Fertility and Sterility On Air: Live from the 2025 ASRM Scientific Congress & Expo (Part 1)

Fertility & Sterility on Air is at the 2025 ASRM Scientific Congress & Expo in San Antonio, Texas (Part 1)! In this episode, our hosts Kate Devine, Blake Evans, and Micah Hill cover:

• Anne Steiner, the current Editor-in-Chief of F&S Reviews, and Ruben Alvero, the incoming Editor-in-Chief, discuss the journal’s submission process and future directions (01:06)
• Complexity of genetics and available testing modalities with Mili Thakur (06:45) 
• Optimizing embryo utilization: cost effectiveness of cryopreservation in the face of legal and financial uncertainty with Allison Eubanks (16:25) 
• Optimizing pain control following oocyte retrieval - a randomized blinded placebo controlled trial assessing ketorolac efficacy and fresh embryo transfer outcomes with Marja Brolinson King and Laura Zalles (21:42) 
• Discussing the Fertility and Sterility Editorial Fellowship program with F&S's first Editorial Fellow, Allison Eubanks (30:22) 
• Beyond expert grading: comparing a fully automated AI system to five expert embryologists for embryo selection with Mohsen Bahrami (36:20)

Welcome to Fertility and Sterility On Air, the podcast where you can stay current on the latest global research in the field of reproductive medicine. This podcast brings you an overview of this month's journal, in-depth discussions with authors, and other special features. FNS On Air is brought to you by the Fertility and Sterility family of journals, in conjunction with the American Society for Reproductive Medicine, and is hosted by Dr. Kurt Barnhart, Editor-in-Chief, Dr. Eve Feinberg, Editorial Editor, Dr. Micah Hill, Media Editor, Dr. Pietro Bortoletto, Interactive Associate-in-Chief, and Associate Editor, Dr. Kate Devine.

Welcome, everyone, to Fertility and Sterility On Air. This is our first interview at ASRM 2025. We're here live with the Media Editor from Fertility and Sterility Reviews, Dr. Blake Evans.

Blake, I'll hand it over to you for the interview. Thank you, Micah. You have such a great intro voice.

I'm so jealous of that. Well, welcome, everyone. I, as Micah had said, I am Blake Evans, FNS Reviews Media Editor, and I am honored today to be chatting with the current Editor-in-Chief, Dr. Ann Steiner.

I asked Dr. Ruben Alvaro, who is going to soon be the next Editor-in-Chief. I made sure I could say that out loud first. He is here with us also, and welcome.

Thank you all for joining. Thanks for having us. Yeah, thank you for having me as well.

Absolutely. I wanted to just do just a few minutes and chat about FNS Reviews, just kind of give a review of what that entails and what type of journal FNS Reviews is. Thank you so much, Blake.

As a reminder, FNS Reviews publishes only review articles. We take solicited and unsolicited reviews in both narrative and systematic reviews. In general, we are looking for any topics related to reproductive medicine.

It can be clinical medicine. It can be basic science, but really focused on reproductive medicine. Excellent.

If someone were to want to submit to this journal, how does that happen? What's the process like? We encourage everyone to submit first a pre-submission proposal where you just basically give us an idea about what you want to do. We want to know kind of what your reason for doing this is, what's the premise, and how you're going to approach the review on the whole. That is reviewed by our editorial board just to make sure that we think it will be a good fit for FNS Reviews.

Then we generally let the author know about that and actually provide some feedback, which we hope makes it for a better article in the end. Then we have the person draft the manuscript and subsequently submit through our Elsevier website the manuscript for review. Excellent.

This is even a good opportunity for fellows or someone who's researching a topic of interest or maybe something that they can't necessarily do a randomized controlled trial for, something that's maybe even controversial, not a lot of data that's currently available, and they can still put together a great publication of what literature is available. I think it's great for people that are getting ready to write their thesis, doing their background research on it, PhD candidates that maybe want to publish their background research. You write a grant and you do all that work.

Let's send us the information, have a discussion about might this be a good fit for your work. I think it is a great, great way for senior faculty to work together with trainees to really develop some good review papers that we hope is going to educate our membership. Excellent.

Then I guess, Dr. Rivera, this would be towards you. Any changes or anything on the horizon for FNS Reviews? No, but first of all, I'd like to really thank Anne for doing such an amazing job over what, is it six years now, I think that you've had? Yes, so six years. She's really taken this from the very beginning to really an incredible journal that is doing great things.

