Session 21 - PGTA: Q&A Triage
Questions for Dr. Gary Harton
- Shaista S: It seems that some software are more prone to make mosaic calls. Do you think this increasing specificity in detecting mosaicism has a benefit? Can you describe the differences in technology that cause these variabilities in mosaicism detection?
- Hosam Zaki: Does WGA contribute to mosaicism?
- Gary: Maybe? I know that isn’t a really good answer but maybe. Certainly there are groups that don’t use WGA but still see mosaicism and we certainly saw mosaicism when using FISH. Mosaicism clearly exists, one only needs to look at data from miscarriage and prenatal (CVS) to see mosaicism is real. The actual question here is what does mosaicism mean for the early embryo. The secondary question is how much of what we do (patient selection, stimulation, embryology, transport, genetics lab, etc.) impacts mosaicism. Remember also I alluded to the bioinformatics that go into PGT today...this probably impacts mosaicism calling as well. If you think about the STAR trial, there was data created that showed wildly different mosaicism rates between genetics labs, which tells me that this is an issue on the bioinformatics/calling side of the equation, what one lab calls mosaic another lab calls euploid or aneuploid. We need to harmonize around these issues and find the best path forward together!
- Mary Ansong Is there a potential of NGS in PGT-M and how do we make it inexpensive?
- Gary: Yes! There are a few ways to do this however, cost will almost always be an issue.
- Labs can begin the WGA process as normal and then split the WGA into two different reactions, one for PGT-A and one for PGT-M where they spike primers for the mutation(s) and linked markers into the WGA reaction. This would then normally lead to some more PCR outside of WGA and then the sample can be diagnosed in any number of ways, including NGS. The issue here is that you are always ‘robbing Peter to pay Paul’. No matter what you do you always run the risk of harming one result or the other (or both) because of the splitting of the sample.
- There is a new product on the market (from my company so I won’t say the name as I’m not here to do a commercial for my company) that has a patented approach to this allowing you to spike specific single gene primers and linked markers into the initial WGA without a split of the samples. Once WGA is completed you take a small aliquot out, perform a secondary PCR for the PGT-M portion of the test, spike this product back into your original WGA material and run everything on one sequencing run.
- No matter which method you choose, however, there is still extra cost, for the extra PCR primers, and PCR reagents, testing of the primers for the patients and their family, genetic counseling etc. Unless and until there are commercial ‘kits’ available for specific diseases made for PGT-M this will continue to be a clinical research and development project for each new patient.
- Brett Reggio How often is the standard "reference" human genome refined? In other words, are we comparing all DNA fragments against a standard that is outdated?
- Gary: Good question. There is no ‘schedule’ but it does get updated a few times per year. If you look closely at any report with NGS it will (should) state clearly which Human Genome Build was used to analyze the data. Interestingly, as we continue to learn more about the human genome, we can go back and reanalyze data from earlier tests against the newest build and find more detail about some areas of the genome.
- khaleel ibrahim Thank you very much for the beautiful lecture,, I would like to ask about the accuracy of NGS and which method or company you recommend
- Gary: Well thank you! I think each system has pros and cons and you and your team need to understand those pros and cons to make the decision that is best for you. A number of platforms have been shown to have very high accuracy in counting chromosomes so when choosing a lab to work with it usually comes down to other services, customer service, patient care etc. There have been a number of papers that look at the accuracy of various systems including SNP array, aCGH/NGS, targeted PCR with NGS and each of them typically comes to an accuracy of around 95 %-97 % for whole chromosome gains and losses, and less for mosaicism and/or segmentals. This data is based on rebiopsy and/or non-selection studies.
- Fernando J Prados Mondéjar Is it possible to calculate a significant mosaicism % taking a sample of four or five cells that are very close together (a clone) out of the 100 cells of the trophectoderm?
- Gary: I don’t think so. Remember that the genetics lab is reporting on the results of the SAMPLE THEY RECEIVED, not the whole embryo. If you gave me the whole embryo I could be a lot more accurate with the result and the % of mosaicism. As mentioned during my talk and above, there is a lot of bioinformatics that goes into PGT-A these days so the result we give you is a representation of the sample you gave us after WGA, library preparation, sequencing and a lot of bioinformatics. As mentioned by Dawn Kelk and Dave McCulloh during their talks, where you choose to biopsy may change your ultimate outcome based on which cells you happen to take. I think we need to start thinking about the PGT-A result in a different way so that it helps guide transfer decisions, not as a diagnostic of the embryo. Clearly some mosaic embryos are euploid and make normal babies, this has been published, however a recent case report by a group in Turkey shows that a low level mosaic can also become a live-born mosaic child! We have a lot to learn...
