CellSighter Test Dataset: CRC Multiplexed Images

Description: This dataset contains multiplexed protein images from colorectal cancer (CRC) patients, used to benchmark the CellSighter algorithm. The images were derived from a subset of the publicly available CRC-FFPE-CODEX collection hosted by The Cancer Imaging Archive (TCIA): https://www.cancerimagingarchive.net/collection/crc_ffpe-codex_cellneighs/ Original dataset citation: Schürch, C. M., Bhate, S., Barlow, G., Phillips, D., Noti, L., Zlobec, I., Chu, P., Black, S., Demeter, J., McIlwain, D., Samusik, N., Goltsev, Y., & Nolan, G. (2020). High-dimensional imaging of colorectal carcinoma and other tumors with 50+ markers [Data set]. The Cancer Imaging Archive. https://doi.org/10.7937/TCIA.2020.FQN0-0326 Derived Annotations: The images were re-segmented using DeepCell (https://www.deepcell.org/). Cell type classifications were generated using CellTune (https://celltune.org/). These segmentations and classifications are original contributions and do not appear in the TCIA dataset. Images, segmentations, and classifications were formatted for use with CellSighter. Usage Policy: This dataset is shared under the terms of TCIA’s data usage policy. It must not be used in any way that could identify individuals. Users must cite both the original dataset and this derived dataset appropriately and use in accordance with TCIA guidelines. TCIA data usage policy: https://www.cancerimagingarchive.net/data-usage-policies-and-restrictions/ Use Case: This dataset was used to evaluate CellSighter, as described in: Amitay, Y., Bussi, Y., Feinstein, B. et al. CellSighter: a neural network to classify cells in highly multiplexed images. Nat Commun 14, 4302 (2023). https://doi.org/10.1038/s41467-023-40066-7