Artificial Intelligence in the Assessment of Female Reproductive Function Using Ultrasound: A Review

The incidence of infertility is continuously increasing nearly all over the world in recent years, and novel methods for accurate assessment are of great need. Artificial Intelligence (AI) has gradually become an effective supplementary method for the assessment of female reproductive function. It has been used in clinical follicular monitoring, optimum timing for transplantation, and prediction of pregnancy outcome. Some literatures summarize the use of AI in this field, but few of them focus on the assessment of female reproductive function by AI-aided ultrasound. In this review, we mainly discussed the applicability, feasibility, and value of clinical application of AI in ultrasound to monitor follicles, assess endometrial receptivity, and predict the pregnancy outcome of in vitro fertilization and embryo transfer (IVF-ET). The limitations, challenges, and future trends of ultrasound combined with AI in providing efficient and individualized evaluation of female reproductive function had also been mentioned.

Elevated levels of nitrous dioxide are associated with lower AMH levels: a real-world analysis

STUDY QUESTION

Are there any associations between environmental pollutants and ovarian reserve, expressed by anti-Mullerian hormone (AMH) serum levels?

SUMMARY ANSWER

In this first real-world approach to demonstrate the relationship between air pollutants and serum AMH levels, adverse associations were observed for nitrogen dioxide (NO2) but not with particulate matter.

WHAT IS KNOWN ALREADY

In recent years, air pollution has emerged as a potential disrupter to the homeostasis of physiological hormones, possibly affecting human reproduction. Although the influence of age and smoking on AMH levels is largely accepted, the relationship between AMH and the environment has not currently been established.

STUDY DESIGN, SIZE, DURATION

A longitudinal, observational, retrospective, real-world study was carried out, including all AMH measurements performed in a single laboratory from January 2007 to October 2017.

PARTICIPANTS/MATERIALS, SETTING, METHODS

Serum AMH data were connected to patients' age and residential address, to include air pollution data after geo-localisation. The air pollution considered daily particulate matter (PM) and NO2 values.

MAIN RESULTS AND THE ROLE OF CHANCE

A total of 1463 AMH measurements were collected (mean 1.94 ng/ml, median 0.90 ng/ml). AMH was inversely related to patients' age in women older than 25 years (adjusted R-squared 0.120, P < 0.001), but not in those younger than 25 years (adjusted R-squared 0.068, P = 0.055). AMH levels were inversely related to environmental pollutants, such as PM10 (Rho = -0.088, P = 0.001), PM2.5 (Rho = -0.062, P = 0.021) and NO2 (Rho = -0.111, P < 0.001). After subdividing the dataset into quartiles for PM10 and PM2.5, the influence of age on AMH serum levels was found to be a stronger influence than that exerted by PM (P = 0.833 and P = 0.370, respectively). On the contrary, considering NO2 quartiles, higher AMH levels were observed in third quartile compared to fourth quartile, even after adjustment for age (P = 0.028), indicating a stronger influence of NO2 exposure on AMH serum levels. Considering an AMH cut-off of 0.3 ng/ml, a significant higher frequency of women with severe ovarian reserve reduction in the fourth quartile was shown only for NO2 (P = 0.010).

LIMITATIONS, REASONS FOR CAUTION

Several limitations should be underlined, such as the lack of information about work and life habits of each patient and the retrospective nature of the analysis performed on real-world data.

WIDER IMPLICATIONS OF THE FINDINGS

Although the genetic component is highly predictive for defining the ovarian reserve at birth, potentially modifiable environmental factors could influence the rate of decline in AMH and ovarian reserve during adulthood.

STUDY FUNDINGCOMPETING INTEREST(S)

Authors have neither funding nor competing interests to declare.

TRIAL REGISTRATION NUMBER

N/A.

