If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Is FAST-SeqS an accurate methodology for preimplantation genetic testing for whole-chromosome aneuploidy (PGT-A), what additional types of chromosomal abnormalities can be assessed, and what are the observed aneuploidy rates in a large clinical cohort?
FAST-SeqS, a next-generation sequencing- (NGS-) based assay amplifying genome-wide LINE1 repetitive sequences, was validated using reference samples. Sensitivity and specificity were calculated. Clinically-derived trophectoderm biopsies submitted for PGT-A were assessed, and aneuploidy and mosaicism rates among biopsies were determined. Clinician-provided outcome rates were calculated.
Sensitivity and specificity were >95% for all aneuploidy types tested in the validation. Comparison of FAST-SeqS to VeriSeq showed high concordance (98.5%). Among embryos with actionable results (n=182,827), 46.2% were aneuploid. Whole-chromosome aneuploidies were most commonly observed (72.9% without or 8.7% with a segmental aneuploidy [SA]), with rates increasing with egg age, while SA rates did not. SAs (n=20,557) were observed on all chromosomes (most commonly deletions), with frequencies roughly correlated with chromosome length. Mosaic-only abnormalities constituted 10.1% (n=3,862/38,145) of samples. Abnormal ploidy constituted 1.8% (n=2,370/128,991) of samples, triploidy being the most common (73.6%). Across 3,297 frozen embryo transfers, the mean clinical pregnancy rate was 62% (range 38-80%); the mean combined ongoing pregnancy and live birth rate was 57% (range 38-72%).
FAST-SeqS is a clinically reliable and scalable method for PGT-A, is comparable to whole genome amplification-based platforms, and detects additional information related to ploidy using SNP analysis. Clinical outcomes suggest an ongoing benefit of PGT-A using FAST-SeqS, consistent with other platforms.
Preimplantation genetic testing for aneuploidy (PGT-A) in conjunction with in vitro fertilization (IVF) is used to identify euploid embryos for transfer to increase the rate of successful pregnancies. The available technologies for PGT-A have evolved and advanced over time, expanding the number of chromosomes interrogated as well as the type of detectable chromosome abnormalities in trophectoderm (TE) biopsies (
). Today, common NGS methods using whole genome amplification (WGA) can provide genome-wide analysis of WCAs and SAs but are unable to determine abnormal ploidy or uniparental isodisomy (UPiD) due to the inability to accurately identify SNP genotypes as a result of low sequencing depth of coverage (ESHRE PGT-SR/PGT-A Working
). In contrast, SNP-based PGT can accurately infer ploidy state and UPiD, although this testing generally requires analysis of parental samples (potentially problematic with donor gametes) and may be suboptimal for detecting some SAs and mosaicism wherever insufficient SNP coverage exists (
Practice Committees of the American Society for Reproductive Medicine and the Society for Assisted Reproductive Technology The use of preimplantation genetic testing for aneuploidy (PGT-A): a committee opinion.
). However, methods used for PGT-A have been laborious, technically complex, prone to error and contamination due to frequent manual manipulations, and limited to testing small numbers of samples per batch. While the use of PGT-A has consistently increased, these method constraints have impeded high-throughput testing. In addition, embryo biopsy and PGT-A analysis have been commonly viewed as another expensive addition to the already prohibitive cost of IVF, which are often not covered by health insurance. These challenges have slowed the introduction of large-scale, lower-cost PGT-A that could benefit families in their reproductive decision making.
With the goal of improving access to PGT-A by lowering cost while maintaining clinically robust performance, the FAST-SeqS assay, originally developed for non-invasive prenatal screening (
). The goal of assay development was to ensure performance at a level equal to or better than the gold standard market assay at the time (aCGH). FAST-SeqS utilizes universal primers to amplify LINE1 (L1) repetitive elements at more than 20,000 locations across the genome. Unlike WGA, this method does not require a library preparation step and therefore can process larger numbers of samples concurrently using standard PCR reactions. FAST-SeqS combines the advantages of both NGS- and SNP-based arrays and is therefore able to simultaneously detect a broad variety of chromosome abnormalities, including WCA, SA, polyploidy, and select UPiD. This study describes the validation of the FAST-SeqS assay, the rates and varieties of chromosomal abnormalities detected in >190,000 TE biopsies, and the clinical pregnancy outcomes in a subset of frozen embryo transfers (FETs) following FAST-SeqS-based PGT-A.
