Proposing a machine-learning based method to predict stillbirth before and during delivery and ranking the features: nationwide retrospective cross-sectional study.
Khatibi Toktam,Hanifi Elham,Sepehri Mohammad Mehdi,Allahqoli Leila
BMC pregnancy and childbirth
BACKGROUND:Stillbirth is defined as fetal loss in pregnancy beyond 28 weeks by WHO. In this study, a machine-learning based method is proposed to predict stillbirth from livebirth and discriminate stillbirth before and during delivery and rank the features. METHOD:A two-step stack ensemble classifier is proposed for classifying the instances into stillbirth and livebirth at the first step and then, classifying stillbirth before delivery from stillbirth during the labor at the second step. The proposed SE has two consecutive layers including the same classifiers. The base classifiers in each layer are decision tree, Gradient boosting classifier, logistics regression, random forest and support vector machines which are trained independently and aggregated based on Vote boosting method. Moreover, a new feature ranking method is proposed in this study based on mean decrease accuracy, Gini Index and model coefficients to find high-ranked features. RESULTS:IMAN registry dataset is used in this study considering all births at or beyond 28th gestational week from 2016/04/01 to 2017/01/01 including 1,415,623 live birth and 5502 stillbirth cases. A combination of maternal demographic features, clinical history, fetal properties, delivery descriptors, environmental features, healthcare service provider descriptors and socio-demographic features are considered. The experimental results show that our proposed SE outperforms the compared classifiers with the average accuracy of 90%, sensitivity of 91%, specificity of 88%. The discrimination of the proposed SE is assessed and the average AUC of ±95%, CI of 90.51% ±1.08 and 90% ±1.12 is obtained on training dataset for model development and test dataset for external validation, respectively. The proposed SE is calibrated using isotopic nonparametric calibration method with the score of 0.07. The process is repeated 10,000 times and AUC of SE classifiers using random different training datasets as null distribution. The obtained p-value to assess the specificity of the proposed SE is 0.0126 which shows the significance of the proposed SE. CONCLUSIONS:Gestational age and fetal height are two most important features for discriminating livebirth from stillbirth. Moreover, hospital, province, delivery main cause, perinatal abnormality, miscarriage number and maternal age are the most important features for classifying stillbirth before and during delivery.
Prediction of an ongoing pregnancy after intrauterine insemination.
Steures Pieternel,van der Steeg Jan Willem,Mol Ben W J,Eijkemans Marinus J C,van der Veen Fulco,Habbema J Dik F,Hompes Peter G A,Bossuyt Patrick M M,Verhoeve Harold R,van Kasteren Yvonne M,van Dop Peter A,
Fertility and sterility
OBJECTIVE:To develop a prognostic model for the outcome of IUI. DESIGN:Retrospective cohort study. SETTING:Four fertility centers in The Netherlands. PATIENT(S):Couples of whom the female partners had a regular cycle and who had been treated with IUI. INTERVENTION(S):Intrauterine insemination with and without ovarian hyperstimulation. MAIN OUTCOME MEASURE(S):Ongoing pregnancy. RESULT(S):Overall, 3371 couples were included who underwent 14968 cycles. There were 1229 (8.2%) pregnancies, of which 1000 (6.7%) pregnancies were ongoing. Logistic regression analysis demonstrated that increasing maternal age, longer duration of subfertility, presence of male factor subfertility, one-sided tubal pathology, endometriosis, uterine anomalies, and an increasing number of cycles were unfavorable predictors for an ongoing pregnancy. Cervical factor and the use of ovarian hyperstimulation were favorable predictors. The area under the receiver operating characteristic curve was 0.59. When couples were divided into four categories based on prognosis, the difference between the predicted and observed chance, that is, the calibration, was less than 0.5% in each of the four groups. CONCLUSION(S):Although our model had a relatively poor discriminative capacity, data on calibration showed that the selected prognostic factors allow distinction between couples with a poor prognosis and couples with a good prognosis. After external validation, this model could be of use in patient counseling and clinical decision making.
Machine learning predicts live-birth occurrence before in-vitro fertilization treatment.
Goyal Ashish,Kuchana Maheshwar,Ayyagari Kameswari Prasada Rao
In-vitro fertilization (IVF) is a popular method of resolving complications such as endometriosis, poor egg quality, a genetic disease of mother or father, problems with ovulation, antibody problems that harm sperm or eggs, the inability of sperm to penetrate or survive in the cervical mucus and low sperm counts, resulting human infertility. Nevertheless, IVF does not guarantee success in the fertilization. Choosing IVF is burdensome for the reason of high cost and uncertainty in the result. As the complications and fertilization factors are numerous in the IVF process, it is a cumbersome task for fertility doctors to give an accurate prediction of a successful birth. Artificial Intelligence (AI) has been employed in this study for predicting the live-birth occurrence. This work mainly focuses on making predictions of live-birth occurrence when an embryo forms from a couple and not a donor. Here, we compare various AI algorithms, including both classical Machine Learning, deep learning architecture, and an ensemble of algorithms on the publicly available dataset provided by Human Fertilisation and Embryology Authority (HFEA). Insights on data and metrics such as confusion matrices, F1-score, precision, recall, receiver operating characteristic (ROC) curves are demonstrated in the subsequent sections. The training process has two settings Without feature selection and With feature selection to train classifier models. Machine Learning, Deep learning, ensemble models classification paradigms have been trained in both settings. The Random Forest model achieves the highest F1-score of 76.49% in without feature selection setting. For the same model, the precision, recall, and area under the ROC Curve (ROC AUC) scores are 77%, 76%, and 84.60%, respectively. The success of the pregnancy depends on both male and female traits and living conditions. This study predicts a successful pregnancy through the clinically relevant parameters in In-vitro fertilization. Thus artificial intelligence plays a promising role in decision making process to support the diagnosis, prognosis, treatment etc.
