Comprehensive chromosome screening is highly predictive of the reproductive potential of human embryos: a prospective, blinded, nonselection study

Transcription

1 ORIGINAL ARTICLES: ASSISTED REPRODUCTION Comprehensive chromosome screening is highly predictive of the reproductive potential of human embryos: a prospective, blinded, nonselection study Richard T. Scott Jr., M.D., a,b Kathleen Ferry, B.S., a Jing Su, M.S., a Xin Tao, M.S., a Katherine Scott, M.S., a and Nathan R. Treff, Ph.D. a,b a Reproductive Medicine Associates of New Jersey, Morristown, New Jersey; and b Division of Reproductive Endocrinology, Department of Obstetrics Gynecology and Reproductive Science, University of Medicine and Dentistry of New Jersey Robert Wood Johnson Medical School, New Brunswick, New Jersey Objective: To determine both the negative and positive predictive values of comprehensive chromosome screening (CCS) results for clinical outcome. Design: Data obtained from two prospective, double-blinded, nonselection studies. Setting: Academic center for reproductive medicine. Patient(s): One hundred forty-six couples with a mean maternal age of years and a mean paternal age of years. Intervention(s): Embryo biopsy for DNA fingerprinting and aneuploidy assessment. Main Outcome Measure(s): Failure rate of embryos predicted aneuploid by CCS (negative predictive value) and success rate of embryos predicted euploid by CCS (positive predictive value). Result(s): A total of 255 IVF-derived human embryos were cultured and selected for transfer without influence from CCS analysis. Embryos were biopsied before transfer, including 113 blastomeres at the cleavage stage and 142 trophectoderm biopsies at the blastocyst stage. Comprehensive chromosome screening was highly predictive of clinical outcome, with 96% of aneuploid predicted embryos failing to sustain implantation and 41% sustained implantation from embryos predicted to be euploid. Conclusion(s): These nonselection data provide the first prospective, blinded, clinical study directly measuring the predictive value of aneuploidy screening for clinical outcome. The clinical error rate of an aneuploidy designation is very low (4%), whereas implantation and delivery rates of euploid embryos are increased relative to the entire cohort of transferred embryos. (Fertil Steril Ò 2012;97: Ó2012 by American Society for Reproductive Medicine.) Key Words: Comprehensive chromosome screening, SNP microarray, aneuploidy, nonselection The inability to accurately prognosticate whether a given embryo possesses true reproductive potential, such that it will implant and progress to delivery of a healthy infant, is one of the most pressing issues in contemporary reproductive endocrinology (1). Morphologic and temporal assessment of embryonic development is critical to successful assisted reproductive technology (ART) practice, but implantation rates remain modest at best. In the most recent Centers for Disease Control and Prevention report, only 19% of transferred embryos delivered (2). The combination of modest implantation rates and the substantial emotional, physical, and financial challenges that couples experience throughout the process lead couples Received September 29, 2011; revised January 5, 2012; accepted January 13, 2012; published online February 1, R.T.S. has received funding from Merck, EMD Serono, and Ferring Pharmaceuticals. K.F. has nothing to disclose. J.S. has nothing to disclose. X.T. has nothing to disclose. K.S. has nothing to disclose. N.R.T. has nothing to disclose. This work was supported by a grant from Merck/Schering Plough. Reprint requests: Nathan R. Treff, Ph.D., Reproductive Medicine Associates of New Jersey, 111 Madison Ave., Suite 100, Morristown, NJ ( ntreff@rmanj.com). Fertility and Sterility Vol. 97, No. 4, April /$36.00 Copyright 2012 American Society for Reproductive Medicine, Published by Elsevier Inc. doi: /j.fertnstert and clinicians to opt for multiple embryo transfer. As a result, and despite improvements in many aspects of ART, the multiple pregnancy rate remains a great challenge (3). Investigators began exploring possible mechanisms that could result in the failure of morphologically normal embryos to implant and develop into healthy infants. Aneuploidy was an excellent candidate. The prevalence of aneuploidy in live births and pregnancy losses rises dramatically with age and parallels the natural decline in fertility (4). Initial studies used fluorescence in situ hybridization (FISH) and demonstrated a high prevalence of aneuploidy in human embryos (5). Logically, this provided an apparent opportunity to improve outcomes by 870 VOL. 97 NO. 4 / APRIL 2012

