Pathology (February 2015) 47(2), pp. 130–133


CD117 and CD43 are useful adjuncts in the distinction of adenoid cystic carcinoma from adenoid basal cell carcinoma B. F. DESSAUVAGIE



Department of Tissue Pathology, PathWest Laboratory Medicine, QEII Medical Centre, Nedlands, and School of Pathology and Laboratory Medicine, University of Western Australia, Crawley, WA, Australia

Summary Distinction of cutaneous adenoid cystic carcinoma (ACC) from adenoid basal cell carcinoma (BCC) is an occasional diagnostic dilemma in dermatopathology. We examined the immunohistochemical staining patterns with CD117 and CD43 in ACCs and BCCs, including BCCs with an adenoid growth pattern, to determine whether a combination of these markers can assist in the differential diagnosis. Fifteen cases each of ACC and BCC, including seven BCCs with a partial or entirely adenoid growth pattern were immunohistochemically stained for CD117 and CD43. The stains were interpreted semi-quantitatively. Staining for CD43 and CD117 was significantly more common in ACC than in BCC. Forty percent of ACCs showed staining for CD43, while no cases of BCC were positive. CD117 was positive in all cases of ACC, with 93% showing moderate or strong staining. BCC were less frequently positive, with only 20% of cases showing labelling of weak or moderate intensity. Immunohistochemical positivity for CD117 and CD43 are likely to be helpful adjuncts in the separation of cutaneous ACC from adenoid BCC. Key words: Adenoid cystic carcinoma, basal cell carcinoma, CD117, CD43, dermatopathology. Received 11 August, revised 5 September, accepted 17 September 2014

INTRODUCTION The morphological distinction of cutaneous adenexal tumours from basal cell carcinoma (BCC) is a frequent diagnostic challenge in dermatopathology. This differential has many guises, one of which is separating cutaneous adenoid cystic carcinoma (ACC) from BCC, particularly those with an ‘adenoid’ growth pattern. Cutaneous ACC is histologically identical to its more common salivary gland counterpart. ACC is characterised by cribriform and less often tubular or solid formations of admixed epithelial and myoepithelial cells. The latter are typically numerous and their round to angular nuclei, scant cytoplasm and indistinct cytoplasmic borders impart a basaloid appearance. These cells surround the punched out pseudoglandular spaces of the characteristic cribriform structures, which contain either eosinophilic basement-membrane like material or basophilic ground substance. Occasional ‘true’ glands, containing eosinophilic secretory products, are lined by low cuboidal epithelial cells with eosinophilic cytoplasm. Cutaneous ACC, like those of the salivary gland and elsewhere, share a propensity for perineural invasion.1,2 Print ISSN 0031-3025/Online ISSN 1465-3931 DOI: 10.1097/PAT.0000000000000209


Basal cell carcinomas are the most common cutaneous malignancy. They are characterised by islands and nests of hyperchromatic basaloid cells with scant, ill-defined pale cytoplasm. In the adenoid variant the tumour islands have a reticulated pattern, resulting in rounded collections of basophilic ground substance, mimicking ACC.3 Histological clues to separate these entities include peripheral palisading of tumour cells, stromal retraction around invasive tumour islands and epidermal attachment, which are features of BCCs not shared with cutaneous ACC. Previous studies have found that staining for EMA, CAM5.2, S100 and CEA is generally positive in ACC but negative in BCC.2,3 The distinction is important. Following complete excision up to 50% of cutaneous ACC may recur.1 In a recent series three of 18 exhibited metastatic potential.1 This contrasts with BCC where, in the absence of aggressive phenotype, perineural invasion and/or involved margins, the outlook following complete excision is excellent, with a low risk of local recurrence and a negligible risk of metastasis.3 We investigated whether markers considered specific for ACC, including CD117 and CD43 could be of value in distinguishing ACC from adenoid BCC.

