Dataset Information

Deep learning as a predictive tool for fetal heart pregnancy following time-lapse incubation and blastocyst transfer.

ABSTRACT: STUDY QUESTION:Can a deep learning model predict the probability of pregnancy with fetal heart (FH) from time-lapse videos? SUMMARY ANSWER:We created a deep learning model named IVY, which was an objective and fully automated system that predicts the probability of FH pregnancy directly from raw time-lapse videos without the need for any manual morphokinetic annotation or blastocyst morphology assessment. WHAT IS KNOWN ALREADY:The contribution of time-lapse imaging in effective embryo selection is promising. Existing algorithms for the analysis of time-lapse imaging are based on morphology and morphokinetic parameters that require subjective human annotation and thus have intrinsic inter-reader and intra-reader variability. Deep learning offers promise for the automation and standardization of embryo selection. STUDY DESIGN, SIZE, DURATION:A retrospective analysis of time-lapse videos and clinical outcomes of 10?638 embryos from eight different IVF clinics, across four different countries, between January 2014 and December 2018. PARTICIPANTS/MATERIALS, SETTING, METHODS:The deep learning model was trained using time-lapse videos with known FH pregnancy outcome to perform a binary classification task of predicting the probability of pregnancy with FH given time-lapse video sequence. The predictive power of the model was measured using the average area under the curve (AUC) of the receiver operating characteristic curve over 5-fold stratified cross-validation. MAIN RESULTS AND THE ROLE OF CHANCE:The deep learning model was able to predict FH pregnancy from time-lapse videos with an AUC of 0.93 [95% CI 0.92-0.94] in 5-fold stratified cross-validation. A hold-out validation test across eight laboratories showed that the AUC was reproducible, ranging from 0.95 to 0.90 across different laboratories with different culture and laboratory processes. LIMITATIONS, REASONS FOR CAUTION:This study is a retrospective analysis demonstrating that the deep learning model has a high level of predictability of the likelihood that an embryo will implant. The clinical impacts of these findings are still uncertain. Further studies, including prospective randomized controlled trials, are required to evaluate the clinical significance of this deep learning model. The time-lapse videos collected for training and validation are Day 5 embryos; hence, additional adjustment would need to be made for the model to be used in the context of Day 3 transfer. WIDER IMPLICATIONS OF THE FINDINGS:The high predictive value for embryo implantation obtained by the deep learning model may improve the effectiveness of previous approaches used for time-lapse imaging in embryo selection. This may improve the prioritization of the most viable embryo for a single embryo transfer. The deep learning model may also prove to be useful in providing the optimal order for subsequent transfers of cryopreserved embryos. STUDY FUNDING/COMPETING INTEREST(S):D.T. is the co-owner of Harrison AI that has patented this methodology in association with Virtus Health. P.I. is a shareholder in Virtus Health. S.C., P.I. and D.G. are all either employees or contracted with Virtus Health. D.G. has received grant support from Vitrolife, the manufacturer of the Embryoscope time-lapse imaging used in this study. The equipment and time for this study have been jointly provided by Harrison AI and Virtus Health.

SUBMITTER: Tran D

PROVIDER: S-EPMC6554189 | biostudies-literature | 2019 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep learning as a predictive tool for fetal heart pregnancy following time-lapse incubation and blastocyst transfer.

Tran D D Cooke S S Illingworth P J PJ Gardner D K DK

Human reproduction (Oxford, England) 20190601 6

<h4>Study question</h4>Can a deep learning model predict the probability of pregnancy with fetal heart (FH) from time-lapse videos?<h4>Summary answer</h4>We created a deep learning model named IVY, which was an objective and fully automated system that predicts the probability of FH pregnancy directly from raw time-lapse videos without the need for any manual morphokinetic annotation or blastocyst morphology assessment.<h4>What is known already</h4>The contribution of time-lapse imaging in effec ...[more]

