Artificial cryopreserved embryo transfer cycle success depends on blastocyst developmental rate and progesterone timing

On the role of artificial intelligence in analysing oocytes during in vitro fertilisation procedures

Nowadays, the most adopted technique to address infertility problems is in vitro fertilisation (IVF). However, its success rate is limited, and the associated procedures, known as assisted reproduction technology (ART), suffer from a lack of objectivity at the laboratory level and in clinical practice. This paper deals with applications of Artificial Intelligence (AI) techniques to IVF procedures. Artificial intelligence is considered a promising tool for ascertaining the quality of embryos, a critical step in IVF. Since the oocyte quality influences the final embryo quality, we present a systematic review of the literature on AI-based techniques used to assess oocyte quality; we analyse its results and discuss several promising research directions. In particular, we highlight how AI-based techniques can support the IVF process and examine their current applications as presented in the literature. Then, we discuss the challenges research must face in fully deploying AI-based solutions in current medical practice. Among them, the availability of high-quality data sets as well as standardised imaging protocols and data formats, the use of physics-informed simulation and machine learning techniques, the study of informative, descriptive yet observable features, and, above all, studies of the quality of oocytes and embryos, specifically about their live birth potential. An improved understanding of determinants for oocyte quality can improve success rates while reducing costs, risks for long-term embryo cultures, and bioethical concerns.

Comparable ongoing pregnancy and pregnancy loss rates in natural cycle and artificial cycle frozen embryo transfers with intensive method-specific luteal phase support; a retrospective cohort study

STUDY OBJECTIVE

The absence of corpus lutea in artificial cycle (AC) frozen embryo transfers (FET) may increase the chances of pregnancy loss. In this retrospective cohort study, the efficacy of AC endometrial preparation was compared natural cycle (NC) endometrial preparation in terms of ongoing pregnancy.

METHODS

One thousand six hundred and eighteen consecutive vitrified-warmed blastocyst FET performed between December 2021 and November 2022 were included, with 1023 compared after exclusions according to the endometrial preparation method; 293 NC-FET, 143 modified NC-FET, 204 unprogrammed AC-FET, and 383 oral contraceptive pill (OCP) programmed AC-FET. Intensive method-specific luteal phase support (LPS) was administered in NC- (human chorionic gonadotropin and micronized vaginal progesterone), mNC- (micronized vaginal progesterone), and in AC-FET (micronized vaginal progesterone, intramuscular progesterone, and oral dydrogesterone).

RESULTS

Clinician choice of endometrial preparation method resulted in the NC- or AC-FET groups having distinct differences, with female age, antral follicle count and body mass index as well as the percentage of DOR or PCOS diagnosed patients significantly different. The unadjusted ongoing pregnancy and total pregnancy loss rates for NC-, mNC-, AC-, and ocp-AC-FET were 61.8 %, 55.2 %, 57.4 %, and 58.5 %, and 19.2 %, 24.0 %, 23.5 % and 23.8 %, respectively. In multivariate logistic regressions to predict the dependent outcomes of ongoing pregnancy and total pregnancy loss, none of the FET methods were selected as independent predictors.

CONCLUSION

Patients undergoing NC- and AC-FET with method-specific progesterone LPS had comparable ongoing pregnancy rates as well as total pregnancy loss rates, with NC-FET ranked first in the regression analysis.

Is there a link between plasma progesterone 1-2 days before frozen embryo transfers (FET) and ART outcomes in frozen blastocyst transfers?

OBJECTIVE

To study the efficacy of combined administration of subcutaneous and vaginal progesterone for priming frozen blastocysts transfers, looking at progesterone levels and ART outcome.

DESIGN

Retrospective study.

SETTING PATIENTS

Three hundred and twenty frozen blastocyst transfer cycles conducted in 213 women aged up to 42 years, BMI between 18 and 30 kg/m², with anatomically normal uterus who underwent frozen embryo transfers (FETs) from February 2019 to December 2019 with a combined luteal-phase support (LPS) associating subcutaneous and vaginal progesterone. Patients with recurrent pregnancy loss (RPL) were excluded.