Individualized decision-making in IVF: calculating the chances of pregnancy.
van Loendersloot L L,van Wely M,Repping S,Bossuyt P M M,van der Veen F
Human reproduction (Oxford, England)
STUDY QUESTION:Are we able to develop a model to calculate the chances of pregnancy prior to the start of the first IVF cycle as well as after one or more failed cycles? SUMMARY ANSWER:Our prediction model enables the accurate individualized calculation of the probability of an ongoing pregnancy with IVF. WHAT IS KNOWN ALREADY:To improve counselling, patient selection and clinical decision-making in IVF, a number of prediction models have been developed. These models are of limited use as they were developed before current clinical and laboratory protocols were established. STUDY DESIGN, SIZE, DURATION:This was a cohort study. The development set included 2621 cycles in 1326 couples who had been treated with IVF or ICSI between January 2001 and July 2009. The validation set included additional data from 515 cycles in 440 couples treated between August 2009 and April 2011. The outcome of interest was an ongoing pregnancy after transfer of fresh or frozen-thawed embryos from the same stimulated IVF cycle. If a couple became pregnant after an IVF/ICSI cycle, the follow-up was at a gestational age of at least 11 weeks. PARTICIPANTS/MATERIALS, SETTING, METHODS:Women treated with IVF or ICSI between January 2001 and April 2011 in a university hospital. IVF/ICSI cycles were excluded in the case of oocyte or embryo donation, surgically retrieved spermatozoa, patients positive for human immunodeficiency virus, modified natural IVF and cycles cancelled owing to poor ovarian stimulation, ovarian hyperstimulation syndrome or other unexpected medical or non-medical reasons. MAIN RESULTS AND THE ROLE OF CHANCE:Thirteen variables were included in the final prediction model. For all cycles, these were female age, duration of subfertility, previous ongoing pregnancy, male subfertility, diminished ovarian reserve, endometriosis, basal FSH and number of failed IVF cycles. After the first cycle: fertilization, number of embryos, mean morphological score per Day 3 embryo, presence of 8-cell embryos on Day 3 and presence of morulae on Day 3 were also included. In validation, the model had moderate discriminative capacity (c-statistic 0.68, 95% confidence interval: 0.63-0.73) but calibrated well, with a range from 0.01 to 0.56 in calculated probabilities. LIMITATIONS, REASONS FOR CAUTION:In our study, the outcome of interest was ongoing pregnancy. Live birth may have been a more appropriate outcome, although only 1-2% of all ongoing pregnancies result in late miscarriage or stillbirth. The model was based on data from a single centre. WIDER IMPLICATIONS OF THE FINDINGS:The IVF model presented here is the first to calculate the chances of an ongoing pregnancy with IVF, both for the first cycle and after any number of failed cycles. The generalizability of the model to other clinics has to be evaluated more extensively in future studies (geographical validation). Centres with higher or lower success rates could use the model, after recalibration, by adjusting the intercept to reflect the IVF success rates in their centre. STUDY FUNDING/COMPETING INTEREST(S):This project was funded by the NutsOhra foundation (Grant 1004-179). The NutsOhra foundation had no role in the development of our study, in the collection, analysis and interpretation of data; in writing of the manuscript, and in the decision to submit the manuscript for publication. There were no competing interests.
Preoperative serum anti-Müllerian hormone level is a potential predictor of ovarian endometrioma severity and postoperative fertility.
