Improving IVF Utilization with Patient-Centric Artificial Intelligence-Machine Learning (AI/ML): A Retrospective Multicenter Experience.

Objectives: In vitro fertilization (IVF) has the potential to give babies to millions more people globally, yet it continues to be underutilized. We established a globally applicable and locally adaptable IVF prognostics report and framework to support patient-provider counseling and enable validated, data-driven treatment decisions. This study investigates the IVF utilization rates associated with the usage of machine learning, center-specific (MLCS) prognostic reports (the Univfy® report) in provider-patient pre-treatment and IVF counseling. Methods: We used a retrospective cohort comprising 24,238 patients with new patient visits (NPV) from 2016 to 2022 across seven fertility centers in 17 locations in seven US states and Ontario, Canada. We tested the association of Univfy report usage and first intra-uterine insemination (IUI) and/or first IVF usage (a.k.a. conversion) within 180 days, 360 days, and "Ever" of NPV as primary outcomes. Results: Univfy report usage was associated with higher direct IVF conversion (without prior IUI), with odds ratios (OR) 3.13 (95% CI 2.83, 3.46), 2.89 (95% CI 2.63, 3.17), and 2.04 (95% CI 1.90, 2.20) and total IVF conversion (with or without prior IUI), OR 3.41 (95% CI 3.09, 3.75), 3.81 (95% CI 3.49, 4.16), and 2.78 (95% CI 2.59, 2.98) in 180-day, 360-day, and Ever analyses, respectively; p < 0.05. Among patients with Univfy report usage, after accounting for center as a factor, older age was a small yet independent predictor of IVF conversion. Conclusions: Usage of a patient-centric, MLCS-based prognostics report was associated with increased IVF conversion among new fertility patients. Further research to study factors influencing treatment decision making and real-world optimization of patient-centric workflows utilizing the MLCS reports is warranted.

An artificial intelligence tool predicts blastocyst development from static images of fresh mature oocytes.

RESEARCH QUESTION: Can a deep learning image analysis model be developed to assess oocyte quality by predicting blastocyst development from images of denuded mature oocytes? DESIGN: A deep learning model was developed utilizing 37,133 static oocyte images with associated laboratory outcomes from eight fertility clinics (six countries). A subset of data (n = 7807) was allocated to test model performance. External model validation was conducted to assess generalizability and robustness on new data (n = 12,357) from two fertility clinics (two countries). Performance was assessed by calculating area under the curve (AUC), balanced accuracy, specificity and sensitivity. Subgroup analyses were performed on the test dataset for age group, male factor and geographical location of the clinic. Model probabilities of the external dataset were converted to a 0-10 scoring scale to facilitate analysis of correlation with blastocyst development and quality. RESULTS: The deep learning model demonstrated AUC of 0.64, balanced accuracy of 0.60, specificity of 0.55 and sensitivity of 0.65 on the test dataset. Subgroup analyses displayed the highest performance for age group 38-39 years (AUC 0.68), a negligible impact of male factor, and good model generalizability across geographical locations. Model performance was confirmed on external data: AUC of 0.63, balanced accuracy of 0.58, specificity of 0.57 and sensitivity of 0.59. Analysis of the scoring scale revealed that higher scoring oocytes correlated with higher likelihood of blastocyst development and good-quality blastocyst formation. CONCLUSION: The deep learning model showed a favourable performance for the evaluation of oocytes in terms of competence to develop into a blastocyst, and when the predictions were converted into scores, they correlated with blastocyst quality. This represents a significant first step in oocyte evaluation for scientific and clinical applications.

A novel system for rapid conversion of Gardner embryo grades to linear scale numeric variables.