MATERIALS AND METHODS
FAST-SeqS assay design
FAST-SeqS was modified to detect WCAs, SAs (≥10 Mb, including those derived from inherited structural rearrangements), whole-chromosome UPiD (excluding chromosomes 17, 19-22 due to the paucity of SNPs), and abnormal ploidy (i.e., haploidy, triploidy, and tetraploidy with varying sex chromosome complements [specifically, XXXX, XXXY, and some XXYY]) in TE biopsies (
). Briefly, FAST-SeqS uses universal PCR amplification of L1 repetitive elements at >20,000 locations across the genome (Figure 1). The resulting amplicons are uniquely tagged with embryo-specific barcodes and sequenced in pooled libraries. The sequences of the L1 elements are conserved enough to be amplified with a single pair of universal PCR primers but sufficiently diverged to be unambiguously aligned to a reference genome. The use of a single primer pair substantially reduces the amplification bias that can be seen with WGA protocols. In addition, subsequent validation with verified cell lines and genomic DNA samples demonstrates a lack of bias. Due to the small amount of cellular input, all measures were taken to prevent contamination, including dedicated staff and a dedicated negative-pressure clean room. Negative controls are included in the assay, with both the control strip run with each plate and clinic-provided blanks for each biopsy set. Contamination can be observed in SNP profiles but rarely is of the level that impacts clinical results.
A custom-designed bioinformatics pipeline and visualization software package (Circular Binary Segmentation [CerBeruS]) was implemented to predict copy number along the length of each chromosome based on sequence read depth and to infer ploidy status based on SNP genotyping (
). In brief, each chromosome is computationally “circularized,” and segments of sequence are compared to neighboring segments for significant differences in copy number proportions. This pipeline was rigorously validated over multiple studies using a combination of verified cell lines, verified genomic DNAs, and 25,000 computational simulated samples. Samples failing to meet quality thresholds were categorized as “non-actionable”. Mosaicism for a suspected abnormality was reviewed and called by a board-certified clinical laboratory director, as sample-to-sample variability can confound automated mosaicism calling. Copy number values of 1.8-2.2 were classified as normal; however, rather than using discrete copy-number thresholds for mosaic changes, each call was made in the context of the remaining genome. This method was utilized to prevent overcalling of mosaicism in noisy samples and omitting mosaic calls in clearly atypical samples due to overly stringent thresholds.
This validation of the FAST-SeqS assay utilized reference specimens with known abnormalities (cell cultures and genomic DNA, Coriell Institute, Camden, NJ; Supplemental Methods, Supplemental Table 1) to establish its performance in detecting the following: WCA, SA, haploidy, and triploidy. In detail, reference samples included 5 euploid genomic DNAs; 6 WCA genomic DNAs; 121 genomic DNAs with SAs ranging from 101 kb to 81.5 Mb in size; 5 fibroblast cell lines with triploidy; 1 fibroblast cell line with whole-genome UPiD; and 2 fibroblast cell lines with single-chromosome UPiD. All samples were analyzed using the CerBeruS bioinformatics analytical pipeline and visualization software.
Reproducibility was determined by analyzing both inter-run and intra-run replicates for concordance. Inter-run studies included 20 reference samples with 2-6 replicates each, for a total of 70 individual samples. Intra-run studies included 41 reference samples with 3-8 replicates each, for a total of 202 individual samples. In addition, reproducibility was confirmed as we validated this assay in two different locations.
Limit of detection (LOD) and mosaicism studies were undertaken using micromaniuplated samples of both euploid and WCA cell lines (Supplemental Methods). To determine the LOD, six reference samples with varying cell inputs (one, two, or five cells) were tested in eight replicates to determine sensitivity and specificity. Finally, mosaicism was simulated with mixed samples with varying combinations (five cells total) of one euploid female and one aneuploid male (trisomy 21) reference samples, with a total of six cell mixture classes. Thirteen replicates of each mixture were tested to determine the sensitivity and specificity of detecting a mosaic abnormality (trisomy 21 or monosomy X) at various calling thresholds.