MATERIALS AND METHODS The pathology archives of PathWest Laboratory Medicine at QEII site were searched for cases of ACC arising at any site and of BCC with a description of adenoid features between the years 1997 and 2014. BCC numbers were supplemented with unselected routine BCC cases. All slides were reviewed by a dermatopathologist (BAW) to confirm the diagnoses. Clinical data were obtained from pathology request forms. Immunohistochemical (IHC) analyses were then performed on additional 4–6 mm sections mounted on positively charged slides from each case on the Ventana BenchmarkXT automated staining machine (Ventana Medical Systems, USA). The processor deparaffinised the slides, carried out antigen retrieval utilising the on-board CCI retrieval solution, and incubated the slides in primary antibody. CD117 (T959 clone, dilution 1:100; Leica Biosystems, Germany) was incubated for 40 min at room temperature with on-board amplification and CD43 (DFT-1 clone, dilution 1:50; Dako, Denmark) was incubated for 32 min at 368C. Visualisation was by the on-board ultraview DAB detection kit. Sections were then counterstained with haematoxylin to provide nuclear detail. IHC staining for CD117 and CD43 was scored semi-quantitatively by estimating percentage tumour cell positivity as: 0, no tumour cell staining; 1þ, 0–5% tumour cells staining; 2þ, 6–49% cell staining; and 3þ, >50% cell staining.

RESULTS Clinical features Fifteen ACCs were identified. The patient age ranged from 19 to 81. Two primary cutaneous forehead ACCs were identified.

2015 Royal College of Pathologists of Australasia

Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.


Another dermal ACC was identified, however it was unclear whether this represented a primary cutaneous ACC or a metastatic tumour deposit. Six cases arose within major salivary glands, five had origin in minor salivary glands and one was a metastatic ACC to the liver (Table 1). Fifteen BCCs were identified. The patient age ranged from 40–86. Eleven cases were on the head and/or neck, two were on the trunk and two were from the upper limb (Table 2). Histological features All cases of ACC showed areas of classic morphology with at least partial cribriform growth containing alternating basophilic and eosinophilic material within the cribriform spaces and comprising a biphasic cell population (Fig. 1). All examples were of low grade, without significant mitotic activity, necrosis, areas of solid growth or dedifferentiation. Seven of the 15 BCCs exhibited complete or partial adenoid growth, characterised by nodules of basaloid cells with a cribriform growth pattern and rounded spaces containing basophilic material (Fig. 2). Where other growth patterns were encountered they included nodular, nodular-cystic, superficial and infiltrating. The remaining eight BCCs exhibited superficial (two cases), mixed superficial and nodular, mixed nodular and infiltrating (three cases), purely infiltrating and purely nodular-cystic growth. Immunohistochemistry Membrane staining with or without a cytoplasmic blush was considered positive for both CD117 and CD43. Positive staining for CD117 was at least focally observed in all cases of ACC, with semiquantitative scores ranging from 1þ to 3þ (Table 1). In 10 of the 15 cases this staining was extensive, with more than 50% of tumour cells positive (Fig. 3A). In contrast, only three BCCs exhibited CD117 membrane positivity and in all cases staining was focal (Fig. 3B). Care was taken not to count scattered positively stained melanocytes enmeshed within the lesional basaloid cells (Fig. 3C, inset). The sensitivity of CD117 for ACC was 100%, with a specificity of 80%. CD43 staining was focally observed in six of the 15 ACCs (Fig. 3C), with semiquantitative scores ranging from 0 to 2þ (Table 2). None of the BCCs exhibited any positivity for CD43 (Fig. 3D). The specificity of CD43 for ACC was 100%, with a sensitivity of 40%.


Table 2 Basal cell carcinoma: clinical features, histomorphology and immunohistochemical staining results Case no. Age Site 16


Right temple

17 18

68 87

19 20 21 22 23 24 25 26 27 28 29 30

59 83 57 63 59 40 43 47 86 71 79 82

Posterior neck Right postauricular Posterior neck Left upper lip Right temple Chest wall Left neck Right nasal bridge Left forehead Forehead Right forearm Left upper chest Right shoulder Right ear

Growth patterns Adenoid, nodular & superficial Adenoid & infiltrating Adenoid & nodular-cystic Adenoid Adenoid & nodular-cystic Adenoid & nodular Adenoid & nodular Nodular Infiltrating Nodular & superficial Nodular & infiltrating Superficial Nodular-cystic Superficial Nodular & infiltrating

