Mucin-producing Sweat Gland Carcinoma of the Eyelid: Diagnostic and Prognostic Considerations


To describe the clinical and pathologic characteristics of mucin-producing sweat gland carcinoma of the eyelid and to determine whether neuroendocrine differentiation is of prognostic significance.


Retrospective interventional case series.


Search of the New York Eye and Ear Infirmary pathology database between 1990 and 2011 identified 16 patients with mucin-producing sweat gland carcinoma. Clinical, histopathologic, and immunohistochemical analyses were performed on all identified cases.


The patients presented with vascularized, focally cystic, nonulcerated eyelid margin lesions. Histopathologic evaluation showed that 4 lesions (25%) had a cystic, papillary, and solid growth pattern with an in situ component, 7 (44%) were pure invasive mucinous carcinomas, and 5 (31%) demonstrated both growth patterns. Immunohistochemical analysis of 15 tumors showed that pure cystic/papillary lesions had a significantly greater percentage of synaptophysin-immunoreactive cells ( P = .036). There was no significant difference in the number of neuroendocrine markers expressed or in the intensity of immunostaining among the 3 different growth patterns. Re-excision for margin clearance was performed in 8 of 13 cases (61.5%). Two of 13 lesions recurred (15%); 1 of these was an in situ tumor with cystic morphology and neuroendocrine differentiation and the other was pure invasive mucinous carcinoma. None of the lesions metastasized.


Mucin-producing sweat gland carcinoma pathologically represents a continuum, from an in situ lesion to a classic, invasive mucinous carcinoma. Immunohistochemical evidence of neuroendocrine differentiation can be observed in all lesions and does not appear to have a prognostic significance, arguing against the utility of immunohistochemical subtyping of mucinous sweat gland carcinomas.

In the past decade, there has been a considerable body of literature focusing on characterization of endocrine mucin-producing carcinoma as a distinct variant of cutaneous mucinous carcinoma. This work has been driven by the developments in the field of breast pathology, which have identified morphologically, immunohistochemically, and biologically distinct subtypes of mammary mucinous carcinoma.

Recent studies have suggested that mucinous carcinomas of the skin have morphologic, immunophenotypic, and biologic heterogeneity paralleling mucinous carcinomas of the breast. Hanby and associates noted 3 distinct morphologic patterns in mucinous sweat gland carcinomas. One tumor variant was composed of rings, ribbons, and cords of cells surrounded by abundant mucin, analogous to type A mucinous carcinoma of the breast. The second variant resembled type B mucinous breast carcinoma with cystic, papillary, and solid growth patterns, and the third variant had mixed morphology. Grimelius stain was positive in the 2 latter mucinous carcinoma variants, consistent with neuroendocrine differentiation in these lesions, which became known as endocrine mucin-producing sweat gland carcinoma.

Subsequent studies focused on further elucidation of clinical, morphologic, and immunohistochemical features of endocrine mucin-producing sweat gland carcinoma. Twenty-one well-documented cases of endocrine mucin-producing carcinoma have been reported in the literature to date. Predilection for older age and female sex has been observed. All reported tumors demonstrated characteristic morphology and immunoreactivity for at least 1 neuroendocrine marker (synaptophysin, chromogranin, neuron-specific enolase, CD56, and CD57). In the largest series of endocrine mucin-producing carcinomas to date, Zembowicz and associates postulated that endocrine mucin-producing carcinoma is a stage in evolution from an in situ tumor to an invasive mucinous carcinoma with neuroendocrine differentiation. Lack of tumor recurrence in their study led the authors to suggest that this mucinous sweat gland carcinoma variant has a favorable prognosis.

In this study we describe the clinical and pathologic characteristics of 16 cases of mucin-producing sweat gland carcinoma of the eyelid and investigate whether neuroendocrine differentiation is associated with a specific tumor growth pattern and if it is of prognostic significance.


The prospective New York Eye and Ear Institutional Review Board approval was obtained to review all clinical and pathology patient records for this retrospective interventional case series. The New York Eye and Ear Infirmary pathology database was searched to identify all histopathologically diagnosed cases of cutaneous mucinous carcinoma between January 1, 1990 and January 1, 2011. Patients’ clinical data, including age, sex, tumor location and appearance, adequacy of surgical excision, and evidence of recurrence or metastases was collected.

