A hybrid artificial intelligence model leverages multi-centric clinical data to improve fetal heart rate pregnancy prediction across time-lapse systems.

STUDY QUESTION: Can artificial intelligence (AI) algorithms developed to assist embryologists in evaluating embryo morphokinetics be enriched with multi-centric clinical data to better predict clinical pregnancy outcome? SUMMARY ANSWER: Training algorithms on multi-centric clinical data significantly increased AUC compared to algorithms that only analyzed the time-lapse system (TLS) videos. WHAT IS KNOWN ALREADY: Several AI-based algorithms have been developed to predict pregnancy, most of them based only on analysis of the time-lapse recording of embryo development. It remains unclear, however, whether considering numerous clinical features can improve the predictive performances of time-lapse based embryo evaluation. STUDY DESIGN, SIZE, DURATION: A dataset of 9986 embryos (95.60% known clinical pregnancy outcome, 32.47% frozen transfers) from 5226 patients from 14 European fertility centers (in two countries) recorded with three different TLS was used to train and validate the algorithms. A total of 31 clinical factors were collected. A separate test set (447 videos) was used to compare performances between embryologists and the algorithm. PARTICIPANTS/MATERIALS, SETTING, METHODS: Clinical pregnancy (defined as a pregnancy leading to a fetal heartbeat) outcome was first predicted using a 3D convolutional neural network that analyzed videos of the embryonic development up to 2 or 3 days of development (33% of the database) or up to 5 or 6 days of development (67% of the database). The output video score was then fed as input alongside clinical features to a gradient boosting algorithm that generated a second score corresponding to the hybrid model. AUC was computed across 7-fold of the validation dataset for both models. These predictions were compared to those of 13 senior embryologists made on the test dataset. MAIN RESULTS AND THE ROLE OF CHANCE: The average AUC of the hybrid model across all 7-fold was significantly higher than that of the video model (0.727 versus 0.684, respectively, P = 0.015; Wilcoxon test). A SHapley Additive exPlanations (SHAP) analysis of the hybrid model showed that the six first most important features to predict pregnancy were morphokinetics of the embryo (video score), oocyte age, total gonadotrophin dose intake, number of embryos generated, number of oocytes retrieved, and endometrium thickness. The hybrid model was shown to be superior to embryologists with respect to different metrics, including the balanced accuracy (P ≤ 0.003; Wilcoxon test). The likelihood of pregnancy was linearly linked to the hybrid score, with increasing odds ratio (maximum P-value = 0.001), demonstrating the ranking capacity of the model. Training individual hybrid models did not improve predictive performance. A clinic hold-out experiment was conducted and resulted in AUCs ranging between 0.63 and 0.73. Performance of the hybrid model did not vary between TLS or between subgroups of embryos transferred at different days of embryonic development. The hybrid model did fare better for patients older than 35 years (P < 0.001; Mann-Whitney test), and for fresh transfers (P < 0.001; Mann-Whitney test). LIMITATIONS, REASONS FOR CAUTION: Participant centers were located in two countries, thus limiting the generalization of our conclusion to wider subpopulations of patients. Not all clinical features were available for all embryos, thus limiting the performances of the hybrid model in some instances. WIDER IMPLICATIONS OF THE FINDINGS: Our study suggests that considering clinical data improves pregnancy predictive performances and that there is no need to retrain algorithms at the clinic level unless they follow strikingly different practices. This study characterizes a versatile AI algorithm with similar performance on different time-lapse microscopes and on embryos transferred at different development stages. It can also help with patients of different ages and protocols used but with varying performances, presumably because the task of predicting fetal heartbeat becomes more or less hard depending on the clinical context. This AI model can be made widely available and can help embryologists in a wide range of clinical scenarios to standardize their practices. STUDY FUNDING/COMPETING INTEREST(S): Funding for the study was provided by ImVitro with grant funding received in part from BPIFrance (Bourse French Tech Emergence (DOS0106572/00), Paris Innovation Amorçage (DOS0132841/00), and Aide au Développement DeepTech (DOS0152872/00)). A.B.-C. is a co-owner of, and holds stocks in, ImVitro SAS. A.B.-C. and F.D.M. hold a patent for 'Devices and processes for machine learning prediction of in vitro fertilization' (EP20305914.2). A.D., N.D., M.M.F., and F.D.M. are or have been employees of ImVitro and have been granted stock options. X.P.-V. has been paid as a consultant to ImVitro and has been granted stocks options of ImVitro. L.C.-D. and C.G.-S. have undertaken paid consultancy for ImVitro SAS. The remaining authors have no conflicts to declare. TRIAL REGISTRATION NUMBER: N/A.