CD117 CD43 staining staining 0


0 0

0 0

0 1þ 0

0 0 0

0 0 0 0 0 2þ 0 1þ

0 0 0 0 0 0 0 0

DISCUSSION The morphological distinction of cutaneous ACC from BCC can be difficult and given its increased propensity for recurrence and metastasis, correctly identifying ACC is clinically important. Therefore, we investigated two newer antibodies considered relatively specific for ACC, to determine their utility in this distinction. CD117 (also known as c-kit) is a transmembrane tyrosine kinase receptor, which upon activation initiates intracellular signals, ultimately resulting in cellular development and growth.4,5 CD117 expression is observed in a range of normal cells, including melanocytes, and in neoplasms, including gastrointestinal stromal tumour, mast cell tumours, seminoma and ACC.4,6 In the Mino et al. series of 66 ACCs and 98 other head and neck neoplasms, using two different CD117 clones, H300 (1:300 dilution; Santa Cruz Biotechnology, USA) and A4502 (1:100 dilution; Dako), almost all ACC exhibited positive cytoplasmic or membrane staining for either one or both CD117 clones.7 In contrast, the morphologically similar

Table 1 Adenoid cystic carcinoma: clinical features and immunohistochemical staining results

Case no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Age 53 56 77 61 67 56 38 34 64 19 81 66 43 43 61

Site Forehead Scalp Right submandibular Endobronchial tumour Left minor salivary gland Maxilla Metastatic deposit in liver Right parotid gland Right submandibular gland Left parotid gland Left submandibular gland Left submandibular gland Left endobronchial Umbilicus (primary/metastasis) Larynx

CD117 staining

CD43 staining

2þ 3þ 3þ 3þ 2þ 2þ 3þ 3þ 3þ 3þ 3þ 1þ 3þ 2þ 3þ

0 2þ 0 1þ 2þ 1þ 0 0 2þ 0 1þ 0 0 0 0

Fig. 1 Primary cutaneous adenoid cystic carcinoma (ACC) of the skin (Case 1) composed of bimorphic cells arranged in cribriform sheets punctuated by rounded spaces containing alternating basophilic and eosinophilic material (right) and tubular structures (central) (H&E).

Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.


Pathology (2015), 47(2), February


Fig. 2 Adenoid basal cell carcinoma (BCC) with a partial adenoid growth pattern (Case 17) characterised by nodules of basaloid cells with a cribriform growth pattern punctuated by rounded spaces containing basophilic material (H&E).

tumours [e.g., pleomorphic adenoma, basal cell adenoma and polymorphous low grade adenocarcinoma (PLGA)] were almost always negative, indicative of a relatively high specificity.7 These observations have application in the distinction of ACC from other morphologically similar salivary gland tumours.4,7 Expression of CD117 in BCC is less well documented. BCCs have been reported as positive for CD117 in 61 of 66 cases in one study6 and completely negative in 11 cases in another.7 In the former, Terada examined 66 consecutive BCCs with a CD117 polyclonal antibody (Dako) at an unspecified dilution. Positive membrane staining was observed in 61 cases.6 In contrast, the study by Mino et al., similar to our own, used

monoclonal antibodies (H300, 1:300 dilution, Santa Cruz Biotechnology; and A4502, 1:100 dilution Dako) and demonstrated a complete absence of staining of BCCs.7 These differing results may be due to the use of a polyclonal antibody and/or in staining interpretation. Of note in that regard, we do not see labelling of normal basal keratinocytes with CD117 in our material, as described by Terada.6 In regard to the latter, melanocytes, which are CD117 positive, are frequently seen as admixed single cells in many cutaneous neoplasms, including BCC.8 These were not interpreted as positive in our material. Our study, with the T959 clone (Leica Biosystems), confirms CD117 is highly sensitive for identifying ACC, with all 15 cases exhibiting positive staining, and highly specific when used in the distinction of ACC from BCC. CD43 (also known as sialophorin or leukosialin) is a transmembrane protein, of poorly characterised function, previously thought limited to expression in haematopoetic cells.9,10 More recently CD43 has been detected in cells of colon carcinoma and non-small cell lung carcinoma.9 Expression of CD43 has emerged as a relatively specific marker for ACC, with positivity ranging from 23.5–100% depending on the clone used.9,10 Seethala et al. examined expression of CD43 in 20 ACCs and 93 non-ACC tumours (42 salivary and 51 non-salivary) using two different clones, DF-T1 (DakoCytomation, Denmark; final concentration 2.10 mg/mL) and MT-1 (Novocastra, UK; dilution 1:50).9 Membranous staining was considered positive. Best results were associated with the MT-1 clone (9/20 ACCs positive) compared to the DF-T1 clone (4/20 ACCs positive). Only two of the 93 non-ACC tumours (a membranous-type basal cell adenocarcinoma positive with both clones and a colorectal adenocarcinoma positive only with the DF-T1 clone) were CD43 positive.9 Woo et al. examined CD43 expression with the L60 clone (Ventana Medical Systems, unspecified dilution) in 40 salivary gland tumours (12 ACCs, 14 PLGAs, and 14 monomorphic