Preconception tests at advanced maternal age

Pregnancies at an advanced reproductive age are increasingly common. However, the safety of pregnancy remains a concern as maternal age is a recognized independent factor for various obstetric complications. Also, age is a risk factor for most systematic health problems and older women are more likely to enter into pregnancy with pre-existing conditions. At the moment there is no separate, structured guidance on preconception tests at advanced maternal age. However, the preconceptual period offers an ideal window to recognize and address underlying health issues, social issues and harmful lifestyle behaviours in order to optimize maternal health ultimately reducing infertility, perinatal morbidity and mortality. Preconception tests should be clinically relevant aiming to identify risk factors and address them to predict and prevent infertility and pregnancy complications. The importance of preconception care is magnified for women of advanced age for whom the risks are higher and the potential benefits greater.

Growth hormone during in vitro fertilization in older women modulates the density of receptors in granulosa cells, with improved pregnancy outcomes

OBJECTIVE

To study the effect of aging and granulosa cell growth hormone receptor (GHR) expression, and the effect of growth hormone (GH) co-treatment during IVF on receptor expression.

DESIGN

Laboratory study.

SETTING

University.

PATIENT(S)

A total of 445 follicles were collected from 62 women undergoing standard infertility treatment.

INTERVENTION(S)

Preovulatory ovarian follicle biopsies of granulosa cells and follicular fluid.

MAIN OUTCOME MEASURE(S)

Older women with a poor ovarian reserve were co-treated with GH to determine the effect of the adjuvant during IVF on the granulosal expression density of FSH receptor (FSHR), LH receptor (LHR), bone morphogenetic hormone receptor (BMPR1B), and GHR. Ovarian reserve, granulosa cell receptor density, oocyte quality, and pregnancy and live birth rates were determined.

RESULT(S)

Growth hormone co-treatment increased the receptor density for granulosal FSHR, BMPR1B, LHR, and GHR compared with the non-GH-treated patients of the same age and ovarian reserve. Growth hormone co-treatment increased GHR density, which may increase GHR activity. The GH co-treatment was associated with a significant increase in pregnancy rate.

CONCLUSION(S)

Growth hormone co-treatment restored the preovulatory down-regulation of FSHR, BMPR1B, and LHR density of the largest follicles, which may improve the maturation process of luteinization in older patients with reduced ovarian reserve. The fertility of the GH-treated patients improved.

The transcriptome of human oocytes is related to age and ovarian reserve

STUDY QUESTION

How does the human oocyte transcriptome change with age and ovarian reserve?

SUMMARY ANSWER

Specific sets of human oocyte messenger RNAs (mRNAs) and non-coding RNAs (ncRNAs) are affected independently by age and ovarian reserve.

WHAT IS KNOWN ALREADY

Although it is well established that the ovarian reserve diminishes with increasing age, and that a woman's age is correlated with lower oocyte quality, the interplay of a diminished reserve and age on oocyte developmental competence is not clear. After maturation, oocytes are mostly transcriptionally quiescent, and developmental competence prior to embryonic genome activationrelies on maternal RNA and proteins.

STUDY DESIGN, SIZE, DURATION

A total of 36 vitrified/warmed MII oocytes from 30 women undergoing oocyte donation were included in this study, processed and analyzed individually.

PARTICIPANTS/MATERIALS, SETTING, METHODS

Total RNA from each oocyte was independently isolated, amplified, labeled, and hybridized on HTA 2.0 arrays (Affymetrix). Data were analyzed using TAC software, in four groups, each including nine oocytes, according to the woman's age and antral follicular count (AFC) (mean ± SD): Young with High AFC (YH; age 21 ± 1 years and 24 ± 3 follicles); Old with High AFC (OH; age 32 ± 2 years and 29 ± 7 follicles); Young with Low AFC (YL; age 24 ± 2 years and 8 ± 2 follicles); Old with Low AFC (OL; age 34 ± 1 years and 7 ± 1 follicles). qPCR was performed to validate arrays.

MAIN RESULTS AND THE ROLE OF CHANCE

We identified a set of 30 differentially expressed mRNAs when comparing oocytes from women with different ages and AFC. In addition, 168 non-coding RNAs (ncRNAs) were differentially expressed in relation to age and/or AFC. Few mRNAs have been identified as differentially expressed transcripts, and among ncRNAs, a set of Piwi-interacting RNAs clusters (piRNAs-c) and precursor microRNAs (pre-miRNAs) were identified as increased in high AFC and old groups, respectively. Our results indicate that age and ovarian reserve are associated with specific ncRNA profiles, suggesting that oocyte quality might be mediated by ncRNA pathways.

