BlastAssist: a deep learning pipeline to measure interpretable features of human embryos.

STUDY QUESTION: Can the BlastAssist deep learning pipeline perform comparably to or outperform human experts and embryologists at measuring interpretable, clinically relevant features of human embryos in IVF? SUMMARY ANSWER: The BlastAssist pipeline can measure a comprehensive set of interpretable features of human embryos and either outperform or perform comparably to embryologists and human experts in measuring these features. WHAT IS KNOWN ALREADY: Some studies have applied deep learning and developed 'black-box' algorithms to predict embryo viability directly from microscope images and videos but these lack interpretability and generalizability. Other studies have developed deep learning networks to measure individual features of embryos but fail to conduct careful comparisons to embryologists' performance, which are fundamental to demonstrate the network's effectiveness. STUDY DESIGN, SIZE, DURATION: We applied the BlastAssist pipeline to 67 043 973 images (32 939 embryos) recorded in the IVF lab from 2012 to 2017 in Tel Aviv Sourasky Medical Center. We first compared the pipeline measurements of individual images/embryos to manual measurements by human experts for sets of features, including: (i) fertilization status (n = 207 embryos), (ii) cell symmetry (n = 109 embryos), (iii) degree of fragmentation (n = 6664 images), and (iv) developmental timing (n = 21 036 images). We then conducted detailed comparisons between pipeline outputs and annotations made by embryologists during routine treatments for features, including: (i) fertilization status (n = 18 922 embryos), (ii) pronuclei (PN) fade time (n = 13 781 embryos), (iii) degree of fragmentation on Day 2 (n = 11 582 embryos), and (iv) time of blastulation (n = 3266 embryos). In addition, we compared the pipeline outputs to the implantation results of 723 single embryo transfer (SET) cycles, and to the live birth results of 3421 embryos transferred in 1801 cycles. PARTICIPANTS/MATERIALS, SETTING, METHODS: In addition to EmbryoScope™ image data, manual embryo grading and annotations, and electronic health record (EHR) data on treatment outcomes were also included. We integrated the deep learning networks we developed for individual features to construct the BlastAssist pipeline. Pearson's χ2 test was used to evaluate the statistical independence of individual features and implantation success. Bayesian statistics was used to evaluate the association of the probability of an embryo resulting in live birth to BlastAssist inputs. MAIN RESULTS AND THE ROLE OF CHANCE: The BlastAssist pipeline integrates five deep learning networks and measures comprehensive, interpretable, and quantitative features in clinical IVF. The pipeline performs similarly or better than manual measurements. For fertilization status, the network performs with very good parameters of specificity and sensitivity (area under the receiver operating characteristics (AUROC) 0.84-0.94). For symmetry score, the pipeline performs comparably to the human expert at both 2-cell (r = 0.71 ± 0.06) and 4-cell stages (r = 0.77 ± 0.07). For degree of fragmentation, the pipeline (acc = 69.4%) slightly under-performs compared to human experts (acc = 73.8%). For developmental timing, the pipeline (acc = 90.0%) performs similarly to human experts (acc = 91.4%). There is also strong agreement between pipeline outputs and annotations made by embryologists during routine treatments. For fertilization status, the pipeline and embryologists strongly agree (acc = 79.6%), and there is strong correlation between the two measurements (r = 0.683). For degree of fragmentation, the pipeline and embryologists mostly agree (acc = 55.4%), and there is also strong correlation between the two measurements (r = 0.648). For both PN fade time (r = 0.787) and time of blastulation (r = 0.887), there's strong correlation between the pipeline and embryologists. For SET cycles, 2-cell time (P < 0.01) and 2-cell symmetry (P < 0.03) are significantly correlated with implantation success rate, while other features showed correlations with implantation success without statistical significance. In addition, 2-cell time (P < 5 × 10-11), PN fade time (P < 5 × 10-10), degree of fragmentation on Day 3 (P < 5 × 10-4), and 2-cell symmetry (P < 5 × 10-3) showed statistically significant correlation with the probability of the transferred embryo resulting in live birth. LIMITATIONS, REASONS FOR CAUTION: We have not tested the BlastAssist pipeline on data from other clinics or other time-lapse microscopy (TLM) systems. The association study we conducted with live birth results do not take into account confounding variables, which will be necessary to construct an embryo selection algorithm. Randomized controlled trials (RCT) will be necessary to determine whether the pipeline can improve success rates in clinical IVF. WIDER IMPLICATIONS OF THE FINDINGS: BlastAssist provides a comprehensive and holistic means of evaluating human embryos. Instead of using a black-box algorithm, BlastAssist outputs meaningful measurements of embryos that can be interpreted and corroborated by embryologists, which is crucial in clinical decision making. Furthermore, the unprecedentedly large dataset generated by BlastAssist measurements can be used as a powerful resource for further research in human embryology and IVF. STUDY FUNDING/COMPETING INTEREST(S): This work was supported by Harvard Quantitative Biology Initiative, the NSF-Simons Center for Mathematical and Statistical Analysis of Biology at Harvard (award number 1764269), the National Institute of Heath (award number R01HD104969), the Perelson Fund, and the Sagol fund for embryos and stem cells as part of the Sagol Network. The authors declare no competing interests. TRIAL REGISTRATION NUMBER: Not applicable.

