Clinicopathologic Features of Eyelid Sebaceous Gland Carcinoma Requiring Immunohistochemical Diagnosis

Abstract

Introduction

The aim of the study was to evaluate the clinicopathological features of eyelid sebaceous gland carcinoma (SGC), which requires immunohistochemical examination for a definitive diagnosis.

Methods

Twenty-seven patients with a final diagnosis of eyelid SGC at Toyama University Hospital between April 2016 and April 2022 were retrospectively studied. In cases with a strong clinical suspicion of SGC, if the initial pathological diagnosis by hematoxylin-eosin staining was non-SGC, additional detailed pathology was performed, including immunostaining for adipophilin (ADP) and androgen receptor (AR).

Results

Five patients (18.5%) had a diagnosis other than SGC, including three with squamous cell carcinoma (SCC), one with basal cell carcinoma, and one with Bowen disease. In these 5 cases, detailed pathology, including immunostaining for ADP and AR, was performed again, which ultimately led to the diagnosis of SGC. ADP was positive in all 5 cases, and AR was positive in 4 cases. The 3 patients diagnosed with SCC were characterized by a high Ki-67 index, active mitosis, and relatively low differentiation.

Conclusion

SGC can be pathologically diagnosed in other cancers, such as SCC and BCC. When SCC was diagnosed, it was often hypo-differentiated and required more attention. Immunostaining for ADP and AR is invaluable for confirming SGC diagnosis.

Introduction

Eyelid sebaceous gland carcinoma (SGC) is a major eyelid malignancy arising from the meibomian gland or Zeiss gland [1]. It is the most common eyelid malignancy, accounting for 40–50% of all eyelid malignancies in Asia, including Japan [2, 3]. The basic treatment for eyelid SGC is adequate surgical excision, often with a 4–5 mm safety margin [1, 4–6]. Nevertheless, local recurrence has been reported in 6–24% of cases and lymph node metastasis in 8–21% of cases, with a relatively high grade of malignancy [5, 7–9]. Tumor size has been reported to be an important poor prognostic factor in eyelid SGC [9–11], requiring early diagnosis and treatment. However, eyelid SGC is commonly misdiagnosed as another condition. It has been reported that the diagnosis of a benign disease such as chalazion, blepharitis, or blepharoconjunctivitis can delay treatment [5, 12, 13]. Another problem is the difficulty in making an accurate pathological diagnosis. It has been reported that poorly differentiated SGC can be misdiagnosed as squamous cell carcinoma (SCC) or basal cell carcinoma (BCC) because sebaceous differentiation is difficult to determine, and squamous metaplasia can also occur [12, 14–16]. Accurate pathological diagnosis is important because it can affect preoperative and postoperative treatment strategies. In this study, we report the clinicopathologic characteristics of a case that was diagnosed as other than SGC on initial pathologic examination with hematoxylin and eosin (HE) staining but was finally diagnosed as SGC by immunohistochemical techniques.

Methods

Subject

Twenty-seven patients diagnosed with eyelid SGC at the University of Toyama Hospital between April 2016 and April 2022 were retrospectively studied. A single ocular oncologist made a clinical diagnosis (TY). In cases where eyelid SGC was strongly suspected, tumor resection and pathological examination were performed. For cases in which a pathology other than SGC was diagnosed during the initial HE-stained pathological examination, additional detailed pathological examinations, including immunohistochemical staining, were requested. The patients in this study were evaluated for age, sex, location, tumor type, tumor size, TNM classification, and presence of recurrence or metastasis. The TNM classification was determined according to the American Joint Committee on Cancer (AJCC) 8th edition [17]. Follow-up positron emission tomography/computed tomography or contrast-enhanced computed tomography was performed to identify metastases after surgery.

Pathological Examination and Immunohistochemistry

The tissues were fixed using 4% paraformaldehyde in phosphate-buffered saline. Subsequently, paraffin-embedded tissues were stained with HE and immunohistochemically examined by the Pathology Institute Corporation (Toyama, Japan). SGCs are classified into three groups based on differentiation: good, moderate, and poor [6]. Mitosis was classified as 0–1 per high-power microscopy field (HPF), 2–5 per HPF, or >5 per HPF [14].

