6th ESACP Congress, Heidelberg, April 7-11, 1999

A136
QUANTIFICATION OF TUMOUR ANGIOGENESIS : PART I - CLINICAL RELEVANCE AND PROBLEMS CAUSED BY MICROVESSEL DETECTION AND QUANTIFICATION
Tran TKN

Centre Francois Baclesse, Caen, Pôle Traitement et Analyse d'Images de Basse Normandie, Caen, France

Angiogenesis is tightly regulated in normal physiological conditions; in cancer its control mechanisms are disturbed, leading to growth of newly formed blood vessels. Folkman J. (1) has first suggested that neoangiogenesis is essential to tumour progression. There is now evidence that tumour mass expansion and dissemination require neovascularisation. Tumour angiogenesis is more often heterogeneous and tightly packed microvessels are localised in so-called ôhot spotsö. Quantification of angiogenesis in these areas has proved to be of clinical interest. Microvascular density strongly correlates with grade, prognosis of numerous cancers and occurrence of metastases. Multivariate analysis of prognosis factors has underlined the particular interest of angiogenesis quantification in breast and ovarian carcinomas (2, 3). These results lead to a new concept in anticancer therapy strategy, targeting both tumour and endothelial cells. Specific antiangiogenic treatments against solid tumours were introduced six years ago and, in 1996, nine molecules were tested in clinical trials. Prognosis and therapeutic assessments require neoangiogenesis quantification. For this purpose, methods which are less tedious, less time consuming and more reproducible than counting through the microscope lens must be developed and tested. The different approaches found in literature and unsolved problems will be discussed here. Reproducible selection of hot spots seems to be the first and most difficult problem to overcome. Manual, stereologic and even semi -automated techniques have been proposed, nevertheless, none of them fully solves this problem. Thus, identification of areas of intensive neovascularisation as well as angiogenesis quantification must be carried out by a fully automated counting method.
1. Folkman J (1971), N. Engl. J. Med., 285:1182-1186
2. Toi M, Kashitani J, Tominaga T (1993), Int. J. Cancer, 55:371-374
3. Schoell W, Pieber D, Reich O, Lahousen M, Janicek M, Guecer F, Winter R (1997), Cancer, 80:2257-62