Background: Breast cancer is 2nd most common malignancy in Indian women. FNAC is first line of investigation for breast lump. Aims: 1)To diagnose and classify breast lesions on cytology according to probabilistic approach. 2) To evaluate accuracy of probabilistic approach by correlating with Histopathological diagnosis wherever possible. Study Design: Prospective study of FNACs classified into positive, suspicious, proliferative with atypia, proliferative without atypia and unremarkable categories. Results: Total 237 cases were studied. The probability of finding carcinoma on histopathology from positive, suspicious and proliferative with atypia categories were 100%, 80% and 43.75% respectively. All the lesions from proliferative without atypia and unremarkable categories were benign on histology. There were no false positive cases. The specificity and sensitivity were 100% and 90.54%. Conclusion: The probabilistic approach with defined diagnostic criteria is an accurate, uniform and reduces inter-observer variability. It minimizes false positives without compromising sensitivity for detecting carcinoma. This scheme of reporting breast FNA provide the rational basis for guidelines for management of breast lesions.

Keywords: Probabilistic Approach; Breast FNAC; Positive for Carcinoma.