A deep convolutional neural network-based automatic delineation strategy for multiple brain metastases stereotactic radiosurgery

• DOI: 10.1371/journal.pone.0185844
• Publication Date: 2017-10-06T17:34:26Z
• Authors: Yan Liu; Strahinja Stojadi...; Brian Hrycushko; Zabi Wardak; Steven Lau; Weiguo Lu; Yulong Yan; Steve B. Jiang; Xin Zhen; Robert Timmerman; Lucien Nedzi; Xuejun Gu
• Keywords: Benchmark (surveying)

Accurate and automatic brain metastases target delineation is a key step for efficient effective stereotactic radiosurgery (SRS) treatment planning. In this work, we developed deep learning convolutional neural network (CNN) algorithm segmenting on contrast-enhanced T1-weighted magnetic resonance imaging (MRI) datasets. We integrated the CNN-based into an segmentation workflow validated both Multimodal Brain Tumor Image Segmentation challenge (BRATS) data clinical patients' data. Validation BRATS yielded average DICE coefficients (DCs) of 0.75±0.07 in tumor core 0.81±0.04 enhancing tumor, which outperformed most techniques 2015 challenge. results patient cases showed DCs 0.67±0.03 achieved area under receiver operating characteristic curve 0.98±0.01. The strategy surpasses current benchmark levels offers promising tool SRS planning multiple metastases.