Validation of a novel computer-assisted sperm analysis (CASA) system using multitarget-tracking algorithms

Abstract

To determine the accuracy and precision of a novel computer-assisted sperm analysis (CASA) system by comparison with existing recommended manual methods. Prospective study using comparative measurements of sperm concentration and motility on latex beads and immotile and motile sperm. Tertiary referral fertility center with strong academic links. Sperm donors and male partners of couples attending for fertility investigations. None. Achievement of Accubead target value for high and low concentration suspensions. Repeatability as demonstrated by coefficients of variation and intraclass correlation coefficients. Correlation and limits of agreement between CASA and manual methods. The CASA measurements of latex beads and sperm concentrations demonstrated a high level of accuracy and repeatability. Repeated Accubead measurements attained the required target value (mean difference from target of 2.61% and 3.71% for high- and low-concentration suspensions, respectively) and were highly reproducible. Limits of agreement analysis suggested that manual and CASA counts compared directly could be deemed to be interchangeable. Manual and CASA motility measurements were highly correlated for grades a, b, and d but could not be deemed to be interchangeable, and manual motility estimates were consistently higher for motile sperm. The novel CASA system was able to provide semen quality measurements for sperm concentration and motility measurements which were at least as reliable as current manual methods.