Machine Learning-Guided Noninvasive Embryo Selection for Clinical in Vitro Fertilization Treatment to Avoid Wasting Potentially Qualified Embryos

doi:10.21203/rs.3.rs-617438/v1

Download PDF

Research Article

Machine Learning-Guided Noninvasive Embryo Selection for Clinical in Vitro Fertilization Treatment to Avoid Wasting Potentially Qualified Embryos

https://doi.org/10.21203/rs.3.rs-617438/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

The success rates of in vitro fertilization (IVF) treatment are limited by the aneuploidy of human embryos. Pre-implantation genetic testing for aneuploidy(PGT-A) is often used to select embryos with normal ploidy but requires invasive embryo biopsy.

Methods

We performed chromosome sequencing of 345 paired blastocyst culture medium and whole blastocyst samples and developed a noninvasive embryo grading system based on the random forest machine-learning algorithm to predict blastocyst ploidy. The system was validated in 266 patients, and a blinded prospective observational study was performed to investigate clinical outcomes between machine learning-guided and traditional niPGT-A analyses. We graded embryos as A, B, or C using machine learning-guided niPGT-A analysis according to their euploidy probability levels predicted by noninvasive chromosomal screening.

Results

We observed higher live birth rate in A- versus C-grade embryos (50.4% versus 27.1%, p=0.006) and B- versus C-grade embryos (45.3% versus 27.1%, p=0.022) and lower miscarriage rate in A- versus C-grade embryos (15.9% versus 33.3%, p=0.026) and B- versus C-grade embryos (14.3% versus 33.3%, p=0.021). The embryo utilization rate was significantly higher through machine learning strategy compared to the conventional dichotomic judgment of euploidy or aneuploidy in the niPGT-A analysis (78.8% versus 57.9%, p<0.001). We observed better outcomes in A- and B-grade embryos versus C-grade embryos and higher embryo utilization rates through machine learning strategies than traditional niPGT-A analysis.

Conclusion

These results demonstrate that the machine learning-guided embryo grading system can optimize embryo selection and avoid wasting potential embryos.

Trial registration

Chinese Clinical Trial Registry,ChiCTR-RRC-17010396.Registered 11 January 2017, http://www.chictr.org.cn/ChiCTR-RRC-17010396

Translational Medicine

embryo selection

noninvasive pre-implantation genetic testing

machine learning

in vitro fertilization

Clinical studies have shown that in vitro fertilization (IVF) outcomes with elective single embryo transfer (eSET) have improved following the development of pre-implantation genetic testing for aneuploidy (PGT-A) [1, 2]. However, this biopsy procedure is detrimental to IVF success rates in some couples with a poor prognosis for IVF[3-5]. Xu et al. [6] recently performed noninvasive chromosome screening (NICS), namely a noninvasive version of PGT-A (niPGT-A), in clinical settings. Moreover, Fang et al. [7] obtained an ongoing pregnancy rate of 58% and reported 27 normal live births in a pilot clinical study by employing NICS with embryo culture medium.

In addition to the omission of embryo biopsy to minimize mechanical harm to the embryo, niPGT-A also comprehensively screens chromosomes and is called embryo biopsy-based PGT-A or PGS2.0. Comprehensive chromosomal screening (CCS) increases the efficacy of PGS2.0 for aneuploidy identification compared to PGS1.0; however, the high-throughput assay of PGS2.0 is sensitive to cause false-positive test results, such as non-specific chromosomal gain or loss due to cells being in S-phase, or apoptotic or dead cells generating profiles resembling mosaicism [8]. This raises the concern that a number of embryos with potential are being discarded for normal euploid pregnancies [9].

Studies evaluating the accuracy of niPGT-A found a fluctuating ploidy concordance rate of 33 %–100 % [10-18]. It is reasonable to speculate that the differences in the results of different clinics can lead to differences in downstream results. Additionally, aneuploid cells are more likely to undergo apoptosis for clearance from the embryos than euploid cells, leading to a high mosaic aneuploidy rate during the chromosomal screening of blastocyst culture medium samples [19]. Huang et al. [17] defined mosaicism of above 60 % as the threshold to distinguish an aneuploid embryo from a euploid embryo, and given that most aneuploid embryos are not expected to be transferred, this may result in a huge waste of potentially qualified embryos and may increase the cancellation rate of IVF cycles.

