[Interest of iDAScore (intelligent Data Analysis Score) for embryo selection in routine IVF laboratory practice: Results of a preliminary study]

INTRODUCTION

Embryo selection is a major challenge in ART, especially since the generalization of single embryo transfer, and its optimization could lead to the improvement of clinical results in IVF. Recently, several Artificial Intelligence (AI) models, based on deep-learning such as iDAScore, have been developed. These models, trained on time-lapse videos of embryos with known implantation data, can predict the probability of pregnancy for a given embryo, allowing automatization and standardization in embryo selection.

MATERIAL AND METHODS

In this study, we have compared the hierarchical categorization of 311 D5 blastocysts of iDAScore v1.0 and the embryologists of our unit. These 311 D5 blastocysts have been classified as top (70.1%), good (Q+: 10.6%) and poor (Q-: 19.3%) quality by embryologists according to Gardner classification. Median iDAScores were [9.9-8.4],]8.4-7.5] and]7.5-2.1] for top, good and poor-quality blastocysts respectively.

RESULTS

We observed a significantly concordant categorization between iDAScore and embryologists for top, good and poor-quality blastocysts (respectively, 89.5, 36.4 and 48.3%, P < 10-4). Moreover, the hierarchical categorization of the three best blastocysts between iDAScore and the embryologists was as follow: 1st rank: 71.9%; 2nd rank: 61.6%; 3rd rank: 56.8% (P=0.07). One hundred and fifty-one blastocysts with known implantation data were analyzed. The iDAScore of blastocysts that implanted was significantly higher than those that did not implant (implantation+: 9.10±0.57; implantation-: 8.70±0.95, P=0.003).

CONCLUSION

This preliminary study shows that iDAScore is able to perform a reproducible, reliable and immediate hierarchical classification of blastocysts. Moreover, this tool can identify the blastocysts with the highest implantation potential. If these results confirmed on a larger scale of embryos and patients, IA could revolutionize IVF laboratories by standardizing embryo hierarchical selection.

Optimal Embryo Selection: The Irreplaceable Role of the Embryologist in an Age of Advancing Technology

Background

Time-lapse incubators allow for ongoing evaluation of embryos without culture condition disruption. The use of time-lapse incubation has been shown to improve outcomes either by improving overall conditions or providing additional information to aid in embryo selection for transfer. Time-lapse incubators can also utilise morphokinetic models to rank embryos based on morphokinetic parameters. We sought to compare a morphokinetic model for embryo comparison to traditional morphologic evaluation.

Aims

The aim of the study is to compare a morphokinetic model for embryo comparison to traditional morphologic evaluation.

Settings and Design

This is a retrospective cohort design.

Materials and Methods

Embryos cultured in a time-lapse culture system that had traditional morphologic evaluation, morphokinetic modelling and known live birth outcomes were included in this study. Embryos with unknown competence were excluded, including when two embryos were transferred with a single live birth resulted.

Statistical Analysis Used

Receiver operating characteristic (ROC) curves were determined for both the morphologic analysis and the morphokinetic model on culture day 3 and day 5. Using the ROC-determined cutoff that optimised both sensitivity and specificity, a binary outcome for each test was analysed using agreement statistics to determine if one method of embryo evaluation was superior to the other.

Results

Morphological and morphokinetic grading were both predictive of embryo competence on days 3 and 5. However, on day 3, morphologic grading was superior to morphokinetic grading with area under the curve (AUC) of 0.66 (P < 0.001) and 0.58 (P = 0.009), respectively. Contrarily, on day 5, the morphokinetic model had a higher AUC of 0.65 (P = 0.03) compared to the morphologic grading, AUC 0.56 (P = 0.02).

Conclusion

Traditional morphology was noted to be a better diagnostic tool (higher AUC) on culture day 3 while a morphokinetic model was superior on day 5.

