—  SYMPOSIUM #51  —

Update of Common Salivary Tumors
Moderators: Dr. John Eveson and Dr. Silloo Kapadia

Section 5 - Adenoid Cystic Carcinoma

John W Eveson
U. K.


Adenoid cystic carcinoma is a basaloid tumour consisting of epithelial and myoepithelial cells in variable morphologic configurations, including tubular, cribriform and solid patterns (WHO 2005).

Adenoid cystic carcinoma (AdCC) accounted for 4.4% of all salivary gland tumors and 11.8% of malignant salivary neoplasms in the largest reported series. It affects a wide age range (13-103 years) with a peak incidence in 50–70 year olds. There is a female preponderance of about 3:2. The most common sites are the parotid (21%), palate (17%) and submandibular gland (15%). Minor glands of the mouth and the upper aerodigestive tract account for about half of all cases. Rare central (intra-osseous) cases have been reported, the large majority involving the body and angle of mandible. [1] Tumors usually present as a slow growing but widely infiltrative mass of long duration and may be mobile or fixed. They may be tender or painful, and cranial nerve lesions, particularly facial nerve palsy, can be the presenting feature. Tumors of minor glands often show ulceration of the overlying mucosa.

AdCC is composed of luminal ductal cells and abluminal, modified myoepithelial cells. The latter predominate and have indistinct cell borders and frequently sparse, amphophilic or clear cytoplasm. The nuclei are uniform in size, basophilic and may be round or angular (peg shaped). The ductal cells surround small and sometimes indistinct lumina; they are cuboidal and have more abundant, eosinophilic cytoplasm and round, uniform nuclei that may contain small nucleoli. There are three main morphological patterns: cribriform (cylindromatous), tubular and solid in that order of frequency. A mixture of these patterns may be seen. The cribriform variant consists of islands of modified myoepithelial cells containing rounded, pseudocytic areas forming a characteristic 'Swiss cheese' appearance. The pseudocysts are basophilic and mucoid, or consist of hyaline, eosinophilic material. They are composed of glycosaminoglycans and reduplicated basement membrane material. Foci of ductal cells are present within the myoepithelial areas but may require careful examination to detect. The tubular variant is double layered and has conspicuous ductal differentiation. There is an inner layer of eosinophilic, duct-lining cells and the abluminal myoepithelial cells may show clear cytoplasm and irregular, angular nuclei or be morphologically similar to the luminal layer. Occasionally, foci of squamous, sebaceous or oncocytic metaplasia are present. The least common solid variant consists of islands or sheets of basaloid cells with larger and less angular nuclei. Duct-lining cells may be few and inconspicuous and comedonecrosis is common. Mitoses are sparse in the cribriform and tubular patterns but may be frequent, often together with extensive apoptosis, in the solid type. Recently, rare but clinically aggressive dedifferentiated variants of AdCC, characterized histologically by the combination of a conventional AdCC and a high-grade carcinoma, have been reported. [2] In addition, AdCC may form part of a hybrid tumor, most commonly in association with salivary duct carcinoma or epithelial- myoepithelial carcinoma. [3] The stroma of AdCC is usually fibrous and cases where hyalinisation is so abundant that tumor cells are attenuated into strands have been reported as a sclerosing variant. [4] Perineural or intraneural invasion is common and frequently conspicuous and the tumor can extend along nerves and their branches over a wide area. The tumor may invade bone extensively before there is any radiological evidence bone destruction. Lymph node involvement is uncommon (~5% of cases) and is usually due to contiguous spread rather than lymphatic permeation or embolization.

The pseudocysts are positive for PAS and Alcian blue and contain basement membrane components such as type IV collagen and laminin isoforms. [5] The epithelial cells are positive for low molecular weight keratins and EMA, and the myoepithelial cells are variably positive for markers such as calponin, SMA, p63, and S100. Expression of S100 and GFAP has been correlated with the presence perineural invasion. [6] High expression of both p53 [7] and bcl-2 has been reported but there was no relation to this observation and the histological tumor types, clinical stage or survival. [8] In the other study, the Ki-67 labelling index was found to be low, in keeping with the natural history of the tumor.

