Increased gene copy number of HER2 and concordant protein overexpression found in a subset of eyelid sebaceous gland carcinoma indicate HER2 as a potential therapeutic target

Abstract

Purpose

To identify crucial molecular alterations of receptor tyrosine kinases that can be used as potential therapeutic targets for eyelid sebaceous gland carcinoma (SbGC).

Methods

The expression levels of HER2, EGFR, C-MET, and FGFR1 were determined by immunohistochemistry (IHC). The copy numbers of the HER2, EGFR, C-MET, and FGFR1 genes were assessed by fluorescence in situ hybridization. The IHC and molecular results were correlated with the clinical parameters.

Results

A total of 49 patients with eyelid SbGC were included in this study. HER2, EGFR, C-MET, and FGFR1 protein expression was detected in 8 of 44 (16.3 %), 8 of 45 (17.8 %), 3 of 35 (8.6 %), and 0 of 45 patient samples, respectively. Increased copy numbers of the HER2 gene were found in 5 of 42 patient samples (11.9 %), including two with amplification (4.7 %) and three with polysomy (7.2 %). EGFR amplification was found in 2 of 33 (6.1 %) and FGFR1 amplification in 4 of 33 patient samples (12.1 %; high-level amplification in one and low-level amplification in three). None of the samples examined exhibited C-MET amplification. Gene copy number of the HER2 gene was correlated with its protein expression (p < 0.0001), whereas copy number of EGFR, C-MET, or FGFR1 was not correlated with protein expression. However, samples with EGFR amplification also exhibited a high level of expression of this protein.

Conclusions

Extra copies of the HER2, EGFR, and FGFR1 genes were identified in a 6–12 % of eyelid SbGCs. A high level of concordant HER2 expression detected by immunohistochemistry can be predictive of a copy number gain of the HER2 gene. Our data suggest that the therapeutic targeting of HER2 might benefit for a subset of patients with periocular SbGCs.

Electronic supplementary material

The online version of this article (doi:10.1007/s00432-015-2009-z) contains supplementary material, which is available to authorized users.

Introduction

The treatment outcomes of cancer patients have remarkably improved since the introduction of personalized medicine based on the molecular alterations of major target genes. However, little progress has been made in the targeted therapy of sebaceous gland carcinoma (SbGC) because of limited knowledge about its molecular pathogenesis. It is a rare, potentially aggressive cutaneous neoplasm that predominantly affects the periocular region, which is densely populated by meibomian glands, particularly in the tarsus (Buitrago and Joseph 2008; Shields et al. 2005). Although the mortality rate of this carcinoma is relatively low, local invasion and metastasis to regional lymph nodes or distant organs occasionally occur in aggressive cases (Shields et al. 2005; Song et al. 2008). The recurrence rate of SbGC has been reported to be up to 10–25 % (Shields et al. 2004; Song et al. 2008; Yoon et al. 2007). The standard frontline treatment of eyelid SbGC is either wide excision or Mohs micrographic surgery; however, the tendency for multicentricity or pagetoid spread remains the main obstacle preventing its complete removal. Moreover, the selection of an optimal treatment modality is not often straightforward due to the risk of severe facial disfigurement. Radiotherapy or chemotherapy plays a limited role as an adjunctive or palliative treatment (Shields et al. 2005).

Receptor tyrosine kinases (RTKs) are key regulators of carcinogenesis and are classified into several families, most notably the human epidermal growth factor receptor 2 (HER2), epidermal growth factor receptor (EGFR), fibroblast growth factor receptor (FGFR), vascular endothelial growth factor (VEGF), and hepatocyte growth factor (HGF) families (Janku et al. 2011; Santarpia et al. 2012; Schmitz et al. 2014). Deregulation of these molecules is not only involved in the pathogeneses of many major epithelial carcinomas, but it also greatly affects patient outcome. Targeting of several RTKs is a widely adapted therapeutic modality for some carcinomas. To select the optimal cases for targeting RTKs, the determination of molecular alterations and protein expression levels is crucial. Notably, HER2 amplification occurs in subsets of breast and gastric carcinomas (Baselga 2010). A mutation or gene amplification of EGFR, along with its high protein expression level, has been found in various tumours, including non-small cell lung cancer (NSCLC), head and neck squamous cell carcinoma, and colorectal cancer (Masuda et al. 2012; Mok et al. 2014; Schmitz et al. 2014). In addition, gene amplification of FGFR1 or C-MET has been observed in various cancers, including NSCLC, and has been found to negatively impact patient survival (Choe et al. 2012; Kim et al. 2013; Lehnen et al. 2013). However, there have been few studies of eyelid SbGC describing the expression levels or molecular alterations of RTKs (Cho et al. 2000; Erovic et al. 2012; Ivan et al. 2010). Overexpression of EGFR has been observed in some cases of cutaneous and conjunctival squamous cell carcinomas, and clinical trials using tyrosine kinase inhibitors (TKIs) or anti-EGFR antibodies are still ongoing (Lewis et al. 2012; Shepler et al. 2006; Yin et al. 2013).

