Prediction model for missed abortion of patients treated with IVF-ET based on XGBoost: establishment and evaluation

doi:10.21203/rs.3.rs-1637185/v1

Download PDF

Research Article

Prediction model for missed abortion of patients treated with IVF-ET based on XGBoost: establishment and evaluation

https://doi.org/10.21203/rs.3.rs-1637185/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Aim: To establish a model based on XGBoost to predict the risk of missed abortion in patients treated with in vitro fertilization-embryo transfer (IVF-ET), evaluate its prediction ability, and compare with the traditional logical regression model.

Methods: The clinical data of 1017 infertile women treated with IVF-ET were collected retrospectively. The independent risk factors were screened by Univariate analysis and binary logistic regression analysis, and then all cases were randomly divided into training set and test set according to the proportion of 7:3 for model construction and verification evaluation. The prediction models are constructed by traditional logical regression method and XGBoost method respectively, and then the prediction performance of the two models is tested by resampling.

Results: Binary logistic regression analysis showed that female age, male age, abnormal ovarian structure, prolactin (PRL), anti-Müllerian hormone (AMH), activated partial thromboplastin time (APTT), anti cardiolipin antibody (ACA) and thyroid peroxidase antibody (TPO-Ab) were the factors that independently influenced missed abortion (P<0.05). The AUC score and F1 score with the training set of XGBoost model were 0.877±0.014 and 0.730±0.019. They were significantly higher than those of logistic model (0.713±0.013 and 0.568±0.026). In the evaluation of the test set, the AUC score and F1 score of XGBoost model were 0.759±0.023 and 0.566±0.042. They were also higher than those of logistic model (0.695±0.030 and 0.550±049).

Conclusions: We established a prediction model based on XGBoost algorithm, which can accurately predict the risk of missed abortion in patients with IVF-ET. The performance of this model is better than the traditional logical regression model.

IVF-ET

Missed abortion

Machine Learning

Prediction model

XGBoost

Missed abortion means that the embryo stops developing for a variety of reasons, but the dead embryo still remains in the uterine cavity[1,2]. With the change of people's living environment and the increase of work pressure, the incidence of missed abortion is increasing year by year, up to 8-13.4% [3]. It has been found that factors affecting embryonic development include female age, genetic factors, endocrine diseases, immune factors, pregnancy infection, pregnancy history, behavioral factors and so on. It is precisely because of the complexity of the etiology, there is no consensus in the pathogenesis at present[4,5,6,7,8].

Missed abortion seriously harms the health of women of childbearing age. If the embryos that stop developing are not found in time and remain in the uterine cavity for a long time, it can cause abnormal maternal blood coagulation, uterine adhesion, DIC, and even pose a threat to maternal life safety[9]. Especially for women who receive assisted reproductive technology, missed abortion brings more physical and mental harm to them, and they are under greater economic and psychological pressure[10]. IVF-ET is one of the most effective reproductive technologies for clinical treatment of infertility. However, many studies have reported that the incidence of missed abortion in the population treated with IVF-ET as high as 18%-30%[11,12]. For families eager to have children, it is a huge strike for them to be diagnosed with missed abortion after the complicated treatment process. So in order to reduce the occurrence of this kind of adverse outcome, we should comprehensively analyze the high risk factors and take targeted preventive measures. At present, many scholars have screened the high risk factors of missed abortion, but there are few studies targeted infertile people, the way to predict high risk groups is even more rare. Therefore, it is necessary to find a method for predicting missed abortion for patients treated with IVF-ET, and take this as a theoretical basis for their targeted personalized diagnosis and treatment, as far as possible to improve the successful pregnancy rate.

In recent years, predictive models have been widely used in the medical field, and common modeling methods include traditional logical regression and machine learning. Machine learning is the core of artificial intelligence, and its ability of data processing, induction and synthesis is far better than other statistical methods. A lifting algorithm based on regression tree called extreme gradient boosting (XGBoost) is one of the hot spots of machine learning. It has been widely used because of its short training time and high accuracy[13]. This study integrates a large number of clinical characteristics of IVF-ET patients to construct a XGBoost prediction model to predict the risk of missed abortion, and compared with the traditional logical regression prediction model to evaluate the prediction performance of the model.

