P-102 Incorporation of an automated sperm counting method: a matter of time

Abstract

Abstract Study question Is a fully-automated system a faster and reliable sperm counting method? Summary answer A fully-automated sperm counting system is a reliable tool that significantly reduces time spent for semen evaluation. What is known already The emergence of automated tools in our everyday life is a reality. Fully-automated systems for the evaluation of seminal samples are available for the andrology laboratories. The expected advantages of such analysers would be a higher reliability and reproducibility of results, with lower inter and intra-individual variability. Although there are few published validation studies, available data have evidenced that results obtained with these methods are, in general, consistent with those obtained with manual/conventional analysis. Moreover, automatic sperm assessment could be useful in reducing the time spent for analysis. However, no data have been published in this regard. Study design, size, duration Observational retrospective study of 220 semen samples analyzed in 2022. Every sample was analyzed using three different methodologies: manual analysis, semi-automated and fully-automated. Manual assessment was performed by senior embryologists undergoing periodical quality control evaluations and included counting with improved Neubauer chamber, motility assessment with slide with phase-contrast optics, and morphology analysis by Diff-Quick stain. The semi-automated analysis was performed by SCATM plus manual morphology assessment. The fully-automated analysis was exclusively performed by SQA-VisionTM. Participants/materials, setting, methods Male patients consulting for infertility and undergoing a spermiogram in a single centre. The sperm parameters evaluated for correlation study among methodologies were: total sperm, motile sperm and progressive sperm concentrations (106 per ml), and morphologically normal forms (%). Duration of analysis (sperm count + motility + morphology) using each method was recorded and compared. Intraclass correlation coefficient (ICC[95%CI]) was used to compare agreement between methodologies and Wilcoxon’s test was used for time (median (p25-p75)). Main results and the role of chance The observed correlation between the three methodologies was moderate (ICC=0.5-0.79) to high (ICC >0.8) for all parameters analysed. Concerning total sperm concentration, both fully and semi-automated systems showed a high correlation with manual assessment (ICC=0.93 (0.90-0.95) fully-automated vs manual; ICC=0.83 (0.73-0.89) semi-automated vs manual, respectively). Concerning motile sperm concentration and progressive motile sperm concentration, both fully and semi-automated systems showed a high or moderate correlation with manual assessment (ICC =0.86 (0.77-0.91) and ICC=0.87 (0.76-0.92) fully-automated vs manual; ICC=0.54 (0.27-0.70) and ICC=0.62 (0.39-0.76) semi-automated vs manual, respectively). In terms of normal forms, fully-automated showed a moderate correlation with manual assessment (ICC=0.69 (0.57-0.78)). Regarding time spent for analysis, both fully and semi-automated analysers showed to be faster than manual method (p < 0.001) (1343 (1199-1472) seconds faster fully-automated vs manual and 608 (429-769) seconds faster semi-automated vs manual). Comparing time spent for analysis between the fully and the semi-automated systems, fully-automated was 707 (610-784) seconds faster (p < 0.001) than the semi-automated. Limitations, reasons for caution Retrospective study. SCATM is a semi-automated operator-dependent system based on image analysis and its accuracy relies on the number of sperm analysed and the operators’ ability in reviewing images. SQA-VisionTM is fully-automated, based on electro-optical signal analysing millions of sperm. Therefore, conclusions may not apply to other software operators. Wider implications of the findings Fully-automated systems for sperm assessment can be reliably introduced in andrology laboratories offering a dramatic reduction in the time spent for analysis. However, such analysers may only be cost-efficient in centres with a very high activity. Trial registration number not applicable