Fine-needle aspiration cytology in the management of salivary gland tumors: an Australian experience.
We conducted a retrospective study of 104 cases of salivary gland tumors that were initially assessed by fine-needle aspiration biopsy (FNAB). Based on subsequent histopathologic analysis of excised specimens, we found that preoperative FNAB was highly sensitive and specific for both benign and malignant tumors--including the most common, pleomorphic adenomas and squamous cell carcinomas, respectively. Despite impossible drawbacks, we conclude that preoperative FNAB is a useful tool in the management of salivary gland tumors.
Salivary gland neoplasms account for fewer than 2% of all human neoplasms. (1) When they do occur, most arise in the parotid gland; a small percentage occur in the submandibular, sublingual, and minor salivary glands. There is a wide variety of histologic types of salivary gland tumors, both benign and malignant.
The management of salivary gland tumors is based on clinical assessment, radiologic investigation, and cytologic study of fine-needle aspiration biopsy (FNAB) specimens. Concerns about the safety of FNAB include bleeding, tumor seeding, and facial nerve injury. Nevertheless, various studies over the past decade have found that FNAB is an accurate and useful investigation. (2-10)
In this article, we describe our study of the accuracy of FNAB cytology in patients with salivary gland tumors at our institution. We also discuss the findings of other researchers who have studied this topic.
Patients and methods
We retrospectively reviewed the medical records of 104 patients who had undergone FNAB biopsy and subsequent surgery for a salivary gland tumor at our institution between Oct. 1, 1996, and Dec. 31, 2002. FNAB had been performed at several different centers, surgery had been performed by 1 of 2 surgeons in the Head and Neck unit of our institution, and histologic analysis of the surgical specimens had been performed by several different pathologists at our institution. Preoperative cytology results were classified as either benign, malignant, or nondiagnostic. In most cases of benign or malignant disease, the cytologist suggested a diagnosis, which was also recorded.
We compared the histopathology of the surgical specimens with the preoperative cytology of the FNAB specimens and calculated the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and overall accuracy of FNAB for diagnosing benign and malignant disease. We also calculated FNAB's ability to identify the most common histopathologies in both categories. Our study protocol was approved by the Human Research Ethics Committee of Northern Sydney Health.
Of the 104 FNABs, 64 (61.5%) were originally classified as benign, 30 (28.8%) as malignant, and 10 (9.6%) as nondiagnostic. Histopathology of the excised specimens revealed that 70 patients (67.3%) had benign disease and 34 (32.7%) had malignant disease (table 1).
Benign disease. FNAB correctly identified 62 of 70 benign lesions (sensitivity: 88.6%) (table 2). Of the eight false-negative results (table 3), seven were actually nondiagnostic rather than malignant. Therefore, when the nondiagnostic results were excluded from the analysis, the sensitivity of FNAB for benign disease was 98.4%. The specificity of FNAB for benign disease was 94.1% overall and 93.5% when nondiagnostie results were excluded (there were two false positives [table 3]). The PPV for benign disease was 0.969, and the NPV was 0.80. Therefore, the overall accuracy of FNAB for benign disease was 90.4% (table 2).
The most common benign disease in this study was pleomorphic adenoma. In diagnosing this tumor, FNAB had a sensitivity of 88.2%, a specificity of 98.1%, and an overall accuracy of 93.3% (table 2).
Malignant disease. The overall sensitivity of FNAB for detecting malignant disease in general was 85.3%. Of the five false-negative results (table 3), three were actually nondiagnostic rather than benign. Therefore, when the nondiagnostic results were excluded, the sensitivity of FNAB for malignant disease was 93.5%. The specificity of FNAB for malignant disease was 98.6% overall and 98.4% when the nondiagnostic results were excluded (there was one false positive [table 3]). The PPV was 0.967, the NPV was 0.932, and the overall accuracy was 94.2% (table 2).
Squamous cell carcinoma (SCC) was the most common malignant pathology. The sensitivity of FNAB was 88.9%, its specificity was 100%, and its overall accuracy was 98.0% (table 2).
