Abstract

Pure ductal carcinoma in situ (DCIS) of the breast is increasing in frequency. These lesions are associated wit5 an increased risk of developing invasive carcinoma, the risk being related to the (subjectively assessed) grade of DCIS. There are significant problems with reproducibility of subjective grading systems and therefore accurate and reproducible diagnostic criteria need to be established to improve the prognostic value. In order to reassess grading in pure DCIS and in DCIS associated with invasive carcinoma (DCISInv) a number of single, sharply defined cytological and architectural features were re-examined in a set of 70 cases of pure DCIS and 59 cases of DCISInv. Nuclear size and shape factors and % area necrosis were obtained by morphomemc analysis of the DCIS component of all cases. The lesions were graded by two experienced histopathologists and the grading compared with reclassification based on the measured parameters. The reclassification showed a 23% discrepancy in the grading of pure DCIS and a 34% discrepancy in the DCISInv cases based on the measured parameters. Comparison between the quantified parameters in pure DCIS and DCISInv revealed that 7 of the mewxed parameters werc similar in both groups. These findings suggest that reproducible, quantifiable nuclear and architectural features may be used to produce a more consistent grading of DCIS and that several of these morphometric parameters may prove useful in predicting the possible hrure development of invasive carcinoma.