Cytohistological correlation of grading in breast carcinoma

Assigning a tumor grade to breast cancer provides important prognostic information and guides optimal therapy. The present study was undertaken to grade breast carcinoma on aspirates by Robinson grading (RGS) and Scarff-Bloom-Richardson grading system (SBR). Histological grading was done according to Nottingham modification of Scarff-Bloom-Richardson method and the two cytological grading systems were compared. Fifty cases of breast carcinoma with preoperative cytologic diagnosis were assigned tumor grade. Histologic grading was done on corresponding mastectomy or partially resected specimens. Statistical analysis was performed; correlation between cytologic and histologic grading was established using Spearman correlation coefficient. Regression analysis was done to assess the significance of each cytological feature. Univariate analysis showed strong correlation (P < 0.01) for all features except dissociation. Multiple regression analysis of cytologic features revealed cell dissociation nucleoli and nuclear margin as the most influential features. A concordance of 72.5% between RGS (cyto) and SBR (cyto), 64% for RGS (cyto) and SBR (histo), 82% for SBR (cyto) and SBR (histo) was noted. There was a significant association (P < 0.001) between the grades assigned to cytologic and histologic specimens. Cytologic grade could be used to predict histologic grade as significant relationship exists between grades assigned to cytologic and histologic specimens.