outline of woman and heart next to stethoscope on blue background

I recently had the pleasure of presenting research on a topic that is of great interest to me and many of the patients who come to see me – egg freezing. The presentation was based on research I recently performed with colleagues at IRMS, and was based on the development of an egg freezing counseling tool that would provide expected live birth rates before oocyte retrieval and any adjustments to the LB rates after retrieval. A key feature of the tool is that it would be personalized to each woman’s health profile, and validated with previous patients’ IVF outcomes. Below, I’ll provide some more context about this study and the conclusions that were drawn, as they greatly reflect specific elements of the future of egg freezing and the decision-making process. Keep reading to learn more!

Background of the Study

The study began with the overarching goal of creating a counseling tool that would meet three main objectives: answering a woman’s egg freezing questions, providing personal prognostics based on the individual patient, and being validated by past IVF outcomes. We developed IVF prediction models based on discussions with patients and experts, as well as machine learning and expert knowledge. 

In the past few years, there has been a significant increase in the number of women who freeze their eggs. ASRM stated in 2018 that “planned oocyte cryopreservation is an emerging but ethically permissible procedure that may help women avoid future infertility. Because planned OC is new and evolving, it is essential that women who are considering using it be informed about the uncertainties regarding its efficacy and long-term effects.” It has previously been accepted that age and the number of mature oocytes are direct predictors of live birth probability. 

Questions to be Answered

I and other reproductive endocrinologists have found that women looking to freeze their eggs have many questions about the process. As this study was driven by patients, some of the patient questions we aimed to address included:

  • What is the probability of having at least 1 baby?
  • Should I do another egg retrieval?
  • How many eggs do I need to freeze?
  • Does my PCOS or endometriosis impact my chance of success with egg freezing?
  • How do I know that my eggs will work as expected when I eventually use them?

As is the goal of any good healthcare provider, reproductive endocrinologists want to answer these questions with proper transparency and personalization. Some of the questions that guided our focus for the study and the study itself included:

  • How can we use our IVF-LB (In Vitro Fertilization Live Birth) track record to counsel about egg freezing?
  • How can we use validated prediction to personalize LBR (Live Birth Rate) from egg freezing?
  • How can we be sure that the patient understands the egg yield range, and that there is attribution at every step?
  • How do I address the potential impact of PCOS, endometriosis, and low AMH (Anti Mullerian Hormone) in patients seeking egg freezing?

Study Materials and Methods

The study utilized linked IRMS IVF-ET data from 1,166 IVF stim cycles for women under 42 in 2015. The data was not restricted to specific types of infertility, and preliminary analysis indicated that the live birth prediction model is much more robust with a larger and more diverse data set. Predictors of both Pre-OR and Post-OR LB models included:

  • Age
  • BMI (body mass index)
  • AMH
  • D3 FSH (Day 3 Follicle Stimulating Hormone – another measure along w/ AMH of ovarian health and egg number)
  • Clinical fertility diagnosis
  • Reproductive history
  • Semen analysis
  • Oocyte yield (Post-OR model only)

Data was applied and cross-validated with an IVF data set from IRMS. Model performance metrics included posterior log odds ratio compared to age, the effectiveness of the prediction model, reclassification, and dynamic range. Clinical predictors Pre and Post-OR can be seen below:

Conclusions of the Study

Through the utilization of the previously mentioned methods and criteria, we were able to develop IVF prediction models that can be used in a clinical setting moving forward. We confirmed through the study that using ML, LB prediction models clinical variables can be utilized beyond age and oocyte number, and most importantly, the use of a more holistic set of clinical variables results in significantly improved AUC and prediction accuracy. 

We drew upon our goals of transparency and personalization through the process of working with women seeking to freeze their eggs, and based on feedback from real patients- improvements will come in the future. We are so excited to continue to counsel patients using the technology and predictors that are available to us. Our main goal is always to provide patients with as much information as possible so they can be comfortable and make an informed decision for themselves and their future families.

The tools and models used in this study are not yet available in our office. At this time, IRMS physicians do draw on a tremendous base of knowledge and data to individually counsel patients about whether or not egg freezing is a good choice for them. We hope that making use of developing technology will make these decisions easier in the very near future. Should you freeze your eggs? Ultimately that is a very personal and individual decision. The most important objective is not necessarily whether you freeze your eggs or not but that you make an informed decision and that regardless of outcome you have no regrets.

If you’re interested in learning more about the study and this topic overall, I encourage you to check out my Instagram video that covers the presentation in further detail! Please reach out to me or connect with me on Twitter or Instagram if you have additional questions. I would love to speak with you about the future of egg freezing and patient counseling.

Facebook Comments
Recent Posts