Conclusions
In the development cohort, the T-bet/GATA3 ratio was significantly lower in women with live birth than those patients with non-live birth [0.148 (0.101, 0.212) vs. 0.246 (0.170, 0.399), P＜0.0001]. In the validation cohort, changes in endometrial T-bet/GATA3 were similar among these groups. The endometrial T-bet/GATA3 ratio was an independent predictor of live birth after correction for patient age, anti-Mullerian hormone (AMH), quality of embryos transferred and other clinical characteristics (aOR = 0.280, 95 % CI: 0.169-0.462, P＜0.001). We developed and validated that an endometrial T-bet/GATA3 ratio at the cut-off of 0.22 had significant predictive value for live birth (developmental cohort: AUC = 0.76, 95 % CI: 0.70-0.81, P < 0.0001. validation cohort: AUC = 0.85 95 % CI: 0.76-0.95, P < 0.0001). Our results suggest that elevated endometrial T-bet/GATA3 ratio is an independent marker of live birth in infertile patients.