Cell-free DNA in human follicular fluid as a biomarker of embryo quality.

STUDY QUESTION: Could cell-free DNA (cfDNA) quantification in individual human follicular fluid (FF) samples become a new non-invasive predictive biomarker for in vitro fertilization (IVF) outcomes? SUMMARY ANSWER: CfDNA level in human follicular fluid samples was significantly correlated with embryo quality and could be used as an innovative non-invasive biomarker to improve IVF outcomes. WHAT IS KNOWN ALREADY: CfDNA fragments, resulting from apoptotic or necrotic events, are present in the bloodstream and their quantification is already used as a biomarker for gynaecological and pregnancy disorders. Follicular fluid is important for oocyte development and contains plasma components and factors secreted by granulosa cells during folliculogenesis. CfDNA presence in follicular fluid and its potential use as an IVF outcome biomarker have never been investigated. STUDY DESIGN, SIZE, DURATION: One hundred individual follicular fluid samples were collected from 43 female patients undergoing conventional IVF (n = 26) or ICSI (n = 17). CfDNA level was quantified in each individual follicular fluid sample. PARTICIPANTS/MATERIALS, SETTING, METHODS: At oocyte collection day, follicles were aspirated individually. Only blood-free follicular fluid samples were included in the study. Follicle size was calculated based on the follicular fluid volume. Each corresponding cumulus-oocyte complex was isolated for IVF or ICSI procedures. Follicular fluid cfDNA was measured by quantitative PCR with ALU-specific primers. MAIN RESULTS AND THE ROLE OF CHANCE: Human follicular fluid samples from individual follicles contain measurable amounts of cfDNA (mean ± SD, 1.62 ± 2.08 ng/µl). CfDNA level was significantly higher in small follicles (8-12 mm in diameter) than in large ones (>18 mm) (mean ± SD, 2.54 ± 0.78 ng/µl versus 0.71 ± 0.44 ng/µl, respectively, P = 0.007). Moreover, cfDNA concentration was significantly and negatively correlated with follicle size (r = -0.34; P = 0.003). A weak significant negative correlation between DNA integrity and 17ß-estradiol level in follicular fluid samples at oocyte collection day was observed (r = -0.26; P = 0.008). CfDNA level in follicular fluid samples corresponding to top quality embryos was significantly lower than in follicular fluid samples related to poor quality embryos (P = 0.022). Similarly, cfDNA level was also significantly lower in follicular fluid samples related to embryos with low fragmentation rate (≤25%) than with high fragmentation rate (>25%) (P = 0.02). LIMITATIONS, REASONS FOR CAUTION: A larger study should be conducted in order to establish the predictive value of cfDNA level for embryo quality and to investigate whether follicular fluid cfDNA levels are correlated with embryo implantation rates and pregnancy outcomes. Moreover, the role of follicular fluid cfDNA on embryo quality should be studied to determine whether high cfDNA concentration in follicular fluid is only a consequence or also a cause of follicular dysfunction. WIDER IMPLICATIONS OF THE FINDINGS: CfDNA evaluation in individual follicular fluid samples might represent an innovative biomarker of embryo quality to use as a supplemental tool to predict embryo quality during IVF. STUDY FUNDING/COMPETING INTERESTS: This study was partially supported by the University Hospital of Montpellier and Ferring Pharmaceuticals. The authors of the study have no competing interests to report.

Male infertility: an obstacle to sexuality?

Interactions between infertility and sexuality are numerous and complex. Infertile men may suffer from sexual dysfunction (SD) when undergoing an assisted reproductive technology programme. We undertook a review both in French and English of the available data on male SD when being diagnosed with a fertility problem with a specific focus on azoospermic men. The review was performed over a 30-year time period using PubMed/Medline. The sexual concerns and needs of infertile/sterile men for whom potential parenting can be compromised were evaluated. When diagnosed with infertility, men usually go through a crisis that can have a deleterious effect on their sexuality with sometimes a feeling of sexual inadequacy. Infertile men will feel stigmatized because they are perceived as being deficient in a specific component of their masculinity. Hence, subsequent SD may occur that can impact the couple sexuality and the infertility management. However, little is known on how the announcement of azoospermia may affect male on a sexual and psychological point of view. The present review suggests that a global management through a healthcare network (biologist, andrologist, sexologist and psychologist) is required which will allow to consider infertility and its subsequent sexual disorders as a whole and not as dichotomized issues.