Consequently, there is a necessity to establish a grading system to prioritize embryos for transfer; in particular, strategies capable of distinguishing actual mosaics and aneuploid embryos from false-positive ones resulting from apoptotic cells are urgently required. The objective of this study was to establish a noninvasive embryo transfer strategy to provide a reference for embryo selection. To this end, we first collected 345 blastocyst embryos and performed CCS in culture media, and selected whole blastocysts from the same embryos. The machine learning technique was used innovatively for the first time, and euploid and aneuploid embryo-specific copy number variation (CNV) patterns in the culture media were deeply mined. With this technique, different chromosomal profiles from culture media could be assigned different aneuploidy probability predictions based on the golden standard of whole blastocyst embryos. Using the predicted euploidy probability, we proposed a prioritized embryo transfer strategy and further designed a prospective and blinded observational clinical transfer experiment with 266 patients to validate this strategy. To the best of our knowledge, this is the first large-scale validation of CCS in culture medium, with both embryonic validation and blinded prospective clinical confirmation using ploidy information in the culture medium as a predictor for selection of embryos for SET in clinical IVF.

Study design

First, an embryo selection strategy pattern based on the euploid probability of the culture medium was established using 345 pairs of culture medium and whole embryo samples. In this strategy, embryos were divided into three grades, A, B, or C, according to euploidy probability using whole embryos as the gold standard. After that, we recruited 266 patients who underwent 278 frozen single blastocyst transfers and validated the NICS artificial intelligence (AI) model and niPGT-A analysis using a blinded prospective approach. The primary endpoint was birth rate, and the secondary endpoints were miscarriage, clinical pregnancy, and ongoing pregnancy rates.

Embryo and culture medium collection

In total, 345 embryos were evaluated at the Reproductive Medicine Center of Nanjing Jinling Hospital (Nanjing, Jiangsu, People’s Republic of China) and Nanjing Drum Tower Hospital (Nanjing City, Jiangsu Province, People’s Republic of China). The ages of the female patients ranged from 20 to 45 years. Institutional Review Board approval (reference number 2016NJKY-028, 2017-09-02) and informed consent from patients were obtained prior to embryo analysis. All embryos were derived from intracytoplasmic sperm injection (ICSI) and cultured to the blastocyst stage (day 5/6). Embryos that were not eligible for transfer according to morphological scoring criteria [20] were donated and used for subsequent analysis. Blastocyst medium (20–25 μL) from each embryo was transferred into RNase–DNase-free PCR tubes, each containing 5 μL of the cell lysis buffer (Yikon Genomics, Suzhou, People’s Republic of China) and stored at −80 °C until use. Matched whole embryo cells were also transferred into 5 μL of the lysis buffer. Cell lysis was performed with 10 μL each of culture medium and whole embryos, according to the manufacturer’s instructions (Yikon Genomics).

Patient selection

All patients who underwent ICSI with NICS and had a single frozen-thawed embryo transfer (FET) between July 2017 and December 2018 at the Reproductive Medicine Department of Nanjing Jinling Hospital were included in the study. The inclusion criteria were as follows: maternal age of less than 45 years and greater than 20 years, embryos derived from ICSI and cultured to the blastocyst stage (D5/6), blastocyst culture fluid collected in all cycles, and at least one blastocyst resuscitation transfer cycle performed. The exclusion criteria included patients with abnormal uterine cavity morphology, endometrial lesions, endometrial injury, intrauterine effusion, untreated hydrosalpinx, and any circumstances considered unsuitable for participation in this study by the researchers.

NICS and FET

All patients underwent controlled ovarian stimulation, oocyte retrieval, ICSI, and embryo culture according to our routine protocols. Embryos were evaluated on the morning of day 5/6 using morphological scoring criteria [20] and vitrified separately. Additionally, the spent culture medium was collected and subjected to analysis as described above.

Hormone replacement was performed to prepare the endometrium during the FET cycles. Estradiol pills were administered orally for 10–18 days, followed by progesterone administration when the endometrial thickness reached 8 mm. A single blastocyst was selected based on a morphological assessment to transfer to each patient per cycle and transferred under transabdominal ultrasound guidance.

Whole-genome amplification and next-generation sequencing

Whole-genome amplification was performed using culture media and whole embryos, followed by library preparation using ChromInst (Yikon Genomics; EK100100724 NICS Inst Library Preparation Kit). Next-generation sequencing was performed on the MiSeq platform (Illumina, San Diego, CA, USA), which yielded ∼2 million sequence reads for each sample [6, 16].