The primary antibodies utilized in this study included mouse monoclonal anti-androgen receptor (clone AR441) (Mob245; 1:100, Diagnostic Biosystems, Inc., Pleasanton, CA, USA), mouse monoclonal anti-Ki-67 (clone MIB-1) (1:100 dilution, Agilent Technologies, Inc., Santa Clara, CA, USA), and mouse monoclonal anti-Perilipin-2/Adipophilin (clone ADFP) (NB-110-40878; 1:200 dilution, Novus Biologicals, LLC, CO, USA). Immunohistochemistry was performed using Leica BOND III Automation (Leica #DS9800) and the BOND Polymer Refine Detection System Kit (Leica Biosystems, Bannockburn, IL, USA). The protocol included in situ deparaffinization and high pH epitope retrieval for 40 min, primary antibody incubation for 20 min, polymer incubation for 8 min, and diaminobenzidine (DAB) as the chromogen for 10 min, followed by a 5-min hematoxylin counterstaining step. Positivity for AR, Ki-67, and ADP was measured by immunohistochemistry.

Statistical Analysis

Differences in patient age, duration from the appearance of symptoms to referral to our hospital, follow-up period after surgery, and Ki-67 index between the groups were analyzed using the Mann-Whitney U test. Comparisons of other parameters between the groups (sex, location, T category, tumor type, presence or absence of pagetoid spread, initial diagnosis at other clinics, presence or absence of metastasis) were analyzed via the χ² test or Fisher’s exact test. All statistical analyses were performed using the JMP 16 software (SAS Institute, Cary, NC, USA).

Results

The mean age of all 27 patients was 76.5 ± 11.1 years (9 males and 18 females). The sites were the lower eyelids in 8 cases, the upper eyelids in 16 cases, and the upper and lower eyelids in 3 cases. The initial pathology results by HE staining were SGC in 22 cases (81.5%) and other malignancies in 5 cases (18.5%). The details of these 5 cases are presented in Table 1. The male to female ratio of the 5 cases was 2:3, with an age range of 38–89 years. There were 3 cases of SCC, 1 case of BCC, and 1 case of Bowen’s disease. The clinical diagnosis at another hospital was suspected malignancy in 3 cases and blepharitis or keratitis in 2 cases. Immunohistochemical staining for ADP and AR and detailed pathology were performed for the 5 cases that resulted in a diagnosis other than SGC. Despite the clinical suspicion of SGC in these 5 cases, sebaceous gland differentiation could not be adequately determined by HE staining alone. In such instances, we concluded that positive immunostaining for ADP and AR highly likely indicated sebaceous gland differentiation. All 5 cases were positive for ADP, and AR was positive in 4 cases. In 5 cases, ADP positivity indicated the presence of fat droplets, and the degree to which the fat droplets showed foamy cytoplasm and scalloping nuclei led to the final diagnosis of SGC, which was determined to have differentiated into sebaceous glands. The anterior ocular findings of patient 1 showed diffuse keratoconjunctivitis-like and blepharitis-like tumor infiltrates (Fig. 1a). HE staining revealed substantial tumor growth, with prominent immature tumor cells and numerous mitotic figures. Atypical cells with darkly stained nuclear chromatin formed a solid alveolar tumor, and the initial pathological diagnosis was SCC. (Fig. 1b). Additional immunohistochemical examination revealed positive ADP in some proliferating foci and positive images in a few scattered foamy cells (Fig. 1c). Similarly, AR was also positive (Fig. 1d). Positive ADP and AR results suggest the presence of fat droplets. Fat droplets are thought to cause the cytoplasm to become foamy and the nucleus to be pushed into a scalloped shape. Therefore, we concluded that there was differentiation in the sebaceous glands, which ultimately led to the diagnosis of SGC. Next, a diffuse tumor lesion resembling blepharitis was observed in patient 3 (Fig. 2a). HE staining revealed intraepidermal infiltration of atypical cells coated with irregularly thickened squamous epithelium. The initial pathological diagnosis was Bowen’s disease (SCC in situ) (Fig. 2b). Additional immunohistochemical examination was positive for ADP (Fig. 2c) but negative for AR (Fig. 2d). Although AR was negative, the presence of an ADP-positive pagetoid spread within the epidermis and atypical cells in the bright endoplasmic reticulum led to a final diagnosis of SGC. The 3 patients diagnosed with SCC (patients 1, 4, and 5) were characterized by a higher Ki-67 index, more prolific mitosis, and relatively low differentiation than the other 2 patients.

