Accuracy of sonographic fetal gender determination: predictions made by sonographers during routine obstetric ultrasound scans

Artificial Intelligence to Determine Fetal Sex

OBJECTIVE

 This proof-of-concept study assessed how confidently an artificial intelligence (AI) model can determine the sex of a fetus from an ultrasound image.

STUDY DESIGN

 Analysis was performed using 19,212 ultrasound image slices from a high-volume fetal sex determination practice. This dataset was split into a training set (11,769) and test set (7,443). A computer vision model was trained using a transfer learning approach with EfficientNetB4 architecture as base. The performance of the computer vision model was evaluated on the hold out test set. Accuracy, Cohen's Kappa and Multiclass Receiver Operating Characteristic area under the curve (AUC) were used to evaluate the performance of the model.

RESULTS

 The AI model achieved an Accuracy of 88.27% on the holdout test set and a Cohen's Kappa score 0.843. The ROC AUC score for Male was calculated to be 0.896, for Female a score of 0.897, for Unable to Assess a score of 0.916, and for Text Added a score of 0.981 was achieved.

CONCLUSION

 This novel AI model proved to have a high rate of fetal sex capture that could be of significant use in areas where ultrasound expertise is not readily available.

KEY POINTS

· This is the first proof-of-concept AI model to determine fetal sex.. · This study adds to the growing research in ultrasound AI.. · Our findings demonstrate AI integration into obstetric care..

First trimester identification of fetal sex by ultrasound

PURPOSE

The hypothesis was fetal sex determination by ultrasound at 11-14 weeks' gestation has sufficient accuracy to be clinically relevant.

METHODS

Fetal sex assessment by transabdominal ultrasound was performed in 567 fetuses at 11-14 weeks' gestation (CRL: 45-84 mm). A mid-sagittal view of the genital region was obtained. The angle of the genital tubercle to a horizontal line through the lumbosacral skin surface was measured. The fetus was assigned male sex if the angle was > 30°, and female sex if the genital tubercle was parallel or convergent (< 10°). At an intermediate angle of 10-30°, the sex was not assigned. The results were divided into three categories based on gestational age: 11 + 2 to 12 + 1, 12 + 2 to 13 + 1, and 13 + 2 to 14 + 1 weeks' gestation. To establish its accuracy, the first trimester fetal sex determination was compared to fetal sex determined on a mid-second trimester ultrasound.

RESULTS

Sex assignment was successful in 534/683 (78%) of the cases. The overall accuracy of fetal sex assignment across all gestational ages studied was 94.4%. It was 88.3%, 94.7%, and 98.6% at 11 + 2 to 12 + 1, 12 + 2 to 13 + 1, and 13 + 2 to 14 + 1 weeks' gestation, respectively.

CONCLUSION

Prenatal sex assignment at the time of first trimester ultrasound screening has a high accuracy rate. The accuracy improved with increasing gestational age, which suggests that if clinically important decisions, such as chorionic villus sampling, are to be made based on fetal sex, they should be delayed until the latter part of the first trimester.

Should non-invasive prenatal testing (NIPT) be used for fetal sex determination? Perspectives and experiences of healthcare professionals

Non-invasive prenatal testing (NIPT) can not only accurately detect early in pregnancy the presence of chromosomal abnormalities but also fetal sex. However, whether fetal sex should be reported after performing NIPT is ethically contentious. In Belgium, NIPT is practically fully reimbursed and offered to all pregnant women as a first-tier screening. In practice, fetal sex is reported upon request of the expectant parents; however, this is not stipulated in guidelines. As more countries are offering NIPT and looking to implement it in public healthcare, challenges and insights of healthcare professionals working in Belgium can be of value for others. We assessed healthcare professionals' experiences with and perspectives on sex determination and reporting following NIPT in Belgium by conducting a semi-structured interview study. We interviewed 30 professionals involved in prenatal screening. While overall healthcare professionals did not consider reporting fetal sex to be an issue if the expectant parents want to know, some consider the reporting of a non-medical trait like fetal sex problematic when it is reimbursed or if it could lead to sex-selective termination of pregnancy. Moreover, the strong desire of expectant parents to know fetal sex risks compromising informed decision-making about NIPT. In this way, fetal sex may distract from the primary aim of NIPT as a test for medical conditions. Improving pre-test counseling both in terms of quality and availability may help overcome some of these issues.