The CNV of each sample was calculated as previously described [6]. R program was used to plot the copy number of each bin to visualize CNV profiles for all 24 chromosomes. The minimum resolution of CNVs using this approach was 10 Mb. The embryo was classified as “euploid” when the degree of mosaicism identified was below 30 %, “aneuploid” when the degree of mosaicism was above 70 % [21], and “mosaic” (euploid-aneuploid mosaics) when the degree of mosaicism was 30 %–70 %. In our ploidy analyses, embryos with five or more abnormal chromosomes were defined as chaotic/abnormal, and the data were not sufficient for their analysis because amplification failure was defined as not available. Following niPGT-A analysis (≥ 60 % mosaic rates were set as the threshold to distinguish aneuploid from euploid), embryos were divided into euploid, aneuploid, and chaotic groups.

Embryo rating and machine learning classifier

The chromosomal euploidy- or aneuploidy-related copy number pattern in spent culture media was explored using the machine learning random forest (RF) algorithm combined with the standard chromosomal screening results of whole blastocysts. Gold standard labels (aneuploid or euploid) of whole blastocysts were determined based on a 50 % mosaicism threshold. Other mosaicism cut-offs (e.g., 30 %, 40 %, 60 %, and 70 %) were also evaluated; however, we found that the mosaicism threshold of 50 % was the best for obtaining ploidy consistency of spent culture media and whole blastocysts. The following 11 features were trained by the machine learning methods: 10M-resolution CNV result, 10M-resolution CNV result redefined by 50 % mosaicism threshold, arm-resolution CNV result, arm-resolution CNV result redefined by 50 % mosaicism threshold, whole chromosome-resolution CNV result, whole chromosome-resolution CNV result redefined by 50 % mosaicism threshold, euploidy number with different resolution result, abnormal chromosome number, highest abnormal mosaicism proportion, largest abnormal fragment size corresponding to the highest mosaicism proportion, and presence of sex chromosome abnormality or not. Multiple machine learning methods, Support vector machine (SVM), k-nearest neighbors (KNN), RF, XGBoost, Nnet were constructed to train and test the paired spent culture media-whole blastocyst data using R package caret 6.0-86. Embryo rating was measured based on euploidy probability obtained from machine learning algorithms. The detailed running process was as follows: (i) training data were split into subsets randomly by bootstrap resampling; (ii) multiple decision trees (500) were generated for each training dataset, and feature subsets; (iii) euploidy probability predictions from 500 decision trees were averaged to obtain the final prediction by the model; (iv) Out of Bag error was used to evaluate the model; and (v) the model was iteratively trained until the optimal one was obtained. The embryos were ranked based on different thresholds of chromosomal euploidy probability obtained from the optimal model of the RF model.

Statistical analysis

The Shapiro-Wilk test was used to test the normality of continuous variables. Continuous variables were presented as mean with standard deviations (normal distribution) or median with interquartile ranges (non-normal distribution). Categorical variables were described as frequencies and percentages. Continuous variables were compared using one-way analysis of variance when normality and homogeneity of variance assumptions were satisfied. Otherwise, the Kruskal-Wallis H test was applied. For categorical variables, differences between groups were compared using chi-squared (χ²) or Fisher’s exact tests. Furthermore, a logistic regression model was used to evaluate the effects of each group after adjusting for confounding variables. All p values were two-sided, and results with p values of less than 0.05 were considered statistically significant. The sample size was calculated based on the comparison of three groups A, B, and C. The test level was set to 0.05, and the test efficiency was set to 80 %. SPSS software program (version 20.0; IBM, Armonk, NY, USA) was used for analysis.

Table 1. Pregnancy outcomes in patients with different noninvasive chromosome screening grades.

Group	A	B	C	P value
Pregnancy rate,	61.7 %	57.0 %	40.7 %	0.010
n (%)	(82/133)	(49/86)	(24/59)	0.010
Pairwise comparisons	B vs. C			0.020
	A vs. C			0.011
	A vs. B			0.590
Miscarriage rate,	15.9 %	14.3 %	33.3 %	0.068
n (%)	(13/82)	(7/49)	(8/24)	0.068
Pairwise comparisons	B vs. C			0.021
	A vs. C			0.026
	A vs. B			0.655
Ongoing pregnancy rate,	51.9 %	48.8 %	27.1 %	0.003
n (%)	(69/133)	(42/86)	(16/59)	0.003
Pairwise comparisons	B vs. C			0.009
	A vs. C			0.001
	A vs. B			0.797
Live birth rate,	50.4 %	45.3%	27.1 %	0.004
n (%)	(67/133)	(39/86)	(16/59)	0.004
Pairwise comparisons	B vs. C			0.022
	A vs. C			0.006
	A vs. B			0.728

