Article Text

Download PDFPDF

14-year case series of eyelid sebaceous gland carcinoma in Chinese patients and review of management
  1. Stacey C Lam1,2,
  2. Emmy Y M Li1,2,
  3. Hunter K L Yuen1,2
  1. 1 Department of Ophthalmology, Hong Kong Eye Hospital, 147K Argyle Street, Kowloon, Hong Kong, China
  2. 2 Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, China
  1. Correspondence to Dr Stacey C Lam, Hong Kong Eye Hospital, 147K Argyle Street, Kowloon, Hong Kong, China; staceylam{at}


Aims Sebaceous gland carcinoma (SGC) of the eyelid is a rare but potentially deadly cancer. The purpose of this study was to determine whether the American Joint Committee on Cancer (AJCC) 7th Edition T category for SGC correlated with metastasis and survival in the Chinese population.

Methods This was a retrospective, single-centre cohort study. Patients with surgically resected eyelid SGC between January 2001 and May 2015 at the Hong Kong Eye Hospital were reviewed. Tumours were staged using the AJCC criteria. The main outcome measures included local recurrence, metastasis and death. Disease-free survival (DFS) was measured from the completion of treatment; overall survival was measured from the date of initial diagnosis.

Results The study included 22 Chinese patients with a mean age of 65.4 years. The majority presented as a nodular lesion (91%) with 12 eyes (54.5%) initially misdiagnosed and a mean presentation time of 1 year. It was found that those with AJCC stage T2b or higher were significantly associated with lymph node metastasis (P=0.002) when compared with those with stage T2a. Older age at diagnosis (P=0.035) and no misdiagnosis (P=0.025) were associated with shorter DFS. Those with stage 3a or higher were associated with shorter DFS (P=0.007) and overall survival (P=0.024).

Conclusion Similar to previous reports, in this Chinese cohort, AJCC staging for SGC correlated with lymph node metastasis, DFS and overall survival. Those with stage 2b or higher on presentation will need closer surveillance for lymph node metastasis and may benefit from sentinel lymph node biopsy.

  • sebaceous carcinoma
  • malignant eyelid tumour
  • Hong Kong
  • lymph node metastasis
  • AJCC

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Sebaceous gland carcinoma (SGC) of the eyelid is an uncommon malignant tumour of the sebaceous glands. It is predominantly a disease of the elderly, more often in females, and affects the upper more than the lower eyelid. It represents 1%–5.5% of malignant eyelid tumours in the Caucasian population.1 In the Asian population, SGC has been reported to account for up to 38.6%, 37.5% and 31.2% of all malignant eyelid tumours in Beijing, India and Japan, respectively.2–4 In Hong Kong, SGC has been reported to comprise 7.1%–11.1% of all malignant eyelid tumours.5 6 This ranks SGC as the second most common eyelid malignancy in the Asian population after basal cell carcinoma.

SGC is uniquely dangerous because it masquerades inflammatory conditions such as blepharitis and chalazia and can also be locally aggressive and can metastasise to lymph nodes and distant organs. The American Joint Committee on Cancer (AJCC) 7th Edition has classified SGC using TNM staging. Previous studies have tried to compare the AJCC classification to clinical outcomes in Caucasians, Japanese, Indian and Korean populations.7–10 Unfortunately, no new research has been published in the English literature on the outcomes in the Chinese population since 1982.11 The purpose of this study was to determine the prognostic value of the AJCC T category on metastasis and survival in Chinese patients. Proper identification of high-risk patients would help determine those who need closer monitoring and better manage patient expectations.


A computerised search of medical records was conducted on all patients diagnosed with SGC at the Hong Kong Eye Hospital, Hong Kong, between January 2001 and May 2015. Patients included were those with histology confirmed SGC, localised tumour (AJCC stage T1–T3) arising from a single focus, who underwent surgical resection and had a minimum 6 months follow-up. Patients who had clinical, laboratory or radiological evidence of metastasis at time of diagnosis were excluded.

