The Rise of AI in Fertility

Between 2015 and 2020, the U.S. artificial reproductive technology (ART) market increased from 231,000 cycles per annum to 326,000 cycles a year - a compound annual growth rate of 7.13%. As demand for ART services grew, so did its success rate; pregnancies resulting from ART rose from around 30% to 45%.

Despite the average success rate of 45% for U.S. IVF treatments, the cost per cycle of $10-20k remains out of reach for many people who are hoping to conceive. In order to address this issue, start-ups such as Alife Health, AIVF, Embryonics, Fairtility, and others have created AI solutions that aim to raise the success rate and reduce the number of cycles needed in a typical IVF treatment course.

These companies offer services like precise embryo selection, implantation prediction and non-invasive PGT (preimplantation genetic testing) which enable embryologists to make accurate decisions without bias while saving them from time consuming manual selection and data entry tasks. This innovation has attracted investments worth over $88M from investors like Lux Capital, Maveron, Union Square Ventures and Insight Partners.

There is clear excitement and conviction around the technology, and we cannot wait to see these technologies become FDA approved to showcase their true potential. Below, we will highlight a couple AI offerings that could make a big impact in an IVF treatment.

Embryo Selection Software:

• Through research, it was found that the speed at which blastocysts (rapidly dividing ball of cells) develop have a statistically significant correlation with high live birth rate (LBR) and implantation rate (IR). Embryos that developed into blastocysts on day 5 and were transferred had a higher LBR than day 6 blastocysts (60.4% vs. 44.8%). Even when holding the quality of the blastocysts constant, day 5 blastocysts had a higher LBR (72.8% vs. 56.5%) than day 6 blastocysts, and a more significantly IR result of (77.7% vs. 58.7%). This is relevant as AI software can select the highest quality embryos based on the growth speed of embryos, and increase the overall success rate of an IVF procedure.

Preimplantation Genetic Testing (PGT) Software:

• Many studies also highlight the importance of speed in blastocysts expansion. It was found that the averaged euploid blastocysts expansion rate was significantly (52.8%) faster than in aneuploid blastocysts (P=.0041, ridiculously predictive). PGT is important because aneuploid embryos typically have a lower chance of implantation, higher chance of miscarry and serious medical problems such as Down Syndrome. During PGTs, the goal is to pick an euploid embryo to maximize an IVF cycle’s success rate. The implication here is that computer vision algorithms can potentially be used to automatically identify and accurately measure the speed of blastocysts expansion that correlates with euploidy/aneuploidy conditions, which ultimately improve implantation and live birth rates.

Oocyte Analysis Software:

• Studies have shown that AI may have the potential to improve in-vitro fertilization (IVF) and egg freezing processes by providing accurate predictions about an oocyte's developmental capacity. A study using a machine learning model analyzing cumulus cells surrounding the oocyte showed an 86% accuracy in predicting its ability to develop into a top-quality day 5 blastocyst. This is impactful as top quality blastocysts present over an 80% chance of resulting in a healthy pregnancy and live birth. This indicates that AI tools could potentially improve IVF success rates and accurately assess egg quality for long term storage, enabling providers to make informed decisions about gamete use even years before it is needed.
  

The potential for AI’s application in IVF is endless, and we can’t wait to see it come into effect post-FDA approvals. At Recharge, we will continue to track the rise of AI in fertility, and invest accordingly following our thesis in building the most comprehensive ecosystem in fertility treatment.

‍