Ultrasound-guided core needle biopsy and incisional biopsy of the parotid gland are comparable in diagnosis of primary Sjögren’s syndrome

Abstract

Objectives

Salivary gland lymphocytic infiltrates are a hallmark of primary SS (pSS), but traditional biopsy techniques hold several disadvantages. Ultrasound-guided core needle (US-guided CN) parotid gland biopsy is minimally invasive and reliable for diagnosis of lymphoma in pSS. This proof-of-concept study aimed to explore this technique in the diagnostic work-up of pSS and is the first to address its value in a consecutive cohort independently of the presence of salivary gland swelling.

Methods

Combined incisional and US-guided CN parotid biopsy was performed in 20 patients with suspected or confirmed pSS from the Belgian Sjögren’s Syndrome Transition Trial (BeSSTT). Surface area and presence of a focus score (FS) of at least one, germinal centres and lymphoepithelial lesions were recorded.

Results

Salivary gland tissue was interpretable in 19 patients. Fourteen patients had ≥4 mm² salivary gland tissue by both techniques, in four US-guided CN biopsies salivary gland tissue was <4 mm². Paired biopsies ≥4 mm² displayed a concordance of 90% for FS ≥ 1. Presence of lymphoepithelial lesions and germinal centres showed absolute concordance. Of four US-guided CN biopsies <4 mm², three interpretable incisional biopsies were available, 2/3 with perfect concordance. When including biopsies of <4 mm² salivary gland tissue, presence of FS ≥ 1 or germinal centres gave a sensitivity of 70% in incisional and of 69% in US-guided CN biopsy.

Conclusions

US-guided CN biopsy of the parotid gland is at least equivalent to incisional biopsy of the parotid gland in the diagnostic work-up of pSS.

Introduction

Salivary gland biopsy is of great importance in the diagnostic work-up of primary SS (pSS) to detect lymphocytic sialadenitis, a hallmark of the disease. The 2016 ACR/EULAR classification criteria are based on the presence of objective sicca and evidence of autoimmunity, which can be demonstrated by either detection of serum anti-SSA/Ro autoantibodies or lymphocytic infiltrates in the minor salivary glands, requiring a focus score (FS) of at least one [1]. Despite several disadvantages, labial salivary gland biopsy is still the gold standard to obtain salivary gland tissue. Besides being an invasive procedure, difficulties can be encountered to collect a sufficient number of glands in atrophic submucosa [2] and permanent sensory loss of the lower lip mucosa has been reported in 1% to 10% of patients [3].

Incisional biopsy of the parotid gland showed to be at least comparable to labial salivary gland biopsy in the diagnostic work-up of pSS [4–6], but has never become common practice despite its clear advantages. Unlike labial glands, the parotid gland allows comparison with diagnostic results of the same gland. In addition, lymphoepithelial lesions (LELs), which are more prevalent in parotid than minor salivary gland tissue, can aid in pSS diagnosis and be indicative of malignant lymphoma [7, 8]. Overall, almost no permanent complications were registered after incisional biopsy [4, 6, 9]. However, the majority of patients who underwent both parotid and labial salivary gland biopsy reported the parotid biopsy to be most burdensome one week and six months post-procedure, which evened out after 12 months [10]. Most reluctance to perform incisional parotid biopsy is due to the theoretical risk of facial nerve damage, development of sialoceles and salivary fistulae. Thus, incisional parotid gland biopsy is still not part of the pSS classification criteria despite being at least comparable to labial salivary gland biopsy, and validated histopathological criteria are lacking [4].

More recently, ultrasound-guided core needle (US-guided CN) biopsy of the parotid gland has proven to be highly sensitive and specific for lymphoma detection in pSS patients with suspected malignancy [9, 11]. This technique is known to be well tolerated and has a minimal risk of complications. Unlike incisional biopsy, the theoretical risk of vascular and facial nerve damage has not been reported [12, 13]. Although US-guided CN biopsy has become common in the oncological diagnostic approach in pSS, its value in classification and diagnosis remains unexplored.

This proof-of-concept study aimed to explore whether US-guided CN biopsy of the parotid gland has a place in the diagnostic work-up of pSS.

