Appearance of the left upper eyelid tumor in patient 2. The irregularity in the surface of the tumor is due to an incisional biopsy.
The tumor in patient 2 was centered in the dermis. Asterisk indicates the tarsus; scale bar, 1 mm (hematoxylin-eosin).
The tumor in patient 2 contained islands of neoplastic basaloid cells ranging from solid to cribriform arrangements within mucinous pools partitioned by thin fibrous septa. Scale bar indicates 100 μm (hematoxylin-eosin).
The tumor in patient 2 had nuclear expression of p63 by scattered tumor cells within invasive nests of carcinoma. Scale bar indicates 50 μm (anti-p63 immunostain).
The in situ component of the tumor in patient 4 was delineated by nuclear expression of p63 in peripherally located, benign myoepithelial cells. Scale bar indicates 100 μm (anti-p63 immunostain).
Papalas JA, Proia AD. Primary Mucinous Carcinoma of the EyelidA Clinicopathologic and Immunohistochemical Study of 4 Cases and an Update on Recurrence Rates. Arch Ophthalmol. 2010;128(9):1160-1165. doi:10.1001/archophthalmol.2010.177
To report 4 cases of primary mucinous carcinoma of the eyelid, demonstrate how p63 expression aids diagnosis, and review the literature regarding the local recurrence potential of these eyelid tumors.
Since 1991, we have examined 4 cases of primary mucinous carcinoma of the eyelid. Histology slides were reviewed and clinical information was obtained from the medical records. Published cases of primary mucinous carcinoma involving the eyelid were identified using Ovid MEDLINE and PubMed and references within the articles.
Our average patient age was 73 years, the female:male ratio was 3:1, and patients had a painless nodular mass for 1 month to 2 years. No patient had another documented malignant tumor at the time of diagnosis. All tumors were dermis based in the upper eyelid, three-fourths invaded the orbicularis oculi muscle, and the average tumor diameter was 3.7 mm. All tumors expressed cytokeratin 7 and gross cystic disease fluid protein 15 and were negative for cytokeratin 20 expression. Tumors demonstrated p63 immunopositivity either focally in the invasive component or peripherally along the in situ component. No recurrences have been evident after an average follow-up of 52 months. Published cases occurring in the eyelid have had a 30% to 40% recurrence rate using surgery without intraoperative evaluation of surgical margins and a 7% recurrence rate after either Mohs surgery or excision with frozen section control.
Primary mucinous carcinoma of the eyelid is a low-grade malignant tumor with the potential for recurrence, although this may be reduced using Mohs surgery or excision with frozen section control.
Primary cutaneous mucinous carcinoma is a rare, low-grade malignant tumor arising preferentially on the head and neck, especially in the eyelid.1 Of the approximately 200 reported cases, 30% were in the eyelid and another 43% were elsewhere on the head or neck.2 This tumor has been reported under many different names, including adenocystic, colloid, gelatinous, and mucinous eccrine carcinoma.1 The latter name is derived in part from older studies that suggested features of eccrine secretory epithelium within the neoplastic component.3- 5 While still somewhat unsettled, the contemporary view is that these tumors actually demonstrate apocrine-type differentiation.6 Primary mucinous carcinoma of the eyelid skin has historically been regarded as having high rates of recurrence (up to 40%) and low metastatic potential.5,7 Here, we describe 4 patients with primary cutaneous mucinous carcinoma of the eyelid, describe how p63 expression aids in differentiating these tumors from mucinous adenocarcinomas metastatic to the skin, review the literature on this tumor occurring in the eyelid, and report lower rates of recurrence when contemporary patient management protocols are used.
We searched the pathology files at our institution from January 1, 1991, through December 21, 2008, and identified 4 cases of primary mucinous carcinoma of the eyelid. Patient 3 was from the consultation files of one of us (A.D.P.). Histology slides were reviewed and clinical information was abstracted from medical records. Immunohistochemical staining was performed on 5-μm paraffin sections following epitope retrieval and application of the antibodies at dilutions listed in Table 1. The chromogen diaminobenzidine tetrachloride was used to visualize the antibody-antigen complex, and sections were counterstained with hematoxylin. Negative control slides consisted of tissue sections of each case processed in parallel with substitution of nonimmune, isotype-specific IgG for the primary antibody. Reported cases of primary mucinous carcinoma involving the eyelid were identified using Ovid MEDLINE and PubMed and references within the articles. This study was approved by the Duke University Health System Institutional Review Board.
Patient 1 was an 80-year-old woman with a 2 × 2-mm lesion on the left upper eyelid that was clinically thought to be an epidermal inclusion cyst. An excisional biopsy revealed primary mucinous carcinoma because results from a systemic workup were negative for any other tumor. The margins were positive for tumor. There was involvement of the orbicularis oculi, but the tarsal plate was free of tumor and no perineural or lymphatic invasion was seen. A reexcision with frozen section margin control was performed; the patient was free of recurrence for a follow-up period of 158 months. A paraffin block was not available for immunohistochemical studies.
