A Prospective Study on Clinical Utility of Preimplantation Genetic Testing for Aneuploidy and Structural Rearrangement Under Advanced Medical Care B in Japan

The protocol was approved by the Ethical Review Board of the University of Osaka Hospital (approval No. 15000038; 13 January 2023).

The authors received no financial support for the research, authorship, or publication of this article.

Between April 2023 and January 2024, 377 couples provided written informed consent (Figure  1 ). After exclusion of two ineligible couples, 375 participants were registered. Of these, 344 underwent PGT‐A/SR. Among these 344 patients, 260 (75.6%) obtained at least one category A or B embryo, and 218 (63.4%) women proceeded to a first frozen–thawed single blastocyst transfer with a category A or B embryo, constituting the embryo‐transfer cohort for the prespecified per‐transfer analyses. The remaining 126 patients (36.6%) did not proceed to embryo transfer: 29 (8.4%) discontinued after PGT‐A/SR despite obtaining a category A/B embryo, 84 (24.4%) did not obtain a category A or B embryo, 13 (3.8%) obtained a category A/B embryo but did not reach transfer for other reasons. When the embryo‐transfer cohort reached 153 patients, the null value recalculated based on the age distribution of enrolled participants was 42.2%, and the corresponding statistical power was 89.2%. Because the recalculated power exceeded 80%, sample size re‐estimation was deemed unnecessary. Flowchart of patient enrollment, PGT‐A/SR procedures, and embryo transfer in the Advanced Medical Care B study. Study flow diagram showing the number of patients at each stage from consent through registration, PGT‐A/SR, and embryo transfer, together with reasons for exclusion or discontinuation at each step. *All 29 patients had obtained at least one category A/B embryo but did not proceed to embryo transfer for reasons unrelated to embryo availability. PGT‐A/SR, preimplantation genetic testing for aneuploidy and structural rearrangements. Baseline characteristics are summarized in Tables  1 and 2 . In the PGT‐A/SR cohort ( n  = 344), indications overlapped and included RIF in 226 patients (65.7%), RPL in 100 (29.1%), and structural chromosomal rearrangements in 24 (7.0%). Median female age was 38.0 years (interquartile range [IQR], 35.0–40.0). A total of 1253 blastocysts were biopsied, of which 532 (42.5%) were classified as category A, 163 (13.0%) as category B, 557 (44.5%) as category C, and 1 (0.1%) as category D. In the embryo‐transfer cohort ( n  = 218), indications for PGT‐A/SR were RIF in 139 women (63.8%), RPL in 65 (29.8%), and structural chromosomal rearrangements in 17 (7.8%). Median female age was 37.0 years (IQR, 33.0–39.0). Endometrial preparation was performed in a natural cycle in 95 cycles (43.6%) and in a hormone replacement cycle in 123 (56.4%). The transferred blastocysts were graded as good (Gardner AA/AB/BA) in 81 cases (37.3%), moderate (BB/AC/CA) in 98 (45.1%), and poor (BC/CB/CC) in 39 (18.0%). Most transfers (198/218, 90.8%) used category A embryos, whereas 20 (9.2%) used category B embryos. Median endometrial thickness at transfer was 10.0 mm (IQR, 8.8–11.7). Baseline characteristics (PGT‐A/SR cohort). Abbreviations: ART, assisted reproductive technology; BMI, body mass index; IQR, interquartile range; PGT‐A/SR, preimplantation genetic testing for aneuploidy and structural rearrangements. Some patients met more than one category. When only the month of infertility onset was missing, June of that year was imputed. Number of prior unsuccessful embryo transfers was defined as the number of embryo transfer cycles that resulted in either a negative hCG test or only a biochemical pregnancy. Baseline characteristics (embryo‐transfer cohort). Abbreviations: ART, assisted reproductive technology; BMI, body mass index; hCG, human chorionic gonadotropin; IQR, interquartile range; PGT‐A/SR, preimplantation genetic testing for aneuploidy and structural rearrangements. Some patients met more than one category. When only the month of infertility onset was missing, June of that year was imputed. Number of prior unsuccessful embryo transfers was defined as the number of embryo transfer cycles that resulted in either a negative hCG test or only a biochemical pregnancy. The primary outcome, oPR, was 48.6% (106/218; 95% CI, 42.0–55.3) in the PGT‐A/SR cohort (Table  3 ). The age‐adjusted cPR expected from the JSOG ART registry in the absence of PGT‐A/SR was 42.4%. The primary comparison did not reach conventional statistical significance ( p  = 0.064). Because the registry benchmark was based on cPR rather than oPR, this finding should be interpreted cautiously as a conservative and asymmetric benchmarking result, rather than as a direct comparative efficacy analysis. When the cPR in the PGT‐A/SR cohort was compared with the same JSOG reference, in a prespecified supportive analysis on the same time scale as the benchmark, the cPR was 57.8% (126/218; 95% CI, 51.2–64.4) versus 42.4% in the registry, demonstrating a significantly higher clinical pregnancy rate with PGT‐A/SR ( p  < 0.001; Table  3 ). The miscarriage rate per clinical pregnancy was 15.9% (20/126; 95% CI, 9.5–22.3), which was significantly lower than the age‐adjusted miscarriage rate of 23.4% estimated from the JSOG registry ( p  = 0.045; Table  3 ). The implantation rate per transfer, including biochemical pregnancies, was 66.1% (144/218; 95% CI, 59.8–72.3; Table  3 ). To distinguish per‐transfer outcomes from patient‐level clinical