Pleomorphic Adenoma: Cytologic Variations and Potential Diagnostic Pitfalls

Pleomorphic Adenoma: Cytologic Variations and Potential Diagnostic Pitfalls Uma Handa, M.D., Neerja Dhingra, M.D., D.N.B.,* Rajan Chopra, M.D., and Harsh Mohan, M.D., M.N.A.M.S., F.I.C.P. The diverse morphological features encountered in pleomorphic adenoma (PA) may cause diagnostic errors in fine needle aspiration cytology (FNAC). The present study was performed to evaluate the variations in the cytological features of pleomorphic adenoma and to assess the efficacy of FNAC in its diagnosis. Fifty cases diagnosed as PA on FNAC were retrieved from the records of the Pathology Department. Cytologic smears and sections were reviewed and the cytologic diagnoses were compared with the definitive histologic diagnoses. In cases correctly diagnosed on aspiration, morphological variables like patterns of the epithelial component, type and extent of the mesenchymal matrix, metaplastic cells, hyaline globules, cystic change, giant cells, crystalline deposits, nuclear inclusions/grooves, and nuclear atypia were evaluated. The extreme diversity in morphologic features seen in histologic sections was reflected in the smears of PA. Metaplastic changes were observed more frequently in sections, while nuclear changes like inclusions/ grooves were more commonly seen in smears. Other morphological features like cylindromatous pattern, giant cells and crystalline deposits were observed with equal frequency in smears and sections. Cytohistologic agreement was present in 45 of the 50 cases (90%). In 5 cases diagnosed as pleomorphic adenoma on FNAC, the histology revealed 1 case each of schwannoma, perineurioma, ectomesenchymal chondromyxoid tumor of tongue, adenoid cystic carcinoma and mucoepidermoid carcinoma. FNAC is a fairly accurate pre-operative procedure for the diagnosis of PA. The cytopathologist needs to be aware of the cytologic variations in pleomorphic adenoma so as to avoid diagnostic errors. Diagn. Cytopathol. 2009;37:11 15. ' 2008 Wiley-Liss, Inc. Key Words: Pleomorphic adenoma; cytology; diagnostic pitfalls Pleomorphic adenoma (PA) is the most common salivary gland tumor and comprises about 79% of the major and Department of Pathology, Government Medical College and Hospital, Chandigarh, India *Correspondence to: Neerja Dhingra, M.D., D.N.B., Department of Pathology, Government Medical College and Hospital, Sector 32-A, Chandigarh 160030, India. E-mail: drneerjadhingra@yahoo.co.in Received 8 April 2008; Accepted 31 July 2008 DOI 10.1002/dc.20951 Published online 29 October 2008 in Wiley InterScience (www. interscience.wiley.com). 72% of the minor salivary gland tumors. 1 Morphologic diversity is the hallmark of this tumor; a variety of patterns showing varying combinations of epithelial and mesenchymal components are seen. FNAC can diagnose a great majority of pleomorphic adenomas with high level of accuracy if established diagnostic criteria are strictly observed. 2 4 In most cases, a combination of bland epithelial cells and characteristic fragments of chondromyxoid stroma with spindled cells is diagnostic. However, in some cases the variations of the morphologic features may cause diagnostic errors. 5,6 The chances of such happening are further enhanced due to limited and selective sampling in FNA. Therefore, it is imperative for the cytopathologist to be cognizant of the variations in the pattern of PA on FNAC smears. This study lists the morphological variations encountered in cytologic smears of pleomorphic adenoma that may sometimes pose problem for the cytopathologist and may lead to an erroneous diagnosis. Materials and Methods Fifty cases diagnosed as pleomorphic adenoma on FNAC in which histopathology was available were retrieved from the archives of the Pathology department. The aspiration procedure was performed by an experienced cytopathologist using a 22/23 gauge needle attached to a 20 ml syringe. The