Improved pregnancy prediction performance in an updated deep-learning embryo selection model: a retrospective independent validation study.

RESEARCH QUESTION: What is the effect of increasing training data on the performance of ongoing pregnancy prediction after single vitrified-warmed blastocyst transfer (SVBT) in a deep-learning model? DESIGN: A total of 3960 SVBT cycles were retrospectively analysed. Embryos were stratified according to the Society for Assisted Reproductive Technology age groups. Embryos were scored by deep-learning models iDAScore v1.0 (IDA-V1) and iDAScore v2.0 (IDA-V2) (15% more training data than v1.0) and by Gardner grading. The discriminative performance of the pregnancy prediction for each embryo scoring model was compared using the area under the curve (AUC) of the receiver operating characteristic curve for each maternal age group. RESULTS: The AUC of iDA-V2, iDA-V1 and Gardener grading in all cohort were 0.736, 0.720 and 0.702, respectively. iDA-V2 was significantly higher than iDA-V1 and Gardener grading (P < 0.0001). Group > 35 years (n = 757): the AUC of iDA-V2 was significantly higher than Gardener grading (0.718 versus 0.694, P = 0.015); group aged 35-37 years (n = 821), the AUC of iDA-V2 was significantly higher than iDA-V1 (0.712 versus 0.696, P = 0.035); group aged 41-42 years (n = 715, the AUC of iDA-V2 was significantly higher than Gardener grading (0.745 versus 0.696, P = 0.007); group > 42 years (n = 660) and group aged 38-40 years (n = 1007), no significant differences were found between the groups. CONCLUSION: The performance of deep learning models for pregnancy prediction will be improved by increasing the size of the training data.

Correlations between a deep learning-based algorithm for embryo evaluation with cleavage-stage cell numbers and fragmentation.

RESEARCH QUESTION: Do cell numbers and degree of fragmentation in cleavage-stage embryos, assessed manually, correlate with evaluations made by deep learning algorithm model iDAScore v2.0? DESIGN: Retrospective observational study (n = 5040 embryos; 1786 treatments) conducted at two Swedish assisted reproductive technology centres between 2016 and 2021. Fresh single embryo transfer was carried out on days 2 or 3 after fertilization. Embryo evaluation using iDAScore v2.0 was compared with manual assessment of numbers of cells and grade of fragmentation, analysed by video sequences. RESULTS: Data from embryos transferred on days 2 and 3 showed that having three or fewer cells compared with four or fewer cells on day 2, and six or fewer cells versus seven to eight cells on day 3, correlated significantly with a difference in iDAScore (medians 2.4 versus 4.0 and 2.6 versus 4.6 respectively; both P < 0.001). The iDAScore for 0-10% fragmentation was significantly higher compared with the groups with higher fragmentation (P < 0.001). When combining cell numbers and fragmentation, iDAScore values decreased as fragmentation increased, regardless of cell number. iDAScore discriminated between embryos that resulted in live birth or no live birth (AUC of 0.627 and 0.607), compared with the morphological model (AUC of 0.618 and 0.585) for day 2 and day 3, respectively. CONCLUSIONS: The iDAScore v2.0 values correlated significantly with cell numbers and fragmentation scored manually for cleavage-stage embryos on days 2 and 3. iDAScore had some predictive value for live birth, conditional that embryo selection was based on morphology.

Using blastocyst re-expansion rate for deciding when to warm a new blastocyst for single vitrified-warmed blastocyst transfer.