Clinical accuracy was estimated by performing two separate dual-biopsy studies. In the first study, dual biopsies from a single embryo were tested with FAST-SeqS and aCGH for WCAs. In the second study, dual biopsies from a single embryo were tested with FAST-SeqS in two study arms for WCAs and SAs to determine intra-laboratory and inter-laboratory concordance rates. In Arm 1, both biopsies were run in the same laboratory in separate runs, while Arm 2 ran one biopsy each in two separate laboratories.
FAST-SeqS was compared to the VeriSeq assay previously validated in our laboratory. Euploid (n=4), WCA (n=5) and SA (n=118 with 127 abnormalities ≥10 Mb) samples were run on both assays to measure the concordance in detection rates.
Study population and data analysis
TE biopsy specimens derived from blastocysts obtained from individuals undergoing IVF among 241 US-based fertility centers analyzed between May 12, 2016 and April 23, 2020 were included. Review and analysis of fully de-identified data were approved by the Western Institutional Review Board (WIRB 1167406). TE biopsies from individuals with known structural rearrangements were excluded from this analysis, as they would be expected to have higher rates of aneuploidy than typical IVF patients of similar age. Testing was performed at two CLIA-certified laboratories, initially at Good Start Genetics (Cambridge, MA) and more recently at Invitae (San Francisco, CA). Specimens tested between May 12, 2016 and December 17, 2017 were assessed for WCAs and SAs only. From December 18, 2017 through the end of the study, ploidy and UPiD was assessed. The overall aneuploidy rate was calculated. Aneuploid biopsies were further defined by the type of abnormality: WCA only, SA only, WCA and SA, and atypical ploidy. WCA and SA rates were calculated and evaluated in relation to egg age. Biopsies that failed quality metrics and were not reported with a clear euploid/aneuploid result were deemed Non-Actionable. These include samples with no results, as well as indeterminate and special considerations samples (those that fail quality metrics, with or without abnormalities detected, respectively).
Ploidy and UPiD rates were calculated from a subset of 134,720 TE biopsy specimens in which the SNP-based analyses were performed. The proportion of IVF cycles with at least one euploid embryo was calculated by age and number of embryos tested in each cycle. Abnormal ploidy rates were compared between intracytoplasmic sperm injection- (ICSI-) and non-ICSI-derived embryos using a Chi-square test (p ≤ 0.05 considered statistically significant).
Mosaicism rates were calculated in a subset of cases (n=38,145) when requested by the clinician. An embryo was defined as mosaic when all abnormalities were present at a mosaic level (i.e., if both a full trisomy and a mosaic trisomy were present, the sample would be classified as non-mosaic).
Lastly, pregnancy outcomes were assessed based on clinician-provided data for frozen embryo transfers (FETs, n=3,297), including ongoing pregnancy and live birth rates.
FAST-SeqS assay validation and concordance with VeriSeq
Validation of FAST-SeqS using 140 reference samples demonstrated 100% sensitivity and specificity for WCAs, ploidy, and UPiD (Supplemental Table 2). For SAs, sensitivity and specificity for abnormalities ≥ 10Mb in size were 97.7% and 100%, respectively, due to false negative results in four samples. One complex sample (NA21883) had adjacent abnormalities on chromosome 15; the 25.8 Mb duplication was not reported, but the 5.2 Mb deletion was evident. Another sample with multiple abnormalities (NA04993) was identified as carrying a 8.3 Mb duplication on chromosome 4, while the 11.1 Mb deletion on chromosome 10 was omitted. Both of these samples were still identified as abnormal on a sample level. FAST-SeqS failed to detect an 11.1 Mb duplication on chromosome 17 (NA23053); this region is only covered by two sequence bins and thus did not reach the four-bin threshold for reporting. Similarly, the assay failed to detect a 14.8 Mb duplication on chromosome 22 (NA02325), which is covered by three sequence bins. Nonetheless, abnormalities as small as 3.6 Mb were able to be detected, albeit not consistently across the genome. Therefore, overall accuracy for detecting SAs ≥ 10Mb in size was 97.8%. Precision was 98% for replicate samples within individual test batches and 100% for replicate samples split across multiple test batches. These results were similar to three validation studies conducted previously in our Cambridge, MA laboratory (
Since embryo biopsies yield a small number of cells and limited amount of DNA, samples with known numbers of cells established the minimum input threshold for the assay. The one-cell and two-cell samples had a greater proportion of test failures. Among one-cell replicates that had reportable results, sensitivity was 96.3% and specificity 99.8% (Supplemental Table 3). Test performance on samples with two or five cells showed a sensitivity of 100% and specificity ranging from 99.7% to 99.8%.