I really, first of all, like to really applaud her for all she's done. I'll give you money later. Going forward, no, I think, first of all, the pre-submission model I think is excellent because I think it both gives folks an opportunity to submit something to see if it's something relevant.

One of the things that we talk about is where are the gaps in knowledge? To see if there is a gap in knowledge that fits our needs and their abilities to produce a project, I think would be a terrific win-win for everybody. I think that that model will continue and be very important. In terms of the kinds of reviews that are done, right now the ones that have been submitted have been narrative reviews and systematic reviews, and I think there's a value to both.

The narrative reviews in particular are something really valuable for somebody more junior perhaps, but also in areas where there's not as much knowledge, which might not lend itself to a systematic review. Systematic review, of course, more rigorous, following a very specific set of guidelines, but also with the ability to really contribute knowledge to the literature insofar as it can take and synthesize knowledge and lead to a policy or to a therapeutic recommendation, which is really important. I think the value of FNS reviews is in that area, is really providing that guidance to societies, to society, and to other areas that might be looking for these answers.

Excellent. Thank you so much. Any last comments that you all want to say to our listeners before we head out? I really would like to encourage, if you're thinking about it, please send me an email.

We'll make that available so that we can talk and see if we can get you to submit something that will be included in the journal, something of good quality, and we look forward to hearing from you. Yeah. Well, thank both of you for all that you do, and we look forward to everything upcoming with FNS Reviews.

Thank you, Blake. Thank you so much. Thanks.

Welcome back to Fertility and Sterility On Air podcast, live from San Antonio ASRM 2025. So I wanted to poll one of our experts in REI genetics, Dr. Millie Thakur from Genome Allies. So for those of you that don't follow her on LinkedIn and other platforms, she really is a voice of expertise and insight into the complexities of genetics in our field.

And so, Millie, I just wanted to invite you here to sort of educate our listeners a little bit about why is it so complex? What's important about the genetic components for what we do in our field? Thank you so much for having me. When we take care of patients who are trying to conceive, we are taking care of a subset of all of the population, one in six. So in those who we are treating and taking care of, there is a nuance about how their genetics work.

So there are some conditions that puts them at risk for infertility, plus also we take care of like couples or women who are advanced age. So with all of that comes the complexity of genetics. And so for an REI practitioner, there is going to be about five to six different tests that we use on a regular basis.

And because of how fast the genomic technology is moving, there is this increase in the complexity of what the doctors and the staff has to navigate in the genetics world. So the different tests that we use on a day to day basis in an REI practice, one of them is carrier screening. And as you know, there is like a wide range of panels in that carrier screening.

The second test that we use is PGT or pre-implantation genetic testing with different types of PGT, PGTA, PGTSR, PGTM, PGTHLA, and now PGTP and whole genome testing. And then there is products of conception testing. So unfortunately, sometimes we will have patients who will miscarry and because of that, you need to have answers from the products.

So there is this whole testing system that's designed around that. The fourth test that we use very commonly is keto type analysis or analysis of the chromosomes in terms of like either recurrent pregnancy loss or severe male factor infertility. And then there are these other tests that are in a way more nuanced and not usually ordered by REIs, but there is testing for single gene disorders either for finding out if the couple that's presenting to us is a genetic carrier of a certain condition that affects them per se, or we sometimes order testing for cancer genetics.

But in the bread and butter everyday genetics in an REI clinic, your carrier screening, PGT, products of conception testing, and then your testing for keto type analysis and why chromosome micro deletion is key. Yeah. The genetic technology is evolving and so we have to navigate that complexity every day.

Yeah. So let's talk about PGTA just because that's something that I guess most of us probably order the most. We use all those things that you just said.

That was a great rundown. When you as an expert look at what platform to use for PGT, like I noticed that there is an abstract here for what people call a non-selection study or a clinical validation study of diagnostic accuracy. What are the things that you look for when you decide what your PGT platform is? You don't have to name specific companies, but just what are the things an REI should be aware of when they're thinking about PGTA testing? Yeah.

So PGTA testing has been available for the last 15 years, has gone through different iterations. And when we are now looking at our primary PGTA platforms, majority of them are next-gen sequencing platforms. And now with integration of the single nucleotide polymorphisms or SNPs, when I'm choosing a PGT platform, I'm looking for number one, that there has been clinical validation.

