Performance of a deep learning based neural network in the selection of human blastocysts for implantation.

Deep learning in in vitro fertilization is currently being evaluated in the development of assistive tools for the determination of transfer order and implantation potential using time-lapse data collected through expensive imaging hardware. Assistive tools and algorithms that can work with static images, however, can help in improving the access to care by enabling their use with images acquired from traditional microscopes that are available to virtually all fertility centers. Here, we evaluated the use of a deep convolutional neural network (CNN), trained using single timepoint images of embryos collected at 113 hr post-insemination, in embryo selection amongst 97 clinical patient cohorts (742 embryos) and observed an accuracy of 90% in choosing the highest quality embryo available. Furthermore, a CNN trained to assess an embryo's implantation potential directly using a set of 97 euploid embryos capable of implantation outperformed 15 trained embryologists (75.26% vs. 67.35%, p<0.0001) from five different fertility centers.

Around one in seven couples have trouble conceiving, which means there is a high demand for solutions such as in vitro fertilization, also known as IVF. This process involves fertilizing and developing embryos in the laboratory and then selecting a few to implant into the womb of the patient. IVF, however, only has a 30% success rate, is expensive and can be both mentally and physically taxing for patients. Selecting the right embryos to implant is therefore extremely important, as this increases the chance of success, minimizes complications and ensures the baby will be healthy. Currently the tools available for making this decision are limited, highly subjective, time-consuming, and often extremely expensive. As a result, embryologists often rely on their experience and observational skills when choosing which embryos to implant, which can lead to a lot of variability. An automated system based on artificial intelligence (AI) could therefore improve IVF success rates by assisting embryologists with this decision and ensuring more consistent results. The AI system could learn how embryos develop over time and then uses this information to select the best embryos to implant from just a single image. This would offer a cheaper alternative to current analysis tools that are only available at the most expensive IVF clinics. Now, Bormann, Kanakasabapathy, Thirumalaraj et al. have developed an AI system for IVF based on thousands of images of embryos. Using individual images, the system selected embryos of a comparable quality to those selected by a human specialist. It also showed a greater ability to identify embryos that will lead to successful implantation. Indeed, the software outperformed 15 embryologists from five different centers across the United States in detecting which embryos were most likely to implant out of a group of high-quality embryos with few visible differences. Artificial intelligence has many potential applications to support expert clinical decision-making. Systems like these could improve success, reduce errors and lead to faster, cheaper and more accessible results. Beyond immediate IVF applications, this system could also be used in research and industry to help understand differences in embryo quality.

Antimüllerian hormone as a predictor of good-quality supernumerary blastocyst cryopreservation among women with levels <1 ng/mL versus 1-4 ng/mL.

OBJECTIVE: To determine whether antimüllerian hormone (AMH) levels predict the availability of good-quality supernumerary blastocysts for cryopreservation. DESIGN: Retrospective study. SETTING: Two fertility centers. PATIENT(S): First fresh IVF cycles (n = 247) grouped as follows: 40 women <35 year old with AMH <1 ng/mL and 77 women with AMH 1-4 ng/mL; 62 women ≥35 year old with AMH <1 ng/mL, and 68 women with AMH 1-4 ng/mL. INTERVENTION(S): AMH level measured before IVF with ovarian stimulation protocols based on patient age and AMH level, including short gonadotropin-releasing hormone (GnRH) agonist, GnRH antagonist, or GnRH agonist microdose flare; supernumerary good-quality blastocysts cryopreserved on days 5 or 6 after retrieval. MAIN OUTCOME MEASURES(S): Supernumerary good-quality blastocysts for cryopreservation in relation to AMH levels. RESULT(S): Among women <35 years of age, there was a statistically significant difference in the number of patients with supernumerary good-quality blastocysts for cryopreservation between the groups with AMH <1 ng/mL and AMH 1-4 ng/mL (30.0% vs. 58.4%) when adjusted for age. Among women ≥35 years of age, there was a statistically significant difference in the number of patients with supernumerary good-quality blastocyst cryopreservation between groups with AMH <1 ng/mL and AMH 1-4 ng/mL (16.1% vs. 42.6%), when adjusted for age. CONCLUSION(S): Low AMH levels are associated with a statistically significantly lower likelihood of blastocysts for cryopreservation as compared with higher AMH levels. This effect was seen among women both <35 and ≥35 years of age. Patient counseling should include realistic expectations for the probability of good-quality supernumerary blastocysts available for cryopreservation.

Blastocyst embryo transfer is the primary determinant for improved outcomes in oocyte donation cycles.

