Cytomorphologic Predictors of Adenoid Cystic Carcinoma: Insights From a Multi-Center Logistic Regression Study.

OBJECTIVES

Diagnosis of adenoid cystic carcinoma (ACC) in salivary gland cytology specimens may be challenging due to overlapping features with cellular pleomorphic adenoma and monomorphic adenomas. Due to this difficulty, some are classified as salivary gland neoplasm of uncertain malignant potential. Criteria exist for the diagnosis of ACC but have not been statistically validated. This study analyzes the utility of published criteria for this purpose.

METHODS

A series of 46 salivary gland tumors (26 ACC and 20 other basaloid neoplasms) were investigated by logistic regression analysis for the utility of 11 criteria used in the diagnosis of ACC.

RESULTS

The reproducibility of these criteria was fair to moderate (Cohen's kappa 0.2473-0.5482). Logistic regression showed that the presence of hyaline stroma and absence of plasmacytoid cells, myxoid stroma, or nuclear palisading supports the diagnosis of ACC.

CONCLUSIONS

The presence of hyaline stroma supports the diagnosis of ACC. While cell rings and "finger in glove" structures were frequently seen in ACCs, they were not found to be useful according to the logistic regression analysis for the diagnosis of ACCs.