Current testing methods are resource-hungry and continue to show poor reproducibility between staff and laboratories in different hospitals. Sperm quality assessments are made as part of the routine work up for the “subfertile” couple, during sperm processing for artificial insemination (AI) or in-vitro fertilisation (IVF), when sperm are frozen (cryopreserved) either for patients who undergo sterilizing therapies such as chemotherapy or for those who donate sperm.
The team’s innovation is a computerised sperm tracking system which uses novel computer vision algorithms to detect, count and evaluate the overall quality of human sperm. Recently, it has already helped the NHS fertility service in Nottingham to improve standards and increase patient throughput. The aim now is to market an affordable CASA system that can provide rapid, accurate and repeatable semen quality data for NHS fertility/IVF clinics, private facilities and District General Hospital (DGH) pathology laboratories.
Early development work and proof of principle was established using both video images and fresh semen samples. It focused on object recognition, confirming that sperm can be tracked using the algorithms specially created for CASA.
Once this was established, the next step was to ensure that “non-sperm” elements, commonly found in semen samples such as leucocytes, germinal cells, epithelia and cell debris could be successful eliminated from inclusion in the analysis and that a large population of sperm could be detected simultaneously in a matter of seconds. Work has since been published in Fertility and Sterility, a well known international fertility journal, to show that sperm concentration as measured by CASA compares favourably with current gold standard manual methods recommended by the World Health Organisation (WHO).
We have already shown accuracy as good, if not better than, even the most diligent human can manage
The current focus is on improving video quality and tracking accuracy. We have already shown this is as good, if not better than, even the most diligent human can manage but we are working on improving things further. Not only have we shown parity with manual testing, but recently generated clinical data demonstrates very well how the results obtained with the system relate to outcome, which is of course pregnancy.
Tracking accuracy can already be verified by simple review of the generated video. This allocates a colour code for the 4 standard grades of sperm swimming speed (fast progressive – red, medium progressive – green, non progressive – blue, and static – yellow). Version II of the prototype is now ready for re-testing and validation, both in Nottingham and other independent test sites which includes clinical trialing in 3 major IVF centres.
When this validation is complete, our aim is to see the system used in the fertility and pathology sectors providing a new standard for semen quality analysis. This will lead to improved confidence in semen quality testing for GPs and fertility specialists in both the UK and internationally.