Dong Zhenzhu,An Jian,Xie Xi,Wang Zhenhong,Sun Pengming
European journal of obstetrics, gynecology, and reproductive biology
OBJECTIVE:To establish a model for predicting revised American Society of Reproductive Medicine (rASRM) scores before endometrioma surgery based on serum anti-Müllerian hormone (AMH) level and to identify factors that might reliably predict postoperative fertility of women diagnosed with endometrioma. STUDY DESIGN:The study population was composed of 134 women with endometrioma, 58 with benign cyst, and 115 with non-ovarian lesion. Preoperative serum AMH level and clinical parameters were compared among three groups. Univariate correlation analyses and multivariate linear regression modeling with a stepwise method were performed for constructing an rASRM scores prediction model. Cox regression analysis was then used to identify predictive variables of spontaneous pregnancy following surgical treatment of endometrioma. RESULTS:Preoperative AMH level were significantly lower in the endometrioma group than in the other two groups (p < 0.001). Multivariate linear regression analysis revealed that age (β=-0.324, p < 0.001), rASRM scores (β=-0.298, p < 0.001) and serum CA125 level (β=-0.176, p = 0.026) independently and negatively correlated with serum AMH level. Cox regression analysis of women with endometrioma who underwent surgical resection indicated that older age (per five-year increase, HR: 0.517; 95% CI, 0.299-0.896) and higher serum AMH level (cut-off value: >3.68 ng/ml, HR: 2.383; 95% CI, 1.093-5.197) were independent predictors for postoperative fertility. CONCLUSION:Patients with advanced staged endometriosis tended to have a lower serum AMH level while postoperative infertility was more likely to occur in older patients with a lower level of serum AMH. Thus, timely detection of AMH levels to assess the severity of ovarian endometriosis and possibility for postoperative pregnancy success is necessary to ensure that optimal medical treatment can be provided.
External validation of a dynamic prediction model for repeated predictions of natural conception over time.
van Eekelen R,McLernon D J,van Wely M,Eijkemans M J,Bhattacharya S,van der Veen F,van Geloven N
Human reproduction (Oxford, England)
STUDY QUESTION:How well does a previously developed dynamic prediction model perform in an external, geographical validation in terms of predicting the chances of natural conception at various points in time? SUMMARY ANSWER:The dynamic prediction model performs well in an external validation on a Scottish cohort. WHAT IS KNOWN ALREADY:Prediction models provide information that can aid evidence-based management of unexplained subfertile couples. We developed a dynamic prediction model for natural conception (van Eekelen model) that is able to update predictions of natural conception when couples return to their clinician after a period of unsuccessful expectant management. It is not known how well this model performs in an external population. STUDY DESIGN, SIZE, DURATION:A record-linked registry study including the long-term follow-up of all couples who were considered unexplained subfertile following a fertility workup at a Scottish fertility clinic between 1998 and 2011. Couples with anovulation, uni/bilateral tubal occlusion, mild/severe endometriosis or impaired semen quality according to World Health Organization criteria were excluded. PARTICIPANTS/MATERIALS, SETTING, METHODS:The endpoint was time to natural conception, leading to an ongoing pregnancy (defined as reaching a gestational age of at least 12 weeks). Follow-up was censored at the start of treatment, at the change of partner or at the end of study (31 March 2012). The performance of the van Eekelen model was evaluated in terms of calibration and discrimination at various points in time. Additionally, we assessed the clinical utility of the model in terms of the range of the calculated predictions. MAIN RESULTS AND THE ROLE OF CHANCE:Of a total of 1203 couples with a median follow-up of 1 year and 3 months after the fertility workup, 398 (33%) couples conceived naturally leading to an ongoing pregnancy. Using the dynamic prediction model, the mean probability of natural conception over the course of the first year after the fertility workup was estimated at 25% (observed: 23%). After 0.5, 1 and 1.5 years of expectant management after the completion of the fertility workup, the average probability of conceiving naturally over the next year was estimated at 18% (observed: 15%), 14% (observed: 14%) and 12% (observed: 12%). Calibration plots showed good agreement between predicted chances and the observed fraction of ongoing pregnancy within risk groups. Discrimination was moderate with c statistics similar to those in the internal validation, ranging from 0.60 to 0.64. The range of predicted chances was sufficiently wide to distinguish between couples having a good and poor prognosis with a minimum of zero at all times and a maximum of 55% over the first year after the workup, which decreased to maxima of 43% after 0.5 years, 34% after 1 year and 29% after 1.5 years after the fertility workup. LIMITATIONS, REASONS FOR CAUTION:The model slightly overestimated the chances of conception by ~2-3% points on group level in the first-year post-fertility workup and after 0.5 years of expectant management, respectively. This is likely attributable to the fact that the exact dates of completion of the fertility workup for couples were missing and had to be estimated. WIDER IMPLICATIONS OF THE FINDINGS:The van Eekelen model is a valid and robust tool that is ready to use in clinical practice to counsel couples with unexplained subfertility on their individualized chances of natural conception at various points in time, notably when couples return to the clinic after a period of unsuccessful expectant management. STUDY FUNDING/COMPETING INTEREST(S):This work was supported by a Chief Scientist Office postdoctoral training fellowship in health services research and health of the public research (ref PDF/12/06). There are no conflicts of interest.
Evaluation of Risk Factors Associated with Endometriosis in Infertile Women.
