(75) MRI in the Assessment of Cervical Anatomy in Müllerian Anomalies

Background: Patients with an obstructed Müllerian structure must undergo surgical management with resection of the structure or anastomosis of the structure to a vagina or neovagina, with the goal of relief of obstruction and preservation of fertility, where appropriate. The decision to offer fertility sparing treatment is often based on imaging findings reflecting the presence or absence of the cervix, however the sensitivity of MRI in assessment of Müllerian anomalies varies widely in the literature (28-100%), and this quoted sensitivity is largely based on fundal contour rather than cervical findings. This study seeks to assess the accuracy of MRI for identification of cervical tissue based on concordance between imaging, intraoperative, and pathologic findings.

Methods: This was an IRB approved retrospective review of patients who underwent resection of a Müllerian structure at a single institution utilizing available electronic medical records from January 1, 2004 to August 1, 2024. Patients who underwent imaging without corresponding surgical intervention were excluded from this study. True positives were defined as presence of a cervix on MRI that corresponded to presence of cervix on pathologic evaluation or intraoperative visualization if no pathology was available. True negatives were defined as the absence of normal cervical tissue on pathology if available, or absence of cervical tissue noted on intraoperative findings if no pathology was available.

Results: A total of 149 patients who underwent resection of a Müllerian structure were identified. Of these, 25 MRI studies were identified commenting on the presence or absence of a total of 46 cervices based on Müllerian anatomy in each of the 25 patients. 24 of these studies were performed post-pubertally, and 1 was performed pre-pubertally. 16 of the cervices identified on MRI were found to be true positives, 8 were found to be false positives. There were 21 true negatives identified by MRI, and 1 false negative. Total sensitivity for MRI in assessment of the cervix was 94.12% (95% CI 71.3-99.8%), and specificity was 72.41% (95% CI 52.7-87.3%).

Conclusions: While accuracy of MRI assessment of uterine fundal contour has been found to be equivalent to that of 3D ultrasound, fundal contour is not a determining factor when it comes to fertility potential of reproductive structures. This is the first study to assess the accuracy of MRI as regards the presence or absence of a cervix or cervices in Müllerian anomalies. The sensitivity and specificity of MRI as it pertains to the cervix should be taken into consideration when determining whether to offer patients with obstructed Müllerian anomalies fertility sparing treatment.