Late gestation MRI to assess maternal pelvimetry, fetal biometry and placental oxygenation: a retrospective pilot...

The role of MRI pelvimetry and fetal size estimation in predicting mode of birth and risk of operative birth have been studied. However, there are no complete MRI studies that assess the maternal pelvis, fetal dimensions, and the placenta in a single protocol, in order to better inform the likelihood of operative vaginal birth or emergency caesarean section. Therefore, the aim of this pilot study was to assess the feasibility of obtaining a comprehensive prelabour assessment of maternal pelvimetry and fetal biometry using 3D MRI reconstructions, in addition to measures of placental function, in one MR examination.

This was a retrospective cohort study of 29 women in late third trimester. Maternal pelvimetry and fetal measurements were performed using structural motion corrected T2 weighted MR images, placental T2* values (an indirect measure of placental oxygenation), and birth outcomes were also collected. Intra and inter-rater variability were calculated for the first 10 patients using the intraclass correlation coefficient. The correlation between manual (measuring the contour area) and calculated circumferences of maternal and fetal structures were also assessed to compare the practicability of performing the two alternative approaches.

People were imaged between 36 + 1 to 38 + 4 weeks' gestation. It was possible to obtain comprehensive maternal and fetal measurements. Intra-rater variability was generally excellent, and inter-rater reliability was moderate to excellent. There was a strong correlation between manually obtained and calculated circumferences; Spearman's ranged from 0.75 to 0.95. Placental volume, mean T2* and kurtosis were available for 23 datasets. The median placental volume was 569.7, the median T2* mean was 44.2, and the median kurtosis was 1.4.

It is possible to perform maternal pelvimetry, fetal biometry and assess placental oxygenation from one late gestation MRI examination. The approach could be employed in a large, prospective study to ascertain whether we can predict the likelihood of assisted birth or caesarean section, with automation of image analysis to minimise inter-rater variability.