To assess the ability to detect mosaicism in a TE biopsy, five samples with mixed chromosome complements were tested in replicate. When one aneuploid cell was present among four euploid ones (simulating 20% mosaicism), sensitivity was 28.6% but specificity was 91.1%. However, when the level of simulated mosaicism increased to 40% (2/5 aneuploid cells), sensitivity increased to 100% and specificity to 96.4% (Supplemental Figure 1). As expected, the ability of the assay to detect mosaicism was greater for chromosome X than for chromosome 21 due to the larger number of sequenced regions on chromosome X.
To estimate the clinical accuracy of the FAST-SeqS assay compared to aCGH (the gold standard assay at the time), dual-biopsy studies were performed as part of the initial validation. Of 238 pairs submitted from four IVF centers, 172 passed quality metrics on both platforms. Sample-level accuracy, using aCGH output as truth, was 97.7%; one false negative and three false positive calls were made using FAST-SeqS. Chromosome-level accuracy was 99.8% (Supplemental Table 4). The false negative call was a segmental gain on the short arm of chromosome 20; this iteration of the FAST-SeqS assay was designed to detect WCAs and not SAs. Intra-laboratory and inter-laboratory FAST-SeqS concordance rates, analyzing approximately 150 biopsy pairs each, were both roughly 90% (Supplemental Table 5). The “discordant” results are most likely due to mosaicism within the trophectoderm but could also represent false negative and false positive calls.
The FAST-SeqS vs. VeriSeq comparison using genomic DNA samples (n=127) demonstrated 98.5% concordance (Supplemental Table 4). Only WCA and SA reference samples were used in the comparison study, as it is established that VeriSeq cannot detect abnormal ploidy or UPiD. Nine DNA samples that were euploid (n=4) or had WCAs (n=5) showed 100% concordant results. Among SAs, each showed 99.2% accuracy for the remaining 118 DNA samples with SAs ranging in size from 3.6-81.5 Mb (each method with one false negative), with a concordance rate of 98.4% for each detectable abnormality. FAST-SeqS failed to detect the same 17p duplication as the initial validation, while VeriSeq failed to detect a 12.6 Mb gain of Y chromosome material in a diploid female sample.
Distribution of clinical PGT-A findings
A total of 191,325 embryos from 40,403 cycles and 31,649 patients were included in the clinical evaluation component of this study. All specimens were assessed for WCAs and SAs, while 70.4% (n=134,720) were also assessed for ploidy and UPiD (Figure 2a). The average clinician-reported egg age was 35 years (range 18 to 55 years) (Supplemental Figure 2). Among embryos with an actionable result (n=182,827), 53.8% were euploid and 46.2% were aneuploid (Supplemental Table 7; Supplemental Figures 3, 4). Among the 84,546 aneuploid embryos, 2.5% (n=4,720) were complex abnormal (five or more abnormalities). The overall aneuploidy rate among donor eggs was 31.2% (n=8,991/28,835). Among all patients in our cohort, 78% of cycles had at least one euploid embryo available for transfer. This chance correlated inversely with egg age but directly with the number of biopsied embryos per cycle (Figure 2b).
Among aneuploid embryos, WCA was the most common observed abnormality, present either without (72.9%) or with (8.7%) an accompanying SA (Figure 2a). As expected, WCA rates increased with egg age (Figure 3a). Aneuploidy rates were slightly lower in biopsy samples from day 5 blastocysts compared to day 6 blastocysts (Figure 3b). A total of 112,763 abnormalities in 69,006 biopsies with at least one WCA were observed across all chromosomes, of which 60,489 (53.6%) were monosomies and 52,274 (46.4%) were trisomies (Figure 4a). The most frequently observed whole-chromosome aneuploidies involved chromosomes 16 and 22 and the least common involved chromosomes 1, 3, and Y.