Methods

Study design and participants

Patients were selected from the Belgian SS Transition Trial (BeSSTT), an observational cohort of patients with a definite diagnosis of pSS based on fulfilment of the 2016 ACR/EULAR pSS classification criteria, and patients suspected of the disease due to positivity of at least one criterion. Demographic as well as clinical data and extensive laboratory parameters with autoimmune serology were collected. Autoimmune serology was determined based on a symphony screen (Thermo Fisher Scientific, Waltham, Massachusetts, USA, REF 14–5671–01) allowing sensitive screening of a mixture of nuclear and cytoplasmic antigens. Positive samples were further investigated using a more specific anti-extractable nuclear antigen line blot (Euroimmun, Lübeck, Germany, REF 1590–1601-5 G), where among others anti-SSA/Ro (anti-Ro52 and anti-Ro60), anti-SSB/La, anti-Sm, anti-Cenp-B, anti-Jo-1 and anti-Scl-70 antibodies could separately be detected. Two experienced head and neck radiologists were involved to perform the ultrasonographic evaluations of the salivary glands with a MyLab60 scanner (Esaote, Genova, Italy) with high resolution probe (4–13 MHz). The assessors alternated each other in the clinical carepath and were well aligned, as described by Deroo et al. [14] Images were scored real-time using Hocevar score (0–48) [15].

Combined incisional and US-guided CN parotid biopsy was performed in 20 patients. Parotid biopsies were performed between February 2021 and April 2022. The Ghent University Hospital ethics committee granted approval for this research project (2019/0452). All participants provided written informed consent. Patients were not involved in the study design.

Biopsy technique

Parotid biopsies were taken from the side with most ultrasonographic abnormalities. For the procedure, the ipsilateral facial halve was disinfected and local anaesthesia with xylocaine 1% and adrenaline 1/100 000 was administered. US-guided automatic CN biopsy was performed by an experienced radiologist (E.G.) using an 18-gauge needle with length of 22 mm. In patients with strongly atrophic parotid glands on ultrasonography, an 18-gauge needle with length of 15 mm was used. The entry point for the biopsy needle was pre-auricular. The biopsy was taken in the superficial lobe of the parotid gland, cranio-caudally orientated, and the needle was angled in the most horizontal way. In this way, the deeper lobe with the main trunk facial nerve, the external carotid artery and the retromandibular vein was avoided. In addition, the biopsy needle was directed parallel to the assumed course of the facial nerve. Two biopsies were taken this way and tissue fragments were stored in formalin immediately after sampling (Fig. 1). Pressure was applied on the entry point of the needle for two min. After local disinfection, the incisional biopsy was performed (F.D.) immediately after the CN biopsy. The biopsy was taken ipsilaterally according to the technique described by Kraaijenhagen [16]. An incision of around 1.5 cm was made pre-auricular, just anterior of the tragus, elliptically shaped caudally towards the lobulus of the earlobe. This is comparable to a small part of the lazy-S incision performed for parotidectomy. Progressive dissection was performed to identify and raise the skin, subcutaneous tissue and the superficial cervicofacial flap. N. auricularis was identified and pre-elevated when necessary to preserve it. The anterior border of the sternocleidomastoid muscle, superficial musculoaponeurotic system and posterior border of the superficial parotid gland were identified. An incisional biopsy was taken in the safest region of the parotid gland to prevent damage of the facial nerve. Tissue was stored on a sterile gauze with NaCl 0.9% until further processing. Local haemostasis was obtained and the wound was closed with Monocryl 3.0 subcutaneous and Dermabond cutaneous. No specific measures were taken to avoid the incisional tract of the CN biopsy.

