College of American Pathologists
Printable Version

  Pitfalls in salivary gland fine-needle
  aspiration cytology





cap today

June 2005
PAP/NGC Programs Review

Michael Henry, MD

Aspiration biopsies of the salivary glands are common specimens in most pathology practices and can present with quite difficult diagnostic challenges. The list of malignant and benign neoplastic lesions as well as other processes that can present as masses is extensive with considerable cytologic overlap. Several large series have documented the utility of FNA of the salivary gland with a reported overall accuracy ranging from 87 percent to 100 percent in distinguishing benign from malignant. However, most of these series were generated at large academic centers and may not represent everyday practice in the community.

With this in mind, Jonathan Hughes, MD, PhD, and colleagues used data derived from the CAP Interlaboratory Comparison Program in Nongynecologic Cytology to identify common diagnostic errors in salivary gland aspirates (Arch Pathol Lab Med. 2005;129:26–31). This study is unique in that the nongyn program assesses the diagnostic accuracy of a diverse group of practicing pathologists including academic centers, commercial laboratories, and large and small private practices.

The study had two parts, looking first at overall accuracy in diagnosing salivary lesions by FNA and then identifying individual cases with a high level of discordance between the reference diagnosis and the participant responses. There were 4,642 responses from cases with a reference diagnosis of benign and 1,607 responses from cases with a reference diagnosis of malignant, for a total of 6,249 responses evaluated for the study. The specificity in making a benign diagnosis was 91 percent with a false-positive rate of eight percent. In these benign cases there was a 76 percent accuracy rate in making a correct specific diagnosis. The sensitivity in making a malignant diagnosis was 68 percent with a false-negative rate of 32 percent. The highest false-negative rates were seen in cases of lymphoma (57 percent), acinic cell carcinoma (49 percent), low-grade mucoepidermoid carcinoma (43 percent), and adenoid cystic carcinoma (33 percent). The accuracy in making a correct specific malignant diagnosis was 48 percent.

Selected review of the most discordant cases provided insight into sources of possible misinterpretation. The failure to recognize diagnostically helpful components such as stromal elements in adenoid cystic carcinoma, oncocytic cells in Warthin’s tumor, the monomorphic nature of lymphocytes in lymphoma, and the architectural disarray of acinar structures in acinic cell carcinoma seemed to contribute to misinterpretation issues. The importance of using a Romanowsky type stain, such as Diff-Quik, along with a Papanicolaou stain was emphasized as well as the usefulness of ancillary studies such as flow cytometry in atypical lymphoid lesions.

The authors concluded that awareness of these potential pitfalls and the application of classic cytologic criteria may help improve the performance of salivary gland FNA in clinical practice.

Dr. Henry, a member of the CAP Cytopathology Committee, is staff pathologist at Cleveland Clinic Florida, Naples.