RESULTS

When using combined vaginal and subcutaneous LPS, SPL >10.50 ng/mL in 95% of cases, with a minimum value of 7.02 ng/mL. CPR, OPR, and global miscarriage rates were 38.4%, 30.9%, and 19.5%, respectively. Analyzing results per quartiles, revealed that miscarriage rates were significantly inferior, and IR were higher in the upper two quartiles of serum progesterone (>21.95 ng/mL) on the day before FET, while there was no difference in CPR and OPR.

CONCLUSIONS

We report ART outcome of frozen blastocyst transfers performed using a combination of vaginal and subcutaneous progesterone for LPS. ART results were honorable and SPL favorable 1-2 days before FET in 99% of cases.

Dydrogesterone versus medroxyprogesterone acetate co-treatment ovarian stimulation for IVF: a matched cohort study of 236 freeze-all-IVF cycles

This matched cohort study was retrospectively performed, with cycles extracted from freeze-all-IVF treatments performed between March and November 2019, to compare the efficacy of flexible-start dydrogesterone (DYG) co-treatment ovarian stimulations (OS) with flexible-start medroxyprogesterone acetate (MPA) co-treatment OS. DYG cycles were matched 1:1 with MPA cycles using female age and antral follicle count, resulting in 236 matched cycles. OS durations and total FSH doses were similar in DYG and MPA OS cycles. The numbers of mature oocytes retrieved were similar; however, the mature oocyte retrieval rate was significantly lower (66.7 vs. 78.2%; p = .001) and the cycle cancellation rates were higher (29.2 vs. 21.2%; p = .056) in DYG co-treatments. A linear regression selected OS co-treatment protocol (0.53 DYG (0.356-0.776), p = .001) into the final model to predict a ≥ 80% mature oocyte retrieval rate. The per transfer (47.2 vs. 49.7; p = .721) and per treatment ongoing pregnancy rates (32.2 vs. 38.1%, p = .210) were similar in the two co-treatment groups. Flexible-start DYG co-treatment OS was as effective in blastocyst freeze-all-IVF cycles as MPA co-treatment, with similar ongoing pregnancy rates; however, mature oocyte retrieval was significantly decreased and cycle cancellation increased in DYG cycles.Impact statementWhat is already known on this subject? Progestin (i.e. artificial progesterone) co-treatment has long been known to be a feasible alternative to conventional GnRH-analogue co-treatment in OS for IVF, because of the long-standing evidence that progestin formulations have in oral contraceptive therapies. The recent evolution of effective freeze-all-IVF (in which high mid-cycle progesterone levels is not of concern because of the postponement of embryo transfer) has now made it possible to investigate progestin co-treatment OS in IVF.What do the results of this study add? Ongoing pregnancy rates from blastocyst frozen embryo transfers in flexible-start dydrogesterone (DYG) co-treatment ovarian stimulation (OS) cycles were similar to rates in flexible-start medroxyprogesterone acetate (MPA) co-treatment OS cycles. The mature oocyte retrieval rate was significantly lower and the cycle cancellation rate higher in DYG than in MPA cycles.What are the implications of these findings for clinical practice and/or further research? The evidence suggests that MPA co-treatment should be preferred in OS for IVF. Further investigation is required to refine progestin co-treatment protocols, because of their potential to reduce the number of viable blastocysts.

Screening and identification of endometrial proteins as novel potential biomarkers for repeated implantation failure

Inadequate endometrial receptivity may be responsible for the low implantation rate of transferred embryos in in vitro fertilization (IVF) treatments. Patients with repeated implantation failure (RIF) impact the clinical pregnancy rate for IVF. We collected endometrial tissue during the implantation window of hysteroscopy biopsies from September 2016 to December 2019 and clinical data were collected simultaneously. Patients were divided into RIF and pregnant controls group according to pregnancy outcomes. A total of 82 differentially expressed endometrial proteins were identified, including 55 up-regulated proteins (>1.50-fold, P < 0.05) and 27 down-regulated proteins (<0.67-fold, P < 0.05) by iTRAQ labeling coupled with the 2D LC MS/MS technique in the RIF group. String analysis found interactions between these proteins which assembled in two bunches: ribosomal proteins and blood homeostasis proteins. The most significant enriched Gene Ontology terms were negative regulation of hydrolase activity, blood microparticle, and enzyme inhibitor activity. Our results emphasized the corticosteroid-binding globulin and fetuin-A as the specific proteins of endometrial receptivity by Western-blot. Our study provided experimental data to establish the objective indicator of endometrial receptivity, and also provided new insight into the pathogenesis of RIF.