LARGE SCALE DATA

Data can be found via GEO accession number GSE87201.

LIMITATIONS, REASONS FOR CAUTION

The oldest woman included in the study was 35 years old, thus our results cannot readily be extrapolated to women older than 35 or infertile women.

WIDER IMPLICATIONS OF THE FINDINGS

We show, for the first time, that several non-coding RNAs, usually regulating DNA transcription, are differentially expressed in relation to age and/or ovarian reserve. Interestingly, the mRNA transcriptome of in vivo matured oocytes remains remarkably stable across ages and ovarian reserve, suggesting the possibility that changes in the non-coding transcriptome might regulate some post-transcriptional/translational mechanisms which might, in turn, affect oocyte developmental competence.

STUDY FUNDING AND COMPETING INTEREST(S)

This work was supported by intramural funding of Clinica EUGIN and by the Secretary for Universities and Research of the Ministry of Economy and Knowledge of the Government of Catalonia. J.H. and A.S. are employees of Affymetrix, otherwise there are no competing interests.

Characterization of a suboptimal IVF population and clinical outcome after two IVF cycles

The number of oocytes retrieved in in vitro fertilization (IVF) cycles is an independent factor influencing pregnancy rate (PR), and optimal number of oocytes would be between 10 and 15. This has led to the hypothesis that the identification of a suboptimal group of responders beforehand (4-9 oocytes retrieved) would allow physicians to optimize their PR. A retrospective observational study counting on 735 women doing an IVF treatment in our center was performed. Multivariable logistic regression was used to analyze the relationship between anti-Mullerian hormone (AMH) and antral follicle count (AFC), within suboptimal and optimal responders. We also analyzed the outcome of those patients with an estimated high probability of having an optimal response and the second cycles of those who did not get pregnant in the first cycle to observe the main significant traits that made them change from one group of responders to the other. Main results are that suboptimal responders account for almost half of our patients. Ovarian reserve markers (AMH and AFC) are significantly different in optimal and suboptimal responders, even when adjusted by age. There is a significant difference in the cumulative PR between both groups. Interestingly, 18.9% shifted from suboptimal to optimal response, and 36.9% from optimal to suboptimal.

Counting ovarian antral follicles by ultrasound: a practical guide

This Consensus Opinion summarizes the main aspects of several techniques for performing ovarian antral follicle count (AFC), proposes a standardized report and provides recommendations for future research. AFC should be performed using a transvaginal ultrasound (US) probe with frequency ≥ 7 MHz. For training, we suggest a minimum of 20-40 supervised examinations. The operator should be able to adjust the machine settings in order to achieve the best contrast between follicular fluid and ovarian stroma. AFC may be evaluated using real-time two-dimensional (2D) US, stored 2D-US cine-loops and stored three-dimensional (3D) US datasets. Real-time 2D-US has the advantage of permitting additional maneuvers to determine whether an anechoic structure is a follicle, but may require a longer scanning time, particularly when there is a large number of follicles, resulting in more discomfort to the patient. 2D-US cine-loops have the advantages of reduced scanning time and the possibility for other observers to perform the count. The 3D-US technique requires US machines with 3D capability and the operators to receive additional training for acquisition/analysis, but has the same advantages as cine-loop and also allows application of different imaging techniques, such as volume contrast imaging, inversion mode and semi-automated techniques such as sonography-based automated volume calculation. In this Consensus Opinion, we make certain recommendations based on the available evidence. However, there is no strong evidence that any one method is better than another; the operator should choose the best method for counting ovarian follicles based on availability of resources and on their own preference and skill. More studies evaluating how to improve the reliability of AFC should be encouraged. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.

Does ovarian reserve affect outcomes in single ideal blastocyst transfers in women less than 40 years of age?