- Dr. Vikash Verma Would Gary know of False Positives and False Negatives Rate in NGS technique such as for Illumina platform? Thank you.
- Gary: I don’t think this comes down to the system but it comes down to the lab using the system. Each laboratory should have done some sort of validation of their platform before they start offering the service clinically. If you have this question you should ask your genetics laboratory for their validation data which should, in my opinion, be shared willingly. My company is currently putting together a validation package for new PGT labs or labs switching over to our platform and this includes putting together the data you ask for as seen below:
With the results and by comparison with the reference data the following table will be calculated
False Positive (FP)
# false pos./total
False Negative (FN)
# false neg./total
No Call Rate
# no result/# samples
Sensitivity (per embryo/cell, TN)
# true pos./(true pos. + false neg.)
Sensitivity (per chromosome)
# true pos./(true pos. + false neg.)
Specificity (per embryo/cell, TN)
# true neg./(true neg. + false pos.)
Specificity (per chromosome)
# true neg./(true neg. + false pos.)
Evaluation criteria will be based on the Percentage of Agreement (PA) between results obtained at the customer site and results provided for the samples included in the kit.
PA (%) = 100 x [TP +TN/ (TP+TN+FP+ FN)]
Questions for Dr. Peter Nagy
- Shaista S: If cell cycle checkpoints are not working ideally in human cells, then mitotic errors would be quite common in embryos, leading to higher mosaicism. Do you think mosaicism is a more common occurrence in embryos than we have previously envisioned? Or is it more common in the TE vs . ICM? What is your opinion on transferring mosaic embryos?
- Peter: Hi Shaista, Thank you for your question. - Yes, - I do think that mitotic errors in human embryos are more common than we realize (that also leads to mosaicism more frequently). I don't know if we have enough evidence to establish if mosaicism is more frequent in TE or in ICM, “traditional belief” was that TE might be more prone for mosaicism. If a patient requests, I would be agreeable to transfer most mosaic embryos.
- Jhon Troya: any information about the nature of mosaic embryos and autocorrection as limitations for PGT practice?
- Peter:: Hi Jhon, Thank you for your question. - I think that most mosaic embryos are viable, - studies demonstrate that they implant and result in viable pregnancies/offspring. However, they seem to have lower implantation potential than “fully” Euploid embryos (non-mosaic). “Auto-correction” at the cellular level (of the embryo), - I don't think it occurs. But at the embryo level, it certainly can happen. For instance if two cells of an 8-cell embryo have chromosomal abnormality (and the remaining 6 cells are chromosomally “correct”); then it can happen that the two cells with abnormal chromosomes will stop dividing, and only the cells with normal chromosomes will divide and participate in the development of a blastocysts (and cells with the abnormal chromosomes will be excluded). And that “correction” can happen at a later embryonic stage too, so, indeed this situation may be a limitation of the PGT practice, as you suggest.
- Jayaprakash Divakaran What is your opinion about those reported cases of aneuploid embryos transferred and resulted in Normal babies? How did the correction take place and at what stage the correction happened? Any data on the abnormal babies born or miscarried in the PGS and Non PGS group?
- Peter: Hi Jayaprakash. Thank you for your question. “Aneuploid embryos” if they are truly aneuploid (100% each cell aneuploid) will not implant. (That is of course an opinion, because we cannot make a study where we test every cell, because then nothing is left to transfer). So, my opinion is that if an “Aneuploid embryo” after transfer is implanted and results in a “normal baby”, then that embryo was not “fully” Aneuploid, but it was mosaic, and when it was biopsied, then only cells with abnormal chromosomal composition were removed and tested. Regarding correction; - I refer to the answer I wrote above (similar question) - and it can happen at any embryonic stage. Babies born after PGS vs non-PGS group, - I dont have data, but I do have the data on miscarriage (it is possible to calculate it from SART) - and I put it in the below table. As you can see, in the age group of less than 35 and 35-37 the miscarrage rate is not very different, but with increasing age the miscarriage rate is increasing in the non-PGT-A group (and remain steady in the PGT-A group)
PGT-A was performed
35 - 37
38 - 40
41 - 42
Number of cycle starts
PGT-A was NOT performed
35 - 37
38 - 40
41 - 42
Number of cycle starts
- Giles Palmer: The HFEA has called this an “add-on” and placed as a “red light”..your thoughts?
- Peter: Hi Giles. Thank you for your question. The benefits (and possible disadvantages) of embryo biopsy/PGT-A have certainly been disputed for a long time. We do have to recognize that this test (PGT-A) does have some advantages but also drawbacks. As long as a patient clearly understands all these points and how embryo biopsy/PGT-A may affect her IVF treatment, I think, it may be valid to let the patient know about this option.
