Automatic segmentation and location learning of neonatal cerebral ventricles in 3D ultrasound data combining CNN and CPPN

Preterm neonates are highly likely to suffer from ventriculomegaly, a dilation of the Cerebral Ventricular System (CVS). This condition can develop into life-threatening hydrocephalus and is correlated with future neurodevelopmental impairments. Consequently, it must be detected and monitored by physicians. In clinical routing, manual 2D measurements are performed on 2D ultrasound (US) images to estimate the CVS volume but this practice is imprecise due to the unavailability of 3D information. A way to tackle this problem would be to develop automatic CVS segmentation algorithms for 3D US data. In this paper, we investigate the potential of 2D and 3D Convolutional Neural Networks (CNN) to solve this complex task and propose to use Compositional Pattern Producing Network (CPPN) to enable Fully Convolutional Networks (FCN) to learn CVS location. Our database was composed of 25 3D US volumes collected on 21 preterm nenonates at the age of 35.8 +/- 1.6 gestational weeks. We found that the CPPN enables to encode CVS location, which increases the accuracy of the CNNs when they have few layers. Accuracy of the 2D and 3D FCNs reached intraobserver variability (IOV) in the case of dilated ventricles with Dice of 0.893 +/- 0.008 and 0.886 +/- 0.004 respectively (IOV = 0.898 +/- 0.008) and with volume errors of 0.45 +/- 0.42 cm(3) and 0.36 +/- 0.24 cm(3) respectively (IOV = 0.41 +/- 0.05 cm(3)). 3D FCNs were more accurate than 2D FCNs in the case of normal ventricles with Dice of 0.797 +/- 0.041 against 0.776 +/- 0.038 (IOV = 0.816 +/- 0.009) and volume errors of 0.35 +/- 0.29 cm(3) against 0.35 +/- 0.24 cm(3) (IOV = 0.2 +/- 0.11 cm(3)). The best segmentation time of volumes of size 320 x 320 x 320 was obtained by a 2D FCN in 3.5 +/- 0.2 s.