The purpose of this study was to investigate genetic abnormalities of major RTKs in eyelid SbGC. The rationale of target gene selection was based on both functional significance and prevalence in the carcinoma, as well as the availability of targeting agents that are currently approved or under investigation.

Materials and methods

Patients

Forty-nine patients with eyelid SbGCs who were diagnosed and treated between 1999 and 2013 were included in this study, based upon tissue availability. An experienced pathologist confirmed the pathologic diagnosis, and clinical data were collected from electronic medical records. The clinical data included demographic information, histopathological diagnosis, TNM staging, treatment details, and outcomes, such as lymph node or distant metastasis and survival. The TNM staging was re-established for all of the cases according to the 2010 American Joint Committee on Cancer (AJCC; 7th edition) (Ainbinder et al. 2009). None of the patients had received previous treatment. This study was approved by the Institutional Review Board of the Seoul Municipal Government-Seoul National University Boramae Hospital.

Immunohistochemistry: HER2, EGFR, C-MET, and FGFR1

Representative paraffin blocks were cut into 4-μm-thick serial sections and mounted on glass slides. Immunohistochemistry (IHC) was performed using anti-HER2 (DAKO, Carpinteria, CA, USA), anti-EGFR (Ventana Medical Systems, Tucson, AZ), anti-C-MET (Ventana Medical Systems), and anti-FGFR1 (Epitomics, Burlingame, CA, USA) antibodies according to the manufacturer’s recommended protocols. After deparaffinization and rehydration, heat-induced epitope retrieval was conducted using the Ventana CC1 mild reagent (Ventana) for 60 min. After epitope retrieval, the slides were treated with 0.3 % hydrogen peroxide for 10 min at room temperature to block endogenous peroxidase activity. Next, the slides were treated with 10 % normal goat serum to block nonspecific antibody binding and then incubated with antibodies against HER2, EGFR, C-MET, and FGFR1 followed by incubation with a horseradish peroxidase-conjugated multimer antibody reagent (Ventana), using an automated immunostainer (Ventana Bench Mark XT). The immunoreaction was developed with diaminobenzidine tetrachloride for 5 min, and the slides were counterstained with haematoxylin. Positive and negative controls were included. HER2 was scored according to the American Society of Clinical Oncology (ASCO)/College of American Pathologists (CAP) guidelines (Wolff et al. 2013) as follows: 0 (no staining is observed in invasive tumour cells), 1+ (incomplete and faint membrane staining in more than 10 % of tumour cells), 2+ (circumferential membrane staining that is incomplete or not intense found in at least 10 % of cells, or complete membrane staining that is intense in ≤10 % of tumour cells), or 3+ (Uniform intense membrane staining of >10 % of invasive tumour cells). As defined in the guideline, scores of 0 or 1+ were considered negative, whereas a score of 2+ was equivocal, and a score of 3+ was positive. For the interpretation of EGFR, C-MET, and FGFR1 expression, only complete membranous staining of at least 10 % of the tumour cells was counted as positive, and the slides were scored as 0, 1+, 2+, or 3+ according to the intensity. Nuclear staining and faint cytoplasmic staining were scored as negative (Lehnen et al. 2013).

In situ hybridization: HER2, EGFR, C-MET, and FGFR1

Fluorescence in situ hybridization (FISH) was performed to determine gene copy number status of HER2, EGFR, C-MET, and FGFR1. Used FISH probes are as follows: PathVysion HER-2 DNA probe kit (Abbott Molecular), LSI EGFR Spectrum Orange/CEP 7 Spectrum Green probe, the Vysis Path Vysion (Abbott Molecular, Abbott Park, IL, USA), MET/CEP7 dual-colour probe (Vysis LSI D7S522 Spectrum Orange/CEP7 Spectrum Green Probes), and FGFR1 (8p11)/SE8 Amplification probe (Kreatech Diagnostics, Amsterdam, Netherland). After the pretreatment procedure, the DNA probe kit was applied to the slides and incubated in humidified air at 73 °C for 5 min to denature the target DNA and probes and then subsequently at 37 °C for 19 h to achieve hybridization. After washing, the slides were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) and p-phenylenediamine.