- Shaista S: Were these samples tested in the same genetics lab? Do you think it would make a difference if they were tested in different labs?
- Peter: Hi Shaista, Thank you for your question. These samples were not tested in the same lab; - but they were tested using the same instruments and similar protocols. Regarding your second question; - I don't think we, as Embryologists, have sufficient experience how different PGT-A laboratories may impact test results (ideally it should not).
Questions for Dr. David McCulloh
- Shaista S: How do we as a community of embryologists tackle quality control amongst embryologists’ handling techniques between the labs
- Do you think day of zona perforation (day 3 vs day 5,6) would have any potential impact on euploidy rates?
- David: Since the analysis presented suggests that there were significant differences in inter-technician euploidy and mosaicism rates when comparing between techs performing zona perforation (for us usually on day 4), my interpretation is that it is quite possible that technicians may be able to affect euploidy and/or mosaicism rates by their handling of embryos at time of zona perforation.
- Kelly: How many cycles do you think an IVF center needs to have to compare technicians variability on procedures?
- David: Probably quite a few. This study was analyzing results of over 4000 cycles. That is a lot. One general rule of statistical analysis is that the more data you get… the more likely you are to find associations between an outcome and one or more independent parameters. The absolute number is difficult to determine; especially since the same euploidy/mosaicims rates are analyzed for each of the steps! The multifactorial nature makes it very difficult to tease things apart.
- Jayaprakash Divakaran Do you have any data on the Aneuploidy of embryos where zona perforation is done just prior to biopsy with those with zona perforation done a day earlier?
- Jose Hernandez David how did you remove the male patient variability at the insemination step to isolate the technician effect? Thanks!
- David: This is really quite preliminary data. Very little scrubbing of the data has been performed. Nothing was done to remove male issues.
- dimitra Christopikou from Embryogenesis Greece !! have you recorded different ploidy or blastulation rates when zona perforation is performed on day 3 day 4 or on day of biopsy
- David: We have not examined this yet.
- Larry Fishel What was the pH of the culture media in the Low O2 environments and in the High O2 environments?
- David: I think that I answered this during the webinar (in writing). We expect little effect of different O2s on pH. (which is to say, I have not checked)
- Shannon Kirkpatrick To be clear, stims with higher use of hMG have higher euploidy rates. Correct?
- David: Yes, up to about 50% hMG. We have few cycles with more hMG than that. It is presented in the paper in Eur. J. Med Genetics.
- hosam zaki is it hMG or LH levels during stimulations?
- David: We obtained data from one center. In the data we received information about FSH and hMG levels administered. Very little of the LH activity in hMG is actually provided by LH. If you examine the area under the (time) curve, the predominant LH activity is provided by hCG in hMG products, since nearly all of these products require exogenous supplementation of LH activity in order to get to “equal activities” of FSH and “LH”
- Frederick Miller David, Have you applied multivariate analysis to your data. Is that a useful approach to figure out what’s going on?
- David: Answered online during the webinar. Answer: not yet but hope to...
- hosam zaki: What in the culture system affects the checking points that you have spoken about?
- David: I can’t answer this - Also it was Dr. Nagy who presented the information about the checkpoints.
- Ahmet Turp: can oocyte retrieval pressure effect euploidy? optimum pressure 100 or 150 mmhg or 120 mmhg?
- David: We did not examine this - however since oocytes are retrieved with a meiotic spindle arrested at MII it does seem logical that this might be a stage susceptible to something like this.
- Fernando J Prados Mondéjar What would happen if the technician with a higher aneuploidy rate performs all the tasks compared with the “best one” doing all tasks?
- David: Seems to me that this is what we would like to try to avoid. If you mean how much of an effect, I am not really sure. I presented the inter-technician variability as a standard deviation. The values varied between about 25% and 50% for both euploidy and mosaicism rates. The typical WITHIN technician variability is about 25%. So the difference between the best and the worst is about an average of 25% of the average euploidy rate or the average mosaicism rate. Our average euploidy rate (of course it varies with patient age) is about 30% euploidy at age 38.4 years. So the +/- 1 standard deviation (worst to best) would be about 23% to 38% average euploidy rates, VERY roughly. Mosaicism (using CooperGenomics which reports higher mosaicism rates than many competitors) averages about 25% so the +/- 1 standard deviation (worst to best) would be about 19% to 31% average mosaicism rates, VERY roughly. HOWEVER, this is such a multifactorial issue that it may be best to pay very little attention to the estimates that I just provided. : )
Questions for Dr. Dawn Kelk
- Yu Sun How does the invasiveness and the integrity of the nucleus of the retrieved TE cells impact biopsy results?
- Shaista Sadruddin: What technique would you recommend for doing a thaw, biopsy, refreeze where the ICM is not clearly visible?