air-dried smears were stained with May- Grünwald Giemsa (MGG) and the alcohol-fixed smears with hematoxylin and eosin (H&E) stains. Surgically excised specimens were processed routinely and stained with H&E. The cellularity was assessed on a scale of 1þ to 3þ and the ratio of epithelial to mesenchymal components was noted. The cellularity categories were based on the cells covering the area of the standard cytology smear: 1þ (<25% of the smear area), 2þ (>25% and <50%), and 3þ (>50%). Epithelial to mesenchymal ratio was categorized as: epithelial predominant when epithelial component formed >60% of the tumor cellularity, mesenchymal pre- ' 2008 WILEY-LISS, INC. Diagnostic Cytopathology, Vol 37, No 1 11

HANDA ET AL. dominant when mesenchymal component formed >60% of the tumor cellularity and equal distribution when either of the component ranged >40% to <60%. Such evaluation was done on all the smears (three smears of compatible size were available in each case) by the same diagnostic team and the average of such evaluation was taken for a particular case. Similar evaluation was also done in the histologic sections. The morphological features like the patterns of the epithelial component, type and extent of the mesenchymal matrix, metaplastic cells, hyaline globules, cystic change, giant cells, crystalline deposits, nuclear inclusions/ grooves, and nuclear atypia were studied. The frequency of the above parameters in cytological smears and histopathological slides was compared. Table I. Cases Erroneously Diagnosed as Pleomorphic Adenoma on Cytology Cytologic diagnosis Histologic diagnosis No. of cases PA Schwannoma 1 PA Perineurioma 1 PA Ectomesenchymal chondromyxoid tumor 1 of tongue PA Adenoid cystic carcinoma 1 PA Mucoepidermoid carcinoma 1 Table II. Cellularity of PA in FNA Smears and Histologic Sections Cellularity FNA smears (n ¼ 45) Histologic sections (n ¼ 45) 1þ 2 1 2þ 20 18 3þ 23 26 Table III. The Relative Amounts of Epithelial (E) and Mesenchymal (M) Components E > M E ¼ M E < M FNA smears (n ¼ 45) 20 15 10 Histologic sections (n ¼ 45) 22 11 12 Results The cases were distributed over a wide age range (12 70 years), the mean age being 41 years. There were 26 female patients and 24 male patients. The lesions were located in the parotid gland in 40 cases (80%), submandibular gland in 6 cases (12%), and oral cavity in 4 cases (8%). The agreement of FNAC diagnoses with the final histologic diagnoses was present in 45 cases. In five cases diagnosed as pleomorphic adenoma on cytology, the histology revealed a different diagnosis. In two of these cases, malignant tumors were misinterpreted as PA on cytology, whereas in three cases the final histopathological diagnosis was that of other benign tumors as shown in Table I. Thus the positive predictive value (PPV) of FNAC in this study was 90%. Review of the 45 cases diagnosed correctly on FNAC as PA, showed varied frequency of the following morphologic features. The cellularity of pleomorphic adenoma was graded as 1þ (scanty), 2þ (moderate), and 3þ (marked). The results as observed in FNA smears and histologic sections are shown in Table II. Fig. 1. The morphologic variations in pleomorphic adenoma as noted in FNA smears and paraffin sections. 12 Diagnostic Cytopathology, Vol 37, No 1

PLEOMORPHIC ADENOMA Fig. 2. Pleomorphic adenoma showing epithelial component admixed with chondromyxoid stroma. An intra-nuclear inclusion (small arrow) and nuclear grooving (larger arrow) also seen (MGG, 3400). Fig. 4. Extensive cylindromatous pattern along with numerous hyaline globules (MGG, 3400). Fig. 3. Giant cell and spindled stromal cells in a pale blue myxoid background (H&E, 3400). Fig. 5. Squamous metaplasia of the epithelial component (MGG, 3400). Assessment of the predominance of epithelial component versus mesenchymal matrix revealed that both these elements were present in all the cases of pleomorphic adenoma. However, there was a wide