Clinical data collected included age at presentation, gender, tumour morphology (solitary nodule or diffuse thickening without distinct nodule), site, size (largest dimension of tumour), initial diagnosis and management. Time to presentation was defined as the time between onset of symptoms and establishment of a correct diagnosis. On presentation, blood panel (complete blood count, liver function test and renal function test) and chest radiographs were taken from all patients. Orbital imaging was performed for all patients (either MRI or  CT with or without whole body positron emission tomography). Liver imaging (ultrasound liver or CT of abdomen and pelvis) was performed on four patients who had a history of other malignancies.

Eyelid margin SGC was excised with full-thickness en-face frozen section-controlled technique as described by Gill et al.12 The main lesion was excised, followed by excision of 1 mm full-thickness medial, lateral and inferior/superior margins to make up for a 5 mm margin. The main specimen was sent for permanent paraffin sections, and intraoperative frozen section analysis was conducted on the three margins. The section was frozen using liquid nitrogen and stained with H&E. If the outermost section had residual tumour cells, then a subsequent 1 mm full-thickness resection was repeated until all margins were clear. Standard exenteration was performed on the other cases.

Follow-up data were collected at approximately 1 week after surgery; then at 1, 2, 3, 6 and 12 months after and at 6–12 month intervals thereafter. At follow-up, physical examination was performed to look for local recurrence, lymphadenopathy and surgical complications. Systemic surveillance included annual blood panel and chest radiographs; other specific tests were guided by history and physical findings. Primary outcome measures of the study were local recurrence, metastasis and death. Disease-free survival (DFS) was defined as time of surgery until the date of local recurrence, metastasis or last follow-up date, whichever came first. Overall survival was defined as date of diagnosis until death or the last follow-up.

Statistical analysis was performed using IBM SPSS Statistics for Windows, V.23.0 (IBM, Armonk, New York, USA). Descriptive statistics were used to summarise demographic data. Continuous variables that were not normally distributed were analysed by the Mann-Whitney U test. Pearson χ² test or Fisher’s exact test was used for categorical data analysis. Kaplan-Meier method was used to draw the survival curve, and log-rank test was used to test for any significant differences between the curves. A P value of ≤0.05 was considered as statistically significant. The study was conducted in accordance with the tenets of the Declaration of Helsinki.


In a period of 14 years, the study included 22 patients with eyelid SGC. The average age was 65.4 years (range, 42.1–93.7 years), with the majority being female (16 patients, 72.7%). The size of the primary lesions ranged from 3 mm to 30 mm. Based on the T category of the AJCC classification, the tumours were classified as T2a (16 eyes, 72.7%), T2b (3 eyes, 13.6%), T3a (2 eyes, 9.1%) and T3b (1 eye, 4.5%). All patients had surgery as the primary treatment. Twenty patients had local excision with orbital reconstruction, including one with mitomycin-C, and two patients had orbital exenteration. One patient had SGC involving the surgical margin in permanent section and required re-excision. The final pathological margins in all cases were clear. Cohort and tumour characteristics are listed in table 1.

Table 1

Cohort characteristics with sebaceous gland carcinoma

Over a mean follow-up period of 5.76 years, three patients (13.6%) developed local recurrence, three patients (13.6%) developed lymph node metastasis and none of our patients had distant metastasis. For the three patients with local recurrence, all were detected by clinical examination. One patient had a recurrence at the same site, one patient on the other lid of the same eye and one patient involving the lower lid and maxilla. The mean primary tumour size was 6.8 mm (range, 5–10 mm), and all three patients had undergone local excision with orbital reconstruction. The mean postoperative time to local recurrence was 18 months (1.2–42 months). For their recurrent disease, one patient underwent palliative radiotherapy, one underwent maxillectomy followed by radiotherapy and one declined further treatment.

There were three patients who developed lymph node metastasis (two patients at the ipsilateral parotid gland node and one patient at the ipsilateral parotid gland and ipsilateral submandibular lymph node). None of these patients had lymph node metastasis on presentation, from physical examination or on imaging (one had CT of the head, one had a MRI of the head and one had a whole body positron emission tomography-CT). Among these patients with lymph node metastasis, two were detected on clinical examination and one by imaging after local recurrence was detected. The mean tumour size was 19 mm (range, 10–25 mm). For the treatment of the primary tumour, two of them had undergone local excision with orbital reconstruction and one had orbital exenteration. The mean postoperative time to lymph node metastasis was 9 months (range, 1.2–18 months). Two patients underwent palliative radiotherapy to their metastatic site and one refused further treatment.