Transgender and gender diverse individuals' perspectives on discussions of fetal sex chromosomes in obstetrics care

In the past decade, prenatal cell-free DNA screening (cfDNA) has become ubiquitous as a screening tool for fetal aneuploidy and sex chromosomes. Healthcare provider (HCP) discussions and public perceptions of sex and gender uniquely impact transgender and gender diverse (TGD) individuals, and existing cfDNA guidelines lack recommendations regarding how to discuss sex and gender prenatally. The aim of this exploratory qualitative study was to examine TGD individuals' opinions regarding fetal sex chromosome disclosure sessions. Twelve semi-structured virtual interviews were conducted with TGD individuals regarding their perspectives on the discussion of fetal sex chromosomes by HCPs within the prenatal setting. Interviews were coded and analyzed using a reflexive thematic approach, generating four major themes: (1) Current practices in prenatal care exclude gender diverse people; (2) HCPs' responsibility to de-gender discussions of sex chromosomes in prenatal care; (3) HCPs' responsibility to acknowledge gender diversity; and (4) HCPs' influence on societal perceptions of sex and gender. More guidance is needed from professional societies regarding best practices for HCP discussions of sex chromosomes, sex, and gender. Participants recommended HCPs educate patients about sex chromosomes and their relevance to health while avoiding the conflation of sex and gender terms. Additionally, there is an acute need for trans-inclusive prenatal healthcare. Ultimately, HCPs' and organizations are in a prime position to deconstruct rigid gender binaries and promote societal inclusion of TGD people.

Prenatal Findings in Postnatal Cases of Disorders of Sex Development: Experience from a Tertiary-Specialized Center in Brazil

INTRODUCTION AND OBJECTIVE

Prenatal suspicion of disorders/differences of sex development (DSDs) is a relatively new phenomenon. The aim of this study was to review the prenatal findings of DSD cases postnatally diagnosed in our tertiary referral center.

METHODS

We evaluated 57 DSD cases with sex ambiguity who had undergone prenatal ultrasound with phenotypic sex assessment and/or cell-free fetal DNA (cffDNA) for genotypic sex assessment.

RESULTS

Prenatal cffDNA had been performed in 32 cases, being positive (suggestive of male genotypic sex) in 26 and negative (suggestive of female genotypic sex) in 6. Five with cffDNA negative had a prenatal ultrasound indicating female external genitalia, in turn, in those with cffDNA positive, only two had a prenatal ultrasound indicating male external genitalia. Our postnatal data showed that when external genitalia were female or poorly virilized, prenatal ultrasound indicated female sex, but in cases of higher degree of virilization, ultrasound showed similar rates of male, female, or undetermined sex. Regarding the karyotype, our data showed those with XY karyotype had positive cffDNA, those with XX karyotype had negative cffDNA, and all five with sex chromosome anomalies had positive cffDNA because they were 45,X/46,XY. We suggested an algorithm to investigate these cases during gestation, including evaluation of uterus, fetal growth, and malformations.

CONCLUSION

We suggest that the parents should be counseled prenatally by a dedicated multidisciplinary team with experience in DSD management and evaluated as soon as possible after birth.

Fetal sex discordance

Fetal sex discordance is an entity that is becoming more frequent due to the expansion of the cfDNA for prenatal diagnosis. Its incidence can be estimated in 1/1500-2000 pregnancies, a frequency as high as that of some common chromosomopathies. The causes of this phenomenon are multiple and diverse, ranging from laboratory errors to important pathologies such as disorders of sexual differentiation. The management of a case of fetal sex discordance must be structured, starting with the review of the clinical history and the tests performed, and may require the performance of invasive tests to reach a diagnosis. Prevention through adequate pretest counseling and ultrasound confirmation can help to reduce its incidence.

Sex ratios at birth in Australia according to mother's country of birth: A national study of all 5 614 847 reported live births 1997-2016

Objectives

Son preference and sex selective practices have resulted in a deficit of girls in several countries, primarily across Asia. Emerging evidence indicates that son preference survives migration to Western high-income countries. The objective of this study was to assess male-to-female (M/F) ratios at birth per mother’s country of birth in Australia 1997–2016, in total and by parity, and by states/territories and over time.

Methods

Data for this national population-based cross-sectional study were obtained from the National Perinatal Data Collection (NPDC) and included all live births in Australia 1997–2016 (N = 5 614 847). M/F ratios with 95% Confidence Intervals were estimated.

