Summary
In this dataset, we provide the raw 3D open-top light-sheet (OTLS) microscope images of prostate biopsies stained with a fluorescent analog of H&E at original resolution, the 2x down-sampled fused OTLS H&E analog images, the synthetic cytokeratin-8 (CK8) images at 2x-downsampled resolution, the 3D binary segmentation masks of glands at 4x down-sampled resolution, and the clinical data for patient outcomes, cancer regions, and Gleason grade groups determined based on 3D visualizations, etc. All the data are from the 50 patient cases that are studied in the paper: Xie et al., Cancer Research, 2022.
This dataset will provide public access to a 3D pathology dataset of prostate cancer, allowing researchers to investigate further on various 3D tissue structures revealed by the H&E analog images of prostate biopsies, and their correlation with prostate cancer disease progression and patient outcome. This will promote more research in the field of computational 3D pathology for clinical decision support.
Additional Information: "The 3D pathology imaging data was generated in our lab with a custom light-sheet microscope, with no clinical metadata within the imaging files. All of the clinical outcomes data that we will provide have already been published in tables within the supplement of our Cancer Research paper. All cases are referred to based on a coded identifier only, so there should be no issues with de-identification."
Acknowledgements
We would like to acknowledge the individuals and institutions that have provided data for this collection:
Research Program (PCRP) through W81XWH-18-10358 (J.T.C. Liu, L.D. True and J.C. Vaughan), W81XWH-19-1-0589 (N.P. Reder), W81XWH-15-1-0558 (A. Madabhushi) and W81XWH-20-1-0851 (A. Madabhushi and J.T.C. Liu).
Support was also provided by the National Cancer Institute (NCI) through K99 CA240681 (A.K. Glaser), R01CA244170 (J.T.C. Liu), U24CA199374 (A. Madabhushi), R01CA249992 (A. Madabhushi), R01CA202752 (A. Madabhushi), R01CA208236 (A. Madabhushi), R01CA216579 (A. Madabhushi), R01CA220581 (A. Madabhushi), R01CA257612 (A. Madabhushi), U01CA239055 (A. Madabhushi), U01CA248226 (A. Madabhushi), and U54CA254566 (A. Madabhushi).
Additional support was provided by the National Heart, Lung and Blood Institute (NHLBI) through R01HL151277 (A. Madabhushi), the National Institute of Biomedical Imaging and Bioengineering (NIBIB) through R01EB031002 (J.T.C. Liu) and R43EB028736 (A. Madabhushi), the National Institute of Mental Health through R01MH115767 (J.C. Vaughan), the VA Merit Review Award IBX004121A from the United States Department of Veterans Affairs (A. Madabhushi), the National Science Foundation (NSF) 1934292 HDR: I-DIRSE-FW (J.T.C. Liu), the NSF Graduate Research Fellowships DGE-1762114 (K.W. Bishop) and DGE-1762114 (L. Barner), the Nancy and Buster Alvord Endowment (C.D. Keene), and the Prostate Cancer Foundation Young Investigator Award (N.P. Reder).
The training and inference of the deep learning models were facilitated by the advanced computational, storage, and networking infrastructure provided by the Hyak supercomputer system, as funded in part by the student technology fee (STF) at the University of Washington.
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation, the National Institutes of Health, the Department of Defense, the Department of Veterans Affairs, or the United States Government.
Additional publications relating to the data
- Renier, Nicolas, et al. "iDISCO: a simple, rapid method to immunolabel large tissue samples for volume imaging." Cell 159.4 (2014): 896-910.
- Glaser, Adam K., et al. "Multi-immersion open-top light-sheet microscope for high-throughput imaging of cleared tissues." Nature communications 10.1 (2019): 1-8.
Data Access
| Data Type | Download all or Query/Filter | License |
|---|---|---|
| Tissue Slide Images (HDF5, TIFF, XML 3.8TB) | (Download requires Aspera plugin) | |
| Clinical data (CSV) |
Click the Versions tab for more info about data releases.