And then also there is an analytical validation of the platform that we are doing a good job of calling the embryos euploid, aneuploid. And then the aneuploid is then further into two or three categories. And basically what we need to do is we need to acknowledge in the field right now that there is this something called intermediate copy numbers, which has been called MOSES for the longest time.

And we need to acknowledge that these intermediate copy numbers have to be dealt with in a certain way, whether you want to improve the accuracy of the PGT platform or whether you want to have more embryos to transfer, the test has to evolve to be able to call the result as is for that embryo. Yeah. I'm so glad you called it intermediate copy number.

One of my pet peeves is that as a field, we call this mosaicism because I guess you're the expert, but how often do you think it represents true biologic mosaicism as opposed to bioinformatic or test platform inability to call with certainty? Yeah. So I looked into this because I was very interested as to where did this word mosaicism even enter the PGT field. Yeah.

In traditional genetics, and I've trained as a medical geneticist, mosaicism means there are two cell lines in a cell population in an organism. Yes. So when we are looking at embryos, if you trace back the history, there is a paper where they actually had access to full embryos that were donated and then they did testing on them and then found true mosaicism in them.

Okay. That is the first paper that found mosaicism and they had access to the whole embryo. Okay.

But now that word has been used in all different ways and it has come to a point where the doctors and the practitioners in the field are understanding mosaicism as a biological phenomenon when it's a test report nuance. Right. So mosaicism should, I think we should retire that word.

Yeah. And we should start calling it what it is, which is intermediate copy number when you use an exigent sequencing platform and you brought up the non-selection studies. So basically when you're trying to clinically validate tests, especially for like PGTA type tests, you know, there are these non-selection studies that some of the industry partners have done and what that has shown us is that when we call an embryo euploid and the test is good, then more than likely it's going to be a euploid embryo.

It's going to have a good pregnancy rate, low miscarriage rate, best pregnancy rate. Right. When we call it aneuploid and it is going to have a failed implantation, it is going to not be a viable pregnancy most of the time if it's a whole chromosome that is involved.

Right. Then there are these other test results which we now have to acknowledge that they exist. It's not a binary system.

Yeah. Then we have to make sure that, you know, we are able to counsel over patients pre-test and post-test. So they are aware that either we are going to get these kind of test results and how we are going to manage them and what are the rules and guidelines.

And I think that is the call to the whole field from my side is that we should be able to acknowledge that there is this nuanced reporting of PGTA. Right. And then the clinicians do a very good job of who they are offering the PGTA.

Yep. Not everybody should be getting offered the PGTA or their goals are different. Patient goals are different.

So we do a good pre-test counseling. We understand what the patient is going to do with embryos with different type of test results. And then we perform the test in a very good way.

Yeah. We have regulations about how we are going to call results. All labs should have the same cutoffs and similar way of reporting in an ideal world.

Yeah. Then we should have a very good way of providing post-test counseling support. Yeah.

The doctors can pick up the embryos and then be able to acknowledge that there is this nuanced way of reporting. Yeah. One of my fellows, Yono Bardos, from several years ago also is a combined REI and geneticist.

And he published a paper looking at four blinded PGT labs. And there were very different rates of aneuploidy and euploidy, very different rates of live birth within those embryos, which I think really highlighted that not all of these platforms are always the same. And I think what's happening there is from a technical standpoint, if you're using an X-Gen sequencing platform and you're doing a due diligence of how you do quality control at the whole genome amplification process, the results should be the same.

Right. At the end of it, how are you calling these intermediate copy numbers? Right. It's going to matter how many embryos you have for transfer.

Yeah. So if you are able to call them in a certain group, either you call them as this intermediate copy number, report it out, and then have a plan of how we will transfer them. Or your test is giving the test results in a way where you are able to call these low mosaic or low intermediate copy numbers as euploids, then you will have a higher embryo transfer.

Okay. Well, that's fantastic. Thank you for coming on with us and sharing your knowledge.

For those who don't follow Dr. Thakur, I encourage you to do that on LinkedIn. I learn every week new things from her. Thank you so much, ma'am, for sharing your expertise.

We are here again at the Fertility and Sterility booth at ASRM 2025 from San Antonio. I'm here with Dr. Allison Eubanks. She's a fellow who has an oral that is looking at the cost effectiveness of minimizing how many eggs are inseminated.