RESEARCH QUESTION: Can Gardner embryo grades be converted to numeric interval variables to improve the incorporation of embryo grading in statistical analyses? DESIGN: An equation that can be used to convert Gardner embryo grades to regular interval scale variables was developed: the numerical embryo quality scoring index (NEQsi). The NEQsi system was then validated with a retrospective chart analysis assessing IVF cycles (n = 1711) conducted at a single Canadian fertility centre between 2014 and 2022. Gardner embryo grades on file were assigned using EmbryoScope and converted to NEQsi scores. Descriptive statistics, univariate logistic regressions and generalized estimating equations with cycle outcomes were prepared to demonstrate the relationship between NEQsi score and probability of pregnancy. RESULTS: NEQsi produces interval numerical scores that range from 2 to 11. Patient case files in which single embryo transfers occurred (n = 1711) were examined and the Gardner embryo grades on file were converted to NEQsi scores. NEQsi scores ranged from 3 to 11, with a median score of 9. A positive linear relationship existed between the NEQsi scores and the probability of pregnancy (as assessed by quantitative ß-HCG). The NEQsi score was a significant predictor of pregnancy (P < 0.001). CONCLUSION: Gardner embryo grades can be converted to interval variables and used directly statistical analyses.

Preimplantation genetic testing for aneuploidy in patients with low embryo numbers: benefit or harm?

PURPOSE: We sought to explore the utility of preimplantation genetic testing for aneuploidy (PGT-A) in a poor prognosis group of women with few embryos available for transfer. METHODS: This was a retrospective matched cohort study examining records for first or second-cycle IVF patients with 1 to 3 blastocysts. The study group comprised 130 patients who underwent PGT-A on all embryos. The control group included 130 patients matched by age, BMI, and blastocyst number and quality who did not undergo PGT-A during the same time period. RESULTS: The live birth rate (LBR) per embryo transfer (ET) were similar in the PGT-A and control groups, and the spontaneous abortion (SAB) rate was the same (23%). However, we found a significantly higher LBR per oocyte retrieval in the control group vs the PGT-A group (43% vs 20%, respectively) likely due to the many no-euploid cycles in the PGT-A group. In a subgroup analysis for age, the similar LBR per ET persisted in women < 38. However, in older women, there was a trend to a higher LBR per ET in the PGT-A group (43%) vs the control group (22%) but a higher LBR per oocyte retrieval in the control group (31%) vs the PGT-A group (13%). CONCLUSIONS: Overall, we observed a significant increase in LBR per oocyte retrieval in women in the control group compared to women undergoing PGT-A, and no difference in SAB rate. Our data suggests that PGT-A has no benefit in a subpopulation of women with few embryos and may cause harm.

Cleavage vs blastocyst stage embryos: how are they interrelating?

PURPOSE: To assess the variables that may predict which cleavage-stage embryo may develop into a blastocyst, and vice versa, to determine whether the cleavage-stage embryo morphology should be taken into consideration when transferring the embryo at the blastocyst stage. METHODS: A single center, retrospective cohort study. The study cohort included 3072 patients undergoing 3607 retrieval cycles and 23,124 embryos at the cleavage stage. We assessed the blastulation rate and evaluated which variables impact the ongoing pregnancy rate. RESULTS: High blastulation rate correlates with higher embryos' grading (I > II > III > IV > V) and higher number of blastomeres (8 > 7 > 6 > 5 > 4). 949 patients had fresh single blastocyst transfers. The ongoing pregnancy rate was 28.9% per transfer. Patients with ongoing pregnancies were significantly younger (34.3 vs. 36 years, p < 0.001), had higher number of oocyte yield (9.8 vs. 9, p = 0.02), and an increased rate of good-quality embryos transferred (70.7% vs. 47.7%, p = 0.001). When evaluating embryos progression, we found that whenever embryo developed to a good-quality blastocyst, its appearance at the cleavage stage did not affect ongoing pregnancy rate. CONCLUSION: Higher the number of blastomeres and better embryo grading were found to correlate with a higher blastulation rate. Nevertheless, if the embryo has already developed to a top-quality blastocyst, its morphology at the cleavage stage did not impact ongoing pregnancy rate.

Can Oocyte Diameter Predict Embryo Quality?

