MICROVESSEL DENSITY QUANTIFICATION IN BREAST CARCINOMAS - ASSESSMENT BY LIGHT-MICROSCOPY VS A COMPUTER-AIDED IMAGE-ANALYSIS SYSTEM

Intratumoral microvessel density (iMVD) has been reported as an indepe ndent prognosticator for breast carcinoma patients. iMVD should be mea sured in areas of highest neovascularization (''hot spots''). Yet, mea suring iMVD in the hot spot yields good, albeit variable, results due to subjectivity in identifying the hot spot. The aim of this study was to compare iMVD determinations by light microscopy (LM) between two p athologists of different experience and by a computer-aided image anal ysis system (CIAS). Also, CIAS determined the total microvessel area ( MVA) and the total microvessel perimeter (MVP) within the same field. Microvessels from 91 node-negative, invasive ductal breast carcinomas were highlighted with anti-CD31. Median patient follow-up was 66.3 mon ths (range 3-99). Because of limited deaths from carcinoma, survival a nalysis was restricted to relapse-free survival (RFS). LM-measured iMV Ds correlated between pathologists (p = 0.0001) but with moderate vari ability (r(2) = 0.4). Also, LM-measured iMVDs correlated with MVA (p = 0.0003) and MVP (p = 0.0001). By univariate analysis, the LM-measured iMVD determined by the experienced pathologist (p = 0.002), MVA (p = 0.009), MVP (p = 0.032), and tumor diameter (p = 0.005) were associate d with RFS, Multivariate analysis revealed that tumor diameter (p = 0. 0016), LM-measured iMVD by the experienced pathologist(p = 0.0062), an d MVA (p = 0.0445) were independently associated with RFS, For inexper ienced pathologists, computer-aided iMVD measurement, which measures n ot only iMVD but also MVA and MVP, may be a more objective way to quan tify the angiogenic activity of tumors.