van Loendersloot Laura, van Wely Madelon, van der Veen Fulco, Bossuyt Patrick, Repping Sjoerd
Center for Reproductive Medicine, Women's and Children's Hospital, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands.
Center for Reproductive Medicine, Women's and Children's Hospital, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands.
Reprod Biomed Online. 2014 Aug;29(2):222-30. doi: 10.1016/j.rbmo.2014.04.016. Epub 2014 May 15.
The selection of embryos based on morphology is still the core of daily laboratory practice in IVF/intracytoplasmic sperm injection. At present, the selection of embryos is primarily based on experience and local protocols. Since an evidence-based ranking strategy for embryos on day 3 is currently lacking, this work constructed a multivariable prediction model to rank embryos according to their implantation potential. A total of 6021 fresh embryo transfers between January 2004 and July 2009 were included, eight potential predictive factors were evaluated and a prediction model was developed using multivariable logistic regression. The model was externally validated with data from couples treated between August 2009 and September 2011 in the same clinic. Five factors were included in the final prediction model: early cleavage, number of blastomeres on days 2 and 3 and morphological score and presence of morula on day 3. With validation, the model showed moderate discriminative capacity (c-statistic 0.70) and calibrated well and was able to distinguish embryos with high ongoing implantation potential from embryos with moderate or low ongoing implantation potential. The model can be used by embryologists as an objective tool to rank embryos according to implantation potential, thereby aiding the selection of embryos for transfer.
基于形态学选择胚胎仍然是体外受精/卵胞浆内单精子注射日常实验室操作的核心。目前，胚胎的选择主要基于经验和当地的方案。由于目前缺乏基于证据的第3天胚胎分级策略，这项工作构建了一个多变量预测模型，根据胚胎的着床潜力对其进行分级。纳入了2004年1月至2009年7月期间共6021例新鲜胚胎移植，评估了8个潜在预测因素，并使用多变量逻辑回归建立了预测模型。该模型在同一诊所对2009年8月至2011年9月期间接受治疗的夫妇的数据进行了外部验证。最终的预测模型纳入了5个因素：早期卵裂、第2天和第3天的卵裂球数量、形态学评分以及第3天桑椹胚的存在情况。经验证，该模型显示出中等的判别能力（c统计量为0.70）且校准良好，能够区分具有高持续着床潜力的胚胎与具有中等或低持续着床潜力的胚胎。胚胎学家可以使用该模型作为一种客观工具，根据着床潜力对胚胎进行分级，从而辅助选择用于移植的胚胎。
