P-204 Cumulus cell derived gene expression models can predict live birth in ICSI patients stimulated with r-hFSH or r-hFSH & r-hLH Article Swipe

Study question Is there a difference in cumulus cell (CC) gene expression between two types of ovarian stimulation with recombinant gonadotropins and can gene-based algorithms predict outcome? Summary answer The CC gene expression is different between r-hFSH and r-hFSH & r-hLH stimulated patients and gene-only based algorithms can potentially predict live birth What is known already One factor influencing ART efficiency is the choice of which embryo to transfer. RNA expression biomarkers derived from CC of individual oocytes of ICSI patients have been used to predict the developmental potential of oocytes and then prioritize an embryo for transfer. A gene expression model (CAMKD1, EFNB2, SASH1) predicting clinical pregnancy was developed earlier for patients stimulated with HP-hMG and clinically validated in a prospective trial. CC expression differs in women stimulated with r-hFSH. It is not known whether CC gene expression differs significantly in women stimulated with r-hFSH or r-hFSH & r-hLH. Study design, size, duration This observational cohort study was performed in a tertiary university hospital IVF centre from January 2021 until June 2023. 137 ICSI patients were recruited and signed informed consent. 24 patients were considered as drop-outs and 113 patients were allocated to one of the two study groups: 47 patients were stimulated with r-hFSH, and 66 patients were stimulated with r-hFSH & r-hLH. For all 113 patients, ICSI was performed and Day 5 SET was scheduled. Participants/materials, setting, methods CC were collected from 113 patients after individual oocyte denudation using Cumulase. RNA extraction of 1135 CC samples was performed with the RNeasy Micro kit, followed by RT-QPCR for 11 preselected oocyte quality biomarker genes (CAMK1D, EFNB2, SASH1, GOT1, SLC6A9, HAS2, PTGS2, HSPH1, VCAN, GSTA4, STC2) and 2 endogenous controls (UBC, B2M). One-parametric (95%CI) and multiparametric analyses (leave-one-out cross-validation and stepwise linear regression) were performed on the gene expression data. Main results and the role of chance Patient characteristics were comparable between both groups, except for serum LH (p < 0,0001) and serum E2 two days prior to oocyte retrieval (p = 0,0002). One-parametric analysis indicated significant differential expression between the two groups for EFNB2 (CC of all analysed oocytes n = 1123, 95%CI 0.27,0.49) and GOT1 (CC of all analysed oocytes n = 1123, 95%CI -0.25,-0.11). Results were confirmed in 4 subsets: MII oocytes n = 868, 2PN oocytes n = 609, oocytes developing into a good quality embryo (GQE) Day3 n = 536, oocytes developing into a GQE Day5 n = 285, and are independent of potential differences in oocyte competence. In patients stimulated with r-hFSH & r-hLH, the CC showed significantly higher EFNB2 expression and significantly lower GOT1 expression. Stimulation-specific predictive gene models were obtained using stepwise linear regression and 11 biomarker genes. Gene expression in CC of oocytes developing into transferred blastocysts was compared based on transfer outcome (live birth (LB) or not). The strongest predictive gene combination for LB in r-hFSH patients (n = 61 fresh + frozen transfers: 33 LB + 28 no LB) was CAMK1D, HAS2 and PTGS2 (AUC 0,6829; accuracy 69%). For r-hFSH & r-hLH patients (n = 85 (36 LB + 49 no LB), GOT1 and HAS2 was the best predictive model for LB (AUC 0,9529; accuracy 88%). Limitations, reasons for caution The new predictive gene models for r-hFSH and r-hFSH & r-hLH treated patients were identified for the first time in this observational cohort study. High accuracy of the predictive gene models merits further validation in prospective clinical studies. The required individual denudation is more time consuming compared to grouped denudation. Wider implications of the findings Gene-only based algorithms hold the potential to predict live birth. Measuring the cumulus cell predictive biomarkers in a treatment specific way is a promising non-invasive technology to enhance the ART efficiency by shortening the time-to-pregnancy in ICSI patients. Trial registration number Yes