Histopathological evaluation

Tissue fragments were fixed in 10% neutral buffered formalin, embedded in paraffin blocks, cut at a thickness of 4 μm and stained with haematoxylin and eosin (H&E). Immunohistochemistry was performed on these sections with a Benchmark XT immunostainer (Ventana Medical Systems, Tucson, AZ, USA). Primary monoclonal antibodies were CD3 (1:500; clone F7 2.38; Agilent), CD20 (1:500; clone L26; Dako, Agilent), CD21(1:100; clone 1F8; Agilent) and CD138 (1:200; clone MI15; Agilent). Visualization was achieved with Ultraview Universal DAB Detection kit (Ventana Medical Systems, Tucson, AZ, USA). Appropriate positive and negative controls were used throughout. All tissue samples were analysed by one experienced pathologist (D.C.) with specific expertise in head and neck pathology. H&E-stained sections were digitally scanned by a Hamamatsu NanoZoomer digital whole slide scanner. Salivary gland tissue areas, including fat, were identified, morphometrically compartmentalized on a 4× magnification, and quantified in surface mm². A minimum of 4 mm² salivary gland tissue, including fat, was necessary to be considered representative [4]. The number of lymphocytic foci of at least 50 lymphocytes was counted, it was determined whether FS was at least one [4, 17], and the T/B cell ratio was assessed (CD3/CD20; predominately T—equal proportion T and B—predominantly B cells). The H&E sections were further screened for presence of atrophy, LELs (lesions that consist of hyperplastic ductal epithelium infiltrated with lymphocytes) [18, 19] and ectopic germinal centres (relatively well-circumscribed inflammatory infiltrates within the lymphoid infiltrate; they are composed of both lymphoid cells and cells with a non-lymphoid nature, macrophages and follicular dendritic cells) [20]. Immunohistochemistry further confirmed germinal centres based on presence of follicular dendritic cell networks (CD21), and identified plasma cells (CD138). Biopsies were considered positive in case of a FS of at least one, or in case of lymphoid infiltrates with FS less than one in combination with LELs [1, 4]. All samples were screened for malignancy and other diseases.

Morbidity

All patients who underwent paired incisional and US-guided CN parotid biopsy were contacted by phone two weeks after the procedure to complete a questionnaire assessing potential complaints, complications and willingness to repeat the procedure.

Statistical analysis

Study data were collected and managed using REDCap electronic data capture tools hosted at the Ghent University Hospital [21, 22]. Statistical analyses were performed using SPSS (IBM Corp. Released 2021. IBM SPSS Statistics for Windows, Version 28.0. Armonk, NY: IBM Corp). Differences concerning epidemiology and morbidity between the study and control group were assessed using Mann–Whitney U, chi-squared or Fisher’s Exact test as appropriate. P-values ≤0.05 were considered statistically significant.

Results

Twenty consecutive patients underwent paired incisional and US-guided CN biopsy of the parotid gland. Their average (s.d.) age was 47.6 (13.0) years, 18 (90.0%) were female, 15 (75.0%) had anti-SSA/Ro antibodies and 15 (75.0%) fulfilled the ACR-EULAR classification criteria independently of the outcome of the histopathology.

Histopathologic features are extensively described in Supplementary Table S1, available at Rheumatology online. Because parotid biopsy requires a minimum of 4 mm² of salivary gland tissue including fat to be considered representative, we first evaluated this item. Via both techniques we obtained 14 representative samples (Tables 1 and 2). US-guided CN biopsy yielded interpretable salivary gland tissue in an additional four patients, but with an area of less than 4 mm². Remarkably, in the residual six patients incisional biopsy only obtained fibroadipose tissue without salivary gland tissue. By contrast, only two US-guided CN biopsy samples showed solely fibroadipose tissue, of whom one also on incisional biopsy. Overall, salivary gland tissue collected using US-guided CN biopsy was available for interpretation in 18/20 patients, compared with 14/20 patients with incisional biopsy.

We then considered the presence or absence of a positive FS in samples of at least 4 mm² salivary gland tissue. On incisional biopsy, half of the patients (7/14) had a positive FS. This was illustrated by an abundancy of B cells (CD20) compared with T cells (CD3) in all patients, except for one in whom both cell types were present in equal proportions. In comparison, in 6/14 patients US-guided CN biopsies a FS of at least one was seen, again all showing an abundancy of B cells compared with T cells, except for one patient in whom both were seen in equal proportions (Fig. 2).

