Slice2Volume: A registered dataset of multimodal medical imaging and light microscopy data in irradiation-injured brain tissue

Recent years have shown that particle therapy offers highly conformal brain irradiation and optimized healthy tissue sparing. Nevertheless, elevated dose levels in healthy tissue, particularly in the distal beam region, can lead to undesirable long-term side effects. The biological mechanisms of such effects, however, remain unclear. A major obstacle towards correlating effects on clinical and cellular imaging levels is the mapping of radiation dose to specific brain regions or individual cell populations.
We present a publicly available dataset of registered, multimodal imaging data of nine mice that received proton brain irradiation of different doses in a clinically relevant setting. It is available open access (https://rodare.hzdr.de/record/801) and comprises a baseline computed tomography (CT) scan, simulated distributions of dose and linear energy transfer, a co-aligned mouse brain atlas as well as magnetic resonance imaging (MRI) follow-up of up to six months. Additionally, we provide registered histological brain sections with eight histological stainings, reflecting all major cell types in adult mice brains. We used the self-developed tool Slice2Volume together with existing methods (Elastix & Big Warp) to fuse image data. The software is available open source: https://github.com/jo-mueller/Slice2Volume
The provided image data spans several orders of magnitude of scale. Images of all modalities can be freely overlaid for every mouse as is demonstrated on Figure 1. This, for instance, allows tracing MRI image changes to specific cell populations. Hence, the dataset enables direct correlations and mechanistic observations regarding effects of proton radiation on the anatomical (atlas), clinical (MRI) and microscopic level (histology).