DISCLAIMER: Today’s Advertiser helped make the production and delivery of this episode possible, for free, to you! But the themes expressed by the guests do not necessarily reflect the views of Inside Reproductive Health, nor of the Advertiser. The Advertiser does not have editorial control over the content of this episode, and the guest’s appearance is not an endorsement of the Advertiser.
When will embryologists be robots?
Dr. Jacques Cohen, Chief Scientific Officer of Conceivable Life Sciences, walks us through the research and development currently underway for the automation of the IVF lab.
Tune in to hear Dr. Cohen discuss:
The next potential game changing innovations in IVF
His opinion on time-lapse incubation and its future in the lab
What the FDA doesn’t like about AI solutions
The $1B project to automating the IVF lab
Dr. Jacques Cohen
LinkedIn
Transcript
[00:00:00] Dr. Jacques Cohen: You don't go to a dentist hoping that your root canal is going to work or not. You go to a dentist and expect it to be a hundred percent successful. Maybe you got a little infection, but that can be treated, but you want it to be a hundred percent successful. And that's what we want in IVF. We want things to be a hundred percent successful, not 98%, not 80%, or what it is now in some clinics over 60%.
No, we want it to be a hundred percent. And we really want that as soon as possible. So, I think all this technology that we discussed today will play a role in that process.
[00:00:33] Sponsor: This episode was brought to you by Future Fertility, the global leaders in AI powered oocyte quality assessment. Discover the power of magenta reports by Future Fertility.
These AI driven reports provide personalized oocyte quality insights to improve treatment planning and counseling for IVF ICSI patients. Magenta can help you to better identify the root cause of failed cycles and counsel patients on next steps to optimize treatment. Download a sample Magenta report plus four key counseling scenarios and see the difference it makes in patient care.
Visit futurefertility.com/ivf. That's futurefertility.com/ivf.
Announcer: Today's advertiser helped make the production and delivery of this episode possible for free to you, but the themes expressed by the guests do not necessarily reflect the views of Inside Reproductive Health nor of the advertiser.
The advertiser does not have editorial control over the content of this episode, and the guest's appearance is not an endorsement of the advertiser.
[00:01:49] Griffin Jones: When will all the embryologists be robots? Soon enough, probably, but that's my speculation. For a more measured walkthrough of what's in the research and development pipeline, For the IVF lab, I bring in veteran lab director, veteran scientific director, Jacques Cohen, Dr. Jacques Cohen, as I should say. And many of you know him very well.
He is now the chief scientific officer of conceivable life sciences. They're working on fully automating the IVF lab. I have Jacques walk us through what they're doing at conceivable at other companies that he's involved with. And I have him walk us through what is preliminary, what's well established, and what's in between.
Is time lapse going to be a must have for embryologists within the next couple years? Dr. Cohen has an opinion. What does the FDA not like about AI solutions? Jacques tells me why and I never knew that. If PGTA and vitrification were among the biggest game changers in the IVF lab in the last decade, what are the next two?
Dr. Cohen walks us through what he thinks might easily be a collective 1 billion project in automating the IVF lab. Enjoy this conversation with Dr. Jacques Cohen. Dr. Cohen, Jacques, welcome to the Inside Reproductive Health podcast.
[00:03:00] Dr. Jacques Cohen: It's a pleasure being here with you, Griff, and really looking forward to it.
[00:03:04] Griffin Jones: Maybe the pleasure should be mine because people say Jacques Cohen is a legend. Jacques Cohen is a legend. And I've worked in the field for nine years. And I think 2023 was the first year that we met in person. So I'm, I'm interesting to, to see if the legend lives up to the hype in this conversation. But many of our listeners are familiar with you already.
And I wanted to go through oftentimes, you know, sometimes I go into the past cause I'm curious about what led to the developments that got us here. I'm more interested in looking at what are today's nice to haves in terms of what's in the R and D pipeline in the IVF. lab that you think are going to be tomorrow's must haves and tomorrow might mean three years from now.
It might mean 11 years from now, but I want to explore that with you. And so maybe just give us a, maybe we just do a little bit in the past are what are a couple things that were nice to haves in the IVF lab a decade ago that are now must haves that any, that the vast majority of them. Embryologists wouldn't even, you know, want to operate in the IVF lab if they didn't have these things.
What are a couple things that were, were nice to have just a few years ago that are now must haves?
[00:04:27] Dr. Jacques Cohen: Yeah, well, it all depends. Well, first of all, that depends, and it's a very good question, but it depends on, on where you are in the world. The philosophy, let's say in Japan, where there's a lot of IVF and a lot of programs, and they're very advanced.
It, it, it very much depends where you are. So, in Japan, they would focus on, on strictly single MBO transfer, nothing else is allowed. They would focus on minimal stimulation, which is done in this country, but only in a few laboratories and a few clinics. So it very much depends where you are. In the U. S., I think in the last 10 years.
The technologies have been kind of the same of the years before. It's always hard to give, to have a hard cut, right? Say it's 10 years, it's a 12 years, 15 years, but in, in that ballpark, I see, I think the most important things to do nowadays are, are, are fitifying at the blaster stage, incredibly successful that took, you know, honestly, that took from the early 80s, the first paper on, on, on, on, uh, Embryo fritification in an animal, in a mammal, that, that, that was, that was published in 1985.
And the results, frankly, were intriguing because nobody had thought it could freeze that fast, but the results weren't great. And that's why for many years, decades, really, nobody looked at fritification. And it's only in the last 10, 15 years that that's been implemented worldwide. Uh, and, and nowadays, uh, It's considered a must to have, not only for spare embryo freezing, but maybe freezing all the embryos.
Because one thing that is obvious and has become obvious slowly over time is that the cycles where the stimulation occurs are good for the ovaries and you get multiple eggs. Well, it's not good for the uterus and, or it's not optimal for the uterus. I should say, because of course there are a lot of fresh embryos that have never been frozen and are being transferred for like the stimulation cycles that just implant.
So that's one area. The other area that is very much now driven in, in, in. in IVF in the United States is of course PGTA, pre imaging genetic testing for aneuploidy. That has had seen a slow process as well. I think we're now close to 50 percent of all cycles where PGTA is being performed. So, some clinics, it's completely routine, and they do a big case as PGTA.
