Tumor Delineation For Brain Radiosurgery by a ConvNet and Non-Uniform Patch Generation

TL;DR

This paper introduces a novel patch-sampling method for deep learning-based brain tumor segmentation, improving detection accuracy for small lesions in stereotactic radiosurgery applications.

Contribution

A new non-uniform patch generation technique tailored for small brain tumors enhances deep learning segmentation performance.

Findings

Improved detection of small lesions.

Enhanced segmentation accuracy on a multi-year dataset.

Solid performance gains demonstrated in experiments.

Abstract

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.