With the recent increased utilization of oocyte vitrification for the purpose of fertility preservation, information regarding the future fertility potential of the frozen oocytes is mandatory. Nowadays, there is a relative lack of data about prediction of assisted reproductive technique (ART) success relying on the retrieved oocytes. In the present study, we therefore aimed to investigate whether oocyte diameter might predict the quality of the developing embryo. A retrospective, single-center cohort study. Oocytes retrieved following controlled ovarian hyperstimulation cycles during 2016 and incubated in a time-lapse incubator system were analyzed. Oocytes were grouped by mean oocyte diameter (MOD) and incubated for 5 days before the final morphological evaluation done by an expert embryologist. A total of 471 cycles which yielded 3355 metaphase II oocytes were included in the analysis. Embryos developed from oocytes with MOD close to the average (Average 1SD < MOD < Average + 1SD) had increased good-quality blastulation rates compared with embryos that developed from very small or very large oocytes. Oocytes with MOD between 105.96 and 118.69 µm have better probability of becoming top-quality D5 blastocysts (17.1-17.4% grade 1 embryos). There is a correlation between oocyte's MOD and the embryo quality at day 5. The oocytes with near average MOD have a better chance to develop to a good-quality embryo. Therefore, the study suggests that MOD might serve as a predictor for embryo grading at day 5.

What is the optimal timing of embryo transfer when there are only one or two embryos at cleavage stage?

Today, most IVF programs have moved to blastocyst transfer but there is still uncertainty regarding when to transfer if there are only one or two embryos at the cleavage stage. The aim of this study was to compare the pregnancy rate of day 3 transfers vs. blastocyst stage transfers in patients who had only one or two embryos on day 3. We conducted a retrospective study of 102 patients with one or two cleavage stage embryos that had their embryos transferred on day 3 and 429 patients had their embryos cultured to day 5 for transfer. The number of mature oocytes (4.0 vs 4.6, p = NS) and number of cleavage stage embryos on day 3 was similar in the two groups (1.3 vs. 1.5, p = NS). The clinical pregnancy rate per retrieval (22% vs. 24.6%, p= NS) and the ongoing pregnancy rate per retrieval (20% vs. 20.2%, p = NS) was comparable between the groups. Fifty seven (13.2%) of the patients had cleavage embryo arrest and did not have an embryo to transfer on day 5. We conclude that the cumulative pregnancy rate is the same for patients with 1-2 cleavage stage embryos regardless of whether the embryo is transferred on day 3 or day 5.

Developmental potential of slow-developing embryos: day-5 morulae compared with day-5 cavitating morulae.

OBJECTIVE: To describe and compare the ongoing pregnancy rate between morulae and cavitating morulae (CAVM) transferred on day 5, to describe and compare the blastulation rate between day 5 morulae and CAVM, and to describe the pregnancy rate of these slow-developing blastocysts during a frozen embryo transfer (FET) cycle. DESIGN: Retrospective cohort study. SETTING: Single tertiary care medical center. PATIENT(S): Delayed-development embryos: 3,321 cycles that included 10,304 embryos on day 5 that were cultured until day 6. INTERVENTION(S): Development of morula and CAVM to the blastocyst stage. MAIN OUTCOME MEASURE(S): Blastulation rate. RESULT(S): The fresh embryo transfers comprised 186 patients with 82 embryos at the morula stage and 104 embryos at the CAVM stage. The pregnancy rate (15.8% vs. 21.1%) and the ongoing pregnancy rate (15.8% vs. 17.3%) were comparable between the groups. The study group included 10,304 day-5 delayed embryos: 5,395 morulae and 4,909 CAVM on day 5. The blastulation rate was statistically significantly higher in the CAVM group compared with the morula group (39.2% vs. 20.4%). We included 201 FET cycles: 77 warmed blastocysts that developed from a morula on day 5 and 124 warmed blastocysts that developed from CAVM on day 5. The clinical pregnancy rate was comparable between the two groups per embryo transfer (21.3% vs. 24.7%). CONCLUSION(S): Transferring of fresh, slow-developing embryos seems to improve the cycle outcomes compared with culturing for another day and then vitrifying and thawing later.

