| Summary: | Volumetric placental measurement using 3D ultrasound has proven clinical utility in predicting adverse pregnancy outcomes. However, this metric can not currently be employed as part of a screening test due to a lack of robust and real-time segmentation tools. We present a multi-class convolutional neural network (CNN) developed to segment the placenta, amniotic fluid and fetus. The ground truth dataset consisted of 2,093 labelled placental volumes augmented by 300 volumes with placenta, amniotic fluid and fetus annotated. A two pathway, hybrid model using transfer learning, a modified loss function and exponential average weighting was developed and demonstrated the best performance for placental segmentation, achieving a Dice similarity coefficient (DSC) of 0.84 and 0.38 mm average Hausdorff distance (HDAV). Use of a dual-pathway architecture, improved placental segmentation by 0.03 DSC and reduced HDAV by 0.27mm when compared with a naïve multi-class model. Incorporation of exponential weighting produced a further small improvement in DSC by 0.01 and a reduction of HDAV by 0.44mm. Per volume inference using the FCNN took 7-8 seconds. This method should enable clinically relevant morphometric measurements (such as volume and total surface area) to be automatically generated for the placenta, amniotic fluid and fetus. Ready availability of such metrics makes a population-based screening test for adverse pregnancy outcomes possible.
|
|---|