Allison, first, tell us, just explain the concept of why you felt like this was an important and timely cost effectiveness study to do. Well, thank you for having me. I'm excited to talk about the study.

We felt like this was an important project because current legislation, both in the U.S. post-obs and internationally, are reflective of the internal struggle, the public struggle that both patients and providers have about supernumerary embryos. So if we create too many embryos and we do not utilize them, is that the ethical thing to do? And we're seeing populations kind of move back and forth between what is the quote-unquote right thing to do. And I think in any situation, the more data, objective data you can add to the situation, the better you can counsel your patients and also understand the process yourself to help the patient make the decision that's best for them.

It sounds like a very timely question given the current political climate. So tell us, how did you go about doing this study? How did you set up your cost effectiveness analysis? Sure, so we simulated 10,000 patients. We used fixed model estimates with a Montevideo simulation and we used estimates for fertilization, blastulation, thaw rates, and transfer rates.

And we did, with those 10,000 simulations, we estimated the outcomes for both the general population and then on an age-based population with start age groups and broke it down into looking both at time to pregnancy as well as the cost of overall time to get to pregnancy. We ran the simulation until either the patient ran out of oocytes or pregnancy. Fantastic, and so what did you find? Overall, without breaking it down by age groups, we found that the, what we called Arm B, which was cryopreservation, so the traditional setup where you take a patient through an IVF cycle, you collect all the eggs that you can, and you fertilize all the eggs that you can, watch them all go to blast, and freeze any viable blast, and then go through repeated frozen embryo transfers versus Arm A, which was through an IVF cycle, retrieve all the eggs, and then freeze them, and then subsequently thaw, attempt fertilization, grow to blast, and transfer if appropriate one by one until pregnancy or use of oocytes.

So Arm A versus Arm B. We found that Arm B, which was standard cryopreservation, was both a shorter time to pregnancy as well as overall cost less than the Arm A, which is sequential insemination. Yeah, that makes a lot of sense. What type of cost differences are we talking about? Overall, the cost for live birth with sequential insemination was about $30,000, and it was about $24,000 based on 2023 pricing in Arm B. So we're about $5,000 to $6,000 per pregnancy.

Now, I would guess if in the ages, as you get older, it probably becomes less cost effective to do one egg at a time, just given the live birth potential of a single egg being inseminated? Absolutely. So we found that over the age of 40, actually the majority of women in that group ran out of eggs before they reached pregnancy. Okay.

Well, that's fascinating. What's your take-home point? How do we apply this knowledge, especially given the current political landscape? Well, I don't think that this paper or this study or any study similar to this should be taken as what's right versus wrong. I think, again, what we need to be doing is counseling our patients with more objective data.

So using figures that estimate cost and estimate time to pregnancy and helping them come to the conclusion that's right for them, balancing their ethical challenges in this situation as well. What about from a legislative standpoint? Should legislatures, should they consider this data if they're considering implementing laws that would minimize the number of eggs that could be inseminated? Personally, I think absolutely. I don't think you should be making laws if you're not taking into account objective data on the subject.

I think that if you're going to make a law that is going to restrict somebody's ability to make embryos, you need to know that you're potentially costing them time to pregnancy as well as actual dollars to get to a pregnancy. Again, the whole goal, I think everybody's goal, both on both sides of the argument, is going to be getting somebody a healthy live baby to take home. And so remembering that that is the end goal here.

Yeah. When you think about David Sable's three pillars of how we need to further improve IVF, this violates two of those because it both increases time to pregnancy rather than decreases it and increases cost per pregnancy. So it's kind of the opposite direction of where we'd like to be taking the field.

This is great. I'm glad you had an oral on this and thank you for coming by Fertility and Sterility and sharing this with us. Of course.

Thank you for having me. We are back here live at Fertility and Sterility on air at ASRM 2025 in San Antonio, Texas. We have one of the prize oral papers, a clinical trial on optimizing pain control.

This is a randomized blinded placebo controlled trial assessing catoralac efficacy and fresh embryo transfer outcomes. Marjah Brolinson and Laura Zayas are here to present their clinical trial. So Marjah, just tell me why did you think this was an important enough topic to do as a clinical trial while you were a fellow? Well, the idea for this trial came in my first year of fellowship.

