EP01.31: Agreement between measurements of the anteroposterior diameter of the levator ani muscle hiatus using 2D versus automated 3D transperineal ultrasound

To assess the agreement between the measurements of the anteroposterior diameter (APD) of the levator ani muscle hiatus using 2D versus automated 3D transperineal ultrasound (TPUS) techniques. We recruited a group of nulliparous women at term before the onset of labour. For each woman, we measured APD in the midsagittal plane using 2D TPUS and acquired a 3D TPUS volume at rest. Subsequently, we measured APD on 3D volumes using an automated software on the axial plane (GE Healthcare, Zipf, Austria). Manual correction of 3D automated measurement was performed in cases where the automated software did not correctly depict the boundaries of the levator hiatus. Intermethod agreement between 2D vs. 3D APD measurements was calculated using intraclass correlation coefficient (ICC). To investigate the presence of fixed bias and/or proportional bias, intermethod agreement was also analysed using the Bland–Altman plot. Student's t-test and Levene's W0 test were used to assess the effect several clinical factors on systematic differences and homogeneity between APD measurements. The analysis included 155 women. In the automated 3D TPUS, APD was measured successfully without need for adjustment in 149 cases (96.1%). Intermethod analyses showed almost perfect agreement between 2D and automated 3D APD measurements with ICC of 0.91 (95% CI 0.87-0.93). A systematic difference was found between 2D and 3D measurements, although the difference was clinically irrelevant (52.8 vs 53.8mm respectively, P= 0.01). The examination of maternal body mass index and maternal age did not show a statistically significant effect on the agreement. We found a very high agreement between 2D and 3D automated transperineal ultrasound measurements of the anteroposterior diameter of the levator hiatus. In nulliparous women at term both techniques can thus be used interchangeably.