A clinical predictive model for live birth in women of advanced age undergoing PGT cycles.

PURPOSE: The trend of delaying childbirth has resulted in a growing number of advanced-aged women who are opting for preimplantation genetic testing (PGT) to screen for monogenic diseases or structural chromosomal rearrangements (PGT-M and PGT-SR). This increase in demand necessitates the development of a clinical predictive model for live birth outcomes in these women. Therefore, the objective of this study is to construct a comprehensive predictive model that assesses the likelihood of achieving a successful live birth in advanced-aged women undergoing PGT-M and PGT-SR treatments. METHODS: A retrospective cohort study of 37-45-year-old women undergoing preimplantation genetic testing for monogenic disease or structural chromosomal rearrangement cycles from 2010 to 2021 was conducted at a university hospital reproductive centre. The purpose was to develop a clinical predictive model for live birth in these women. The main outcome studied was the cumulative live birth rate in the first or subsequent cycles. Developing a decision tree enabled a comprehensive study of clinical parameters and expected outcomes. RESULTS: The analysis included 158 women undergoing 753 preimplantation genetic testing cycles. The cumulative live birth rate was 37.342% (59/158). Decision tree analysis revealed that women aged ≤ 40.1 or women > 40.1 with one or more top-quality transferable embryos in their first cycle had the best chance for a live baby (56% and 41%, respectively). Those older than 40.1 without top-quality embryos and seven or fewer dominant follicles had no live births. A Kaplan-Meier curve showed that for autosomal dominant diseases, there was a negligible increase in live birth rate after three cycles, compared to six cycles in autosomal recessive inheritance. CONCLUSION: In older women, the chance of delivering after repeated cycles is higher in those with at least one top-quality unaffected embryo in their first preimplantation genetic testing cycle. Additional preimplantation genetic testing cycles after three in carriers of an autosomal dominant disorder and six in those with an autosomal recessive disorder should be considered prudently.

The association between a carrier state of FMR1 premutation and numeric sex chromosome variations.

PURPOSE: Women carriers of FMR1 premutation are at increased risk of early ovarian dysfunction and even premature ovarian insufficiency. The aim of this study was to examine a possible association between FMR1 permutation and numeric sex chromosome variations. METHODS: A retrospective case-control study conducted in the reproductive center of a university-affiliated medical center. The primary outcome measure was the rate of sex chromosomal numerical aberrations, as demonstrated by haplotype analyses, in FMR1 premutation carriers compared to X-linked preimplantation genetic testing for monogenic/single gene defect (PGT-M) cycles for other indications that do not affect the ovarian follicles and oocytes. RESULTS: A total of 2790 embryos with a final genetic analysis from 577 IVF PGT-M cycles were included in the final analysis. Mean age was similar between the groups, however, FMR1 carriers required more gonadotropins, and more women were poor responders with three or less oocytes collected. The ratio of embryos carrying a numeric sex chromosome variation was similar: 8.3% (138/1668) of embryos in the FMR1 group compared to 7.1% (80/1122) in the controls. A subgroup analysis based on age and response to stimulation has not demonstrated a significant difference either. CONCLUSIONS: Although carriers of FMR1 premutation exhibit signs of reduced ovarian response, it does not seem to affect the rate of numeric sex chromosomal variation compared to women undergoing PGT-M for other indications. This suggests that the mechanism for chromosomal number aberrations in women at advanced maternal age are different to those FMR1 premutation carriers with poor ovarian reserve.