Table 1.

Clinicopathologic characteristics of cases with the initial pathology diagnosis other than SGC

	Patient 1	Patient 2	Patient 3	Patient 4	Patient 5
Site	L upper and lower	L upper	L upper and lower	R upper	R lower
Initial diagnosis at other clinic	Blepharitis, keratoconjunctivitis	Suspected malignant tumor	Blepharitis	Suspected malignant tumor	Suspected malignant tumor
Metastasis	+	−	−	−	−
Pagetoid spread	+	−	+	−	+
Ki67 index	64.8	46.6	40.6	73.2	60.2
Differentiation	Poor	Good	Good	Moderate	Moderate
Mitosis	>5/HPF	0–1/HPF	0–1/HPF	>5/HPF	2–8/HPF
Initial pathological diagnosis	SCC	BCC	Bowen’s disease	SCC	SCC
Immunostaining	AR (+)	AR (+)	AR (−)	AR (+)	AR (+)
	ADP (+)	ADP (+)	ADP (+)	ADP (+)	ADP (+)

SGC, sebaceous gland carcinoma; M, male; F, female; R, right; L, left; HPF, high power field; SCC, squamous cell carcinoma; BCC, basal cell carcinoma; AR, androgen receptor; ADP, adipophilin.

Fig. 1.

Slit-lamp and pathology images of patient 1. a Slit-lamp microphotograph of SGCs invading the eyelid and cornea. b HE staining (original magnification, ×40). c Immunohistochemical analysis of ADP. d Immunohistochemical analysis of AR (original magnification, ×40).

Fig. 2.

Slit-lamp and pathology images of patient 3. a Slit-lamp microphotograph of diffusely invasive SGCs of the eyelid. b HE staining (original magnification, ×40). c Immunohistochemical analysis of ADP. d Immunohistochemical analysis of AR (original magnification, ×40).

The patients were divided into two groups, one with an initial pathological diagnosis of SGC (group A) and the other with a diagnosis other than SGC (group B), and their clinicopathological characteristics were compared (Table 2). There were no significant differences in sex, location, T classification, presence or absence of pagetoid spread, time from symptom onset to hospital visit, initial diagnosis at other hospitals, presence or absence of metastasis, or postoperative follow-up period. Patients in group B tended to be younger, with more upper and lower eyelid lesions and diffuse types (p = 0.07, p = 0.07, and p = 0.06, respectively).

Table 2.

Clinicopathological characteristics of the SGC group (group A) and group with a diagnosis other than SGC (group B)

	Group A (n = 22)	Group B (n = 5)	p value
Age, years	78.0±7.5	70.6±19.8	0.07
Sex	M: 7, F: 15	M: 2, F: 3	0.73
Site			0.07
Lower	7	1
Upper	14	2
Lower and upper	1	2
T category			0.78
T1	16	4
T2	4	1
T3	2	0
Type			0.06
Nodule	20	3
Diffuse	2	2
Pagetoid spread			0.09
Presence	5	3
Absence	17	2
Time from symptom onset to visit our hospital, months	4.3±3.9	5.2±3.4	0.33
Initial diagnosis at another hospital			0.56
Benign disease	12	2
Malignant tumor	10	3
Postoperative follow-up period	13.8±11.6	19.6±14.7	0.17
Postoperative local recurrence or metastasis (+)	3	1	0.71

M, male; F, female.

Discussion

SGC is a relatively common and significant eyelid malignancy in Japan. If not diagnosed early, tumor growth and lymph node metastasis may occur, resulting in poor prognosis. However, it has been reported that the diagnosis and treatment of eyelid SGC can be delayed due to misdiagnosis of benign conditions such as chalazion, blepharitis, and eyelid conjunctivitis [5, 12, 13]. Furthermore, achieving accurate pathology results has proven to be challenging, with instances of incorrect diagnoses such as SCC and BCC [12, 14–16]. Historically, there have been reports of erroneous pathological diagnoses in 5 of 30 cases (16.7%) [14]. Additionally, it was observed that only 50% of biopsies conducted at other hospitals were correctly histopathologically diagnosed as SGC, while 18% were misdiagnosed as SCC [5]. These findings underscore the complexity of achieving an accurate pathological diagnosis of SGC. It has been suggested that as many as 25% of pathological diagnoses made by general pathologists who are not well versed in ophthalmic pathology may lack precision [15]. Similarly, in this study, initial HE staining resulted in a pathological diagnosis other than SGC in 18.5% of the cases. Pathological diagnoses based only on HE staining may have limited accuracy.