Study design

This was a retrospective study that was approved by the Ethics Committee of the Affiliated Women and Children’s Hospital of Qingdao University. Because of the retrospective nature of the study, informed patient consent was waived by the Ethics Committee of the Affiliated Women and Children’s Hospital of Qingdao University.

Participants

Our subjects were 1017 infertile women who were treated in the Reproductive Center of Qingdao Women's and Children's Hospital from September 2019 to February 2022. The inclusion criteria were patients with intrauterine singleton pregnancy after IVF treatment. The exclusion criteria include: (1) couples receiving donated sperm or eggs; (2) pregnancy failure due to special reasons such as medication or trauma; (3) incomplete clinical data. According to the pregnancy outcome, the missed abortion population was located in the observation group (n=340), and the normal pregnancy population was taken as the control group (n=677).

Diagnostic criteria

According to the ultrasonic diagnosis, the missed abortion accorded with one of the following criteria[14,15]: (1) the head and hip length of the embryo ≥ 7mm, no heartbeat; (2) the diameter of the gestational sac ≥ 25mm and no embryo; (3) for the gestational sac without yolk sac, there was no heartbeat after 2 weeks; (4) for the gestational sac with yolk sac, there was still no heartbeat after 11 days.

Variables included

The included predictive variables are determined according to literature review and expert group discussion, and all of the following data were obtained from medical records: (1) genetic factors: chromosomal abnormalities in both men and women; (2) female endocrine levels: Follicle-stimulating hormone (FSH), luteinizing hormone (LH), estradiol (E2), progesterone (P), testosterone (T), prolactin (PRL), AMH; (3) thyroid hormone levels: FT3, FT4, TSH, TG-Ab, TPO-Ab; (4) immune factors: anticardiolipin antibody (ACA), antinuclear antibody (ANA); (5) coagulation function tests: PT, APTT, TT, FIB, D-dimer; (6) infection factors: leucorrhea test, chlamydia trachomatis, neisseria gonorrhoeae, TORCH test, etc; (7) abnormal ovarian structure: ultrasonography showed abnormal number, size, polycystic changes or space occupying lesions of the ovary; (8) uterine structural lesions: uterine malformation, leiomyoma, adenomyosis, scar diverticulum, etc; (9) male sperm abnormality; (10) other factors: age of both men and women, female BMI, fasting blood glucose, blood type, history of uterine surgery, type of infertility, years of infertility, type of cycle, number of sinus follicle and number of eggs obtained, number of high quality embryos, thickness and type of endometrium on the day of transfer, quality of embryo. Data processing and analysis

All the data were carried out with SPSS 20.0 statistical software firstly. Continuous data in accordance with normal distribution are expressed as average±standard deviation, continuous variables that do not conform to normal distribution are expressed as median (25%~75% percentile), and classified variables are expressed as frequency (percentage). According to the characteristic analysis of all the data, T-test was used for continuous variables in accordance with normal distribution, Mann-Whitney U test was used for continuous variables that did not conform to normal distribution, and chi-square test was used for classified variables. P<0.05 is statistically significant. The independent risk factors were screened by binary logistic regression analysis.

Prediction methods and model evaluation

All the independent risk factors were processed by Python (version 3.0) to develop the prediction model based on XGBoost and traditional logical regression. The train_test_split function of the sklearn package was used to split the data set into 70% as the training data set and 30% as the test data set[16]. The data of the training set are used to construct the prediction model. The predictability of the established model is evaluated by resampling method. Reference indicators include AUC score and F1 score. The feature importance function of the better model was used to rank the importance of variables.

Univariate analysis

In this study, a total of 1017 subjects were included according to the strict standard, including 48 influencing factors, among which Rh blood group, Chlamydia Infections, gonorrhea, trichomonas vaginitis and TORCH test were eliminated because of the large proportion gap between the classification groups. The remaining results are shown in the table 1 and table 2. A total of 12 indicators with statistical differences were selected. Compared with the control group, the average age of both female and male in the observation group was higher, and there were more chromosome abnormalities. In the observation group, there were more abnormal ovarian structure, higher PRL level and lower AMH level, shorter PT and APTT, higher proportion of positive anti cardiolipin antibody (ACA) and antinuclear antibody (ANA), lower serum free thyroid level and higher Anti-thyroid peroxidase antibody (TPO-Ab) positive rate.

