Written by: Eduardo Hariton, MD, MBA, and Kevin Loewke, PhD,

Eduardo Hariton, MD, MBA, reproductive endocrinology and infertility fellow at UCSF Center for Reproductive Health, joining RSC Bay Area later this year. Kevin Loewke, PhD, head of data science for Alife Health.

Deciding when to administer the trigger shot is one of the most important decisions a clinician makes during an IVF cycle, and machine learning (more broadly referred to as artificial intelligence, or AI) can help make that decision. But, first, why is the trigger shot so important?

The timing of the trigger shot may make a big difference in how many eggs are retrieved during an IVF cycle. During IVF, a patient is prescribed medications to stimulate the growth of multiple follicles in the ovaries. This process is called “ovarian stimulation.” The follicles must grow to a certain size before the eggs inside of them can be retrieved. Once the follicles have reached the right size, the doctor prescribes a “trigger shot,” which starts the maturation process and helps physicians time the egg retrieval so eggs are harvested before ovulation. Since the timing of the trigger shot affects how many mature eggs are collected, it also has implications on how many eggs are fertilized and how many blastocysts (embryos) are available for transfer. Triggering too early may result in immature eggs, while triggering too late may result in post-mature eggs or fewer eggs collected, as well as increased risk of ovarian hyperstimulation.

According to the study we recently published in Fertility and Sterility, our research suggests that machine learning can help clinicians choose the optimal trigger day. In this case, machine learning – in which computer algorithms learn to recognize patterns from large datasets – may help doctors decide on the optimal trigger day based on information from the outcomes of thousands of previous ovarian stimulation cycles.

Study results indicate that Alife’s machine learning model could help doctors optimize the trigger in the 50% of patients who are triggered too early or too late, and help retrieve up to three more mature oocytes (eggs), two more fertilized oocytes (eggs fertilized by sperm), and one more usable blastocyst (embryo) on average, for that subset of patients. Each additional mature egg retrieved helps to increase the opportunity for a patient to have a successful outcome from their IVF cycle.

While our analysis in this study was drawn from Alife’s dataset of over 30,000 historical IVF cycles, it does have limitations. The retrospective nature of the study means that researchers only looked back at IVF cycles that were already completed. The next step for this technology is to perform real-world, forward-looking studies to confirm these findings.

This is exciting because the study demonstrates that AI has the potential to make IVF more effective. While more research is needed, it’s encouraging to have early proof of the difference that AI can make in a patient’s IVF cycle.