Also silver in situ hybridization (SISH) was done for reconfirming HER2 amplification status for the INFORM HER2 DNA/chromosome 17 (Roche) probe using an automatic immunostainer (Bench Mark XT; Ventana Medical Systems).Interpretation of both FISH and CISH was performed according to the criteria of the ASCO/CAP guideline for HER2 (Wolff et al. 2013), the University of Colorado Cancer Center (UCCC) criteria for EGFR FISH assay (Varella-Garcia et al. 2009), a reading system suggested by Schildhaus et al. (2012) for FGFR1 and the Cappuzzo scoring system for the MET gene (Cappuzzo et al. 2009).

Statistical analysis

Fisher’s exact test was used to compare nominal variables, and the Mann–Whitney U test was used to compare continuous variables without normality and ordinal variables. The Spearman correlation was used to determine the correlation between IHC and FISH. Statistical analyses were performed using SPSS software version 20.0 (SPSS Inc., IBM, NY, USA). p values <0.05 were considered statistically significant.

Results

Patient demographics

This study included 49 patients with eyelid SbGC with a mean age of 62.4 ± 13.8 years (range, 37–93 years) and a male–female ratio of 15:34. As determined by AJCC staging, 35 patients had T2N0 tumours, 12 had T3N0 tumours, one possessed a T2N1 tumour, and another one had a T3N1 tumour at initial diagnosis. All of the patients received wide surgical excision as the first treatment, and four patients required exenteration. Most of the patients (47 of 49, 95.9 %) showed classic histologic features of sebaceous gland carcinoma, including lobular or solid growth of tumour cells with foamy, vacuolated cytoplasm. Anaplastic or undifferentiated features were found in two cases (2 of 49, 4.1 %; Fig. 1). The detailed clinical features, immunohistochemistry results, and molecular results for all patients are described in Table 1 and also Supplementary Table 1. The mean follow-up period was 25.6 ± 22.5 months (range, 1–94 months). Lymph node metastasis was detected in eight patients at the time of treatment or during follow-up, and lung metastasis was noted in one patient during follow-up. There were no significant differences regarding sex (p = 0.702, Fisher’s exact test) or follow-up periods (p = 0.144, Mann–Whitney U test) between the metastasis and non-metastasis groups. However, the patients with metastasis were significantly younger than those without metastasis (mean age 53.8 ± 16.8 vs. 64.3 ± 12.5, p = 0.023, Mann–Whitney U test).

Fig. 1.

Histologic features of sebaceous gland carcinoma. Lobular aggregation of tumour cells showing numerous, vacuolated cytoplasm is observed in the classic type (a). The anaplastic variant is characterized by marked pleomorphism and brisk mitotic figures (b). (Haematoxylin–Eosin, ×400)

Table 1.

Clinical profiles of patients with eyelid sebaceous gland carcinoma (n = 49)

Variables	No. of cases (%)
Age, years (median)
37–93 (62)	–
Sex
Male	15 (30.6)
Stage
T2N0	35 (71.4)
T3N0	12 (24.5)
T2N1	1 (2.0)
T3N1	1 (2.0)
Initial treatment
Wide excision and eyelid reconstruction	45 (91.8)
Exenteration	4 (8.2)
Progression
Local recurrence	4 (8.2)
Lymph node metastasis	6 (12.2)
Distant metastasis	1 (2.0)
Follow-up, months (median)
1–98 (20)	–
Dead/alive/missed	2/45/2
Histologic variants
Classic	47 (95.9)
Undifferentiated	2 (4.1)