Table 1

Comparison of continuous variables between the observation group and control group

Variables	Observation group (n=340)	Control group (n=677)	t/z	P
Female age,years	34.11±4.65	32.12±3.75	-7.378	＜0.001
Male age,years	33.58±5.23	32.95±4.37	-2.046	0.041
BMI,kg/m²	23.53±2.35	23.17±3.28	-1.775	0.076
Years of infertility	3.60±1.47	3.38±2.28	-1.606	0.109
FSH,mIU/ml	7.34±2.19	7.27±3.34	-0.360	0.719
LH,mIU/ml	5.65±3.67	6.23±5.97	1.649	0.099
PRL,ng/ml	23.15±16.38	20.48±9.71	-3.252	0.006
E2,pmol/l	176.65±77.67	187.10±142.22	1.251	0.211
T,ng/ml	0.75±0.26	0.78±0.40	1.378	0.168
P,nmol/l	0.72±0.39	0.78±0.61	1.701	0.089
AMH,ng/ml	2.78±1.48	3.58±2.78	4.964	＜0.001
D-Dimer,mg/l	0.25±0.16	0.25±0.25	-0.128	0.882
PT,s	11.16±0.48	11.26±0.80	2.047	0.041
APTT,s	24.50±2.32	25.32±3.34	4.067	＜0.001
TT,s	18.05±0.67	18.00±1.48	-0.507	0.612
FIB,g/l	2.67±0.38	2.71±0.89	0.823	0.411
ESR,mm/h	13.34±5.69	12.65±7.89	-1.443	0.149
FBG,mmol/l	5.00±0.45	4.94±0.48	-1.770	0.077
FT3,pmol/l	4.80±0.45	4.83±0.53	0.837	0.403
FT4,pmol/l	15.60±1.56	15.89±2.36	2.041	0.042
TSH,μIU/ml	2.09±0.72	2.13±1.21	0.603	0.547
Left basal sinus follicle	7.11±4.01	7.56±4.29	1.596	0.111
Right basal sinus follicle	7.58±4.48	7.77±4.21	0.696	0.487
Total number of eggs obtained	10.24±3.79	9.82±4.54	-1.468	0.142
Endometrial thickness in transfer date,mm	10.26±1.73	10.34±1.60	0.752	0.452
Number of embryos implanted	2 [1,2]	1 [1,2]	-1.000	0.317
Number of high-quality embryos	3 [2,5]	3 [2,4]	-1.323	0.186

Table 2

Comparison of classified variables between the observation group and control group

Variables	Classification	n(%)		χ²	P
		Observation group	Control group
Cycle type	Fresh embryo transfer	190(55.9%)	396(58.5%)	0.632	0.427
	Frozen embryo transfer	150(44.1%)	281(41.5%)
Chromosome abnormality	NO	289(85.0%)	605(89.4%)	4.056	0.044
	YES	51(15.0%)	72(10.6%)
Infertility type	Secondary infertility	147(43.2%)	286(42.2%)	0.091	0.763
	Primary infertility	193(56.8%)	391(57.8%)
Male sperm abnormality	NO	250(73.5%)	513 (75.8%)	0.609	0.435
	YES	90 (26.5%)	164(24.2%)
Abnormal ovarian structure	NO	265(77.9%)	573(84.6%)	6.999	0.008
	YES	75(22.1%)	104(15.4%)
Abnormal uterine structure	NO	257(75.6%)	528(78.0%)	0.742	0.389
	YES	83(24.4%)	149(22.0%)
History of uterine cavity surgery	NO	274(80.6%)	559(82.6%)	0.600	0.439
	YES	66(19.4%)	118(17.4%)
ABO blood group	A	111(32.6%)	199(29.4%)	3.670	0.299
	B	93(27.4%)	223(32.9%)
	O	100(29.4%)	181(26.7%)
	AB	36(10.6%)	74(10.9%)
ACA	(-)	321(94.4%)	661(97.6%)	7.083	0.008
	(+)	19(5.6%)	16(2.4%)
ANA	(-)	330(97.1%)	669(98.8%)	4.030	0.045
	(+)	10(2.9%)	8(1.2%)
TG-Ab	(-)	281(82.6%)	568(83.9%)	0.257	0.612
	(+)	59(17.4%)	109(16.1%)
TPO-Ab	(-)	233(68.5%)	542(80.1%)	16.592	＜0.001
	(+)	107(31.5%)	135(19.9%)
Cleaning degree of vagina	I°	90(26.5%)	200(29.5%)	2.612	0.455
	II°	219(64.4%)	432(63.8%)
	III°	29(8.5%)	42(6.2%)
	IV°	2(0.6%)	3(0.4%)
VVC	NO	337(99.1%)	672(99.3%)	0.060	0.807
	YES	3(0.9%)	5(0.7%)
Quality of implanted embryos	A	311(91.5%)	633(93.5%)	1.400	0.237
	B	29(8.5%)	44(6.5%)
Endometrial type in implant date	A	17(5.0%)	69(10.2%)	9.468	0.050
	A-B	10(2.9%)	12(1.8%)
	B	10(2.9%)	10(2.8%)
	B-C	33(9.7%)	55(8.1%)
	C	270(79.4%)	522(77.1%)