utility, we also summarized outcomes among all 344 patients who underwent PGT‐A/SR. On this patient‐level basis, the clinical pregnancy rate was 36.6% (126/344) and the ongoing pregnancy rate was 30.8% (106/344). In a prespecified supportive analysis at the planned sample size (153 transfers), 82 women had an ongoing pregnancy (oPR, 53.6% [82/153; 95% CI, 45.7–61.5]), which was significantly higher than the threshold value ( p  = 0.043) (Table  S2 ). In the same cohort, 93 women achieved a clinical pregnancy, yielding a cPR of 60.8% (93/153; 95% CI, 53.0–68.5), which also significantly exceeded the threshold ( p  < 0.001) (Table  S2 ). To explore the attenuation of oPR after the planned sample‐size analysis, we compared baseline and cycle characteristics between the first 153 transfers and the additional 65 transfers (Table  S3 ). No statistically significant differences were observed in the analyzed variables. However, reproductive‐history variables showed borderline or numerical imbalances, including prior clinical pregnancies ( p  = 0.052) and prior miscarriages or stillbirths ( p  = 0.065), both of which were higher in the additional 65 transfers. Pregnancy outcomes in the embryo‐transfer cohort. Note: * p  < 0.05, ** p  < 0.001. Clinical pregnancy rate was prespecified as a supportive analysis because the registry benchmark was based on clinical pregnancy rather than ongoing pregnancy. Abbreviation: CI, confidence interval. At the visit between 12 + 0 and 13 + 6 weeks of gestation. At the time of gestational sac confirmation. Age‐stratified analyses showed that the favorable association of PGT‐A/SR with oPR and cPR was most evident in women aged ≥ 35 years. For oPR, the observed rates in the PGT‐A/SR cohort versus the age‐adjusted JSOG clinical pregnancy rates were 47.0% (31/66) versus 53.9% ( p  = 0.258) in women < 35 years, 53.0% (53/100) versus 42.6% ( p  = 0.036) in women aged 35–39 years, and 42.3% (22/52) versus 27.5% ( p  = 0.017) in women aged 40–42 years (Table  4 ). For cPR, the corresponding values were 60.6% (40/66) versus 53.9% ( p  = 0.275), 60.0% (60/100) versus 42.6% ( p  < 0.001), and 50.0% (26/52) versus 27.5% ( p  < 0.001), respectively (Table  5 ). Thus, both oPR and cPR were significantly higher than registry‐based expectations in women aged 35–39 and 40–42 years, but not in those aged < 35 years. Subgroup analysis of ongoing pregnancy rate at 12 + 0–13 + 6 weeks in the embryo‐transfer cohort. Note: * p  < 0.05. For non‐age subgroups, results were interpreted descriptively by comparing the 95% CI with the overall registry benchmark value used in the primary analysis (42.4%; Table  3 ); no p ‐values are provided for these comparisons. Abbreviations: CI, confidence interval; PGT‐A/SR, preimplantation genetic testing for aneuploidy and structural rearrangements. Some patients met more than one category. Subgroup analysis of clinical pregnancy rate at gestational sac confirmation in the embryo‐transfer cohort. Note: ** p  < 0.001. For non‐age subgroups, results were interpreted descriptively by comparing the 95% CI with the overall registry benchmark value used in the primary analysis (42.4%; Table  3 ); no p ‐values are provided for these comparisons. Abbreviations: CI, confidence interval; PGT‐A/SR, preimplantation genetic testing for aneuploidy and structural rearrangements. Some patients met more than one category. For non‐age subgroups, oPR and cPR were interpreted descriptively against the age‐adjusted registry benchmark used in the primary analysis (42.4%; Table  3 ) by assessing whether the lower bound of the 95% CIs exceeds this value. For oPR (Table  4 ), only the good‐quality embryo subgroup had a 95% CI entirely above 42.4% (58.0% [47.3–68.8]); the RIF (49.6% [41.3–58.0]) and category A embryo (48.5% [41.5–55.4]) subgroups narrowly overlapped the benchmark, and all other estimates overlapped it. For cPR (Table  5 ), the 95% CIs for RIF (59.7% [51.6–67.9]), good‐quality embryos (71.6% [61.8–81.4]), and category A embryos (58.1% [51.2–65.0]) lay entirely above 42.4%, whereas the moderate‐quality embryo subgroup narrowly overlapped (51.0% [41.1–60.9]) and other estimates overlapped the benchmark. For miscarriage rate per clinical pregnancy, the overall registry benchmark was 23.4% (Table  3 ). The moderate‐quality embryo (12.0% [3.0–21.0]), category A embryo (16.5% [9.7–23.3]), and 35–39‐year (11.7% [3.5–19.8]) subgroups had upper 95% CI bounds below 23.4% (Table  6 ). By contrast, the RIF subgroup narrowly overlapped the benchmark (16.9% [8.8–24.9]), and the good‐quality embryo subgroup also did not exclude it (19.0% [8.9–29.1]); other subgroup estimates similarly overlapped 23.4%. Implantation rates were broadly similar across infertility indications and between category A and B embryos. In contrast, implantation rates decreased across embryo grade (good 77.8%, moderate 62.2%, poor 51.3%) (Table  S4 ). Subgroup analysis of miscarriage rate at 12 + 0–13 + 6 weeks in the embryo‐transfer cohort. Abbreviations: CI, confidence interval; PGT‐A/SR, preimplantation genetic testing for aneuploidy and structural rearrangements. Some patients met more than one category. Miscarriage rate was calculated as the number of miscarriages divided by the number of clinical pregnancies in each subgroup. For non‐age subgroups, results were interpreted descriptively by comparing the 95% CI with the overall registry benchmark value for miscarriage rate per clinical pregnancy (23.4%; Table  3 ); no p ‐values are provided for these comparisons.