However, HE staining of SGC in pathology is highly important. Key features of HE staining of SGC include the following: amphophilic staining of tumor nuclei, inflammation beneath the pagetoid spread of SGC, and sloughing of epithelium in pagetoid spread, where the tumor cells tend to slough off, leaving only a basal layer of keratinocytes. Furthermore, the biopsy procedure is crucial, and in the case of superficial biopsies, a full-thickness biopsy of the eyelid is warranted to evaluate the meibomian glands [18]. While a full-thickness biopsy was performed in this study, it is thought that there are cases that cannot be adequately diagnosed by HE staining alone.

In this study, SGCs that were difficult to diagnose by initial pathology, that is, HE staining alone, tended to be younger, had upper and lower eyelid lesions, and were more diffuse in type. It has been reported that the non-nodular type of SGC is more invasive and likely to show squamous epithelialization [19]. SGC is considered a difficult cancer for histological diagnosis because of the difficulty in determining sebaceous gland differentiation [16]. In addition, SGCs are often poorly differentiated and may differentiate into the squamous epithelium, rendering it difficult to differentiate them from SCC and BCC [14, 16, 20, 21]. Furthermore, hypo-differentiated SGCs lack sebaceous gland structures and have been reported to have impaired sebaceous gland metabolism at the genetic level as well [22, 23]. It has been reported that 84% of SGCs are hypo-differentiated, of which 75% resemble SCC and 7% resemble BCC [24]. Further, 11 of 30 cases of SGC showed squamous differentiation and that individual cell necrosis with pyknotic nuclei, which can be observed in BCC and SCC, rendered pathological diagnosis difficult [14]. Bowen disease is also an in situ SCC that causes intraepithelial invasion [25]. Poorly differentiated SGC is prone to Bowen-like proliferation [24], and the histological features of intraepithelial SGC and Bowen disease are very similar, rendering diagnosis difficult [26, 27]. Therefore, it may be difficult to diagnose SGC accurately using HE staining alone.

Accurate pathological diagnosis of SGC, especially its differentiation from BCC and SCC, is important. SGC is an eyelid tumor with a relatively high risk of metastasis, with local recurrence reported in 6–24% of cases and lymph node metastasis reported in 8–21% of cases [5, 8, 9]. Local recurrence of 9–11% and cervical lymph node metastasis of 9–13% have been reported in Japan [6, 7, 28]. In contrast, with regard to BCC and SCC, recurrence rates of 3–5.6% for periocular BCC [29, 30] and 3–12.1% for periocular SCC [31, 32] have been reported, suggesting that the recurrence rate is relatively low compared to that of SGC. Furthermore, it has been reported that SGC is more malignant and has more metastases than BCC and SCC [3, 5, 33, 34]. Therefore, an accurate diagnosis of SGC is important in terms of prognosis and follow-up of postoperative recurrence and metastasis. Complete surgical excision of eyelid malignancies is the basic method, and adequate resection with a safety margin is necessary. It has been reported that SGC has a low recurrence rate when a safety margin of 4 mm or more is taken [1, 4, 5]. In contrast, the resection margin for BCC is often reported to be approximately 3 mm [35–37]. Thus, there may be differences in preoperative surgical plans and eyelid reconstruction methods. Accurate pathological diagnosis of SGC is important, and it would be desirable to have an accurate diagnosis before extended resection.