Binary logistic regression analysis

A binary logistic regression model was established with pregnancy outcomes as the dependent variable and the above statistically significant variables as independent variables. The results showed that independent influencing factors of missed abortion included female age, male age, abnormal ovarian structure, PRL, AMH, APTT, ACA and TPO-Ab.

Table 3

Binary logistic regression analysis

Variables	B	SE	OR	95%CI	P
Female age	0.188	0.028	1.207	1.143~1.275		＜0.001
Abnormal ovarian structure	0.622	0.185	1.863	1.295~2.679		0.001
PRL	0.020	0.007	1.020	1.006~1.035		0.006
AMH	-0.121	0.037	0.886	0.824~0.952		0.001
APTT	-0.059	0.025	0.943	0.898~0.990		0.018
ACA	0.903	0.377	2.468	1.180~5.163		0.016
TPO_Ab	0.594	0.162	1.812	1.318~2.490		＜0.001
Male age	-0.115	0.024	0.891	0.851~0.934		＜0.001

Construction and evaluation of prediction model

We use all the influence factors in Table 3 to establish the prediction model. The hyperparameters used in the XGBoost model are as follows：objective-binary:logistic, learning rate-0.01, max depth-4, min child weight-2, reg lambda-1. The hyperparameters used in the logical regression model are:Regularization factor (C)-1,max iter-100, penalty-l2, tol-0.0001.AUC score and FI score were used to evaluate the performance of XGBoost model and Logistic regression model (table 4). The AUC score of XGBoost model was 0.877±0.014. It was significantly higher than that of logistic model (0.713±0.013). ROC curves of two prediction models are drawn in Figure 1. The F1 score of XGBoost model was 0.730±0.019. It was also significantly higher than that of logistic model (0.568±0.026). XGBoost model performed better than the logistic model. The predictability of the two models is evaluated by resampling method, and XGBoost still has a better performance (table 5 and Figure 2). Therefore, the ranking of importance is also based on the XGBoost model ( Figure 3).

Table 4

training set results

	Model	AUC (95%CI)	Accuracy	Sensitivity	Specificity	F1 Score
Mean	logistic	0.713 (0.673~0.753）	0.668	0.657	0.676	0.568
SD	logistic	0.013 (0.014~0.013）	0.021	0.093	0.073	0.026
Mean	XGBoost	0.877 (0.851~0.904）	0.802	0.795	0.807	0.730
SD	XGBoost	0.014 (0.015~0.012）	0.021	0.038	0.045	0.019

Table 5

test set results

	Model	AUC (95%CI)	Accuracy	Sensitivity	Specificity	F1 Score
Mean	logistic	0.695 (0.633~0.758）	0.635	0.659	0.657	0.550
SD	logistic	0.030 (0.032~0.028）	0.042	0.074	0.065	0.049
Mean	XGBoost	0.759 (0.698~0.820）	0.705	0.587	0.861	0.566
SD	XGBoost	0.023 (0.026~0.021）	0.021	0.071	0.045	0.042

Assisted reproductive technology has been used in clinic for many years, IVF-ET technology is also known as the "last hope" of many infertile couples. Missed abortion as an important factor to reduce the success rate of pregnancy brings the sense of frustration and disappointment to patients. Only clarifying the clinical influencing factors of missed abortion after IVF-ET can we provide a basis for etiological treatment. In this study, 8 independent influencing factors were discovered.