There were five non-tumour-related deaths and one tumour-related death within the study period. The patient who died had metastasis to the ipsilateral parotid and submandibular gland with facial swelling and difficulty swallowing; the time to death was 3.2 years from initial presentation. Sixteen patients, including two patients with metastasis, were alive at the end of the study.

The AJCC T category was found to be associated with lymph node metastasis, DFS and overall survival. Tumours with AJCC stage T2b or higher were significantly associated with lymph node metastasis when compared with those with T2a (P=0.002; table 2). Moreover, those with T3a or higher were associated with shorter DFS (P=0.007) and overall survival (P=0.024) (table 2). The median DFS for those with T2a/T2b was 3.55 years, while DFS for T3a/T3b was 0.82 years with a HR of 5.88 (95% CI, 1.362 to 25.48) (figure 1). Those with T3a or higher were also associated with shorter overall survival (P=0.033). The median overall survival for those with T2a/T2b was 5.56 years while those with T3a/T3b was 2.86 years with a HR of 4.27 (95% CI, 1.09 to 16.8) (figure 2).

Table 2

Associations between tumour size, local recurrence, lymph node metastasis, disease-free survival and overall survival

Figure 1

Kaplan-Meier curve showing the difference in disease-free survival between stage T2a/T2b and stage T3a/T3b.

Figure 2

Kaplan-Meier curve showing the difference in overall survival between stage T2a/T2b and stage T3a/T3b.

Other factors that were found to be associated with DFS were age and initial misdiagnosis. Those with older age at diagnosis had a shorter DFS, with a HR 1.04 (95% CI, 1.00 to 1.08; P=0.035) (table 3). Those who had an initial misdiagnosis had a longer DFS with a HR of 5.47 (95% CI, 1.2 to 8.74; P=0.025) (table 3). Gender, laterality, location, morphology, poorly differentiated pathology and presence of pagetoid spread were not found to be associated with DFS or overall survival.

Table 3

Predictive factors associated with survival


This study aims to determine the prognostic value of the AJCC 7th Edition staging system for eyelid SGC on Chinese patients. We found that in this ethnic group, the T category in the classification correlated with lymph node metastasis, DFS and overall survival. Moreover, older age at diagnosis and no prior misdiagnosis were correlated with a shorter DFS.

In keeping with previous publications, we found that SGC larger than 10 mm was correlated with lymph node metastasis. In our series, no tumours 9 mm or smaller had lymph node metastasis. Our study had a regional lymph node metastasis rate of 13.6%, which was in line with previously reported rates of 8%–18%.7 13 Similar to the Caucasian cohort, we found that tumours T2b or higher (>10 mm) were associated with lymph node metastasis and tumours T3a or higher were associated with shorter DFS and overall survival.7 Similarly, in the Korean cohort, T2b or worse were associated with lymph node metastasis.10 The Indian cohort also found that the risk of systemic metastasis and death was related to an increasing T category.9 However, in the Japanese cohort, the authors found that tumours even smaller than T2a were associated with tumour metastasis.8 This may be because of the difference in the primary modality of treatment employed. Our Chinese cohort, along with the Caucasian, Korean and Indian cohort all used surgical excision as primary method of treatment (margins not specified), while the Japanese cohort also included patients who underwent surgical excision without frozen section and radiotherapy in their study.

Our study’s findings can provide prognostic information and help clinicians design a rational surveillance schedule for patients. Sentinel lymph node biopsy (SLNB) for eyelid carcinoma has been suggested to detect microscopic regional lymph node metastasis that would not have been detected by clinical examination or imaging otherwise. Current indications for SLNB for SGC are not well established. Based on Esmaeli’s study on Caucasian patients, the authors suggested SLNB for tumours stage ≥T2b and tumour size >10 mm in diameter due to the associated increased risk of metastasis.7 With our current findings, the use of SLNB can also be considered in the Chinese population for tumour size >10 mm or greater than or equal to a T2b lesion.

