Purpose
Imageâguided radiotherapy provides images not only for patient positioning but also for online adaptive radiotherapy. Accurate delineation of organsâatârisk (OARs) on Head and Neck (H&N) CT and MR images is valuable to both initial treatment planning and adaptive planning, but manual contouring is laborious and inconsistent. A novel method based on the generative adversarial network (GAN) with shape constraint (SCâGAN) is developed for fully automated H&N OARs segmentation on CT and lowâfield MRI.
Methods and material
A deep supervised fully convolutional DenseNet is employed as the segmentation network for voxelâwise prediction. A convolutional neural network (CNN)âbased discriminator network is then utilized to correct predicted errors and imageâlevel inconsistency between the prediction and ground truth. An additional shape representation loss between the prediction and ground truth in the latent shape space is integrated into the segmentation and adversarial loss functions to reduce false positivity and constrain the predicted shapes. The proposed segmentation method was first benchmarked on a public H&N CT database including 32 patients, and then on 25 0.35T MR images obtained from an MRâguided radiotherapy system. The OARs include brainstem, optical chiasm, larynx (MR only), mandible, pharynx (MR only), parotid glands (both left and right), optical nerves (both left and right), and submandibular glands (both left and right, CT only). The performance of the proposed SCâGAN was compared with GAN alone and GAN with the shape constraint (SC) but without the DenseNet (SCâGANâResNet) to quantify the contributions of shape constraint and DenseNet in the deep neural network segmentation.
Results
The proposed SCâGAN slightly but consistently improve the segmentation accuracy on the benchmark H&N CT images compared with our previous deep segmentation network, which outperformed other published methods on the same or similar CT H&N dataset. On the lowâfield MR dataset, the following average Dice's indices were obtained using improved SCâGAN: 0.916 (brainstem), 0.589 (optical chiasm), 0.816 (mandible), 0.703 (optical nerves), 0.799 (larynx), 0.706 (pharynx), and 0.845 (parotid glands). The average surface distances ranged from 0.68Â mm (brainstem) to 1.70Â mm (larynx). The 95% surface distance ranged from 1.48Â mm (left optical nerve) to 3.92Â mm (larynx). Compared with CT, using 95% surface distance evaluation, the automated segmentation accuracy is higher on MR for the brainstem, optical chiasm, optical nerves and parotids, and lower for the mandible. The SCâGAN performance is superior to SCâGANâResNet, which is more accurate than GAN alone on both the CT and MR datasets. The segmentation time for one patient is 14Â seconds using a single GPU.
Conclusion
The performance of our previous shape constrained fully CNNs for H&N segmentation is further improved by incorporating GAN and DenseNet. With the novel segmentation method, we showed that the lowâfield MR images acquired on a MRâguided radiation radiotherapy syste...