Circulating microRNAs in follicular fluid, powerful tools to explore in vitro fertilization process.

Circulating or "extracellular" microRNAs (miRNAs) detected in biological fluids, could be used as potential diagnostic and prognostic biomarkers of several disease, such as cancer, gynecological and pregnancy disorders. However, their contributions in female infertility and in vitro fertilization (IVF) remain unknown. This study investigated the expression profiles of five circulating miRNAs (let-7b, miR-29a, miR-30a, miR-140 and miR-320a) in human follicular fluid from 91 women with normal ovarian reserve and 30 with polycystic ovary syndrome (PCOS) and their ability to predict IVF outcomes. The combination of FF miR-30a, miR-140 and let-7b expression levels discriminated between PCOS and normal ovarian reserve with a specificity of 83.8% and a sensitivity of 70% (area under the ROC curve, AUC = 0.83 [0.73-0.92]; p < 0.0001). FF samples related to low number of mature oocytes (≤2) contained significant less miR-320a levels than those related to a number of mature oocytes >2 (p = 0.04). Moreover, FF let-7b predicted the development of expanded blastocysts with 70% sensitivity and 64.3% specificity (AUC = 0.67 [0.54-0.79]; p = 0.02) and FF miR-29a potential to predict clinical pregnancy outcome reached 0.68 [0.55-0.79] with a sensitivity of 83.3% and a specificity of 53.5% (p = 0.01). Therefore, these miRNAs could provide new helpful biomarkers to facilitate personalized medical care during IVF.

[Comparison of frozen embryo transfer outcomes at blastocyst stage according to freezing method and type of endometrial preparation]. / Comparaison des issues de transferts d'embryons congelés au stade blastocyste selon la technique de congélation et le type de préparation endométriale.

OBJECTIVE: This study intended to compare frozen embryo transfer (FET) outcomes at blastocyst stage according to freezing methods, slow freezing versus vitrification and according to the type of endometrial preparation. PATIENTS AND METHODS: A total of 172 FET at blastocyst stage (day 5 or 6) were included retrospectively from April, 2007 to December, 2012. The FET outcomes from slow freezing (group 1, n=86) were compared with those from vitrification (group 2, n=86). More particularly, the survival rate after thawing, as well as implantation and pregnancy rates (clinical and ongoing pregnancy rates) were compared respectively between these two groups, after matching on women's age at freezing day, embryo number and embryo development stage for transfer. Furthermore, for each freezing method, FET outcomes were compared according to the type of endometrial preparation, i.e. natural cycle (group N) versus stimulated cycle (group S). RESULTS: The survival rate as well as implantation and clinical pregnancy rates were significantly higher for FET after vitrification compared to FET after slow freezing (97% vs 85%, P<0.0001; 32% vs 20%, P=0.02; 43% vs 28%, P=0.04, respectively). By taking into account the number of transferred embryos for each group, the multiple pregnancy rate was three-fold higher in the group of FET after vitrification compared to the group of FET after slow freezing but not significantly (27.3% vs 8.3%, NS). However, FET outcomes were not affected significantly by the type of endometrial preparation whatever freezing methods. Nevertheless, the early spontaneous abortion (ESA) rate was lower in the case of embryos that were frozen by vitrification and transferred in natural cycle (group N2 vs group S2: 20% vs 47%, NS). DISCUSSION AND CONCLUSION: Our study confirms that the survival rate after thawing at blastocyst stage (day 5 or 6) is significantly improved after freezing by vitrification compared to slow freezing method. Likewise, implantation and clinical pregnancy rates are significantly increased in the case of FET at blastocyst stage when these embryos were frozen by vitrification. The results obtained by vitrification are very satisfactory but are also associated with an increased multiple pregnancy rate. Moreover, FET associated with natural or stimulated cycle does not modify significantly the outcomes of attempts, whatever the freezing method. However, the risk of ESA is reduced in the case of FET with natural cycle and after embryo vitrification.

[Circulating nucleic acids and infertility]. / Les acides nucléiques circulants et infertilité.