Immunohistochemistry for HER2, EGFR, C-MET, and FGFR1

As previously reported, EGFR immunoreactivity is observed in normal skin within epidermal basal cells and to a lesser extent within suprabasal keratinocytes (Kodama et al. 1995). EGFR expression is also observed within the eccrine ducts and cells of the outer root sheaths of hair follicles. Basal and germinative cells at the periphery of the sebaceous glands have been found to exhibit increased EGFR expression. However, no HER2, C-MET, or FGFR1 expression was observed in the normal eyelid skin or sebaceous glands. Of the 49 SbGC samples tested, grade 3+ HER2 immunoreactivity according to the ASCO/CAP guidelines was found in two (4.1 %) samples (Fig. 2). Two samples (4.1 %) showed grade 2+, and four (8.2 %) showed grade 1+ staining. In particular, intratumoural heterogeneity was prominent in the expression of HER2. EGFR was positively expressed in 8 of 45 SbGC samples, 4 (8.9 %) of which showed grade 3+ staining and 4 (8.9 %) of which showed grade 2+ staining. C-MET expression was detected in 3 of 35 carcinoma samples (8.6 %), with only one showing grade 2+ positivity, while another two exhibited a faint focal reaction. Membranous expression of FGFR1 was not found in any of the carcinoma samples tested, although focal to diffuse cytoplasmic positivity was observed in a few, which was considered to indicate negativity.

Fig. 2.

EGFR protein expression was found in the normal eyelid sebaceous gland (a), whereas no expression of HER2, C-MET, or FGFR1 was found. In eyelid sebaceous gland carcinoma, positive immunoreactivity to HER2 (b), EGFR (c), and C-MET (d) is found in a subset of tumour cells. FGFR1 immunohistochemistry revealed weak and focal cytoplasmic reaction in some tumour cells, which were considered as a false positivity (e). (all, ×400)

In situ hybridization for HER2, EGFR, C-MET, and FGFR1

HER2 amplification was detected in 2 of the 42 samples (4.7 %), and increased CEP17 gene indicating polysomy 17 was found in three (7.2 %; Fig. 3). The concordance rate between the FISH and SISH results for the HER2 gene was 100 %. Intratumoural heterogeneity was more easily identified by the SISH method (Fig. 4). Amplification of EGFR was found in 2 of 33 carcinoma samples (6.1 %), with one showing high-level amplification and the other exhibiting low-level amplification (high polysomy). C-MET amplification was not observed in any of the samples examined in this study. With regard to FGFR1, 4 of the 33 samples (12.1 %) exhibited gene amplification, with high-level amplification in one and low-level amplification in three. One sample with a low level of EGFR amplification simultaneously harboured a low level of FGFR1 amplification. The detailed clinical features, immunohistochemistry results, and molecular results of cases with gene extra copies are summarized in Table 2.

Fig. 3.

Dual-colour FISH analysis of HER2, EGFR, C-MET, and FGFR1 in eyelid sebaceous gland carcinoma. Representative images showing HER2 (a) and EGFR (b) amplification. c Tumour cells showed disomy pattern for C-MET. d FGFR1 signal patterns showed low polysomy

Fig. 4.

Immunohistochemistry and silver in situ hybridization (SISH) of HER2 in eyelid sebaceous gland carcinoma. Strong HER immunopositivity (grade 3+) was found; intratumoural heterogeneity was prominent (×400) (a). HER2 gene amplification, as demonstrated by SISH. The copy number of HER2 (black signals) exceeded that of CEP17 (red), with a ratio of over 2.2 (arrow; b). However, intratumoural heterogeneity was also evident, with adjacent tumour cells exhibiting disomy (SISH, ×1000)

Table 2.

Characteristics of patients with extracopies of HER2, EGFR, or FGFR1

Case no.	Age (years)	Sex	Stage	FU (mo)	Mets	HER2 IHC	HER2 FISH	EGFR IHC	EGFR FISH	C-MET IHC	C-MET FISH	FGFR1 IHC	FGFR1 FISH
4	63	M	T3N0	50	N	1	Poly	0	Di	NA	NA	NA	Di
23	84	M	T2N0	3	N	3	Amp	1	Di	0	Di	0	Di
24	78	F	T2N0	8	N	2	Poly	NA	NA	NA	NA	0	NA
27	58	F	T2N0	30	N	2	Poly	1	Di	0	NA	0	Di
37	68	F	T2N0	15	N	3	Amp	1	Di	0	Di	0	Di
39	65	F	T3N0	7	N	0	Di	3	High amp	0	Di	0	Di
3	65	F	T2N0	2	N	0	Di	3	Low amp	0	Di	0	Low amp
7	88	M	T2N0	8	N	0	Di	1	Di	0	NA	0	Low amp
14	58	M	T2N0	17	N	0	Di	2	Di	0	NA	NA	Low amp
30	42	M	T2N0	9	N	1	Di	0	Di	0	NA	0	High amp