Prediction of embryo euploidy probability using RF machine learning

We developed an embryo grading strategy according to the euploidy probability of embryos using 345 paired blastocyst culture medium and whole blastocyst samples with confirmed chromosomal ploidy status and features such as chromosomal mosaicism proportion, segmental aneuploidy size, multiple chromosomal abnormalities, and sex chromosome information, which were considered when predicting the euploidy probability of each embryo using spent culture medium. SVM, KNN, RF, extreme gradient boosting (XGBoost), and neural network (Nnet) algorithms were used and compared using 70 % training/30 % testing split strategy (Table S1). In a clinical application scenario, maintaining a low false-negative rate is especially important to facilitate the implantation of a normal embryo. Hence, the negative predictive value (NPV) is given more consideration than other measures. To this end, combined with the overall performance value and area under the receiver operating characteristic curve (AUC), RF, which showed an AUC of 0.92 and NPV of 0.93. RF was selected for further analysis. Consequently, the ensemble learner with multiple decision trees based on a bagging strategy was trained to obtain the optimal tree in which the chromosomal ploidy concordance between samples from culture media and whole blastocysts was the highest (Figure 1A). Each culture medium was assigned a euploidy probability at the end of the analysis and as expected, the higher the euploidy probability predicted by NICS, the higher the likelihood of euploidy in the whole blastocyst (Figure 1B).

Development of an embryo grading strategy using euploidy probability predicted by NICS

We graded embryos as A, B, or C according to their euploidy probabilities predicted by NICS, with euploidy probabilities of ≥ 0.94, 0.7–0.94, and ≤ 0.7 for grades A, B, and C, respectively. This grading strategy balanced both the euploidy predictive value and specificity, two important indices for embryo selection for implantation (Figure 2A).

With a 10-fold cross-validation study of 345 paired training data, each sample could be predicted once. Using the grading above criterion, we observed that 37 % of the embryos were categorized as grade A, whereas 29 % and 34 % of the embryos were categorized as grades B and C, respectively (Figure 2B). A-grade embryos (91.4 %) from spent culture media showed consistent chromosome euploidy with corresponding whole blastocysts, with the proportion of B- and C-grade embryos being 76.4 % and 34.6 %, respectively (Figure 2C).

Furthermore, we explored the ploidy pattern of different grade-labeled embryos from spent culture media. The chromosomal copy number pattern of A-grade embryos showed absolute euploidy, low mosaicism, or small segmental aneuploidy compared to B- or C-grade embryos (Figure S1). As expected, C-grade embryos showed large-scale irregularities in fragment size, such as those at the whole chromosomal level and high mosaicism or 100 % aneuploidy pattern (Figure S1).

NICS grading system as a predictor to reduce the waste of potentially qualified embryos

The clinical outcomes of the patients (n = 266) who underwent 278 single frozen embryo resuscitation transplantation based on morphological assessment were observed (Figure 3). The transferred embryos were divided into three groups according to the NICS pattern. Demographics among groups, including women’s age, body mass index, serum sex hormone levels, endometrial thickness on transplantation day, and men’s age and infertility type, are shown in Table S2.

For each blastocyst transfer performed, we collected the corresponding culture medium and performed NICS. The NICS results were blinded to the investigator until clinical outcomes were obtained and revealed. Then, a total of 278 FET embryos were assessed using the NICS grading system, and three grades, A, B, or C, were assigned. We found that the clinical outcomes were not significantly different between A- and B-grade embryos in terms of pregnancy rate (61.7 % versus 57.0 %, P = 0.590), ongoing pregnancy rate (51.9 % versus 48.8 %, P = 0.797), live birth rate (50.4 % versus 45.3 %, P = 0.728), and miscarriage rates (15.9 % versus 14.3 %, P = 0.655). However, the pregnancy rate, ongoing pregnancy, and live birth rate in grade C were 40.7 %, 27.1 %, and 27.1 %, respectively, which were significantly lower than those in grades A and B (Table 1).