Fig. 3 Immunohistochemical staining. (A) All ACCs exhibited at least focal membranous CD117 positivity. In this example (Case 15) widespread positivity was observed. (B) CD117 showed focal membranous positivity in only two cases of BCC (Case 20). Care was taken not to count scattered positively stained melanocytes amongst the lesional basaloid cells (inset, Case 18). (C) Focal membranous CD43 positivity was observed in five cases of ACC (Case 5). (D) All 15 BCCs were entirely negative for CD43 (Case 18).

Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.


adenomas).10 Membranous and cytoplasmic staining was considered positive. All 12 ACCs were positive while only one PLGA and three monomorphic adenomas were positive.10 The results of these two studies indicate CD43 has the potential to be highly sensitive and specific for ACC, depending on the clone employed. CD43 expression has not previously been examined in BCCs. Using the DF-T1 CD43 clone, six of our 15 ACCs were positive and all of our BCCs (regardless of subtype) were negative, corresponding to a perfect specificity, but suboptimal sensitivity. Sensitivity in our study could potentially be improved with use of an alternate CD43 clone, as best results thus far are associated with the MT-1 and L60 clones. In conclusion, ACC and adenoid BCC share morphological similarities, but have distinctly different clinical outcomes, necessitating their histopathological distinction. Positive CD117 and CD43 immunohistochemical staining, routinely available in most histopathology laboratories, is potentially useful in this distinction. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Address for correspondence: Dr Benjamin Dessauvagie, Department of Tissue Pathology, PathWest Laboratory Medicine, QEII Medical Centre, Nedlands, WA 6009, Australia. E-mail: [email protected]


References 1. Ramakrishnan R, Chaudhry IH, Ramdial P, et al. Primary cutaneous adenoid cystic carcinoma: a clinicopathologic and immunohistochemical study of 27 cases. Am J Surg Pathol 2013; 37: 1603–11. 2. Wick MR, Swanson PE. Primary adenoid cystic carcinoma of the skin. A clinical, histological, and immunocytochemical comparison with adenoid cystic carcinoma of salivary glands and adenoid basal cell carcinoma. Am J Dermatopathol 1986; 8: 2–13. 3. Calonje E, Brenn T, Lazar A, McKee PH. McKee’s Pathology of the Skin. 4th ed. China: Elsevier Saunders; 2012; 1088-1101. 4. Miettinen M, Lasota J. KIT (CD117): a review on expression in normal and neoplastic tissues, and mutations and their clinicopathologic correlation. Appl Immunohistochem Mol Morphol 2005; 13: 205–20. 5. Gibson PC, Cooper K. CD117 (KIT): a diverse protein with selective applications in surgical pathology. Adv Anat Pathol 2002; 9: 65 – 9. 6. Terada T. Expression of NCAM (CD56), chromogranin A, synaptophysin, c-KIT (CD117) and PDGFRA in normal non-neoplastic skin and basal cell carcinoma: an immunohistochemical study of 66 consecutive cases. Med Oncol 2013; 30: 444. 7. Mino M, Pilch BZ, Faquin WC. Expression of KIT (CD117) in neoplasms of the head and neck: an ancillary marker for adenoid cystic carcinoma. Mod Pathol 2003; 16: 1224–31. 8. Florell SR, Zone JJ, Gerwels JW. Basal cell carcinomas are populated by melanocytes and Langerhans [correction of Langerhan’s.] cells. Am J Dermatopathol 2001; 23: 24–8. 9. Seethala RR, Pasha TL, Raghunath PN, et al. The selective expression of CD43 in adenoid cystic carcinoma. Appl Immunohistochem Mol Morphol 2008; 16: 165–72. 10. Woo VL, Bhuiya T, Kelsch R. Assessment of CD43 expression in adenoid cystic carcinomas, polymorphous low-grade adenocarcinomas, and monomorphic adenomas. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2006; 102: 495–500.

Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.

CD117 and CD43 are useful adjuncts in the distinction of adenoid cystic carcinoma from adenoid basal cell carcinoma.

Distinction of cutaneous adenoid cystic carcinoma (ACC) from adenoid basal cell carcinoma (BCC) is an occasional diagnostic dilemma in dermatopatholog...
332KB Sizes 0 Downloads 12 Views