Blastocyst age, expansion, trophectoderm morphology, and number cryopreserved are variables predicting clinical implantation in single blastocyst frozen embryo transfers in freeze-only-IVF

PURPOSE

To determine which blastocyst assessment variables predict clinical implantations in single blastocyst frozen embryo transfers (FET) of freeze-only-IVF cycles, following improved vitrified-warmed blastocyst survival and developmental competence preservation.

METHOD

In this retrospective cohort study performed at a single private IVF center, the pregnancy outcomes of 1795 single blastocyst FET cycles were analyzed, from freeze-only-IVF retrievals performed between January 2017 and January 2020. Stepwise forward logistic regressions with clinical implantation (i.e., normal gestational sac and cardiac activity) as dependent variable were performed to identify the significant predictors. All blastocysts were vitrified using Cryotop technology, with before transfer (post-warming) blastocyst morphology scores used in all analyses.

RESULT(S)

The 1795 blastocysts transferred were vitrifıed on embryo days 4 (1057), 5 (716), and 6 (22). The overall clinical implantation rate was 50.9%; however, using blastocyst age and blastocyst morphological score the clinical implantation rates increased from 49.0% (day-4 1 and 2) and 25.2% (day-5 1 and 2) to 71.2% (day-4 4AA) and 64.3% (day-5 4AA), respectively. Whereas full (≥3) blastocysts with scores of AA and BA had similar clinical implantation rates (66.2 vs. 66.7%), the rate of full blastocysts with scores of AB was lower (58.9%). In stepwise forward logistic regressions, female age, blastocyst age, blastocyst expansion score, blastocyst trophectoderm score, and number of blastocysts vitrified were significant predictors of clinical implantation.

CONCLUSION(S)

Using blastocyst age and before transfer blastocyst expansion and trophectoderm morphology scores to select blastocysts, clinical implantation rates greater than 70% could be achieved for top-scoring blastocysts.

Medroxyprogesterone acetate used in ovarian stimulation is associated with reduced mature oocyte retrieval and blastocyst development: a matched cohort study of 825 freeze-all IVF cycles

PURPOSE

To compare the effectivity of flexible-start medroxyprogesterone acetate (MPA) co-treatment ovarian stimulations (OS) with flexible-start gonadotropin-releasing hormone antagonist (GnRH-ant) co-treatment OS, in blastocyst freeze-all IVF cycles.

METHOD

This matched cohort study was performed at a single IVF center. Study cycles were extracted from freeze-all IVF cycles performed between February 2015 and June 2018 with cycles grouped according to the co-treatment protocol (MPA and GnRH-ant groups) used. MPA cycles were matched 1:1 using antral follicle count, female age, infertility duration, and female body mass index, with GnRH-ant cycles, resulting in 825 matched cycles. MPA or CET co-treatment was started when leading follicles reached 11-12 mm.

RESULTS

Duration of OS was significantly longer, and total FSH dose was significantly higher in the MPA group. Numbers of mature oocytes retrieved were similar; however, the mature oocyte retrieval rate (83.8 vs. 97.1%; p < 0.001), number of blastocysts, blastocyst rate (36.4 vs. 41.4%; p < 0.001) and > 2 viable blastocyst rate were all significantly lower in the MPA group. The live birth (LB) per transfer rates (51.6 vs. 55.7%; p = 0.155) were similar; however, the LB rate per treatment was significantly lower (40.9 vs. 45.8%; p = 0.05). A linear regression included the OS co-treatment protocol (GnRH-ant; 1.4 (1.07-1.81); p = 0.013) in the final model to predict having > 2 viable blastocysts.

CONCLUSION

Flexible-start MPA co-treatment OS was as effective in freeze-all IVF cycles as GnRH-ant co-treatment, with similar LB per transfer rates; however, increased cycle cancellation and reduced blastocyst numbers reduced LB per treatment rates significantly.

Large-for-gestational age is male-gender dependent in artificial frozen embryo transfers cycles: a cohort study of 1295 singleton live births

RESEARCH QUESTION

What is the effect of frozen embryo transfer (FET) on infant birth weight outcomes and which variables predic large-for-gestational age (LGA) infants.