PURPOSE

There is much debate whether diminished ovarian reserve is purely a quantitative issue, or if quality as determined by pregnancy potential of the ensuing oocytes is also affected. The purpose of this study was to determine whether diminished ovarian reserve, as established by one of three ways described below, affects pregnancy outcomes of women under 40 years old undergoing a single ideal blastocyst transfer.

MATERIALS AND METHODS

This was a retrospective cohort study, including 507 women undergoing an ideal quality single embryo transfer between August 2010 and March 2014. Logistic regression was used to control for age, duration of infertility, parity, body mass index, and smoking status. For analysis, women were stratified for: antral follicle counts (≤ 5 vs. > 5), basal serum FSH levels (< 13 vs. ≥ 13 IU/L), and quartile of total FSH dose required for stimulation.

RESULTS

In stratifying women by antral follicle count (AFC) ≤ 5 vs. > 5, the pregnancy rate (40 vs. 53%, p = 0.04), clinical pregnancy rate (29 vs. 46%, p = 0.02), and live birth rate (13 vs. 43%, p = 0.001) were superior with AFC > 5. Using FSH levels (< 13 vs. ≥ 13 IU/L), the pregnancy rate (50 vs. 31%, p = 0.27), clinical pregnancy rate (40 vs. 13%, p = 0.45), and live birth rate (38 vs. 13%, p = 0.48) were similar. Examining quartiles of FSH stimulation, the pregnancy rates were similar (from lowest to highest: 45, 52, 54, 41%, p = 0.13); however, clinical pregnancy rate (36, 43, 47, 25%, p = 0.003) and live birth rate (32, 38, 44, 20%, p = 0.005) were superior in lower quartiles vs. the highest quartile.

CONCLUSION

Ovarian reserve may affect embryo pregnancy potential and outcomes when measured by AFC and exogenous stimulation but not by basal FSH levels.

Low oocyte yield during IVF treatment and the risk of a trisomic pregnancy

A low number of antral follicles may result in the selection of suboptimal oocytes that are prone to meiotic errors. The aim of this case-control study was to evaluate women receiving IVF treatment with low oocyte yield (defined as three or fewer oocytes retrieved after ovarian stimulation) who are at an increased risk of a trisomic pregnancy. Data were obtained from Danish and Dutch medical registries between 1983 and 2011. Analyses were carried out in 105 cases and 442 controls matched by age and year of IVF treatment. Cases were women with a trisomic pregnancy (trisomies 13, 18 or 21) resulting from fresh IVF treatment and confirmed by karyotyping. Cases were included regardless of pregnancy outcome. Controls were women with a live born child without a trisomy, resulting from fresh IVF treatment. Low oocyte yield was observed in 6.6% (29/440) of the women, of which 8.4% (7/83) were cases and 6.2% (22/357) controls. Low oocyte yield in IVF treatment was not associated with a higher risk of trisomic pregnancy (OR 1.43, 95% CI 0.64 to 3.19). Stratification for female age, adjustment for history of ovarian surgery, and gonadotrophin-releasing hormone protocol used did not change the results.

A putative role for anti-Müllerian hormone (AMH) in optimising ovarian reserve expenditure

The mammalian ovary has a finite supply of oocytes, which are contained within primordial follicles where they are arrested in a dormant state. The number of primordial follicles in the ovary at puberty is highly variable between females of the same species. Females that enter puberty with a small ovarian reserve are at risk of a shorter reproductive lifespan, as their ovarian reserve is expected to be depleted faster. One of the roles of anti-Müllerian hormone (AMH) is to inhibit primordial follicle activation, which slows the rate at which the ovarian reserve is depleted. A simple interpretation is that the function of AMH is to conserve ovarian reserve. However, the females with the lowest ovarian reserve and the greatest risk of early reserve depletion have the lowest levels of AMH. In contrast, AMH apparently strongly inhibits primordial follicle activation in females with ample ovarian reserve, for reasons that remain unexplained. The rate of primordial follicle activation determines the size of the developing follicle pool, which in turn, determines how many oocytes are available to be selected for ovulation. This review discusses the evidence that AMH regulates the size of the developing follicle pool by altering the rate of primordial follicle activation in a context-dependent manner. The expression patterns of AMH across life are also consistent with changing requirements for primordial follicle activation in the ageing ovary. A potential role of AMH in the fertility of ageing females is proposed herein.