Other clinics are more careful or more selective, I should say, and do it maybe in a proportion of patients, whereas in some clinics, it may not be done at all, but the average is close to 50 percent in this country. It's very different from the rest of the world. There we kind of stand out, and this has not been happening overnight.
The data is very good. The data that we have gotten over the years is coming very slowly. There has been tremendous debate back and forth. Debate isn't finished yet, particularly internationally, on PGTA. But we see major advantages of this in this country, and particularly because it gives you a higher chance early on in your adventure as a patient having having MBLs transferred because what is striking with the, looking at the data now from, from SART, what is striking the, the, the, is that A lot of patients don't come back after one or two attempts, irrespective of their economic or insurance situation, they just don't come back.
And so you want to, you want to strike it when the iron is hot. You want to get an embryo transfer now, or the embryo is frozen and you get an embryo transfer in a couple of months or next month or three months from now. That is when people are not just motivated, but not exhausted yet, and yet unfortunately A very exhaustive process and most, and most, and most patients experience it like this.
Not everyone does, but most patients do. I think those are two areas where these are now considered must haves. You don't have to do PGTA in each patient. Also it's expensive and it's, it's, it's cumbersome. It's very time consuming for ambiologists and doctors and nurses. And so we want to maybe do it a bit more selective than some clinics do, but I do think it's of the total package.
It's not going to disappear anytime soon. So those two, fitification, PGTA, and, and they go hand in hand. I think PGTA wouldn't have happened on this scale without the success of fitification. They're very much tied in together. So those are two examples.
[00:08:58] Griffin Jones: Your point that what is a nice to have in some areas might be a must to have, must have in other geographic areas and vice versa.
Makes me think of what I've been starting to learn about time-lapse incubation. I'm not a a scientist, I'm not an embryologist, so I don't know enough about the cost benefit. But all I can observe is that for some people, time lapse incubation appears to be an absolute must have for some people. It, it, they would, they would never work in a lab that didn't have TLI.
And there are many countries where TLI is the norm, but in the United States it seems like it hasn't really taken off. So can you tell me why that is?
[00:09:42] Dr. Jacques Cohen: Yeah. Thank you for bringing up TL. I, I probably, uh, I'm, I'm more leaning towards the people who, who couldn't do without it. Not necessarily because I think it improves pregnancy, although I don't see a reason why it shouldn't.
It's nice to leave the MBOs alone for the entire period. Um, you're, you're sitting in, you know, the MBOs are basically in a, it is a robot, it's an incubation robot, and, and they're being photographed every few minutes or. Or every minute and each one at a time, you get a timeless video at the end. What is really, really good about this.
You have a permanent record of that patients and BOS at all times. It also, these incubators have been sought through with so much detail that they kind of are. on the high end side, and they have very, very good results. So, so why it hasn't happened in this country as much as, let's say, some countries in Europe, particularly in Scandinavia, and then England?
I think, I think that is because maybe of the expenditures, and also we are very much data driven in this country, and that's because we have the luxury of looking up our own data. The data of our competitors and clinics in SAR and the CDC, and that is something, don't take it for granted because there is only maybe five, six countries in the world where we have data reporting that is, that's mandated.
And, and, and in most countries, and particularly in Europe, you, you see some data reporting, but it's very, very cursory. And so when we look at those other countries that have data reporting and we compare it, we try to compare it, it's a difficult process because there's so many other factors involved when you analyze data.
But if you try to compare it. I think we're a little better per embryo. I think we're a little better than, than let's say most of the European countries. And I know I'm sticking my head out here and I hope, hope nobody from Europe is watching. But if you are, I think that is the reason why we haven't jumped onto time lapse because all the time lapse, the initial five, six, seven years all came from European countries.
And but I, I think time lapse is, is here to stay. I think this is now the norm. But the reason I didn't mention it is because you set that 10 year limit and time lapse is now 15 years. We've had time lapse for 15 years, hundreds and hundreds of papers. I think it's pretty convincing. Um, things have been discovered we didn't know about before and there's still a long way to go.
So I think time lapse is not going to disappear. Yeah, I think it's the standard to leave the MBOs alone while they're being watched by a machine. It's just a wonderful thing. You don't have to take them back and forth to an incubator. It's, it's, it's, it's an absolute must, but you know, they're expensive and they have to be maintained.
So there's an extra cost as well. I don't know if clinics charge an extra fee for it. I would be, that would be unusual, but maybe, maybe that is the case. I'm not, I don't know enough about that, but yeah, at least it drives up the cost for the clinic as well. Definitely. It's not just the investment.
[00:12:46] Griffin Jones: Is, is the use of PGTA somehow related to adoption of, of time lapse incubators to that other countries don't, or they use time lapse incubators more because they don't use PGTA.
I've heard something like that, but I don't understand, but I don't understand the rationale. Can you explain that?
[00:13:07] Dr. Jacques Cohen: Well, there are a few papers that have suggested that if you look at MBO development using time lapse, not using, using the, the archaic manual systems, if you use time lapse, there is a correlation with euploidy.
is normal chromosome detection and abnormal chromosome detection. It's being debated. There's very few papers about this, but that's one of, you hear people, indeed, you're quite right. You hear people say, well, specialists say, you hear say, well, I do, I do time lapse. I don't need to, I don't need PGTA. I hear that less the other way around, but I hear, hear that, hear you say, hear that.
that occasionally, but I think, I think our reaction in this country of not using time lapse is mostly associated because we have the data to show we have so much detail. There's so much information going inside a CDC that's not published in the national report that you do not see in the individual clinic reporting of SARC, which is fairly extensive, very detailed.
It's not, we don't see that in any country, including, including the UK. But it has been data reporting for less, less time, but data reporting nationally has been happening in 1988. It's quite an, in 1987. I mean, it's, it's, it's unbelievable, 35 years of it. And, and if you compare it to our Southern neighbor, Mexico, where there are a lot of good clinics.
There is no national data reporting. That is the norm for, for 80 or 90 percent of all countries in the world, including the ones that do a lot of IVF, including China, where there probably now is much more IVF than anywhere in the world, and including India. But there's also an enormous ton of patients.
