Developing and using the computer based sperm motion analysis system to predict the pregnancy rate

INTRODUCTION: Infertility is defined as 1 year of regular unprotected intercourse without conception. Semen analysis is the gold standard method for the evaluation of male factor in infertile couples. Semen analysis is performed manually and/or with automated systems (computer assisted sperm analysis - CASA). Sperm motility is an important parameter affecting fertility. OBJECTIVES: In our country we perform manuel semen analysis due to CASA systems costs. CSCTAS software developed by Yıldız Technical Universty Department of Computer Engineering, which analyzes the motility parameters measured by CASA. It provides faster, standardized and reproducible results with less cost. This system is intended to be used to predict pregnancy rates. METHOD: The study was conducted between January 2018 and December 2019 at Division of Reproductive Endocrinology and Infertility, İstanbul University Faculty of Medicine. The study group consisted of 184 unexplained infertile couples willing to participate in the study. Women were under 35 years old, men were under 41 years old. Detailed sperm motility analyzes of the samples prepared for intrauterine insemination (IUI) after ovulation induction (OI) were performed by manual method and CSCTAS. The relationship between the results and pregnancy was investigated. RESULTS: The number of couples who can get pregnant is 29/184 (%15,76) during the study period. The distribution of the groups was similar in terms of female age, male age, infertility duration, female and male smoking, female body mass index and used induction agents. In the analysis of movement parameters, WOB (wobble) is significantly higher in cases with pregnancy (p=0.004). According to the decision tree architecture created by Kohen kappa analysis, age is the most important root value. LIN (linearity) was evaluated as the most important sperm motility characteristic determining pregnancy, pregnancy rate increases significantly at values above 0.73. The success of the tree model was measured as 95,65% with the evaluation made using the confusion matrix (kappa=0.845). CONCLUSION: Whereas the most important demographic parameter in couples with the highest probability of pregnancy was age; The most important sperm movement parameters were determined as linearity and wobble. Our results need to be confirmed by analyzes involving larger patient groups.