This Advanced Medical Care B study is the first to provide prospectively monitored multicenter benchmarking data on PGT‐A/SR in Japan under the Japanese indication framework in place at the time of the study. In the primary analysis, oPR at 12–13 weeks did not significantly exceed the age‐adjusted registry benchmark ( p  = 0.064). Exploratory comparison of baseline and cycle characteristics between the first 153 transfers and the additional 65 transfers did not identify any statistically significant imbalance, although a slightly greater reproductive‐history burden in the additional 65 transfers may have contributed in part to the attenuation of oPR (Table  S3 ). No protocol, procedural, or outcome‐ascertainment change was identified that would explain the attenuation of oPR; however, this post hoc comparison cannot exclude random variation or unmeasured case‐mix differences. Nevertheless, prespecified supportive analyses showed higher cPR and lower miscarriage rates than the registry benchmarks on a per‐transfer basis. Because this was a single‐arm study using historical registry‐derived benchmarks rather than a contemporaneous non‐PGT control group, these findings indicate association rather than direct causal superiority of PGT‐A/SR over standard care. In age‐stratified analyses, women aged ≥ 35 years showed more favorable associations than registry‐based expectations, whereas no benefit was demonstrated in women aged < 35 years. These findings suggest that any potential benefit of PGT‐A/SR may be concentrated in older patients, but they should be interpreted cautiously within the limits of a benchmarking design. The unconventional choice of endpoints and comparators in this study reflects the regulatory and health‐system context in Japan rather than a purely academic design preference. Following the April 2022 revision of the national fee schedule, many infertility treatments became covered by public health insurance; however, PGT‐A/SR was not included despite prior use as a self‐funded service. Because mixed billing is prohibited, patients who undergo PGT‐A/SR must pay the full cost of the treatment cycle out of pocket; demonstrating benefit under realistic Japanese practice was therefore a prerequisite for considering public coverage. In addition, several groups were preparing similar Advanced Medical Care B protocols using different commercial PGT platforms, so a simple, reproducible design that could be implemented rapidly at multiple centers was required. For these pragmatic reasons, we were compelled to use oPR at 12–13 weeks as the primary outcome for the PGT‐A/SR cohort and age‐adjusted cPR from the JSOG registry as the null comparator. We also prespecified cPR as a supportive outcome on the same scale as the registry benchmark, because live birth rates could not feasibly be assessed in both groups within this study framework. However, this policy context does not remove the interpretive limitation created by the mismatch in outcome definitions and the use of historical registry benchmarking. This approach created a structurally disadvantageous comparison for PGT‐A/SR, because miscarriages occurring between gestational sac confirmation and 12–13 weeks reduce oPR but are not reflected in the registry‐derived cPR. Accordingly, the cPR results should be read as supportive rather than corrective evidence, and the oPR result remains the primary finding. Within these constraints, the overall pattern of cPR and miscarriage results was directionally favorable on a per‐transfer basis. Beyond the asymmetry of the primary comparison, residual pregnancy loss between clinical pregnancy and 12–13 weeks in the PGT‐A/SR cohort likely attenuated the observed oPR. Although the miscarriage rate after gestational sac confirmation was significantly lower in the PGT‐A/SR cohort than in the age‐adjusted JSOG registry benchmark, 15.9% of clinical pregnancies still ended in loss by 12–13 weeks in this study. This rate is lower than that reported for non‐PGT pregnancies in some cohorts, such as approximately 28% between 6 and 12 weeks in a U.S. IVF population [ 24 ], but markedly higher than the 5%–6% loss reported in a Chinese randomized trial of PGT‐A in younger women [ 13 ]. One possible explanation for the residual pregnancy loss is the relatively high proportion of poor‐quality blastocysts transferred in this study (18%), reflecting a patient population highly motivated to biopsy and transfer embryos even when morphology was suboptimal. However, in exploratory subgroup estimates (Table  6 ), the poor‐quality embryo subgroup did not show a miscarriage rate above the registry benchmark. Taken together, these observations may suggest that residual pregnancy loss in our cohort was not confined to poor‐quality embryos and that non‐embryonic factors may have contributed in at least a subset of patients. Consistent with this interpretation, a systematic review reported lower live birth rates after euploid embryo transfer in patients with RPL than in those without RPL [ 25 ]. In our cohort, enrichment of patients with multiple prior failed transfers and/or RPL—more prognostically unfavorable than the general ART population captured in the JSOG registry—may also have contributed to the residual miscarriage risk despite PGT‐A/SR. In addition, the apparently higher miscarriage rate observed in good‐quality embryos than in moderate‐quality embryos is difficult to interpret mechanistically, because subgroup sizes were small and confidence intervals overlapped substantially (Table  6 ). Importantly, the biological interpretation of these residual losses remains limited because chromosomal analysis data of products of conception (POC) were not systematically available in this multicenter study; therefore, the currently available data in this study do not allow definitive distinction between embryonic, maternal/non‐embryonic, and possible technical factors. In addition to the higher‐than‐expected early loss described above, two further considerations may help interpret the observed outcomes. First, compared with the earlier Japanese pilot study used to inform assumptions [ 22 ], our Advanced Medical Care B cohort had broader eligibility criteria and fewer exclusions: for example, in RPL, pilot studies rigorously screen for and exclude non‐embryonic contributors such as uterine anomalies or antiphospholipid syndrome, whereas our study did