Semen Analysis with Machine Intelligence: evolution of technology

Despite the availability of computer-assisted semen analysis (CASA) systems over the past 25 years, they are by no means universally used for clinical andrology, and laboratories continue to favour manual methods that on the whole are time-consuming, error-prone and subjective. At the same time, automation in other areas of laboratory medicine has become main-stream. The high level of subjectivity (and therefore uncertainty) associated particularly with motility and morphology is exacerbated by the lack of a suitable standard, comparator or calibrant.

The results of external quality assessment (EQA) schemes continue to highlight an industry-wide problem. These are designed to provide reassurance both to the laboratories and to users (ie clinicians, patients) that the service is capable of providing reliable and clinically relevant results no matter which operator provides it or in which laboratory. However, the example shown in Figure 1, even for a measure with lower associated subjectivity such as sperm concentration, shows a clear lack of parity across almost 250 laboratories, and not only demonstrates an unacceptable level of uncertainty but fails to instil the level of confidence required in the industry.

The situation for other parameters such as motility and morphology, where there are no universally agreed or traceable standards and no calibrant, is even more challenging. The World Health Organization (WHO)¹ has taken the view that motility needs to be simplified and therefore has changed from four to three grades of motility to improve operator parity. The end-result may well be improved comparability between operators but how does it affect the clinical value of the measurement? Most studies assessing the clinical value of sperm motion characteristics, or even WHO, suggest that sperm swimming speed is of more clinical value than motility alone.^2,3 Therefore, just describing sperm simply as ‘progressive’ is misleading. In other words, the question is not how many sperm are swimming but how fast they swim.