NICE: A Two‐Step Non‐Invasive Framework for Embryo cfDNA Read Enrichment and Quality Assessment

Non-invasive preimplantation genetic testing (niPGT) using cell-free DNA (cfDNA) extracted from spent embryo culture medium (SECM) has shown great potential, providing economic and practical advantages for embryo ploidy testing and quality assessment, while minimizing the risk of embryo damage. However, maternal DNA contamination, which can result in sex discordance and false-negative findings, remains a critical barrier to its clinical application in embryo prioritization. In this study, we present the NICE (Non-Invasive CfDNA-based Embryo assessment) framework, designed to enable contamination-resistant evaluation and support accurate embryo selection. The workflow integrates a two-step strategy: first, embryonic cfDNA is effectively purified using DECENT-plus-an enhanced version of our previously established deep CNV reconstruction algorithm (DECENT)-that minimizes interference from polar body-derived maternal DNA and improves signal resolution between maternal and embryonic origins. Subsequently, machine learning models based on biometric features extracted from the purified cfDNA are constructed to classify embryo quality, providing intelligent decision support for clinical embryo prioritization. By addressing the challenge of maternal contamination, NICE establishes a more automated, standardized and non-invasive paradigm for embryo quality assessment and underscores the critical role of cfDNA-based analysis in facilitating non-invasive selection of high-quality embryos to improve outcomes in assisted reproductive technology.