Ashrafi Mahnaz,Sadatmahalleh Shahideh Jahanian,Akhoond Mohammad Reza,Talebi Mehrak
International journal of fertility & sterility
BACKGROUND:Endometriosis affects women's physical and mental wellbeing. Symptoms include dyspareunia, dysmenorrhea, pelvic pain, and infertility. The purpose of this study is to assess the correlation between some relevant factors and symptoms and risk of an endometriosis diagnosis in infertile women. MATERIALS AND METHODS:A retrospective study of 1282 surgical patients in an infertility Institute, Iran between 2011 and 2013 were evaluated by laparoscopy. Of these, there were 341 infertile women with endometriosis (cases) and 332 infertile women with a normal pelvis (comparison group). Chi-square and t tests were used to compare these two groups. Logistic regression was done to build a prediction model for an endometriosis diagnosis. RESULTS:Gravidity [odds ratio (OR): 0.8, confidence interval (CI): 0.6-0.9, P=0.01], parity (OR: 0.7, CI: 0.6-0.9, P=0.01), family history of endometriosis (OR: 4.9, CI: 2.1-11.3, P<0.001), history of galactorrhea (OR: 2.3, CI: 1.5-3.5, P=0.01), history of pelvic surgery (OR: 1.9, CI: 1.3-2.7, P<0.001), and shorter menstrual cycle length (OR: 0.9, CI: 0.9-0.9, P=0.04) were associated with endometriosis. Duration of natural menstruation and age of menarche were not correlated with subsequent risk of endometriosis (P>0.05). Fatigue, diarrhea, constipation, dysmenorrhea, dyspareunia, pelvic pain and premenstrual spotting were more significant among late-stage endometriosis patients than in those with early-stage endometriosis and more prevalent among patients with endometriosis than that of the comparison group. In the logistic regression model, gravidity, family history of endometriosis, history of galactorrhea, history of pelvic surgery, dysmenorrhoea, pelvic pain, dysparaunia, premenstrual spotting, fatigue, and diarrhea were significantly associated with endometriosis. However, the number of pregnancies was negatively related to endometriosis. CONCLUSION:Endometriosis is a considerable public health issue because it affects many women and is associated with the significant morbidity. In this study, we built a prediction model which can be used to predict the risk of endometriosis in infertile women.
Use of In Vitro Fertilisation Prediction Model in an Asian Population-Experience in Singapore.
Saha Laxmi,Fook-Chong Stephanie Mc,Rajesh Hemashree,Chia Diana Sf,Yu Su Ling
Annals of the Academy of Medicine, Singapore
INTRODUCTION:This retrospective study was conducted to perform an external validation of the in vitro fertilisation (IVF) predict model developed by Scott Nelson et al in an Asian population. MATERIALS AND METHODS:All IVF cycles registered in the study centre from January 2005 to December 2010 were included. Observed and predicted values of at least 1 live birth per cycle were compared by discrimination, calibration. Hosmer-Lemeshow test was used to assess the goodness-of-fit of the model calibration and Brier score was used to assess overall model performance. RESULTS:Among 634 IVF cycles, rate of at least 1 live birth was 30.6%. Causes of infertility were unexplained in 35.5% cases. Fifty-seven percent of women came for their first IVF treatment. First IVF cycle showed significantly higher success in comparison to subsequent cycles. The odds ratio of successful live birth was worse in women with endometriosis. Observed outcome was found to be more than the prediction of the model. The area under the curve (AUC) in this study was found to be 0.65 that was close to that of Nelson model (0.6335) done in internal validation. Brier score (average prediction error) of model was 0.2. Chi square goodness-of-fit test indicated that there was difference between the predicted and observed value (x² =18.28, df = 8, P = 0.019). Overall statistical findings indicated that the accuracy of the prediction model fitted poorly with the study population. CONCLUSION:Ovarian reserve, treatment centre and racial effect on predictability cannot be excluded. So it is important to make a good prediction model by considering the additional factors before using the model widely.
Revisiting the Risk Factors for Endometriosis: A Machine Learning Approach.
Journal of personalized medicine
Endometriosis is a condition characterized by implants of endometrial tissues into extrauterine sites, mostly within the pelvic peritoneum. The prevalence of endometriosis is under-diagnosed and is estimated to account for 5-10% of all women of reproductive age. The goal of this study was to develop a model for endometriosis based on the UK-biobank (UKB) and re-assess the contribution of known risk factors to endometriosis. We partitioned the data into those diagnosed with endometriosis (5924; ICD-10: N80) and a control group (142,723). We included over 1000 variables from the UKB covering personal information about female health, lifestyle, self-reported data, genetic variants, and medical history prior to endometriosis diagnosis. We applied machine learning algorithms to train an endometriosis prediction model. The optimal prediction was achieved with the gradient boosting algorithms of CatBoost for the data-combined model with an area under the ROC curve (ROC-AUC) of 0.81. The same results were obtained for women from a mixed ethnicity population of the UKB (7112; ICD-10: N80). We discovered that, prior to being diagnosed with endometriosis, affected women had significantly more ICD-10 diagnoses than the average unaffected woman. We used SHAP, an explainable AI tool, to estimate the marginal impact of a feature, given all other features. The informative features ranked by SHAP values included irritable bowel syndrome (IBS) and the length of the menstrual cycle. We conclude that the rich population-based retrospective data from the UKB are valuable for developing unified machine learning endometriosis models despite the limitations of missing data, noisy medical input, and participant age. The informative features of the model may improve clinical utility for endometriosis diagnosis.
