Multimodal Volume-Aware Detection and Segmentation for Brain Metastases Radiosurgery
This work addresses the need for automated detection and segmentation in SRS treatment planning for brain metastases, representing a domain-specific improvement.
The researchers tackled the problem of manually intensive tumor delineation in stereotactic radiosurgery for brain metastases by developing a deep learning approach with multimodal imaging and ensemble neural networks, achieving performance that surpasses current benchmark levels.
Stereotactic radiosurgery (SRS), which delivers high doses of irradiation in a single or few shots to small targets, has been a standard of care for brain metastases. While very effective, SRS currently requires manually intensive delineation of tumors. In this work, we present a deep learning approach for automated detection and segmentation of brain metastases using multimodal imaging and ensemble neural networks. In order to address small and multiple brain metastases, we further propose a volume-aware Dice loss which optimizes model performance using the information of lesion size. This work surpasses current benchmark levels and demonstrates a reliable AI-assisted system for SRS treatment planning for multiple brain metastases.