At the same time, the embryos were assessed using niPGT-A (≥ 60 % mosaic rates were set as the threshold to distinguish aneuploid from euploid) and divided into euploid, aneuploid, and chaotic groups. Of the 278 embryos with conclusive outcomes, 161 embryos were classified into the euploid group, 76 embryos into the aneuploid group, and 41 embryos into the chaotic group. Considering A- and B-grade embryos as transplantable, we observed an improved embryo transplantable rate of A- and B-grade embryos compared to that for euploid embryos (78.8 % (219/278) versus 57.9 % (161/278), respectively). Similar clinical outcomes were observed with the comparison of A- and B-grade embryos and the euploid group. They were as follows: pregnancy rate: 59.8 % versus 60.9 %, ongoing pregnancy rate: 50.70 % versus 51.6 %, live birth rate: 48.4 % versus 49.1 %, and miscarriage rate: 15.3 % versus 15.3 % (Figure 4. According to the results of the niPGT-A analysis, aneuploid and chaotic embryos were considered untransplantable embryos. The miscarriage rate of C-grade embryos was higher than that in aneuploid and chaotic groups (33.3 % versus 22.8 %), while clinical pregnancy, ongoing pregnancy, and live birth rates of C-grade embryos were lower than those in aneuploid and chaotic groups (40.7 % versus 48.7%, 27.1 % versus 37.6 %, 27.1 % versus 36.8 %, respectively) (Figure 4).

At present, the typical method used for embryo selection depends on its morphological score. Despite its convenience, morphology is a poor indicator of embryo chromosomal composition and is highly subjective [22]. PGT-A has been used extensively in clinical IVF to avoid the transfer of aneuploid embryos; however, its invasive nature and embryo mosaicism are controversial [23, 24]. A mosaic embryo possesses two or more chromosomally distinct cell lines within the embryo, and trophectoderm (TE) biopsy may not truly reveal the chromosomal status of the inner cell mass (ICM) that develops into the fetus. Bolton et al. [19] showed that the source of cell-free DNA from apoptotic cells in the culture fluid is mainly the ICM, which is more representative of the fetus than the TE-biopsied cells. However, owing to the mosaic nature of human embryos, it is likely that DNA in the culture medium is a mixture of DNA obtained from both ICM and apoptotic cells cleared from the embryo, which can result in a false-positive during chromosomal screening [19]. Therefore, labeling of embryos as ‘aneuploid’ or ‘euploid’ based on dichotomous screening outcomes may result in the false exclusion of several embryos, increasing the cancellation rate of IVF cycles.

Mosaicism has a reported incidence rate ranging from 2.0 %–2.9 % [25, 26] to 14.0 %–17.3 % [27, 28], even as high as 48 % [29] because different clinics have different internal standards, thus yielding a fluctuating ploidy concordance of 33 %–100 %. Additionally, the definitions of ploidy concordance are different, especially if all chromosome aberrations (such as fragment aneuploidy and chimerism) are considered. It is not clear whether the segmental or mosaic abnormality is defined as aneuploidy. Moreover, the gold standard samples are different in different studies. Kuznyetsov et al. [30], Jiao et al. [16], Xu et al. [6], and Li et al. [11] found that the consistency between culture medium and the whole embryo was greater than 90 %. Using whole embryos as a gold standard, Huang et al. [17]even observed a higher consistency in culture medium DNA than in TE-biopsied samples, indicating that culture medium-based PGT-A could potentially be more accurate than TE-based PGT-A. For niPGT-A, mosaicism is defined as the percentage of abnormal cells of ≥ 40 % [16] or ≥ 50 % [14]. According to this standard, Huang et al. [17] defined mosaicism of above 60 % as the threshold to distinguish aneuploid from euploid for the same population, and the NICS grading system and 60 % threshold method were used to evaluate the embryos at the same time. The embryos < 60 % were regarded as euploid embryos, and we found that the clinical outcomes of A- and B-grade embryos were similar to those of euploid embryos. Surprisingly, the number of euploid and A- and B-grade embryos was found to be 57.9 % and 78.8 %, respectively.

NiPGT-A is a suitable method for selecting embryos with normal chromosomal ploidy; however, embryos with aneuploid, mosaic, and chaotic CNV are often not considered for transfer. Although the pregnancy rate of the aneuploid group was significantly lower than that of the euploid group, a 36.8 % live birth rate was achieved. The NICS grading system is more accurate as it considers features such as chromosomal mosaicism proportion, segmental aneuploid size, and multiple chromosomal abnormality information for the RF model. To verify the correlation between this scoring system and the clinical outcome, embryos were graded as A, B, or C based on different cut-off criteria of their euploidy probability. Different grades of embryos were found to have different characteristics. A-grade embryos showed absolute euploidy, low mosaicism, or small segmental aneuploidy; B-grade embryos showed a low rate of mosaicism; C-grade embryos exhibited a high mosaicism proportion or a 100 % aneuploid pattern (Figure S1). Generally, smaller abnormal fragments and lower mosaic ratios tend to predict higher euploid probabilities.