Next, we evaluated a FS less than one, but with presence of lymphocytic foci of ≥50 lymphocytes. On incisional biopsy, 2/14 patients demonstrated a B-cell predominant infiltrate, one patient had a germinal centre, the other a clear plasma cell infiltrate. In comparison, this was the case in 3/14 patients on US-guided CN biopsy. In one we noticed mixed T cells and B cells, in two there was an abundancy of B cells compared with T cells, while germinal centres were present in all three and a clear plasma cell infiltrate in two.

As LELs are a hallmark of pSS in parotid gland tissue, we assessed its presence in both incisional and US-guided CN biopsies of at least 4 mm² salivary gland tissue. Presence of LELs showed an absolute concordance between incisional and CN biopsy. Of interest, in the small subset of patients where no salivary gland tissue was found on incisional biopsy, one additional patient demonstrated LELs on US-guided CN biopsy. Perfect agreement was also seen for presence of germinal centres when both biopsy types were representative. In addition, three patients without salivary gland tissue on incisional biopsy showed germinal centres on US-guided CN biopsy. The same was true for one incisional biopsy lacking salivary gland tissue on US-guided CN biopsy.

As mentioned above, 4/20 US-guided CN biopsy samples consisted of less than 4 mm² salivary gland tissue. However, one biopsy showed clear histopathological features of pSS illustrated by two foci predominantly consisting of B cells, LELs, germinal centres and a clear plasma cell infiltrate. In the other three biopsies only minimal T-, B- and plasma cell infiltrates were identified.

Collectively, paired incisional and US-guided CN biopsies of at least 4 mm² salivary gland tissue were available in ten patients, of which nine showed the same results (FS <1 in five, FS ≥1 in four patients) and one a discordant result (FS <1 on incisional biopsy vs FS ≥1 on US-guided CN biopsy). However, incisional biopsy in this patient showed two clear foci, but did not reach a FS of one. Presence of LELs and germinal centres showed absolute agreement between incisional and US-guided CN biopsy in all ten patients. Of the three patients with less than 4 mm² salivary gland tissue on US-guided CN biopsy, assessment of FS corresponded in two patients (FS <1 in one, FS ≥1 in one), but not in one (FS <1 on US-guided CN biopsy vs FS ≥ 1 on incisional biopsy) (Fig. 3).

We then investigated the sensitivity of the FS for pSS classification according to the 2016 ACR/EULAR classification in samples with at least 4 mm² salivary gland tissue. Sensitivity for incisional biopsy was 6/10 (60%) and for US-guided CN biopsy 5/11 (45%). When also considering biopsies of less than 4 mm² salivary gland tissue or presence of germinal centres, the sensitivity of incisional biopsy increased to 7/10 (70%) and for US-guided CN biopsy to 9/13 (69%). Of interest, Hocevar scores ware significantly higher in patients with, compared with patients without presence of germinal centres on US-guided CN biopsy [mean (s.d.) 16.7 (12.5) vs 31.2 (11.1); P = 0.024].

Parotid biopsy did not impact the pSS classification of any patient. Nonetheless, one patient could not be classified as pSS according to the 2016 ACR-EULAR criteria, but both biopsies were clearly positive. This patient was clinically diagnosed as pSS based on anti-Cenp-B reactivity, subjective sicca symptoms without Raynaud’s phenomenon, and a Hocevar score of 32/48. Taking this into account increases both sensitivity and specificity.

None of the salivary gland tissue samples obtained by either technique were suspected of malignancy or other pathologies.

Eighteen of the 20 patients completed the post-procedure questionnaire. Haematoma formation was the most prevalent complication (12/18), and a minority of patients reported post-procedural bleeding (2/18). No major complications were reported. The majority of patients were willing to repeat the procedure if necessary (15/18). Because incisional parotid biopsy and US-guided CN biopsy were performed simultaneously, individual data on the burden of US-guided CN biopsy cannot be given.

Discussion

To our knowledge, this is the first report exploring the value of US-guided CN parotid biopsy in the diagnostic work-up of pSS. We demonstrate for the first time that this could be a valuable tool in the diagnostic work-up of pSS and in clinical trial design.

US-guided CN parotid biopsy has already proven to be a reliable instrument in the diagnostic oncological approach in pSS [11–13]. However, its place in pSS diagnosis and classification has not yet been studied. Our results show that US-guided CN biopsy enables collection of sufficient salivary gland tissue to assess lymphocytic infiltrates, LELs and other histopathologic features suggestive of pSS and correlates well with incisional biopsy which is known to be more invasive.

