Abstract
Chronic liver diseases (CLD) account for more than 2% of deaths worldwide. Extensive research has been conducted to better understand CLD, generating vast amounts of data. However, only a small fraction of raw preclinical data are publicly available, posing a significant challenge for transparency, reproducibility, and data reuse. Therefore, we built a preclinical liver imaging dataset, the first of its kind to our knowledge. The database contains longitudinal liver MRI scans from mice with hepatocellular carcinoma, metabolic dysfunction-associated steatohepatitis (MASH, formerly NASH), and fibrosis, as well as CT scans of mice with MASH and mice carrying a dysfunctional ICAM-1 gene. Superimposable MRI and CT scans bridge the gap between the modalities. Some of the 222 murine scans have annotated segmentations. Metadata containing both scan and mouse parameters are organized using a tailored metadata profile in ISA-Tabs. This dataset enables advanced image analysis, such as building tools for automated segmentation, train radiomics analysis tools, or can be used as a reference control dataset.