Fig. 5: Impact of follicle size profile and elevated progesterone on live birth rate. | Nature Communications

Fig. 5: Impact of follicle size profile and elevated progesterone on live birth rate.

From: Explainable artificial intelligence to identify follicles that optimize clinical outcomes during assisted conception

Fig. 5

Partial dependence plots under logistic regression modeling (n = 9843) to demonstrate the impact of (a) the percentage of follicles sized 13–18 mm on the day of trigger (DoT) (OR: 1.03 (1.00 − 1.06) per 10 percentage points change; p = 0.048), and (b) mean follicle size on the DoT (OR: 0.95 (0.93 − 0.98) per 1 mm change; p =  0.001), on predicted live birth rate (LBR). Both models were adjusted for: age, total follicle count on the DoT, and type of trigger administered (hCG or GnRH agonist). The mean LBR and its 95% confidence interval (CI) are plotted with 100 bootstrapped simulations. c The association (mean and 95% CI) between increasing progesterone elevation on the DoT versus mature oocyte yield (n = 646) and LBR (n = 427). d The association (mean and 95% CI) between an increasing number of follicles sized >18 mm on DoT versus serum progesterone levels on the DoT (n = 994). Having at least 7 follicles of >18 mm in size presented a significant elevation in progesterone with reference to patients with <2 follicles >18 mm on the DoT using the two-way Dunnett’s multiple comparisons test reported with adjusted p values.

Back to article page