Validation of a system for computerized measurements of spermatozoal velocity and percentage of motile sperm.

Abstract

An objective method for measuring the percentage and velocity of motile sperm was evaluated. Super-8 motion picture films of bull sperm in buffer or TEST extender (30–40 × 106 sperm/ml) in a special chamber 5 μm deep were prepared at 18 or 25 frames/sec using phase-contrast optics. Films were projected frame-by-frame onto a screen viewed by a video camera. The image for each frame, containing 20–45 sperm, in a scene was digitized at 65,536 points and the computer determined the location of each sperm head. The position of each sperm head in the 3–7 subsequent frames of the scene was correlated with its position in the first frame. The computer determined 1) the velocity of each sperm and designated it as motile (>35 mm/sec) or immotile, 2) the mean velocity for all motile sperm, and 3) the percentage of motile sperm. Sperm which were not located in all frames of a scene, usually because they left the field of view or turned edgewise, were considered to be motile. Analyses based on 4 frames/scene were more precise than those based on 5–8 successive frames. Variation among scenes within slides was significant, but if 5 scenes for each of 2 slides per sample were evaluated, the predicted standard deviation was <5.5 percentage units for motile total sperm. By suspending sperm in buffer containing Ficol, it was shown that a 15–20% decrease in spermatozoal velocity could be detected. Correlations between the ratio of fresh to killed sperm in mixed samples and computer evaluations of the percentage of motile sperm were >0.94. However, for samples containing only killed sperm, 31% were evaluated as motile defined sperm and, thus, computer evaluations were inaccurate for samples containing few motile sperm. For samples containing a high percentage of motile sperm, the accuracy was good and measurements by computer agreed with subjective evaluations of the same frames of film made by technicians. The present system for computerized measurement of spermatozoal motility is useful for evaluating fresh samples of semen having a high percentage (≥50%) of motile sperm. For samples containing few motile sperm, other objective methods are more accurate. With refinements to locate more precisely the midpoint of each sperm head, computer evaluations of spermatozoal motility might prove to be useful for evaluation of sperm samples containing either a high or a low percentage of motile sperm.