"Pseudoseptum" of the uterine cervix on MRI

Abstract

Purpose: To determine the prevalence of a longitudinal " pseudoseptum" on T2-weighted MR images within the cervices of women who do not have a uterine anomaly. Materials and Methods: We reviewed 317 consecutive female pelvic MR examinations performed at a single institution over a four-month period. All examinations included T2-weighted sequences in at least two orthogonal planes. Of the 317 studies, 57 were excluded due to prior radical hysterectomy. Axial and coronal T2-weighted images of the remaining 260 examinations were evaluated for the presence of a longitudinal low T2 signal intensity structure within the endocervical lumen that mimicked the appearance of a septum. Interpretations were performed independently by two MR radiologists and kappa analysis of interobserver agreement was performed. Results: In 50 (19%) of the 260 women, both readers noted the presence of a pseudoseptum on at least one imaging plane. In 162 (62%), neither reader noted a pseudoseptum. Overall, there was 81% agreement between the readers. Kappa analysis yielded a value of 0.55, indicating a moderate degree of interobserver agreement beyond chance. Conclusion: A pseudoseptum was depicted in 20% to 30% of women's cervices on T2-weighted imaging. We hypothesize that chance long-axis depiction of the endocervical folds can mimic a cervical septum. The presence of a pseudoseptum on MRI should be considered a normal finding and not a feature of a developmental anomaly of the uterus or cervix. © 2007 Wiley-Liss, Inc.