The impact of fragile X premutation carrier status on embryo morphokinetic development.

RESEARCH QUESTION: Does inheritance of the fragile X mental retardation 1 (FMR1) premutation allele affect embryo morphokinetic development? DESIGN: A retrospective cohort analysis of 529 embryos from 126 IVF cycles of 39 FMR1 premutation female carriers undergoing preimplantation genetic testing for monogenic/single gene defects (PGT-M). Morphological and morphokinetic parameters obtained using a time-lapse monitoring system were compared between embryos that inherited the FMR1 premutation allele (FMR1 group, n = 271) and those who received the normal allele (normal group, n = 258). The following embryo outcome measures were compared: morphokinetic parameters up to day 3, start of blastulation time (tSB) for day 5 embryos and the rate of top-quality embryos on days 3 and 5. RESULTS: No differences were found in morphokinetic parameters between the groups from the time of intracytoplasmic sperm injection (ICSI) until a biopsy on day 3. The blastulation rate in the two groups was comparable. However, the start of blastulation was delayed in FMR1 embryos compared to that in the genetically normal embryos (median tSB: 104.2 h [99.3-110.3] versus 101.6 h [94.5-106.7], P = 0.01). In addition, the rate of top-quality FMR1 embryos was lower than that of genetically normal embryos (25.6% versus 38.8%, P = 0.04). CONCLUSION: Embryos that inherit the FMR1 premutation allele are of lower quality at the blastocyst stage compared with those that do not inherit the mutated allele.

Similar fertilization rates and preimplantation embryo development among testosterone-treated transgender men and cisgender women.

RESEARCH QUESTION: What are the effects of testosterone treatment on oocyte fertilization and preimplantation embryo development among transgender men who have undergone fertility preservation? DESIGN: A retrospective study was undertaken in a university-affiliated tertiary hospital between April 2016 and November 2021. Embryos were divided into three groups by source: 210 embryos from 7 testosterone-exposed transgender men, 135 from 10 cisgender women who cryopreserved embryos, and 276 from 24 cisgender women who underwent fertility treatment. Statistical analyses compared assisted reproductive technology outcomes between the group of transgender men and both groups of cisgender women. Morphokinetic and morphological parameters were compared between the embryos derived from these three groups. RESULTS: The transgender men (30.2 ± 3.5 years of age) were significantly younger than the cisgender women who cryopreserved embryos (35.1 ± 1.8 years; P = 0.005) and the cisgender women who underwent fertility treatment (33.8 ± 3.2 years; P = 0.017). After adjusting for participant age, the fertilization rate was comparable between the transgender men and both groups of cisgender women (P = 0.391 and 0.659). There were no significant differences between the transgender men and the cisgender women who preserved fertility in terms of number of cryopreserved embryos (7.2 ± 5.1 and 3.5 ± 2.6; P = 0.473) or the distribution of embryo age at cryopreservation (P = 0.576). All morphokinetic parameters evaluated by time-lapse imaging, as well as the morphological characteristics, were comparable for the embryos in all three groups. CONCLUSIONS: Testosterone exposure among transgender men has no adverse impact upon fertilization rates or preimplantation embryo development and quality.

A decision tree analysis applied to women aged 43-45: who should be referred for ovum donation?

