Purpose
To describe the clinical and microscopic features of pigmented basal cell carcinomas (pBCC) of the eyelid.
Design
Retrospective observational case series collected at one institution.
Methods
An analysis of clinical records, photographs, and histopathologic characteristics of 257 BCCs with a review of the literature. The frequencies of clinically pigmented, and of microscopically pigmented but clinically nonpigmented, BCCs were determined. Cytochemical stains (Fontana-Masson, Prussian blue) and immunohistochemical probes (S-100, microphthalmia-associated transcription factor [MiTF], HMB-45, MART-1, CK20, synaptophysin, chromogranin, CD1a, Ki-67) were then employed and the findings correlated with the degree of clinical pigmentation.
Results
Histopathologically, 13 of 257 cases (5.06%) were found to have pigment; of these 13, 6 (all white patients) had clinically apparent pigmentation (2.33%), either focal or diffuse. Eight of 13 lesions developed on the lower eyelids. All stained positively for melanin but negatively for iron. MiTF highlighted numerous melanocytic nuclei in the tumor lobules, while MART-1 and HMB-45 revealed the dendritic shapes of the entrapped melanocytes. There was a subtotal blockage of melanin transfer to the surrounding basaloid cells. Intralobular S-100-positive cells included CD1a-positive Langerhans cells, while CK20 did not identify any Merkel cells.
Conclusions
Only 1 of 6 lesions was uniformly clinically pigmented, whereas the other 5 were only focally brown-black. The clinical pigmentation was imparted by varying densities and distributions of melanocytes with arborizing dendrites, which were present in all BCCs. Melanophages within the stroma and basaloid cell melanization also contributed to pigmentation. No behavioral or biologic differences in pBCC were documented compared with clinically nonpigmented lesions.
Basal cell carcinoma (BCC) is the most common cutaneous epithelial malignancy, almost exclusively originating in hair-bearing skin. In the largest series of excised benign and malignant eyelid lesions (5504 specimens), BCC comprised 14% of all lesions and 86% of malignancies. Conjunctival BCC arising at the limbus or mucocutaneous junction is vanishingly rare, and its pigmented variant is unrecognized to date. The classical constituent “basaloid” cells of BCC are small and ovoid or spindled, with a high nuclear-to-cytoplasmic ratio, thereby creating the microscopic appearance of a hematoxylophilic or “blue” tumor; they lack the cytoplasmic eosinophilia of epidermoid (squamous) cells. The tumors are characteristically organized into variably sized lobules exhibiting peripheral palisading and clefting artifact. Some foci of BCC display cells with more ample eosinophilic cytoplasm; such tumors have been termed metatypical if the cells are of intermediate size, or basosquamous if the cells are more ample (the latter has a greater metastatic potential).
Over 20 architectural and cytologic subtypes of BCC have been described, ranging from the relatively common nodular, micronodular, superficial (pseudomultifocal), adenoidal (pseudoglandular), cystic, and infiltrating examples to the extremely rare expressions of granular cell, signet ring cell, clear cell, pleomorphic (giant cell), fibroepitheliomatous (of Pinkus), ductular, neuroendocrine, trabecular, and amyloid-producing tumors. Only the ductular variety seems to have a predilection for the eyelids . Not infrequently confused with BCC microscopically are much more ominous hematoxylophilic tumors that have a strong metastatic potential. The most common examples of such lesions are nodular sebaceous carcinoma, which often exhibits a yellow clinical coloration (attributable to multivacuolated cytoplasmic lipid) occurring together with a unilateral blepharoconjunctivitis, and nodular Merkel cell carcinoma (neuroendocrine or trabecular carcinoma), with a violaceous hue and subcellular neurosecretory, dense-core granules representing neuropeptides.