Several methods have been used to diagnose SGC pathologically. Fat stains such as Oil Red and Sudan III are useful for detecting lipid accumulation [14]. However, it has been noted that frozen sections from fresh tissue are required and that Oil-Red staining is often negative in poorly differentiated lesions [14, 21]. Recent advances in immunohistochemical diagnostic techniques have improved the diagnostic accuracy of SGC. Anti-ADP antibody recognizes proteins associated with lipid droplets and is a marker of lipid droplet accumulation in various diseases (fatty liver, liposarcoma, SGC, etc.) [38]. Previous studies have reported an anti-ADP positivity rate of 88–100% in the immunostaining of SGC, making it a useful diagnostic marker [38, 39]. Similarly, in this study, ADP was positive in all 5 cases in which the initial pathological diagnosis was not SGC and was very useful in confirming the diagnosis of SGC. ADP staining appears to be a very effective diagnostic tool in cases where pathological diagnosis is difficult despite strong clinical suspicion of SGC.

However, it has been reported that ADP is often positive in SCC and other conditions [40, 41]. Therefore, a combination of two or more immunohistochemical staining techniques is effective. AR is typically expressed in the sebaceous glands and has been reported to be a useful marker for assessing sebaceous gland differentiation [42]. The AR positivity rate is very high in SGC [43, 44]. AR has been reported to be useful in differentiating SGC from SCC and BCC in the diagnosis of SGC [45, 46]. Both the sensitivity and specificity of SGC have been reported to be very high when both ADP and AR are positive [43], and combining ADP and AR staining may further improve diagnostic accuracy. In this study, patient 3, whose pathology was initially diagnosed with Bowen’s disease using HE staining, was positive for ADP but negative for AR. In the past, it has been reported that all 10 cases of Bowen’s disease, which is a form of SCC in situ, were found to be ADP negative [47]. In patient 3, the presence of diffuse blepharitis-like proliferation and pagetoid spread of ADP-positive cells were considered diagnostic of SGC.

The diagnoses of the 5 patients in group B were SCC (3), Bowen’s disease (1), and BCC (1). The 3 patients diagnosed with SCC had lower differentiation, a higher ki67 index, and more active mitosis than the other 2 patients. SGC, which is difficult to differentiate from SCC, may have a poorer prognosis than SGC, which is difficult to differentiate from Bowen’s disease or BCC. However, the major limitation of this study is its small sample size. Further accumulation of cases and long-term follow-ups are necessary.

This study had several limitations. We determined the presence or absence of ADP and AR immunostaining; however, a more detailed pathological diagnosis requires quantifying and scoring the intensity of ADP immunostaining and nuclear expression of AR in tumor cells. In addition, ADP and AR are not SGC-specific markers; ADP is reported to be expressed in non-lipid gland tumors such as renal cell carcinoma [48], while AR is reported to be expressed in conditions like Paget’s disease, apocrine tumors, etc. [49, 50]. It has also been reported that the foci of positive granular staining observed in this study may also appear in BCC and SCC [51]. Therefore, interpretation in poorly differentiated cancers should be approached with caution, and the consideration of other immunostains is advisable.

Finally, SGC is difficult to diagnose pathologically and may be diagnosed as other cancers such as SCC. Immunohistochemical diagnostic techniques would enable accurate diagnosis. Ophthalmologists should inform pathologists of the possibility of SGC based on clinical findings and order immunostaining for ADP and AR, which would greatly improve diagnostic accuracy.

Acknowledgments

We thank the Pathology Institute Corporation (Toyama, Japan) for performing some of the immunopathological diagnostics.

Statement of Ethics

The procedures conformed to the tenets of the Declaration of Helsinki of the World Medical Association. Written informed consent was obtained from all patients after providing them with sufficient information about the procedures. This study was approved by the Institutional Review Board of the Toyama University Hospital (No. R2023048).

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This study was supported, in part, by a Grant-in-Aid for Scientific Research (22K09786) from the Japan Society for the Promotion of Science.

Author Contributions

Drs. Tatsuya Yunoki and Atsushi Hayashi planned the study, reviewed the clinical and pathological materials, and wrote and edited the manuscript. Dr. Akio Miyakoshi reviewed the clinical data material and edited the manuscript.

Funding Statement

This study was supported, in part, by a Grant-in-Aid for Scientific Research (22K09786) from the Japan Society for the Promotion of Science.

Data Availability Statement

All data generated or analyzed during this study are included in this article and its online supplementary material files. Further inquiries can be directed to the corresponding authors.