RESEARCH QUESTION: In women at the advanced age of 43-45 years undergoing repeated IVF cycles with autologous oocytes, who has the highest chance for birth and who should be referred early to receive donor oocytes? DESIGN: A retrospective cohort study was conducted at a university hospital reproductive centre. The computerized database of 394 women aged 43-45 years undergoing 1528 non-donor IVF or intracytoplasmic sperm injection cycles between 2010 and 2019 was analysed. A decision tree was developed, enabling a comprehensive study of a set of clinical parameters and the expected outcomes. RESULTS: The cumulative clinical pregnancy rate was 15.0% (59/394) and the cumulative live birth rate was 8.4% (33/394). The decision tree developed to predict women who should be offered egg donation included age, poor ovarian response to stimulation, the number of top-quality embryos, dominant follicles, previous pregnancy or live birth, fertilized oocytes and body mass index. The model showed that a good ovarian response in the first cycle was the best predictor for live birth (13.3% gave birth). However, among women with poor responses, 7.1% of those who were younger than 43.5 years gave birth, and none of the women who were older than 43.5 years did. CONCLUSIONS: Women over 43.5 years old with fewer than four oocytes collected in their first IVF cycle should be offered ovum donation, since their live birth rate in subsequent cycles is negligible.

Why fertility preservation rates of transgender men are much lower than those of transgender women.

RESEARCH QUESTION: What are the fertility preservation rates of transgender women and transgender men, and what are the factors that affect their decision-making? DESIGN: This prospective study included 97 transgender women and 91 transgender men referred to the Gender Clinic of the study medical centre's Endocrinology Institute and to the Gan Meir Community Health Care Centre. The responders completed a 28-item questionnaire during 2018. RESULTS: Most of the transgender women and transgender men wished to parent a child (67.4% and 61.9%, respectively, P = 0.447), but only 40.4% of the transgender women and 5.8% of the transgender men used fertility preservation (P < 0.001). The main reasons for not pursuing fertility preservation were unwillingness to postpone gender-affirming treatment (58.8% and 74.0%, respectively, P = 0.076), preference to adopt a child (58.8% and 60.9%, respectively, P = 0.818) and cost (44.9% and 60.9%, respectively, P = 0.086). Factors related to the fertility preservation process itself were specifically chosen by transgender men compared with transgender women as the reason for not pursuing this treatment, including distress caused by the fertility preservation technique (60.3% versus 29.3%, respectively, P = 0.006), fear of gender dysphoria caused by hormonal treatment (63.5% versus 28.3%, P = 0.002) and concern over the attitude of medical staff (44% versus 19%, P = 0.027). CONCLUSIONS: Fertility preservation rates were considerably lower among transgender men than transgender women, strongly related to the fertility preservation process itself. Finding ways to overcome the obstacles confronted by transgender individuals, especially transgender men, will enhance their future biological parenting.

Maternal underweight does not adversely affect the outcomes of IVF/ICSI and frozen embryo transfer cycles or early embryo development.

OBJECTIVE: To compare assisted reproductive technology (ART) outcomes and preimplantation embryo development between underweight and normal-weight women. METHODS: This retrospective cohort study included 26 underweight women (body mass index [BMI] < 18.50 kg/m2) and 104 normal-weight women (BMI >20 and <24.9 kg/m2) who underwent a total of 204 in vitro fertilization/intracytoplasmic sperm injection (IVF/ICSI) cycles and 358 fresh/frozen embryo transfers (ET) in our institution between January 2016 and December 2018. Statistical analyses compared selected ART outcomes (ovarian stimulation, fertilization, and pregnancy) between both weight groups. Morphokinetic and morphological parameters were also compared between 346 and 1467 embryos of underweight and normal-weight women, respectively. RESULTS: The mean ± standard deviation age of the underweight and normal-weight women was similar (31.6 ± 4.17 vs 32.4 ± 3.59 years; p = .323). There were no differences in the peak estradiol levels, the number of retrieved oocytes, the number of metaphase II oocytes, and the oocyte maturity rates between the two groups. The IVF/ICSI fertilization rates and the number of embryos suitable for transfer or cryopreservation were similar for both groups. All morphokinetic parameters that were evaluated by means of time-lapse imaging as well as the morphological characteristics were comparable between low and normal BMI categories. There were no significant differences in pregnancy achievement, clinical pregnancy, live births, and miscarriage rates between the suboptimal and optimal weight women. CONCLUSION: Underweight status has no adverse impacts on the outcomes of IVF/ICSI with either fresh or frozen ET or on preimplantation embryo development and quality.

