2. Can we further refine cytologic criteria for atypia in breast fine needle aspirates? a study of three years

(1) To determine frequency of use of the term ‘atypia’. (2) To evaluate various cyto-morphologic criteria to further separate the benign from malignant in our series of ‘atypical’ breast FNA cases. Breast FNA cases reported as atypical/indeterminate performed in our department from January 2004 to December 2006 were retrieved. Patients who underwent surgical resection and/or core-biopsieswereincluded. AllPAPand DQ stained FNA smears were reanalysed. Published criteria were evaluated to define their significance in potentially benign and malignant breast diseases. Our set of criteria included: cellularity, nuclear enlargement, crowding and eccentricity, size of nucleolus, chromatin pattern, cytoplasmic lumina, myoepithelial cells, dispersed intact cells, necrosis, mitoses, foamy and apocrine cells along with various architectural criteria. Cytohistologic correlation and statistical analysis were performed. The total number of breast cytology cases was 2494; 158 were atypical/indeterminate; 118 patients underwent surgical intervention. Histologic follow-up was malignancy in 54, benign in 61 and atypical ductal hyperplasia in 3 cases. Ten of 18 morphologic criteria were statistically significant. For malignant lesions, nuclear enlargement, crowding and eccentricity, dispersed cells, mitoses, cytoplasmic lumina and coarse chromatin were significant. For benign lesions, presence of myoepithelial nuclei and flat sheets proved significant. Using our set of criteria we could recategorise 21 cases as suspicious or malignant and 14 cases as benign, thereby reducing numbers. Thus, the true diagnostic rate for atypia which was initially 4.7%, became 3.3% after recategorisation. ‘Atypia’ in breast cytology represents a true grey zone. Although there is an overlap of criteria amongst cases diagnosed as ‘true atypia’, we strongly recommend the use of strict criteria for excluding benign and malignant cases from those of ‘atypia’.