Human embryo stage classification using an enhanced R(2 + 1)D model and dynamic programming with optimized datasets

Biomedical Signal Processing and Control, Journal Year: 2025, Volume and Issue: 107, P. 107841 - 107841

Published: March 28, 2025

Language: Английский

---

Predictive potential of combined secretomics and image-based morphometry as a non-invasive method for selecting implanting embryos

Reproductive Biology and Endocrinology, Journal Year: 2025, Volume and Issue: 23(1)

Published: April 12, 2025

Non-invasive selection of human embryos for in vitro fertilization purposes is still a major challenge to pursue. Therefore, this study aims identify non-invasive morphometric and secretomic parameters that reliably select the with highest likelihood implantation prior embryo transfer (ET). Prospective single-centre cohort study. Thirty-two day 5 blastocysts derived from 28 couples undergoing intracytoplasmic sperm injection (ICSI) ET between January 2023 April 2023. Patients were split according their outcome, confirmed serum beta-human chorionic gonadotropin (b-hCG) levels > mIU/mL nine days post-SET. Ninety-two proteins involved embryonic developmental programming measured spent blastocyst media (SBM) using protein extension assay. Sparse PLS-DA (sPLS-DA) was used principal component analysis. Forty-seven related trophoblast, inner cell mass blastocele dimension evaluated microphotographs ImageJ software. T-test Mann-Whitney tests respectively compare measurements normalized expression secreted (NPx) implanted or not. Predictive value models based on proteins. Chi-squared showed no significant differences transferred stage, quality, state subgroups. Implanting (n = 14) presented significantly different shape descriptors (i.e., internal circularity, roundness, axis ratio, angle trophoblast mean width) than non-implanting 13). Among quantifiable (86/92) SBM eleven implanting blastocysts, NPx sPLS-DA analysis revealed three differentially expressed Matrilin-2 (MATN2) legumain (LGMN) elevated (p < 0.01 both cases) while thymosin beta-10 (TMSB10) decreased 0.05) embryos. exclusively profiles accurately discriminated outcomes (AUC 0.71). The model integrating blastocysts' ratio MATN2 TMSB10 had exceptional negative positive predictive power (100% 90.91%, respectively; AUC 0.93). Morphometric emerge as promising candidate markers selection.

Language: Английский

---

Ensemble learning for fetal ultrasound and maternal–fetal data to predict mode of delivery after labor induction

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: July 3, 2024

Abstract Providing adequate counseling on mode of delivery after induction labor (IOL) is utmost importance. Various AI algorithms have been developed for this purpose, but rely maternal–fetal data, not including ultrasound (US) imaging. We used retrospectively collected clinical data from 808 subjects submitted to IOL, totaling 2024 US images, train models predict vaginal (VD) and cesarean section (CS) outcomes IOL. The best overall model only (F1-score: 0.736; positive predictive value (PPV): 0.734). imaging employed fetal head, abdomen femur showing limited discriminative results. images 0.594; PPV: 0.580). Consequently, we constructed ensemble test whether could enhance the model. included 0.689; 0.693), presenting a false negative interesting trade-off. accurately predicted CS 4 additional cases, despite misclassifying 20 VD, resulting in 6.0% decrease average accuracy compared Hence, integrating into latter can be new development assisting counseling.

Language: Английский

---

A review of artificial intelligence applications in in vitro fertilization

Journal of Assisted Reproduction and Genetics, Journal Year: 2024, Volume and Issue: 42(1), P. 3 - 14

Published: Oct. 14, 2024

Language: Английский

---

Opportunities and limitations of introducing artificial intelligence technologies into reproductive medicine

Obstetrics Gynecology and Reproduction, Journal Year: 2025, Volume and Issue: unknown

Published: March 14, 2025

Given the increasing problem of infertility in Russian Federation, assisted reproductive technologies (ART) have proven to be one most effective treatments for this condition. Notably, introduction ART methods, particularly vitro fertilization (IVF), has led markedly increased birth rates over past two decades. Studies show that machine learning algorithms can process images embryos assess their quality, thus facilitating selection viable among them transfer. There are ethical and technical barriers hindering widespread adoption artificial intelligence (AI) clinical practice, including concerns data privacy as well a need train specialists deal with new technologies. AI analyze vast amounts data, medical histories research results, more accurately predict pregnancy outcomes. This enables doctors make justified decisions. In future, will able patient efficiently, helping identify causes at earlier stages.