Periductal lymphocytic infiltrates consisting of T and B cells are a hallmark of pSS, and can be seen in both minor and major salivary glands of most patients [23]. pSS classification criteria only take into account foci consisting of at least 50 lymphocytes, with FS being the number of foci per 4 mm² salivary gland tissue [24, 25]. An FS of at least one is considered positive both in incisional parotid and labial salivary gland biopsy, and is an important pSS classification criterion [1, 7]. Assessment of FS of at least one, LEL’s and germinal centres positivity showed an almost perfect agreement between incisional and US-guided CN parotid biopsy. Previous studies already demonstrated comparable sensitivity and specificity of incisional parotid and labial salivary gland biopsy, and transcriptomic analysis showed shared enriched immune pathways in both types of salivary gland tissue [4, 26, 27].

However, there are some drawbacks to the use of FS. Revision by several pathologists often leads to different conclusions [28], progression of fibrotic and atrophic changes in pSS can coincide with extinction of inflammation leading to a ‘burnt-out’ biopsy with a negative FS [29], and most importantly FS is only based on the number and not the size of the foci [30, 31]. In addition, FS is calculated per 4 mm² while it is inherent to the technique of US-guided CN biopsy that a smaller amount of tissue is collected compared with incisional parotid or labial salivary gland biopsy. This way, not all tissue samples obtained via US-guided CN biopsy consisted of the 4 mm² salivary gland tissue required for FS calculation. One of the main concerns is sampling error in case of limited tissue or underestimation, especially in case of a low number of foci [23]. However, this risk can be partially mitigated with US-guided CN biopsy because of ultrasound guidance, as biopsies can be taken in the apparently most affected tissue. Recent research in pSS patients with clinical suspicion for salivary gland malignancy showed that patients with suspicious characteristics on US had an increased probability to be diagnosed with a lymphoma by US-guided CN biopsy [32], and even that in some patients CN biopsies taken from sonographic lesions confirmed MALT lymphoma, while this was not apparent in those taken from surrounding parenchyma [9]. In addition, sampling error might be reduced as US-guided CN biopsy samples salivary gland tissue over a depth of 22 mm, in contrast to incisional parotid biopsy and labial salivary gland biopsy in which sampling is more localized. Moreover, two spatially separated biopsies are taken.

This illustrates the limitations of FS as key to assess US-guided CN biopsy of the parotid gland in the diagnostic work-up of pSS. To implement US-guided CN biopsy on large scale, larger studies should be performed to allow development of a validated set of histopathologic criteria for parotid gland tissue, also taking into consideration other features suggestive of pSS. In this context, LELs, striated ducts infiltrated by lymphocytes with concurrent hyperplasia of the epithelial cells, are more frequently seen in parotid than minor salivary gland tissue. Incisional parotid biopsy is deemed positive not only when FS is at least one, but also when both foci and LELs are detected [4]. However, the sensitivity and specificity of LELs with regard to pSS diagnosis is not known [23]. Also, a relative decrease of IgA-positive plasma cells due to a relative increase in IgG- and IgM-positive plasma cells is related to pSS with a threshold of <70% IgA being highly sensitive and specific for pSS [33]. Combination of FS and percentage of IgA‐containing plasma cells proved to be more accurate than either examination alone [34]. Moreover, the area of infiltration can be evaluated for a better estimation of the inflammation level in the salivary gland biopsy [35]. Additionally, it has been shown that incidence of CD4(+)-T cells, interdigitating dendritic cells and T/B cell ratio were negatively, and of B cells and macrophages positively associated with infiltration grade and FS in labial salivary gland biopsies [36]. In the present study, detection of a focus on US-guided CN biopsy with presence of a germinal centre would substantially increase sensitivity to levels comparable to labial salivary gland biopsy [23] without impacting the specificity. In support thereof, we observed a strinkingly high association between presence of germinal centres and levels of Hocevar scores. Further research is necessary to elucidate how ultrasonography of the salivary glands reflects specific histopathologal features related to pSS [9, 32].

