From the birth of first IVF baby in 1978, more than 8 million babies have born worldwide through IVF. Due to unabated developments in this field, the global IVF market size is expected to reach USD 36 Billion by 2030. In India, IVF is currently a USD 746 Million industry and expected to double in the next 5 years. However, the success rate of IVF globally is merely in the range of 35-45% varying depending on women’s age, BMI, quality of egg and sperm to name a few. This low success rate means that there is still a lot of opportunity to improve success rates where a small advancement in technology can lead to a higher chance of starting a family. Apart from the clinical parameters, IVF success rate is also dependant on the clinics patients choose for undergoing IVF. The clinical infrastructure in the IVF lab, conditions like temperature and air quality within the lab and judgement of embryologist greatly affects the success rate across clinics. The judgements of embryologists are largely based on their experience, skill and a set of rudimentary rules leading to huge subjectivity in decision making at multiple junctures during an IVF cycle. 

This is where Artificial Intelligence can change the course of IVF. IVF process typically involves 8 stages viz. clinical assessment, pre treatment medications, stimulation protocol, egg retrieval, semen collection, embryo culture, embryo transfer and pregnancy test. As patients go through this process, the success rate decreases with every stage. Artificial Intelligence finds its importance in these strategic decision-making points which impact the overall success of IVF through an objective decision-making approach.

AI in Fertility and IVF Treatments:

  • Clinical Diagnosis to determine the Treatment Protocol: AI tech is armed with longitudinal outcome data from thousands of patient parameters which can be used to predict the most suitable treatment outcomes for the patients. This increases the overall efficiency of the clinicians to select tailor-made treatment for the couple and removes dependency on individual subjectivity to advise treatment. 
  • Defining the Stimulation Protocol: Stimulation protocol is typically defined based on patient’s age, AMH and medical history to name a few. Thus, an IVF cycle might differ from patient to patient in terms of stimulation protocol. AI could aid the IVF Specialists take a more objective decision based purely on the patient’s clinical parameters to optimize for best outcomes. It could be applied to data-mine patient records to identify novel markers which can predict pregnancy and live birth.
  • Grading of Oocytes and Sperms: Using static images or time lapse images for direct inspection of oocytes and sperms often gives varied results based on imaging equipment used and subjectivity of embryologists. AI technology can grade the oocytes based on markers of oocyte quality like follicle size, oocyte morphology and cytoplasmic characteristics. Likewise, sperm grading can be conducted to grade basis sperm morphology, sperm count. AI also has the potential to determine the optimal oocyte-sperm combination having the highest potential for success.
  • Grading of Embryos: Currently the embryos are monitored and graded through direct visualization using a light microscope or through time-lapse embryoscope. It grades embryos depending on their quality & ability to reach different stages of development using personal judgement. AI, using routinely generated embryo images performs embryo assessment and grading to offer a non-subjective decision making in choosing the healthy embryos. This reduces the need for scouting for highly skilled embryologists. The technology can also improve the overall throughput per embryologist as it leads to fast computation, decision-making and EMR process reducing the manual workload. Furthermore, AI has proved to improve the accuracy of manual grading and predicting development of blastocyst & pregnancy. It has also proven to precisely quantify attributes like size, shape, area, proportion and symmetry. That’s definitely not something a human can do. To summarize, by using AI, we can reduce the number of embryos transferred, reduce the risk of multiple pregnancies and reduce number of repeat cycles by transferring only the highest quality embryos which ultimately impacts the pregnancy outcome
  • Genetic Diagnosis: AI can be used to analyse genetic data and predict potential genetic issues before conception. This helps the clinicians, and couples make more informed decisions about their treatment options

India currently has close to negligible penetration of AI technology within the IVF industry despite myriad of applications as there are certain barriers to entry which need to be crossed to boost integration of AI in reproductive medicine. 

Challenges:

  • The magnitude of data required to train the algorithm to give accurate results is very high. Compare this with the lack of digitization due to higher reliance on physical medical records in unorganized sector in India. Thus, getting large-scale high-quality data would be one of the biggest hurdles to cross in India 
  • Adoption of AI Technology in Indian IVF clinics might witness resistance towards change as this is perhaps a mentality shift in terms of revalidating existing clinical and operational workflows
  • For the local IVF clinics which account for almost 60-70% of the market share in India, implementing AI in their clinics might not come at low cost and hence investments could be one of the barriers
  • Currently there is lack of prospective research on efficacy of AI implementation which could stone-carve guarantee the magnitude of impact on the live birth rate using AI

Despite the challenges, AI, undoubtedly has a critical role to play in the journey towards personalized reproductive medicine and higher pregnancy outcomes for patients. However, a lot still needs to be done in research and development to integrate AI completely in reproductive medicine.

Linkedin
Disclaimer

Views expressed above are the author's own.

END OF ARTICLE