Recently, there have been several studies employing machine learning-guided approaches for the selection of human blastocysts via imaging or video, including morphometric analysis by time-lapse imaging [31], mathematical and statistical tools [32, 33], and computer-assisted scoring [34, 35]. However, these methods are based on subjective developmental and morphological characteristics.

Indeed, morphology has a weak association with embryo ploidy status, which is the most critical factor for sustained implantation. Morphological analysis cannot be relied on to ensure the transfer of chromosomally normal embryos [25, 36]. The recent implementation of time-lapse devices, which allows continuous embryo monitoring in stable culture conditions, has led some teams to test the effectiveness of morphokinetic parameters as predictors of embryo ploidy, but the conclusions remain controversial [37, 38]. The advantage of our method is that the model predicts the euploidy probability based on the CNV features of each embryo in the spent culture medium.

We compared many machine learning models such as SVM, KNN, RF, XGBoost, and Nnet, and finally selected RF for classifier construction based on favorable AUC and NPV performance. RF is an ensemble learner with multiple decision trees based on a bagging strategy. It corrects overfitting in a decision tree model and offers better results than those predicted using an individual model.

In conclusion, our results suggest that the AI NICS grading system may be ideal for optimizing the selection of a single embryo for transfer, thus maximizing the chance of live birth, avoiding the waste of potentially qualified embryos, and serving as an effective noninvasive approach. In the future, with the accumulation of comparative studies on NICS and large clinical data, including morphological grade, DNA concentration, aneuploidy screening, mitochondrial copy number, clinical outcome, and patient information, an AI system based on machine learning may be developed to evaluate the implantation potential of each embryo

AI, Artificial intelligence; CCS, Comprehensive chromosomal screening; CM, Culture medium; CNV, Copy number variation; FET, Frozen-thawed embryo transfer; ICM, Inner cell mass; KNN, K-nearest neighbors; NGS, Next-generation sequencing; NPV, Negative predictive value; PGD, Pre-implantation genetic diagnosis; PPV, Positive predictive value; RF, Random forest; SCM, Spent culture media; SVM, Support vector machine; WGA, Whole-genome amplification

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Nanjing Jinling Hospital (reference number: 2016NJKY-028; trial registration number: ChiCTR-RRC-17010396).

Consent for publication

Not applicable.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests

Yaxin Yao, Yangyun Zou, Menglin Ma, Jieliang Ma, and Sijia Lu are employees of Yikon Genomics.

Funding

This work was supported by the National Key Research and Development Program of China (grant nos. 2018YFC1003800 and 2018YFC1003100), the Jiangsu Provincial Key R&D Programme (grant no. BE2018714), the Six Talent Peaks Project in Jiangsu Province (grant no. 2017-WSW-033), Special Research on Military Family Planning (grant no. 18JS004), and the State Key Laboratory of Reproductive Medicine of Nanjing Medical University (grant no. SKLRM-K201803).

Authors' contributions

Li Chen, Wen Li, Yuxiu Liu, Zhihang Peng, Xiaoming Teng, Sijia Lu, Bing Yao, and Haixiang Su made substantial contributions to conception and design. Liyi Ca, Ningyuan Zhan, Juanjuan Xu, Liang Wang, Yaxin Yao, and Menglin Ma performed experiments. Li Chen, Wen Li, Liyi Cai, Ningyuan Zhang, Juanjuan Xu, Xiaoming Teng, Yaxin Yao, Yangyun Zou, Bing Yao, and Haixiang Sun contributed to data acquisition. Li Chen, Wen Li, Yuxiu Liu, Yaxin Yao, Yangyun Zou, Sijia Lu, and Bing Yao performed data analysis. Li Chen, Wen Li, Yuxiu Liu, Yaxin Yao, Yangyun Zou, Sijia Lu, and Bing Yao drafted the manuscript. All authors revised and approved the manuscript.

Acknowledgments

The authors would like to thank Shiping Bo and Shujie Ma for their assistance in NGS data analysis.

Authors' information

1. Scott RT, Ferry K, Su J, Tao X, Scott K, Treff NRJF, et al. Comprehensive chromosome screening is highly predictive of the reproductive potential of human embryos: a prospective, blinded, nonselection study. Fertil Steril. 2012;97(4):870-5.