Amp amplification, Di Disomy, F female, M male, Mets metastasis, NA not available, Poly polysomy

Correlation between protein expression and gene copy number

HER2 gene ISH status significantly correlated with the protein expression level demonstrated by IHC (p < 0.0001, Spearman’s coefficient = 0.804). Both samples with grade 3+ immunoreactivity also had HER2 gene amplification, and both samples with grade 2+ immunoreactivity had CEP17 polysomy. Only one sample showed both CEP polysomy and grade 1+ HER2 immunoreactivity by IHC, whereas the other five had disomy (Table 3). EGFR protein expression was not associated with any genetic alteration, such as an amplification or mutation. However, two samples with EGFR amplification detected by FISH showed strong EGFR immunopositivity. Likewise, there was no correlation between protein expression and gene copy number status for C-MET or FGFR1.

Table 3.

HER2 in situ hybridization results according to protein expression score in eyelid sebaceous gland carcinoma

	Protein expression				p a
	+3	+2	+1	+0
ISH (n = 49)
Gain of CN	2 amp	2 poly	1 poly	0	<0.001
Disomy	0	0	3	41
Total	2	2	4	41

Amp amplification, CN copy number, ISH in situ hybridization, Poly polysomy

^aSpearman’s correlation analysis

Clinicopathologic correlation of immunohistochemistry and FISH results

Univariate analyses of clinical features, including the development of metastasis, did not reveal any statistical correlation with gene copy number changes and protein expression of HER2, EGFR, or C-MET (Supplementary Table 2).

Discussion

Our study is the first to investigate alterations in the major RTKs involved in eyelid SbGC. In this study, we observed extra copies of HER 2, EGFR, and FGFR1 in a subset of eyelid SbGCs (6–12 %) and HER2 protein overexpression is predictive of an increased gene copy number. However, there is intratumoural heterogeneity of both protein and genetic levels, warranting caution in the interpretation of both IHC and FISH results.

Four patient tissue samples (8.2 %) exhibited strong HER2 expression with staining of over grade 2 in our study. The results of previous studies have been controversial in terms of the rate of positive HER2 expression in eyelid SbGC, and it has been reported to range from 0 to 85.7 % (Cabral et al. 2006; Cho et al. 2000; Kwon et al. 2014). We found a copy number gain of the HER2 gene in approximately 12 % of patients with eyelid SbGC in this study. Our results contradict the recent paper which reported much higher HER2 protein expression (12 of 14 patients) and gene amplification (8 of 14 patients) in eyelid SbGC (Kwon et al. 2014). Although we cannot clearly explain the reasons for these differences in results, the most probable factors are intratumoural heterogeneity and differing numbers of enrolled patients. We performed SISH to reconfirm the FISH results, showing identical results. As shown in Fig. 4, a variegated pattern of gene copy numbers even in the same tumour area could lead to remarkable interobserver variability. In breast carcinoma, the HER2 gene is amplified in approximately 15 % of patients, which is a similar incidence to that which was found for eyelid SbGC in our study.

HER2 amplifications considered a poor prognostic indicator. However, in terms of a therapeutic approach, the use of trastuzumab, a specific chemotherapeutic agent targeting HER2, dramatically improves disease-free survival and overall survival in HER2-positive patients with breast carcinoma (Li and Li 2013). Because there is no consensus with regard to the selection criteria for HER2-targeted therapy in eyelid SbGC, the ASCO/CAP guidelines for breast cancer may represent an alternate or a reference (Wolff et al. 2013). Breast carcinoma and SbGC share several characteristics, such as the histologic features, biological behaviours, and ontogenic proximities. Both have skin adnexal origins because the mammary gland is a type of modified sweat gland. Their histopathological features are partially similar, especially for the undifferentiated type of SbGC, characterized by comedo-like necrosis and sometimes by the pagetoid spread of tumours. In addition, a significant portion of eyelid SbGCs harbours sex hormone receptors, such as those for oestrogen and progesterone (Cho et al. 2000). These findings indicate that the molecular pathogeneses of these two cancers are common. At least a subset of SbGCs showing high HER2 protein expression or copy number gain can be candidates for trastuzumab therapy. In our study, the IHC and FISH results were 100 % concordant for the patient tissue samples with high HER2 expression (positivity of over grade 2); however, IHC cannot predict gene copy number status in cases of grade 1 positivity. Possible causes of this discrepancy are thought to include intratumoural heterogeneity, polysomy 17, and low-level amplification (Ruschoff et al. 2012; Yoshida et al. 2014). We found polysomy 17 in three cases (two showing 2+ and one showing 1+ positivity by IHC), and intratumoural heterogeneity was more frequent in the polysomy cases compared with those in which HER2 was obviously amplified. The clinical significance of polysomy 17 in eyelid SbGC should be determined in future studies. In breast cancer, CEP 17 polysomy has also been associated with a good therapeutic response to trastuzumab (Hanna et al. 2014).