Tons of patients that are being treated there, although the accessibility for the country's population is very, very limited because it's all, it's all out of pocket. So it's still a small population, but because it's so many people, there's a lot of IVF cycles being done. None of that is nationwide reported.
We do not know how well these clinics do.
[00:15:10] Griffin Jones: I want to make sure I understand this relationship between the comprehensiveness of data reporting and time lapse incubation. Is it that other countries where there isn't this national level of reporting where they can see other clinic success rates and the other data points?
Is it, is it they're getting something from time lapse incubators that, They're getting a level of data from time lapse incubators that that they need because they're not getting from a wider pool of data. Or is the United States, because we have a wider pool of data, we're not convinced by the value of time lapse incubators.
I'm, I'm, I want to make sure that I understand the relationship and I don't think that I do.
[00:15:56] Dr. Jacques Cohen: No, no, and I think that maybe I've, I've slightly misled it, misled you, because, because listen, you need to know the data in order to go forward and understand how well, how, how, where you, where you lag or how well you are doing.
You need to have data, data feedback so that you can compare with your colleagues and other clinics. Time, that's data and actual fact. It's not really entered in the SART data reports in the, you know, that, that would overload the system so much because timelapse, as you know, generates an enormous amount of data on, on an M, on the individual MBO level, on the individual oocyte and sperm level, the data that goes into SART and other national reporting sites in other country.
is, is limited or none. So that, that data is very independent from, from the argument I made, which is, you need to know that data. And I think that data has driven this process. In our country, we've just looked at like, look how we are doing. We have a national report. And if we look at that national report, yeah, we are slightly better than other countries.
I only, Do the comparison looking at individual embryos, because if you look at it on a patient level, well, some patients will have two embryos, quite a lot still. Most will have now one embryo, which is what has changed in the last 10 years. But it's difficult then to compare. What you really have to do is compare on each embryo that's being transferred, how many led to live births.
How many implanted, how many led to live births. You're going to get a live birth rate per embryo that's transferred. And if you compare that to other dead populations that are out there, I think we are clearly better corrected for confounders and confounder is a factor that affects the outcomes.
Maternal age is the most important confounder, but there are probably hundreds. My colleague of mine called Rusty Poole from Texas has, has published this years ago and he came with more than 200 co founders and he was probably being modest. There are probably more, in other words, those are all factors that affect the outcome.
So, it is a little difficult to look at another country and say, well, this is what you're They're not doing as well. But if we just look at if they report on maternal age groups, we can make the comparison. And that's what we do. Often counties will only compare patients over 40 and lower than 40. If you look at it per age, 35, 36, 37, 38, they compare all age groups.
You see that drop off in panacea rate and an increase in abnormal chromosomes. and aneuploidy. Highly correlated with each other. That's why we have gone to the PGTA route. We, and also the sync. We think we're syncing PGTA because, yeah, you may be, you may be living in a country. So I'm originally from the Netherlands and in the Netherlands, you get three free treatments for everybody.
It's for everybody. You have three, three treatments. That is three egg retrievals. You can have 10 transfers, all included and free. So you could have 12. Okay. Also there, you see a drop off and how patients returning. It's just, it's just, it's striking that not everybody necessarily the pleats, all the embryos that have been frozen.
It's striking. And that may even be, they may, it may even have PGTA. So they know they have no embryos that look nice, that have normal chromosomes and they do not return. And so, therefore, you need to get, you need to get the first shot is the most important thing. The first and the second, the second attempt are the most important.
Some patients react differently to this. I'm, I'm, I'm not trying to generalize. Some patients say, well, no, I'm going to go for this. I will take a look at every embryo that I have and have that transferred one at a time. And if that doesn't work, I'll have another act of retrieval. But that's, that is not the norm.
[00:19:45] Griffin Jones: Do you think that time lapse will become the norm in the United States, that it will be a must have in the next some years that Embryologists will demand it if they if they've gotten a taste of it elsewhere And they then perceive it as the is the standard or do you think it will continue to be an option?
[00:20:06] Dr. Jacques Cohen: That, that is a hard prediction to make. I think, I unfortunately don't have the data saying, well, how many clinics out of the 400 clinics, how many of those have time lapse and use it all the time?
[00:20:17] Griffin Jones: I'm guessing it's less than 20%, right? And I don't, I don't know. I don't know how much it is, but it's probably, it's maybe 10%, maybe between 10 and 20.
Yeah. That's, that's a, that's a guess.
[00:20:28] Dr. Jacques Cohen: Yeah. But let's not forget that if we would know those numbers, which we don't, if we would know that number and would know how much it is in the Netherlands, right? So how many, how many time lapse clinics are there in the Netherlands? How many are there in the UK? Well, there are frankly, we don't have the numbers.
We think everybody in Europe is using time lapse, I can assure you they don't. And it's the same for them saying, well, every, every, every patient in the United States gets BGTA. They were saying that about us 20 years ago, and it was only a few percent. Right. And now it's just climbing up to 50%. So it's hard.
So once you know those numbers, it's always striking to see that there's not necessarily the norm and that it's just the frequency. Their frequency is probably higher than ours. But I think, I think you have to, now, now we're driven by large clinic networks. And, and, and so they often look at the bottom line and time lapse is more expensive, not just buying an incubator.
The incubator is just one expense. It's just embryology spent more time analyzing the data that comes unless you have a fully automated process and data analysis, which, which, which involve AI, artificial intelligence, and those packages have been approved in Europe or are used. experimentally and they have not been approved in the United States necessarily.
So the European clinics have much newer versions of their software and AI analyses than we do. We, we still have to do it kind of by hand. I think that may be changing and maybe I'm a few months behind and it was approved, but it's very difficult to get. An IVF related AI or any clinical AI that's based on, on, on machine learning, uh, and neural networks.
It's very difficult. to get those, uh, approved by the FDA, simply because the FDA loves algorithms that are stuck. So in other words, it's the same algorithm that's approved, but if you're changing the algorithm because you have AI feedback, well, then you have an intelligence system. And that they, they haven't gone into that very much in that they're, they're, they're worried about it, I guess.
I, it's hard to tell, but I think that they're worried about it. So. The Europeans have that advantage over us. They have more updated time lapse software than we do. So that is a big difference.