not exclude patients based on the presence of such factors. Consequently, some participants may have had maternal or uterine conditions that reduce the likelihood of pregnancy continuation even after transfer of euploid embryos. This may partly explain why the cPR observed in our study (57.8%) was lower than the pilot‐study cPR (66.7%), and also suggests that our study design—parameterized based on the pilot benchmark—posed a statistically demanding target. Second, because all transfers in our cohort were single‐embryo transfers, rates “per transfer” effectively reflect rates “per embryo,” whereas the JSOG registry aggregates outcomes from practice that may include multiple‐embryo transfers. This could inflate registry‐based “per‐transfer” pregnancy rates and thereby render our comparisons more conservative. Overall, these factors may indicate that our results reflect real‐world performance in a more severe population under a stringent benchmarking approach. Our results are consistent with the emerging consensus that PGT‐A/SR should not be used indiscriminately but may be beneficial in selected high‐risk groups, particularly women of advanced maternal age. The ESHRE good practice recommendations on add‐ons emphasize that current evidence does not support routine PGT‐A for all ART patients, that patient‐level benefit has not been consistently demonstrated in randomized trials, and that time‐to‐pregnancy and cost remain clinically important unresolved issues [ 18 ]. The HFEA classifies PGT‐A as an “add‐on” that does not improve live birth rates for most patients but may reduce miscarriage in some groups [ 17 ]. The 2024 ASRM committee opinion similarly concludes that routine PGT‐A for all ART patients is not supported by current evidence, while acknowledging possible advantages for older women and certain high‐risk indications [ 16 ]. ESHRE guidance and consensus statements from the Indian Society for Assisted Reproduction (ISAR) also recommend restricting PGT‐A to indications such as RIF, RPL, and advanced maternal age, rather than offering it routinely to young patients with good prognosis [ 15 , 26 , 27 ]. For couples carrying balanced chromosomal rearrangements, ESHRE further considers PGT‐SR an appropriate option, and a recent systematic review reported that although cumulative live birth rates with PGT‐SR are similar to those with natural conception, miscarriage rates are significantly reduced, supporting PGT‐SR as a useful strategy to mitigate the psychological and physical burden associated with pregnancy loss in this population [ 15 , 28 ]. Consistent with this, the ISAR consensus document characterizes PGT‐SR as an established technique that should be actively offered to eligible couples [ 27 ]. Randomized trials and large observational studies have reported little or no benefit of PGT‐A in younger women but potential improvements in ongoing pregnancy or live birth in women ≥ 35–38 years [ 11 , 13 , 29 , 30 ]. Recent observational evidence, including the large national cohort reported by Harris et al., suggests that the apparent benefit of PGT‐A may be age‐dependent, with little or no improvement in younger women and more favorable outcomes in women aged 35 years and older, particularly 38–40 years [ 14 ]. In our study, both oPR and cPR were significantly higher than age‐adjusted registry expectations in women aged 35–39 and 40–42 years, whereas no benefit was demonstrated in women < 35 years. These findings are broadly concordant with an age‐dependent framework, but they remain observational and do not establish causality. From a practical standpoint, PGT‐A/SR also requires embryo biopsy, freeze‐all logistics, and additional laboratory costs, and this study did not directly evaluate time‐to‐pregnancy or cost‐effectiveness. Taken together, these findings may support considering approximately 35 years of age as one practical factor in selecting candidates for PGT‐A/SR, consistent with the addition, from 2025, of “around ≥ 35 years” as a new consideration for PGT‐A eligibility in Japan. The inclusion of both PGT‐A and PGT‐SR increases real‐world relevance, but these indications are biologically and clinically distinct. Patients with RIF or RPL and those carrying structural chromosomal rearrangements differ in mechanism, prognosis, and expected benefit. In an exploratory comparison of baseline and cycle characteristics, the PGT‐SR indication group differed from the PGT‐A indication group in several clinically relevant respects, including age, prior oocyte retrievals, prior unsuccessful embryo transfers, and prior miscarriages or stillbirths (Table  S5 ). These findings indicate that the combined PGT‐A/SR cohort should be interpreted as a pragmatic clinical cohort rather than as a biologically homogeneous population. At the same time, the structural chromosomal rearrangement subgroup was very small ( n  = 17), and this study was not designed or powered to estimate separate efficacy profiles for PGT‐A and PGT‐SR. Therefore, outcomes in the PGT‐SR subgroup should be regarded as descriptive only, and the inability to formally compare PGT‐A and PGT‐SR represents an important limitation of this study. The role of PGT‐A in the management of RPL remains incompletely characterized. Some reports suggest that patients with RPL have similar or higher rates of aneuploid embryos compared with infertile controls [ 4 , 31 ], whereas large registry‐based analyses and recent meta‐analyses indicate that PGT‐A can improve live birth rates and reduce miscarriage in RPL, especially at advanced maternal age [ 25 , 32 ]. The recent meta‐analysis in unexplained RPL showed that, even when restricting analysis to presumed euploid transfers, RPL was associated with lower live birth despite similar clinical pregnancy and miscarriage rates, highlighting the contribution of non‐embryonic factors [ 25 ]. The relatively high rate of early miscarriage in our younger subgroup may be compatible with a contribution of non‐embryonic factors and may also underscore the importance of a broader evaluation of uterine and systemic risk factors in RPL patients, even when PGT‐A is employed. At the same time, our trial was not powered to detect interaction effects between age and indication, and these results should not be interpreted as formal evidence of differential efficacy by indication. This interpretation also remains tentative in the absence of