Technical Verification and Assessment of Independent Validation of Biomarker Models for Endometriosis.
O Dorien F,Fassbender Amelie,Van Bree Rita,Laenen Annouschka,Peterse Daniëlle P,Vanhie Arne,Waelkens Etienne,D'Hooghe Thomas M
BioMed research international
There is a great need for a noninvasive diagnosis for endometriosis. Several biomarkers and biomarker panels have been proposed. Biomarker models consisting of CA-125, VEGF, Annexin V, and glycodelin/sICAM-1 were previously developed by our group. The objective of our current study was to assess the impact of technical and biological variability on the performance of those previously developed prediction models in a technical verification and a validation setting. The technical verification cohort consisted of peripheral blood plasma samples from a subset of the patients included in the original study of Vodolazkaia (99 women with and 37 women without endometriosis). The validation study was done in plasma samples of an independent patient cohort (170 women with and 86 women without endometriosis). Single immunoassays were used for CA-125, VEGF-A, sICAM-1, Annexin V, and glycodelin. Statistical analyses were done using univariate and multivariate (logistic regression) approaches. The previously reported prediction models for endometriosis had a low performance in both the technical verification and validation setting. New prediction models were developed, which included CA-125, Annexin V, and sICAM-1, but CA-125 was the only marker that was retained in the models across the technical verification and validation study. Overall, successful validation of a biomarker model depends on several factors such as patient selection, collection methods, assay selection/handling, stability of the marker, and statistical analysis and interpretation. There is a need for standardized studies in large, well-defined patient cohorts with robust assay methodologies.
Negative Sliding Sign during Dynamic Ultrasonography Predicts Low Endometriosis Fertility Index at Laparoscopy.
Alfaraj Sukainah,Noga Heather,Allaire Catherine,Williams Christina,Lisonkova Sarka,Yong Paul J,Bedaiwy Mohamed A
Journal of minimally invasive gynecology
STUDY OBJECTIVE:Endometriosis fertility index (EFI) is a robust tool to predict the pregnancy rate in patients with endometriosis who are attempting non-in vitro fertilization conception. However, EFI calculation requires laparoscopy. Newly established imaging techniques such as sliding sign, which is used to diagnose pouch of Douglas obliteration, could provide a promising alternative. The objective of this study was to investigate the practicality of using ultrasound data to predict a low EFI (score ≤6). DESIGN:Observational study from a prospective registry (Endometriosis Pelvic Pain Interdisciplinary Cohort, clinicaltrials.gov #NCT02911090). Analyzed data were captured from December 2013 to June 2017. SETTING:Tertiary referral center at British Columbia Women's Hospital. PATIENTS:We analyzed data for 2583 participants from the Endometriosis Pelvic Pain Interdisciplinary Cohort. In this cross-sectional study, we included 86 women aged <40 years. INTERVENTIONS:Dynamic ultrasonography for the sliding sign testing and EFI calculation during laparoscopic surgery. MEASUREMENTS AND MAIN RESULTS:Logistic regression was used to obtain receiver operating characteristic area under the curve (AUC) for the prediction models. Significance was p <.05. Patients with a negative sliding sign were older and had severe endometriosis and longer duration of infertility. Patients with a negative sliding sign had significantly lower total EFI scores and lower surgical factors scores than patients with a positive sliding sign. Logistic regression showed that a negative sliding sign and EFI historic factors score can predict an EFI score ≤6 (sensitivity = 87.9%, specificity = 81.1%, AUC = 0.93 [95% confidence interval, 0.88-0.98]). Adding the diagnosis of endometrioma to the previous prediction model resulted in AUC = 0.95 (95% confidence interval, 0.90-0.995), sensitivity = 84.8%, and specificity = 92.5%. CONCLUSION:The sliding sign could be a potential alternative to the EFI surgical factors, and it could be used in combination with EFI historic factors and the diagnosis of endometrioma to predict an EFI score ≤6 for patients who are not scheduled for immediate surgery.
Markers of deep infiltrating endometriosis in patients with ovarian endometrioma: a predictive model.
Perelló Maria,Martínez-Zamora Maria A,Torres Ximena,Munrós Jordina,Llecha Silvia,De Lazzari Elisa,Balasch Juan,Carmona Francisco
European journal of obstetrics, gynecology, and reproductive biology
OBJECTIVE:The purpose of the study was to develop an easily applicable predictive model to predict deep infiltrating endometriosis in patients with ovarian endometrioma. STUDY DESIGN:We performed a retrospective analysis of 178 consecutive women with ovarian endometrioma who underwent surgery, with histological confirmation and complete removal of endometriosis in the Hospital Clinic of Barcelona. Several markers were prospectively obtained and compared between the group of patients presenting deep infiltrating endometriosis associated with ovarian endometrioma and women with only ovarian endometrioma. Multiple logistic regression analysis was performed to create a model to predict the presence of deep infiltrating endometriosis and internal validation was later performed. RESULTS:Of the 178 patients studied, 80 (45%) were classified in the ovarian endometrioma group and 98 (55%) in the group of patients presenting deep infiltrating endometriosis associated with ovarian endometrioma. The independent variables to predict deep infiltrating endometriosis were: at least one previous pregnancy, a past history of surgery for endometriosis and the mean endometriosis-associated pelvic pain score. The area under the ROC curve was 0.91 (95% confidence interval: 0.86-0.94), with an optimal cut-off of the predicted probability of 0.54. The sensitivity of the model was 80% and the specificity 84%. CONCLUSIONS:This model predicts the development of deep infiltrating endometriosis in patients with ovarian endometriomas allowing prioritization of women for referral to specialized centers.