[00:22:44] Sponsor: In IVF, the sperm, embryo, and endometrium are all evaluated to set and optimize treatment plans. Yet we remain in the dark about oocyte quality, a critical factor in successful outcomes.
To date, IVF There has been no standard of care for assessing oocytes. At best, general age based estimates are provided, which incorrectly assume that all patients of the same age have the same quality. This creates so much untapped potential for clinicians to better manage patient expectations for success, while creating treatment plans tailored for each patient's unique oocyte quality.
Future fertility has addressed this gap in care, enabling your clinic to drive a superior patient experience while optimizing IVF outcomes. Future Fertility developed the world's first clinically validated AI powered Oocyte assessment. Magenta, made for IVF ICSI patients, integrates seamlessly with any lab, with or without time lapse, to deliver objective and personalized Oocyte quality insights.
Magenta leverages the industry's largest oocyte dataset to provide oocyte quality scores that directly correlate with blastocyst development and blastocyst quality. These unprecedented insights are a game changer in helping you uncover the root cause for unsuccessful cycles, enabling you to plan more precise strategies for future cycles while delivering a superior data driven patient experience.
Let Magenta become your partner in fertility planning, transforming the way your clinic approaches treatment and counseling for IVF ICSI patients. Visit their website today at futurefertility.com/ivf to receive your free sample report plus four key patient scenarios where Magenta Insights can transform your treatment planning.
That's futurefertility.com/ivf.
[00:24:56] Griffin Jones: Is it that European regulators have accounted for the machine learning feedback in algorithms and they, they have a different criteria for, for algorithms where the FDA? prefers one set algorithm or, or what, uh, or already set algorithms as opposed, as opposed to adjustable algorithms.
Is there a difference in, in how the European regulators look at it?
[00:25:22] Dr. Jacques Cohen: That, that I don't know. I presume there is. There's one or two countries, there may be more that have said, well, let's see, let's wait and see about this and accept it as it is. I see the UK is one of those, whereas in the U S it's no. No, no, we're interested and we will evaluate this, but we need to know more about it.
So I think the FDA is more conservative than the European regulators, but I don't really have numbers and it may be different from country to country, but I know some countries have said, well, let's, you know, let's look at, let's not panic, which is what we do about AI a lot. We panic and we say, well, this is going to take over from us, so it's going to do.
funny stuff. You know, you don't get checked GPT, you know, but it hallucinates, as they say, well, AI packages in healthcare do that too. And I think, I think you can avoid that because you can have a very solid mechanism that, that catches that. And those AIs are not comparable to the large language model, but I think the Europeans have a more, a little bit more open minded about AI than we are.
[00:26:23] Griffin Jones: As we look into the research pipeline, does it make sense to talk about robotics first or to talk about AI first?
[00:26:31] Dr. Jacques Cohen: They go hand in hand. Yeah, they go hand in hand. So, so the, the, the time, some of the time lapse incubators, this doesn't apply to all time lapse incubators, but, but some of them have, have, it is a robotic system, right?
You put a little dish inside an opening and then that dish is taken away and goes inside incubator and it's photographed all the time. And AI is applied to it, at least to one or two of them. Or two types of incubators. And, and, and you got basically a reading tells you, well, we recommend that you transfer the following or freeze these four embryos and recommend that you can choose one of this is the best one.
According to us, but it doesn't mean you have to do that, of course, but that, that level is, is very different from what we have and where we are. I hope that answers your question.
[00:27:19] Griffin Jones: So it, well, they, they go hand in hand and, and whenever there's a chicken and egg question, I remember that I remember a quote that David Sable told me like two or three years ago.
He said the entrepreneur's job is to solve the chicken and the egg. And so our, right now is our, where are we further ahead in your view? Are we further ahead in the development of the robotics or are we further ahead in the development of the AI? Bye.
[00:27:48] Dr. Jacques Cohen: Okay. So just, yeah, no, very nice. It's a very good point.
We, I see, I think, uh, I think there's a lot of effort in AI and also now a lot of effort in robotics. Robotics started earlier. The first papers are from 2007, 2008. Uh, mostly coming from Montreal from new songs, a program professor, you saw at Toronto university, you know, at, at university of Montreal and, and, and he and his group have been building this up for the last 15 years slowly and more and more interest now, at least five, six efforts in companies that have started up in the last few years, starting different aspects of IVF or studying the entire IVF process and see if all of it can be automated or maybe should you just focus at one particular aspect.
AI is, has had a tremendous interest because robotics. It's on a different economic level, robotics is relatively expensive, whereas AI is very doable. And you can, you can develop nice AI packages for relatively limited amounts of money. And so there are a lot of AI companies, uh, David Sable, you just mentioned him, he and his colleagues calculate that the.
There was a few months ago, there were 35. Uh, I wouldn't be surprised if we're over 40 and also wouldn't be surprised that the researchers missed several of them because they are basically, basically not noticeable. These could be companies or, or clinics or university groups that are not noticeable because they haven't published yet or they haven't really been loud enough as a matter of speaking and they're being loud enough that you know about them.
There could be as many as 50. So that is an explosion and most of them focus on MBO selection. Well, I can tell you that, that, that it's going to end up in the typical civic and valley frequency, 95 percent of them will fail. Maybe higher, but this cannot be sustained, particularly because they're competing with established methods.
The established methods are time lapse, morphology analysis, development rate. If you don't lose time lapse, there are other methods and also PGTA. So you're going, going up against 40 years of IVF and to see that replaced overnight, it's just not going to happen easily. But what is nice about an AI is you can just ignore it.
I think AIs need to be either for free or affordable. It's very affordable. So if PCTA means that you charge, let's say, 100 per MVO, an AI should be a few dollars per MVO. Also free. That's, that's my opinion. Because all of this is just somebody's opinion. An intelligence system that's looked at a lot of data and has come to conclusions that are maybe holding up or not.
We don't, we don't have enough papers yet that it is really making much of a difference. I've been involved in one effort and that seems very interesting. But. That's just one and a couple more that have been published, but you know, the advantage of AI is it's simply to, it has to be simple in terms of installing.
You just download a program and you get an interface to work or your, or your system in the lab and it shouldn't involve hardware. So it's very easy. It's just like an app on a phone, but you need input of data, morphological data. You need Maybe photographic or video data input, but if it's time lapse, then you need that kind of input, more, more hard, more computer hardware needed.
