clear imaging demonstration of large artery involvement or involvement of cranial arteries other than the temporal artery. A negative biopsy should also prompt the clinician to complement the assessment with imaging, particularly in cases of moderate to high suspicion.

The limitations to TAB sensitivity may be technical or due to the involvement of vascular territories other than the temporal artery [6].

Technical limitations

Technical limitations may be related to suboptimal sampling, insufficient sectioning, or lack of awareness of incomplete findings.

Suboptimal sampling

During the surgical procedure, sometimes, no arterial tissue is obtained. This may be overcome by better training of the physician carrying out the biopsy or by performing an US-guided biopsy in cases with no detectable pulsation or absence of thickening.

The length of the biopsy matters according to some but not all studies [9, 10]. Variability may be due to the cross-sectional nature of the existing studies: in heavily inflamed arteries, a single section may provide sufficient evidence for GCA lesions, but in arteries with segmental involvement, a longer segment increases the probability of detecting disease features. Depending on the proportion of heavily inflamed vs segmental arteries included in these studies, the results may vary. A minimal length of 0.5–2 cm is recommended [6, 8, 11].

In the case of a negative biopsy and moderate to high suspicion of GCA, a contralateral TAB may be performed: several studies have shown that discordance between the two sides occurs in 3–13% of cases [12]. There are low but real chances of confirming diagnosis in a second biopsy.

If other cranial arteries appear more thickened than the temporal artery, it may be worth doing biopsies of other branches. Reports on retroauricular, occipital, or facial (at the submandibular level) biopsies have been reported as disclosing GCA lesions. If those sites are biopsied instead of or in addition to the temporal artery, an increase in the diagnostic yield is likely, but this has not been formally tested.

Suboptimal processing

Examining a few sections only may miss GCA in specimens with segmental lesions [12]. Therefore, it is important to examine multiple cross-sections at various levels of the artery segment. Agostino et al. examined 20 sections at serial levels 100 µm apart. They found GCA lesions at the first 3 levels in 92% of cases, but further sectioning was needed to evidence GCA in the remaining 8% of patients [12]. Therefore, it is advisable to obtain and examine multiple sections at various levels.

Limited awareness of incomplete findings

Some TABs do not show all the characteristic features of the disease. For instance, it is well established that the characteristic giant cells are observed in approximately half of the biopsies. Some TABs do not disclose transmural inflammation (Fig. 1A), but inflammation at the adventitia and intima with a spared media, or even inflammation limited to the adventitia (Fig. 1B) [13]. Clear inflammatory infiltrates in the adventitia in most of the arterial circumference may support the diagnosis, but this is controversial [13, 14]. A recently published consensus document recommends that pathologists should not report TABs dichotomously as being positive or negative. They should rather describe the precise findings observed [14].

Discordance among pathologists examining the same TAB have been reported, indicating limited awareness of the relevance of incomplete or less typical abnormalities, which can be improved by training [2].

When the findings are subtle (i.e. slight inflammatory changes in some areas of the adventitia, particularly in the surrounding vasa vasorum), there is some diagnostic uncertainty. It adds to the clinical suspicion, and sometimes further sectioning, a contralateral biopsy, or imaging may provide clearer findings. However, these changes have been observed in other diseases (i.e. isolated PMR, small-vessel vasculitis, malignancy, infection), usually accompanied by a marked systemic inflammatory reaction [15].The finding of subtle inflammatory changes is not irrelevant but is insufficient to fully support the diagnosis of GCA.

Lack of temporal artery involvement

The temporal artery is involved in a high percentage of patients with GCA, but TAB provides a small sample of the vascular tree and, as mentioned, other cranial or extracranial arteries may be involved. A prospective study performing US and TAB on the same patients within 1 week after the beginning of glucocorticoid treatment, and considering clinical diagnosis as the standard, found that, by exploring various bilateral cranial arteries and the axillary arteries, US had higher diagnostic sensitivity than the TAB, but TAB had the highest specificity—100% in this study [16]. It is important to remark that this study describes one of the lowest reported sensitivities of the TAB (32%), as compared with others, and that the standard was the clinical diagnosis, which is highly influenced by individual expertise. Noticeably, earlier studies assessing the sensitivity of US of the temporal artery only, compared with TAB results, showed that US had lower sensitivity [17, 18]. This makes sense, since it is like examining an arterial segment under a magnifying lens vs a microscope. Sensitivity, of course, increases by examining multiple artery segments. A recent meta-analysis reported an overall sensitivity of 77% [16].

