Research question

Is it possible to establish a visualized clinical model predicting good quality blastocyst (GQB) formation for patients in their first IVF/intracytoplasmic sperm injection (ICSI) cycle?

Design

A total of 4783 patients in their first IVF/ICSI cycle between January 2015 and December 2019 were retrospectively included and randomly divided into the training set (n = 3826) and the testing set (n = 957) in an 8:2 ratio. The least absolute shrinkage and selection operator (LASSO) regression was adopted to select the most critical predictors for GQB formation to construct a visualized nomogram model based on the data of patients in the training set. Receiver operating characteristic and calibration curves were used to evaluate the predictive accuracy and discriminative ability. The performance of the model was also validated on independent data from patients treated in the testing set.

Results

Maternal age, maternal serum anti-Müllerian hormone (MsAMH) concentration and the number of oocytes retrieved were highlighted as critical predictors of GQB development and were incorporated into the nomogram model. Based on the area under the curve (AUC) values, the predictive ability for ≥1, ≥3 and ≥5 GQB were 0.831, 0.734 and 0.748, respectively. The calibration curve also showed high concordance between the observed and predicted results. The AUC for predicting ≥1, ≥3 and ≥5 GQB in the testing set were 0.805, 0.695 and 0.707, respectively, which were similar to those for the training set.

Conclusions

The visualized nomogram model provides great predictive value for GQB development in patients in their first IVF/ICSI cycle and can be used to improve clinical counselling.

中文翻译：

---

预测第一个 IVF/ICSI 周期中优质囊胚发育的可视化临床模型。

研究问题

是否有可能建立一个可视化的临床模型来预测第一个 IVF/卵胞浆内单精子注射 (ICSI) 周期中患者的优质囊胚 (GQB) 形成？

设计

回顾性纳入 2015 年 1 月至 2019 年 12 月期间第一个 IVF/ICSI 周期的 4783 名患者， 并 以 8:2 的比例随机分为训练集（n = 3826）和测试集（n = 957）。采用最小绝对收缩和选择算子（LASSO）回归选择最关键的GQB形成预测因子，以基于训练集中患者的数据构建可视化的列线图模型。接受者操作特征和校准曲线用于评估预测准确性和辨别能力。该模型的性能还通过来自测试集中治疗的患者的独立数据进行了验证。

结果

母亲年龄、母亲血清抗苗勒管激素 (MsAMH) 浓度和取回的卵母细胞数量被强调为 GQB 发展的关键预测因素，并被纳入列线图模型。根据曲线下面积 (AUC) 值，≥1、≥3 和 ≥5 GQB 的预测能力分别为 0.831、0.734 和 0.748。校准曲线还显示出观察结果和预测结果之间的高度一致性。在测试集中预测≥1、≥3和≥5 GQB的AUC分别为0.805、0.695和0.707，与训练集相似。

结论

可视化列线图模型为患者在第一个 IVF/ICSI 周期中的 GQB 发展提供了巨大的预测价值，可用于改善临床咨询。