loss‐specific pathological data. The management of mosaic embryos remains an area of ongoing debate. Large retrospective series and narrative reviews have consistently shown lower implantation and higher miscarriage rates after mosaic embryo transfer compared with euploid embryos, although perinatal and neonatal outcomes among live‐born infants appear broadly similar [ 33 , 34 , 35 , 36 ]. Professional guidance, including ASRM committee opinions, emphasizes that mosaic embryos may be considered for transfer when no euploid embryos are available, but should generally be ranked after euploid embryos and followed by appropriate prenatal testing and counseling [ 33 ]. In our study, the subgroup of transfers using category B embryos was very small ( n  = 20), and the corresponding outcome estimates should therefore be interpreted as exploratory and hypothesis‐generating only. Within this limitation, the observed oPR and miscarriage rates after transfer of category B embryos were numerically similar to those after category A embryos, and this pattern is consistent with the possibility that many transferred category B embryos represented low‐level mosaicism, in line with relatively favorable outcomes reported for low‐level mosaic embryo transfer in another cohort [ 34 ]. Nevertheless, given the limited standardization of mosaicism calling and inter‐laboratory variability, our data do not support any conclusion that category B embryos perform comparably to category A embryos. In addition, because our follow‐up ended at 12–13 weeks and did not extend to birth or early childhood, we cannot draw conclusions regarding obstetric, perinatal, neonatal, or long‐term safety after transfer of category B embryos in this population. Taken together, these considerations underscore the need for cautious interpretation and individualized counseling when considering transfer of category B embryos. Several limitations should be acknowledged. First, we used age‐adjusted outcomes from the JSOG ART registry rather than a contemporaneous randomized or matched control group. Because the benchmark was based on 2018–2019 registry data rather than contemporaneous non‐PGT data within the same institutions, both chronological bias and inter‐institutional differences in clinical practice may have influenced the observed outcomes. Also, the registry cohort includes a mixture of embryo stages, fresh and frozen transfers, and many good‐prognosis cases, whereas our Advanced Medical Care B cohort was enriched for women with RIF and/or RPL. Therefore, our results should be interpreted as benchmarking per‐transfer performance in a relatively poor‐prognosis population rather than as a direct estimate of treatment effect. Second, the primary comparison contrasted oPR in the PGT‐A/SR cohort with cPR in the registry, resulting in a deliberately conservative and asymmetric test. Although the Japanese regulatory context explains why this approach was chosen, it does not remove the interpretive limitation created by the non‐equivalent outcome definitions. Accordingly, the cPR comparison should be regarded as prespecified supportive evidence rather than as a replacement for the primary oPR result. Third, outcomes were analyzed per embryo transfer rather than per retrieval cycle. Accordingly, outcomes could not be assessed for participants who discontinued or did not proceed to embryo transfer, and this may have influenced interpretation of per‐transfer results. This distinction is clinically important: although the per‐transfer cPR and oPR were 57.8% (126/218) and 48.6% (106/218), respectively, the corresponding patient‐level estimates among all patients who underwent PGT‐A/SR were 36.6% (126/344) and 30.8% (106/344). These lower patient‐level estimates reflect the 126 patients who did not proceed to embryo transfer, including 84 who did not obtain a category A or B embryo. Fourth, this study did not evaluate time‐to‐pregnancy, treatment cost, or cost‐effectiveness. Therefore, the broader clinical utility of PGT‐A/SR, particularly as an add‐on intervention requiring embryo biopsy, freeze‐all logistics, and additional laboratory procedures, cannot be inferred from pregnancy outcomes alone. Fifth, although the combined PGT‐A/SR cohort reflects pragmatic clinical practice, PGT‐A and PGT‐SR differ in etiology and expected mechanisms of benefit, which limits pooled interpretation. Because the structural chromosomal rearrangement subgroup was small ( n  = 17), the study was not designed or powered for indication‐specific efficacy inference, and PGT‐SR outcomes should be interpreted descriptively only. Sixth, POC chromosomal data were not systematically available in this multicenter study; therefore, the causes of residual pregnancy loss could not be distinguished definitively as embryonic, maternal/non‐embryonic, or technical factors. Finally, follow‐up did not extend to delivery or beyond, precluding evaluation of obstetric, perinatal, neonatal, and long‐term child outcomes, particularly after transfer of category B embryos. Despite these limitations, the present study has several notable strengths. It is a prospective, multicenter evaluation conducted under the Advanced Medical Care B framework, with strict eligibility criteria, standardized NGS‐based PGT‐A/SR procedures, prespecified endpoints, and third‐party central monitoring to ensure data completeness and protocol adherence. When our outcomes are compared with those of the JSOG special clinical study [ 19 ], they appear slightly less favorable. However, retrospective registry‐based analyses may be affected by selection and information biases as well as missing data, and discrepancies between prospective and retrospective findings have been reported, which may partly explain this apparent discrepancy [ 37 ]. In addition, the use of age‐stratified JSOG registry data enabled benchmarking of PGT‐A/SR outcomes against real‐world ART performance in Japan, rather than against selected single‐center controls. Finally, the study population also reflects the current Japanese indication set for PGT‐A/SR—RIF, RPL, and structural chromosomal rearrangements—making these findings directly relevant to clinical decision‐making and reimbursement discussions in this setting.

Written informed consent was obtained from all the patients to participate in this study, and for their anonymized information to be published in this article.

