Can A.I. predict egg quality?

There are two important concepts about eggs when it comes to fertility: 1) quantity, and 2) quality. Both inevitably decline with age. However, as we would expect, there is variation between individuals, so tools to assess personalized levels are very useful.

 

The ovaries contain the maximum amount of eggs before birth, with the numbers rapidly declining over the years, such that by the age of 35 only 10% of the eggs that were initially present remain. Once the number of eggs reaches a minimal lower threshold, menopause ensues. While we can never directly measure the absolute number of eggs remaining (ovarian reserve in medical lingo), tests such as antral follicle count, or Anti-Mullerian Hormone (AMH) level are helpful in comparing an individual’s relative egg quantity to what we would be typically expected for their age.

 

Egg quality (oocyte competence in technical parlance) is much more elusive. What we really want to know is the egg’s ability to mature properly, fertilize normally, and develop into a healthy embryo. The decline in egg quality is one of the major reasons that it is harder to become pregnant, but also why the risks of pregnancy losses increase, as well as the chances of pregnancies affected by chromosomal disorders such as Down syndrome. Unfortunately, poor egg quality can be a significant factor even in young patients who are struggling with infertility. Yet, there is no blood test or imaging study that can tell us how good your eggs are.

 

It is not until one undergoes IVF that we can get some insight into egg quality, and even then, simply observing the eggs themselves under the microscope is not terribly informative. We have to attempt to fertilize with sperm and observe embryo development. Only then, once we have put the eggs to the task, can we make some inferences about quality. (Of course, sperm quality also plays a role, a topic for another day).

 

This puts our egg freezing patients in a tricky situation: because we will not be fertilizing the eggs until some uncertain time in the future, it is difficult to say how good their eggs really are. If we wait until years later, only to find out then that the eggs are not healthy, valuable time would have been lost. We do try to make some predictions using age and the number of frozen eggs to estimate the probability of eventually leading to live birth (check out our egg freezing estimator here), but accounting for only these two factors, all patients of the same age and same egg number together are clumped together. For example, all 35-year-olds with any 10 eggs will be given the exact same odds for success. Such estimations, while better than nothing, are admittedly very crude and may not be too meaningful for any given individual.

 

Wouldn’t it be great if we had a tool that actually allowed us to make more personalized predictions, not just based on age and number of eggs, but also on the characteristics of individual eggs?

 

Well, maybe we do: introducing Violet, an image analysis tool from the Canadian start-up, Future Fertility, that leverages the power of artificial intelligence to noninvasively assess oocyte quality. Violet’s algorithm is based on hundreds of thousands of images in the training dataset. Each egg is analyzed independently and a probability distribution is generated to predict its chances to develop into a blastocyst and result in a live birth. Check out these images from a sample Violet report:

 

Timeline

Description automatically generated

 

Violet is the only objective tool that provides personalized insights about egg quality. If it pans out, it could be a game-changer.

 

How does Violet really help? Knowledge is power as they say, and Violet offers a more personalized knowledgebase than we have ever had before. Violet is not a crystal ball (yet), but at this point, I would say it’s a better way to make an educated guess. Instead of playing the odds just based on age and egg number, every patient, every cycle, and every egg can be considered independently. With objective and individualized data, you can make more informed decisions about your fertility preservation plans.  

 

Always looking to embrace technologies that add value, Olive is proud to be one of the first clinics to offer Violet for our egg freezing patients. This is the first of the many AI tools that will inevitably shift the paradigm and change how we practice medicine (and succumb to our robot overlords).