BCC is typically found in the lower eyelids or medial canthus. Clinically and histopathologically, the most common variants are the nodulo-ulcerative, micronodular, superficial, and infiltrating. Some confusion still persists regarding the terminological distinction between infiltrating and sclerosing BCC. In the former there are elongated, narrow strands of basaloid cells situated between bundles of loose collagen with a minimal increase in the number of fibroblastic nuclei but without compaction and augmentation of the collagen. This pattern can be encountered deeper in the dermis beneath more superficial nodular or micronodular masses. The sclerosing type is also referred to as fibrosing, scirrhous, desmoplastic, and morpheic; it displays compressed basaloid strands that are embedded in a dense fibrous stroma, often resembling a keloid. Pigmentation can be seen in all types of BCC except for the infiltrating and sclerosing subtypes.
While conjunctival pigmented squamous dysplasias, carcinomas in situ, and invasive carcinomas have been well described, palpebral pigmented basal cell carcinoma (pBCC) has not been separately designated in a large histopathologic series and has been described only infrequently from a clinical perspective, with little attention paid to cytomorphologic composition. It is alleged to be more frequently, but not exclusively, encountered in patients with darker complexions (Mediterranean, Asian, Hispanic, and North American black patients). The origin of the pigment has been demonstrated to arise from dendritic melanocytes located within the tumor, rather than being intrinsically synthesized within the basaloid tumor cells themselves. The melanocytes in the tumor are usually inconspicuous in hematoxylin-eosin-stained sections, but special stains demonstrate increased numbers of melanosomes within ramifying dendritic networks. A relative but not absolute blockage exists in the normal transfer of the pigment via the dendritic processes of the melanocytes to the surrounding basaloid tumor cells.
We reviewed microscopically all cases of BCC on file in our laboratory collected over a 3-year period in order to determine the incidence of pBCC referred to our specialty hospital. We subsequently correlated the histopathologic findings with the clinical presence or absence of pigmentation, and compared our results with those previously published for pBCC in the ophthalmic and dermatologic literatures. In addition to the identification of intralobular and intrastromal melanophages, an appreciation of the role played by the density of intratumoral melanocytes, the scaffolding created by their dendrites, and the extent of cytoplasmic accumulation of melanin granules by the basaloid cells was enhanced by the application of special cytochemical stains and immunohistochemical techniques.
Methods
A retrospective observational case series of all cases diagnosed as BCC accumulated over a 3-year period was performed using the diagnostic files of the Cogan Ophthalmic Pathology Laboratory of the Massachusetts Eye and Ear Infirmary, Boston, Massachusetts. Pathologic glass slides that were in the files were retrieved and reviewed in all cases to establish the accuracy of the diagnoses, and in questionable cases were shown to a dermatopathologist at the Massachusetts General Hospital. Information about the patient’s complexion and the degree of clinical lesional pigmentation was obtained from clinicians’ medical records, clinical photographs, or direct interviews with the clinicians. Cases with insufficient clinical information or inadequate pathologic sections were excluded. A total of 257 BCCs were found that satisfied the criteria for inclusion in this investigation. The incidence of microscopic pigmentation was documented in hematoxylin-eosin-stained sections. When blocks were available, special stains to identify focal or diffuse cellular pigmentation were employed to differentiate between melanin (Fontana-Masson) and hemosiderin (Prussian blue iron stain). Immunohistochemistry, conducted in the Diagnostic Immunopathology Division of the Department of Pathology at the Massachusetts General Hospital, Boston, Massachusetts, was likewise performed on additional sections from the blocks using probes for S-100, microphthalmia-associated transcription factor (MiTF), MART-1, HMB-45 (all for detecting melanocytes); pancytokeratin (for the intermediate cytoplasmic filaments of squamous or epidermal cells); Ber-EP4 (BCC cells); CD1a (Langerhans cells); CK20, synaptophysin, chromogranin (all for detecting Merkel cells); and Ki-67 (for highlighting nuclei in S-phase of DNA replication prior to mitosis). Whereas MiTF disclosed the presence of melanocytes by staining their nuclei, HMB-45 and MART-1 brought out their cytoplasmic shapes and prominence of their dendrites. Seven nonclinically and nonmicroscopically pigmented tumors were also immunostained to serve as controls. A search of the published literature indexed in PubMed was performed without any limit on date or language of publication using the following keywords in various combinations: pigmented basal cell carcinoma, eyelids, skin, melanocytes, dendrites, immunohistochemistry, epidermal-melanin unit. Studies relevant to the main aspects of our investigations were selected for citation and served as a backdrop for interpreting our results.