2 Fertility and Sterility screening embryos for aneuploidy before transfer. Unfortunately, clinical results have been disappointing. A large number of prospective, randomized clinical trials have consistently failed to show any improvement in delivery rates using FISH-based aneuploidy screening (6). Despite the fact that FISH-based aneuploidy screening has not succeeded in improving clinical outcomes, the concept remains valid. Embryos that are trisomic and monosomic should principally fail to implant, and those that do implant will generally result in pregnancy loss, with very few aneuploid gestations actually resulting in live births. A 24-chromosome aneuploidy screening system has been developed that may provide a more comprehensive and accurate approach to improve outcomes (7). A key issue is how best to validate new technologies that are proposed as embryo diagnostics. There is no universally accepted paradigm, but one model has recently been described (8). In that paradigm, three fundamental types of data are needed. First,the predictive value of the technique must be confirmed on cell lines of known abnormal genetic complements. These should then be repeated on human embryos that have been discarded and donated for research, to demonstrate that consistent results are possible with both blastomeres and trophectoderm. Next in the paradigm, the positive and negative predictive values of the test would be determined. Given that no test is ever absolutely accurate, it is inevitable that some embryos labeled as aneuploid will actually be normal and have normal reproductive potential, and vice versa. By directly determining the predictive value of these tests, it will be possible for embryologists and clinicians to incorporate that precision into their decision making about when to offer chromosomal screening and for patients to be accurately counseled about the risk of misdiagnosis. A true nonselection study provides a rigorous method to accurately calculate the positive and negative predictive values of the test. Patients and their embryos undergo routine care up through the point at which embryos are selected for transfer. Immediately before transfer, the embryos are biopsied. Transfer occurs without the use of results from genetic analysis. Each embryo is followed individually to determine whether it resulted in delivery, with the use of embryonic DNA fingerprinting. This type of nonselection study has not been described previously and may provide a unique opportunity to directly evaluate the predictive values of these technologies before clinical implementation. Finally, a prospective, randomized, controlled clinical trial would then be necessary to determine whether using chromosome screening results to select embryos for transfer will actually significantly improve delivery rates. The first of these three steps in this paradigm has previously been reported (7). This article details the second step in this process: a prospective, blinded, nonselection study that allows direct assessment of the positive and negative predictive values of the technique. MATERIALS AND METHODS Population and ART Cycles The study population consisted of couples attempting conception in which the female partner was aged years. Cycles using an oocyte donor were also included. All patients had to have basal follicle counts of 10 or more and serum basal FSH concentrations <12 IU/L. Couples with two or more prior failed IVF retrievals (does not include cryopreserved ET cycles) were excluded. Couples with a history of endometrial insufficiency or with obstructive azoospermia requiring surgical sperm retrieval were also excluded. There were no restrictions on the type of follicular stimulation used in the study. Both partners had peripheral blood samples collected, which allowed for isolation of their DNA. It was essential to have parental DNA to allow for accurate DNA fingerprinting of the embryos and the infants. All aspects of retrieval and oocyte recovery were performed using established routine laboratory procedures. All cycles used intracytoplasmic sperm injection to ensure that an embryo biopsy could be accomplished without risk of contamination from spermatozoa that had been bound to the zona pellucida (ZP). The patients were selected for either day-3 or day-5 transfer independent of participation in the study. Per standard protocol in this