Language: Английский

---

Artificial intelligence in human reproduction

Archives of Medical Research, Journal Year: 2024, Volume and Issue: 55(8), P. 103131 - 103131

Published: Nov. 29, 2024

Language: Английский

---

FertilitY Predictor—a machine learning-based web tool for the prediction of assisted reproduction outcomes in men with Y chromosome microdeletions

Journal of Assisted Reproduction and Genetics, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 9, 2024

Language: Английский

---

Noninvasive testing of preimplantation embryos in assisted reproductive technology

IntechOpen eBooks, Journal Year: 2024, Volume and Issue: unknown

Published: May 6, 2024

One approach to improving the success of assisted reproductive technology (ART) is careful selection embryos prior implantation. Although preimplantation genetic testing (PGT) widely employed for embryo selection, it needs biopsy and detrimental embryos. Thus, noninvasive offers new possibilities evaluating quality. Here, we reviewed current progression technologies, including use microscopy images combined with artificial intelligence (AI) select based on morphology, minimally invasive PGT blastocoel fluid spent culture medium, omics analysis molecules in medium assess developmental potential More importantly, using AI various type data each will greatly improve assessments. these cutting-edge technologies offer fresh insights into have enhance quality efficiency ART procedures.

Language: Английский

---

Predicting implantation by using dual AI system incorporating three‐dimensional blastocyst image and conventional embryo evaluation parameters—A pilot study

Reproductive Medicine and Biology, Journal Year: 2024, Volume and Issue: 23(1)

Published: Jan. 1, 2024

Abstract Purpose To investigate the usefulness of an original dual artificial intelligence (AI) system, in which first AI system eliminates background sliced tomographic blastocyst images, then second predicts implantation success using three‐dimensional (3D) reconstructed images sequential and conventional embryo evaluation parameters (CEE) such as maternal age. Methods Patients (from June 2022 to July 2023) could opt out there was additional information on Web site clinic. Implantation non‐implantation cases numbered 458 519, respectively. A total 10 747 a time‐lapse incubator with CEE were obtained. Results The statistic values by 0.774 ± 0.033 (mean standard error) for area under characteristic curve, 0.727 sensitivity, 0.719 specificity, predictive value positive test, negative 0.723 accuracy, Conclusions predicting handling 3D data demonstrated. This may be feasible option clinical practice.

Language: Английский

---

Применение цифровых продуктов в области вспомогательных репродуктивных технологий

Published: Nov. 20, 2024

Целью подготовки данного обзора является изучение применения цифровых продуктов в рамках программы вспомогательных репродуктивных технологий (ВРТ) с точки зрения рассмотрения информатизации регистров и методологической поддержки проведения ВРТ анализа использования подходов на разных этапах цикла ВРТ. В отечественных международных базах данных были отобраны проанализированы две группы статьей, посвященные регистрам алгоритмам машинного обучения за последние 5 лет. Исследования алгоритмов распределены по основным этапам ВРТ, также выделены основные преимущества недостатки выполненных работ. Разработка требует формирования набора последующего признаков, при этом данный процесс будет зависеть от рассматриваемого назначения алгоритма вида анализированных данных. данной работе был приведен этапы разработки моделей для предсказания исхода На основании выполненного опубликованных работ установлены ограничения исследований их перспектива. Было показано, что использование качестве принятия решения врачами отборе эмбрионов демонстрировало большую точность. Для внедрения должно быть проведено подтверждение безопасности эффективности разрабатываемых систем проспективных рандомизированных клинических исследований, которые обладают наивысшей степенью доказательности. Также выявлен недостаток исследования экономической целесообразности ИИ, которая должна оценена отдельных научных исследований.

Language: Русский

---