Circulating nucleic acids (cell-free DNA and microRNAs) have for particularity to be easily detectable in the biological fluids of the body. Therefore, they constitute biomarkers of interest in female and male infertility care. Indeed, in female, they can be used to detect ovarian reserve disorders (polycystic ovary syndrome and low functional ovarian reserve) as well as to assess follicular microenvironment quality. Moreover, in men, their expression levels can vary in case of spermatogenesis abnormalities. Finally, circulating nucleic acids have also the ability to predict successfully the quality of in vitro embryo development. Their multiple contributions during assisted reproductive technology (ART) make of them biomarkers of interest, for the development of new diagnostic and/or prognostic tests, applied to our specialty. Circulating nucleic acids would so offer the possibility of personalized medical care for infertile couples in ART.

[Does embryo morphology constitute a reliable criterion for embryo selection?]. / La morphologie embryonnaire a-t-elle encore un intérêt à l'heure actuelle ?

Embryo selection based exclusively on embryo morphological criteria does not allow currently to obtain satisfactory implantation rates. So, new time-lapse systems have been proposed to improve the rates of in vitro fertilization success. However, the profit/investment ratio of time-lapse systems and the interest of embryo morphokinetics evaluation remain clearly to be established by clinical, robust, prospective, and randomized trials. Consequently, morphological and morphokinetic approaches showed their limitations and justify the development of new non-invasive "omics" approaches. These approaches are very promising and have for main aim to identify new non-invasive biomarkers, in oocyte microenvironment and/or in embryo culture medium, predictive of oocyte and/or embryo quality as well as pregnancy rates. These approaches thus open the way to develop new diagnostic and/or prognostic tests for embryo viability based on the expression and/or the quantification of biomarkers of interest in assisted reproductive technology.

[Circulating nucleic acids and in vitro fertilization]. / Acides nucléiques circulants et fécondation in vitro.

During the last years, the use of circulating nucleic acids (microRNAs and cell-free DNA) as diagnostic and/or prognostic tools in cancerology was widely documented. Likewise, in obstetrics and gynecology, the development of non-invasive prenatal testing based on the assessment of these biomarkers confirmed their growing interest in this speciality. In human reproduction, several studies were interested in the microRNAs, small non-coding RNA sequences, present in the ovarian follicle and their implication in folliculogenesis. Some of these microRNAs, as well as the vesicles which transport them, are easily detectable in the bloodstream and could be used as reliable biomarkers of interest in infertility care. Cell-free DNA level varies according to physiopathology and reflect the proportion of apoptotic and/or necrotic events occurring in the body. As a result, its quantification could give an additional help to the practitioners for ovarian functional status evaluation. Furthermore, these circulating nucleic acids could also constitute new predictive biomarkers of oocyte and/or embryo quality and represent a promising perspective for the prevention of in vitro fertilization implantation failures. In conclusion, these circulating nucleic acids open the way to the development of new diagnostic and/or prognostic innovative tests in order to improve in vitro fertilization outcomes.

Cell-free nucleic acids as non-invasive biomarkers of gynecological cancers, ovarian, endometrial and obstetric disorders and fetal aneuploidy.

BACKGROUND: Proper folliculogenesis is fundamental to obtain a competent oocyte that, once fertilized, can support the acquisition of embryo developmental competence and pregnancy. MicroRNAs (miRNAs) are crucial regulators of folliculogenesis, which are expressed in the cumulus-oocyte complex and in granulosa cells and some can also be found in the bloodstream. These circulating miRNAs are intensively studied and used as diagnostic/prognostic markers of many diseases, including gynecological and pregnancy disorders. In addition, serum contains small amounts of cell-free DNA (cfDNA), presumably resulting from the release of genetic material from apoptotic/necrotic cells. The quantification of nucleic acids in serum samples could be used as a diagnostic tool for female infertility. METHODS: An overview of the published literature on miRNAs, and particularly on the use of circulating miRNAs and cfDNA as non-invasive biomarkers of gynecological diseases, was performed (up to January 2014). RESULTS: In the past decade, cell-free nucleic acids have been studied for potential use as biomarkers in many diseases, particularly in gynecological cancers, ovarian and endometrial disorders, as well as in pregnancy-related pathologies and fetal aneuploidy. The data strongly suggest that the concentration of cell-free nucleic acids in serum from IVF patients or in embryo culture medium could be related to the ovarian hormone status and embryo quality, respectively, and be used as a non-invasive biomarker of IVF outcome. CONCLUSIONS: The profiling of circulating nucleic acids, such as miRNAs and cfDNA, opens new perspectives for the diagnosis/prognosis of ovarian disorders and for the prediction of IVF outcomes, namely (embryo quality and pregnancy).