Segmentation of Vertebral Metastases in MRI Using an U-Net like Convolutional Neural Network
This study’s objective was to segment vertebral metastases in diagnostic MR images by using a deep learning-based approach. Segmentation of such lesions can present a pivotal step towards enhanced therapy planning and implementation of minimally-invasive interventions like radiofrequency ablations. For this purpose, we used a U-Net-like architecture trained with 38 patient-cases. Our proposed method has been evaluated by comparison to expertly annotated lesion segmentations via Dice coeffcients, sensitivity and specificity rates. While the experiments with T1-weighted MRI images yielded promising results (average Dice score of 73:84 %), T2-weighted images were in average rather insufficient (53:02 %). To our best knowledge, our proposed study is the first to tackle this particular issue, which limits direct comparability with related works. In respect to similar deep learning-based lesion segmentations, e.g. in liver MR images or spinal CT images, our experiments with T1-weighted MR images show similar or in some respects superior segmentation quality.
Unable to display preview. Download preview PDF.
- 4.Liu J, Li M, Wang J, et al. A survey of MRI-based brain tumor segmentation methods. Tsinghua Sci Technol. 2014;19(6):578-595.Google Scholar
- 5.Christ PF, Ettlinger F, Grün F, et al. Automatic liver and tumor segmentation of ct and mri volumes using cascaded fully convolutional neural networks. arXiv:170205970. 2017;.
- 9.Ççicek Ӧ, Abdulkadir A, Lienkamp SS, et al.; Springer. 3D U-Net: learning dense volumetric segmentation from sparse annotation. Proc MICCAI. 2016; p. 424-432.Google Scholar
- 10.Isensee F, Kickingereder P, Bonekamp D, et al. Brain tumor segmentation using large receptive field deep convolutional neural networks. Proc BVM. 2017; p. 86-91.Google Scholar
- 11.Li X, Chen H, Qi X, et al. H-DenseUNet: hybrid densely connected UNet for liver and liver tumor segmentation from CT volumes. arXiv:170907330. 2017;.