Our study found that there was abnormal coagulation function in people with missed abortion, such as the shortening of APTT and PT, suggesting the hypercoagulable state of blood. The possible reason is that the hypercoagulable state of blood can selectively affect the blood circulation of uterus and placenta, form microthrombus in the placenta, cause local placental infarction, decrease of placental blood supply, embryo or fetal ischemia and hypoxia, leading to embryo grow arrest[17].

The hypothalamus-pituitary-ovary axis regulates the complex endocrine system of the body. When any part of the axis is abnormal, it can lead to adverse pregnancy outcome. In this study, it was found that the level of PRL was also significantly increased in the observation group, and hyperprolactinemia was common in pituitary dysfunction or pituitary space occupying lesions. Increased prolactin can inhibit the synthesis and release of gonadotropin, affect the development of follicles, cause ovulation disorders, and affect the development of embryos, resulting in infertility or missed abortion[18]. In addition, the level of AMH in the placenta group decreased and was often accompanied by structural lesions of the ovary. Some studies found that AMH was negatively correlated with the risk of early spontaneous abortion[19], and the structural changes of the ovary would affect its function. Progesterone is an important condition for successful implantation of fertilized eggs and pregnancy, and good ovarian function is an important condition for normal pregnancy[20].

Age has always played an important role in the success of pregnancy. With the increase of age, the quality of oocytes decreases, mainly manifested in the occurrence of errors in the process of oocyte meiosis, the formation of aneuploidy, chromosome translocation, inversion and so on. Embryos formed by such gametes are at greater risk of spontaneously stopping development[21,22]. Our study not only confirms the above point of view, but also finds that male age is also an important influencing factor. Some other studies have shown that the male sperm quality decreases with the increase of male age, and the probability of gene mutation increases[23,24]. In our study, sperm abnormality has no effect on the results, the possible reason is that the standard of abnormal sperm in is based on clinical diagnosis, such as oligospermia, asthenospermia and sperm deformity. Rather than the quality of sperm for in vitro fertilization in the laboratory. Recently, China's three-child policy has been implemented but the reproductive desire of young women has been decreasing. The proportion of advanced pregnant women in China will increase. So it is worth thinking about what policies should be adopted to encourage women to give birth at the appropriate age.

Another finding of our study is that the level of FT4 in people with embryo termination was lower than that in normal people, and the positive rate of TPO-Ab increased significantly, suggesting that thyroid dysfunction may have adverse effects on embryos. Previous studies have shown that hypothyroidism may cause adverse effects such as embryo termination and fetal malformations[25]. TPO-Ab and TG-Ab are specific indicators of thyroid autoimmunity. Abnormal levels of TPO-Ab and TG-Ab can cause autoimmune hypothyroidism. Some studies[26,27] have shown that positive TPO-Ab early pregnant women have a corresponding increased risk of abortion. Therefore, we should pay attention to pre-pregnancy thyroid function examination, and give reasonable treatment to women who are diagnosed with thyroid diseases and abnormal laboratory indexes.

The last influencing factor is immune factor. ACA and ANA are human autoimmune antibodies. This study found that these two antibodies showed an increasing trend in missed abortion people. It was found that ACA may act on the membrane phospholipids of placental vascular endothelial cells and platelets in the early stage of pregnancy, block prostacyclin synthesis, and lead to placental embolism. ANA may be related to its influence on DNA replication and immune abnormalities[28].

Synthesize the above factors, our study attempts to establish a XGBoost-based missed abortion risk prediction model in patients treated with IVF-ET, and found that its prediction performance is better than the traditional logical regression model. Machine learning is a new field of artificial intelligence, the XGBoost algorithm in this study belongs to the integrated tree algorithm, and the integrated tree algorithm is a popular algorithm in the field of machine learning in recent years. It may replace the classical logistic model to predict the risk of missed abortion, and take this as a theoretical basis to make targeted preventive measures. When predicting the risk of missed abortion, doctors should intervene the high-risk factors in advance and monitor these people in time after confirming clinical pregnancy to avoid missing diagnosis leading to long-term retention of embryos in the uterus.

