G. Gherardi et C. Marveggio, Cytologic score and DNA-image analysis in the classification of borderlinebreast lesions: A prospective study on 47 fine-needle aspirates, DIAGN CYTOP, 20(4), 1999, pp. 212-218
We prospectively evaluated the accuracy of the cytologic score system devel
oped by Masood et al. combined with DNA-image analysis in the subclassifica
tion of 47 fine-needle aspiration samples with cytologic features of border
line breast lesions. Cytologic scores ranged between 12-18. All cases under
went surgical excision of the lesion, and histology revealed 24 cases of fl
orid hyperplasia, 8 of atypical hyperplasia, and II noninvasive and 4 invas
ive ductal carcinomas. DNA-image analysis demonstrated 33 diploid and 14 an
euploid cases. Diploid samples were divided into slowly proliferating (S G2/M less than or equal to 13%) and rapidly proliferating (S + G2/M > 13%)
cases. By considering florid hyperplasia a "low-risk" lesion and by amalgam
ating atypical hyperplasia, and in situ and invasive ductal carcinoma in th
e category of "high-risk" lesions, the positive predictive value of a score
value > 16 was 100%. In cases scoring less than or equal to 16, the slowly
proliferating pattern had a negative predictive value of 95%, while the an
euploid and rapidly proliferating patterns had a positive predictive value
of 100% and 63%, respectively. We conclude that a combination of cytologic
scare evaluation and DNA-image analysis is very useful in differentiating "
low-risk" from "high-risk" cases in the field of breast borderline lesions,
thus improving the impact of fine-needle aspiration diagnosis on patient,
management. (C) 1999 Wiley-Liss, Inc.