Tumor Delineation For Brain Radiosurgery by a ConvNet and Non-Uniform Patch Generation
This work addresses segmentation for small brain tumors in radiosurgery, an incremental improvement over existing methods.
The authors tackled the problem of brain tumor segmentation for stereotactic radiosurgery, where small lesion sizes cause high false negative and positive rates with existing methods like DeepMedic. They proposed a new patch-sampling procedure, achieving solid improvements in segmentation accuracy on a 6-year dataset across three tumor types.
Deep learning methods are actively used for brain lesion segmentation. One of the most popular models is DeepMedic, which was developed for segmentation of relatively large lesions like glioma and ischemic stroke. In our work, we consider segmentation of brain tumors appropriate to stereotactic radiosurgery which limits typical lesion sizes. These differences in target volumes lead to a large number of false negatives (especially for small lesions) as well as to an increased number of false positives for DeepMedic. We propose a new patch-sampling procedure to increase network performance for small lesions. We used a 6-year dataset from a stereotactic radiosurgery center. To evaluate our approach, we conducted experiments with the three most frequent brain tumors: metastasis, meningioma, schwannoma. In addition to cross-validation, we estimated quality on a hold-out test set which was collected several years later than the train one. The experimental results show solid improvements in both cases.