US-guided CN parotid biopsy has comparable features as incisional parotid biopsy: detection of B-cell malignancy which most frequently affects the parotid glands, comparability with other diagnostic tests of the same gland (e.g. salivary gland US) and possibility of repeated biopsies [7]. In addition, US-guided CN biopsy in the hands of an experienced radiologist is minimally invasive and very well tolerated without any reported long-term complications [9, 11, 13, 37]. Moreover, this approach can be easily perfomed in an outpatient setting, reducing societal costs. These major advantages over other techniques to obtain salivary gland tissue may lower the threshold for repeated biopsy of salivary gland tissue. This would not only be of interest in case of inadequate tissue collection, but particularly for follow-up of indolent malignancies or in patients with high malignancy risk. Moreover, US-guided CN biopsy might spark a revolution in clinical trial design in which there is major need of objective biomarkers. Histopathologic assessment of salivary gland response to treatment holds great potential for this purpose [29]. Several studies have shown evidence of treatment response in several histopathologic parameters such as of reduction in CD20+ B cells, LELs and germinal centres after rituximab treatment [30] and reduced presence of GCs, associated with improvement of the glandular domain, after abatacept treatment [38]. At this moment, labial salivary gland biopsy is usually only optional before trial inclusion. Implementation of US-guided CN biopsy could substantially lower the barrier to perform biopsies before and after treatment, and thus provide a rich source of information.

This is the first time that US-guided CN biopsy of the parotid gland was performed paired with incisional biopsy in a population of patients suspected for pSS. The limitations of this study include a relatively small cohort, the lack of labial salivary gland biopsies in the same patients and no information on the burden of US-guided CN biopsy as both biopsies were performed simultaneously. To implement US-guided CN biopsy on a large scale in pSS diagnosis and in assessment of treatment response, larger studies are needed and specific histological criteria have to be developed. In addition, further research performing repeated US-guided CN biopsies of the parotid gland is necessary to assess its tolerability and safety in the context of pSS.

In sum, when performed by an expert, US-guided CN biopsy of the parotid gland is a relatively easy technique that combines the histopathological characteristics and repeatability of incisional parotid biopsy into a less invasive, well-tolerated procedure. As such, US-guided CN biopsy of the parotid gland may become useful in the field of pSS, not only in the clinical work-up, but particularly in clinical trial development.

Supplementary material

Supplementary material is available at Rheumatology online.

Data availability

De-identified data underlying this article are available for researchers who provide a methodologically sound proposal. Data are available up to 5 years after publication. Data may be requested to achieve aims of the approved research proposal or for individual participant data meta-analysis. For further information, please contact the corresponding author.

Contribution statement

I.P., D.E., L.D., E.G., F.D. and D.C. conceptualized and designed the study. All authors collected original patient data. L.D., E.G., F.D., D.C., H.A., I.P., F.VdB. and D.E. contributed to the analysis and interpretation of the data. L.D. prepared the manuscript. All authors critically revised the manuscript for important intellectual content and approved the final version of the manuscript for publication.

Funding

This work was supported by a grant from the Fund for Scientific Research on Rheumatology (2018-J5820590-210858). These financial resources have been used to support data collection and management. There was no involvement of any kind in the study conduction. The corresponding author has full access to all the data and takes the final responsibility for the decision to submit for publication.

Disclosure statement: L.D., E.G., F.D., D.C., H.A., K.DB., W.B. and D.R. report no competing interests. F.VdB. has received consultancy fee honoraria from Abbvie, Amgen, Eli Lilly, Galapagos, Gilead, Janssen, Merck, Moonlake, Novartis, Pfizer and UCB. D.E. has received consultancy fee honoraria from Abbvie, Amgen, Eli Lilly, Galapagos, Gilead, Janssen, Merck, Novartis, Pfizer and UCB. I.P. has received consultancy and/or speaker fee honoraria from Amgen, Argenx, Galapagos and Novartis; and participated in advisory boards for Astra-Zeneca, Amgen, GSK, Argenx, BMS, Galapagos, UCB and Novartis.

Acknowledgements

We thank all the participants, clinicians and others who have contributed to this study.

References

Author notes