RESEARCH QUESTION: Can predictive post-warm parameters that support the decision to transfer a warmed blastocyst or to warm another blastocyst be identified in women with multiple frozen-vitrified blastocysts? DESIGN: Retrospective single-centre observational cohort analysis. A total of 1092 single vitrified-warmed blastocyst transfers (SVBT) with known Gardner score, maternal age and live birth were used to develop live birth prediction models based on logistic regression, including post-warm re-expansion parameters. Time-lapse incubation was used for pre-vitrification and post-warm embryo culture. A dataset of 558 SVBT with the same inclusion criteria was used to validate the model, but with known clinical pregnancy outcome instead of live birth outcome. RESULTS: Three different logistic regression models were developed for predicting live birth based on post-warm blastocyst re-expansion. Different post-warm assessment times indicated that a 2-h post-warm culture period was optimal for live birth prediction (model 1). Adjusting for pre-vitrification Gardner score (model 2) and in combination with maternal age (model 3) further increased predictability (area under the curve [AUC] = 0.623, 0.633, 0.666, respectively). Model validation gave an AUC of 0.617, 0.609 and 0.624, respectively. The false negative rate and true negative rate for model 3 were 2.0 and 10.1 in the development dataset and 3.5 and 8.0 in the validation dataset. CONCLUSIONS: Clinical application of a simple model based on 2 h of post-warm re-expansion data, pre-vitrification Gardner score and maternal age can support a standardized approach for deciding if warming another blastocyst may increase the likelihood of live birth in SVBT.

Comparing performance between clinics of an embryo evaluation algorithm based on time-lapse images and machine learning.

PURPOSE: This article aims to assess how differences in maternal age distributions between IVF clinics affect the performance of an artificial intelligence model for embryo viability prediction and proposes a method to account for such differences. METHODS: Using retrospectively collected data from 4805 fresh and frozen single blastocyst transfers of embryos incubated for 5 to 6 days, the discriminative performance was assessed based on fetal heartbeat outcomes. The data was collected from 4 clinics, and the discrimination was measured in terms of the area under ROC curves (AUC) for each clinic. To account for the different age distributions between clinics, a method for age-standardizing the AUCs was developed in which the clinic-specific AUCs were standardized using weights for each embryo according to the relative frequency of the maternal age in the relevant clinic compared to the age distribution in a common reference population. RESULTS: There was substantial variation in the clinic-specific AUCs with estimates ranging from 0.58 to 0.69 before standardization. The age-standardization of the AUCs reduced the between-clinic variance by 16%. Most notably, three of the clinics had quite similar AUCs after standardization, while the last clinic had a markedly lower AUC both with and without standardization. CONCLUSION: The method of using age-standardization of the AUCs that is proposed in this article mitigates some of the variability between clinics. This enables a comparison of clinic-specific AUCs where the difference in age distributions is accounted for.

Towards Automation in IVF: Pre-Clinical Validation of a Deep Learning-Based Embryo Grading System during PGT-A Cycles.

Preimplantation genetic testing for aneuploidies (PGT-A) is arguably the most effective embryo selection strategy. Nevertheless, it requires greater workload, costs, and expertise. Therefore, a quest towards user-friendly, non-invasive strategies is ongoing. Although insufficient to replace PGT-A, embryo morphological evaluation is significantly associated with embryonic competence, but scarcely reproducible. Recently, artificial intelligence-powered analyses have been proposed to objectify and automate image evaluations. iDAScore v1.0 is a deep-learning model based on a 3D convolutional neural network trained on time-lapse videos from implanted and non-implanted blastocysts. It is a decision support system for ranking blastocysts without manual input. This retrospective, pre-clinical, external validation included 3604 blastocysts and 808 euploid transfers from 1232 cycles. All blastocysts were retrospectively assessed through the iDAScore v1.0; therefore, it did not influence embryologists' decision-making process. iDAScore v1.0 was significantly associated with embryo morphology and competence, although AUCs for euploidy and live-birth prediction were 0.60 and 0.66, respectively, which is rather comparable to embryologists' performance. Nevertheless, iDAScore v1.0 is objective and reproducible, while embryologists' evaluations are not. In a retrospective simulation, iDAScore v1.0 would have ranked euploid blastocysts as top quality in 63% of cases with one or more euploid and aneuploid blastocysts, and it would have questioned embryologists' ranking in 48% of cases with two or more euploid blastocysts and one or more live birth. Therefore, iDAScore v1.0 may objectify embryologists' evaluations, but randomized controlled trials are required to assess its clinical value.

Development and validation of deep learning based embryo selection across multiple days of transfer.

