This retrospective experimental trial included 207 oocyte retrieval cycles and 124 cycles of single vitrified-warmed blastocyst transfer performed between August 2020 and March 2023 at a single in vitro fertilization (IVF) center (Kinutani Women’s Clinic). A total of 557 blastocysts underwent PGT-A. All embryos were monitored using a time-lapse system from fertilization to the blastocyst stage. However, cases exhibiting structural chromosomal rearrangements on pre-implantation genetic testing and multiple embryo transfers, including two-step embryo transfers, were excluded from the analysis. The main outcome measures were clinical pregnancy (presence or absence of a gestational sac) and live birth. This study was approved by the Ethics Committee of Kinutani Women’s Clinic (certification number: 17000119). The approval number assigned to this study by the Institutional Review Board was 20191120. All procedures were conducted in accordance with the ethical standards of the institutional research committee and the 1964 Helsinki Declaration and its later amendments. Informed consent was obtained from all patients who participated in the study.

Standard regimens of controlled ovarian hyperstimulation were employed based on established clinical protocols. Patients were treated with a gonadotropin-releasing hormone (GnRH) agonist using a short or long protocol following progestin-primed ovarian stimulation, in accordance with each patient’s ovarian response and medical history of IVF treatment. Serum estradiol levels were monitored, and follicular growth was evaluated by transvaginal ultrasonography. In some cycles, a minimal stimulation protocol or a natural cycle was used. When a follicle reached a mean diameter ≥18 mm, human chorionic gonadotropin (Aska Pharmaceutical), 250 g of choriogonadotropin alpha (Ovidrel; Merck Serono), or a GnRH agonist was administered. Transvaginal oocyte retrieval was performed 35 to 37 hours after administration.

The retrieved cumulus–oocyte complexes were cultured for approximately 4 hours in G-IVF medium (Vitrolife) at 37 °C under an atmosphere of 6% CO₂, 5% O₂, and 89% N₂ before conventional IVF (c-IVF) or intracytoplasmic sperm injection (ICSI). For c-IVF, sperm was prepared at a concentration of 7.5–10×10⁶/mL. A total of 5 to 50 μL of sperm was added to the dish containing the oocytes. The sperm and oocytes were then co-cultured at 37 ºC under 6% CO₂, 5% O₂, and 89% N₂ for approximately 18 to 20 hours. For ICSI, cumulus–oocyte complexes were denuded via pipetting with 80 IU/mL hyaluronidase (Irvine Scientific). The maturity of denuded oocytes was assessed by visualization of the first polar body. Only metaphase II oocytes were used for ICSI. Embryos fertilized by c-IVF or ICSI were cultured in DNP dishes (Dai Nippon Printing Co. Ltd.) with SAGE™ 1-Step (Origio) at 37 °C under an atmosphere of 6% CO₂, 5% O₂, and 89% N₂ for 5 to 7 days and evaluated for development to the blastocyst stage.

Trophectoderm (TE) biopsy was conducted on expanded blastocysts on days 5 to 7, regardless of morphological grade. First, artificial shrinkage was performed using a laser system (Zilos-TK; Hamilton Thorne Biosciences Inc.) to create a space between the zona pellucida and TE cells. While securing the blastocyst with the inner cell mass (ICM) positioned at approximately 8 to 9 o’clock, serial laser pulses were used to create a small hole (approximately 10 µm) in the zona pellucida at a position of approximately 3 to 4 o’clock. A biopsy pipette was then inserted into the hole created by the laser to aspirate the TE cells. Five to 10 TE cells were aspirated into the biopsy pipette and collected by flicking the holding pipette and biopsy pipette with the aid of several laser pulses. Biopsy samples were processed for whole-genome amplification and NGS at the Varinos Laboratory using the VeriSeq PGS kit (Illumina K.K.) on a MiSeq system (Illumina K.K.) in 24-sample runs, according to the manufacturer’s protocol. The copy number of each sample was analyzed using BluFuse Multi Software (Illumina K.K.). A molecular karyotype profile consistent with mosaicism was defined when a whole chromosome or chromosomal segment showed intermediate copy number levels of 20% to 80% between whole-number states, according to the guidelines of the Preimplantation Genetic Diagnosis International Society [24]. The mosaic embryo group was subclassified into low-level (between 20% and 49% abnormal cells) and high-level (between 50% and 80% abnormal cells) mosaicism. Mosaic embryos were further subgrouped into whole chromosome (loss or gain), segmental, and complex mosaic embryos. When mosaicism was present on more than one chromosome, the chromosome with the highest level of mosaicism was assigned as the mosaic type.