If it's just still pictures, it's very easy to do, and you really literally could do it on your phone. And then you get an opinion, and if you don't like that opinion, as an embryologist or a doctor, It's usually nowadays that, that's a, that's a team decision which I'm able to transfer or which I'm able to solve first.
Then you, then you, you can use an AI. If you don't like it, you got another AI. And if you like them both for different reasons, and you have tested two, why not use them both? Then you have two opinions. That's all it is. It's an opinion and it's an assistant. It's not, this is not what you should necessarily be doing.
You, you're, you're, you're. You're the end point of hundreds of millions of years of evolution. AIs cannot compete with you, but they can do one particular thing, one particular thing very well, particularly if they are based on experience of other clinics and you may be in a small clinic and you could use that experience and that, that, so it is democratizing in a way the solutions that you're building in the IVF lab by making use of an AI or it is an assistant.
So that's. A very big, big difference with robotics, where you have to develop, you have to imitate what an ambiologist is doing, what the lab technician is doing. The lab technicians that we, that, that are nowadays, they have been trained for years and experience counts, and you can just see that in a lab.
You don't see many publications showing. results of embryologist expressed of how they perform in the, in the lab, but it can assure you there are differences. And of course, as a lab manager, a lab director, you try to minimize that, but it's, it's, it's, it's, it's amazing that the robot is basically being put in place, replacing that kind of experience.
I personally think it's doable and I think it's going to happen. The timeline, I'm not sure about. Some of these applications could be a couple of years away. Others may take longer. It's hard to say. I hear all sorts of numbers out there. Some people think it will take a generation or two. Others are saying it's going to be a few years.
The truth, the truth we'll find out later. This is going to happen. There are some procedures that embryologists either don't like doing. Or maybe not so good at, you know, you get tired in the lab. When you do a lot of procedures, you get tired and you have good days and bad days. A robot, if it's well developed and well tested, it doesn't have good base, good days and bad days.
It doesn't bring out what it's experiencing at home into the lab. It doesn't look at this phone and pick up the phone or text. It's not distracted. It's, it's, it's an idiotic system that's very, very focused on one particular task and, and that, and that's how you use it. And then it can be very, very helpful.
We developed the sperm selection AI. I don't see if you, once you have had that, I think that is such a wonderful thing to have. It, it, it actually makes your decision faster and you know, you know, you can use it all the time. But you can ignore it. And that's the beauty of AI, whereas if you have a robot in place, well, you would have to stop the process and go into the robot and take the embryos out, or the eggs out, and interrupt the process.
So a robot has to be very, very well tested before it's implemented on a routine basis. It's a very different process. I, I, I think it, it will literally take hundreds of millions of dollars to develop robotic systems. And it probably, if you add it all up, and once you're done, let's say in five years from now, you add it all up, what all our efforts have been, this could be a billion dollar project, maybe more.
So, so AI, where if you can get data from different clinics, you're in, you're in a good place to develop an AI product that could make, could make a difference. Thanks.
[00:35:04] Griffin Jones: So, when you say that you think that the AI should be either affordable or, or, or very, very low cost or free, do you mean as, uh, as a pass on to the patient?
And do you mean for as long as it is simply as good as an opinion as, uh, an embryologist or a clinician?
[00:35:27] Dr. Jacques Cohen: Yeah, it goes both ways, right? So if it's a, if it's an add on cost to the clinic, it often is passed on to the patient and discounted. I think this is on a level that it shouldn't be. I've always been surprised that if you are able.
at some point in time to make maybe a difference with a new technology in terms of success rate, whether that's higher fertilization, whether it means that you can get more embryos to develop by changing things in your culture system. We don't pass that on to our patients directly. But there's sometimes these, these develops, like PGTA is of course a good example because it's so labor intense and costly.
But there are others, like assisted hatching used to be in the past, and, and clinics would charge a fee for something that takes a few minutes. And I don't think, I, I personally never felt comfortable about that. I think that, that is, that's often a decision of administrators, but the practitioners may not feel comfortable about these things.
So, we need to tinker on the, with the culture system, which is still the major. research line that exists, right? We're talking usually about sexy things like robotics and AI, and gametogenesis, artificial syntax, making synthetic sperm and eggs. I mean, those are the big sexy projects out there. Most of our research is about how can we make things better and safer?
And those are spreading tiny little steps and suggestions in the scientific literature. And that's where we focus most of our energy. It goes back to your earlier question, because that's really, that's really improving the culture system is never going to change. We will always think of Mr. Culture system.
That is a research line. That's incredibly important. Big breakthroughs in the last 40 years in that area. But because You know, if you change the culture medium ingredients and test different culture media against each other, and I've been hundreds of those trials, people don't get overwhelmed by that.
The lay people out there, they don't, they don't see that as something they're necessarily interested in, but that's why we got better. The cultures making changes to the culture systems, why we have gotten better over, over the decades. That will, that will not stop. That's not going to stop. That's going to continue.
[00:37:44] Griffin Jones: Will the AI not get to a point where it's better than an opinion, where it's better than the average opinion of the average embryologist and average clinician? Will we not get to a point where the AI has the closest to certainty?
[00:38:00] Dr. Jacques Cohen: Well, that's a loaded question. It all depends on, on what your end point is.
If your end point is helping an, uh, an ambriologist setting up instruments and timing themselves, uh, you could develop an AI. We have developed an AI that's tracking the ambriologist. And I think there, you're probably going to say at some point, well, this is your guide. It's basically somebody who's keeping the books, right?
It's telling you, well, that those tools are, this tool is not looking good. Get another tool. You know, you need to position this differently. Oh, well, one second. You don't see there's a hole in this zone of Pellucida, you know, their AIs can actually take over and, and, or take over, help you to, in such an extent, you're going to ignore it.
Definitely. The decision AIs, those that are not observing and just helping you, but are making decisions, not necessarily for you, but making decisions like this is the best sperm. This is the best ag. This is the best MBO in their opinion. That's an opinion. Is that going to be equivalent to what you would come up at some point?