laboratory, embryos from patients with four or more high-quality embryos on day 3 (six or more cells with <15% fragmentation) were placed into extended culture and designated for blastocyst transfer. Embryos deemed to have the best morphologic development were selected for transfer per routine. Shortly before transfer, the embryos that had been selected underwent embryo biopsy. A single blastomere was removed after creation of a breech on the ZP using a laser for day-3 biopsies. Embryos selected for transfer on day 3 were transferred within minutes of biopsy and were not placed into extended culture. If the patients were planned for day-5 transfer, then a small opening in the ZP was made using the laser before placing the embryos into extended culture. This would allow herniation of some trophectoderm through the ZP at the time of blastocyst expansion and facilitate biopsy on day 5. All embryos had attained an expansion score of R3 using Gardner s scoring system (9). Embryo transfer was completely per routine. No results of any kind were available. The genetic analysis of the biopsied cells was performed in batches, typically several weeks later. In those cases in which a pregnancy was established, it was necessary to obtain a DNA sample on the conceptus. There were two fundamental approaches to obtaining DNA from conceptus. First, some patients were available to have peripheral blood samples obtained through routine phlebotomy at approximately 9 gestational weeks. A prior study has demonstrated that cell-free fetal DNA is sufficient to allow precise DNA fingerprinting at that time (10). Alternatively, a DNA sample could be attained by buccal swab from the newborn after delivery (11). In the unfortunate circumstance of a clinical loss, villous material was the source of DNA from the conceptus. In cases of biochemical loss or clinical loss in which no pathology specimen was available, no fingerprinting analysis was possible. The latter circumstance did not impair the analysis because the embryos clearly did not result in a viable gestation and were designated as having failed. VOL. 97 NO. 4 / APRIL

3 ORIGINAL ARTICLE: ASSISTED REPRODUCTION Once the genotype of the conceptus was known, it was possible to compare those results with the genotype of the embryos that were transferred and to determine which implanted and sustained development as previously described (10, 11). Experimental Design The data in this study were extracted from two separate studies, both registered with ClinicalTrials.gov under the identifiers NCT and NCT ( Both studies included institutional review board approval and informed patient consent. Data from both studies were included, to maximize the relevant sample size for the current investigation. These studies had in common the fact that the embryos were biopsied and transferred without knowledge of any screening results. The first nonselection study was designed to assess the predictive value of multiple technologies. Patients went through their ART care without regard to participation in the study. During their laboratory care, multiple specimens were collected as a part of normal embryology procedures. Specimens collected for future analyses that would be considered noninvasive included granulosa cells, leftover spermatozoa, and spent culture media. Oocytes underwent first polar body biopsy at the time of intracytoplasmic sperm injection, second polar body biopsy at the time of fertilization check, and embryo biopsy at the time of transfer. It is possible that the three biopsies might adversely impact the ultimate implantation potential of these embryos, which limits the ability to directly extrapolate the implantation rates attained in this study to routine clinical care. Nevertheless, all oocytes and embryos were treated equally, which means that those that were eventually diagnosed as being either aneuploid or euploid would be similarly impacted by the biopsy. Differences in implantation rates amongst embryos attained in the study may not be attributed to their having undergone three biopsies. Data from a second nonselection study were also used in this analysis of the predictive value of comprehensive chromosome screening (CCS). In this investigation, patients underwent a two-embryo transfer. Immediately before transfer, one of the two embryos was biopsied. As before, no analysis of the biopsied cells was done before transfer. When a pregnancy occurs, it is then possible to use DNA fingerprinting to determine whether the biopsied embryo