In contrast to an egg age-dependent increase in WCA rates, SA rates were independent of egg age (Figure 3a). In the 20,557 embryos with at least one SA, 26,449 events were observed across all chromosomes, and their frequencies per chromosome were roughly correlated with chromosome length (Figure 4b, Supplemental Figure 5). SAs on the long arms of chromosomes constituted the majority (67.3%), and deletions were more common than duplications, with the exception of the short arm of chromosome 16. SAs involving whole chromosome arms comprised 24.5% of deletions and 22.7% of duplications (Supplemental Table 5). Of the ten most commonly observed SA breakpoints (n=3,086), 78.2% involved whole arms, primarily of chromosomes 9, 1, and 7. SAs, including partial chromosomal arms, most commonly involved breakpoints at 7q21, 7q31, and Xq21. A small but distinct subset of SAs in our study were interstitial in nature (5.9% of deletions, 18.7% of duplications), in contrast to those reported in other studies of segmental abnormalities (
). The most frequent interstitial SA observed was del(5)(p13p14) (n=51). When controlling for arm size, the genomic regions with the highest rates of segmental abnormalities were 8p (15.71 events/Mb), 9q (13.32 events/Mb), and 5p (12.93 events/Mb).
Mosaic-only abnormalities accounted for 10.1% (n=3,862/38,145) of evaluated embryos, with events observed on all chromosomes. Mosaicism rates varied by clinic (range 6.7% to 33.3%). Mosaic-only abnormalities were observed in all egg-age groups, ranging from 5.3% in those aged 40 years or older to 14.2% in those aged 20 to 24 years. Mosaicism involving WCAs were slightly more frequent than that involving only SAs (44.4% vs 37.6%, respectively; Figure 4c). Nearly an eighth (11.6%) of embryos with mosaicism involved both WCAs and SAs.
SNP-based ploidy analysis detected abnormal results in 1.8% of 128,991 embryos (Figure 2a, Figure 5a-c). Abnormal ploidy rates did not vary widely by egg age, ranging from 1.5% to 2.6%. In embryos that were polyploid and for whom the fertilization method was known, 90.9% were derived from ICSI rather than other methods (compared to 90.0% in our overall cohort). ICSI did not appear to specifically influence the rates of triploidy (1.30% with ICSI vs. 1.39% without ICSI; p = 0.4079) or tetraploidy (0.19% with ICSI vs. 0.21% without ICSI; p = 0.6330). Notably, in contrast, a significant difference in the rate of haploidy or whole-genome UPiD was observed between embryos derived from ICSI versus non-ICSI (0.31% vs 0.1%, respectively, p < 0.0001). Among embryos with ploidy abnormalities, triploidy was the most common (73.6%), with haploidy (15.7%) and tetraploidy (10.6%) comprising the rest. Among triploid embryos, the frequency of XXX and XXY sex chromosomes were observed at approximatly equal distributions (42.0% vs 47.8%, respectively; Supplemental Figure 6). Tetraploid embryos most frequently had a sex chromosome complement of XXXX (41.7%) or XXXY (44.8%) (Supplemental Figure 6).
Without SNP-based analysis, 53.9% (1,276/2,369) of the ploidy samples would have appeared diploid using generic count-based analysis due to “normal” sex chromosome ratios. Just over half (56.8%, n=725) of these had additional abnormalities that would have been detected without SNP analysis and classified as aneuploid (n=184) or as mosaic (n=541) (Supplemental Figures 7, 8). The remaining 43.2% (n=551) had no additional abnormalities and therefore would have been incorrectly reported as euploid and considered suitable candidates for embryo transfer.
Single-chromosome UPiD was observed in a small proportion of embryos (47/128,991, 0.04%), with 48 instances (Supplemental Figure 9). However, among the analyzable chromosomes, at least a single instance of UPiD was observed for the majority of chromosomes . UPiD of chromosome 15 was most commonly observed (n=18) (Figure 5d).