There was a quality improvement project that was occurring that was looking at decreasing narcotic use after egg retrieval. And the question then expanded into what are ways that we can try to decrease narcotic use? One of which would be optimizing pain control following egg retrieval. Toradol is a NSAID that is widely used in many other surgical specialties.

However, people have been hesitant to use Toradol in the fertility world due to the concern for possible risk of post-op complications, bleeding, and the theoretical concern for decrease in implantation due to the NSAID effects on the endometrium. Great. So for either of you, how did you go about setting this up? Just give us the 30,000 foot view.

We understand clinical trials. So how did you set up the study design? So this was a study that we did a power calculation to determine how many patients we needed to enroll. We were looking for a 50% decrease in narcotic use after egg retrieval.

And to accomplish that, we needed to enroll at least 320 patients. So we set our goal at 400 patients to account for dropout. We randomized 200 to the placebo group and 200 to receive Catorlac at the time of the retrieval.

The anesthesia provider administered the Catorlac. Therefore, the patient, the provider doing the retrieval, the post-op nurse, and all of our statisticians were blinded to who received the Catorlac versus who received the placebo. Great.

Sounds like a nice rigorous trial design. Well, everyone wants to know, what did you find? Is this helpful? Absolutely. So our key takeaway points are specifically that we found that administration of Catorlac at the end of the egg retrieval was a simple intervention that was efficacious in improving pain control, decreased narcotic use, was safe with no increase in post-op complications, and most importantly, no difference in fresh embryo transfer reproductive outcomes following the administration of Catorlac at retrieval.

How large of an effect size are we talking about when we're saying it reduced pain or reduced pain narcotic use? Was it from one pill to zero, 10 pills to nine? What are we talking about from an effect size standpoint? Yes. So it reduced the amount of narcotic use after egg retrieval from 42% in the control group to 25% in the Catorlac group. Okay.

And then for administration of IV narcotic, we made a calculation in morphine milliequivalents and the control group required 9.4 morphine milliequivalents, whereas the Catorlac group required 5.2 morphine milliequivalents in Pacu. So I think we would all hypothesize that Toradol would be helpful. I think we all understand that from other surgeries that we do.

And I think a lot of places use this when they're not planning on doing a fresh embryo transfer. So you said with those that got a fresh embryo transfer, there was no reduction in pregnancy. What was the absolute numbers on that? And were you powered to be able to detect a difference in that? Unfortunately, we were not powered to detect a difference in reproductive outcomes, but we did find that for those who proceeded to fresh embryo transfer, which is roughly a third in each group, there were no differences in biochemical pregnancy loss, implantation, clinical pregnancy, clinical pregnancy loss, or live birth.

So we did follow all of the patients out to live birth and there were no differences in the live birth outcomes. Great. So for both of you, Laura, I know you had a big role in helping complete this trial.

Sometimes you start a trial in fellowship and you can't finish it in fellowship. That's just the reality of recruitment. So a question for both of you, what were challenges? What were things that came up that maybe you didn't expect? If you were giving advice to a fellow out there that wanted to do a clinical trial, what would that advice be? I throw all three of those questions to both of you.

Sure. Yes. So I came in, Marjorie brought me in to help a lot with the recruitment, which was probably the biggest challenge.

Even with this being a medication that is routinely used that a lot of patients have heard of, just it being a research study, there was a lot of hesitancy. Initially, we were reaching out to patients when they were triggered. And so they were already at a time in their treatment cycle where they were considering a lot and under stress.

And so that short time frame really made it hard to get them excited about participating in research. I think a really big game changer was when we started working with Engaged MD, who put together a video kind of describing the study that we sent out to patients when they started their stimulation cycle. And so it introduced the study.

Marjorie and I were able to then touch base with them throughout their cycle, answer their questions. And we just saw a much higher proportion of patients being interested and were able to complete the recruitment process that way. So I'd say advice-wise, getting creative with recruitment, seeing where you're having challenges, and really reaching out to patients early, giving a lot of time for opportunities for questions, and then working with other sources that are around you.

Yes, I think if you have a question that you're interested in and passionate about, it's absolutely worth designing a trial and seeking out an answer. I think it's very important to remain flexible in your approach and kind of roll with the punches as they come, because they certainly will come when you're completing a RCT. We have learned so much through this process, and it has made us better clinicians, better researchers, and I think better at giving back to patients through our work completing this trial.

