Early CASA systems proved to be ‘clunky’, inaccurate and provided sperm kinetic calculations which are dependent on the frame rate of the video used and to this day have little or no solid clinical validation data behind them. As such CASA failed to gain early the confidence of the industry and as a consequence some of the negative experiences of users of early CASA systems has had a lasting influence on the users of today. More recent CASA systems like SAMi have been developed which are more user-friendly and have slowly gained acceptance yet many laboratories continue to resist change and find it difficult to part with their well-established manual methods even if they are prone to error and associated with uncertainty. This is perhaps understandable if the CASA is marketed without demonstrating at least parity with the haemocytometer for concentration and were prone to missing sperm, detecting debris and non-sperm cells. Despite providing intricate motility parameters, confidence in the measurements was always likely to be difficult if sperm detection itself was a fundamental problem (Tomlinson et al, 2010).
More recent systems like SAMi have provided a simpler and more intuitive solution, allow editing of the captured sperm screen to ensure that a. only sperm are detected and b. missed sperm can be replaced which gives a more reliable sperm concentration and one which is in line with the haemocytometer (Tomlinson et al, 2010). Moreover if object identification can be considered more accurate then it stands to reason that the ratio of motile to immotile objects will also be classified with more reliability.
Automation of ‘other’ sperm parameters
Some CASA manufacturers have attempted to provide automated capture systems for other parameters such as DNA fragmentation or sperm morphology yet again without having prior clinical validity. DNA fragmentation appears to have some relationship with male fertility yet there are at least 4 commonly used assays available for its measurement, all giving different clinical thresholds. As a result, public bodies such as ESHRE or ASRM failed to reach consensus as to which is the gold standard test (Tomlinson et al, 2013) and as such there is little justification in selecting just one of these for use in a CASA analysis.
Sperm morphology modules have also been offered by CASA manufacturers for more than 20 years but possibly even more than with sperm concentration and motility, laboratories remain skeptical. In this case, users are perhaps right to remain cautious especially since (as discussed previously) the definitions employed for normal and abnormal sperm are associated with such a high level of uncertainty (Tomlinson, 2016). By automating the process the end- result might well be a more consistent outcome but one which could be consistently incorrect. Most are based on the Kruger strict criteria approach first devised in the 1980s, however recent attempts to recreate this data and in a much larger patient setting have largely discredited the early definitions (Van den Hoven et al, 2014). Moreover the Kruger-defined sperm size criteria differ significantly to those mentioned in the WHO handbook of 2010. So neither manual morphology has much credibility and any attempt to automate based on the same criteria is just as flawed. Until the andrology industry as a whole understands better the definition of what is considered ‘normal’, how defects are associated with sperm function and how ‘borderline’ forms should be managed then there is little or no advantage to automating the process.