Clinical prediction models for in vitro fertilization outcomes: a systematic review, meta-analysis, and external validation.

This article systematically reviews and meta-analyzes prognostic models predicting live birth outcomes following in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI), with a focus on identifying the best-performing model. Among 72 studies encompassing 86 models and 132 predictors, McLernon's post-treatment model demonstrated superior discriminative ability, with an area under the receiver operating characteristic curve (AUC) of 0.73 in meta-analysis and showed the best calibration in external validation using a large Chinese IVF cohort. The study highlights that while many models exist, most exhibit moderate to low predictive performance and suffer from methodological limitations such as high risk of bias and heterogeneity across populations. The findings support the clinical utility of McLernon's post-treatment model but emphasize the need for further development and external validation of IVF prognostic models to improve individualized patient counseling and treatment decisions.