This work describes the development and validation of a fully automated deep learning model, iDAScore v2.0, for the evaluation of human embryos incubated for 2, 3, and 5 or more days. We trained and evaluated the model on an extensive and diverse dataset including 181,428 embryos from 22 IVF clinics across the world. To discriminate the transferred embryos with known outcome, we show areas under the receiver operating curve ranging from 0.621 to 0.707 depending on the day of transfer. Predictive performance increased over time and showed a strong correlation with morphokinetic parameters. The model's performance is equivalent to the KIDScore D3 model on day 3 embryos while it significantly surpasses the performance of KIDScore D5 v3 on day 5+ embryos. This model provides an analysis of time-lapse sequences without the need for user input, and provides a reliable method for ranking embryos for their likelihood of implantation, at both cleavage and blastocyst stages. This greatly improves embryo grading consistency and saves time compared to traditional embryo evaluation methods.

Does embryo categorization by existing artificial intelligence, morphokinetic or morphological embryo selection models correlate with blastocyst euploidy rates?

RESEARCH QUESTION: Does embryo categorization by existing artificial intelligence (AI), morphokinetic or morphological embryo selection models correlate with blastocyst euploidy? DESIGN: A total of 834 patients (mean maternal age 40.5 ± 3.4 years) who underwent preimplantation genetic testing for aneuploidies (PGT-A) on a total of 3573 tested blastocysts were included in this retrospective study. The cycles were stratified into five maternal age groups according to the Society for Assisted Reproductive Technology age groups (<35, 35-37, 38-40, 41-42 and >42 years). The main outcome of this study was the correlation of euploidy rates in stratified maternal age groups and an automated AI model (iDAScore® v1.0), a morphokinetic embryo selection model (KIDScore Day 5 ver 3, KS-D5) and a traditional morphological grading model (Gardner criteria), respectively. RESULTS: Euploidy rates were significantly correlated with iDAScore (P = 0.0035 to <0.001) in all age groups, and expect for the youngest age group, with KS-D5 and Gardner criteria (all P < 0.0001). Additionally, multivariate logistic regression analysis showed that for all models, higher scores were significantly correlated with euploidy (all P < 0.0001). CONCLUSION: These results show that existing blastocyst scoring models correlate with ploidy status. However, as these models were developed to indicate implantation potential, they cannot accurately diagnose if an embryo is euploid or aneuploid. Instead, they may be used to support the decision of how many and which blastocysts to biopsy, thus potentially reducing patient costs.

Correlation between an annotation-free embryo scoring system based on deep learning and live birth/neonatal outcomes after single vitrified-warmed blastocyst transfer: a single-centre, large-cohort retrospective study.

PROPOSE: Does an annotation-free embryo scoring system based on deep learning and time-lapse sequence images correlate with live birth (LB) and neonatal outcomes? METHODS: Patients who underwent SVBT cycles (3010 cycles, mean age: 39.3 ± 4.0). Scores were calculated using the iDAScore software module in the Vitrolife Technology Hub (Vitrolife, Gothenburg, Sweden). The correlation between iDAScore, LB rates, and total miscarriage (TM), including 1st- and 2nd-trimester miscarriage, was analysed using a trend test and multivariable logistic regression analysis. Furthermore, the correlation between the iDAScore and neonatal outcomes was analysed. RESULTS: LB rates decreased as iDAScore decreased (P < 0.05), and a similar inverse trend was observed for the TM rates. Additionally, multivariate logistic regression analysis showed that iDAScore significantly correlated with increased LB (adjusted odds ratio: 1.811, 95% CI: 1.666-1.976, P < 0.05) and decreased TM (adjusted odds ratio: 0.799, 95% CI: 0.706-0.905, P < 0.05). There was no significant correlation between iDAScore and neonatal outcomes, including congenital malformations, sex, gestational age, and birth weight. Multivariate logistic regression analysis, which included maternal and paternal age, maternal body mass index, parity, smoking, and presence or absence of caesarean section as confounding factors, revealed no significant difference in any neonatal characteristics. CONCLUSION: Automatic embryo scoring using iDAScore correlates with decreased miscarriage and increased LB and has no correlation with neonatal outcomes.

Robust and generalizable embryo selection based on artificial intelligence and time-lapse image sequences.

