A proof of concept for a deep learning system that can aid embryologists in predicting blastocyst survival after thaw.

The ability to understand whether embryos survive the thaw process is crucial to transferring competent embryos that can lead to pregnancy. The objective of this study was to develop a proof of concept deep learning model capable of assisting embryologist assessment of survival of thawed blastocysts prior to embryo transfer. A deep learning model was developed using 652 labeled time-lapse videos of freeze-thaw blastocysts. The model was evaluated against and along embryologists on a test set of 99 freeze-thaw blastocysts, using images obtained at 0.5 h increments from 0 to 3 h post-thaw. The model achieved AUCs of 0.869 (95% CI 0.789, 0.934) and 0.807 (95% CI 0.717, 0.886) and the embryologists achieved average AUCs of 0.829 (95% CI 0.747, 0.896) and 0.850 (95% CI 0.773, 0.908) at 2 h and 3 h, respectively. Combining embryologist predictions with model predictions resulted in a significant increase in AUC of 0.051 (95% CI 0.021, 0.083) at 2 h, and an equivalent increase in AUC of 0.010 (95% CI -0.018, 0.037) at 3 h. This study suggests that a deep learning model can predict in vitro blastocyst survival after thaw in aneuploid embryos. After correlation with clinical outcomes of transferred embryos, this model may help embryologists ascertain which embryos may have failed to survive the thaw process and increase the likelihood of pregnancy by preventing the transfer of non-viable embryos.

Semen parameters on the day of oocyte retrieval predict low fertilization during conventional insemination IVF cycles.

PURPOSE: Poor fertilization during conventional IVF is difficult to predict in the absence of abnormal semen parameters; large-scale studies are lacking. The purpose of this study is to evaluate factors associated with low fertilization rates in conventional insemination IVF cycles. METHODS: A retrospective cohort study evaluating demographic, reproductive evaluation, and IVF cycle characteristics to identify predictors of low fertilization (defined as 2PN/MII ≤ 30% per cycle). Participants were included if they were undergoing their first IVF cycle utilizing fresh autologous oocytes and conventional insemination with male partner's sperm (with normal pretreatment semen analysis). They were randomly divided into a training set and a validation set; validation modeling with logistic regression and binary distribution was utilized to identify covariates associated with low fertilization. RESULTS: Postprocessing sperm concentration of less than 40 million/ml and postprocessing sperm motility < 50% on the day of retrieval were the strongest predictors of low fertilization in the training dataset. Next, in the validation set, cycles with either low postprocessing concentration (≤ 40 million/ml) or low postprocessing progressive motility (≤ 50%) were 2.9-times (95% CI 1.4, 6.2) more likely to have low fertilization than cycles without either risk factor. Furthermore, cycles with low postprocessing concentration and progressive motility were 13.4 times (95% CI 4.01, 45.06) more likely to have low fertilization than cycles without either risk factor. CONCLUSIONS: Postprocessing concentration and progressive motility on the day of oocyte retrieval are predictive of low fertilization in conventional IVF cycles with normal pretreatment diagnostic semen analysis parameters.