Patient 2 was a 59-year-old man with a 6 × 6-mm left upper eyelid lesion of 2 years' duration (Figure 1). A biopsy disclosed mucinous carcinoma (Figure 2 and Figure 3), and a pentagonal wedge resection was performed because there was no evidence of a primary lesion elsewhere using computed tomographic and positron emission tomographic scans. The tumor was centered in the dermis and invaded the orbicularis oculi muscle. The tarsal plate was free of tumor, and no perineural or lymphatic invasion was seen. Tumor cells were immunopositive for cytokeratin (CK) 7 and gross cystic disease fluid protein 15, with focal p63-positive cells (Figure 4) noted in the invasive tumor. The tumor cells did not express CK20 (Table 2). There was adjacent foreign body giant cell reaction associated with extravasated mucin. Owing to positive margins in the pentagonal wedge resection, Mohs micrographic surgery was subsequently performed with clear margins of resection; no recurrence was discerned at follow-up examination 7 months later.
Patient 3 was a 73-year-old woman with a right upper eyelid lesion of unknown duration (owing to cognitive impairment) that was increasing in size. The nodule was subcutaneous, firm, and thought clinically to be benign. An incisional biopsy revealed a mucinous carcinoma with positive margins. No perineural or lymphatic invasion was seen. Tumor cells were immunopositive for CK7 and gross cystic disease fluid protein 15, with focal p63-positive cells noted in the invasive tumor (Table 2). The tumor cells were negative for CK20. Mohs micrographic surgery was subsequently performed with clear margins of resection. No recurrence was evident at follow-up examination 6 months later.
Patient 4 was a 78-year-old woman with a painless, mobile, 3 × 3-mm subcutaneous nodule on the left upper eyelid of 1 month's duration. The preoperative diagnosis was an apocrine hidrocystoma. Excisional biopsy revealed a mucinous carcinoma involving the dermis and orbicularis oculi muscle, but the tarsal plate was free of tumor and no perineural or lymphatic invasion was seen. Tumor cells were immunopositive for CK7 and gross cystic disease fluid protein 15, with p63 staining of peripheral cells in the in situ component (Figure 5) and focal staining in the invasive tumor. The tumor cells did not express CK20. Margins were positive for tumor, but the patient did not receive additional treatment. There was no evidence of any other malignant neoplasm in the patient at this time. No recurrence was reported in the eyelid after 36 months, at which time the patient was diagnosed as having invasive ductal carcinoma of the breast without a mucinous component. Immunohistochemical staining results were not available for the breast carcinoma.
Primary mucinous carcinoma of the skin was first reported by Lotzbeck in 1859.8 By the 1950s, a histogenetic link between primary cutaneous mucinous carcinoma and the sweat gland was recognized and accepted.9,10 Starting in the 1970s, several case series emerged11,12 and the name adenocystic carcinoma was given to the entity we now recognize as primary cutaneous mucinous carcinoma. Primary cutaneous mucinous carcinomas are sweat gland–derived tumors,13 currently classified by the World Health Organization under the category of malignant tumors with apocrine and eccrine differentiation.1 These tumors were originally believed to be derived from eccrine coil epithelium4,5,14 and until recently were referred to as mucinous eccrine carcinoma.7 It is currently believed, however, that these lesions demonstrate apocrine differentiation6; thus, the term mucinous eccrine carcinoma is inappropriate.
Mendoza and Helwig12 provided the first contemporary report of mucinous carcinoma of the eyelid. Soon thereafter, Wright and Font5 published the largest study to date of mucinous carcinoma of the eyelid (21 cases), including 2 previously reported cases. Clinically, primary cutaneous mucinous carcinomas of the eyelid are slow-growing tumors with a duration of several months to several years, usually occurring in patients in their seventh decade.15 These papulonodular tumors range from 0.7 cm to 2.5 cm in diameter, about one-third the size of their trunk-based counterparts.7,12,16 The clinical impression is often that of a cyst, basal cell carcinoma, keratoacanthoma,12 nevus,14 or apocrine hidrocystoma, and in some instances the clinical differential includes vascular lesions such as Kaposi sarcoma.15 Histopathologic examination reveals an unencapsulated, dermis-based tumor that when located outside the eyelid may show extension into subcutaneous planes.6 In the eyelid, the tumor often invades the orbicularis oculi muscle, as exemplified by our patients. The tumor contains islands of neoplastic basaloid cells having solid to cribriform arrangements within slightly basophilic, periodic acid–Schiff–positive mucinous pools, sometimes partitioned by thin fibrous septa.1 The mucin is Alcian blue positive (pH 2.5) and stains with mucicarmine and colloidal iron. As a sialomucin, it is hyaluronidase resistant and sialidase labile.12 Cytologically, the tumor cells are characteristically round to cuboidal with moderate amounts of cytoplasm and typically with a low mitotic count and little nuclear atypia.1
Immunohistochemical analysis of primary mucinous carcinoma and differentiation of these neoplasms from adenocarcinomas metastatic to the skin are areas of active investigation. The tumor cells of primary mucinous carcinoma stain positively for carcinoembryonic antigen, epithelial membrane antigen, gross cystic disease fluid protein 15, low-molecular-weight cytokeratin, α-lactalbumin, salivary-type amylase, β2-microglobulin, thyroid transcription factor 1, thyroid transcription factor 3, and estrogen and progesterone receptors.14,16- 18 Primary mucinous carcinoma tumor cells are CK7 positive and CK20 negative,19 similar to breast cancer but different from gastrointestinal adenocarcinoma, which is CK7 negative and CK20 positive.19 Use of CK7 and CK20 immunostains thus allows approximately one-half of cases of mucinous carcinomas (ie, those from the gastrointestinal tract) to be effectively eliminated from consideration.19 Other CK7-positive and CK20-negative tumors such as adenocarcinoma of the lung and gallbladder may metastasize to the skin19; these can be differentiated from a primary skin tumor based largely on clinical investigation with potential assistance from other immunohistochemical stains.