In Japan, approximately 11% of all births in 2023 resulted from assisted reproductive technology (ART), yet treatment outcomes remain suboptimal. Pregnancy rates decline and miscarriage rates rise sharply with increasing maternal age [ 1 ]. A major biological driver of these age‐related outcomes is the steep increase in embryo chromosomal abnormalities—from roughly one quarter of embryos in women in their late twenties to about 90% at age 45 [ 2 ]—which are estimated to account for approximately 67% of miscarriages [ 3 ]. By contrast, higher rates of chromosomally abnormal embryos have been reported in clinically relevant groups even at younger ages, including patients with recurrent pregnancy loss (RPL) and couples carrying structural chromosomal rearrangements [ 4 , 5 ]. Moreover, among women without known uterine factors that impair implantation, transfer of up to five chromosomally normal embryos yielded a cumulative live birth rate of 98.1% [ 6 ], suggesting that embryo aneuploidy is a frequent contributor to implantation failure in refractory infertility, including recurrent implantation failure (RIF). Taken together, these findings underscore the central importance of embryo chromosomal status for pregnancy success. Nevertheless, routine embryo assessment still relies on morphological grading, although chromosomally abnormal embryos may appear morphologically normal and chromosomally normal embryos may appear morphologically poor [ 7 ]. Thus, morphology alone provides limited information on embryo ploidy, underscoring the need for a pre‐transfer assessment method that directly evaluates embryo chromosomal status. Preimplantation genetic testing (PGT) involves biopsy of trophectoderm (TE) cells from the blastocyst followed by analysis of the chromosomal or genetic status of the embryo. PGT for aneuploidy (PGT‐A) and PGT for structural rearrangements (PGT‐SR) are used to detect numerical and structural chromosomal abnormalities [ 8 ]. Observational studies have reported that PGT‐A–selected embryos yield higher live birth rates and lower miscarriage rates than embryos transferred without PGT, especially in women of advanced maternal age [ 9 , 10 , 11 ]. However, the evidence is not uniform: a meta‐analysis showed a reduction in miscarriage rate without a clear improvement in pregnancy rate, and a randomized controlled trial in younger women did not demonstrate benefit [ 12 , 13 ]. Recent observational evidence also suggests that any apparent benefit of PGT‐A may be age‐dependent, with little or no improvement in younger women and more favorable outcomes in women aged 35 years and older [ 14 ]. Reflecting these mixed data and concerns about diagnostic accuracy and technical limitations, major international societies, including the European Society of Human Reproduction and Embryology (ESHRE), the American Society for Reproductive Medicine (ASRM), and the Human Fertilisation and Embryology Authority (HFEA), advise against unrestricted use of preimplantation genetic testing for aneuploidy and structural rearrangements (PGT‐A/SR) in all infertile patients and instead recommend selective use in subgroups such as those with RIF or advanced maternal age [ 15 , 16 , 17 ]. These considerations are particularly relevant when PGT‐A/SR is discussed as an add‐on, because patient‐level benefit, time‐to‐pregnancy, and cost‐effectiveness remain incompletely established [ 18 ]. In Japan, a special clinical study led by the Japan Society of Obstetrics and Gynecology (JSOG) suggested that PGT‐A/SR may improve pregnancy rates and reduce miscarriage rates per embryo transfer cycle [ 19 ]. However, rigorously monitored clinical studies remain scarce, and the clinical utility of PGT‐A/SR in the Japanese population has not been firmly established. Although JSOG initially withheld approval of PGT as comprehensive genetic screening, advances in genome analysis have enabled quantitative assessment of all chromosomes. Accordingly, PGT‐A/SR is now widely used in Europe and the United States [ 20 , 21 ]. In Japan, a JSOG‐led pilot preimplantation genetic screening study and a subsequent PGT‐A special clinical study suggested the potential efficacy and safety of PGT [ 19 , 22 ]. Following these studies, JSOG revised its position from withholding approval of PGT to permitting its use under specified conditions, and the 2021 JSOG guideline assigned PGT‐A/SR a recommendation level of B (recommended) [ 23 ]. Nevertheless, when public insurance coverage for infertility treatment was introduced, PGT‐A/SR was not included because rigorously monitored domestic trials evaluating its effectiveness were lacking. Because mixed medical care is prohibited in Japan, offering PGT as a self‐funded option makes the entire ART treatment self‐funded. This imposes a substantial financial burden and leads many patients to forgo PGT, consequently creating strong demand for insurance coverage of PGT‐A/SR. To provide evidence to inform reimbursement decisions within the framework of Advanced Medical Care B, the University of Osaka and JSOG‐designated collaborating institutions conducted a clinical study in which strictly monitored PGT‐A/SR cases were compared with age‐adjusted non‐PGT cases constructed from the JSOG ART Data Book. The aim was to clarify whether PGT‐A/SR is associated with higher pregnancy rates and lower miscarriage rates in Japan within a prospectively monitored single‐arm benchmarking framework, rather than to estimate a causal treatment effect relative to contemporaneous non‐PGT care.

The first author (Tsuyoshi Takiuchi) is affiliated with a Joint Research Division funded by Takara Bio at The University of Osaka. This affiliation may be perceived as a potential conflicts of interest. All other authors declare no conflicts of interest.