Overall. For benign and malignant tumors combined, sensitivity, specificity, and accuracy were 86.9, 96.3, and 92.3%, respectively (table 2).
The sensitivity and specificity of FNAB for benign and malignant disease in our study were comparable to or better than previously reported rates. (2-10)
The overall accuracy of cytologic diagnosis based on FNAB has been reported to range from 56 to 98%; our overall accuracy rate was 92.3%. For benign lesions, the reported sensitivity of FNAB ranges from 76 to 97%; our rate was 88.6%. The specificity of FNAB has been reported to range from 61 to 83%; our rate was 94.1%.
For malignant lesions, reported rates of sensitivity range from 55 to 87%; our rate was 85.3%. Specificity has been reported to range from 90 to 100%; our rate was 98.6%.
In another study from Australia, Que Hee and Perry concluded that FNAB is unreliable. (11) They analyzed the diagnostic accuracy of FNAB in 169 patients at a city teaching hospital similar to our own institution. They found that the specificity of FNAB for malignant disease overall and for SCC was 100 and 99%, respectively--rates that are similar to our own (98.6 and 100%). However, they also reported a markedly lower sensitivity for malignant disease (57%) and for SCC (52%); our rates were 85.3 and 88.9%, respectively. With respect to benign disease, they reported a comparable sensitivity (88 vs. 88.6%) but a much lower specificity (61 vs. 94.1%). Also, they found that FNAB had a lower sensitivity for pleomorphic adenoma (78 vs. 88.2%), but a similar specificity (95 vs. 98.1%). Que Hee and Perry based their conclusion that FNAB is unreliable on the low sensitivity for malignant disease.
When FNAB is nondiagnostic--which occurred in 9.6% of our cases--some authors recommend performing an ultrasound-guided core biopsy to obtain a tissue diagnosis. (12) However, it is our usual practice in such cases to perform a repeat FNAB. In our study, the sensitivity of FNAB for benign and malignant disease improved substantially when nondiagnostic findings were excluded, while its specificity remained about the same. It is important to acknowledge that a nondiagnostic result is not a negative result; rather, it represents an invalid test.
Some of the false-negative and false-positive results in our study deserve mention. First, two of the five specimens that were falsely negative for malignancy were lymphomas. These were the only cases of lymphoma in our group of patients. The fact that both were not detected as malignant on FNAB is consistent with other reports that lymphoma is difficult to diagnose by FNAB. (4-6, 13) If a lymphoma is suspected on clinical grounds, flow cytometry of FNAB can be a useful diagnostic aid. (13)
Second, the only false positive for malignancy in our study was reported as "large malignant cells suggesting carcinoma" on cytology. This tumor was later found to be a pleomorphic adenoma on histopathology. Conversely, a finding of pleomorphic adenoma on FNAB does not completely exclude the presence of an adenocarcinoma on histology. In rare cases, adenocarcinoma arises in a pre-existing pleomorphic adenoma, such as occurred in one of our patients. Such a tumor may not be detected because of a sampling error or an incorrect cytologic interpretation. (14) The great variety of histologic appearances of salivary gland tumors does make cytologic diagnosis a challenging and often difficult process. It is important that the pathologist be familiar with the atypical features of some tumors. (12)
There are other concerns about the use of FNAB preoperatively. Some authors have suggested that FNAB can cause histologic changes that can obscure a histologic diagnosis. (15, 16) However, the findings of at least one large study appear to disprove this idea. (17) Mukunyadzi et al found that while FNAB may produce changes such as infarction and hemorrhage, these changes were usually not extensive enough to compromise a histologic diagnosis. (17)
We did not use adjunctive techniques to improve the diagnostic accuracy of FNAB in our study, but some have been shown to be helpful. Megerian and Maniglia reported that preoperative FNAB combined with intraoperative frozen-section analysis resulted in a higher rate of diagnostic accuracy than either procedure alone. (18) Likewise, Takashima et al suggested that a combination of clinical assessment, magnetic resonance imaging, and FNAB with cytology and flow cytometry improves accuracy. (10)
We believe that FNAB is useful in preoperative planning for several reasons. As already discussed, lymphoma can be diagnosed by flow cytometry of an FNAB specimen. Appropriate chemotherapy can then be administered and surgery avoided. Our study has also shown that FNAB was highly accurate at diagnosing SCC, the most common malignant salivary gland neoplasm in Australia. Making this diagnosis preoperatively allows for tumor excision and neck dissection to be performed in one operation when this is deemed beneficial. Likewise, a diagnosis of melanoma may allow neck dissection to be performed during the same operation as the tumor excision. Alternately, if Warthin's tumor is diagnosed in a patient deemed to be a high surgical risk, a nonsurgical approach with regular follow-up can be taken.