Limitations

One of the limitations of this study is that the inclusion index does not reach an ideal state. Embryonic chromosomes, bad living habits, history of close contact with pets and environmental factors are not included, because this study is a retrospective study. there are some deficiencies in clinical data, and the subjective indicators are difficult to be standardized. If there is an opportunity to complete the inclusion of indicators in the future, a more accurate forecasting model may be constructed. Another limitation is that the research population has strong regional characteristics and does not have wide representativeness. It is hoped that the data from multiple medical centers can be used for the external verification of the model.

In summary, we found that independent influencing factors of missed abortion included female age, male age, abnormal ovarian structure, PRL, AMH, APTT, ACA and TPO-Ab. And we established a prediction model based on XGBoost algorithm, which can accurately predict the risk of missed abortion in patients with IVF-ET. The performance of this model is better than the traditional logical regression model, and can provide a theoretical basis for the prevention of people at high risk of missed abortion.

Ethics approval and consent to participate

The protocol for this research project has been approved by the Ethics Committee of the Affiliated Women and Children’s Hospital of Qingdao University, with approval number QFELL-YJ-2022-18. Because of the retrospective nature of the study, informed patient consent was waived by the Ethics Committee of the Affiliated Women and Children’s Hospital of Qingdao University. The research conforms to the provisions of the Declaration of Helsinki (as revised in Brazil 2013).

Consent for publication

Not applicable.

Availability of data and materials

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

Competing interests

The authors declare that they have no competing interests.

Funding

This study was supported by the National Nature Fund (Grant No. 81671416).

Author’s contributions

XD and CH made substantial contributions to the conception. GY and BL designed the work. GY, HZ, MZ and YL made contributions to the acquisition of data. GY and BL made contributions to the analysis and interpretation of data. GY drafted the work, XD and CH substantively revised it. All authors read and approved the final manuscript.

Acknowledgement

We sincerely thank the network engineer Xia for technical support and all the patients in this study.