Assessing and selecting the most viable embryos for transfer is an essential part of in vitro fertilization (IVF). In recent years, several approaches have been made to improve and automate the procedure using artificial intelligence (AI) and deep learning. Based on images of embryos with known implantation data (KID), AI models have been trained to automatically score embryos related to their chance of achieving a successful implantation. However, as of now, only limited research has been conducted to evaluate how embryo selection models generalize to new clinics and how they perform in subgroup analyses across various conditions. In this paper, we investigate how a deep learning-based embryo selection model using only time-lapse image sequences performs across different patient ages and clinical conditions, and how it correlates with traditional morphokinetic parameters. The model was trained and evaluated based on a large dataset from 18 IVF centers consisting of 115,832 embryos, of which 14,644 embryos were transferred KID embryos. In an independent test set, the AI model sorted KID embryos with an area under the curve (AUC) of a receiver operating characteristic curve of 0.67 and all embryos with an AUC of 0.95. A clinic hold-out test showed that the model generalized to new clinics with an AUC range of 0.60-0.75 for KID embryos. Across different subgroups of age, insemination method, incubation time, and transfer protocol, the AUC ranged between 0.63 and 0.69. Furthermore, model predictions correlated positively with blastocyst grading and negatively with direct cleavages. The fully automated iDAScore v1.0 model was shown to perform at least as good as a state-of-the-art manual embryo selection model. Moreover, full automatization of embryo scoring implies fewer manual evaluations and eliminates biases due to inter- and intraobserver variation.

Predicting Embryo Viability Based on Self-Supervised Alignment of Time-Lapse Videos.

With self-supervised learning, both labeled and unlabeled data can be used for representation learning and model pretraining. This is particularly relevant when automating the selection of a patient's fertilized eggs (embryos) during a fertility treatment, in which only the embryos that were transferred to the female uterus may have labels of pregnancy. In this paper, we apply a self-supervised video alignment method known as temporal cycle-consistency (TCC) on 38176 time-lapse videos of developing embryos, of which 14550 were labeled. We show how TCC can be used to extract temporal similarities between embryo videos and use these for predicting pregnancy likelihood. Our temporal similarity method outperforms the time alignment measurement (TAM) with an area under the receiver operating characteristic (AUC) of 0.64 vs. 0.56. Compared to existing embryo evaluation models, it places in between a pure temporal and a spatio-temporal model that both require manual annotations. Furthermore, we use TCC for transfer learning in a semi-supervised fashion and show significant performance improvements compared to standard supervised learning, when only a small subset of the dataset is labeled. Specifically, two variants of transfer learning both achieve an AUC of 0.66 compared to 0.63 for supervised learning when 16% of the dataset is labeled.

Aneuploidy rates and morphokinetic parameters of embryos cultured in distinct culture media: a sibling oocyte study.

STUDY QUESTION: Do embryos from sibling oocytes assigned to distinct single-step media culture systems demonstrate differences in early embryo development, morphokinectics or aneuploidy rates? SUMMARY ANSWER: Embryo quality, morphokinetic parameters and aneuploidy rates from trophectoderm biopsy were similar between sibling embryos cultured in distinct media systems from the time of gamete isolation. WHAT IS KNOWN ALREADY: Studies on the effect of commercially available embryo culture media systems have demonstrated inconsistent impact on human embryonic development, morphokinetics, aneuploidy rates and clinical outcomes. In addition, these studies have been primarily randomized at the level of the embryo or the patient to culture media. STUDY DESIGN, SIZE, DURATION: Prospective sibling oocyte cohort derived from 200 subjects undergoing IVF at a tertiary academic medical center between February 2018 and November 2019. PARTICIPANTS/MATERIALS, SETTING, METHODS: Sibling oocytes were allocated to Global® or SAGE® media system based upon laterality of ovary from which they were retrieved. All embryos were cultured in a time-lapse incubator. Blastocysts underwent trophectoderm biopsy for preimplantation genetic testing for aneuploidy using next-generation sequencing. MAIN RESULTS AND THE ROLE OF CHANCE: One hundred twenty-seven subjects (n = 127) had paired blastocysts for biopsy in each culture media system. There was no difference in top quality blastocyst formation (47.1 ± 31.0 vs 48.1 ± 27.2%; P = 0.87) nor aneuploidy rate (62.3 ± 34.0 vs 56.1 ± 34.4%; P = 0.07) for sibling embryos cultured in Global versus SAGE media system. Embryo morphokinetic parameters including time to each cell division from two cells (t2) to eight cells (t8), time to morula stage (tM), time to blastocele formation (tSB), time to fully formed blastocyst (tB) and time to expansion of the blastocyst (tEB) were similar between paired blastocysts from each culture media system. LIMITATIONS, REASONS FOR CAUTION: Pregnancy outcomes and offspring health data were not available for analysis. WIDER IMPLICATIONS OF THE FINDINGS: Commercially available culture media may not have a differential impact on embryo development and blastocyst aneuploidy rate when patient and stimulation-related factors are held constant. STUDY FUNDING/COMPETING INTEREST(S): There was no external funding for this study. C.H. is owner of a consultancy company, IVF Professionals, Chief Scientific Officer at Apricity, Executive Director at TMRW and co-owner and shareholder of Aria Fertility. She has received speaker fees, consulting fees and travel support from Cooper Surgical and Vitrolife. J.B. is an employee and shareholder of vitrolife. TRIAL REGISTRATION NUMBER: N/A.