Follicle pool, ovarian surgery and the risk for a subsequent trisomic pregnancy

STUDY QUESTION

Is there an association between trisomic pregnancy, a marker for decreased oocyte quality, and the reduced oocyte quantity that follows ovarian surgery?

SUMMARY ANSWER

Previous ovarian surgery is not associated with an increased risk for a subsequent trisomic pregnancy.

WHAT IS KNOWN ALREADY

Ovarian surgery diminishes the number of oocytes. The risk for a trisomic pregnancy is suggested to be higher in women with fewer oocytes, independent of their chronological age.

STUDY DESIGN, SIZE, DURATION

This is a matched case-control study. Cases are women with a confirmed trisomic pregnancy occurring between 1 January 2000 and 31 December 2010 regardless of pregnancy outcome and controls are women that had a live born child without a trisomy. In total, there were 8573 participants in the study; 1723 cases and 6850 controls.

PARTICIPANTS/MATERIALS, SETTING, METHODS

Data were obtained from Danish medical registries. Matching criteria were maternal age and year of conception. Number of controls matched per case ranged from one to four. Among cases and controls with a trisomic pregnancy, 2.7% (46/1723) versus 2.5% (172/6850) had undergone ovarian surgery before pregnancy.

MAIN RESULTS AND ROLE OF CHANCE

History of ovarian surgery is not associated with a higher risk for a subsequent trisomic pregnancy (odds ratio = 1.00, 95% confidence interval 0.99-1.01). Subgroup analyses by indication of surgery and interval between ovarian surgery and pregnancy do not show an effect on trisomic pregnancy risk.

LIMITATIONS, REASONS FOR CAUTION

The medical registries used to select cases and controls did not contain information on surgical technique nor volume of ovarian tissue resected, previous trisomic pregnancy prior to the ovarian surgery or long-term use of oral contraceptives. Therefore, correction for these factors was not performed.

WIDER IMPLICATIONS OF THE FINDINGS

We did not confirm the hypothesis that ovarian surgery, a marker for decreased oocyte quantity, is related to trisomic pregnancy, a marker for decreased oocyte quality. This suggests that ovarian surgery, which has a direct reductive effect on the size of the follicle pool, may affect oocyte quality differently when compared with the reduction in follicle pool size due to ageing.

STUDY FUNDING/COMPETING INTERESTS

The study was supported by grants from the Gratama Stichting, University of Groningen and the University Medical Center Groningen, The Netherlands. Ø.L. has within the last 3 years received honoraria for speeches in pharmacoepidemiological issues, not related to this study. The Department of Obstetrics and Gynaecology receives unrestricted educational grants from Ferring Pharmaceuticals. A.H. received a grant from ZonMW (i.e. National Dutch Scientific funding) for a RCT not related to this publication. Dr A.H. received speakers fee from MSD for an educational presentation. All other authors have no conflict of interest.

[Epidemiology of loss pregnancy]

OBJECTIVES

Study of epidemiology of pregnancy loss.

MATERIALS AND METHOD

A systematic review of the literature was performed using Pubmed and the Cochrane library databases and the guidelines from main international societies.

RESULTS

The occurrence of first trimester miscarriage is 12% of pregnancies and 25% of women. Miscarriage risk factors are ages of woman and man, body mass index greater than or equal to 25kg/m(2), excessive coffee drinking, smoking and alcohol consumption, exposure to magnetic fields and ionizing radiation, history of abortion, some fertility disorders and impaired ovarian reserve. Late miscarriage (LM) complicates less than 1% of pregnancies. Identified risk factors are maternal age, low level of education, living alone, history of previous miscarriage, of premature delivery and of previous termination of pregnancy, any uterine malformation, trachelectomy, existing bacterial vaginosis, amniocentesis, a shortened cervix and a dilated cervical os with prolapsed membranes. Fetal death in utero has a prevalence of 2% in the world and 5/1000 in France. Its main risk factors are detailed in the chapter.