Histopathology and immunohistochemistry were reviewed. The parameters evaluated included in situ vs invasive distribution (based on tumor morphology and immunostaining for myoepithelial markers smooth muscle actin and p63), evidence of sweat gland differentiation (based on immunoreactivity for low-molecular-weight cytokeratin, carcinoembryonic antigen, gross cystic disease fluid protein 15 (BRST-2), and estrogen and progesterone receptors), neuroendocrine differentiation (based on immunoreactivity for synaptophysin, chromogranin, CD56, and neuron-specific enolase), and proliferative activity (Ki-67 immunolabeling index). Immunoreactivity was scored semi-quantitatively for staining intensity and for number of immunoreactive cells as “0” (no staining), “1+” (weak intensity / ≤25% immunoreactive cells), “2+” (moderate intensity / between 26% and 50% immunoreactive cells), “3+” (strong intensity / between 51% and 75% immunoreactive cells), and “4+” (immunoreactivity in nearly all cells).

Statistical analysis using Fisher exact test and analysis of variance was carried out to determine the strength of relationship between the tumor growth pattern and the number of neuroendocrine markers expressed, intensity of immunostaining, and the percentage of immunoreactive cells.


The pathology database search yielded 6975 eyelid biopsies, performed between January 1, 1990 and January 1, 2011. Of these, 16 cases (0.2%), diagnosed as mucinous sweat gland carcinoma, were selected for review. Clinical, histopathologic, and immunohistochemical data are summarized in the Table .


Clinical, Histopathologic, and Immunohistochemical Characteristics of Mucin-producing Sweat Gland Carcinomas of the Eyelid

Clinical Characteristics Growth Pattern Neuroendocrine Markers a
Age/Sex Site Size (cm) Appearance Clinical Suspicion EMPSGC IMSGC NSE Syn Chr CD56 Margin Re-excision Recurrence
1 69 M LLL 0.4 × 0.3 × 0.3 Horny, papilloma BCCA vs sweat duct CA No Yes 2+/4+ 2+/1+ 2+/1+ 0 Positive Yes No at 6 y
2 64 F LLL 0.6 × 0.6 × 0.3 Pigmented, raised, poorly circumscribed Melanoma Yes No 1+/1+ 0 0 0 Positive NA NA
3 72 M Lid 0.4 × 0.4 × 0.3 NA NA Yes Yes 2+/3+ 0 0 0 Positive No No at 8 mo
4 71 M LLL 0.8 × 1 Well-circumscribed, raised, vascularized, pearly Chalazion vs sebaceous CA vs BCCA vs SCCA No Yes 3+/2+ 0 3+/4+ 0 Positive No No at 5 y
5 68 M RLL 0.5 × 0.4 × 0.4 NA Sebaceous cyst No Yes 1+/1+ 0 3+/2+ 2+/2+ NA NA NA
6 61 F LLL 0.5 × 0.5 × 0.4 NA NA No Yes 2+/3+ 0 3+/2+ 0 Positive No No at 4 d
7 76 M RUL 1.2 × 0.8 × 0.5 NA NA No Yes NA NA NA NA Positive Yes Yes at 10 mo
8 70 M LUL 0.8 × 0.5 × 0.3 Pedunculated BCCA No Yes 2+/4+ 0 3+/3+ 0 Positive No No at 2 mo
9 80 M RUL 0.4 × 0.2 × 0.2 NA BCCA No Yes 1+/3+ 0 0 2+/2+ Positive NA NA
10 53 M RUL 0.2 × 0.2 × 0.2 NA BCCA Yes No 2+/4+ 1+/3+ 1+/3+ NA Positive Yes No at 2 mo
11 70 F RUL 0.6 × 0.4 × 0.4 Cystic BCCA vs Merkel cell CA Yes Yes 2+/4+ 1+/1+ 3+/1+ 2+/1+ Positive No No at 1 y
12 61 M LUL 0.2 × 0.2 × 0.1 Well-circumscribed, no ulceration NA Yes Yes 1+/3+ 0 0 0 Positive Yes No at 1.5 y
13 78 F RUL 0.4 × 0.3 × 0.1 Pedunculated cystic BCCA Yes No 1+/3+ 2+/2+ 2+/1+ 0 Positive Yes–twice Yes at 3 y
14 87 M LLL 0.2 × 0.2 × 0.1 Well-circumscribed, vascularized BCCA Yes No 3+/3+ 2+/4+ 3+/4+ 0 Positive Yes No at 10 mo
15 66 F LUL 0.6 × 0.6 × 0.6 NA Cyst/wart Yes Yes 1+/2+ 0 0 2+/1+ Positive Yes No at 8 mo
16 69 M LUL 0.8 × 0.6 × 0.4 NA Atypical chalazion Yes Yes 1+/1+ 2+/2+ 0 3+/3+ Positive Yes No at 4 mo