Alterations in chromosomes 6q, 9p and 17p12-13 are the most frequent cytogenetic alterations reported. There are frequent losses at 12q, 6q23-qter, 13q21-q22 and 19q regions. A high frequency of LOH at 6q23-25 has been reported and, although not specific for AdCC, this has been correlated with both tumor grade and behaviour. Over half of cases show genomic deletions of chromosome 6 [9] and candidate suppressor genes have also been mapped to chromosome 12. [10] Hypermethylation of the p16 promotor gene was found in 4/22 of AdCC and was associated with higher histological grades of malignancy. Alterations in p53 and rb genes have been reported and correlated with behaviour. [11] Recently, microarrays and comparative genomic hybridization have been used to identify candidate genes for AdCC. [12]

It is important to distinguish AdCC from polymorphous low-grade adenocarcinoma (PLGA) in tumors from minor salivary glands. PLGA consists of a uniform cell population with cytologically bland, round or oval and vesicular nuclei and pale eosinophilic cytoplasm, whereas cells in AdCC often have clear cytoplasm and angular, hyperchromatic nuclei and may show mitotic activity. The MIB-1 proliferative index is reported to be nearly 10x higher in AdCC than PLGA with no overlap zone. [13] Smooth muscle markers of myoepithelial differentiation are positive in AdCC but negative in PLGA. In addition, PLGA shows a much wider spectrum of histomorphological differentiation. Even when cribriform areas are present in PLGA they are typically focal and AdCC does not show papillary differentiation or the single cord arrangement of cells seen in PLGA. Although both AdCC and PLGA show neural invasion, in PLGA this is often associated with a striking whorling arrangement of single file cords of cells or small ducts, and is seen predominantly within or very close to the main tumor mass. It has been postulated that staining with c-kit or galectin 3 aids the distinction between PLGA and ACC, but recent studies suggest the value of these immunoagents is limited. [14, 15] In addition, c-kit expression in AdCC does not appear to correlate with behavior. [16] Occasional foci in pleomorphic adenoma can resemble AdCC but the presence of typical myxochondroid matrix and plasmacytoid or spindle-shaped cells helps to avoid confusion. [17] Basaloid squamous cell carcinoma can resemble solid variants of AdCC but typically involves the hypopharynx and glottal region, which are uncommon, sites for AdCC. Both can show islands with cribriform configurations, hyaline material surrounding tumor nests and solid areas with comedonecrosis, but basaloid squamous carcinoma also has evidence of squamous differentiation and usually involves the overlying mucosa. In addition, the different patterns of p63 staining may aid distinction. [18] Both epithelial-myoepithelial carcinoma and the tubular variant of AdCC can show double-layered duct-like structures with an abluminal layer of clear cells.

The average 5 and 10-year survival rates are about 60% and 40% respectively but the majority of patients usually die of, or with, the tumor. Local recurrence is very common, especially in the first 5 years after surgery. The main prognostic factors are site, clinical stage and histological pattern. Bone involvement and failure of primary surgery are associated with poor prognosis. Correlations between tumor morphology, grading and outcome have yielded conflicting results, particularly after 10 year follow-up due to the overall poor long-term prognosis. [19] The tubular and cribriform variants have been reported to have a better outcome than tumors with a solid component, especially if this exceeds 30% of the tumor volume. Some studies suggest that tumor size and clinical stage are more reliable indicators of prognosis. [20] Tumors in the submandibular gland have a poorer prognosis than those in the parotid. The relationship between perineural invasion (PNI) and survival is also contentious but invasion of larger nerves appears to correlate with more aggressive behavior. [21] In a review of the literature, Barrett and Speight (2002) [22] found 23% (range 2-86%) of patients had clinical evidence of neurological deficit. Histological evidence of PNI was reported in 51% (range 8-86%) of cases. Clinical evidence of nerve involvement such as facial nerve palsy was an indicator of poor prognosis but histological evidence of PNI appeared not to be an independent factor but was associated with other adverse features such as clinically large and more aggressive tumors. In addition, Spiro and co-workers in large follow-up studies of patients with AdCC found that both histological grades and PNI are unreliable predictors of behavior and they suggest that clinical stage is a better guide to prognosis. Lymph node involvement is relatively uncommon but distant metastases to lung, bone, brain and liver are seen in 40 – 60% of cases. Wide local excision, together with adjuvant radiotherapy, offers the best hope of local control.

References
  1. Al-Sukhun J, Lindqvist C, Hietanen J, et al. Central adenoid cystic carcinoma of the mandible: case report and literature review of 16 cases. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2006;101:304-8.

  2. Nagao T, Gaffey TA, Serizawa H, et al. Dedifferentiated adenoid cystic carcinoma: a clinicopathologic study of 6 cases. Mod Pathol 2003;16:1265-72.

  3. Nagao T, Sugano I, Ishida Y, et al. Hybrid carcinomas of the salivary glands: report of nine cases with a clinicopathologic, immunohistochemical, and p53 gene alteration analysis. Mod Pathol. 2002;15:724-33.