Blastocyst vitrification was performed as reported by Hiraoka et al. [25] within 1 hour after embryo biopsy. Blastocysts were placed in an equilibration solution containing 7.5% (v/v) ethylene glycol (Sigma-Aldrich) and 7.5% (v/v) dimethyl sulfoxide (Nacalai Tesque Inc.) in modified human tubal fluid (mHTF; Irvine Scientific) with 20% (v/v) serum protein substitute (SPS; Origio) at 37 °C. Blastocysts were then transferred into a vitrification solution containing 15% (v/v) ethylene glycol, 15% (v/v) dimethyl sulfoxide, and 0.5 M sucrose (Sigma-Aldrich) in mHTF with 20% SPS for 1 minute at 37 °C. Blastocysts were loaded onto Cryotops (Kitazato Corporation) in a minimal volume and immediately immersed in liquid nitrogen at –196 °C. For the warming protocol, the tip of the Cryotop was immersed directly in 1.0 M sucrose solution for 1 minute at 37 °C. Blastocysts were then transferred to 0.5 M sucrose solution for 3 minutes and washed twice in mHTF with 20% SPS for 5 minutes at 37 °C. Prior to embryo transfer, the warmed blastocyst was cultured for approximately 3 to 4 hours in EmbryoGlue (Vitrolife). Luteal phase support was started using intravaginal micronized progesterone or in natural cycles. Embryo transfers were performed under ultrasound guidance using a soft catheter (Origio).

Oocytes were cultured individually in a time-lapse system (CCM-iBIS; Astec). A total of 1 to 25 embryos were placed in a 100 µL drop of medium, and images of the embryos were recorded automatically at 15-minute intervals. Kinetic timing and compaction dynamics were annotated by two well-trained embryologists, who were blinded to the PGT-A results, using a Linkid image analyzer (Astec). The exact timing of two events was annotated as follows: completion of the compaction process (tM) and formation of a full blastocyst (tB). All timing parameters were expressed in hours after the fading of the pronucleus. Compaction was observed and classified into two patterns: ‘whole,’ indicating compaction through normal development, and ‘partial,’ indicating blastomeres left behind during compaction. In addition, an instrument was used to measure the diameter of the blastocyst at the time of embryo biopsy.

Statistical analysis was performed using JMP ver. 14.0 (SAS Institute Inc.) and GraphPad Prism ver. 6.03 (GraphPad Inc.). Patient characteristics were presented as mean and standard deviation. Proportions were analyzed using the Fisher exact probability test or the chi-square test to determine statistical differences. Continuous variables were compared using the Student t-test or one-way analysis of variance, and significance was determined using the Tukey–Kramer test. As the time-lapse data were not normally distributed, we used the Wilcoxon rank-sum test or Kruskal–Wallis test, and significance was determined using the Dunn test. To adjust for the potential confounding effect of maternal age on developmental kinetics, analysis of covariance (ANCOVA) was performed with maternal age included as a covariate. Differences were considered statistically significant at p<0.05.

Table 1 shows the participant characteristics for oocyte retrieval cycles and NGS PGT-A results. Overall, 687 blastocysts were obtained from 207 oocyte retrieval cycles. Of these 687 blastocysts, 641 were biopsied and 578 were analyzed by NGS. The results revealed that 105 embryos were euploid and 364 were aneuploid. There were 88 embryos with mosaicism, categorized as low- or high-level according to the proportion of abnormal cells detected. All NGS-tested samples were subjected to ploidy analysis, and 16 embryos were found to have abnormal ploidy (haploidy or triploidy). In addition, 0.9% (five embryos) had inconclusive or failed NGS results. Table 2 shows the characteristics of embryo transfer participants. A total of 124 single vitrified-warmed blastocyst transfers were performed with euploid and mosaic embryos, primarily under hormone replacement cycles. Of these 124 cycles, 60 clinical pregnancies were achieved, 50 of which resulted in live births.

