Predictive Factors for the Formation of Viable Embryos in Subfertile Patients with Diminished Ovarian Reserve: A Clinical Prediction Study

Patient Characteristics General Characteristics

This retrospective study included records of 1016 blastocyst cultures, which were randomly divided into a training set and a validation set at a ratio of 7:3. The baseline demographic and clinical characteristics of the study population are summarized in Table 1. The characteristics include maternal age, paternal age, infertility duration, infertility type, number of blastocysts cultured, normal fertilization, the blastomere number of D2, embryo fragmentation of D2, day 3 embryo grade, fusion embryos, the blastomere number of D3, embryo fragmentation of D3, day 2 embryo grade, ovarian stimulation protocol, type of fertilization, maternal BMI, basal FSH, basal LH, basal PRL, basal E2, basal T, basal P, AMH, AFC, total Gn dosage, duration of stimulation, initial FSH dosage, serum E2 level on the HCG trigger day, serum LH level on the HCG trigger day, serum P level on the HCG trigger day, volume of semen after treatment, semen density after treatment, semen recovery rate, and semen recovery rate. Overall, the baseline characteristics were generally well-balanced between the training cohort and the internal test cohort, with non-significant p-values for most comparisons, suggesting that the two cohorts were suitable for predictive research.

Table 1 Patient demographics and baseline characteristics Predictive ModelLASSO Regression Model

The candidate predictors, maternal age, paternal age, infertility duration, infertility type, number of blastocysts cultured, normal fertilization, the blastomere number of D2, embryo fragmentation of D2, day 3 embryo grade, fusion embryos, the blastomere number of D3, embryo fragmentation of D3, day 2 embryo grade, ovarian stimulation protocol, type of fertilization, maternal BMI, basal FSH, basal LH, basal PRL, basal E2, basal T, basal P, AMH, AFC, total Gn dosage, duration of stimulation, initial FSH dosage, serum E2 level on the HCG trigger day, serum LH level on the HCG trigger day, serum P level on the HCG trigger day, volume of semen after treatment, semen density after treatment, and semen recovery rate, were included in the original model, which were then reduced to 9 potential predictors using LASSO regression analysis performed in the training cohort. The coefficients are shown in the following table, and a coefficient profile is plotted in Fig. 1. A cross-validated error plot of the LASSO regression model is also shown in Fig. 2. The most regularized and parsimonious model, with a cross-validated error within one standard error of the minimum, included 9 variables. As shown in Fig. 3, the ROC analysis of the abovementioned variables yielded AUC values greater than 0.5.

Fig. 1

Lasso regression cross-validation plot

Fig. 2

Lasso regression coefficient path plot

Fig. 3

ROC curve analysis of 9 candidate diagnostic indicators

Multivariate Logistic Analyses

Using multivariate logistic regression analysis, further analysis was conducted on the optimal matching factors identified by Lasso regression. The results revealed that five variables—female age, normal fertilization, day 2 cleavage-stage embryo grading, day 3 cleavage-stage embryo grading, and method of fertilization—were independent predictors of blastocyst formation in patients with diminished ovarian reserve (DOR), as shown in Table 2

Table 2 Results of multivariate logistic regressionNomogram Prediction Model Construction of a Nomogram Predicting Usable Blastocyst Formation in DOR Patients

Based on the results of the multivariate logistic regression analysis, a nomogram was constructed to predict the formation of usable blastocysts in DOR patients, incorporating female age, normal fertilization, day 2 cleavage-stage embryo grading, day 3 cleavage-stage embryo grading, and method of fertilization. The model for this nomogram is shown in Fig. 4. A nomogram is a graphical calculating device, a two-dimensional diagram designed to allow the approximate graphical computation of a function. It is based on the principles of multivariable regression analysis and integrates multiple prognostic indicators by employing scaled lines drawn proportionally on the same plane to express the interrelationships among various variables in a predictive model. Each variable is represented by a line segment with marked scales, indicating the range of possible values for that variable, while the length of the line segment reflects the impact of that factor on the outcome event. As illustrated in Fig. 4, the age of patients with diminished ovarian reserve (DOR) is the most significant factor affecting the formation of usable blastocysts. This is followed in importance by day 3 embryo grade, normal fertilization, type of fertilization, and day 2 embryo grade.

Fig. 4

Nomogram prediction model. Normal fertilization: 0 represents non-2PN (pronuclear), and 1 represents 2PN (pronuclear). Type of fertilization: 1 represents in vitro fertilization (IVF), and 2 represents intracytoplasmic sperm injection (ICSI)

ROC Curves of the Nomogram Prediction Model Analysis of Calibration of the Nomogram for Predicting Usable Blastocyst Formation in DOR Patients

The area under the receiver operating characteristic (ROC) curve (AUC) for the model constructed from the training set was 0.832. Internal validation of the nomogram model was performed, and the AUC for the validation set was 0.793, as seen in Fig. 5. In the nomogram prediction model, the individual ROC for each of the five included factors was ≤ 0.63.The calibration plots of the nomogram in the different cohorts are plotted in the following figures, which demonstrate a good correlation between the observed and predicted Usable blastocyst. The results showed that the original nomogram was still valid for use in the validation sets, and the calibration curve of this model was relatively close to the ideal curve, which indicates that the predicted results were consistent with the actual findings.

Fig. 5

ROC curves of the nomogram prediction model

Decision Curve Analysis

The following figure displays the DCA curves related to the nomogram. A high-risk threshold probability indicates the chance of significant discrepancies in the model’s prediction when clinicians encounter major flaws while utilizing the nomogram for diagnostic and decision-making purposes. This research shows that the nomogram offers substantial net benefits for clinical application through its DCA curve (Figs. 6, 7, 8 and 9).

Fig. 6