Microvessel density compared with the Chalkley count in a prognostic study of angiogenesis in breast cancer patients S Hansen, 1,4 F B Sørensen, 2 W Vach, 3 D A Grabau, 4 M Bak 5 & C Rose 4,6 1 Department of Oncology, Odense University Hospital, Odense, 2 Department of Pathology, Aarhus University Hospital, Aarhus, 3 Department of Statistics and Demography and 4 Oncological Research Centre, Odense University, Odense and 5 Department of Pathology, Odense University Hospital, Odense, Denmark, and 6 Department of Oncology, Lund University Hospital, Lund, Sweden Date of submission 18 March 2003 Accepted for publication 22 January 2004 Hansen S, Sørensen F B, Vach W, Grabau D A, Bak M & Rose C (2004) Histopathology 44, 428–436 Microvessel density compared with the Chalkley count in a prognostic study of angiogenesis in breast cancer patients Aims: Evaluation of angiogenesis by intratumoral ves- sel profiles can be performed by different methods. The aim of this study was to investigate the prognostic value of estimates obtained by the intratumoral microvessel density (MVD) method and then to com- pare with corresponding estimates obtained by the Chalkley method. Methods and results: A total of 330 patients treated for primary, unilateral, invasive breast carcinoma were included. The median follow-up time was 14 years and 4 months. The microvessels were immunohistochemi- cally stained by antibodies to CD34. MVD was not significantly correlated with any clinicopathological variables. By univariate analysis, MVD showed no prognostic value with regard to recurrence-free survi- val (RFS) or overall survival (OS), while the Chalkley count had significant prognostic value (P< 0.0001; RFS and OS). In the Cox multivariate analysis, MVD had no prognostic impact {median HR [confidence interval (CI)] was 0.93 [0.66, 1.32] for RFS; and HR [CI] was 0.86 [0.62, 1.19] for OS}, while the Chalkley count [median HR (CI) was 2.12 (1.48, 3.06) for RFS; and HR (CI) was 1.71 (1.23, 2.37) for OS] provided independent prognostic value when adjusted for age, menopausal status, axillary lymph node status, tumour size, histological grade, adjuvant systemic treatment and radiation therapy. In comparing the results obtained by MVD in our study with those from other published studies we find good agreement. Conclusions: The Chalkley count technique seems to be preferable for estimating angiogenesis with regard to the prognostic stratification of breast cancer patients, based on its strong prognostic impact, and acceptable reproducibility. Keywords: breast neoplasms, neovascularization, prognosis, survival analysis Abbreviations: HR, hazard ratio; MVD, microvessel density; OS, overall survival; RFS, recurrence-free survival; vp, vessel profiles Introduction The increasing number of contradictory prognos- tic studies 1 of angiogenesis in breast cancer has not been able to clarify the clinical usefulness of this prognosticator. Generalization from these studies is mainly complicated by the application of different techniques to assess angiogenesis. In earlier reports we have evaluated the reliability of some of these techniques 2,3 and documented substantial variation, mostly related to methodology. Published studies differ by more than methodology alone, and thus the reasons for inconsistencies may be multiple and are yet not completely explained. Conse- quently, we have previously evaluated the prognostic Address for correspondence: Steinbjørn Hansen MD, PhD, Department of Oncology, Odense University Hospital, Sdr. Boulevard 29, DK-5000 Odense, Denmark. e-mail: steinbjoern.hansen@dadlnet.dk Ó 2004 Blackwell Publishing Limited. Histopathology 2004, 44, 428–436