2. Rubio C, Bellver J, Rodrigo L, Castillón G, Guillén A, Vidal C, et al. In vitro fertilization with preimplantation genetic diagnosis for aneuploidies in advanced maternal age: a randomized, controlled study. 2017; doi: 10.1016/j.fertnstert.2017.03.011.

3. Zeng Y, Lv Z, Gu L, Wang L, Sha JJC, Research T. Preimplantation genetic diagnosis (PGD) influences adrenal development and response to cold stress in resulting mice. 2013;354(3):729-41.

4. Wu Y, Lv Z, Yang Y, Dong G, Yu Y, Cui Y, et al. Blastomere biopsy influences epigenetic reprogramming during early embryo development, which impacts neural development and function in resulting mice. 2014;71(9):1761-74.

5. Guzman L, Nuñez D, López R, Inoue N, Portella J, Vizcarra F, et al. The number of biopsied trophectoderm cells may affect pregnancy outcomes. 2018; doi: 10.1007/s10815-018-1331-1.

6. Xu J, Fang R, Chen L, Chen D, Xiao JP, Yang W, et al. Noninvasive chromosome screening of human embryos by genome sequencing of embryo culture medium for in vitro fertilization. 2016:11907.

7. Fang R, Yang W, Zhao X, Xiong F, Guo C, Xiao J, et al. Chromosome screening using culture medium of embryos fertilised in vitro: a pilot clinical study. 2019;17.

8. Coonen E, Rubio C, Christopikou D, Dimitriadou E, Open MRJHR. ESHRE PGT Consortium good practice recommendations for the detection of structural and numerical chromosomal aberrations. 2020;2020(3).

9. Gleicher N, Albertini DF, Barad DH, Homer H, Noyes NJRB, Endocrinology. The 2019 PGDIS position statement on transfer of mosaic embryos within a context of new information on PGT-A. 2020;18(1).

10. Feichtinger M, Vaccari E, Carli L, Wallner E, Mädel U, Figl K, et al. Non-invasive preimplantation genetic screening using array comparative genomic hybridization on spent culture media: a proof-of-concept pilot study. 2017;34(6):583-9.

11. Li P, Zhe S, Yao Y, Huang T, Mao R, Huang J, et al. Preimplantation genetic screening with spent culture medium/blastocoel fluid for in vitro fertilization. 2018;8(1):9275.

12. Ho JR, Arrach N, Rhodes-Long K, Ahmady A, Mcginnis LKJF, Sterility. Pushing the limits of detection: investigation of cell-free DNA for aneuploidy screening in embryos. 2018;110(3).

13. Vera-Rodriguez M, Diez-Juan A, Jimenez-Almazan J, Martinez S, Navarro R, et al. Origin and composition of cell-free dna in spent medium from human embryo culture during preimplantation development. Hum Reprod. 2018;33(4):745-56.

14. Rubio C, Rienzi L, Navarro-Sánchez L, Cimadomo D, Fertility CSJ, Sterility. Embryonic cell-free DNA versus trophectoderm biopsy for aneuploidy testing: concordance rate and clinical implications. 2019;112(3):510-9.

15. Yeung QSY, Zhang YX, Chung JPW, Lui WT, Kwok YKY, Gui B, et al. A prospective study of non-invasive preimplantation genetic testing for aneuploidies (NiPGT-A) using next-generation sequencing (NGS) on spent culture media (SCM). J Assis Reprod Genet. 2019;36(8):1609–21.

16. Jiao J, Shi B, Sagnelli M, Yang D, Wang XJHR. Minimally invasive preimplantation genetic testing using blastocyst culture medium. 2019;34(7).

17. Huang L, Bogale B, Tang Y, Lu S, Racowsky CJPotNAoS. Noninvasive preimplantation genetic testing for aneuploidy in spent medium may be more reliable than trophectoderm biopsy. 2019;116(28):201907472.

18. Victor AR, Tyndall JC, Brake AJ, Lepkowsky LT, Murphy AE, Griffin DK, et al. One hundred mosaic embryos transferred prospectively in a single clinic: exploring when and why they result in healthy pregnancies. 2019;111(2):280.

19. Bolton H, Graham SJL, Niels VDA, Kumar P, Theunis K, Fernandez Gallardo E, et al. Mouse model of chromosome mosaicism reveals lineage-specific depletion of aneuploid cells and normal developmental potential. 2016;7:11165.