PMID: 31111884

Similar Datasets

Project description:Background0PN zygotes have a low cleavage rate, and the clinical outcomes of cleavage-stage embryo transfers are unsatisfactory. Blastocyst culturing is used to screen 0PN embryos, but whether the cell number of 0PN embryos on day 3 affects the clinical outcomes following single blastocyst transfer is unknown and would be helpful in evaluating the clinical value of these embryos.MethodsThis retrospective study compared 46,804 0PN zygotes, 242 0PN frozen-thawed single blastocyst transfers, and 92 corresponding 0PN singletons with 232,441 2PN zygotes, 3563 2PN frozen-thawed single blastocyst transfers, and 1250 2PN singletons from January 2015 to October 2019 at a tertiary-care academic medical centre. The 0PN and 2PN embryos were divided into two groups: the group with < 6 cells on day 3 and that with ≥ 6 cells. Embryo development, subsequent pregnancy and neonatal outcomes were compared between the two groups.ResultsThe cleavage and available blastocyst rates of the 0PN zygotes were much lower than those of the 2PN zygotes (25.9% vs. 97.4%, P < 0.001; 13.9% vs. 23.4%, P < 0.001). In the < 6 cells group, the available blastocyst rate of the cleaved 0PN embryos was significantly lower than that of the 2PN embryos (2.5% vs. 12.7%, P < 0.001). However, in the ≥ 6 cells group, the available blastocyst rate of the 0PN cleaved embryos significantly improved, although it was slightly lower than that of the 2PN embryos (33.9% vs. 35.7%, P = 0.014). Importantly, compared to those of the 2PN single blastocyst transfers, the clinical pregnancy rate, live birth rate, Z-score and malformation rate of the 0PN single blastocyst transfers were not significantly different in either the < 6 cells group (30.4% vs. 39.8%, P = 0.362; 30.4% vs. 31.3%, P = 0.932; 0.89 ± 0.90 vs. 0.42 ± 1.02, P = 0.161; 0% vs. 2.6%, P = 1.000) or the ≥ 6 cells group (50.7% vs. 46.6%, P = 0.246; 39.7% vs. 38.3%, P = 0.677; 0.50 ± 1.23 vs. 0.47 ± 1.11, P = 0.861; 2.4% vs. 1.8%, P = 1.000).ConclusionsThe cell number on day 3 of 0PN embryos affected the subsequent formation of blastocysts but did not influence the subsequent pregnancy and neonatal outcomes of 0PN single blastocyst transfers, which may be beneficial to clinicians counselling patients on the clinical value of 0PN embryos.

Project description:BackgroundTime-lapse imaging combined with embryo morphokinetics may offer a non-invasive means for improving embryo selection. Data from clinics worldwide are necessary to compare and ultimately develop embryo classifications models using kinetic data. The primary objective of this study was to determine if there were kinetic differences between embryos with limited potential and those more often associated with in vitro blastocyst formation and/or implantation. We also wanted to compare putative kinetic markers for embryo selection as proposed by other laboratories to what we were observing in our own laboratory setting.MethodsKinetic data and cycle outcomes were retrospectively analyzed in patients age 39 and younger with 7 or more zygotes cultured in the Embryoscope. Timing of specific events from the point of insemination were determined using time-lapse (TL) imaging. The following kinetic markers were assessed: time to syngamy (tPNf), t2, time to two cells (c), 3c (t3), 4c ( t4), 5c (t5), 8c (t8), morula (tMor), start of blastulation (tSB); tBL, blastocyst (tBL); expanded blastocyst (tEBL). Durations of the second (cc2) and third (cc3) cell cycles, the t5-t2 interval as well as time to complete synchronous divisions s1, s2 and s3 were calculated. Incidence and impact on development of nuclear and cleavage anomalies were also assessed.ResultsA total of 648 embryos transferred on day 5 were analyzed. The clinical pregnancy and implantation rate were 72% and 50%, respectively. Morphokinetic data showed that tPNf, t2,t4, t8, s1, s2,s3 and cc2 were significantly different in embryos forming blastocysts (ET or frozen) versus those with limited potential either failing to blastulate or else forming poor quality blastocysts ,ultimately discarded. Comparison of embryo kinetics in cycles with all embryos implanting (KID+) versus no implantation (KID-) suggested that markers of embryo competence to implant may be different from ability to form a blastocyst. The incidence of multinucleation and reverse cleavage amongst the embryos observed was 25% and 7%, respectively. Over 40% of embryos exhibiting these characteristics did however form blastocysts meeting our criteria for freezing.ConclusionsThese data provide us with a platform with which to potentially enhance embryo selection for transfer.