Results

The M/F ratio for births to Australian-born mothers was within the expected range (1.03–1.07) regardless of parity and time period. M/F ratios were elevated above the expected range for births to mothers born in China in the total sample (M/F ratio 1.084, 95% confidence interval 1.071–1.097) and at parity 2 (1.175, 1.120–1.231), and for births to mothers born in India at parity 2 (1.146, 1.090–1.204). Parity 2 births were the most consistently male-biased across time. Across states, elevated M/F ratios were identified for both groups in New South Wales (China parity 2: 1.182, 1.108–1.260; India parity 2: 1.182, 1.088–1.285), for births to Chinese-born mothers in Victoria (total births: 1.097, 1.072–1.123; parity 1: 1.115, 1.072–1.159) and Australian Capital Territory (total births: 1.189, 1.085–1.302) and births to Indian-born mothers Western Australia (parity 2: 1.307, 1.122–1.523).

Conclusions

Son preference persists in some immigrant communities after migration to Australia. The consistent pattern of elevated M/F ratios across the larger states indicates that sex imbalances at birth are largely independent of restrictiveness of local abortion laws. Drivers and consequences of son preference in Western high-income settings should be explored to further promote gender equality, and to strengthen support for women who may be vulnerable to reproductive coercion.

[Genital abnormalities: Contextualization of a neglected area in prenatal diagnosis]

OBJECTIVE

To reflect on how the area of genital abnormalities has fallen behind in prenatal diagnosis.

METHODS

Based on the thesis that prenatal diagnosis of genital abnormalities has scarcely developed, a comparison with other areas of prenatal diagnosis and with its postnatal counterpart is presented; different explanations for this situation are examined; and a reflection is presented on ways to expand this field of knowledge.

CONCLUSIONS

Compared to other disciplines, prenatal diagnosis of genital abnormalities finds itself lagging behind in terms of diagnostic tools, management protocols and scientific literature. Potential causes include a perception of low prevalence and limited importance, or exploration challenges. Integration of current knowledge, together with the acquisition of the appropriate tools and translation to clinical medicine, would be a way to make this discipline stronger.

Discordant fetal sex on NIPT and ultrasound

Prenatal diagnosis of sex discordance is a relatively new phenomenon. Prior to cell-free DNA testing, the diagnosis of a disorder of sexual differentiation was serendipitous, either through identification of ambiguous genitalia at the midtrimester morphology ultrasound or discovery of genotype-phenotype discordance in cases where preimplantation genetic diagnosis or invasive prenatal testing had occurred. The widespread integration of cfDNA testing into modern antenatal screening has made sex chromosome assessment possible from 10 weeks of gestation, and discordant fetal sex is now more commonly diagnosed prenatally, with a prevalence of approximately 1 in 1500-2000 pregnancies. Early detection of phenotype-genotype sex discordance is important as it may indicate an underlying genetic, chromosomal or biochemical condition and it also allows for time-critical postnatal treatment. The aim of this article is to review cfDNA and ultrasound diagnosis of fetal sex, identify possible causes of phenotype-genotype discordance and provide a systematic approach for clinicians when counseling and managing couples in this circumstance.

Sex selection and non-invasive prenatal testing: A review of current practices, evidence, and ethical issues

Non-invasive prenatal testing (NIPT) can determine the sex of the fetus very accurately and very early in gestation. There are concerns that the ease, timing, and accuracy of NIPT sex determination will facilitate sex-selective termination of pregnancy (TOP). Here, we review current practices, the evidence for a link between NIPT and sex-selective TOP, and associated ethical issues. Sex-selective TOP, usually motivated by son preference, has had serious demographic consequences in countries such as India and China. Currently, ultrasound is the primary method by which parents determine the sex of the fetus. The diffusion of ultrasound technology has had a direct impact on the rates of sex-selective TOP. Although NIPT is currently more costly, it is feasible that increased uptake of this technology could have a similar effect. Partly because NIPT is a relatively recent development in prenatal screening, there is little data on the impact of NIPT on sex selection practices. Evidence that NIPT is playing a role in sex-selective TOP remains largely anecdotal. Further research is required to assess and quantify TOP resulting from NIPT sex determination. The use of these technologies for sex selection raises a number of ethical issues, in addition to practical demographic consequences.

Prenatal imaging of genital defects: clinical spectrum and predictive factors for severe forms

OBJECTIVES

To report the clinical spectrum of genital defects diagnosed before birth, identify predictive factors for severe phenotypes at birth, and determine the rate of associated malformations.

