Introduction

Sebaceous carcinoma (SC) is the second most common malignant eyelid tumor following basal cell carcinoma in the Asian population [1] and the third most common following basal cell carcinoma and squamous cell carcinoma in the Caucasian population [2]. The reported incidence is higher in the Asian population than in the Caucasian population [3]. The incidence was found to be 0.32 in male and 0.1 in female per 100,000 person-years in the US [4]. SC constituted 1–5% of all malignant eyelid tumors in the US [5], 7.1–11.1% in Hong Kong SAR [3, 6], 31.7% in China [7], and 32.6% in India [8].

It has been coined as the great masquerader due to its variable clinical presentations and multifocal nature in some cases. Difficulties in early recognition combined with its potentially aggressive behavior can lead to poor prognosis. The tumor-related mortality rate was reported at 6% during mean follow-up of 39.2 months with a 5-year disease-specific survival rate of 92% in a study of 100 patients with SC [9]. With better understanding of the disease and improved histopathological techniques, early diagnosis has improved the overall outcome [10]. The mainstay of treatment is surgical excision with the aim of achieving negative microscopic margins. External beam radiotherapy, cryotherapy, and topical chemotherapy have been used as adjuvant treatment [5]. However, there is no clear consensus on many aspects of management of SC and clinical practice can vary greatly between different populations. Our study aims to identify some of these differences and explore the current practice patterns amongst oculoplastic surgeons in the Asia Pacific region in the management of eyelid SC.

Methods

A survey containing 40 questions regarding the demographics, training background, and practice patterns for management of SC of eyelid was constructed on the website kwiksurveys.com. The survey was in the format of multiple-choice questions and when appropriate, respondents were given the option of selecting one-best answer, multiple answers, or to specify their answer in free-text format. A copy of the survey can be found in Appendix A. Out of the 40 questions, 12 of them could have multiple answers. Respondents could skip questions if they wished to do so. The survey was electronically distributed via e-mail to members of the Asia Pacific Society of Oculoplastic and Reconstructive Surgery (APSOPRS), Hong Kong Society of Oculoplastic and Reconstructive Surgery (HKSOPRS), and individual oculoplastic society directories of the corresponding author. A follow-up e-mail was sent 1 month later as a reminder. The responses were extracted from the online platform and analyzed.

The study protocol was approved by the local institutional research ethics committee, Institutional Review Board of the Hospital Authority Hong Kong West Cluster. The study firmly adhered to the tenets of the 1964 Helsinki declaration and its later amendments.

Results

A total of 263 respondents participated in the survey. Among the 263 respondents, 9 were not from the Asia Pacific region and were excluded, while 63 responders answered fewer than 10 questions and were also excluded. The data from the remaining 192 responders were included in the study.

Geographic distribution, training, and experience with SC of survey respondents

See Table 1, Q.1–7. All respondents were from Asia Pacific countries, primarily Hong Kong (23%), China (18%), South Korea (10%), and India (9%). 68 (38%) respondents practiced primarily in an academic institution and 62 (35%) respondents in the public health care system. The respondents received their oculoplastic surgical training in various regions including the US (34, 18%), Hong Kong (27, 14%), and South Korea (18, 9%). 86 (67%) respondents managed fewer than five cases of SC per year, 26 (20%) managed 5–10 cases per year while 17 (13%) managed more than 10 cases annually. 30 (24%) respondents had managed more than 40 cases of SC in their career. 96 (36%) respondents thought that the most common presentation of SC was a painless eyelid mass and 81 (30%) selected recurrent chalazion as the most common presentation.

Table 1 Demographics and training background

Reported practice patterns

See Table 2, Q.8–31. In patients with suspected SC, 74 (55%) surgeons selected an incisional biopsy as their preferred first intervention and 47 (35%) respondents selected full thickness excisional biopsy as the initial step. 106 (35%) respondents undertook physical examination with emphasis on lymph node palpation as work up for newly diagnosed SC. Of the remaining respondents, 81 (27%) performed radiological imaging and 47 (16%) ordered blood work up. Chest radiographs (CXR) and computed tomography (CT) orbit were the most preferred investigative modality regardless of size. 82 (30%) respondents performed conjunctival map biopsy when there was clinical suspicion of pagetoid involvement of conjunctiva, 56 (20%) in recurrent disease, and 52 (19%) if both upper and lower eyelids were involved.

Table 2 Practice patterns and preferences in management of sebaceous carcinoma

Intraoperatively, 36 (32%) determined surgical margin for excision of SC based on frozen section findings, while 30 (27%) respondents preferred to use 5 mm of clinical margin clearance. Margin clearance was based on frozen section evaluation by a pathologist according to 63 respondents (57%) and Mohs micrographic surgery according to 20 respondents (18%). After excision of SC, reconstruction of posterior lamella using lid sharing procedures was the preferred option of 51 (46%) respondents for upper eyelid and 59 (55%) respondents in lower eyelid.

