The variability of measuring sperm concentration and motility as determined by computer assisted image analysis and visual estimation.

Systems of computer-assisted image processing (CAIP) offer good reproducibility and low intra-assay variability between replicate analyses of the same sample. The correlation between these values and those obtained by directly microscopic observation can be well demonstrated. The comparability between systems of different manufacturers is sufficient and satisfactory. Thus, they are well-suited for the routine work in the andrologic laboratory. Since sperm parameters are used for the prediction of male fertility, it is important to know whether computer-assisted analysis improves the predictive power as compared to visual estimation. In our study we compared the variation of semen parameters between two semen samples of the same individual taken in an interval of at least 3 months, by using the CAIP and visual methods. Significantly close correlations between the two values were obtained by both methods. Values of cell concentration showed the highest correlation in both methods (r = 0.75). The lowest correlation, but also statistically significant, was observed in the values of locally motile cells (r = 0.54 for CAIP and r = 0.30 for visual determination). Although apparently higher in CAIP within method correlation coefficients did not show statistically meaningful differences between both methods. We conclude from our study that if values of sperm concentration and motility as measured by CAIP are compared to those obtained by visual measurement, the first method is not superior to the latter in terms of retest reliability.

[The use of computer-assisted semen analysis (CASA) in the andrological laboratory]. / Der Einsatz von computer-assistierter Samenanalyse (CASA) im andrologischen Labor.

Computer-assisted systems for automated semen analysis have been developed to simplify the analysis of spermiograms and to eliminate the subjective influences of the investigator. Such systems are suitable for routine laboratory use only if their results correlate with the results of the conventional spermiogram evaluated by microscopy. We compared the sperm count and motility of 250 patients following analysis with the computer and the microscope. The values for sperm count and total motility evaluated by the computer and the microscope show close correlations, while those for linear and local motility are less closely correlated. The differences can be explained by methodological variations. In general, automated semen analysis appears to be a suitable alternative to the conventional method, even at this early stage of development.