variation in their proportion. In cytological smears, the epithelial component was more than, equal to and less than the mesenchymal component in 20, 15, and 10 cases, respectively. In histological sections the corresponding values were 22, 11, and 12 cases (Table III). Frequency of many morphological variations in PA differed in the FNA smears and biopsy sections (Figs. 1 5). Metaplastic changes like squamous, oncocytic, sebaceous, and mucinous change in the epithelial component were observed more frequently in sections, whereas the nuclear changes like inclusions/ grooves were more commonly seen in FNA smears. Other morphological features like presence of cylindromatous pattern, giant cells, and crystalline deposits were observed with equal frequency in smears and sections. Focal nuclear atypia was seen in smears in one case. However, the histologic sections of this case did not reveal any evidence of anaplasia or mitotic activity. Cytological material revealed myxoid stroma in most of the cases (44 cases). The variations in the appearance of the stromal component like chondroid, hyaline, fibrous, and osseous could not be discerned in smears but were observed in histological sections. Hyaline globules of Diagnostic Cytopathology, Vol 37, No 1 13

HANDA ET AL. stromal material were apparent in FNA smears in three cases. Discussion Fine-needle aspiration cytology is a widely accepted safe and effective technique for the preoperative diagnosis of salivary gland lesions. The reliability of FNAC in diagnosing pleomorphic adenoma has been reported as 89.5% 96.2%. 3 The cytological diagnosis of PA is relatively simple. A combination of bland epithelial cells in regular aggregates and metachromatic fragments of fibrillary chondromyxoid stroma with spindle cells is very characteristic. However, PA is well known for a variety of architectural and cytomorphological patterns. Not only does the proportion between epithelium and chondromyxoid stroma vary considerably, but there are also variations in the appearance of the epithelial cells and the stromal component. 3,5 10 The morphological variations do not often pose a difficulty in the histological diagnosis as the whole of the tumor is available for interpretation. In FNAC only a portion of the tumor is sampled, and therefore it is not surprising that PA can be mistaken for several other types of tumors. 8 In the present study, we studied various morphological variations in PA. Metaplastic changes like squamous metaplasia and oncocytic changes were seen in 8 and 4% of the FNA smears, respectively. Giant cells were seen in 4% of the cases. Nuclear changes like inclusions/grooves were a common finding in smears, present in 28 and 56% of the cases, respectively. Das and Anim 5 in their study of 25 cases of PA observed squamous metaplasia and oncocytic change in 12% cases each, giant cells in 4%, nuclear inclusions and grooves in 36 and 84% of the cases, respectively. In addition we observed crystalline deposits in 2% of our cases. The incidence of crystals in PA varies from 1.5 to 21% in different studies. 11 Cystic degeneration and mucin production are common in PA. 4,12 14 In our series cystic change was present in 42% of the cases. Low grade mucoepidermoid carcinoma, Warthin s tumor, acinic cell carcinoma, and benign nonneoplastic lesions of salivary glands may show cystic changes with aspirates dominated by abundant mucinous material and scant cellularity. PA should be suspected when epithelial and stromal elements are identified within the mucinous material. In the present series a case of low grade mucoepidermoid carcinoma was misinterpreted as PA on FNAC. The smears showed clusters of epithelial cells with bland nuclear chromatin in a background of scanty mucinous material. Cells with obvious squamous differentiation and the characteristic stroma of PA were not seen. Because of scantiness of mucin and predominance of epithelial cells, mucoepidermoid carcinoma was not suspected on FNAC. Retrospectively we feel such case should have been reaspirated for more representive cellularity of mucoepidermoid carcinoma, instead of it being labeled as epithelial predominant PA. If the epithelial cells in a FNA smear show squamous differentiation, the possibility of mucoepidermoid carcinoma needs to be considered. Selective sampling of an area of squamous metaplasia in PA can cause diagnostic problems. 