Adjuvant treatment was selected by 71 (23%) of respondents in the presence of positive margins after surgery. Among the adjuvant therapies, 44 (42%) used radiotherapy and 39 (38%) used topical Mitomycin-C (MMC). Regarding the use of MMC, 50 (48%) respondents followed the regimen 0.04% MMC four times daily for a week, followed by 1 week of rest. In cases with pagetoid spread but without metastasis, complete excision with frozen section control of margins combined with adjuvant therapy was the preferred approach of 84 (78%) respondents and exenteration was preferred by 14 (13%) respondents.

46 (42%) respondents never performed sentinel lymph node biopsy (SLNB), while 3 (3%) performed SLNB in all of their cases. The remaining respondents performed SLNB in selected cases. Among those who performed SLNB, 26 (24%) performed SLNB in patients with stage AJCC 7th edition TNM criteria T2b SC (greater than 10 mm).

In locally advanced non-metastatic SC, 53 (51%) respondents preferred surgical excision combined with adjuvant treatment. Among the neo-adjuvant agents used in globe preserving surgical excision for locally advanced SC, 21 (21%) respondents preferred MMC. For recurrent disease, 50 (48%) preferred surgical excision with adjuvant therapy and 24 (23%) preferred exenteration combined with radiotherapy.

Outcomes

See Table 3, Q.32–34. According to their experience, 42 (40%) respondents estimated a recurrence rate of less than 10% in patients with localized SC. 52 (50%) of respondents estimated the overall mortality rate from SC as less than 5%. 62 (61%) of respondents selected distant metastatic disease as the most common cause of death in patients with SC.

Table 3 Results and outcomes based on personal experience

Follow-up

See Table 4, Q.35–37. From our survey, 65 (64%) respondents routinely referred patients for assessment by a medical oncologist and 37 (36%) did not. 38 (37%) of respondents used physical examination alone during surveillance. 75 (74%) of respondents reported that long-term surveillance for local control was carried out by oculoplastic surgeons.

Table 4 Practice patterns in follow-up care and surveillance of sebaceous carcinoma

Discussion

Consistent with the well-known higher incidence of SC in the Asia Pacific countries [1, 5, 11, 12], many of our respondents seem to be more experienced in managing SC compared with the average non-Asian oculoplastic surgeon and up to 24% have managed more than 40 cases in their career so far.

In keeping with previous studies in the Asia Pacific population, the main presenting symptom of SC as reported by our respondents was a solitary eyelid nodule (36%) [13,14,15]. Diffuse eyelid thickening mimicking blepharoconjunctivitis was reported by 13% of our respondents, which is less than the Caucasian population [16, 17]. Since nodular presentation such as recurrent chalazion is more common, incisional biopsy together with incision and curettage is commonly performed. This could explain why our respondents most commonly performed incisional biopsy (55%) followed by full thickness excisional biopsy (35%). This is different from a recently published British cohort in which primary excisional biopsy was the most commonly performed diagnostic intervention (42%) followed by incisional wedge biopsy (33%) [18].

We found there was no consensus amongst respondents on the investigative modality used for disease staging. More respondents chose CT orbit over MRI orbit, perhaps due to the availability and affordability of CT over MRI. Some series suggest map biopsies for all cases because even in solitary nodular SC with no conjunctival involvement clinically, half had biopsy-proven conjunctival involvement [19]. Others perform map biopsy only if there is clinical suspicion of diffuse palpebral and bulbar conjunctival involvement [20]. Our respondents performed conjunctival map biopsy in selected cases, most commonly those with clinical signs suggestive of pagetoid involvement of conjunctiva (30%), recurrent disease (20%), and both upper and lower eyelid involvement (19%). Sa et al. [9] carried out a retrospective study of predominantly White and Hispanic subjects in the US and found among the 100 patients diagnosed with SC, 31 patients had intraepithelial neoplasia and this finding was mostly based on evaluation of the surgical specimen rather than conjunctival map biopsies; these authors advocate for conjunctival map biopsies only in selected patients who have clinically suspicious signs of involvement of conjunctiva adjacent to the main tumor. Yoon et al. [15] found only 8.3% of their Korean patients with SC had pagetoid spread. SC may behave differently in terms of its clinical presentation and tendency for pagetoid spread between different ethnicities.

We found that most respondents rely on physical examination as the main method for the detection of lymph node metastasis. For lesions greater than 10 mm, 24% of respondents would perform SLNB and 9% would perform SLNB for lesions greater than 5 mm. The relatively low percentage of SLNB performed in the Asia Pacific region may be related to the lack of technical expertise at certain centres and it may also be due to the fact that SLNB is logistically time consuming to set up. Among our respondents, 10% and 11% of respondents would evaluate regional lymph nodes with ultrasonography and perform fine needle aspiration as needed in lesions less than 10 mm and greater than 10 mm, respectively. Ultrasonography is a relatively inexpensive and accurate way of assessing for nodal involvement [21] and can be considered if SLNB is not available. Since SC metastasize to regional and systemic lymph nodes, ultrasonography of parotid, submandibular and cervical lymph nodes, and SLNB have been suggested as a way to evaluate regional lymph nodes to rule out metastasis. Studies have demonstrated that tumors larger than 10 mm correlate with nodal metastasis and this can occur in up to 18% [14, 16, 21]. Watanabe et al. [13] found that even tumors less than 5 mm can be associated with regional nodal metastasis among their Japanese cohort. Routine SLNB requires special expertise and can yield false negative results early in the learning curve, but the accuracy of SLNB for eyelid tumors has improved considerably in the past decade [22, 23]. The overall SLNB positivity rate was reported at 13% by the MD Anderson Cancer Center and all cases with positive SLNB had SC lesions greater than 10 mm [23, 24]. In a large series of 51 eyelid and conjunctival melanomas, 30% of patients with eyelid melanoma had positive SLNB [22]. This highlights the importance of SLNB in ocular adnexal malignancies.