AIM: Using oocyte donation cycles as an ideal model, we sought to compare pregnancy and implantation rates in cleavage stage (day 3) versus blastocyst stage (day 6) embryo transfers (ET); assess the predictive value of blastocyst formation rates based on cleavage cell stage and morphology grade; and evaluate the ability to predict formation of high quality (HQ) blastocysts. METHODS: Ninety three consecutive oocyte donation cycles from July 2003 to August 2005 were retrospectively evaluated and analyzed to determine if either resulted in a cleavage stage (n = 30) or blastocyst (n = 45) ET. The primary outcomes measured pregnancy rates, the percent development of HQ blastocysts based on day 3 embryo status, and the ability to select day 3 embryos suitable for transfer among four blinded evaluators by assessing their day 6 embryo outcome. RESULTS: Cleavage stage ET resulted in significantly lower pregnancy rates, clinical pregnancy rates, and implantation rates (47% [n = 14/30]; 40% [n = 12/30] and 27 + or - 7%) compared to blastocyst stage (82% [n = 37/45]; 73% [n = 33/45] and 64 + or - 6% [+ or -SE], P < 0.01). In total, HQ blastocysts resulted from high and good quality day 3 embryos 35% (191/546) and 17% (93/546), respectively. Blinded evaluation revealed at least one, two or all three day 3 embryos were correctly selected for ET on day 6, 97%, 67% and 19%, respectively. CONCLUSION: Day 6 ET resulted in significantly better clinical outcomes compared to day 3 ET. While day 3 status is not predictive of blastocyst quality, the selection of at least one day 3 embryo ultimately suitable for blastocyst ET underscores the significance of optimal endometrial receptivity.

Interim analysis of a randomized clinical trial comparing abdominal versus transvaginal ultrasound-guided embryo transfer.

AIM: Ultrasound-guided embryo transfer appears to improve overall pregnancy outcomes for in vitro fertilization embryo transfer cycles. Most reports are done using a transabdominal ultrasound-guided approach, in contrast to a transvaginal ultrasound, which does not require a full bladder. We sought to determine if either approach was better with respect to clinical pregnancy outcomes. METHODS: This randomized clinical trial took place in a university-based practice with an enrollment of 186 subjects. Secondary measures also included embryo transfer time and degree of cramping and pain, which subjects ranked as none = 1; mild = 2; moderate = 3; or severe = 4. RESULTS: There were no differences in implantation rates (33.1 + or - 4.1%, + or - standard error vs 31.1 + or - 3.9%, P = 0.78), pregnancy rates (60% vs 54%, P = 0.38), clinical pregnancy rates (48% vs 45%, P = 0.77), and live-birth rates (30% vs 39%, P = 0.37) between transabdominal and transvaginal ultrasound-guided embryo transfer groups. No differences were noted between groups for time required for embryo transfer (157 + or - 279 seconds vs 130 + or - 176 seconds, P = 0.92), uterine cramping (1.2 + or - 0.5 vs 1.2 + or - 0.4, P = 0.4), and the degree of pain (1.4 + or - 0.7 vs 1.3 + or - 0.5, P = 0.13). CONCLUSIONS: Neither transabdominal nor transvaginal ultrasound-guided embryo transfer is more beneficial in optimizing pregnancy outcomes. While delays often occur while waiting for bladder distension for transabdominal ultrasound-guided embryo transfer, uterine position, parity and level of physician comfort should dictate the choice of either approach.

Follicular and luteal phase endometrial thickness and echogenic pattern and pregnancy outcome in oocyte donation cycles.

PURPOSE: To study the effect of endometrial thickness (ET) and echogenic pattern (EP) in oocyte donation cycles upon pregnancy outcomes. METHODS: Seventy-nine cycles resulting in blastocyst embryo transfer were evaluated. Donors underwent ovarian hyperstimulation using rFSH and GnRH-antagonist. Recipients were synchronized to donors using GnRH-agonist down-regulation followed by fixed dose of estrogen (E2) and progesterone (P4) following hCG. Transvaginal ultrasound (US) obtained ET and EP 10-11 days after initiation of E2 and on day of embryo transfer. Primary outcome was ET and EP in pregnant and non-pregnant cycles. Stimulation and embryology data was analyzed in donors to assess differences prior to transfer. RESULTS: Fifty-nine cycles resulted in clinical pregnancy. No differences were observed in pregnant vs. non-pregnant cycles in proliferative or secretory ET and EP. Similar baseline and stimulation characteristics were found in pregnant and non-pregnant cycles. Regression analysis showed end thickness were not predictive of pregnancy outcomes. CONCLUSIONS: Endometrial characteristics in recipients prior to and following progesterone were not predictive of pregnancy outcomes.