- Fernanda S. Peruzzato Hello! Which day do you think is the best FOR EMBRYO to open ZP before biopsy? (D3, D5 or at biopsy time for blastocysts)
- Do you recommend the laser be checked before each case or only once per day?
- Dawn K: The laser rarely changes in our experience so we test the first case of the day. I have travelled to other labs where the focus on the laser was completely off and they wondered why they were having a difficult time performing biopsies. If you test the zona from a discard oocyte or embryo, you should see a very crisp, infocus obliteration of the zona.
- Kathleen Tucker This is a question for Dawn. Great presentation. I was particularly impressed with the control you showed with your biopsies. Can you tell me what type of manipulators you use? Thank you.
- Dawn K: We use Narishige manipulators with an Eppendorf Cell Tram Vario injector. This injector goes back to the days when we were doing blastomere biopsies and fine control was absolutely essential to keep the cell intact. I think TE biopsy is much more forgiving if you use the right size biopsy pipet for the job. Large biopsy pipettes (30-40µm) don’t allow for the best seal on the cells and when you do cut, the biopsy tends to shoot up the pipet especially when you suction well to pull and stretch while performing the biopsy. Our preference is a pipet (20-25µm) which allows the biopsied cells to plug the opening when you stretch & cut and the biopsy doesn’t shoot up. We did try 15µm pipettes (the size we would use for polar body biopsy) and found they were too small. They suctioned on too small of an area and the cells would break & collapse the blastocyst rather than aspirating gently into the opening of the pipet.
- Estrella Rosemberg Is there something you do in your biopsy that increases the noise of the sample and then the result is that the embryo is classified as abnormal
- Peter: Hi Estrella, Thank you for your question. (This question may be for Dawn, but I can try to answer). I think, how one is performing embryo biopsy is critical in getting a reliable/accurate test result. One possible cause for getting abnormal (or mosaic abnormal) results can be if you include “non-incorporated” cell(s) in your biopsy (when doing blastocyst biopsy). Those tend to be larger, and not be part of the embryo, and likely have abnormal chromosome composition (this can happen also not intentionally).
- Mike Large: Do any of the great speakers have any hunches about the likeliness that the many variables associated with growing blastocysts may have an impact on NIPGTA?
- Peter: Hi Mike, Thank you for your question. I think it is very likely that many of the variables of embryo culture have impact/or are associated with differences in different testing platforms, and most likely NIPGTA is not an exception either.
- Anabella Marconetto Instead of the micromanipulation expertise that could be acquired learning TE biopsy, now that the niPGT is in advent...do you recommend learning and developing a TE biopsy skills for someone that doesn't perform it now? (Should we still be teaching TE biopsy if the future of PGT is non-invasive? Is the future of PGT noninvasive?)
- Gary: I don’t think that invasive biopsy is going to be totally replaced by non-invasive PGT so I’d suggest continuing to teach and hone the skill of biopsy. (Agree - Peter)
- Dimitra Christopikou: Do you think based on clinical data that pgta should be offered to all patients, none or should it be an indication base? do get referrals for pgta based on indications or to all ivf patients?
- Gary: Yes, I think every patient should be given a thorough counseling on the pros and cons of PGT A including the current data on the use of PGT-A (increase in IR/LBR, decrease in miscarriage rate, decrease in time to pregnancy) and false positive/false negative rates, mosaicism, segmental errors, transfer of mosaic embryos, etc. Following this discussion, the clinical team and the patient should make a decision on what is best for this cycle considering past cycles, future plans, number of embryos available for biopsy etc. PGT-A is not the best option for every patient but I think that every patient deserves to have a fair and balanced assessment of the test ahead of making the decision.
- dimitra Christopikou Do you think that patient reports should include low and high levels of mosaicism?
- Gary: I’m not sure...I really don’t think we know enough about this designation yet. PGD IS is trying to put together a database of known mosaic transfers with outcomes to help here. Most people think low level (< 40 %-50 %) are similar to euploid embryos however, a case report just came out of Turkey showing a 35 % mosaic embryo at PGT-A was transferred and resulted in a mosaic live born with various levels of mosaicism for the chromosome diagnosed at PGT-A affected in different tissues.
- Sara El Gzeiry do you think the inner diameter of embryo affecting on PGTA results
- Peter: Hi Sara, Thank you for your question. Honestly, I have no idea. I would think, if we consider blastocyst stage embryo, and relate inner diameter with the expansion stage of the blastocysts, then I can imagine, that a more expanded blastocyst (larger inner diameter) has more cells, which may by some is more likely to be a “viable” embryo, thus if tested by PGT-A, then it may be associated also with a higher likelihood of euploidy (but this is just pure speculation).