The expression of p63 by primary cutaneous mucinous carcinoma and its diagnostic utility are controversial.19- 24 When p63 immunostaining highlights a peripheral layer of myoepithelial cells as exemplified by our patient 4 (Figure 5), this is considered diagnostic of an in situ component of the tumor and conclusive evidence of an origin in the skin.19,23 The significance of p63 expression by invasive tumor cells, as seen in all 3 of our cases in which immunostaining was performed (Table 2), is less certain because occasional cutaneous metastases from breast22,24 and lung24 cancer may express this protein. To further analyze the p63 staining characteristics of our cases of primary mucinous carcinoma of the eyelid, we stained 1 case each of mucinous breast and colon adenocarcinoma metastatic to the skin as well as a case of primary vulvar mucinous adenocarcinoma arising from a Bartholin gland (Table 2). The selective staining of primary mucinous carcinoma of the skin by p63 in this limited study indicates that additional investigation of mucinous carcinoma metastatic to the skin will be worthwhile. In the meantime, we concur with Kazakov and colleagues that “full clinical investigation remains the gold-standard approach to establish the origin” of mucinous carcinoma in the skin.19
Primary mucinous carcinomas of the eyelid are low-grade tumors. Morbidity in these tumors is associated with incompletely resected lesions, and tumors located at the inner canthus may be especially difficult to excise completely.12,25 Rarely, primary cutaneous mucinous carcinoma of the eyelid will metastasize to preauricular or submandibular nodes.5,26 As noted earlier, paramount in the clinical evaluation of patients with cutaneous mucinous carcinoma are a detailed history, physical examination, and appropriate radiographic studies to eliminate the possibility of a visceral mucinous tumor (mainly from the breast and intestinal tract19) that may have manifested with cutaneous metastasis. The recurrence rate of these tumors located on the eyelids has traditionally been estimated at 40% based mainly on the 1979 study by Wright and Font.5 As recently as 2004, an estimated recurrence rate of 34% in the eyelid27 was stated, even though that recurrence rate was generated in a study published in 1992.7
In 1988, Weber et al28 first reported the use of Mohs micrographic surgery for the treatment of primary mucinous carcinoma of the eyelid. Since then, 30 cases of primary mucinous carcinoma of the eyelid (including the current cases and 2 lateral canthal lesions) have been published (Table 3).7,14,18,19,23,27- 36 The rate of reported recurrence derived from the cases in which follow-up is provided is 26% (6 of 23 cases) (Table 3), with an average follow-up period of 26.3 months (range, 1-158 months).7,14,27,29,33 This recurrence rate is significantly less than that reported previously for primary mucinous carcinoma of the eyelid.5,7 Three of the cases that recurred after conventional excision were subsequently treated successfully by Mohs surgery with average disease-free follow-up of 52 months (range, 36-60 months).7,27,28 Of the cases reported since 1988 (including our patients 1, 2, and 3) whose primary or recurrent tumor was treated either by Mohs surgery7,27,31,32,36 or with frozen section control,29,34 only 1 of 14 tumors (7%) recurred (Table 4),29 with a mean disease-free follow-up of 36.5 months (range, 6-158 months). Although the numbers are small, these data suggest that if frozen section control or Mohs micrographic surgery is used, the recurrence rate of primary mucinous carcinoma of the eyelid is much less than that reported previously. However, prolonged follow-up for recurrence is warranted owing to the small number of reported cases excised with frozen section control of margins or using Mohs micrographic surgery and to the potential for late local recurrence by these tumors.
Correspondence: Alan D. Proia, MD, PhD, Department of Pathology, Duke University Medical Center, DUMC 3712, Durham, NC 27710 (email@example.com).
Submitted for Publication: August 24, 2009; final revision received February 21, 2010; accepted February 23, 2010.
Financial Disclosure: None reported.
Additional Contributions: Michael Richard, MD, Duke University Eye Center, provided the clinical photograph of patient 2. R. Dax Hawkins, MD, Eye Associates of Wilmington, Wilmington, North Carolina, provided the clinical synopsis for patient 3. Steven R. Conlon and Susan Reeves prepared the photomicrographs for submission.