20. Munné S, Wells D, Cohen JJF, Sterility. Technology requirements for preimplantation genetic diagnosis to improve assisted reproduction outcomes. 2010;94(2):408-30.

21. Pagliardini L, Viganò P, Alteri A, Corti L, Papaleo EJRbo. Shooting STAR: reinterpreting the data from the 'Single Embryo TrAnsfeR of Euploid Embryo' randomized clinical trial. 2020;40(4).

22. Popovic M, Dh Eed En EA, Christodoulou C, Taelman J, Dhaenens L, Van NF, et al. Chromosomal mosaicism in human blastocysts: the ultimate challenge of preimplantation genetic testing? 2018;(7):7.

23. Antonio C, Laura R, Danilo C, Roberta M, Thomas E, Graham W, et al. Correlation between standard blastocyst morphology, euploidy and implantation: an observational study in two centers involving 956 screened blastocysts. 2014;(6):1173-81.

24. Murray MJJMHR. Comprehensive analysis of karyotypic mosaicism between trophectoderm and inner cell mass. 2010;16(12):944-9.

25. Fragouli E, Alfarawati S, Daphnis DD, Goodall NN, Mania A, Griffiths T, et al. Cytogenetic analysis of human blastocysts with the use of FISH, CGH and aCGH: scientific data and technical evaluation. 2011;26(2):480-90.

26. Northrop LE, Treff NR, Levy B, Scott RT. SNP microarray-based 24 chromosome aneuploidy screening demonstrates that cleavage-stage FISH poorly predicts aneuploidy in embryos that develop to morphologically normal blastocysts. 2010;16(8):590-600.

27. Bielanska M, Tan SL, Ao A. Chromosomal mosaicism throughout human preimplantation development in vitro: incidence, type, and relevance to embryo outcome. Hum Reprod. 2002;17(2):413-9.

28. Valeriy K, Svetlana M, Ran A, Rina A, Gelareh M, Zenon I, et al. Evaluation of a novel non-invasive preimplantation genetic screening approach. 2018;13(5):e0197262.

29. Rubio I, Galán A, Larreategui Z, Ayerdi F, Bellver J, Herrero J, et al. Clinical validation of embryo culture and selection by morphokinetic analysis: a randomized, controlled trial of the EmbryoScope. 2014;102(5):1287-94.e5.

30. Iwata K, Sato M, Matsumoto I, Shimura T, Yumoto K, Negami A, et al. Deep learning based on images of human embryos obtained from high-resolusion time-lapse cinematography for predicting good-quality embryos. 2018;110(4).

31. Curchoe CL, Bormann CL. Artificial intelligence and machine learning for human reproduction and embryology presented at ASRM and ESHRE 2018. J Assist Reprod Genet. 2019;36(4):591-600.

32. Santos FE, Noble JA, Poli M, Griffiths T, Emerson G, Wells DJ. A method for semi-automatic grading of human blastocyst microscope images. 2012;(9):2641-8.

33. Vermilyea M, Hall J, Diakiw SM, Johnston A, Perugini MJ. Development of an artificial intelligence-based assessment model for prediction of embryo viability using static images captured by optical light microscopy during IVF. 2020;35(4).

34. Alfarawati S, Fragouli E, Colls P, Stevens J, Gutiérrez-Mateo C, Schoolcraft WB, et al. The relationship between blastocyst morphology, chromosomal abnormality, and embryo gender. 2011;95(2):520-4.

35. Zaninovic N, Irani M, Meseguer MJF, Sterility. Assessment of embryo morphology and developmental dynamics by time-lapse microscopy: is there a relation to implantation and ploidy? 2017;108(5):722-9.

36. Reignier A, Lammers J, Barriere P, Freour TJ. Can time-lapse parameters predict embryo ploidy? A systematic review. 2018;36(4):380-7.

37. Gardner DK, Schoolcraft WB, Beyond G. In vitro culture of human blastocyst. 1999.

38. Lawrenz B, Khatib IE, Lián A, Bayram A, Fatemi HM. The clinician dilemma with mosaicism - An insight from inner cell mass biopsies. 2019;34(6).

Download PDF

Version 1

posted

You are reading this latest preprint version

Machine Learning-Guided Noninvasive Embryo Selection for Clinical in Vitro Fertilization Treatment to Avoid Wasting Potentially Qualified Embryos

Status:

Version 1

Abstract

Figures

Background

Methods

Results

Discussion

Conclusion

Abbreviations

Declarations

References

Status:

Version 1