Outcome data following PGT-A was requested from referring IVF clinics. Seventeen clinics provided data for 3,297 frozen embryo transfers (FET). Clinical pregnancy rates varied across clinics within a range of 38%-80% (mean 62%). The combined pregnancy plus live birth rate per transfer (813 ongoing pregnancies and 1,057 reported live births) was 57%, with a clinic-specific rate ranging from 38% to 72%. The total loss for the population (biochemical, spontaneous abortion and later term loss) was 9%.DISCUSSION
The FAST-SeqS method modified for PGT-A combines the benefits of WGA-NGS- (coverage) and SNP-based (ploidy) PGT-A assays, matching or exceeding the performance of these alternative methods. FAST-SeqS has been shown to accurately detect WCAs, SAs ≥10Mb, most forms of polyploidy, and select whole-chromosome UPiDs. Additionally, similar findings from previous validation studies at multiple locations demonstrate the reproducibility and robustness of the FAST-SeqS technology (
). Furthermore, our clinical validation strategy, which utilized distinct biopsies rather than a single WGA product, aimed to mimic real-world conditions.
FAST-SeqS is an internally developed assay that minimizes the need to purchase commercial kits. Additionally, FAST-SeqS allows for the inclusion of more samples per run (higher multiplexing). The automation and fewer steps involved prior to sequencing allows FAST-SeqS to be performed at a relatively low cost per sample (∼75% less than VeriSeq based on our internal experience). These advances provide a clinically robust and affordable PGT-A option that significantly improves access for more individuals who can benefit from PGT-A as part of their IVF journey.
In one of the largest studies of PGT-A results, FAST-SeqS yielded clinical results comparable to those reported in studies using other technologies. WCA and SA rates and associations with egg age align with previous reports (
). The average clinical pregnancy rate across all ages was higher than the average FET-specific outcome without PGT-A (47.3%) and comparable to that with PGT-A (64.4%) described in the National Summary Report issued by the Society for Assisted Reproductive Technology (SART) in 2017. The differences in outcomes from PGT cycles are likely due to multiple variables, such as multiple technologies, sample sizes, and clinical indications. Additionally, clinical protocols, laboratory technology, and procedural competency contribute to pregnancy and loss rates. The clinical outcomes described in this study are consistent with studies using other PGT-A platforms (
). Together, these data establish FAST-SeqS as a reliable and accurate assay for PGT-A that can be incorporated within IVF workflows in clinics to improve pregnancy outcomes.
The size of the cohort provides a high degree of confidence in age-related euploid rates and provides estimates of how often at least one embryo in an IVF cycle is euploid based on age and number of embryos for biopsy. These observations can help inform clinicians and genetic counselors who provide likelihood estimates of embryo transfer following PGT-A to patients. Given the consistent rates of observed abnormalities with previous reports, these data could be applicable for all patients, regardless of PGT-A platform.
The mosaic-only aneuploidy rate was 10.1% and relatively consistent across egg age. Published mosaic rates vary widely, ranging up to 30% (
). Our observed rate is likely due to our method of reporting, in which embryos with mosaic-only abnormalities are classified as mosaic, rather than reporting mosaicism for each individual abnormality. Future analyses defining mosaic embryos with WCA versus SA would provide more knowledge to the field, especially in light of recent studies demonstrating an increased likelihood of “rescue” of mosaic SAs in comparison to mosaic WCAs (
Transfer outcomes of embryos with preimplantation genetic testing for aneuploidy (PGT-A) diagnosis of undetermined reproductive potential: Results from a prospective, blinded, multi-center non-selection study.
Our observed SA rates in blastocysts are consistent with previous reports (e.g., most SAs involved whole chromosome arms, with larger chromosomes having more segmental events). However, we also detected interstitial SAs, demonstrating that these do occur though, to our knowledge, they have not been reported previously in studies reporting PGT-A findings. This may be due to differences in resolution and reporting thresholds. While large SAs account for 4-6% of chromosomally abnormal miscarriages (
), smaller SAs, including the most frequently seen in our cohort, are associated with known syndromes that cause birth defects and intellectual disability in a liveborn (Xq21 [Turner syndrome phenotype], 4p15 [Wolf-Hirschhorn syndrome], and 1p36 [1p36 deletion syndrome]), demonstrating that detecting SAs are important for clinical decision making.
More than 1,000 embryos with ploidy abnormalities could have been incorrectly classified as euploid or mosaic using traditional NGS-based technologies. Had these embryos been selected for transfer, a miscarriage or molar pregnancy would have been the outcome. In addition, no other studies have analyzed the rate of tetraploidy, which we found to be 0.20%; however, this number is likely an underestimate, as not all forms of tetraploidy are identified by this assay. Correctly assigning ploidy status is imperative to preventing transfer of embryos that not only are inviable but also may have detrimental health implications for a pregnant woman (molar pregnancy).