Prolonged culture of blastocysts after thawing as a tool for improving prediction of success.

PURPOSE: A few years ago, we started to use a new freeze-thaw protocol for the frozen embryo transfer cycles. Instead of thawing the embryos 2-4 h prior to the transfer, we started thawing the embryos 20-22 h prior to the transfer. The aim of this study was to compare the pregnancy rate in cases of embryos that continued to develop in the post-thawing culture to that of embryos that did not. METHODS: A retrospective cohort study of blastocyst freeze/thaw cycles vitrified on day 5, thawed and transferred after 20-22 h in the culture, between January 2012 and December 2016. RESULTS: A total of 375 patients were included in the analysis. Two hundred twenty-eight embryos graded as good, 87 graded as fair, and 60 graded as poor embryos were transferred. The clinical pregnancy rate (50% vs. 19.5% vs 3.3% p < 0.01) and the ongoing pregnancy rate (38.5% vs. 13.6% vs 1.7% p < 0.01) were higher in cases of good embryo quality compared with fair and poor-quality embryos, respectively. For good embryos, progressing to a better grade during the culture did not change the clinical pregnancy rate (51.3% vs. 46.2% p = NS) or the ongoing pregnancy rate (38.5% vs. 37.5% p = NS). For fair embryos, progressing to a better grade during the culture resulted in a higher clinical pregnancy rate (25.4% vs 9% p = 0.05). CONCLUSIONS: The development of the fair embryos in the culture has a highly positive impact on the pregnancy rate and this factor should be taken into consideration before deciding how many embryos to transfer.

The VEGF and PEDF levels in the follicular fluid of patients co- treated with LETROZOLE and gonadotropins during the stimulation cycle.

BACKGROUND: Previous studies have shown that androgens, in addition to serving as precursors for ovarian estrogen synthesis, also have a fundamental role in primate ovarian follicular development by augmentation of FSH receptor expression on granulosa cells. Recent studies have shown that aromatase inhibitor, letrozole, improves ovarian response to FSH in normal and poor responder patients, possibly by increasing intraovarian androgen levels. Studies in mice also showed an effect of letrozole to increase pigment epithelium-derived factor (PEDF) and to lower vascular epithelial growth factor (VEGF), which might be expected to reduce the risk of ovarian hyperstimulation syndrome (OHSS) with stimulation. The aim of this study was to compare the VEGF and PEDF levels in the follicular fluids of normal responders treated with letrozole and gonadotropins during the ovarian stimulation with patients treated with gonadotropins only. METHODS: A single center, prospective clinical trial. We collected follicular fluid from 26 patients, on a GnRH antagonist protocol, dual triggered with hCG and GnRH agonist. The patients in one group were co-treated with letrozole and gonadotropins during the ovarian stimulation and the patients in the other group were treated with gonadotropins only. VEGF, PEDF, estrogen, progesterone and testosterone levels were measured by ELISA kits. RESULTS: The age of the patients, the total dose of gonadotropins and the number of oocytes were comparable between the two groups. In the follicular fluid, the estrogen levels (2209 nmol/l vs. 3280 nmol/l, p = 0.02) were significantly decreased, and the testosterone levels (246.5 nmol/l vs. 40.7 nmol/l, p < 0.001) were significantly increased in the letrozole group compared to the gonadotropin only group. The progesterone levels (21.4 µmol/l vs. 17.5 p = NS) were comparable between the two groups. The VEGF levels (2992 pg/ml vs. 1812 pg/ml p = 0.02) were significantly increased and the PEDF levels (9.7 ng/ml vs 17.3 ng/ml p < 0.001) were significantly decreased in the letrozole group. CONCLUSIONS: Opposite to observations in the mouse, we found that VEGF levels were increased and PEDF levels were decreased in the follicular fluid in patients treated with letrozole during the stimulation cycles. Further investigation is required to determine if patients treated with letrozole during the IVF stimulation protocol are at increased risk for developing OHSS as a result of these findings.

