Three-dimensional transvaginal ultrasound offers superior live birth prediction after hysteroscopic adhesiolysis.

RESEARCH QUESTION: Can three-dimensional transvaginal ultrasound (3D-TVUS) characteristics be used to predict live births in patients with intrauterine adhesions (IUA)?

DESIGN

Patients with IUA (n = 688) after surgical abortion undergoing hysteroscopic adhesiolysis were retrospectively enrolled as the derivation cohort. A total of 426 patients with IUA were prospectively enrolled for external validation from another participating hospital. Patients were followed up over 2 years after hysteroscopic adhesiolysis for reproductive outcomes. Logistic regression analysis was used to develop prediction models. Performance was compared by calculating the area under curve (AUC).

RESULTS

Live births were achieved by 242 patients in the derivation cohort compared with 150 in the validation cohort. Logistic regression analysis revealed that general clinical features were statistically significant between live and no live birth cohorts. These included age, gravidity, parity, previous spontaneous miscarriage and 3D-TVUS features (endometrial thickness, endometrial echo homogeneous, segmentation of scar contraction middle, segmentation of scar contraction lower, segmentation of the endometrial deficiency upper and visibility of fallopian tube ostia-bilateral invisible). The AUCs for the prediction model, including these indicators in the derivation and validation cohorts, were 0.798 and 0.857, respectively; these were higher than the AUCs for the hysteroscopy pre-hysteroscopic adhesiolysis characteristics model (derivation cohorts: AUC = 0.706; validation cohorts: AUC = 0.69).

CONCLUSIONS

The use of preoperative 3D-TVUS is an accurate approach to predict live births for patients with IUA after hysteroscopic adhesiolysis, and can provide valuable reproductive guidance for these patients in the clinic.