implanted. Thus, the same types of data are available for the biopsied embryos in this study as from the first study because the genetic result is only known afterward. Obviously, no information is obtained from the nonbiopsied embryo, and its outcome is not included in or relevant to the results of the present analysis. However, preliminary results of directly assessing the impact of biopsy have been presented and indicate that blastomere biopsy significantly negatively impacts implantation potential (31% with biopsy and 53% without biopsy; P¼.035), whereas blastocyst biopsy does not (52% with biopsy and 54% without biopsy; P¼.80) (12). Determining the Karyotype of the Embryo: Microarray-Based CCS Individual blastomeres or trophectoderm biopsies were processed as previously described (7). Briefly, cells were lysed in alkaline solution and underwent whole-genome amplification using GenomePlex WGA4 (Sigma Aldrich), followed by single nucleotide polymorphism (SNP) microarray based analysis of copy number and genotypes using NspI SNP genotyping microarrays, copy number analysis tool, and GTYPE software (Affymetrix). DNA Fingerprinting to Determine Which Embryos Implanted and Progressed Through Delivery Parental genomic DNA was genotyped on the NspI microarray as recommended by the supplier (Affymetrix) and used to identify informative SNPs for performing genotyping comparisons between embryonic and conceptus-derived DNA, as previously described (10, 11). Statistics Calculation of the predictive values of normal and abnormal preimplantation genetic screening (PGS) results was straightforward. The outcomes for each embryo were determined. Embryos that implanted and progressed to delivery were considered to have a successful outcome. All other embryos, whether they failed to implant or resulted in some type of loss, were considered to have failed. The predictive value of an abnormal result was calculated by dividing the total number of embryos that had been designated as genetically abnormal by PGS and that ultimately delivered by the total number of embryos that had been transferred that had abnormal PGS results (abnormal delivered/all abnormals). The results are expressed as a percentage. The predictive value of a normal result was calculated in the classic form by expressing the percentage of embryos that delivered as a function of the total number of embryos that were designated as having normal genetics. Other analyses included evaluating the outcomes of embryos that screened normal or abnormal after day-3 vs. day-5 transfer. The impact of age was also evaluated. These outcomes were compared through contingency table analyses. An a error of 0.05 was considered significant for all comparisons. RESULTS Clinical Outcomes One hundred forty-six patients aged years age participated in the study. The mean maternal age of patients for whom transferred embryos were evaluated was years, and the mean paternal age was years. Eighty-one (55%) achieved a biochemical pregnancy (þbhcg), 62 (42.4%) obtained a clinical pregnancy (þsac), and 55 (37.6%) either delivered or have been discharged with an ongoing pregnancy. Of 255 embryos transferred, 72 (28.2%) resulted in a clinical implantation. The clinical implantation rate for the patients participating in the triple-biopsy study was 25.5% (38 of 149). The clinical implantation rate for patients participating in the single-biopsy study was 41.0% (34 of 83). After controlling for the embryonic stage at transfer, there were no differences in implantation rates between the two groups. As such, there is no evidence that the additional 872 VOL. 97 NO. 4 / APRIL 2012

4 Fertility and Sterility polar body biopsy adversely impacted implantation rates, and both data sets are combined for the remainder of the analysis. FIGURE 1 Human Embryo DNA Characterization A total of 255 embryos were biopsied before transfer, including 113 blastomeres at the cleavage stage and 142 trophectoderm biopsies at the blastocyst stage. Seven blastomeres and five trophectoderm biopsies (4.7%) failed to amplify. Five blastomeres and six trophectoderm biopsies (4.3%) resulted in nonconcurrent copy number assignments (7). A total of 99 of the 232 evaluable microarray results (42.7%) were aneuploid. Fifty-two (22.4%) had more than one abnormality, 26 (11.2%) had a single monosomy, and 21 (9.1%) had a single trisomy. Every autosome was found at least once in the monosomy state and at least once in the trisomy state. Microarray Predictive Value for Reproductive Potential of