In case of moderate to high suspicion, and in the setting of a negative biopsy, imaging of cranial and extracranial arteries may be essential. Obviously, in the absence of temporal artery involvement, imaging outperforms TAB.

Temporal arteritis or temporal artery thickening is not always GCA

The temporal artery can be involved in other vasculitis, including ANCA-associated vasculitis and other diseases. When the temporal artery is involved, cranial symptoms suggesting GCA or local discomfort may be present.

In the case of temporal artery involvement in other vasculitis, histopathological examination of the TAB reveals useful clues suggesting the existence of other systemic vasculitis. Fibrinoid necrosis and a marked presence of neutrophils suggest necrotizing vasculitis, rather than GCA [14, 19]. The presence of eosinophils may suggest eosinophilic granulomatosis with polyangiitis. Predominant involvement of small arteries rather than the temporal artery itself also indicate that patients may have small-vessel vasculitis (Fig. 1C) [20].

The temporal artery may be involved in other diseases that may mimic cranial, systemic or local abnormalities typical of GCA. These include systemic amyloidosis (Fig. 1D), thromboangiitis obliterans, kimura disease, arterial calcification, and infiltration by B lymphocytes in chronic lymphocytic leukaemia. In some of these diseases, and even in severe arteriosclerosis, imaging may detect thickening of the artery wall [21].

TAB may detect the footprint of GCA in patients treated for months or even years

The current recommendations advise immediate initiation of glucocorticoid treatment upon suspicion of GCA, and frequently, patients have already been treated for several days or weeks when they are evaluated in referral centres. Imaging findings may fade or decrease with high-dose glucocorticoid treatment in a substantial proportion of patients, reducing its diagnostic sensitivity [22]. Although high-dose glucocorticoid induces functional changes in inflammatory mediator production in GCA arteries, histopathological lesions may persist for weeks or months, as repeatedly reported in the literature [8, 23, 24]. Clusters of inflammatory cells, along with disruption of the normal vessel wall architecture, can be detected even years after the initiation of treatment, in temporal artery or aortic specimens [8, 23]. Although not rigorously tested, imaging may be less sensitive than TAB during the first weeks or months after the initiation of treatment.

TAB is an important research tool

Immunopathology and molecular pathology studies in the remaining biopsy tissue have, over the years, provided useful information supporting the current pathogenesis model of GCA [25]. More recently, high-throughput techniques applied to GCA tissue have confirmed existing information obtained from hypothesis-driven, candidate molecule approaches, and have revealed a much higher level of complexity, identifying unexpected potentially relevant molecules and interactions [26–28]. Experimental models allowing exploration of the specific impact of certain molecules and pathways on GCA immunopathology have been developed and have provided proof of concept of some pathogenic principles derived from association studies. Existing models include the development of patient-derived artery xenografts into SCID mice [29], adoptive inflammation of donor-derived arterial xenografts with peripheral blood mononuclear cells from patients with GCA [30], and ex-vivo culture of GCA-involved arteries in tridimensional matrix [30], and isolation and expansion of primary myofibroblasts [31]. Preclinical studies in these models have identified candidate molecules (i.e. GM-CSF receptor α, IL-17A) for targeted treatments [30, 32–34], and their efficacy have already been confirmed in phase 2 clinical trials [35, 36].

Conclusion

TAB is the method with the highest specificity for the diagnosis of GCA and may be useful in providing histopathological clues for diagnosis of alternative conditions that may mimic GCA. Imaging may be useful to complement the limitations in its diagnostic sensitivity, and is less invasive and can be more easily implemented in fast-track systems in some centres. Regardless of the method used to evaluate the cranial arteries (TAB or imaging), systematic evaluation of large-vessel involvement, present in ∼70% of patients with biopsy-proven GCA, may be advisable based on a prospective, long-term follow-up study showing that patients with a positive PET scan at diagnosis are at higher risk of thoracic aortic aneurysms [37].

Data availability

Not applicable.

Funding

Information generated by the authors has been funded by Ministerio de Ciencia, innovación y universidades (Agencia Estatal de Investigación): PID2020-114909RB-I00 and PID2023-152265OB-I00, and Agencia de Gestió d’Ajuts Universitarisi de Recerca (AGAUR): 2021 SGR 01561.

Disclosure statement: G.E.F. has received consulting/lecturing fees from GSK and CSL-Vifor and a research grant from GSK. M.C.C. has received consulting/lecturing fees from GSK, AbbVie, Astrazeneca and CSL-Vifor and a research grant from Kiniksa Pharmaceuticals Ltd.

References

Author notes