Development of a prediction model to aid primary care physicians in early identification of women at high risk of developing endometriosis: cross-sectional study.
Verket Nina Julie,Falk Ragnhild Sørum,Qvigstad Erik,Tanbo Tom Gunnar,Sandvik Leiv
OBJECTIVES:To identify predictors of disease among a few factors commonly associated with endometriosis and if successful, to combine these to develop a prediction model to aid primary care physicians in early identification of women at high risk of developing endometriosis. DESIGN:Cross-sectional anonymous postal questionnaire study. SETTING:Women aged 18-45 years recruited from the Norwegian Endometriosis Association and a random sample of women residing in Oslo, Norway. PARTICIPANTS:157 women with and 156 women without endometriosis. MAIN OUTCOME MEASURES:Logistic and least absolute shrinkage and selection operator (LASSO) regression analyses were performed with endometriosis as dependent variable. Predictors were identified and combined to develop a prediction model. The predictive ability of the model was evaluated by calculating the area under the receiver operating characteristic curve (AUC) and positive predictive values (PPVs) and negative predictive values (NPVs). To take into account the likelihood of skewed representativeness of the patient sample towards high symptom burden, we considered the hypothetical prevalences of endometriosis in the general population 0.1%, 0.5%, 1% and 2%. RESULTS:The predictors and demonstrated the strongest association with disease. The model based on logistic regression (AUC 0.83) included these two predictors only, while the model based on LASSO regression (AUC 0.85) included two more: and . For the prevalences 0.1%, 0.5%, 1% and 2%, both models ascertained endometriosis with PPV equal to 2.0%, 9.4%, 17.2% and 29.6%, respectively. NPV was at least 98% for all values considered. CONCLUSIONS:External validation is needed before model implementation. Meanwhile, endometriosis should be considered a differential diagnosis in women with frequent absenteeism from school or work due to painful menstruations and positive family history of endometriosis.
Predicting the likelihood of successful medical treatment of early pregnancy loss: development and internal validation of a clinical prediction model.
Human reproduction (Oxford, England)
STUDY QUESTION:What are clinical predictors for successful medical treatment in case of early pregnancy loss (EPL)? SUMMARY ANSWER:Use of mifepristone, BMI, number of previous uterine aspirations and the presence of minor clinical symptoms (slight vaginal bleeding or some abdominal cramps) at treatment start are predictors for successful medical treatment in case of EPL. WHAT IS KNOWN ALREADY:Success rates of medical treatment for EPL vary strongly, between but also within different treatment regimens. Up until now, although some predictors have been identified, no clinical prediction model has been developed yet. STUDY DESIGN, SIZE, DURATION:Secondary analysis of a multicentre randomized controlled trial in 17 Dutch hospitals, executed between 28 June 2018 and 8 January 2020. PARTICIPANTS/MATERIALS, SETTING, METHODS:Women with a non-viable pregnancy between 6 and 14 weeks of gestational age, who opted for medical treatment after a minimum of 1 week of unsuccessful expectant management. Potential predictors for successful medical treatment of EPL were chosen based on literature and expert opinions. We internally validated the prediction model using bootstrapping techniques. MAIN RESULTS AND THE ROLE OF CHANCE:237 out of 344 women had a successful medical EPL treatment (68.9%). The model includes the following variables: use of mifepristone, BMI, number of previous uterine aspirations and the presence of minor clinical symptoms (slight vaginal bleeding or some abdominal cramps) at treatment start. The model shows a moderate capacity to discriminate between success and failure of treatment, with an AUC of 67.6% (95% CI = 64.9-70.3%). The model had a good fit comparing predicted to observed probabilities of success but might underestimate treatment success in women with a predicted probability of success of ∼70%. LIMITATIONS, REASONS FOR CAUTION:The vast majority (90.4%) of women were Caucasian, potentially leading to less optimal model performance in a non-Caucasian population. Limitations of our model are that we have not yet been able to externally validate its performance and clinical impact, and the moderate accuracy of the prediction model of 0.67. WIDER IMPLICATIONS OF THE FINDINGS:We developed a prediction model, aimed to improve and personalize counselling for medical treatment of EPL by providing a woman with her individual chance of complete evacuation. STUDY FUNDING/COMPETING INTEREST(S):The Triple M Trial, upon which this secondary analysis was performed, was funded by the Healthcare Insurers Innovation Foundation (project number 3080 B15-191). TRIAL REGISTRATION NUMBER:Clinicaltrials.gov: NCT03212352.