Human Embryos There were 99 embryos that would have been diagnosed as aneuploid using microarray-based screening. Four of these embryos (4%) led to delivery of healthy euploid children. This was significantly lower than the overall rate of delivery of all embryos transferred (28.2%; P<.0001). The negative predictive value of a microarray-based aneuploidy diagnosis was therefore 96% (failure of delivery per aneuploid embryo transferred). There were 133 embryos that would have been diagnosed as euploid using microarray-based screening. Fifty-five of these embryos led to delivery of healthy euploid children. The positive predictive value of a microarray-based euploid result for the reproductive potential of each embryo was therefore 41.4% (delivery per euploid embryo transferred). This was significantly higher than the overall rate of delivery of all embryos transferred (28.2%; P<.0002). Five embryos with evaluable microarray results led to a clinical loss (2.2%). Two were aneuploid (47,XX,þ17 and 47,XY,þ17), and three were euploid (46,XY). All predicted karyotypes were confirmed by analysis of the products of conception. The percentage of aneuploid embryos leading to a clinical loss (2.0%; 2 of 99) was not significantly different from the percentage of euploid embryos leading to a clinical loss (2.3%; 3 of 132; P¼1.000). Comparison of Blastomere and Trophectoderm Biopsy Predictive Values The identification of the most appropriate stage of embryo biopsy for PGS of aneuploidy remains of great interest. Embryo-specific outcomes were segregated according to the type of embryo biopsy performed, to compare the negative and positive predictive values of blastomere and trophectoderm biopsy (Fig. 1). Of 131 blastocyst-stage embryos with a trophectoderm biopsy, 85 (64.9%) would have been given a euploid diagnosis by microarray. Forty-one of these (48.2%) led to a delivery or an ongoing pregnancy. Of 101 cleavage-stage embryos with a blastomere biopsy, 48 (40.6%) would have been given a euploid diagnosis by microarray. Fourteen of these (29.2%) led to a delivery or ongoing Predictive values of CCS for clinical outcome with either cleavagestage blastomere biopsy or blastocyst-stage trophectoderm biopsy. Scott. Clinical predictive value of CCS. Fertil Steril pregnancy. This was significantly lower than the trophectoderm positive predictive value (P¼.0016). Forty-six blastocyst-stage embryos with trophectoderm biopsy would have been given an aneuploid diagnosis by microarray analysis. Three of these led to delivery of a healthy euploid newborn, resulting in a 93.5% negative predictive value (43 of 46). Fifty-three cleavage stage embryos with a blastomere biopsy would have been diagnosed as aneuploid by microarray analysis. One of these led to delivery of a healthy euploid newborn, resulting in a 98.1% negative predictive value (52 of 53). This was not significantly different (P¼.1525). Age-Related Euploid Embryo Delivery Rates Because the age-related diminution in reproductive potential is largely attributed to an increase in aneuploidy incidence, we evaluated whether euploid embryos are equally capable of delivery after transfer to young women compared with those of advanced maternal age. Figure 2 displays the delivery rates per embryo for patients within each of the two age groups. Results indicate that there remains a significant decline in the reproductive potential of embryos as a function of reproductive age, even with removal of aneuploid embryos from the analysis. DISCUSSION The present work represents the first-ever direct clinical evaluation of the predictive values of both a normal and an abnormal result when screening human embryos for aneuploidy. These types of data are critical to understanding the precision of embryonic aneuploidy screening and to be able to adequately counsel patients about the risks of potentially discarding a normal embryo. It may be particularly important in those patients who have limited numbers of embryos. If a given technology mistakenly documents a chromosomal imbalance, then that patient has the potential to be harmed VOL. 97 NO. 4 / APRIL