8 The identification of characteristic admixture of epithelial and chondromyxoid stroma allows a correct diagnosis in such cases. Cylindromatous presentation of PA can cause diagnostic dilemma. Such cases may resemble adenoid cystic carcinoma (ACC) cytologically. 3,4,7,15 Furthermore, hyaline globules of stromal material can be seen in both the tumors, therefore it is not a fool proof differential morphological feature. 8,10 Viguer et al. 3 in their study found cylindromatous pattern in 9 of the 212 (4.2%) cases of PA. Smears showing extensive cylindromatous pattern with hyaline globules was noted in two (4%) of our cases. In such cases the chromatin pattern of the tumor cell nuclei must be closely examined. The nuclei of PA cells show bland chromatin with uniform granularity whereas nuclei of ACC cells are hyperchromatic, the chromatin is coarse and some may show prominent nucleoli. 3,4,16 In addition, lack of single cell dispersion and presence of stromal fragments with spindled cells favor PA. The one case in our study, which was misinterpreted as PA, had nuclear atypia on review and lacked stromal fragments with spindled cells. Although one case with prominent cylindromatous pattern correctly diagnosed on FNAC had spindle-celled stroma and also lacked nuclear atypia and single-cell dispersion. Other benign tumors such as monomorphic adenoma and myoepithelioma can be erroneously suspected as PA on cytology. 17 Especially in cases showing highly cellular smears with scanty stromal elements, a possibility of monomorphic adenoma and even ACC may be suggested. 5 Another important feature that may be observed in aspirates from PA is the presence of atypical cells having irregular often bizarre nuclei. These elements are probably degenerative in nature. Malignancy should only be considered when the atypical cells are abundant, show abnormal chromatin pattern and are accompanied by necrosis. 4,18 We observed significant nuclear atypia in FNA smears in 1 of the 50 (2%) cases. This is in contrast to other studies that show a higher incidence of nuclear atypia, ranging from 5.3% to 48.3%. 3,19,20 Aspirates from the other myxoid lesions in the salivary gland locations can be misdiagnosed as PA 21 24 ; as has been our experience in three cases in the present study. Two cases of peripheral nerve sheath tumors-schwannoma, perineurioma and one case of ectomesenchymal chondromyxoid tumor were mistaken for stroma rich PA on FNAC. This calls for caution when examining a myxoid lesion so as not to miss the correct diagnosis. Clues 14 Diagnostic Cytopathology, Vol 37, No 1

PLEOMORPHIC ADENOMA such as epithelial cell aggregates of PA which were lacking in the aforementioned cases should be carefully looked for. To conclude, the PA is unique in its morphological complexity, which is reflected in the cytologic material. Though the variations in the morphological pattern do not influence the tumor behavior and prognosis, they may pose diagnostic difficulties for the cytopathologist. Thorough examination of the smears and awareness of the variations in morphology are important measures to avoid diagnostic errors. References 1. Kapila K, Mathur S, Verma K. Schwannomas: A pitfall in the diagnosis of pleomorphic adenomas on fine needle aspiration cytology. Diagn Cytopathol 2002;27:53 59. 2. Orell SR, Sterrett GF, Whitaker D, Klijanienko J. Manual and atlas of fine needle aspiration cytology. London, Churchill Livingstone; 2005. p 53 82. 