Yeah, that's great. My last question is, would you feel comfortable now giving Toradol to someone with a fresh embryo transfer, or do you think we need a larger trial? I guess it would need to be non-inferiority designed to show that it's non-inferior for live birth or at least pregnancy. Do you feel like we need that bigger trial, or do you feel based upon this that you'd be comfortable giving Toradol to patients with a fresh transfer? I do feel personally, based on the results of this trial, that I am comfortable administering Toradol to patients undergoing retrieval and completing fresh transfer.

The half-life of Toradol is such that it's cleared from the patient's system within one day of administration, so regardless of practices that are completing day three transfers or day five transfers, the medication should be cleared and therefore no longer mediate continued effects on the endometrium. It would be wonderful to have the time and the resources to, of course, complete a larger study, a 2.0, and maybe that could be in the works for the future, but I do feel comfortable based on the results from our trial. Well, I think you gave a good answer from the biologic plausibility standpoint.

Based upon that, combined with your data, I think that's a reasonable answer, and certainly I think our patients would appreciate that. I know patients feel much more comfortable afterwards if they get that dose of Toradol in the OR. Thank you both so much for coming by Fertility and Sterility OnAir and sharing the prize paper oral on your clinical trial.

Thank you, Micah. It was a pleasure. I'm back here live again at ASRM 2025 Fertility and Sterility OnAir podcast.

I'm with the first ever Fertility and Sterility Editorial Fellow, Allison Eubanks. So, Allison, tell me, what is the FNS Editorial Fellowship? Thank you for having me, and I'm excited to talk about the role that I've been so fortunate to be the first to experience. So, I'm here with Allison Eubanks, and she's my editor.

So, my job, in some ways, I feel like it's a little bit of an internship. You know, I'm observing the editorial board meetings. I'm kind of seeing how the process happens behind the scenes.

What are the discussions like when papers are controversial, or how do we encourage writers to write better articles or to edit appropriately? But I also have been fortunate enough to be given the chance to develop ideas for the journal, and I also see this role as a mediator between the experienced, you know, well-involved editors, editors-in-chiefs that have been doing this for 20, 30 years, and then the newer fellows, the newer physicians who don't have this experience, and how do we connect those two groups of people? So, there's a weekly meeting with the editorial leadership across the family of FNS journals, and Allison comes to that. So, that's what she's describing there, is we tackle challenging editorial questions, controversial papers, or papers with split decisions. So, you've got to be involved in learning about that, and your perspective and insights have been fantastic to listen to.

So, you mentioned that you've taken on some projects. Tell us about some of the projects you've taken on, and what those are looking like, and what that might bring to the journal in the future. Yeah.

So, one of the ones I'm most excited about is for FNS Reports. We are creating a new series. There's several series like in other journals like it, and we're going to do a clinical expert series, and so we're taking the challenges that we face as REIs on a regular basis and asking the top people in that field at that specific area of our job, and to help us kind of navigate those situations.

My focus was, you know, as much fun as it is to read about the most rare crazy zebras, I feel like there's a real role for those gray zones, the patients that we encounter all the time that we don't really have, is this the right thing to do for this patient? So, we have four of those already in play, and we're going to get a couple more rolling here to start publishing with FNS Reports. So, these are questions that obviously have less than what would go in FNS Reviews, so they don't have enough for a whole narrative review or a whole systematic review. I guess they're a little bit less data than a mini-review that already exists in FNS Reports.

These are more like sort of niche clinical questions that are data-lite, but we want experts. How do you handle this sort of thing? Right, so if you had this situation as a fellow, if you had a situation like this patient, and you had that one staff who, you know, just navigates this, you know, recurrent pregnancy loss or embryo arrest, and they just seem to do it better than everybody else, kind of how are they navigating that? So, that's really what we're aiming for there. Okay, that sounds excellent.

Is there anything else that you've been working on as far as projects for the journals? Yes, so I got to be involved with all the 75th anniversary production. I got to read all the most cited articles in FNS and kind of pick the top ones that I felt like were most relevant. I've been fortunate enough to also work with the Keeper Program, and I was able to develop a series through the Keeper Program, which is the ASRM's program to mentor residents and fellows and medical students, and we, with that program staff, I was able to develop this series of videos that allowed me to interview people who had written articles either on huge topics like PCOS or POI, but also the people that helped develop the guidelines for those, and I think as a med student resident fellow new staff, some of the hardest parts of the job is how do you take a guideline and use it in your everyday practice, and I really wanted to break that down, so I interviewed some of the people that have written guidelines, the people that were on the paper that tells us how to write guidelines, and break it down for people to understand how should guidelines be used in your everyday life and when to not follow the guidelines.