Early prediction of live birth for assisted reproductive technology patients: a convenient and practical prediction model.
Gao Hong,Liu Dong-E,Li Yumei,Wu Xinrui,Tan Hongzhuan
Live birth is the most important concern for assisted reproductive technology (ART) patients. Therefore, in the medical reproductive centre, obstetricians often need to answer the following question: "What are the chances that I will have a healthy baby after ART treatment?" To date, our obstetricians have no reference on which to base the answer to this question. Our research aimed to solve this problem by establishing prediction models of live birth for ART patients. Between January 1, 2010, and May 1, 2017, we conducted a retrospective cohort study of women undergoing ART treatment at the Reproductive Medicine Centre, Xiangya Hospital of Central South University, Hunan, China. The birth of at least one live-born baby per initiated cycle or embryo transfer procedure was defined as a live birth, and all other pregnancy outcomes were classified as no live birth. A live birth prediction model was established by stepwise multivariate logistic regression. All eligible subjects were randomly allocated to two groups: group 1 (80% of subjects) for the establishment of the prediction models and group 2 (20% of subjects) for the validation of the established prediction models. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of each prediction model at different cut-off values were calculated. The prediction model of live birth included nine variables. The area under the ROC curve was 0.743 in the validation group. The sensitivity, specificity, PPV, and NPV of the established model ranged from 97.9-24.8%, 7.2-96.3%, 44.8-83.8% and 81.7-62.5%, respectively, at different cut-off values. A stable, reliable, convenient, and satisfactory prediction model for live birth by ART patients was established and validated, and this model could be a useful tool for obstetricians to predict the live rate of ART patients. Meanwhile, it is also a reference for obstetricians to create good conditions for infertility patients in preparation for pregnancy.
Development and Validation of Prediction Model for High Ovarian Response in Fertilization-Embryo Transfer: A Longitudinal Study.
Computational and mathematical methods in medicine
OBJECTIVE:To develop and validate a prediction model for high ovarian response in fertilization-embryo transfer (IVF-ET) cycles. METHODS:Totally, 480 eligible outpatients with infertility who underwent IVF-ET were selected and randomly divided into the training set for developing the prediction model and the testing set for validating the model. Univariate and multivariate logistic regressions were carried out to explore the predictive factors of high ovarian response, and then, the prediction model was constructed. Nomogram was plotted for visualizing the model. Area under the receiver-operating characteristic (ROC) curve, Hosmer-Lemeshow test and calibration curve were used to evaluate the performance of the prediction model. RESULTS:Antral follicle count (AFC), anti-Müllerian hormone (AMH) at menstrual cycle day 3 (MC3), and progesterone (P) level on human chorionic gonadotropin (HCG) day were identified as the independent predictors of high ovarian response. The value of area under the curve (AUC) for our multivariate model reached 0.958 (95% CI: 0.936-0.981) with the sensitivity of 0.916 (95% CI: 0.863-0.953) and the specificity of 0.911 (95% CI: 0.858-0.949), suggesting the good discrimination of the prediction model. The Hosmer-Lemeshow test and the calibration curve both suggested model's good calibration. CONCLUSION:The developed prediction model had good discrimination and accuracy via internal validation, which could help clinicians efficiently identify patients with high ovarian response, thereby improving the pregnancy rates and clinical outcomes in IVF-ET cycles. However, the conclusion needs to be confirmed by more related studies.
Development and validation of a pregnancy prediction model based on ultrasonographic features related to endometrial receptivity.
Shui Xujuan,Yu Caicha,Li Jianxin,Jiao Yan
American journal of translational research
OBJECTIVE:Our aim was to identify multiple endometrial receptivity related factors by applying non-invasive, repeatable multimodal ultrasound methods. We further established a practical prediction model for pregnancy prediction. MATERIALS AND METHODS:Our study included 152 participants from Wenzhou People's Hospital, Wenzhou Maternal and Child Health Care Hospital, and the Third Affiliated Hospital of Wenzhou Medical University. Clinical information including age and ultrasonographic data were collected. By applying t-test and Wilcoxon rank sum tests, we obtained endometrial receptivity related factors, and by using logistic regression, we established a prediction model for possibility of successful pregnancy. RESULTS:Among all the factors associated with endometrial receptivity, uterine peristaltic wave frequency, uterine spiral artery resistant index, endometrial flow index, ultrasound elastography strain radio (SR), and age showed significant statistical difference between nonpregnant and pregnant volunteers. Consequently, we developed and validated a nomogram prediction model with its value of area under the receiver operating curve up to 0.949 for predicting pregnancy by using age and ultrasonographic factors including uterine peristalsis, uterine spiral artery, and ultrasound elastographic features. The sensitivity was 0.83 and specificity was 0.96. In addition, its performance was better than that of a direct scoring system. CONCLUSION:By employing the pregnancy prediction model with endometrial receptivity associated ultrasonographic factors, clinicians can give a quantitative evaluation and a real time screen of the uterus condition as well as optimal guiding, treatment, and management recommendations for infertility-related patients.