  4. Albores-Saavedra J, Wu J, Uribe-Uribe N. The sclerosing variant of adenoid cystic carcinoma: a previously unrecognized neoplasm of major salivary glands. Ann Diagn Pathol 2006;10:1-7.

  5. Kumamoto M, Kuratomi Y, Yasumatsu R, et al. Expression of laminin 5 basement membrane components in invading and recurring adenoid cystic carcinoma of the head and neck. Auris Nasus Larynx 2006; 33: 167-72.

  6. Luo X-L, Sun M-Y, Lu C-T, et al. The role of Schwann cell differentiation in perineural invasion of adenoid cystic and mucoepidermoid carcinoma of the salivary gland. Int J Oral Maxillofac Surg 2006; [Epub ahead of print]

  7. Yamamoto Y, Wistuba II, Kishimoto Y, et al. DNA analysis at p53 locus in adenoid cystic carcinoma: comparison of molecular study and p53 immunostaining. Pathol Int 1998;48:273-80.

  8. Carlinfante G, Lazzaretti M, Ferrari S, et al. p53, bcl-2 and Ki-67 expression in adenoid cystic carcinoma of the palate. A clinico-pathologic study of 21 cases with long-term follow-up. Pathol Res Pract 2005; 200: 791-9.

  9. Rutherford S, Yu Y, Rumpel CA, et al. Chromosome 6 deletion and candidate tumor suppressor genes in adenoid cystic carcinoma. Cancer Lett 2006;236:309-17.

  10. Rutherford S, Hampton GM, Frierson HF, et al. Mapping of candidate tumor suppressor genes on chromosome 12 in adenoid cystic carcinoma. Lab Invest 2005;85:1076-85.

  11. Yamamoto Y, Virmani AK, Wistuba II, et al. Loss of heterozygosity and microsatellite alterations in p53 and RB genes in adenoid cystic carcinomas of the salivary glands. Hum Pathol 1996; 27: 1204 -10

  12. Kamsamatsu A, Endo Y, Uzawa K, et al. Identification of candidate genes associated with salivary adenoid cystic carcinomas using combined comparative genomic hybridization and oligonucleotide microarray analyses. Int J Biochem Cell Biol 2005;37: 1869-80.

  13. Skalova A, Simpson RH, Lehtonen H, et al. Assessment of proliferative activity using the MIB1 antibody help to distinguish polymorphous low grade adenocarcinoma from adenoid cystic carcinoma of salivary glands. Pathol Res Pract 1997;193:695-703.

  14. Jeng YM, Lin CY, Hsu HC. Expression of the c-kit protein is associated with certain subtypes of salivary gland carcinoma. Cancer Lett 2000;154:107-11.

  15. Edwards PC, Bhuiya T, Kelsch RD. C-kit expression in the salivary gland neoplasms adenoid cystic carcinoma, polymorphous low-grade adenocarcinoma, and monomorphic adenoma. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2003;95:586-93.

  16. Aslan DL, Oprea GM, Jagush SM, et al. c-kit expression in adenoid cystic carcinoma does not have an impact on local or distant tumor recurrence. Head Neck 2005: 27: 1028-34.

  17. Ogawa I, Miyauchi M, Matsuura H, et al. Pleomorphic adenoma with extensive adenoid cystic carcinoma-like cribriform areas of parotid gland. Pathol Int 2003;53:30-4.

  18. Emanuel P, Wang B, Wu M, et al. p63 Immunohistochemistry in the distinction of adenoid cystic carcinoma from basaloid squamous cell carcinoma. Mod Pathol 2005;18:645-50.

  19. Szanto PA, Luna MA, Tortoledo ME, et al. Histologic grading of adenoid cystic carcinoma of the salivary glands. Cancer 1984 Sep 15;54:1062-9.

  20. Spiro RH, Huvos AG. Stage means more than grade in adenoid cystic carcinoma. Am J Surg 1992;164:623-8.

  21. Garden AS , Weber RS, Morrison WH, et al. The influence of positive margins and nerve invasion in adenoid cystic carcinoma of the head and neck treated with surgery and radiation. Int J Radiat Oncol Biol Phys 1995;32:619-26.

  22. Barrett A, Speight P. Chapter 19. The controversial adenoid cystic carcinoma. The implications of histological grade and perineural invasion. In: Controversies in the management of salivary gland disease. Ed M McGurk, A Renehan. Oxford University Press 2002. pp 211-7.