Does daily co-administration of letrozole and gonadotropins during ovarian stimulation improve IVF outcome?

BACKGROUND: For the last year we have been treating normal responders with gonadotropins and letrozole during the whole stimulation in order to improve response to FSH by increasing the intrafollicular androgen concentration, and to reduce circulating estrogen concentrations. The aim of this study was to compare the IVF outcome of normal responders treated with letrozole and gonadotropins during ovarian stimulation with patients treated with gonadotropins only. METHODS: A single centre retrospective cohort study of 174 patients (87 in each group). RESULTS: The age of the patients was comparable between the groups. Estradiol levels were significantly higher in the control group (6760 pmol/L vs. 2420 pmol/L respectively, p < 0.01), and the number of follicles ≥15 mm at the trigger day was significantly lower in the control group (7.9 vs. 10, p = 0.02). The number of retrieved oocytes (10 vs. 14.5, p < 0.01), MII oocytes (7.9 vs. 11.2, p < 0.01) and blastocysts (2.7 vs. 4.0, p = 0.02) was significantly higher in the study group. We found no significant differences in the cumulative pregnancy outcome between the two groups (65.2% vs 58.3% p = NS). CONCLUSIONS: We conclude that co-treatment with letrozole improves the IVF outcome in normal responders in terms of increased number of blastocysts obtained without increasing the pregnancy rate or the risk of OHSS.

Clinical pregnancy rate following frozen embryo transfer is higher with blastocysts vitrified on day 5 than on day 6.

PURPOSE: The aim of this study was to compare the pregnancy rates between good quality blastocysts vitrified on day 6 versus blastocysts vitrified on day 5 after fertilization. METHODS: This is a retrospective cohort study of 791 freeze-thaw cycles of blastocysts vitrified either on day 5 or on day 6 and transferred between January 2012 and October 2015. Five hundred and thirty-seven cycles included blastocysts vitrified on day 5, and 254 cycles included blastocysts vitrified on day 6. RESULTS: The age of the patients and the proportion of embryos that survived the thawing process were comparable between the two groups. More good quality embryos were transferred in the group in which blastocysts were vitrified on day 6 (1.2 vs. 1.3, p = 0.005), but the clinical pregnancy rate (44 vs. 33 %, p = 0.002) and the ongoing pregnancy rate (41 vs. 28 %, p < 0.001) were higher in the group in which blastocysts were vitrified on day 5. Multivariate regression analysis adjusting for patient's age, number of good quality embryos transferred (≥3BB), and treatment protocol demonstrated that the day 6 vitrified group had a significantly lower clinical pregnancy rate compared to the day 5 vitrified group (OR 0.54, 95 % CI 0.38-0.76). CONCLUSIONS: The clinical pregnancy rate following frozen embryo transfer is significantly lower with blastocysts vitrified on day 6 compared to blastocysts vitrified on day 5.

Human sperm rheotaxis: a passive physical process.

A long-standing question in natural reproduction is how mammalian sperm navigate inside female reproductive tract and finally reach the egg cell, or oocyte. Recently, fluid flow was proposed as a long-range guidance cue for sperm navigation. Coitus induces fluid flow from oviduct to uterus, and sperm align themselves against the flow direction and swim upstream, a phenomenon termed rheotaxis. Whether sperm rheotaxis is a passive process dominated by fluid mechanics, or sperm actively sense and adapt to fluid flow remains controversial. Here we report the first quantitative study of sperm flagellar motion during human sperm rheotaxis and provide direct evidence indicating that sperm rheotaxis is a passive process. Experimental results show that there is no significant difference in flagellar beating amplitude and asymmetry between rheotaxis-turning sperm and those sperm swimming freely in the absence of fluid flow. Additionally, fluorescence image tracking shows no Ca(2+) influx during sperm rheotaxis turning, further suggesting there is no active signal transduction during human sperm rheotaxis.