Results
Clinical Findings
Of the 257 cases selected with slides of eyelid BCCs retrievable for review, 13 lesions contained variable amounts of pigment histopathologically. All of these cases had available clinical information. In 10 cases in which the paraffin blocks could be located, enough tissue remained to permit the special stains and immunohistochemical studies. In only 6 patients did clinically detectable pigmentation of an eyelid BCC nodule correlate with the histopathologic presence of melanin. All of these patients were white. There were 4 men and 2 women, with a median age of 72 (range, 54–82). Four of the cases were located on the lower eyelids near the eyelashes or mucocutaneous junction (2 on the right and 2 on the left lower eyelids) and 2 on the upper eyelids. In several lesions elevated edges could be discerned. The amount of clinical pigment in the lesions was variable, ranging from diffuse (1 case) to focal (5 cases) ( Figure 1 , Top left and right and Middle left). Of the 7 patients with subclinical microscopic pigmentation (not of sufficient degree to create clinically pigmented lesions), 5 were female and 2 were male. The median age was 69 years (range, 42–79). Four of these lesions were found on the lower eyelids and 3 on the upper eyelids. No recurrences after anterior lamellar dissections (10 cases) were reported, with follow-ups ranging from 3 to 41 months (median 36). The size of the excised specimens ranged from 0.9 × 0.7 × 0.3 cm to 0.2 × 0.1 × 0.1 cm.
Histopathologic and Immunohistochemical Findings
Three patients underwent full-thickness eyelid resections and 10 had anterior lamellar excisions. The 13 lesions of BCC were classified as nodular (4), combined nodular and micronodular (3), micronodular (2), adenoidal (2), superficial (1), and nodular with metatypical features (1) ( Figure 1 , Middle right). On standard hematoxylin-eosin-stained sections, irregular foci of pigmentation located in melanophages or basaloid tumor cells were haphazardly discovered, usually within a subset but not all of the tumor cell lobules, cords, or stromal strands in any given lesion ( Figure 1 , Bottom left and right). Such pigmentation was generally near the epidermis rather than more deeply in the dermis. Rarely in hematoxylin-eosin-stained sections one could detect slender, dendritic melanocytic processes insinuated among the basaloid tumor cells, whereas the melanophages found in the stroma were rounded or polyhedral with obscured nuclei because of the coarse clumping of the engulfed melanin granules ( Figure 1 , Middle right and inset). The 1 diffusely pigmented lesion had a central zone of stromal melanophages ( Figure 1 , Middle right), perhaps the result of earlier necrosis. Melanophages engorged with coarsely agglutinated melanin granules were also observed in thin strands of stroma or in the setting of small foci of intratumoral necrosis in many of the lesions ( Figure 2 , Top left and right). The Fontana-Masson stain was positive for melanin by usually changing its coloration from brown to black ( Figure 2 , Middle left) in the melanophages, and by highlighting small numbers of melanin granules in some of the basaloid tumor cells. The Prussian blue stain for iron was negative (no blue reaction product was seen) in all clinically pigmented and nonpigmented tumors that contained microscopic pigment ( Figure 2 , Middle right).
