Objective To construct a risk prediction model for contrast agent reflux in transvaginal four-dimensional hysterosalpingo-contrast sonography (TVS-4D-HyCoSy) and validate its efficacy.

Methods A total of 271 infertile patients who underwent TVS-4D-HyCoSy were retrospectively included as the study subjects. According to the presence or absence of contrast agent reflux in the ultrasonic images, they were divided into observation group (121 cases) and control group (150 cases). Clinical data of patients in the two groups were collected. Logistic regression analysis was used to screen the influencing factors of contrast agent reflux and construct a prediction model. The goodness- of-fit between the prediction model and the observed data was evaluated using the Hosmer-Lemeshow test. The receiver operating characteristic (ROC) curve was plotted to analyze the predictive efficacy of the model, and the area under the curve (AUC) was calculated to quantify the discriminatory ability.

Results There were statistically significant differences between the two groups in terms of infertility type, injection pressure, fallopian tube patency, history of previous uterine cavity surgery, intrauterine adhesion, and myometrial echo uniformity (P < 0.05). Multivariate Logistic regression analysis showed that injection pressure (OR=1.840, 95%CI, 1.336 to 2.790), fallopian tube patency (OR=3.411, 95%CI, 2.576 to 4.728), history of previous uterine cavity surgery (OR=4.522, 95%CI, 3.682 to 5.085), intrauterine adhesion (OR=1.893, 95%CI, 1.416 to 2.951), and myometrial echo uniformity (OR=2.380, 95%CI, 1.683 to 3.584) were independent influencing factors for the occurrence of contrast agent reflux (P < 0.05). A Logistic regression prediction model was constructed based on the independent influencing factors. The ROC curve showed that the AUC of the model for predicting contrast agent reflux was 0.836 (95%CI, 0.647 to 0.911), with a sensitivity of 93.02% and a specificity of 86.41%. The Hosmer-Lemeshow test showed χ²=4.326, P=0.503, indicating good predictive efficacy of the model.

Conclusion Injection pressure, fallopian tube patency, history of previous uterine cavity surgery, intrauterine adhesion, and myometrial echo uniformityare independent influencing factors for the occurrence of contrast agent reflux during TVS-4D-HyCoSy in infertile patients. The constructed risk prediction model for contrast agent reflux based on these factors shows good fit and can provide an accurate tool for risk assessment of contrast agent reflux, enhancing the safety and efficacy of the examination.