Fantastic, so you got extended, you did such a good job that Dr. Barnhart extended you for a second year for this fellowship. Next year when the applications open up, who would you encourage to apply for this, and what have been the benefits that you've seen from being part of this process? I think the benefits are honestly uncountable for me, but I think, you know, the original intention of applying for this role was that I hope to be editor-in-chief someday, and I think that picking somebody that or somebody who feels that that's part of their journey in this process, that that's the person should apply. I mean there's a lot of nitty-gritty, there's a lot of smaller bits to it that I think without actually experiencing the behind-the-scenes work, you wouldn't understand what the job of editor-in-chief entails.

I think learning statistics by those discussions, learning the finer details of publication, I think is really invaluable. So did this experience scare you away from wanting to have that role in the future? Absolutely not, I think the challenges are very fascinating. I think those discussions and I think the other thing I've also learned is there's always more to learn, unfortunately or fortunately.

Well that is outstanding, we've enjoyed having you as the Editorial Fellow for Fertility and Sterility, and thanks for sharing the insights on your experience and what you've been learning. Good morning Dr. Barami, thank you so much for joining us here at F&S OnAir from ASRM Live in San Antonio. Good morning, thanks for having me.

I really enjoyed listening to you present your work yesterday in the prize paper session, so congratulations on your research and having been selected to present in the prize paper session. So you talked about your abstract called Beyond Expert Grading, comparing a fully automated AI system to five expert embryologists for embryo selection. Tell me, bottom line up front, what did you find? Yeah actually what we found was that AI can outperform senior embryologists in a statistically significant way, but one interesting thing that we found was that by including the grades from embryologists, we get the best result.

And I think this is because AI can pick all kinds of subtle features to improve the prediction, but still it picks some noise along the way, and by including grades from embryologists, we get this, you know, we include this polished, abstract, experience-based, you know, evaluation from embryologists that improves the performance. Got it, so I like that you already have thought through why that might be. How did you figure this out? Did your AI model test just the AI grading versus just the embryologist grading versus a combination? What was the methodology for your study? Yeah, we actually did three main comparisons.

The first one we just used the grades from embryologists in a simple just machine learning model to predict ploidy status, and then we compared this to a fully automated AI system that uses time-lapse imaging data. So this was like the first comparison between the AI and embryologists or the grades from embryologists. And then we added maternal age and see how, because that's a very important factor, and maybe that could help, you know, embryologists to have a better performance.

But it turned out that still AI performs better, you know, both performances increased, but still AI had a better performance. And finally, yeah, we included the grades from embryologists into our AI system, and that turned out to be the best result, at least for now. Got it, and is this a commercially available AI that you use or one that you developed in your own center? No, this is not a commercial app now.

We started this whole project actually as a research project because we've seen all kinds of apps and all kinds of, you know, interesting studies actually that show AI could be useful, but we had this very, very large data set, and that's how we started this project. We're just interested to see if it really improves, you know, the embryo selection. Got it, and so you also mentioned that you found that the AI was the most helpful tool in terms of predicting euploidy.

Were you also able to look at other outcomes, for example, implantation, live birth, clinical outcomes down the path? You know, the outcome that we were interested because euploidy, you know, the ploidy status is an objective and biologically meaningful measure of the viability of the embryo regardless of transfer-related issues or some other factors like maternal factors that could influence or bias the pregnancy outcome, and that's why we decided to start with ploidy status, but I didn't say that this is, you know, most useful for ploidy status. Maybe it is, you know, it could be for other outcomes, and that's what we're working on right now for ongoing pregnancy, the fetal and heart pregnancy, but I think it's, you know, the reason that we started was that it was a more objective outcome to predict and see a fair comparison between these two. Understood.

So when do you see this actually being implemented clinically as a standard of care at your center? We're planning to do this in the next three to six months. That's the plan. It's ready, you know, everything's been implemented, and it's just some final test that we're doing because I think the best, you know, outcome is when you tailor it to the specifics of the clinic, and that's what we're working on right now.