This prospective multicenter study was conducted as an Advanced Medical Care B program to evaluate the clinical utility of PGT‐A/SR in Japan. The protocol was approved by the Ethical Review Board of the University of Osaka Hospital (approval No. 15000038; 13 January 2023). The study was carried out in accordance with the Declaration of Helsinki and the Ethical Guidelines for Medical and Biological Research Involving Human Subjects in Japan. All couples provided written informed consent before enrolment. This study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement for cohort studies. This was a single‐arm, prospective, multicenter benchmarking study of couples undergoing ART with PGT‐A/SR, with outcomes compared against age‐adjusted reference values derived from the national ART registry of JSOG. The registry comparator was used to contextualize observed outcomes under routine Japanese practice and was not intended to provide a contemporaneous estimate of treatment effect or direct causal inference. The study was conducted at four JSOG‐approved facilities: The University of Osaka Hospital, HORAC Grand Front Osaka Clinic, IVF Osaka Clinic, and IVF Namba Clinic. Patient enrolment started in April 2023 and continued until the number of embryo transfers was expected to reach the pre‐specified target sample size. Couples requesting PGT‐A/SR at any of the participating centers were screened for eligibility. Eligible participants were infertile couples covered for ART under the national health insurance scheme, in whom both the woman and her partner were ≥ 18 years of age, and who met at least one of the following indications based on the 2021 JSOG statement on preimplantation genetic testing: a history of RIF, a history of RPL, or a confirmed structural chromosomal rearrangement in either partner. The main exclusion criterion was any previous participation in this study with PGT‐A/SR (i.e., previous embryo biopsy within this protocol), irrespective of whether embryo transfer had been performed, in order to ensure statistical independence of the cases. After registration, ovarian stimulation, oocyte retrieval, fertilization, and embryo culture were performed according to each center's standard protocols. When at least one biopsy‐eligible blastocyst was obtained, TE biopsy was performed and the biopsied cells were sent to a central clinical laboratory (Takara Bio Inc., Shiga, Japan) for PGT‐A/SR. Blastocysts were classified as category A (euploid), category B (euploid with suspected mosaic aneuploidy), category C (aneuploid), or category D (no reliable diagnosis), according to the JSOG “Guidelines for Embryo Diagnosis” and the standardized PGT‐A/SR reporting system. This classification was based on autosomal findings; sex chromosome abnormalities were not considered in the A–D categorization. The “embryo‐transfer cohort” for outcome analyses was defined as women who underwent their first single frozen–thawed blastocyst transfer using a category A or B embryo. Clinical information, including subsequent ART treatment and pregnancy outcomes, was obtained from medical records and entered into electronic case report forms. The planned sample size was 383 patients in the PGT cohort and 153 patients in the embryo‐transfer cohort. The target size for the embryo‐transfer cohort was calculated using a one‐sample proportion test with a normal approximation. We used an age‐stratified clinical pregnancy rate per embryo transfer of 41.4% in women younger than 43 years in the non–PGT‐A/SR population, derived from the 2018–2019 JSOG ART registry (Table  S1 ), as the reference value, and assumed a clinical pregnancy rate of 66.7% in the PGT‐A/SR cohort based on the results of a previous Japanese pilot study [ 22 ]. With a two‐sided significance level of 0.05 and 90% power, 153 embryo transfers were required. The probability of obtaining at least one category A or B embryo per patient was estimated as 40% based on interim results of the JSOG special clinical study of PGT‐A. Accordingly, approximately 383 registered couples were required to achieve 153 embryo transfers. A pre‐planned sample size re‐estimation was implemented to prevent loss of power if the age distribution of the embryo‐transfer cohort differed from that assumed in the initial design. When the embryo‐transfer cohort reached 153 patients, the null value for the ongoing pregnancy rate was recalculated as an age‐weighted mean of clinical pregnancy rates from the JSOG registry. If, under this recalculated null value, the power to detect an ongoing pregnancy rate of 55% with 153 patients (two‐sided α  = 0.05) was < 80%, the required sample size was to be recalculated with a target power of 80%, keeping all other design parameters unchanged. Because informed consent, ovarian stimulation, and PGT‐A/SR require several months, embryo transfers were allowed to continue beyond 153 patients when this number had already been exceeded at the time of re‐estimation. To ensure the integrity of case enrolment and data quality, the study was prospectively monitored by an independent third party. Central monitoring was performed based on the electronic case report forms and data accumulated at the data center. When necessary, on‐site monitoring visits were conducted to verify source documents and confirm adherence to the protocol and to national ethical guidelines. TE biopsy was performed when the blastocyst reached an appropriate stage of expansion, typically on day 5 or 6 after fertilization. Approximately 5–10 TE cells were obtained using laser‐assisted biopsy, and the biopsied blastocyst was vitrified immediately after sampling. Genomic DNA from TE samples was extracted and subjected to whole‐genome amplification using a commercial PGT‐A kit (Embgenix PGT‐A Kit, Takara Bio Inc., Shiga, Japan). Amplified DNA libraries were prepared according to the manufacturer's instructions and analyzed by low‐coverage whole‐genome next‐generation sequencing on a MiSeq platform (Illumina, San Diego, CA, USA). Sequencing reads were processed with dedicated analysis software (Embgenix Analysis System, Takara Bio Inc.) to estimate copy number across the genome and to detect copy number variations ≥ 8 Mb, as well as mosaicism, defined as intermediate copy‐number states corresponding to an abnormal cell fraction of approximately 30%–80%. Based on the JSOG “Guidelines for Embryo Diagnosis,” embryos were classified into four categories (A–D). Information on sex chromosomes was not disclosed to clinicians unless a sex‐chromosome abnormality was detected. Endometrial preparation for frozen–thawed embryo transfer was performed either in a natural cycle or in a hormone replacement cycle, according to each center's routine. All transfers were single blastocyst transfers using a category A or B embryo. The first frozen–thawed single blastocyst transfer using a