A general limitation of PGT-A is that it is a screen based on the analysis of 4-10 cells from the TE (not the inner cell mass), but considered diagnostic for the whole embryo by many clinicians and patients. “Misdiagnoses” are typically not due to technical error, but likely attributed to biological factors (i.e., mosaicism) or, rarely, human error. Recent studies have shown that if there is a discrepancy between the TE and inner cell mass, it is more likely to be for SAs than for WCAs, likely due to their post-zygotic origin (
), the FAST-SeqS technology cannot provide exact breakpoints for SAs or results at the gene level. From a clinical perspective, precise base-pair level SA breakpoint information is not needed to determine the presence of a gain or loss. In addition, they are often not necessary when making a decision regarding transfer, due to the size (≥10Mb). While the overall cohort was very large, the cohorts of embryos evaluated for mosaicism and abnormal ploidy were relatively small. Additional studies in larger cohorts will help to more accurately determine these rates within the PGT-A population. The outcome data reported here is limited. As outcome data is the best way to truly determine the benefits and accuracy of PGT analysis, prospective studies collecting IVF outcome data following PGT-A across clinical indications will provide additional insights into the clinical utility of this assay. Additionally, as with all PGT assays, we recognize the best way to determine the true clinical utility of FAST-SeqS would be a randomized control trial. While not feasible during commercial validation, a future non-selection study would provide more evidence towards this end. Lastly, although there are several benefits of FAST-SeqS vs WGA-based PGT, a downside to not using WGA is that PGT for monogenic disorders (PGT-M) cannot be performed on the same sample as PGT-A. Therefore, for the small subset of patients who require PGT-M, a separate biopsy sample would be required.
PGT-A studies have demonstrated shortened time to pregnancy and improved patient outcomes. PGT-A itself, and certainly in addition to IVF, has historically been very expensive and self-funded, and therefore out of reach for many. Even in instances of insurance coverage, individuals have often still faced significant out-of-pocket expenses. To overcome these barriers, especially in poor responders or advanced maternal age, clinicians often recommend “banking and batching” embryos over multiple IVF cycles and testing all embryo biopsies at once to avoid the need to pay for multiple PGT-A analyses (
). However, this approach may lengthen the time to pregnancy as multiple cycles are required to reach some threshold number of embryos before PGT-A. Since FAST-SeqS can be performed at a lower cost compared to other assays, its introduction has been instrumental in encouraging clinicians and their patients not to batch, thereby reducing time to a successful pregnancy with a euploid embryo. As technology advances and costs drop, coupled with progress towards IVF insurance coverage, the practice of banking and batching embryos may no longer become a necessity simply because of cost.
Our data has demonstrated that a modified FAST-SeqS assay is an accurate, reliable PGT technology that can detect chromosomal abnormalities relevant to decision making for embryo transfers, and it also combines the benefits of WGA-NGS-based (mosaicism detection) and SNP-based (ploidy detection) methods. Other methods without the capability to detect these types of abnormalities could result in the transfer of embryos that do not actually have a normal chromosome complement. The FAST-SeqS platform demonstrates that advancing technology can provide an accurate, scalable, and automated PGT-A assay that, as a result, may widen accessibility.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Declaration of Interest
All authors are employees and stockholders of Invitae Corporation.
We thank Betty Abelev, Eric Boyden, Ben Breton, Natasa Dzidic, Katilyn Germain, Jeff Gole, Athurva Gore, Karine Hovanes, Katya Kosheleva, Xu Li, Tom Mullen, Greg Porreca, Kristina Robinson, Trilochan Sahoo, Eric Tsung, Mark Umbarger, and Mei Zhu for contributions toward validating the FAST-SeqS assay.
Transfer outcomes of embryos with preimplantation genetic testing for aneuploidy (PGT-A) diagnosis of undetermined reproductive potential: Results from a prospective, blinded, multi-center non-selection study.
Dr. Walters-Sen is board-certified in clinical cytogenetics and clinical molecular genetics by the American Board of Medical Genetics and Genomics. She completed her fellowships at Nationwide Children's Hospital and received her doctorate in human genetics from Vanderbilt University. Dr. Walters-Sen has an interest in prenatal and early postnatal genetic diagnostics.