DESIGN

In a large cohort study, the birth weight of 1295 singleton live births from blastocyst freeze-all-IVF treatments carried out between February 2015 and February 2017 at a single IVF centre were analysed. All embryo transfers were vitrified-warmed blastocyst transfers in artificial FET cycles, with patients having one (n = 864) or two (n = 431) blastocysts transferred. All live births were from ultrasound confirmed single fetal heart pregnancies.

RESULTS

The mean gestational age at delivery was 38.2 (±1.7) weeks, with a 1.11 : 1 female to male ratio for infants delivered. The small and large-for-gestational age rates were 5.02 and 13.28%, with 81.7% of infants appropriate for gestational age. In a multiple logistic regression analysis, the independent variables selected in the model to predict having an LGA infant were maternal parity, infant gender and maternal body mass index (BMI). The risk for LGA at term was significantly higher for male infants when adjusting for maternal parity and BMI (2.8 OR 1.805 to 4.450; P < 0.001).

CONCLUSION

The present study showed that fetal growth of artificial cycle FET pregnancies resulted in an 13.28% LGA infant rate that was mostly male gender dependent.

Prediction of live birth and cumulative live birth rates in freeze-all-IVF treatment of a general population

PURPOSE

To investigate the cumulative live birth (cLB) rate of one complete freeze-all-IVF cycle in a general infertile population and to investigate patient and treatment variables that predict blastocyst development and live birth (LB).

METHOD

In a retrospective observational study, the data of all IVF cycles performed between 1 February 2015 and 31 January 2016 at a single IVF centre was investigated. In the study, patient-couples were followed up for 18 months following oocyte retrieval. After exclusions, the patient and treatment variables of 1582 patient-couples who underwent treatment were included in the analyses.

RESULTS

The median time interval between the oocyte retrieval attempt and the frozen embryo transfer (FET) in which LB was achieved was 38.0 (35.0-67.0) days. The variables of freeze-all-IVF cycles with single blastocyst FET selected by multiple logistic regression to predict LB significantly were female age, infertility duration, FET number (i.e. 1st, 2nd, or ≥ 3rd FET), and blastocyst quality. In a regression adjusting for female age, the number of blastocysts transferred, and oocyte number group (1-3, 4-9, 10-15, and > 15), none of the oocyte number groups were selected to predict LB of 1st FET, significantly. While the per transfer LB rates decreased linearly from the 1st (56.5%) to the 3rd (36.4%) FET, the cLB rate increased from 47.3% after the 1st FET to 55.0% after a 3rd possible FET.

CONCLUSION

The cLB rate of one complete freeze-all-IVF cycle of a general infertile population, with 18-month follow-up, was 55.0%. In freeze-all-IVF, ovarian reserve variables were not selected by regression models to predict LB, significantly.

Placental volume and other first-trimester outcomes: are there differences between fresh embryo transfer, frozen-thawed embryo transfer and natural conception?

RESEARCH QUESTION

Does mode of conception influence placental volume and other first-trimester outcomes?

DESIGN

This retrospective single-centre case-control study led in Dijon University Hospital included 252 singleton pregnancies (84 IVF with either fresh embryo transfer or frozen-thawed embryo transfer [FET] and 168 natural conceptions). First-trimester placental volume, uterine artery pulsatility index and maternal serum PAPP-A and beta-HCG were measured. Statistical analyses were adjusted for gestational age, the newborn's gender, maternal age, parity, body mass index and smoking status.

RESULTS

Placental volume was significantly greater in the FET group than in the control group (P = 0.043) and fresh embryo transfer (P = 0.023) groups. At birth, fresh embryo transfer newborns were significantly smaller than controls (P = 0.01) and FET newborns (P = 0.008). Postpartum haemorrhage was far more frequent in FET than in controls and fresh embryo transfer group (38.1%, 2.6% and 1.9%, respectively; P < 0.0001). Placental volume positively correlated with PAPP-A, beta-HCG and the newborn's birth weight, and negatively correlated with uterine artery pulsatility index.

CONCLUSIONS

Placental volume and other first-trimester parameters are modified by IVF with fresh embryo transfer and FET compared with natural conception, but with opposite trends. Given the different protocols used for these techniques, hormonal treatment per se may have a major effect on pregnancy outcomes through the modification of placental invasiveness.