In conclusion, we found that FNAB is an accurate means of diagnosing both benign and malignant salivary gland tumors at our institution. It is especially useful in excluding SCC. Although it is important to be mindful of the possible pitfalls of FNAB, we believe it plays a useful role in the assessment of salivary gland tumors.
Table 1. Pathology of the 104 excised benign and malignant specimens Pathology n (%) Benign (n = 70) Pleomorphic adenomac 51 (72.9) Warthin's tumor 8 (11.4) Adenoma (other) 4 (5.7) Chronic sialadenitis 2 (2.9) Epidermoid cyst 2 (2.9) Oncocytoma 2 (2.9) Benign myofibroblastic tumor 1 (1.4) Malignant (n = 34) Squamous cell carcinoma 18 (52.9) Acinic cell carcinoma 4 (11.8) Melanoma 3 (8.8) Adenocarcinoma 2 (5.9) Lymphoma 2 (5.9) Adenoid cystic carcinoma 1 (2.9) Malignant histiocytoma 1 (2.9) Myoepithelial carcinoma 1 (2.9) Poorly differentiated carcinoma 1 (2.9) Sarcoma 1 (2.9) Table 2. Accuracy of FNAB Pathology Sensitivity Specificity PPV * NPV * Accuracy Benign (n = 70) 88.6% 94.1% 0.969 0.800 90.4% Pleomorphi 88.2% 98.1% 93.3% adnoma (n = 51) ([dagger]) Malignant (n = 34) 85.3% 98.6% 0.967 0.932 94.2% Squamous cell 88.9% 100% 98.0% carcinoma (n = 18) ([double dagger]) Overall 86.9% 96.3% 0.968 0.866 92.3% * PPV = positive predictive value; NPV = negative predictive value. ([dagger]) Pleomorphic adenoma was the most common benign disease in this study. ([double dagger]) Squamous cell cacinoma was the most common malignant disease in this study. Table 3. False negatives and false positives for benign and malignant disease Pathology n False negatives, benign (n = 8) Pleomorphic adenoma 3 Oncocytoma 2 Cystadenoma 1 Myofibroblastic tumor 1 Warthin's tumor 1 False positives, benign (n = 2) Adenocarcinoma (arising 1 in pleomorphic adenoma) Lymphoma 1 False negatives, malignant (n = 5) Lymphoma 2 Adenocarcinoma (arising 1 in pleomorphic adenoma) Adenoid cystic carcinoma 1 Melanoma 1 False positive, maglinant (n = 1) Pleomorphic adenoma 1
(1.) Cotran RS, Kumar V, Collins T. Robbins Pathologic Basis of Disease. 6th ed. Philadelphia: W.B. Saunders, 1999:769.
(2.) Atula T, Greenman R, Laippala P, Klemi PJ. Fine-needle aspiration biopsy in the diagnosis of parotid gland lesions: Evaluation of 438 biopsies. Diagn Cytopathol 1996;15:185-90.
(3.) Weinberger MS, Rosenberg WW, Meurer WT, Robbins KT. Fine-needle aspiration of parotid gland lesions. Head Neck 1992; 14:483-7.