Understood. So, you know, just to kind of play devil's advocate here because there are a lot of skeptics about AI and really specifically about how AI can help improve outcomes when we're talking about embryo selection because at the end of the day, you know, the supply of embryos that any particular patient has is going to be limited and sometimes extremely limited, and ultimately with the tools we have in terms of morphology and PGTA at the end of the day, you know, really AI only stands to benefit the patient where the embryologist and the AI would disagree, right? And then even if the embryologist was wrong, quote-unquote, and the AI was correct, you're ultimately going to transfer all the embryos in sequence anyway, so the final cumulative live birth rate is unchanged. So given that perspective, how helpful is this really? I agree with what you said, and I think we should be very transparent about, you know, what AI can do.

You mentioned PGTA. That's true. We don't have the accuracy that the PGTA has for this outcome.

I think we should be transparent, and even with patients, I think we should mention that if we have this tool, it's not going to have like an accuracy of a lab test result, but on the other side, we have increasing data and studies that show PGTA may not be as useful for younger patients as it is for older patients, and it's invasive, it's costly, and for example, in our study, we show that, you know, with AI and combining the grades, we can get a very good performance in ranking these embryos, and I think that will have some important clinical implications. Was one of the outcomes that you looked at the proportion of embryos for which the AI and the embryologist would have selected different embryos, and if so, what was the percentage? The percentage, actually, so we just looked at the cohort that we had and ranked them based on the scores from grades and from the AI system, and the AI system, the percentage that we got in which the top-ranked embryo was deployed was 79%, for example. We just used AI, and this was 71% for the grades of embryologists, like an 8% improvement, so it you have eight patients more, you know, with the right ranking, you know, that's in terms of the euploid being the top one, and this, actually, for younger patients, this went up to 87%, which is a significant probability, and I think it just brings more confidence to tell them that, okay, we have this tool, and it ranks it, you know, and you have 87% chance.

So that's interesting, so just so I understand, it was more predictive among younger patients for the endpoint of euploidy, is that right? It was more efficient for ranking embryos, you know, among young patients, that's what we did, but for the AUC, and this is because for younger patients, they usually have more euploid embryos, and that plays a role here, but still, I mean, we compared the two models, and what that was the purpose of this study, still, you know, that's true for grades as well, they have more euploids, but still, the AI did a better job in ranking those embryos. So we've seen that we've approached something of an asymptote in terms of IVF success rates, and haven't really been able to break through that barrier of, you know, somewhere 60%-ish, depending on which clinic you look at and who you ask. What do you think AI can do to change that, not just for embryo selection, but across the board? Yeah, that's a great question, and I think there are multiple factors here, like maternal age, sperm and egg quality, and of course, embryo selection, and I think there is a lot of potential for AI to be used in some of these areas.

We have interesting studies to analyze sperm and egg, and that's another area that we can use AI, but also for embryo selection, and just to be a little more specific for embryo selection, for example, and how this could really, you know, how AI could help, we use the expertise of embryologists, they look at the morphology of the blastocyst, but the problem here is that these grades are subjective, and we don't have necessarily a good consistency, and also we have these arguments that the morphology at a single time point may not necessarily reflect the true viability of the embryo, and this is where AI can really come in and could be helpful, because if we train AI models, at least it brings consistency to the evaluations that we have, and also by analyzing the entire sequence, they can pick a lot of features that could be predictive of the outcome and improve the prediction. Got it. Well, standardization is actually my middle name, so I'm definitely with you on the value that AI has the potential to add there, so thanks again for joining us, Dr. Mohsen Barami from North Carolina, and we'll look for more innovation from you in the AI space.

Thank you, thanks for having me. This concludes our episode of Fertility and Sterility on Air, brought to you by Fertility and Sterility in conjunction with the American Society for Reproductive Medicine. This podcast is produced by Dr. Molly Kornfield, Dr. Adriana Wong, Dr. Elena HogenEsch, Dr. Selina Park, Dr. Carissa Pekny, and Dr. Nicholas Raja.

This podcast was developed by Fertility and Sterility and the American Society for Reproductive Medicine as an educational resource in service to its members and other practicing clinicians. While the podcast reflects the views of the authors and the host, it is not intended to be the only approved standard of practice or to direct an exclusive course of treatment. The opinions expressed are those of the discussants and do not reflect Fertility and Sterility or the American Society for Reproductive Medicine.