PATIENTS AND METHODS

A retrospective study (2008-2017) of 4580 fetuses, identified prenatally with abnormalities evaluated by our Reference Center for Fetal Medicine, included cases with fetal sonographic findings of abnormal genitalia or uncertainty of fetal sex determination. Familial, prenatal and postnatal data were collected via a standardised questionnaire.

RESULTS

In all, 61 fetuses were included. The positive predictive value (PPV) of the prenatal diagnosis of genital defects was 90.1%. Most cases were 46,XY-undervirilized boys, 42 cases (68.8%), which included 29 with mid-penile or posterior hypospadias, nine with anterior hypospadias, and epispadias, micropenis, scrotal transposition, and buried penis (one each). In all, 46,XX-virilized girls were identified in seven cases (11.5%), which included four with congenital adrenal hyperplasia, two with isolated clitoromegaly, and one with ovotestis. Other defects included prune belly syndrome and persistent cloaca (six cases). Early detection during the second trimester (58.1% vs 18.8%, P = 0.03), intra-uterine growth restriction (IUGR) (45.2% vs 9.1%, P = 0.06), and curvature of the penis (38.7% vs 0%, P = 0.02), were more frequently related to severe defects in male newborns. Associated malformations (14 cases, 22.9%) and genetic defects (six) were frequent in undervirilized boys.

CONCLUSION

Prenatal imaging of genital defects leads to a wide range of phenotypes at birth. Its PPV is high and extra-urinary malformations are frequent. Early diagnosis during the second trimester, associated IUGR, and curvature of the genital tubercle, should raise suspicion of a severe phenotype and may justify delivery near a multidisciplinary disorders/differences of sex development team.

Maternal Intuition of Fetal Gender

Purpose

Fetal gender speculation is a preoccupation of many expecting parents, and pregnant women commonly profess to intuitively know the gender of their unborn babies. This study objectively compared pregnant mothers' perceptions of fetal gender to sonographically proven gender determinations. Also, success rates from previously published studies, noninvasive prenatal testing and a myriad of gender determination methods were observed and reported for context.

Methods

All pregnant women presenting for second-trimester screening ultrasound (at 17-23 weeks gestation) in the obstetrics department of a single health center were asked to participate. A medical sonographer described the ultrasound examination, obtained appropriate consent and medical history. Each mother was asked if she had any perception as to the fetal gender and her answer documented. Mothers who had foreknowledge of fetal gender were excluded. Frequencies of actual gender were compared with observed frequencies of the maternal prediction using chi-squared test.

Results

Approximately 40% (n=411) of our study population (N=1,026) indicated having an intuition or perception of fetal gender. These women correctly predicted the gender of their babies 51% of the time (P=0.6571). Women who expressed a "strong" degree of intuition (n=53) fared better, accurately predicting fetal gender at a rate of 62%, though the difference in this smaller subcohort also failed to demonstrate statistical significance (P=0.0741).

Conclusions

Intuition of fetal gender is professed by almost half of mothers though, when present, is no better at accurately predicting fetal gender than flipping a coin.

Accuracy of fetal sex determination on ultrasound examination in the first trimester of pregnancy

PURPOSE

The aim of this study was to evaluate the feasibility and success rate of sex determination on transabdominal sonographic examination at 11-13 weeks' gestation and to identify factors influencing accuracy.

METHODS

In this prospective observational evaluation of 672 fetuses between 11 weeks' and 13 weeks + 6 days' gestational age (GA), we determined fetal sex according to the angle of the genital tubercle viewed on the midsagittal plane. We also analyzed maternal, fetal, and operator factors possibly influencing the accuracy of the determination.

RESULTS

Fetal sex determination was feasible in 608 of the 672 fetuses (90.5%), and the prediction was correct in 532 of those 608 cases (87.5%). Fetal sex was more accurately predicted as the fetal crown-rump length (CRL), and GA increased and was less accurately predicted as the maternal body mass index increased. A CRL greater than 55.7 mm, a GA more than 12 weeks + 2 days, and a body mass index below 23.8 were identified as the best cutoff values for sex prediction. None of the other analyzed factors influenced the feasibility or accuracy of sex determination.

CONCLUSIONS

The sex of a fetus can be accurately determined on sonographic examination in the first trimester of pregnancy; the accuracy of this prediction is influenced by the fetal CRL and GA and by the maternal body mass index. © 2015 Wiley Periodicals, Inc. J Clin Ultrasound 44:272-277, 2016.