Detailed Description
Image Statistics | Pathology Image Statistics |
|---|---|
Modalities | Hierarchical Data Format V5, Tiff stack |
Number of Patients | 50 |
Number of Images | 118 |
| Images Size (tB) | 3.8 |
Our 3D imaging method is entirely slide-free and non-destructive. We image intact tissue specimens that have been fluorescently labeled and optically cleared (to make them transparent to light). We do not use any commercial scanners since the imaging is all done with our custom-developed in-house open-top light-sheet (OTLS) microscopes.
Our OTLS microscopes use an sCMOS camera to collect images of optically cleared tissues in a slice-by-slice manner as we translate the specimens through an illumination “light sheet.” The sCMOS cameras generate raw 16-bit TIFF files (grayscale) that are assembled into volumetric imaging “tiles” within the RAM of the acquisition computer. These imaging tiles are like volumetric “bricks” of data, as shown in Fig. 1G of this paper in Nature Communications. Each tile is then saved to the hard drive in the form of a multi-resolution HDF5 file. As adjacent imaging tiles are scanned within the specimen, they are appended into the same HDF5 in the hard drive.
The HDF5 file is basically like a 3D TIFF stack except it contains multiple “downsampled” versions of the dataset so that users can quickly visualize or access the 3D data at whatever resolution they need. The HDF5 file is also “chunked” so that smaller volumetric regions of interest can be quickly accessed. The HDF5 file has an associated XML file that contains microscope metadata (e.g. stage coordinates for each tile, acquisition parameters, etc.). For our project, we also used a lossless compression method, called B3D compression, to make the HDF5 files ~10X smaller than the raw TIFF stacks in RAM. We do not save the raw TIFF stacks. Therefore, the B3D-compressed HDF5 file is the “rawest” form of our imaging data.
Before performing any sort of computational analysis on the imaging data, we also need to “fuse” the 3D imaging data contained in the raw HDF5 file. This is done with an open-source program called “BigStitcher” that finely aligns all of the volumetric imaging tiles, shears the dataset, and blends the edges of those imaging tiles such that a “seamless” 3D image is generated. In this project, we also downsample the imaging data by a factor of 2X in all three dimensions (8X-reduced dataset size) during this fusion process. The fused dataset is saved as a TIFF stack so that we can analyze the datasets with our Python-based codes.
Citations & Data Usage Policy
Users must abide by the TCIA Data Usage Policy and Restrictions. Attribution should include references to the following citations:
Data Citation
DOI goes here. Create using Datacite with information from Collection Approval form
Publication Citation
Xie, W., Reder, N. P., Koyuncu, C., Leo, P., Hawley, S., Huang, H., Mao, C., Postupna, N., Kang, S., Serafin, R., Gao, G., Han, Q., Bishop, K. W., Barner, L. A., Fu, P., Wright, J. L., Keene, C. D., Vaughan, J. C., Janowczyk, A., … Liu, J. T. C. (2021). Prostate Cancer Risk Stratification via Nondestructive 3D Pathology with Deep Learning–Assisted Gland Analysis. In Cancer Research (Vol. 82, Issue 2, pp. 334–345). American Association for Cancer Research (AACR). https://doi.org/10.1158/0008-5472.can-21-2843
TCIA Citation
Clark, K., Vendt, B., Smith, K., Freymann, J., Kirby, J., Koppel, P., Moore, S., Phillips, S., Maffitt, D., Pringle, M., Tarbox, L., & Prior, F. (2013). The Cancer Imaging Archive (TCIA): Maintaining and Operating a Public Information Repository. In Journal of Digital Imaging (Vol. 26, Issue 6, pp. 1045–1057). Springer Science and Business Media LLC. https://doi.org/10.1007/s10278-013-9622-7
Other Publications Using This Data
TCIA maintains a list of publications which leverage TCIA data. If you have a manuscript you'd like to add please contact the TCIA Helpdesk.