Our goal is to develop a methodology for accurate diagnosis of the fertility of human semen, a methodology with sufficient sensitivity as to be capable of detecting changes in the semen that reflect initial exposure to environmental hazards. This methodology employs both new technology and new techniques of biostatistical analysis. Using automatic, computer-based image analysis technology we will comprehensively characterize the spermatozoa in terms of the morphometric details of their size and shape and the kinetic details of their swimming motion. We will study groups of men with recent proven fertility and infertility and will employ multivariate statistical techniques to determine which parameters of the semen, combinations thereof, best distinguish the two groups. We will also study a group of clinic couples for whom the wives have normal reproductive function but the husbands initially have unknown fertility. We will make use of detailed clinical information on these men, and will analyze a two year followup semen specimen from a subgroup of them. Applying new formulations of the statistical technique of survival analysis, we will develop models that predict future fertility on the basis of present semen quality.
Showing the most recent 10 out of 14 publications