Pregnancy prediction performance of an annotation-free embryo scoring system on the basis of deep learning after single vitrified-warmed blastocyst transfer: a single-center large cohort retrospective study.

OBJECTIVE: To analyze the performance of an annotation-free embryo scoring system on the basis of deep learning for pregnancy prediction after single vitrified blastocyst transfer (SVBT) compared with the performance of other blastocyst grading systems dependent on annotation or morphology scores. DESIGN: A single-center large cohort retrospective study from an independent validation test. SETTING: Infertility clinic. PATIENT(S): Patients who underwent SVBT cycles (3,018 cycles, mean ± SD patient age 39.3 ± 4.0 years). INTERVENTION(S): None. MAIN OUTCOME MEASURE(S): The pregnancy prediction performances of each embryo scoring model were compared using the area under curve (AUC) for predicting the fetal heartbeat status for each maternal age group. RESULT(S): The AUCs of the <35 years age group (n = 389) for pregnancy prediction were 0.72 for iDAScore, 0.66 for KIDScore, and 0.64 for the Gardner criteria. The AUC of iDAScore was significantly greater than those of the other two models. For the 35-37 years age group (n = 514), the AUCs were 0.68, 0.68, and 0.65 for iDAScore, KIDScore, and the Gardner criteria, respectively, and were not significantly different. The AUCs of the 38-40 years age group (n = 796) were 0.67 for iDAScore, 0.65 for KIDScore, and 0.64 for the Gardner criteria, and there were no significant differences. The AUCs of the 41-42 years age group (n = 636) were 0.66, 0.66, and 0.63 for iDAScore, KIDScore, and the Gardner criteria, respectively, and there were no significant differences among the pregnancy prediction models. For the >42 years age group (n = 389), the AUCs were 0.76 for iDAScore, 0.75 for KIDScore, and 0.75 for the Gardner criteria, and there were no significant differences. Thus, iDAScore AUC was either the highest or equal to the highest AUC for all age groups, although a significant difference was observed only in the youngest age group. CONCLUSION(S): Our results showed that objective embryo assessment by a completely automatic and annotation-free model, iDAScore, performed as well as or even better than more traditional embryo assessment or annotation-dependent ranking tools. iDAScore could be an optimal pregnancy prediction model after SVBT, especially in young patients.

Comparing prediction of ongoing pregnancy and live birth outcomes in patients with advanced and younger maternal age patients using KIDScore™ day 5: a large-cohort retrospective study with single vitrified-warmed blastocyst transfer.