Reproductive Outcomes and Overall Prognosis of Women with Asherman's Syndrome Undergoing IVF.

STUDY OBJECTIVE: To evaluate the impact of Asherman syndrome (AS) following hysteroscopic adhesiolysis on reproductive outcomes and the time to achieve pregnancy in women with infertility undergoing in vitro fertilization (IVF) treatment. DESIGN: Case-control study. SETTING: Tertiary university-affiliated medical center. PATIENTS: Fifty-one infertile women who were treated for AS and underwent IVF (study group) matched for age and etiology of infertility with non-AS controls at a 1:1 ratio. INTERVENTIONS: Medical records search, chart review, and phone survey were used to assess reproductive outcomes. MEASUREMENTS AND MAIN RESULTS: A multivariate logistic regression analyses was used to assess live birth, accounting for patient age at stimulation cycle start, parity, number of embryos transferred, and endometrial thickness. A survival analysis was performed to assess the times that had lapsed from interventions to conception. The study group of 51 women included 38 (74.5%) with moderate to severe disease. The mean number of embryo transfers per woman was similar for the study and control groups (4.9 ± 4.6 vs 6.22 ± 4.3, respectively, p = .78). The controls had a significantly higher mean endometrial thickness before embryo transfer (8.7 ± 1.8 mm vs 6.95 ± 1.7 mm, p = .001). The overall time to achieve live birth was significantly longer in women with AS (p = .022). In a logistic regression analysis, the presence of moderate to severe AS was shown to be an independent factor for achieving a live birth (adjusted odds ratio 0.174, 95% confidence interval [CI], 0.032-0.955, p = .004). Women with AS who had live births had a significantly thicker mean endometrial thickness (8.2 ± 1.4 mm vs 6.9 ± 1.2, p = .001). CONCLUSION: Moderate and severe AS has a detrimental effect on reproductive performance in infertile women. Endometrial thickness is an important predictor for live births among women with AS who undergo IVF.

Testicular microlithiasis defines a subgroup of azoospermic men with low rates of sperm retrieval.

OBJECTIVE: To investigate the prevalence of testicular microlithiasis and its association with sperm retrieval rates and histopathology in men with non-obstructive azoospermia. METHODS: A total of 120 men underwent scrotal ultrasonography prior to microsurgical testicular sperm extraction. Sperm retrieval rate, testicular histopathology, testicular size, reproductive hormones, karyotyping, Y chromosome microdeletion analyses, and presence of varicoceles and hydroceles were compared between men with and without testicular microlithiasis. RESULTS: The total sperm retrieval rate was 40%. Ten men with normal spermatogenesis were excluded. The remaining 110 men with non-obstructive azoospermia were analyzed and testicular microlithiasis was detected in 16 of them (14.5%). The sperm retrieval rate in that subgroup was only 6.2% (1/16) as opposed to 39.4% (37/94) in men with non-obstructive azoospermia and no evidence of microlithiasis (P = 0.009). The mean right and left testicular diameters were significantly lower in the microlithiasis group (P = 0.04). On multivariate logistic regression analysis, the presence of mictolithiasis (odds ratio 7.4, 95% confidence interval 2.3, 12.2; P = 0.01) was the only independent predictor of unsuccessful sperm retrieval. The 15 patients with microlithiasis and without successful sperm extraction were diagnosed by histopathology as having Sertoli cells only. The 16th patient with successful sperm retrieval had a histopathology of mixed atrophy and was diagnosed with Klinefelter syndrome. CONCLUSION: The presence of testicular microlithiasis is associated with low sperm retrieval rates among our cohort of men with non-obstructive azoospermia undergoing scrotal ultrasonography prior to microsurgical testicular sperm extraction. Larger, prospective studies should be conducted to confirm these findings.