Embryo culture conditions are significantly improved during uninterrupted incubation: A randomized controlled trial

A parallel group superiority prospective randomised controlled trial was devised to compare the culture characteristics of human pre-implantation stage embryos during uninterrupted culture in a time lapse incubator (TLI) versus the conventional model of interrupted culture in a standard incubator (SI) under low oxygen tension using a single step medium. 221 patients aged 35-and-under, 124 patients aged between 36 and 39 and 86 patients aged 40-and-over years were randomised and cultured either in a SI or in a TLI. Patients in the three age groups were distributed between the TLI and SI in a 1:1 ratio. The development of embryos on days 2, 3 and 5, and the clinical pregnancy and implantation rates were recorded. The fertilisation rate, development of day 2 and clinical pregnancy rates were similar in both treatments but the 8-cell development rate in all age groups combined (p = 0.016), blastocyst development rate (p = 0.0022) and the implantation rate (p = 0.0022) was significantly higher for the uninterrupted culture. These findings demonstrated significant differences between the two incubation groups. It also indicated less efficacious embryonic development with age in both treatments which appeared more pronounced in the conventional incubator. In conclusion uninterrupted culture is superior compared to the interrupted incubation culture system.

Anti-Mullerian hormone (AMH) and embryo quality assessed by time-lapse imaging (TLI): a cross-sectional observational study

PURPOSE

Serum anti-Mullerian hormone shows a strong positive correlation to the quantitative ovarian reserve but its correlation to embryo quality is unclear. This study assessed the association between serum AMH as a marker of ovarian reserve and embryo quality, using the technology of time-lapse imaging of the embryos in women undergoing in vitro fertilisation (IVF) treatment.

METHODS

304 embryos from 198 women undergoing IVF were included in the study. Serum AMH was assessed for all women. Embryo quality was assessed with the known implantation data (KID) score generated by the time-lapse imaging system.

RESULTS

There was no statistically significant difference in mean serum AMH among different KID score categories (p = 0.135). This remained non-significant after controlling for confounding variables (p = 0.305).

CONCLUSIONS

The results of our study show no significant association between serum AMH and embryo quality in women undergoing IVF treatment when embryo quality was assessed using the KID scores generated by time-lapse imaging which is a better method of embryo assessment rather than conventional morphological assessment.

Do morphokinetic data sets inform pregnancy potential?

Purpose

The aim of this study was to create a model to predict the implantation of transferred embryos based on information contained in the morphokinetic parameters of time-lapse monitoring.

Methods

An analysis of time-lapse recordings of 410 embryos transferred in 343 cycles of in vitro fertilization (IVF) treatment was performed. The study was conducted between June 2012 and November 2014. For each embryo, the following data were collected: the duration of time from the intracytoplasmic sperm injection (ICSI) procedure to further division for two, three, four, and five blastomeres, time intervals between successive divisions, and the level of fragmentation assessed in successive time-points. Principal component analysis (PCA) and logistic regression were used to create a predictive model.

Results

Based on the results of principal component analysis and logistic regression analysis, a predictive equation was constructed. Statistically significant differences (p < 0.001) in the size of the created parameter between the implanted group (the median value: Me = −5.18 and quartiles: Q₁ = −5.61; Q₃ = −4.79) and the non-implanted group (Me = −5.69, Q₁ = −6.34; Q₃ = −5.16) were found. A receiver operating characteristic (ROC) curve constructed for the considered model showed the good quality of this predictive equation. The area under the ROC curve was AUC = 0.70 with a 95 % confidence interval (0.64, 0.75). The presented model has been validated on an independent data set, illustrating that the model is reliable and repeatable.

Conclusions

Morphokinetic parameters contain information useful in the process of creating pregnancy prediction models. However, embryo quality is not the only factor responsible for implantation, and, thus, the power of prediction of the considered model is not as high as in models for blastocyst formation. Nevertheless, as illustrated by the results of this study, the application of advanced data-mining methods in reproductive medicine allows one to create more accurate and useful models.