Project description:Study questionAre there any differences in live birth rates (LBR) following fresh blastocyst transfer in natural or clomiphene-stimulated cycles, or after elective blastocyst freezing in clomiphene-stimulated cycles followed by thawing and transfer at different time-points?Summary answerClomiphene citrate (CC) administration adversely affected the LBR after single fresh blastocyst transfer (SBT) in CC cycles compared with that in natural cycles, while this adverse effect of CC is not present when a single vitrified-warmed blastocyst transfer (SVBT) is performed in subsequent natural ovulatory cycles, regardless of the duration between CC administration and the day of SVBT.What is known alreadyCC affects uterine receptivity associated with a thinning of the uterine endometrium through an antioestrogenic effect. However, the duration that this adverse effect of CC on uterine endometrium persists after initial use is still unknown.Study design size durationA retrospective cohort study of 157 natural cycle IVFs followed by SBT and 1496 minimal ovarian stimulation with CC IVF cycles followed by SBT (n = 24) or SVBT (n = 1472) from January 2010 to December 2014 was conducted. SVBT cycles were classified into two groups according to the period between the last day of CC administration and the day of SVBT (A: ≤60 d and B: ≥61 d). All groups were then compared based on pregnancy outcomes (natural-SBT group: n = 157, CC-SBT group: n = 24, SVBT-A: n = 1143, SVBT-B: n = 329).Participants/materials setting methodsWomen were aged 30-39 years at oocyte retrieval. In SVBT cycles, blastocysts were vitrified and warmed using a Cryotop safety kit. SVBT was performed in subsequent natural ovulatory cycles. The main outcomes were LBR and neonatal outcome, and both were compared among the groups.Main results and the role of chanceThe LBR in the CC-SBT group (29.2%, 7/24) was significantly lower compared with the natural-SBT (56.1%, 88/157) (P = 0.01) and SVBT-A (50.0%, 572/1143) (P = 0.04), but not SVBT-B (47.4%, 156/329), groups. Furthermore, multivariate logistic regression analysis revealed that the LBR was comparable among the natural-SBT and SVBT groups, but was significantly lower in the CC-SBT group (adjusted odds ratio: 0.324, 95% CI: 0.119-0.800, P = 0.01). No significant differences among all groups were observed for gestational age (P = 0.19), birthweight (P = 0.41) and incidence of malformation (P = 0.53).Limitations reasons for cautionIn this study we analysed a biased sample, based on clinical judgement regarding endometrial thickness, and the study was limited by its retrospective nature. The low statistical power caused by the group size disparity was also a limitation, especially in the CC-SBT group. Although the outcome showing inferiority of CC-SBT compared to natural-SBT is consistent with general findings in the literature, further large-scale clinical studies, ideally RCTs, are necessary to validate our results and clarify the prolonged effect of CC in SVBT cycles on pregnancy and neonatal outcomes.Wider implications of the findingsOur observation suggests that CC administered in minimal ovarian stimulation cycles affects adversely the pregnancy outcomes when SBT is performed. Therefore, for a CC-based minimal stimulation IVF cycle, we suggest that frozen embryo transfer should be performed in a subsequent natural ovulatory cycle to avoid the possibility of implantation failure associated with CC administration.Study funding/competing interestsThe authors have no conflicts of interest to declare. No external funding was either sought or obtained.

Dataset Information

Deep learning as a predictive tool for fetal heart pregnancy following time-lapse incubation and blastocyst transfer.

Publications

Deep learning as a predictive tool for fetal heart pregnancy following time-lapse incubation and blastocyst transfer.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets