ARTIFICIAL-INTELLIGENCE IN MEDICINE AND MALE-INFERTILITY

Main problem: fertility data is inadequately assessed by traditional s tatistical methods for a variety of reasons. First, the principal test of male fertility potential, the Semen Analysis (SA) is a composite o f several dissimilar parameters, and the SA and other laboratory tests of fertility potential reflect physiological mechanisms that interact in complex ways. Second, patient data is often fragmented, obtained f rom multiple sources. Importantly, 2 patients are required for the fin al result. Methods: Novel and powerful computational method, the neura l network, was explored to analyze fertility data. An integrated serie s of programs was written in the C computer language to implement a ba ck propagation algorithm. A model data analysis system was chosen, pre dicting the penetration of zona-free hamster ova by sperm (Sperm Penet ration Assay (SPA)) and the distance travelled by the farthest swimmin g sperm (Penetrak Assay) from the SA, for these 2 assays are generally believed by the reproductive medical community to be independent of t he SA. The classification accuracy of the neural network was compared to 2 standard statistical methods, linear discriminant function analys is (LDFA) and quadratic discriminant function analysis (QDFA). Results : A neural network could be trained to correctly predict the Penetrak result in over 80% of assays it had not previously encountered, and an other network could predict the SPA outcome in nearly 70%. The neural network was superior to LDFA and QDFA in predicting both assay outcome s (for Penetrak: LDFA = 64%, QDFA = 69%; for SPA: LDFA = 65%, QDFA = 4 5%). Conclusion: The neural network is a powerful method of data analy sis, and may discover correlations in fertility data that are not appa rent by standard statistical analysis. The next step will be to apply neural networks to solve 'real' fertility problems such as predicting male fertility potential.