Wu H, Marwah S, Wang P, et al. Misoprostol for medical treatment of missed abortion: a systematic review and network meta-analysis[J]. Scientific reports, 2017,7(1): 1-9.
Segawa T, Kuroda T, Kato K, et al. Cytogenetic analysis of the retained products of conception after missed abortion following blastocyst transfer: a retrospective, large-scale, single-centre study[J]. Reproductive biomedicine online, 2017, 34(2):203-210.
Zhi Z, Yang W, Liu L, Jiang X, Pang L. Early missed abortion is associated with villous angiogenesis via the HIF-1α/VEGF signaling pathway. Arch Gynecol Obstet. 2018 Sep;298(3):537-543.
Fang J, Xie B, Chen B, et al. Biochemical clinical factors associated with missed abortion independent of maternal age: A retrospective study of 795 cases with missed abortion and 694 cases with normal pregnancy[J]. Medicine, 2018, 97(50):e13573.
Zhao Yanhui, Pang Hong,Guo Shuaishuai,et al. Identification of cryptic structural chromosomal aberrations in parents through de tection of copy number variations in miscarriage tissues[J].Chinese Journal of Medical Genetics,2019,36:1 123-1 126.
Zhao Ranran, Wu Yuelian,Zhao Fangfang,et al. The risk of missed abortion associated with the levels of tobacco, heavy metals and phthalate in hair of pregnant woman: A case control study in Chi nese women[J].Medicine(Baltimore),2017,96:9388.
Zhou Qing, Liu Shuaimei,Zhang Ruijin, et al. Research progress on the related factors of embryo damage [J]. China Medical Herald,2019,16(23):53-56.
Guo Liyan, Zhai Min,Wang Shasha,,et al.Meta-analysis of risk factors related to the induction of embryo damage[J].Chinese Journal of Child Health Care2016,24(02):166-169.
Fang Jie, Xie B, Chen B,et al. Biochemical clinical factors associated with missed abortion independent of maternal age: A retrospective study of 795 cases with missed abortion and 694 cases with normal pregnancy[J]. Medicine, 2018, 97(50) : e13573.
Hu L, Du J, Lv H.et al. Influencing factors of pregnancy loss and survival probability of clinical pregnancies conceived through assis ted reproductive technology[J]. Reprod Biol Endocrinol, 2018,16(1):74.
Jie Qiao, Hongzhen Li.Improvement of Successful Rate of ART[J].Journal of International Reproductive Health, 2013,32(1):1-2.
Shiling Chen. Assisted reproductive technology and spontaneous abortion[J]. Journal of Practical Obstetrics and Gynecology, 2016,32(2): 89-91.
Chen T, Guestrin C . XGBoost: A Scalable Tree Boosting System[J]. ACM, 2016:785-794.
ACOG Practice Bulletin No.200 SummaryEarly pregnancy loss[J]. Obstet Gynecol,2018,132(5):1311-1313.
Huchon C, Deffieux X, Beucher G,et al.Pregnancy loss:French clinical practice guidelines[J].Eur J Obstet Gynecol Reprod Biol,2016,201:18-26.
Lapolla A, Dalfrà MG, Ragazzi E, et al. New International Association of the Diabetes and Pregnancy Study Groups (IADPSG) recommendations for diagnosing gestational diabetes compared with former criteria: a retrospective study on pregnancy outcome. Diabet Med. 2011;28:1074–7.
Dong Xuemei,Du Xiaozhong.Research progress of thrombosis tendency and recurrent spontaneous abortion[J].Laboratory Medicine and Clinic,2013,10(2):197-200.
Grigg Jasmin, Roisin Worsley , Caroline Thew, et al. Antipsychotic-induced hyperprolactinemia: synthesis of world-wide guidelines and integrated recommendations for assessment, management and future research.[J]. Psychopharmacology, 2017, 234(22) : 3279-3297.
Bruno Tarasconi, Teddy Tadros, Jean-Marc Ayoubi, et al. Serum antimüllerian hormone levels are independently related to miscarriage rates after in vitro fertilization–embryo transfer[J]. Fertility and Sterility, 2017, 108(3) : 518-524.
Patel BG, Rudnicki M,Yu J, et al.Progesterone resistance in endometriosis:origins, consequences and interventions[Jl.Acta Obst et Gynecol Scand,2017,96(6):623-632.
Hansen JP. Older maternal age and pregnancy outcome[J]. Obstet Gynecol Surv,1986,41 (11): 726－742.
Jie Qiao，Rui Yang.Clinical outcome of assisted reproductive technology in women at advanced maternal age[J].Chinese Journal of Practical Gynecology and Obstetrics，2017.33(1);64-67
Brahem S, Mehdi M, Elghezal H, et al. The effects of male aging on semen quality, sperm DNA fragmentation and chromosomal abnormalities in an infertile population. J Assist Reprod Genet. 2011 May;28(5):425-32.
Cooper TG, Noonan E, von Eckardstein S, Auger J, et al. World Health Organization reference values for human semen characteristics. Hum Reprod Update. 2010 May-Jun;16(3):231-45.
Alexander EK, Pearce EN, Brent GA, et al. 2017 Guidelines of the American Thyroid Association for the Diagnosis and Management of Thyroid Disease During Pregnancy and the Postpartum[J]. Thyroid. 2017 Mar;27(3):315-389.
Ji Wei, Wang Dan, Song Yunxia. Correlation analysis between thyroid dysfunction and embryonic cessation[J]. International Journal of Laboratory Medicine.2019,40(Z2):149-152.
Grossmann M. Measuring thyroid peroxidase antibodies on the day nulliparous women pre sent for management of miscarriage:a descriptive cohort study[J].Reprod Biol Endocrinol,2013,11(1):40.
Yue Qian, Wang Lu, Wang Wen, et al. Risk factors of adverse pregnancy outcome and clinical significance of detection of antinuclear antibody and anticardiolipin antibody in serum[J]. Journal of Guangxi Medical University,2020,37(3):454-458.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Prediction model for missed abortion of patients treated with IVF-ET based on XGBoost: establishment and evaluation

Status:

Version 1

Abstract

Figures

Background

Methods

Results

Discussion

Conclusion

Declarations

References

Additional Declarations

Status:

Version 1