Targeted therapy against the major RTKs is now widely applied for many carcinomas. In EGFR-targeted therapy, the selection of patients who will likely achieve a good therapeutic response requires careful consideration based on the assessment of molecular, such as EGFR-activating mutations, genes with increased copy numbers, and overexpressed proteins. In this study, amplification of EGFR was detected in approximately 6 % of the samples tested. We also examined EGFR mutation using PNA Clamp mutation detection (Panagene, Daejeon, Korea) and failed to detect mutations in all 35 cases tested (data not shown). One previous report also described EGFR expression without detectable mutations in periocular and extraocular SbGCs (Ivan et al. 2010). Therefore, EGFR mutation seems a rare event in particular periocular SbGCs. Several studies have suggested that an increased copy number of EGFR is a potential predictor of a better treatment response to EGFR TKIs in NSCLC and is also a poor prognostic indicator in the absence of TKI therapy (Bethune et al. 2010). However, we did not find any clinical significance of EGFR amplification because of the limited number of positive cases.

Despite the low incidence of the amplification of EGFR gene, its expression was frequently observed in eyelid SbGC. There have been limited reports of EGFR protein expression in eyelid SbGC. One previous report has demonstrated that periocular tumours exhibit markedly lower expression compared with extraocular tumours in terms of distribution (16 vs. 88 %) and intensity (21 vs. 77 %) (Ivan et al. 2010). In one report with 20 cases of head and neck SbGCs, 76 % of those cases exhibited high EGFR expression without its prognostic significance (Erovic et al. 2012). In general, EGFR protein expression data do not provide rationale for EGFR TKI application not only because of their discordance with molecular alterations but also because of the disappointing results obtained by clinical trials (Keedy et al. 2011). Our study also showed a lack of correlation between EGFR expression and patient outcome. The prognostic impact of EGFR amplification in eyelid SbGC, therefore, is uncertain at present.

Amplification of C-MET, one of the mechanisms responsible for acquired resistance to EGFR TKIs, was not detected in our study, although some cases exhibited the weak or moderately strong expression of this protein (Ainbinder et al. 2009). This type of discrepancy has been previously reported, and other mechanisms of C-MET regulation, such as ligand-mediated paracrine stimulation, have been suggested in eyelid SbGC (Choe et al. 2012; Organ and Tsao 2011).

FGFR inhibitors demonstrated marked cell growth inhibition and suppression of clonogenicity in the non-small cell lung cancer cell line (Ren et al. 2013). We found the amplification of this gene in approximately 12 % of the patients with eyelid SbGC, at a low level. The prognostic impact of low-level amplification is not as convincing as that of high-level amplification (Schildhaus et al. 2012). The four patient samples with FGFR amplification did not show the overexpression of this protein, and such a discrepancy has been observed in other tumour types, although the reason is unknown (Lehnen et al. 2013; Thomas et al. 2014).

In conclusion, among four receptor tyrosine kinases (HER2, EGFR, C-MET, and FGFR1) studied, we observed significant correlation in the extra copies of HER2 gene and concordant protein overexpression, in particular with HER2 gene amplification, in a subset of eyelid SbGCs. These results suggest that the current therapy targeting HER2 might be beneficial for eyelid SbGC patients with HER2 amplification, which can be detected easily by a simple immunohistochemistry.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Compliance with Ethical Standards

Conflict of interest

We declare that we have no conflict of interest.

Ethical approval

All procedures performed in this study were in accordance with the 1964 Helsinki Declaration and its later amendments. The written informed consent was waived due to the study’s retrospective nature, and the data were analysed anonymously. The protocol for this study was approved by Institutional Review Board of the Seoul Municipal Government-Seoul National University Boramae Hospital.