Yes, I think it will be. I think it will not only be an equivalent, it will be better than what we have come up with. But this is a development. Is that going to be a year from now? I'll be very surprised. Five or 10 years from now. It's going to be, it's going to be there. And look how long it take, took to get PGTA somewhat accepted in this country.
It took 15 years, maybe longer. With AI, it's going to be in the same timeline. So for every, every clinician, every embryologist to be, to accept that technology will take a long time. Uh, but I have little doubt that it's going to be at least as good as what we do, if not better.
[00:39:46] Griffin Jones: Are you using it right now in the IVF lab, or do you use it to grade cases?
[00:39:51] Dr. Jacques Cohen: Yeah, I'm not running an IVF lab anymore since, since at least a year. But when you're consulting? Yeah, definitely. Yeah, I definitely, I definitely suggest it. There are AIs you can get for free or for very little. There are some that are charging hundreds of dollars per MBO. I don't understand that. It's a changing algorithm.
And I, I don't understand why it has to be that expensive, certainly wouldn't have cost that much to develop. So, so I think, I think should be for very little or for free. And I, I am consulting people say, well, these should get for you for very little or for free. And you could use several of them. That's, that's my advice.
Don't, don't use one MBO selection AI, but use several. If it's, if it's reasonably priced or for free, then that's what you should do. And you got, you got, you got, and then you can basically keep track of that data. See what you thought as an embryologist, for instance, what did you think should be transferred?
What did the two AIs think? And then you can get some analysis later on. You've done a thousand of those after a year or two years and then analyze that data. See if it has worked for you. Are you just kicking AI out, right? It's just turning over an app, just turning over an app. It's not a big deal. I think, I think a lot of it should be for free.
[00:41:05] Griffin Jones: With regard to robotics, you said that this will end up being a hundreds of millions of dollars, possibly a billion dollar project to fully automate the IVF lab. How far into that billion dollar project are we?
[00:41:22] Dr. Jacques Cohen: I think we're over 100 million, but they're probably between 100 and 200 million right now. I mean, if you just look at Overture, that's already 150 million, I think, so, so we're probably at a quarter, quarter, quarter billion or 300 million in that ballpark, and I really don't have figures.
You know, that's the amazing thing, really, really hard to find out, but we're already probably 300, 400 million up there. I'm changing the numbers as I speak, but, but, but it's, it's a, it's a guess. I think within a few years we'll be at a billion. That, that's, so that includes all the companies. That doesn't mean that the, that one company that is serious about robotics is spending hundreds of millions of dollars, that you could actually focus into robotics.
And if you only are interested, let's say in, in finding eggs during egg retrieval to automate that process, you're probably looking at procedures that probably could be quite inexpensive to apply. But if you're looking at a fully automated robotic. Existation, which doesn't exist yet, or at least has not been published.
There you're looking at a massive amount of AIs, and you're looking at very intricate, very, very subtle and tested robotics and automation. There you're probably looking at a relatively expensive instrument to develop. So that will cost you many, many millions but yeah, if you look at the total effort, really a billion is not so you know, I'm being pushed back all the time when I say this, but is it really if you're already up to 300 million now, by the end before it's fully automated, which I think will take a while, fully automated will take a while.
Easy to predict it will be.
[00:42:59] Griffin Jones: And so within that system, what pieces have we established in the last two to four years? And what pieces are still missing?
[00:43:12] Dr. Jacques Cohen: Okay. So what we have established are preliminary data in most procedures, except for one. And that is tomorrow, the tomorrow system based in New York City.
I'm on the advisory board and I've been associated with them since the early days in 2018. So they have developed two robots that will label MBOs or their little devices that they're held in during the verification process and cryo store all the samples. So cryo storage, which was which has been notorious in terms of mishaps over time.
These, these refrigerators, we call them dealers. These refrigerators can fail as all machines can. And so they, they are under a very harsh and then a very harsh environment and they will fail at some point, but it could take 20, 25 minutes before, before these fail. When that happens, it's a disaster. Also, what happens a lot, it's a lot of errors being made.
Because there are all sorts of good reasons for that. Almost all labs will have errors, at least in communication or errors in, in, uh, in the data processing of individual embryos and eggs. And so it's very common. So we want a more secure method. And RFID chips, which is of course an electronic way of labeling, Each MBO separately.
That, that had to be introduced. And it has been done, and TAMUA uses that technology. And then takes, takes a tube filled with a device that has the MBO stuck to it, that's already fittified, keeps it cold. And then sticks it in a pre programmed place. The advantage for the clinic is that they have immediately a log.
If you tell ambiologists, let's audit our, our units. Doers. Let's order a cryo storage lab. It could be 60 doers. There could be thousands and thousands of patients, MBOs in there. Everybody looks for the exit. All the ambiologists are looking for the exit because it's so much work.
[00:45:09] Griffin Jones: Yeah.
[00:45:10] Dr. Jacques Cohen: And you're going to find things you don't like.
And so. Here and all that is, literally, you take your, you take your, your phone, you take your phone, you click on it, you have done your audit. It doesn't matter if there are 10, 000 MBOs there or a million MBOs. It will be a second thing after audit. You know exactly where they are and that they are still there.
That system has been put together by tomorrow. That is a robot that is in place and that's available now. There is, there are two other robots that have been developed. for our field, except for time lapse, which of course is a robotic system. Two other robots which have to do with part of the fitification procedure.
Fitification consists of four or five parts, and one of those has been been available already from Overture in Spain and Genia in Australia. The Genia one is at least 10 years old, but because it only does one in four, of the aspects of the procedure and biologists, including me, frankly, have never been interested.
Why would you have a robot where you do the other three procedures and the robot does the fourth part? I want one that hears the dish, frees this, and I then want the frozen embryos to come out and go in something like a tumoral system also automatically. So I don't have to be worried about it, and I get the data in my EMR.
That's really what I, what I want as an embryologist, because that'd be very, very helpful. Fidification is one of these things where experienced embryologists get very, very good at it. But it's, it takes a while to teach somebody to understand all the little details. details of it, and really start being excellent about it.
And so there, robotics would make a major difference. And Xe would make a major difference in things like egg finding, sperm prep, all of those procedures, yeah, so it would make a difference.
[00:47:07] Griffin Jones: Are there people working on each of those areas right now? Automating AXE, AXE automating, egg freezing, is that, are we in sort of a race to see which company develops that first?