3. Viguer JM, Vicandi B, Jiménez-Heffernan JA, López-Ferrer P, Limeres MA. Fine needle aspiration cytology of pleomorphic adenoma: An analysis of 212 cases. Acta Cytol 1997;41:786 794. 4. Viguer JM, Jiménez-Heffernan JA, Vicandi B, López-Ferrer P, Navarro M. Cytologic diagnostic accuracy in pleomorphic adenoma of the salivary glands during 2 periods. A comparative analysis. Acta Cytol 2007;51:16 20. 5. Das DK, Anim JT. Pleomorphic adenoma of salivary gland: To what extent does fine needle aspiration cytology reflect histopathological features? Cytopathology 2005;16:65 70. 6. Verma K, Kapila K. Role of fine needle aspiration cytology in diagnosis of pleomorphic adenomas. Cytopathology 2002;13:121 127. 7. Boccato P, Altavilla G, Blandamura S. Fine needle aspiration biopsy of salivary gland lesions: A reappraisal of pitfalls and problems. Acta Cytol 1998;42:888 898. 8. Orell SR, Nettle WJS. Fine needle aspiration biopsy of salivary gland tumors. Problems and pitfalls. Pathology 1988;20:332 337. 9. Orell SR. Diagnostic difficulties in the interpretation of fine needle aspirates of salivary gland lesions: The problem revisited. Cytopathology 1995;6:285 300. 10. Klijanienko J, Vielh P. Fine-needle sampling of salivary gland lesions. I. Cytology and histology correlation of 412 cases of pleomorphic adenoma. Diagn Cytopathol 1996;14:195 200. 11. Salem F, Siddiqui N. Tyrosine crystals in pleomorphic adenoma. Diagn Cytopathol 2006;34:419 420. 12. Siddiqui NH, Wu SJ. Fine needle aspiration biopsy of cystic pleomorphic adenoma with adnexa like differentiation mimicking mucoepidermoid carcinoma: A case report. Diagn Cytopathol 2005; 32:229 232. 13. Nishimura T, Furukawa M, Kawahara E. Pleomorphic adenoma of parotid gland with cystic degeneration. J Laryngol Otol 1994;108: 446 448. 14. Stanley MW, Lowhagen T. Mucin production by pleomorphic adenoma of the parotid gland: A cytologic spectrum. Diagn Cytopathol 1990;6:49 52. 15. Lee S-S, Cho K-J, Jang J-J, Ham EK. Differential diagnosis of adenoid cystic carcinoma from pleomorphic adenoma of the salivary gland on fine needle aspiration cytology. Acta Cytol 1996;40:1246 1252. 16. Klijanienko J, Vivelh P. Fine needle sampling of salivary gland lesions. III. Cytology and histology correlation of 75 cases of adenoid cystic carcinoma. Review and experience at Institut Curie with emphasis on cytologic pitfalls. Diagn Cytopathol 1997;17:36 41. 17. Kawahara A, Harada H, Akiba J, Yokoyama T, Kage M. Fine needle aspiration cytology of basal cell adenoma of the parotid gland: Characteristic cytological features and diagnostic pitfalls. Diagn Cytopathol 2007;35:85 90. 18. Klijanienko J, El-Naggar AK, Vivelh P. Fine needle sampling findings in 26 carcinomas ex pleomorphic adenomas. Diagnostic pitfalls and clinical considerations. Diagn Cytopathol 1999;21: 163 166. 19. Chan MKM, McGuire FJ, King W, Li AKC, Lee JCK. Cytodiagnosis of 112 salivary gland lesions: Correlation with histologic and frozen section diagnosis. Acta Cytol 1992;36:353 363. 20. Eneroth CM, Zajizek J. Aspiration biopsy of salivary gland tumors. III. Morphologic studies on smears and histologic sections from 386 mixed tumors. Acta Cytol 1966;10:440 454. 21. Smith BC, Ellis GL, Meis-Kindblom JM, Williams SB. Ectomesenchymal chondromyxoid tumors of the anterior tongue. Nineteen cases of a new clinicopathologic entity. Am J Surg Pathol 1995;19: 519 530. 22. Kannan R, Damm DD, White DK, Marsh W, Allen CM. Ectomesenchymal chondromyxoid tumors of the anterior tongue. A report of 3 cases. Oral Surg Oral Med Oral Pathol 1996;82:417 422. 23. Saad RS, Takei H, Lipscomb J, Ruiz B. Nodular fasciitis of parotid region: A pitfall in the diagnosis of pleomorphic adenomas on fine needle aspiration cytology. Diagn Cytopathol 2005;33:191 194. 24. Chhieng DC, Cohen JM, Cangiarella JF. Fine needle aspiration of spindle cell and mesenchymal lesions of the salivary gland. Diagn Cytopathol 2000;23:253 259. Diagnostic Cytopathology, Vol 37, No 1 15