Reproducibility of dynamic MR imaging pelvic measurements: a multi-institutional study

Area of Science:

• Obstetrics and Gynecology
• Radiology
• Medical Imaging

Background:

• Pelvic floor disorders are a significant concern for women after childbirth.
• Dynamic magnetic resonance (MR) imaging offers detailed assessment of pelvic structures.
• Standardization of imaging protocols is crucial for multicenter research.

Purpose of the Study:

• To evaluate the reproducibility of bone and soft-tissue pelvimetry measurements from dynamic MR imaging.
• To assess inter-reader agreement across multiple clinical sites.

Main Methods:

• Prospective study of 198 primiparous women across six clinical sites.
• Standardized dynamic pelvic 1.5-T multiplanar T2-weighted MR imaging performed post-birth.
• Blinded readers performed initial measurements, followed by rereadings after additional training.

Main Results:

• Initial readings showed acceptable agreement (ICC ≥ 0.7) for 8 of 19 parameters.
• Bone measurements demonstrated better reproducibility than soft-tissue measurements.
• Additional training improved agreement for 12 of 19 parameters, but variability remained.

Conclusions:

• High variability exists in dynamic pelvic MR imaging measurements among readers, particularly for soft tissues.
• Measurement variability limits the utility of many MR imaging parameters for multicenter research.
• Further standardization efforts are needed to enhance reliability in pelvic floor disorder research.