利用人工智能（AI）技术进行标准化的胚胎植入前遗传学检测（PGT-A）与单冻融整倍体胚胎移植（STEET）周期中妊娠结局的改善相关。
Utilization of standardized preimplantation genetic testing for aneuploidy (PGT-A) via artificial intelligence (AI) technology is correlated with improved pregnancy outcomes in single thawed euploid embryo transfer (STEET) cycles.
机构信息
New York University Grossman School of Medicine, 550 First Avenue, NBV 9E2, New York, NY, 10016, USA.
CooperSurgical, Inc., 75 Corporate Drive, Trumbull, CT, 06611, USA.
出版信息
J Assist Reprod Genet. 2023 Feb;40(2):289-299. doi: 10.1007/s10815-022-02695-7. Epub 2023 Jan 7.
PURPOSE
To investigate the role of standardized preimplantation genetic testing for aneuploidy (PGT-A) using artificial intelligence (AI) in patients undergoing single thawed euploid embryo transfer (STEET) cycles.
METHODS
Retrospective cohort study at a single, large university-based fertility center with patients undergoing in vitro fertilization (IVF) utilizing PGT-A from February 2015 to April 2020. Controls included embryos tested using subjective NGS. The first experimental group included embryos analyzed by NGS utilizing AI and machine learning (PGTai Technology Platform, AI 1.0). The second group included embryos analyzed by AI 1.0 and SNP analysis (PGTai2.0, AI 2.0). Primary outcomes included rates of euploidy, aneuploidy and simple mosaicism. Secondary outcomes included rates of implantation (IR), clinical pregnancy (CPR), biochemical pregnancy (BPR), spontaneous abortion (SABR) and ongoing pregnancy and/or live birth (OP/LBR).
RESULTS
A total of 24,908 embryos were analyzed, and classification rates using AI platforms were compared to subjective NGS. Overall, those tested via AI 1.0 showed a significantly increased euploidy rate (36.6% vs. 28.9%), decreased simple mosaicism rate (11.3% vs. 14.0%) and decreased aneuploidy rate (52.1% vs. 57.0%). Overall, those tested via AI 2.0 showed a significantly increased euploidy rate (35.0% vs. 28.9%) and decreased simple mosaicism rate (10.1% vs. 14.0%). Aneuploidy rate was insignificantly decreased when comparing AI 2.0 to NGS (54.8% vs. 57.0%). A total of 1,174 euploid embryos were transferred. The OP/LBR was significantly higher in the AI 2.0 group (70.3% vs. 61.7%). The BPR was significantly lower in the AI 2.0 group (4.6% vs. 11.8%).
CONCLUSION
Standardized PGT-A via AI significantly increases euploidy classification rates and OP/LBR, and decreases BPR when compared to standard NGS.
目的
研究人工智能（AI）在接受单冻融整倍体胚胎移植（STEET）周期的患者中进行标准化胚胎植入前遗传学检测（PGT-A）的作用。
方法
这是一项在单一大型大学附属生育中心进行的回顾性队列研究，该中心于 2015 年 2 月至 2020 年 4 月期间对接受体外受精（IVF）的患者使用 PGT-A。对照组包括使用主观 NGS 检测的胚胎。第一个实验组包括使用 AI 和机器学习（PGTai 技术平台，AI 1.0）进行 NGS 分析的胚胎。第二个组包括使用 AI 1.0 和 SNP 分析（PGTai2.0，AI 2.0）分析的胚胎。主要结局包括整倍体率、非整倍体率和单纯镶嵌率。次要结局包括种植率（IR）、临床妊娠率（CPR）、生化妊娠率（BPR）、自然流产率（SABR）和持续妊娠和/或活产率（OP/LBR）。
结果
共分析了 24908 个胚胎，比较了 AI 平台的分类率与主观 NGS。总体而言，通过 AI 1.0 检测的胚胎显示出显著增加的整倍体率（36.6%比 28.9%）、降低的单纯镶嵌率（11.3%比 14.0%）和降低的非整倍体率（52.1%比 57.0%）。总体而言，通过 AI 2.0 检测的胚胎显示出显著增加的整倍体率（35.0%比 28.9%）和降低的单纯镶嵌率（10.1%比 14.0%）。与 NGS 相比，AI 2.0 检测的非整倍体率降低不具有统计学意义（54.8%比 57.0%）。共移植了 1174 个整倍体胚胎。AI 2.0 组的 OP/LBR 显著更高（70.3%比 61.7%）。AI 2.0 组的 BPR 显著更低（4.6%比 11.8%）。
结论
与标准 NGS 相比，通过 AI 进行标准化 PGT-A 可显著提高整倍体分类率和 OP/LBR，并降低 BPR。
Zotero 插件