(4.) Boccato P, Altavilla G, Blandamura S. Fine needle aspiration biopsy of salivary gland lesions. A reappraisal of pitfalls and problems. Acta Cytol 1998;42:888-98.
(5.) Cajulis RS, Gokaslan ST, Yu GH, Frias-Hidvegi D. Fine needle aspiration biopsy of the salivary glands. A five-year experience with emphasis on diagnostic pitfalls. Acta Cytol 1997;41: 1412-20.
(6.) Zurrida S, Alasio L, Tradati N, et al. Fine-needle aspiration of parotid masses. Cancer 1993;72:2306-11.
(7.) Al-Khafaji BM, Nestok BR, Katz RL. Fine-needle aspiration of 154 parotid masses with histologic correlation: Ten-year experience at the University of Texas M.D. Anderson Cancer Center. Cancer 1998;84:153-9.
(8.) Deans GT, Briggs K, Spence RA. An audit of surgery of the parotid gland. Ann R Coll Surg Engl 1995;77:188-92.
(9.) Lussier C, Klijanienko J, Vielh P. Fine-needle aspiration of metastatic nonlymphomatous tumors to the major salivary glands: A clinicopathologic study of 40 cases cytologically diagnosed and histologically correlated. Cancer 2000;90:350-6.
(10.) Stewart CJ, MacKenzie K, McGarry GW, Mowat A. Fine-needle aspiration cytology of salivary gland: A review of 341 cases. Diagn Cytopathol 2000;22:139-46.
(11.) Que Hee CG, Perry CF. Fine-needle aspiration cytology of parotid tumours: Is it useful? ANZ J Surg 2001;71:345-8.
(12.) Buckland JR, Manjaly G, Violaris N, Howlett DC. Ultrasound-guided cutting-needle biopsy of the parotid gland. J Laryngol Otol 1999;113:988-92.
(13.) Tworek JA, Michael CW. Parotid gland mucosa associated lymphomas and their cytologic mimics. Semin Diagn Pathol 2001; 18:147-50.
(14.) Verma K, Kapila K. Role of fine needle aspiration cytology in diagnosis of pleomorphie adenomas. Cytopathology 2002;13: 121-7.
(15.) Li S, Baloch ZW, Tomaszewski JE, LiVolsi VA. Worrisome histologic alterations following fine-needle aspiration of benign parotid lesions. Arch Pathol Lab Med 2000; 124:87-91.
(16.) Di Palma S, Simpson RH, Skalova A, Michal M. Metaplastic (infarcted) Warthin's tumour of the parotid gland: A possible consequence of fine needle aspiration biopsy. Histopathology 1999;35:432-8.
(17.) Mukunyadzi P, Bardales RH, Palmer HE, Stanley MW. Tissue effects of salivary gland fine-needle aspiration. Does the procedure preclude accurate histologic diagnosis? Am J Clin Pathol 2000;114:741-5.
(18.) Megerian CA, Maniglia AJ. Parotidectomy: A ten year experience with fine needle aspiration and frozen section biopsy correlation. Ear Nose Throat J 1994;73:377-80.
(19.) Takashima S, Takayama F, Wang Q, et al. Parotid gland lesions: Diagnosis of malignancy with MRI and flow cytometric DNA analysis and cytology in fine-needle aspiration biopsy. Head Neck 1999;21:43-51.
From the Royal North Shore Hospital, St Leonards, Sydney, Australia. Reprint requests: Nicholas Stow, 63 Beatrice St., Balgowlah Heights, N.S.W. 2093, Australia. Phone: 61-4-0163-2762; fax: 61-2-9210-3153; e-mail: firstname.lastname@example.org
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Original Article|
|Publication:||Ear, Nose and Throat Journal|
|Date:||Feb 1, 2004|
|Previous Article:||Leiomyosarcoma of the maxillary sinuses: report of two cases.|
|Next Article:||CME test.|