MiTF staining revealed many more melanocytic nuclei within the tumoral units than could be appreciated in standard hematoxylin-eosin-stained sections ( Figure 2 , Bottom left). These melanocytes were either randomly scattered within the nodules (some micronodules had MiTF positively staining nuclei, signaling the presence of melanocytes, whereas other nests in the same lesion often lacked positivity), or regularly positioned among the outermost basaloid cells when a finger-like, micronodular, or lacy pattern was adopted. The melanocytes were scattered throughout the large nodules, instead of being restricted to the periphery as in the smaller units ( Figure 2 , Bottom right). MiTF-positive-staining melanocytic nuclei were present in both the microscopically pigmented and nonpigmented control tumors. In the nonpigmented lesions there were fewer such nuclei within the lobules, and micronodules often completely lacked positive staining.
As one proceeded more deeply into the invasive tumor the number of melanocytes tended to diminish or totally disappear. According to the results of HMB-45-positive-staining melanocytes, which more clearly than MART-1 outlined the cytoplasmic shapes of the melanocytes, we divided the presence of intratumoral melanocytes into 3 categories: dense, intermediate, and sparse. The uniformly pigmented lesion was the only one to display a dense melanocytic network with elaborate engorged dendrites ( Figure 3 , Top left and inset). This pattern was not manifested in the hematoxylin-eosin-stained sections ( Figure 1 , Middle right). The lesions with intermediate melanocytic densities displayed smaller, less engorged dendrites and clinically exhibited only focal pigmentation. This melanocytic category was found in the 5 clinically partially pigmented lesions and in 1 of the allegedly clinically nonpigmented lesions ( Figure 3 , Top right and inset). In all of the nonpigmented portions of the lesions and in all of the control specimens small numbers of HMB-45-positive melanocytes displayed blunted or inconspicuous dendrites ( Figure 3 , Middle left and right).
Pancytokeratin strikingly highlighted the epidermis but was negative in the basaloid nests, although focally positive in 1 case exhibiting intermediate epidermal differentiation (a metatypical lesion) ( Figure 1 , Middle right). Ber-EP4 was positive within the basaloid tumor cells but not in the metatypical region or intratumoral melanocytes. If S-100 positivity within the tumor lobules was greater than expected from the results of MiTF nuclear immunostaining for melanocytes, CD1a was used and revealed the presence of numerous positively staining dendritic Langerhans cells dispersed randomly throughout the tumor nests ( Figure 3 , Bottom left). Merkel cells could not be identified immunohistochemically within either the pigmented or nonpigmented BCCs using CK20, chromogranin, and synaptophysin probes. Ki-67 was consistently more strikingly positive within the tumors than among the basal germinal cells of the intact overlying epidermis, and was most conspicuous toward the periphery of the lobules than in their centers ( Figure 3 , Bottom right).
Results
Clinical Findings
Of the 257 cases selected with slides of eyelid BCCs retrievable for review, 13 lesions contained variable amounts of pigment histopathologically. All of these cases had available clinical information. In 10 cases in which the paraffin blocks could be located, enough tissue remained to permit the special stains and immunohistochemical studies. In only 6 patients did clinically detectable pigmentation of an eyelid BCC nodule correlate with the histopathologic presence of melanin. All of these patients were white. There were 4 men and 2 women, with a median age of 72 (range, 54–82). Four of the cases were located on the lower eyelids near the eyelashes or mucocutaneous junction (2 on the right and 2 on the left lower eyelids) and 2 on the upper eyelids. In several lesions elevated edges could be discerned. The amount of clinical pigment in the lesions was variable, ranging from diffuse (1 case) to focal (5 cases) ( Figure 1 , Top left and right and Middle left). Of the 7 patients with subclinical microscopic pigmentation (not of sufficient degree to create clinically pigmented lesions), 5 were female and 2 were male. The median age was 69 years (range, 42–79). Four of these lesions were found on the lower eyelids and 3 on the upper eyelids. No recurrences after anterior lamellar dissections (10 cases) were reported, with follow-ups ranging from 3 to 41 months (median 36). The size of the excised specimens ranged from 0.9 × 0.7 × 0.3 cm to 0.2 × 0.1 × 0.1 cm.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