SGC is considered as a masquerade disease and is associated with misdiagnosis and delayed treatment. Among our 22 cases, 45.4% had a correct diagnosis on initial presentation. This was higher than the reported rates in the Caucasian studies, which had a correct diagnosis in 18.6%–32% among their patients.13 14 This discrepancy could be due to the difference in the clinical presentation of SGC. There was a higher proportion (23%–2%–56.7%) of Caucasians patients who presented with a diffuse eyelid thickening leading to a misdiagnosis of blepharoconjunctivitis.13 14 In our study, 91% presented with a nodular lesion, compared with 94% of the Japanese and 55% of the Korean cohort.8 15 Chalazion was the most commonly reported misdiagnosis, representing 45% in our study, lying in the range of 29%–50% reported in other Asian studies.8 15 Higher rates of correct initial diagnosis in our cohort may be due to an increased awareness of SGC in its varied clinical manifestations and histological features among Asian ophthalmologists and pathologists.

Paradoxically, those who had a correct diagnosis on presentation had a shorter DFS. Those who already had malignant features on clinical presentation may facilitate the diagnosis of the tumour, but could also represent a more aggressive lesion, which by nature might lead to shorter DFS despite a correct initial diagnosis. Shorter DFS was also associated with older age. Older age at diagnosis is a known risk factor for incidence of SGC. The average age of presentation in our cohort was 65 years old, which falls in the average age of presentation to be 57–72 years old in other studies.16 Those with older age would probably warrant closer monitoring postoperatively.

Our cohort focused on those with eyelid SGC who underwent complete surgical excision. The main treatment method is surgical excision with adequate margins. Prior studies have suggested at least a 5 mm margin, combined with frozen section control either by standard method or Moh’s micrographic surgery to check for residual tumour until the histopathological margins were clear.17 Frozen section allows for prompt intraoperative evaluation, though pagetoid spread may be difficult to identify because of poor resolution and tissue artefacts of freezing. Permanent sections with a staged procedure have been suggested as a more reliable method to confirm absence of tumour at the margins, though this procedure takes longer.18 Patients with orbital invasion, multifocal tumour or extensive pagetoid spread may need orbital exenteration. In our cohort, the two patients who underwent orbital exenteration had tumours stage T2b or higher. Topical adjuvant chemotherapy such as mitomycin-C has been suggested for conjunctival pagetoid invasion.19 This was combined with excision in one of our patients who had pagetoid spread in histopathology. Other methods that we did not employ include radiotherapy, chemotherapy and cryotherapy. Radiation as a primary treatment has been associated with higher rates of recurrence and mortality when compared with surgical excision.20 Case series have demonstrated the benefit of radiation for those who were not surgical candidates or who refused surgery.21 Adjuvant radiotherapy after orbital exenteration provides a better disease control than exenteration only.22 The role of chemotherapy is limited, though studies have suggested the use of 5-fluorouracil and cisplatin in recurrent cases.23 The use of conjunctival cryotherapy is controversial.24 Finally, in the management for metastasis, regional metastasis is typically treated with regional lymph node dissection (including parotidectomy) with adjuvant radiation.25 Radiation was used in four of our patients at the local recurrence site and metastatic site as palliative treatment.

Our study has investigated the course of surgical treatment of eyelid SGC. As a result, a limitation of this study is that significant predictors mentioned in this study are mainly applicable to those who were treated by surgery as their primary therapy. Furthermore, despite being the second most common malignant eyelid tumour in Asians, SGC is still an uncommon tumour, and the number of patients recruited in this cohort is still low. There are also inherent limitations of a retrospective study. Nonetheless, this study is important as it is the first to determine the prognostic value of AJCC for eyelid SGC on Chinese patients in the English literature. Ophthalmologists should be aware of the risk of metastasis in patients with tumours with stage T2b or higher and provide appropriate surveillance.


We would like to thank Mr Edward Choi for his help in verifying the statistical results.



  • Contributors All authors have contributed to the implementation of the research. SCL took the lead in the analysis and writing of the manuscript.

  • Competing interests None declared.

  • Patient consent Detail has been removed from this case description/these case descriptions to ensure anonymity. The editors and reviewers have seen the detailed information available and are satisfied that the information backs up the case the authors are making.

  • Ethics approval Institutional Review Board of the Kowloon Central Cluster, Hospital Authority, Hong Kong.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost B Jonas