Apport potentiel des lames virtuelles et de l’analyse d’images automatique dans l’estimation de la vascularisation des sutures crâniennes

Abstract

Determining age at the time of death is a difficult problem in forensic practice. The study of the vascularization of cranial sutures is an original approach, which may mark the process of synostosis associated with aging. Counting sections of blood vessels on a histological section of cranial suture, raises however a number of practical problems related to the quality of the preparation and to the method of quantification. The purpose of this study is to illustrate the potential contribution of an automatic analysis of virtual slides to overcome these problems. The proposed method of analysis is illustrated from three samples of frontosphenoidal suture whose vessels were immunostained after decalcification of bone tissue. A high resolution image (x 20 objective and microscopic resolution of 0.5 microm) of each microscopic preparation was acquired through a microscopic scanning device. The automatic image analysis protocol takes advantage of a processing of virtual slides at two resolutions. The chosen strategy ensures the identification of the specific area of interest, the enumeration of blood vessel sections on the whole preparation and the visual control of the detected structures. The quantitative estimate of the vascularization of a large structure, such as a cranial suture, can benefit from scanning and fully automatic processing of virtual slides. The automatic analysis requires however an optimal preparation of tissues; decalcification and immunohistochemical staining must be done with great care.