CRISMAS - a new system for computer-assisted semen analysis

L. Larsen1, P. Gade-Nielsen2, S. Ziebe3 and A. Giwercman1

1Department of Growth and Reproduction, Copenhagen University Hospital, Denmark
2Image House A/S, Copenhagen, Denmark
3Fertility Clinic, Copenhagen University Hospital, Denmark

 

Introduction

Assessment of sperm motility by manual semen analysis is a subject of great inter-observer variation. Therefore systems for computer-assisted semen analysis (CASA) have been developed in order to obtain a more objective assessment of sperm motility. By CASA several parameters of velocity and linearity, defined by the World Health Organization, are calculated. Until now, however, existing CASA-systems have not fulfilled previous expectations as better predictors of male fertility than conventional semen analysis. One major problem is that the CASA-systems cannot always discriminate between immotile spermatozoa and other static cells in the semen, which leads to incorrect estimations of sperm concentration and motility. We have developed a new CASA-system called "CRISMAS". The main purpose of introducing a new CASA-system is to limit the error rate by optimizing the image quality and implementing tail detection. An ‘edit function’ allows the user to play back the image sequence and correct errors by deleting and connecting tracks. As part of a validation of CRISMAS we have made a pilot study which compares CRISMAS with a well known and widely used CASA-system.

Aim of study

To compare the outcome of two CASA-systems: CRISMAS and the Hamilton-Thorne Motility Analyzer - Integrated Visual Optics System (HTMA-IVOS) (version 10.7k).

Materials and methods

Semen samples were collected from 20 men. The samples were diluted in a phosphate buffer, placed in Makler chambers and analyzed simultaneously with CRISMAS and HTMA-IVOS. The results were examined by the Spearman rank correlation coefficient (rs) and the Wilcoxon matched pairs test.

Results

The measurements of the two systems were generally very well correlated, but at different levels, which often resulted in significant P-values in the Wilcoxon test. The calculated sperm concentrations were significantly correlated (rs=0.86, P=0.0001). However, CRISMAS generally measured concentrations that were lower than HTMA-IVOS, but differed less from conventional sperm concentration than HTMA-IVOS did (median differences = 6.7 x 106 and 35.5 x 106, respectively). The proportions of progressive and non-progressive spermatozoa also showed good correlations, although the absolute values differed slightly (P<0.05). The velocity parameters were clearly correlated, but CRISMAS generally measured lower velocities than HTM-IVOS.

Discussion

We have developed a new CASA-system, CRISMAS, and compared it with the well known HTMA-IVOS system. CRISMAS measured more accurate sperm concentrations than HTMA-IVOS, which is probably partly due to the use of the ‘edit function’ of CRISMAS, but not of HTMA-IVOS. CRISMAS measured a lower proportion of immotile spermatozoa and lower sperm velocities than HTMA-IVOS. This can be explained by the differences in the settings and definitions of the two systems.

Conclusion

This pilot study has proven a new CASA-system, CRISMAS, a reliable alternative to existing systems. Furthermore, the study clearly illustrates the need for a standardization of the settings and algorithms of CASA-systems in order to avoid variation between the measurements of different systems.

Poster presented at Measuring Behavior '98, 2nd International Conference on Methods and Techniques in Behavioral Research, 18-21 August 1998, Groningen, The Netherlands

© 1998 Noldus Information Technology b.v.