PGT‐A/SR–classified category A or B embryo was considered the index transfer for analysis. The primary outcome was the ongoing pregnancy rate per embryo transfer (oPR) in the embryo‐transfer cohort. The oPR was calculated as the proportion of embryo transfers resulting in an ongoing pregnancy, defined as a clinical pregnancy with fetal cardiac activity confirmed by transvaginal ultrasonography between 12 + 0 and 13 + 6 weeks of gestation. Because the external comparator provided by the national registry is based on clinical pregnancy (gestational sac confirmation) rather than ongoing pregnancy at 12–13 weeks, the primary analysis necessarily involved a non‐equivalent benchmark. Accordingly, the clinical pregnancy rate per embryo transfer (cPR) was prespecified as a supportive outcome on the same scale as the registry benchmark to aid interpretation of the primary analysis; it was not intended to replace the primary endpoint. The cPR was defined as the proportion of embryo transfers resulting in at least one intrauterine gestational sac visualized on ultrasonography. Secondary outcomes were: 1. the miscarriage rate per clinical pregnancy, defined as the proportion of clinical pregnancies that resulted in pregnancy loss before 22 weeks of gestation; and 2. the implantation rate per embryo transfer including biochemical pregnancies, defined as the proportion of embryo transfers that resulted in biochemical or clinical pregnancy. Biochemical pregnancy was defined as a serum human chorionic gonadotropin (hCG) level ≥ 5 mIU/mL measured between 4 + 0 ± 2 and 5 + 0 ± 2 weeks of gestation. Clinical and ongoing pregnancies were followed at the participating ART centers or affiliated institutions at least through the visit at 12–13 weeks of gestation. Participant baseline characteristics collected at enrolment included female and partner age, dates of birth, height, weight, duration of infertility, indication for ART, number of prior clinical pregnancies, number of live births, number of previous miscarriages or stillbirths, number of previous oocyte retrievals, number of prior embryo transfers under insurance coverage, number of previous unsuccessful embryo transfers (embryo transfer cycles that resulted in either a negative hCG test or only a biochemical pregnancy), and the presence of structural chromosomal abnormalities in the woman or her partner. For the index transfer cycle, we recorded the date and type of endometrial preparation, use of hCG preparations, blastocyst morphology according to the Gardner grading system, PGT‐A/SR result, and endometrial thickness at the time of embryo transfer or progesterone initiation. Embryo damage was assessed descriptively by summarizing biopsy‐related damage among all embryos submitted for PGT‐A/SR and post‐thaw damage among embryos selected for transfer. All efficacy analyses were performed in the embryo‐transfer cohort. Proportions are presented with 95% confidence intervals (CIs) based on the normal approximation to the binomial distribution. The primary analysis compared the observed oPR with a null value representing expected performance in the absence of PGT‐A/SR using a one‐sample test for a proportion (two‐sided α  = 0.05). The null value was derived from the JSOG ART registry for 2018–2019 by obtaining age‐specific clinical pregnancy rates per first embryo transfer after oocyte retrieval and then computing an age‐weighted average according to the age distribution of the embryo‐transfer cohort. The same procedure was applied to define the null value for cPR. Because the JSOG registry does not include ongoing pregnancy rates at 12 weeks, the primary comparison relied on registry‐derived cPR as an intentionally conservative but non‐equivalent benchmark. For the miscarriage rate per clinical pregnancy, the point estimate and 95% CI were calculated, and a one‐sample test for a proportion was performed against a null value derived from age‐adjusted JSOG ART registry miscarriage rates up to 22 weeks. To approximate the miscarriage rate by 12 weeks of gestation, the age‐weighted miscarriage rate up to 22 weeks was multiplied by 0.90, assuming that approximately 90%–94% of miscarriages occur before 12 weeks, as previously reported [ 24 ]. For the implantation rate including biochemical pregnancies, point estimates and 95% CIs were summarized. Pre‐specified exploratory subgroup analyses in the embryo‐transfer cohort were performed for the primary and secondary outcomes according to infertility category (RIF, RPL, structural chromosomal rearrangement), blastocyst morphology (good [Gardner AA/AB/BA], moderate [AC/CA/BB], poor [BC/CB/CC]), PGT‐A/SR classification of the transferred embryo (category A vs. category B), and female age group (< 35, 35–39, and 40–42 years). For the age subgroups, age‐specific JSOG registry data were again used to calculate subgroup‐specific weighted null values for oPR and cPR, and one‐sample proportion tests using a normal approximation were applied (two‐sided α  = 0.05). For non‐age subgroups of oPR and cPR, estimates were interpreted descriptively by comparing their 95% CIs with the overall registry benchmark used in the primary analysis (42.4%). For miscarriage rate per clinical pregnancy, subgroup estimates were interpreted descriptively by comparing the 95% CI with the overall registry benchmark (23.4%). We performed two comparisons of baseline and cycle characteristics: one between the first 153 transfers and the subsequent transfers accrued after the planned sample‐size analysis, and another between the PGT‐A and PGT‐SR indication groups. Continuous variables were compared using t ‐tests. Categorical variables were compared using Fisher's exact test or Pearson's chi‐squared test. All p values were exploratory and were not adjusted for multiple comparisons. Given the heterogeneity of indications and the anticipated small number of structural chromosomal rearrangement carriers, indication‐specific subgroup analyses were specified as exploratory and descriptive. In addition to the prespecified per‐transfer analyses, patient‐level outcomes were descriptively summarized using all patients who underwent PGT‐A/SR as the denominator. Statistical analyses were performed using SAS software, version 9.4 (SAS Institute Inc., Cary, NC, USA).

Table S1: Age‐specific clinical pregnancy and miscarriage rates after the first embryo transfer following oocyte retrieval in the JSOG ART registry 2018–2019. Table S2: Pregnancy outcomes at the planned sample size ( N  = 153). Table S3: Comparison of baseline and cycle characteristics between the first 153 transfers and the additional 65 transfers. Table S4: Subgroup analysis of implantation rate (including biochemical pregnancies) in the embryo‐transfer cohort. Table S5: Descriptive comparison of baseline and cycle characteristics between the PGT‐A and PGT‐SR indication groups.