5 ORIGINAL ARTICLE: ASSISTED REPRODUCTION FIGURE 2 Delivery rate per embryo transferred and according to the overall outcome of all embryos transferred or those specifically predicted to possess a euploid karyotype by CCS. Scott. Clinical predictive value of CCS. Fertil Steril by precluding the transfer of the only embryo with the only real chance to conceive a healthy conceptus from that cohort. Additionally, inadvertent transfer of an aneuploid embryo that had been identified as euploid may result in failed treatment, pregnancy loss, or rarely the development of an ongoing abnormal gestation. The predictive value of a normal result was excellent. No embryo designated as normal developed into an aneuploid evaluable clinical loss or ongoing abnormal gestation. Of course, it is not possible to evaluate the outcomes of those embryos that failed to implant. Although it would be ideal, the reality is that the predictive value of aneuploid testing results was not flawless. Four percent of the embryos designated as abnormal implanted and developed into healthy chromosomally normal infants. This error rate is an important piece of information when providing clinical counseling because patients need to know that results are not absolutely predictive of outcome. Both laboratory and clinical sources of error may contribute to these clinical misdiagnoses. One possibility that must be acknowledged is that the test result may have been wrong. A normal euploid cell might have been biopsied, and the laboratory assessment might be inaccurate and provide a misdiagnosis of aneuploidy. A prior study has reported that the laboratory error rate is less than 2% (7). Although relatively low, it is definitely not zero. It is not possibly to directly compare results with those attained with other technologies because the predictive values of any other technology have never been directly evaluated. One study has estimated that an aneuploid result after FISH-based aneuploidy screening may have been inaccurate as often as 60% of the time, although with a limited sample size (13). In any event, the possibility of laboratory misdiagnosis is small but real. Another possibility is that the laboratory diagnosis is correct but that a clinical error still results. Mosaicism is an established phenomenon, and some mosaics have ongoing development and survive to term with normal infants. Embryonic mosaicism is currently an inescapable source of error in this setting and may lead to a clinical error independent of the accuracy of the laboratory diagnosis. Interestingly, all four misdiagnosed embryos also had both polar bodies biopsied. Although polar body analysis was not the purpose of this study, these were evaluated for aneuploidy. In each case, both polar bodies were determined to be euploid. Although this is not definitive evidence of mosaicism, elimination of maternal meiotic aneuploidy as the source increases the likelihood that the observed embryonic aneuploidy was postzygotic mitotic in origin. It may seem reasonable to assume that, as a result of this study, all array-based aneuploidy screening technologies will provide adequate negative predictive values. However, there are a number of reasons why this is not the case. Different array manufacturers use markedly different strategies for determining copy number assignments, and these have been shown to produce different results when applied to the same sample (14). Different whole-genome amplification strategies, including polymerase chain reaction or isothermal rolling circle based amplification, also display unique results from the same sample (15). Data analysis strategies and preclinical validation study designs must also be considered. Extrapolation from this study to all forms of array-based aneuploidy screening would be unsafe and is therefore unwarranted. Clearly, these data do not indicate that this is the only valid approach to 24-chromosome PGS or that this laboratory is the only one that can perform this type of analysis with precision. This methodology applied in an equivalent fashion would work in any laboratory. However, other methodologies, which might include different amplification strategies, microarray or other laboratory platforms, or statistical algorithms, may not be presumed to have equivalent predictive values. This is no different than any other diagnostic testing scenario in which multiple methodologies exist for making the same measurement. Each must be rigorously validated before being presumed correct. This study provides significant insight into reproductive aging. As anticipated, there was an age-related diminution in euploidy rates. In contrast, there was also a significant decline in implantation rates among euploid embryos with increasing age. Clearly there is more to the age-related decline in reproductive potential than aneuploidy. Further studies evaluating other factors that increasingly limit reproductive outcomes with advancing maternal age are needed. Studies of the transcriptome, secretome, and metabolome are currently underway. These data do not address whether clinical outcomes will be improved through the application of this technology. This was a true prospective, blinded, nonselection study. The data on aneuploidy were not available until long after the completion of the entire treatment cycle and did not direct any 874 VOL. 97 NO. 4 / APRIL 2012