BCCA = basal cell carcinoma; CA = carcinoma; Chr = chromogranin; EMPSGC = characteristic endocrine mucin-producing sweat gland carcinoma; F = female; IMSGC = characteristic invasive mucinous sweat gland carcinoma; LLL = left lower lid; LUL = left upper lid; M = male; NA = not available; NSE = neuron-specific enolase; RLL= right lower lid; RUL = right upper lid; SCCA = squamous cell carcinoma; Syn = synaptophysin.

a Immunoreactivity scoring (staining intensity/percentage of immunoreactive cells): 0 = none; 1+ = weak/1%-25% of cells; 2+ = moderate/26%-50% of cells; 3+ = strong/51%-75% of cells; 4+ = immunoreactivity in nearly all cells.

Demographic and Clinical Data

The mean patient age at diagnosis was 70 years (range, 53-87 years). Eleven patients (69%) were male. Specific lesion location was available in 15 patients. Sixty percent (9/15) of the lesions involved left periocular skin. The upper eyelid was involved in 9 of 15 patients (60%). The lower eyelid was involved in the remaining 6 patients (40%). The lesions presented as nodular or pedunculated, vascular, occasionally cystic, nonulcerated growths at the eyelid margin, ranging in size from 1 mm to 1.2 cm ( Figure 1 , Table ).

Figure 1

Representative clinical photographs of patients with mucinous sweat gland carcinoma of the eyelid. (Left) Patient 4. (Middle) Patient 14. (Right) Patient 11. The tumors appear as nodular, flesh-colored, mildly lobulated, nonulcerated eyelid margin lesions with focally telangiectatic vessels. Translucent clear or bluish foci, suggestive of a cystic component, are present in all lesions. Two lesions are associated with madarosis (left and right photographs).

Histopathology and Immunohistochemistry

Histopathology was available for review in 16 cases ( Table ). Morphologic findings of endocrine mucin-producing carcinoma, characterized by cystic and papillary/solid growth patterns, were identified in 9 of 16 lesions (56%). Four lesions (25%) were composed entirely of cystic and papillary/solid growth patterns with an identifiable in situ component, 7 (44%) were composed entirely of invasive mucinous carcinoma nests suspended in abundant pools of mucin, and 5 (31%) demonstrated both growth patterns ( Figures 2-4 ). The neoplastic cells in all tumors had mild nuclear pleomorphism and scant mitotic activity ( Figures 2-4 ). All lesions were initially excised with positive surgical margins.

Figure 2

Histopathologic and immunohistochemical findings of recurrent endocrine mucin-producing carcinoma in situ of the eyelid in Patient 13. (Top left) Histopathologic evaluation of the primary biopsy demonstrates a cystic lesion, lined by mildly pleomorphic cuboidal cells, arranged in a cribriform pattern with focal intracellular and extracellular mucin (asterisk). Collection of luminal keratin is also observed, consistent with proximity to intraepidermal segment of sweat gland duct. (Top middle) Histopathologic evaluation of recurrent tumor similarly demonstrates a cystic lesion originating from a sweat gland duct (arrow). (Top right) Higher-magnification photomicrograph of recurrent lesion shows that the cyst is lined focally by a hidrocystoma-type bilayer, and focally by a multilayered epithelium with nuclear resetting and intracytoplasmic and extracytoplasmic mucin (arrows). Immunohistochemical evaluation of the primary tumor demonstrates focal immunoreactivity of the neoplastic cells for carcinoembryonic antigen antibodies (Middle left) and BRST-2 (Middle center). Myoepithelial cells at the periphery of neoplastic epithelium are highlighted with smooth muscle actin (Middle right; arrows) and p63 immunostains (Bottom left; arrows). (Bottom middle) There is a moderate, patchy cytoplasmic immunoreactivity for synaptophysin. (Bottom right) Ki-67 immunostain labels approximately 10% of neoplastic cells. (Top row: hematoxylin-eosin stain; Middle left: carcinoembryonic antigen antibodies; Middle center: BRST-2 antibodies; Middle right: smooth muscle actin antibodies; Bottom left: p63 antibodies; Bottom middle: synaptophysin antibodies; Bottom right: Ki-67 antibodies. Top middle, original magnification ×25; Middle right, original magnification ×100; all other images, original magnification ×50.)

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Jan 9, 2017 | Posted by in OPHTHALMOLOGY | Comments Off on Mucin-producing Sweat Gland Carcinoma of the Eyelid: Diagnostic and Prognostic Considerations

Full access? Get Clinical Tree

Get Clinical Tree app for offline access