Or is that in very preliminary stages?
[00:47:24] Dr. Jacques Cohen: It's right now, if I'm to guess and going by the literature, which is maybe only a couple of dozen papers, it's in preliminary stages, but it's getting closer. I think we'll see entirely a series of robotic systems being published in the next year or two, the first stages of that, before robotics becomes really implemented on a routine basis.
also involving the regulatory aspects that are sometimes needed for that, depending on the situation, depending on the type of robotics. I think you'll see, you'll see that that will take, of course, always a lot longer before something comes in team, but within the next months or years, you're going to, you're going to get papers where people are planning.
I can do X finding, not find all that. And I should find out maybe finding more X than I thought that worked. So, so that those things are going to happen probably sooner than, than. And then later, because there are quite a few efforts worldwide. I mean, I said five or six early on. It may be more than that.
Maybe a lot of, a lot of things, but the, so there are two, there are two types of robotics initiatives, companies that are looking at every aspect. And then there are companies, uh, looking at particular application. I don't know what's the better, best approach, but that's, that's, that seems to be what's, what's going on right now.
[00:48:40] Griffin Jones: How about with regard to non invasive genetic testing, non invasive biopsying of the embryo? And I have to give credit to your colleague, Cynthia Hudson, for planting this idea in my mind, because after her interview, she said, shoot, I wish I Thought and talked more about that and, and so she gets credit for, for putting the, the idea in my mind to ask the question, but how, how close do we are, are, are, are, are there preliminary papers about that or, or are we really far away?
[00:49:12] Dr. Jacques Cohen: Now, I say about preliminary paper, Stephanie, um, what are two approaches if you have an AI that selects embryos based on the development of the embryo and, and use machine vision to analyze embryos, that is kind of non invasive, right? That's non invasive embryo selection. And, and that could be trained on, on, on whether embryos are genetically normal or not.
I mean, I've been involved in an effort, it's a company called IVF 2. 0 based in Mexico. We've developed an AI called Erica and Erica was trained, really only trained on embryos. On a lot of MBLs, looking at whether they were normally, whether they had normal chromosomes or abnormal chromosome counts. So whether they were euploid or aneuploid.
Yes. The data was also provided there, which one of those MVOs would make a pregnancy or not, and which one miscarries or not. But the basic training set was euploidy versus aneuploidy. And so that is a non invasive way of doing PGTA, but probably Cynthia was hinting not at that, but at taking a sample from the culture medium.
Where the blastocyst has been, provided the blastocyst was by itself. And then, and then analyzing that chemically or maybe taking a sample of the fluid that's inside the blastocyst, as you know, the blastocyst is fluid filled and the cells are on the outside. Taking a sample from that. Both of those approaches have been done.
Interesting data. But for me, the most interesting paper is if you find DNA there that comes from ambiose, you could wonder, well, why, why is that? Why did that DNA come there? And the group in Bologna in on the, on the Luca Girodi's leadership in Bologna, Italy, they have found recently and published that if you find DNA, The culture fluid that the chances of fantasy of dose ambose is actually significantly lower than the embryos that do not have DNA in their culture media.
And so embryonic, DNA in the culture media, so that tells you, you may be finding DNA and that may help you what anomaly you're gonna find, but it also means what the DNA is there, that's already not a good sign. The advantage of this finding is that you could just test for DNA and that's very affordable.
Just looking at DNA. Rather than getting information back, you have to confirm it has to be embryonic DNA. Once you confirm that, that's all you need to know. If that's there, that embryo probably should be chosen not up front compared to embryos where you could not find the DNA, the embryonic DNA. So, because why would they lose cells?
Well, that means something's going wrong in that embryo. That means that cells die. or lice, and all the, all the content comes out, including the chromosomes and the DNA. That's why they end up in the medium. It's probably not the best sign that it's there. So in my, in my opinion, that kind of non invasive DNA assessment, chromosome assessment, if you like, has a future.
Particularly if you can just, in an easy way, sample the culture medium, say, in a 15 minute test, there's embryonic DNA there, yes or no, and that has a future. To get details of that embryonic DNA, I think that is far, far short. I would, I would go with the AIs looking for embryo selection based on just data and morphology, PGTA data, and, and choose those AIs, and, and already have a dozen of them.
I would look for those. the answers that those have to offer. That's also non invasive PGTA. So whether it's a very good point, non invasive PGTA, getting rid of biopsy is something that we need to try. We really need to focus on that because biopsy is difficult. It's difficult and it's expensive.
[00:53:09] Griffin Jones: So, it sounds like getting rid of biopsying is on the preliminary end of the spectrum, on the very preliminary end of the spectrum, whereas it sounds like something more like the robotic labeling of embryos and the cryo storage inventory of tomorrow is on the mature side of the, of the spectrum.
What's in the middle right now?
[00:53:32] Dr. Jacques Cohen: Well, I think the efforts on ICSI, one of them has been published, the Overture Group in Spain, and MBA Tools Lab in Barcelona. They looked, there was an editorial with that paper, they looked and the editor calculated how many of the steps were actually automated. It was a modest number, but nevertheless, that's never been done before and had, had fantasies.
So this was published. Just a few months ago or half a year ago. And I think that that tells you that there is a lot of work done in that area. There's work done on all aspects. I think on the fertification side, I think there's work done to complete those procedures, not look at one part of the procedure, but the entire set of procedures, and it's the same of all aspects.
So we have done, the field has accomplished making culture, the culture system. Robotic. It has accomplished making the acquired storage systems through tomorrow robotic and, and, and it's, it's, it's looking, it's obviously looking at all the other aspects, which means sperm prep, automation, sperm prep, and, and that, that, that's going forward in strides or making it at least so simple that only involves one or two activities by embryologists on andrology donations.
At finding in the laboratory, there's of course an egg finding or egg retrieval. There's two. There's two efforts going on. It's the surgeon, it's the gynecologist or the, the IE extracting follicular fluid. And then the, that follicular fluid goes through the lab and the embryologist looks through the follicular fluid in very shallow layers, so they decant it into battery dishes and look very quickly for acts.