The characteristics of PGT-A embryos observed by time-lapse are shown in Table 3. Maternal age in the euploid group was significantly lower than in the mosaic and aneuploid groups (37.3±3.79 years vs. 39.1±3.31 years vs. 40.6±3.08 years, respectively; p<0.05). However, paternal age was highest in the aneuploid group and did not differ significantly between the euploid and mosaic groups (38.8±4.92 years vs. 38.9±5.00 years vs. 42.1±5.34 years, respectively; p<0.05). Time to completion of compaction did not differ significantly among the three groups. Time to blastocyst development was earliest in the euploid group compared with the mosaic and aneuploid groups (83.3 hours [range, 69.0–111.7] vs. 87.3 hours [range, 71.5–120.8] vs. 88.2 hours [range, 71.0–139.0], respectively; p<0.05). The percentage of good-quality embryos decreased significantly in the order of euploid, mosaic, and aneuploid (61.9% vs. 43.2% vs. 25.8%, p<0.05), while the percentage of poor-quality embryos was highest in the aneuploid group (15.2% vs. 25.0% vs. 33.2%, respectively; p<0.05). ANCOVA was performed with maternal age as a covariate to adjust for potential age-related effects on blastocyst development. The effect of ploidy remained statistically significant (F=7.53, p=0.0006), whereas maternal age showed no significant association (F=1.90, p=0.168). These results indicate that differences in blastocyst development timing among euploid, mosaic, and aneuploid embryos are independent of maternal age. Figure 1 shows the timing of blastocyst development according to mosaic level. No significant difference was noted in the timing of compaction among the mosaic groups (euploid, 63.8 hours [range, 45.6–97.7]; low-level mosaic, 66.2 hours [range, 39.1–88.0]; high-level mosaic, 64.4 hours [range, 49.8–90.3]; aneuploid, 66.5 hours [range, 33.3–105.2]; not significant) (Figure 1A). Blastocyst development was significantly faster in the euploid group than in the aneuploid group (83.3 hours [range, 69.0–111.7] vs. 88.2 hours [range, 71.0–139.0]; p<0.01) (Figure 1B). However, no significant difference was found in the timing of blastocyst development between low- and high-level mosaicism (87.3 hours [range, 72.9–120.8] vs. 89.1 hours [range, 71.5–118.4], respectively; not significant) (Figure 1B). Mosaic embryos exhibited blastocyst development patterns intermediate between euploid and aneuploid embryos, but no significant differences were observed between the mosaic subgroups after subdivision by mosaicism level. Overall, differences in developmental kinetics and morphological grades were evident between the euploid and aneuploid embryos, whereas the kinetics of mosaic embryos resembled those of aneuploid embryos.

The PGT-A results for 124 cycles, including 71 mosaic embryos, were classified by mosaic type and level. Figure 2 shows the clinical outcomes after single vitrified-warmed blastocyst transfer stratified by mosaic type and level. Embryos with low-level mosaicism were classified into four types and analyzed for clinical outcomes. Pregnancy and live birth rates were highest for embryos with segmental mosaicism, but no statistically significant differences were found among mosaic types (Figure 2A, 2B). High-level mosaic embryos were separated into four categories, and the clinical pregnancy and live birth rates of these categories were compared (Figure 2C, 2D). Clinical pregnancies were observed even in mosaic embryos with high degrees of chromosomal mosaicism. The live birth rate for embryos with high-level chromosome loss (monosomy) was 37.5% (3/8). However, no full-term live births were recorded in the group showing high-level chromosomal gain (trisomy) or segmental aneuploidy. Clinical outcomes differed depending on whether the mosaicism involved monosomy or trisomy (Figure 2C, 2D).

Table 4 compares neonatal outcomes after single vitrified-warmed blastocyst transfer according to mosaic group. Neonatal outcomes were divided into three groups: euploid, low-level mosaicism, and high-level mosaicism. No significant differences were observed in clinical pregnancy rates among these groups. Miscarriage rates were significantly higher in the low-level and high-level mosaic groups than in the euploid group (euploid, 0% [0/27]; low-level mosaic, 33.3% [5/25]; high-level mosaic, 62.5% [5/8]; p<0.05). The live birth rate in the high-level mosaic group was 17.6% (3/17), significantly lower than that of the fully euploid embryos (52.9% [27/51]). No significant differences in beta-human chorionic gonadotropin levels or maternal age were found. Furthermore, no significant differences in mean birth weight were noted among the groups (euploid, 3,117±373.5 g; low-level mosaic, 3,073±502.0 g; high-level mosaic, 2,621±730.9 g).