6 Fertility and Sterility treatment decisions. It will now be necessary to perform a selection trial to specifically determine whether adding this information to the paradigm used to select embryos for transfer will ultimately result in higher implantation rates as well as lower loss rates and an increase in live births. A randomized clinical trial is underway at this time (16). In conclusion, the application of sophisticated molecular genetic screening seems to reliably diagnose human embryonic aneuploidy. If the randomized clinical trial demonstrates enhanced implantation rates similar to that seen among euploid embryos in this study, this technology could provide a critical next step to improving clinical outcomes while possibly reducing transfer order. Those studies are ongoing at present. Acknowledgments: The authors thank the participating patients, the Reproductive Medicine Associates of New Jersey Embryology Team, the Colorado Center for Reproductive Medicine, and Reproductive Medicine Associates of Connecticut, Michigan, and New York. REFERENCES 1. Jones HW Jr. Seven roads traveled well and seven to be traveled more. Fertil Steril 2011;95: Centers for Disease Control and Prevention. National summary report. Assisted reproductive technology (ART) report. Atlanta: CDC; Adamson D, Baker V. Multiple births from assisted reproductive technologies: a challenge that must be met. Fertil Steril 2004;81: Hassold T, Hunt P. To err (meiotically) is human: the genesis of human aneuploidy. Nature Rev 2001;2: Munne S, Alikani M, Tomkin G, Grifo J, Cohen J. Embryo morphology, developmental rates, and maternal age are correlated with chromosome abnormalities. Fertil Steril 1995;64: Fritz MA. Perspectives on the efficacy and indications for preimplantation genetic screening: where are we now? Hum Reprod 2008;23: Treff NR, Su J, Tao X, Levy B, Scott RT Jr. Accurate single cell 24 chromosome aneuploidy screening using whole genome amplification and single nucleotide polymorphism microarrays. Fertil Steril 2010;94: Scott RT Jr, Treff NR. Assessing the reproductive competence of individual embryos: a proposal for the validation of new -omics technologies. Fertil Steril 2010;94: Gardner DK, Schoolcraft WB. In vitro culture of human blastocysts. In: Jansen R, Mortimer D, editors. Towards reproductive certainty: infertility and genetics beyond. Carnforth, United Kingdom: Parthenon Press; 1999: Treff NR, Tao X, Su J, Lonczak A, Northrop LE, Ruiz A, et al. Tracking embryo implantation using cell-free fetal DNA enriched from maternal circulation at 9 weeks gestation. Mol Hum Reprod 2011;17: Treff NR, Su J, Tao X, Miller KA, Levy B, Scott RT Jr. A novel single-cell DNA fingerprinting method successfully distinguishes sibling human embryos. Fertil Steril 2009;94: Treff NR, Ferry KM, Zhao T, Su J, Forman EJ, Scott RT. Cleavage stage embryo biopsy significantly impairs embryonic reproductive potential while blastocyst biopsy does not: a novel paired analysis of cotransferred biopsied and non-biopsied sibling embryos. Fertil Steril 2011;96:S Northrop LE, Treff NR, Levy B, Scott RT Jr. SNP microarray-based 24 chromosome aneuploidy screening demonstrates that cleavage-stage FISH poorly predicts aneuploidy in embryos that develop to morphologically normal blastocysts. Mol Hum Reprod 2010;16: Levy B, Treff NR, Nahum O, Su J, Tao X, Scott RT Jr. The accuracy and consistency of whole genome preimplantation genetic diagnosis (PGD): a comparison of two independent methods microarray PGD (mpgd) and comparative genomic hybridization (CGH). Fertil Steril 2008;90:S Treff NR, Su J, Tao X, Northrop LE, Scott RT Jr. Single-cell whole-genome amplification technique impacts the accuracy of SNP microarray-based genotyping and copy number analyses. Mol Hum Reprod 2011;17: Scott RT Jr, Tao X, Taylor D, Ferry K, Treff N. A prospective randomized controlled trial demonstrating significantly increased clinical pregnancy rates following 24 chromosome aneuploidy screening: biopsy and analysis on day 5 with fresh transfer. Fertil Steril 2010;94:S2. VOL. 97 NO. 4 / APRIL