And sometimes those are hiding, sometimes they sit in blood clots. So it's a bit of an art. It needs to be done in, in a, in a. In a timely fashion, you can't take hours, you need to do this in minutes. So that can be automated, the laboratory part can be automated. I stay away from the clinical part, I think in true course that can be automated too.
But the laboratory part can be automated and you'll probably see the first data sets coming out in the next year.
[00:55:36] Griffin Jones: How about the systems being developed by Conceivable for automating the IVF lab, where does that fall in the, in the spectrum of preliminary to mature? for listening.
[00:55:46] Dr. Jacques Cohen: Okay, so Consiglio Rouvas is now 12 months old.
I'm the Chief Scientific Officer. So we are looking at trying to automate all aspects of the IVF procedure. And there's at least one other group out there that's trying to do the same. So we're looking at egg retrieval, sperm preparation, we're looking at, you know, Denudation, the process where you strip cumulus cells away from the eggs before they go to eggsheep.
Automation, the full automation of the entire eggsheep process we're looking at, we're looking at full fertification. We're also looking at automation in, in the embryo culture system because we feel that the culture systems are very expensive. So we want to come up It's culture systems and timelines that are much more affordable.
So we're working on that and we're, we're working on full certification with the tomorrow system at the end to cryo store the MVO. So it is, it is the idea is to do all of these processes and then string them together.
[00:56:49] Griffin Jones: I feel like I've gotten a really good look into the pipeline today, and you've also made a few points to me that really educated me on why time lapse hasn't been adopted to the level that it has in other countries, why the FDA has not approved it.
Uh, AI algorithms. And so I want to give you the concluding floor. How would you like to conclude about the, the research and development pipeline in the IVF lab?
[00:57:22] Dr. Jacques Cohen: Well, I think, I think overriding what we do in the United States is the fact that funding is so difficult to come by. The largest funding agency in the world is NIH.
And they found, found, what is it, 60, 60, 70 billion in healthcare research. Thank you very much. Why don't they fund the IVF lab? Since IVF started in the early 1980s in this country, I think the first lab is from 1981, the Norfolk lab. There has not been a single. experiment. A single observational set has been funded by NIH.
There's a moratorium on embryo research since the, since the late 1970s, since 1979. That's now by law. So there cannot be any public money spent on human embryo research. It's outrageous because Everything we try to do in IVF, we have to actually go and do this on patients and spend private monies rather than public money.
And so, yes, we can study IVF and do IVF related research in animal systems. But at the end, if you look, for instance, at chromosomal anomalies, there's no animal system that's helpful. You have to find out in the human. So embryo biopsy studies have been very slow to come by. It's because there's not been any.
NIH funding available. I think we have to frame it like that. People saying, well, this is going very slowly. There is progress each year. If you, if you look at the data analysis and, and two of my colleagues, Alex Bissignano and Mina Alikani and I published a paper in 2012, where we looked at fine combed, uh, um, SAR data and found that there is progression in outcomes per MBO of, of 0.
9 percent a year, year by year. That was only a 10 year analysis and it includes all the clinic, but I've, I've looked since then. And it's going up by 0. 9 percent a year. It's more profound in young patients. In patients younger than 35, it's one and a half percent per year, but it is going up. So in other words, we're doing a lot of good things, but it is very, very slow.
Do we have the patience? Do we have the patience to go, to go this slowly? to where it becomes as good as dentistry, right? We go to a dentist, and you don't go to a dentist hoping that your wood canal is going to work or not. You go to a dentist and expect it to be 100 percent successful. Maybe you got a little infection, but that can be treated.
But you want it to be 100 percent successful. And that's what IVF. We want things to be 100 percent successful. Not 98%, not 80%, or what it is now in some clinics over 60%. No, we want it to be 100%. And we really want that as soon as possible. So I think all this technology that we discussed today will, will, will help us.
Play a role in that process, but it's not only technology driven. It's not only technology driven that we go up by 0. 9 percent per MBO in this country each year in terms of implantation and life births. It's not just technology. It's also communication. So what you do, communicating to the community. Uh, conferences, other webinars.
Training is very important in this country. There has always been a lot of emphasis in training. Our doctors s are trained, that's very unique in the world. In, in other countries is usually usually an OB GYN or, or a GP that becomes an IBF specialist. And could they become good at it? Oh yeah, they could become really good at it, but it's a little bit more tedious to do it that way.
And so. So, I think training also of embryologists has changed a lot over the last 20 years. All of those factors, particularly the communication and the awareness, creating the awareness of all this and having a discussion and comparing our data and comparing our methodology, that is making as much a difference to just saying what's all driven by new technology.
It's not just new technology. But the new technology could be introduced a lot faster if we had NIH funding. And we don't.
[01:01:33] Griffin Jones: Dr. Jacques Cohen, I look forward to having you back on to look at the updates on the research and development pipeline in the IVF lab. I enjoyed this conversation today. Thank you for coming on the Inside Reproductive Health podcast.
[01:01:46] Dr. Jacques Cohen: It was a pleasure. Pleasure. Thank you, Griffin.
[01:01:50] Sponsor: This episode was brought to you by Future Fertility, the global leaders in AI powered oocyte quality assessment. Discover the power of magenta reports by Future Fertility. These AI driven reports provide personalized oocyte quality insights to improve treatment planning and counseling for IVF ICSI patients.
Magenta can help you to better identify the root cause of failed cycles and counsel patients on next steps to optimize treatment. Download a sample Magenta report plus four key counseling scenarios and see the difference it makes in patient care. Visit futurefertility.com/ivf. That's futurefertility.com/ivf.
Announcer: Today's advertiser helped make the production and delivery of this episode possible for free to you. But the themes expressed by the guests do not necessarily reflect the views of Inside Reproductive Health nor of the advertiser. The advertiser does not have editorial control over the content of this episode and the guest's appearance is not an endorsement of the advertiser.
This has been another episode of the Inside Reproductive Health Podcast. Tune in for a new lineup of episodes premiering in June, where we'll be taking a tour of the C suite with a powerful new series featuring CEOs from some of the largest fertility networks in the world. We can't wait to share these inspiring conversations with you.
Until then, stay informed of the latest fertility news with our weekly digest, delivering curated content straight to your inbox every Thursday. Stay tuned for more updates and thank you for listening to Inside Reproductive Health.