BACKGROUND: The KIDScore™ Day 5 (KS-D5) model, version 3, is a general morphokinetic prediction model (Vitrolife, Sweden) for fetal heartbeat prediction after embryo transfer that was developed based on a large data set that included implantation results from a range of clinics with different patient populations, culture conditions and clinical practices. However, there was no study to comparing their pregnancy and live birth prediction ability among different maternal age. The aim of this study is to analyze the performance of KS-D5 in predicting pregnancy and live birth in various maternal age groups after single vitrified-warmed blastocyst transfer (SVBT). METHODS: A total of 2486 single vitrified-warmed blastocyst transfer (SVBT) cycles were analyzed retrospectively. Confirmed fetal heartbeat positive (FHB+) and live birth (LB+) rates were stratified by Society for Assisted Reproductive Technology (SART) maternal age criteria (< 35, 35-37, 38-40, 41-42 and ≥ 43 years of age). Within each age group, the performance of the prediction model was calculated using the AUC, and the results were compared across the age groups. RESULTS: In all age groups, the FHB+ rates decreased as the KIDScore decreased (P <  0.05). Conversely, the AUCs increased as the maternal age increased. The AUC of the < 35 age group (0.589) was significantly lower than the AUCs of the 41-42 age group (0.673) and the ≥43 age group (0.737), respectively (P <  0.05). In all age groups, the LB+ rates decreased as the KIDScore decreased (P <  0.05). Conversely, the AUCs increased as the maternal age increased. The AUC of the ≥43 age group (0.768) was significantly higher than the AUCs of other age groups (P <  0.05; < 35 age group = 0.596, 35-37 age group = 0.640, 38-40 age group = 0.646, 41-42 age group = 0.679). CONCLUSIONS: In the present study, we determined that the KIDScore model worked well for prediction of pregnancy and live birth outcomes in advanced age patients.

Automatic grading of human blastocysts from time-lapse imaging.

BACKGROUND: Blastocyst morphology is a predictive marker for implantation success of in vitro fertilized human embryos. Morphology grading is therefore commonly used to select the embryo with the highest implantation potential. One of the challenges, however, is that morphology grading can be highly subjective when performed manually by embryologists. Grading systems generally discretize a continuous scale of low to high score, resulting in floating and unclear boundaries between grading categories. Manual annotations therefore suffer from large inter-and intra-observer variances. METHOD: In this paper, we propose a method based on deep learning to automatically grade the morphological appearance of human blastocysts from time-lapse imaging. A convolutional neural network is trained to jointly predict inner cell mass (ICM) and trophectoderm (TE) grades from a single image frame, and a recurrent neural network is applied on top to incorporate temporal information of the expanding blastocysts from multiple frames. RESULTS: Results showed that the method achieved above human-level accuracies when evaluated on majority votes from an independent test set labeled by multiple embryologists. Furthermore, when evaluating implantation rates for embryos grouped by morphology grades, human embryologists and our method had a similar correlation between predicted embryo quality and pregnancy outcome. CONCLUSIONS: The proposed method has shown improved performance of predicting ICM and TE grades on human blastocysts when utilizing temporal information available with time-lapse imaging. The algorithm is considered at least on par with human embryologists on quality estimation, as it performed better than the average human embryologist at ICM and TE prediction and provided a slightly better correlation between predicted embryo quality and implantability than human embryologists.

Symposium: innovative techniques in human embryo viability assessment. Human oocyte respiration-rate measurement--potential to improve oocyte and embryo selection?

Oocyte and embryo selection are not highly successful, with fewer than 10% of oocytes in assisted reproduction resulting in a delivery. Techniques for oocyte and embryo selection rely on highly subjective morphology assessment, with few true quantitative techniques available. One aspect of oocyte health that could be considered is the ability to produce ATP through respiration. Using a non-invasive technology, the respiration rates of individual human oocytes were recorded in an attempt to correlate respiration and oocyte health with probable subsequent development. Oocytes used were either immature or mature, non-fertilized oocytes from a clinical assisted reproduction programme. Differences in respiration rates between oocytes within a cohort and between cohorts of oocytes were recorded. The differences between cohorts reflected many of the currently known differences in oocyte health, related to age and FSH concentrations. However, within a cohort, differences between oocytes were observed, with some having high rates and others low. Oocytes with respiration rates of between 0.48 and 0.55 nl O(2)/h were viable, with lower rates consistent with lack of continued in-vitro maturation or atresia. This technology may have a future in the clinical laboratory as a predictor of oocyte health and ability to develop into an embryo with greater potential of delivery.