Abstract
Introduction: Cytological diagnosis of borderline breast lesions remains challenging, and interobserver variability exists in their interpretation. The Masood scoring index (MSI) has been proposed to help in the subgrouping of breast lesions using objective criteria. Aim: The aim of the present study was to assess the interobserver variability in the scoring of breast lesions according to MSI and to see the utility of a modified scheme for discriminating benign and atypical lesions. Study Design: Papanicolaou-stained smears (100 cases) that underwent fine needle aspiration for a palpable breast lump were independently evaluated by 2 observers, and the cases were categorized as per MSI. Percent agreement beyond chance score between both observers was calculated. Sensitivity analysis was performed by comparing the scores using models containing different parameters of MSI. Results: The agreement amongst the 2 observers for scores was found to be 0.88 and it was 69% for category-wise diagnosis. Sensitivity analysis showed that the model with only 3 cytological parameters (cell arrangement, pleomorphism, and nucleoli) had similar discrimination ability in the classification of breast disease as benign or atypical as the standard MSI model. Conclusion: Further simplified models of MSI should be tested for improved diagnostic accuracy and wider acceptability.