According to our survey, the majority of responders (32%) utilized frozen section to determine the size of excision margin with 27% using a 5 mm clinical excision margin. This is consistent with previous studies using an excision margin of 5 mm [14]. To assess for margin involvement, frozen section is the most commonly employed technique (57%) followed by Mohs micrographic surgery (18%). Most oculoplastic surgeons use a similar technique and principle to Mohs surgery in the eyelid area but removing the main tumor specimen as a full thickness “wedge” resection and then taking small en-face (adjacent) margins next to the main tumor specimen until negative margins are achieved. Our survey respondents used Mohs surgery more frequently than the British cohort, in which only 8% utilized Mohs surgery [18]. Mohs surgery, in common with all techniques that use frozen section evaluation of margins carries the risk of missing intraepithelial pagetoid spread and skip lesions [25]. The long-term benefits of Mohs surgery compared with other techniques is still unknown [26]. Full-thickness en-face frozen section has also been shown to be effective with comparable results to Mohs surgery for basal cell carcinoma [27, 28]. In a recent report from MD Anderson Cancer Center, using complete surgical excision of SC with en-face margin evaluation on frozen section, the local recurrence rate was only 6% [9]. In 4 of the 6 patients who experienced local recurrence in this 100-patient cohort, the surgeon had knowingly left a microscopically positive margin in an effort to avoid an orbital exenteration and MMC was given instead. This suggests that frozen section control using en-face margins can yield comparable results to Mohs surgery [29]. Despite not being commonly employed by our respondents, local resection with rapid paraffin section analysis has been employed in some studies and offered low recurrence rates with good cosmetic outcome. Its advantage is that paraffin sections may be more reliable than frozen sections for detection of intraepithelial pagetoid spread and margin evaluation according to some reports [30].

In our survey, radiotherapy, topical MMC, and cryotherapy were the most preferred adjuvant therapies (42%, 38%, and 14%, respectively). Shields et al. used cryotherapy after local excision in 68% of their patients and report a recurrence rate of 18% [16]. The use of MMC was adopted by 47% of our respondents. The most common regimen used was 0.04% Mitomycin-C four times daily for a week followed by one week off in one cycle used for four cycles. This is similar to the regimen used by Shields et al. [31] for patients with proven pagetoid involvement of the conjunctiva. Chemotherapy combined with surgical excision may be used in cases with extensive disease, nodal involvement, metastasis and occasionally in recurrent cases [32]. However, 46% of our respondents do not use neo-adjuvant chemotherapy for globe-preservation in locally advanced SC. Neo-adjuvant chemotherapy such as 5-fluorouracil and cisplatin may have a role in downstaging disease and avoiding more radical surgery, but current evidence is mostly limited to case reports and small case series [33, 34].

Based on their experience, the majority of our respondents report a recurrence rate of under 20% and a mortality rate of under 10%, with distant metastatic disease as the most common cause of death. This appears to be in keeping with other studies [9, 10, 14, 16]. Our survey suggests that most practitioners prefer a multidisciplinary approach in managing patients with SC, with up to 64% of surgeons routinely referring patients for assessment by an oncologist.

Our study is the first study to date to analyze the various practice patterns of oculoplastic surgeons in the Asia Pacific region. Limitations of this study include the nature of survey, which relies on the surgeon’s impressions and estimation of events, rather than actual surgical outcomes. There is a definite possibility of selection and recall bias.

SC can have variable clinical presentations and may also behave differently in different ethnicities, explaining the variation in practice and management. Our survey highlights some differences in practice patterns in the Asia Pacific region compared with other populations. In particular, SC favors a nodular presentation in the Asia Pacific region. Initial diagnostic intervention and approach to conjunctival map biopsy may be different. We also found that SLNB is less commonly performed, possibly due to lack of access and expertise.

Summary

What was known before

  • Sebaceous carcinoma has variable clinical presentations and management practices can vary greatly between surgeons and institutions.

  • Mainstay treatment includes excision with adequate margins, radiotherapy, cryotherapy, and topical chemotherapy.

What this study adds

  • This study explores the practice patterns of surgeons in Asia Pacific and how this compares with other populations.

  • Incisional biopsy is commonly performed, conjunctival map biopsies are performed in selected cases, while Mohs micrographic surgery and sentinel lymph node biopsy are less widely available.