Determination of Cut Off for Endometrial Thickness in Couples with Unexplained Infertility: Trustable AI.
Studies in health technology and informatics
Endometrial thickness in assisted reproductive techniques is one of the essential factors in the success of pregnancy. Despite extensive studies on endometrial thickness prediction, research is still needed. We aimed to analyze the impact of endometrial thickness on the ongoing pregnancy rate in couples with unexplained infertility. A total of 729 couples with unexplained infertility were included in this study. A random forest model (RFM) and logistic regression (LRM) were used to predict pregnancy. Evaluation of the performance of RFM and LRM was based on classification criteria and ROC curve, Odd Ratio for ongoing Pregnancy by EMT categorized. The results showed that RFM outperformed the LRM in IVF/ICSI and IUI treatments, obtaining the highest accuracy. We obtained a 7.7mm cut-off point for IUI and 9.99 mm for IVF/ICSI treatment. The results showed machine learning is a valuable tool in predicting ongoing pregnancy and is trustable via multicenter data for two treatments. In addition, Endometrial thickness was not statistically significantly different from CPR and FHR in both treatments.
Gradient Boosting Machine Learning Model for Defective Endometrial Receptivity Prediction by Macrophage-Endometrium Interaction Modules.
Frontiers in immunology
Background:A receptive endometrium is a prerequisite for successful embryo implantation. Mounting evidence shows that nearly one-third of infertility and implantation failures are caused by defective endometrial receptivity. This study pooled 218 subjects from multiple datasets to investigate the association of the immune infiltration level with reproductive outcome. Additionally, macrophage-endometrium interaction modules were constructed to explore an accurate and cost-effective approach to endometrial receptivity assessment. Methods:Immune-infiltration levels in 4 GEO datasets (n=218) were analyzed and validated through meta-analysis. Macrophage-endometrium interaction modules were selected based on the weighted gene co-expression network in GSE58144 and differentially expressed genes dominated by GSE19834 dataset. Xgboost, random forests, and regression algorithms were applied to predictive models. Subsequently, the efficacy of the models was compared and validated in the GSE165004 dataset. Forty clinical samples (RT-PCR and western blot) were performed for expression and model validation, and the results were compared to those of endometrial thickness in clinical pregnancy assessment. Results:Altered levels of Mϕs infiltration were shown to critically influence embryo implantation. The three selected modules, manifested as macrophage-endometrium interactions, were enrichment in the immunoreactivity, decidualization, and signaling functions and pathways. Moreover, hub genes within the modules exerted significant reproductive prognostic effects. The xgboost algorithm showed the best performance among the machine learning models, with AUCs of 0.998 (95% CI 0.994-1) and 0.993 (95% CI 0.979-1) in GSE58144 and GSE165004 datasets, respectively. These results were significantly superior to those of the other two models (random forest and regression). Similarly, the model was significantly superior to ultrasonography (endometrial thickness) with a better cost-benefit ratio in the population. Conclusion:Successful embryo implantation is associated with infiltration levels of Mϕs, manifested in genetic modules involved in macrophage-endometrium interactions. Therefore, utilizing the hub genes in these modules can provide a platform for establishing excellent machine learning models to predict reproductive outcomes in patients with defective endometrial receptivity.
Internal validation and comparison of predictive models to determine success rate of infertility treatments: a retrospective study of 2485 cycles.
Infertility is a significant health problem and assisted reproductive technologies to treat infertility. Despite all efforts, the success rate of these methods is still low. Also, each of these methods has side effects and costs. Therefore, accurate prediction of treatment success rate is a clinical challenge. This retrospective study aimed to internally validate and compare various machine learning models for predicting the clinical pregnancy rate (CPR) of infertility treatment. For this purpose, data from 1931 patients consisting of in vitro fertilization (IVF) or intra cytoplasmic sperm injection (ICSI) (733) and intra uterine insemination (IUI) (1196) treatments were included. Also, no egg or sperm donation data were used. The performance of machine learning algorithms to predict clinical pregnancy were expressed in terms of accuracy, recall, F-score, positive predictive value (PPV), brier score (BS), Matthew correlation coefficient (MCC), and receiver operating characteristic. The significance of the features with CPR and AUCs was evaluated by Student's t test and DeLong's algorithm. Random forest (RF) model had the highest accuracy in the IVF/ICSI treatment. The sensitivity, F1 score, PPV, and MCC of the RF model were 0.76, 0.73, 0.80, and 0.5, respectively. These values for IUI treatment were 0.84, 0.80, 0.82, and 0.34, respectively. The BS was 0.13 and 0.15 for IVF/ICS and IUI, respectively. In addition, the estimated AUCs of the RF model for IVF/ICS and IUI were 0.73 and 0.7, respectively. Some essential features were obtained based on RF ranking for the two datasets, including age, follicle stimulation hormone, endometrial thickness, and infertility duration. The results showed a strong relationship between clinical pregnancy and a woman's age. Also, endometrial thickness and the number of follicles decreased with increasing female age in both treatments.