3DCPath-datasets is a compilation of publicly available datasets in 3D Computational Pathology (3D CPath), spanning morphological data (3D tissue images) and molecular 3D data. It is organized into three main categories: volumetric morphological/imaging data, volumetric genomic data, and combined volumetric morphological and genomic data. The goal of this repository is to accelerate research in 3D CPath and provide the community with a centralized place to access 3D pathology datasets for future studies.
Note
Contributions are welcome! Please check the contributions section below!
- 08/25/26: 3DCPath-datasets is now live and will be continuously updated as new datasets become available.
| Added | Venue / Date | Dataset | Title | Modality | Resolution | Organ | Volumes | Size |
|---|---|---|---|---|---|---|---|---|
| 08/25/26 | Nature Biomedical Engineering (2026) |
TRICARE [Paper] |
Deep-learning triage of three-dimensional pathology datasets for comprehensive and efficient pathologist assessments | Open-Top Light-Sheet Microscopy (OTLS) | Prostate: ~1 µm lateral, ~3.5–4 µm axial; Esophagus: ~0.6 µm lateral, ~2.75–3 µm axial | Prostate cancer biopsies and endoscopic biopsies of Barrett’s esophagus | Esophageal: 29 patients (125 specimens); Prostate: 59 patients (59 specimens) | 191.3GB |
| 08/25/26 | Cell (2026) |
ULTRA [Paper] |
Ultrarapid deep 3D histology enables intraoperative mapping of glioma infiltration | Volumetric stimulated Raman scattering (SRS) lipid-channel images using the ULTRA pipeline | Lateral resolution: 0.414 μm/pixel, axial resolution: 1 μm | Three major glioma subtypes: oligodendroglioma, astrocytoma, and glioblastoma | 3 volumes with over a depth of 1 mm | 9.4GB |
| 07/15/26 | Royal Society Interface (2026) |
VISTACT [Paper] |
VIrtual histological STAining of micro-Computed Tomography (VISTACT) | Synchtron-based phase-contrast microCT and serial H&E and EvG-stained histological sections | Human lung microCT scans: 1.63 μm/voxel ; Mouse lung microCT scan: 0.88 μm/voxel | Human lung tissue from patients with pulmonary hypertension and mice heart-lung tissue | 10 from human tissue (4 patients), and 1 from mice tissue | 869GB |
| 08/25/26 | Cell Press Blue (2026) |
CODA matrices 1 CODA matrices 2 [Paper] |
Quantitative 3D histology reveals localized immune remodeling during early pancreatic cancer progression | CODA-derived 3D tissue and cell matrices from serial H&E histology, with IHC-derived immune-cell information for selected samples | Released matrices: 12 × 12 × 12 µm³/voxel; underlying CODA tissue segmentation generated at 2 µm/pixel | Human pancreas | 48 samples / 48 volumes; two samples additionally have IHC-enhanced matrix versions | 99.7 GB |
| 07/15/26 | Nature Communications (2025) |
Holotomography [Paper] |
Revealing 3D microanatomical structures of unlabeled thick cancer tissues using holotomography and virtual H&E staining | Holotomography and H&E-virtually-stained images | Lateral resolution: 156 nm; Axial resolution: 1.07 μm | Human colon cancer samples | N/A | 24.74GB |
| 08/25/26 | eLife (2024) |
OT-TP-LSM [Paper] |
Open-top Bessel beam two-photon light sheet microscopy for three-dimensional pathology | Open-top Bessel-beam two-photon light-sheet microscopy | Isotropic 0.9 µm/voxel | Human skin—basal-cell carcinoma; pancreas—normal, IPMN/premalignant, and cancer; prostate—benign and cancer | 7 human tissue volumes: 2 skin BCC, 3 pancreas (normal, IPMN, PDAC), and 2 prostate (benign, adenocarcinoma) | 11.10 GB |
| 07/15/26 | Cancer Research(March 2023) |
PCa_Bx_3Dpathology [Paper] |
3D pathology of prostate biopsies with biochemical recurrence outcomes: raw H&E-analog datasets and image translation-assisted segmentation in 3D (ITAS3D) datasets | OTLS | 0.88 μm/pixel | Human prostate | 118 (50 patients) | 3.8TB |
| 07/15/26 | MICCAI challenge(2023) |
ACROBAT |
AutomatiC Registration Of Breast cAncer Tissue (ACROBAT) | H&E and IHC (ER, PGR, HER2, KI67) | Multiple. Validation conducted at 1.25x | Human breast cancer | 750 (1 H&E and 1-4 matched routine IHC) | n/a |
| 08/25/26 | European Journal of Cancer (2023) |
3D melanoma histology model [Paper] |
A 3-dimensional histology computer model of malignant melanoma and its implications for digital pathology | Serial H&E histological sections, reconstructed into a 3D volume using TrakEM2/Fiji | Original scans: 0.25 µm/pixel at 40×; alignment images downsampled to 1 µm/pixel; 3 µm section thickness | Human cutaneous malignant melanoma | 1 volume (66 serial H&E sections) | 732.7MB |
| 07/15/26 | Optica(December 2022) |
Data [Paper] |
Label- and slide-free tissue histology using 3D epi-mode quantitative phase imaging and virtual hematoxylin and eosin staining | 3D epi-mode quantitative phase imaging | Lateral resolution of 0.6µm and cross-sectional/axial resolution of 3.5 µm | Mouse liver, rat gliosarcoma, and human gliomas | Mouse liver: 8, rat gliosarcoma: 14, human gliomas: 5 | 26 GB |
| 07/15/26 | Cell(2022) |
3DCRC [Paper] |
Multiplexed 3D atlas of state transitions and immune interaction in colorectal cancer | H&E and t‑CyCIF serial histological sections | Tissue sections with 5μm thickness and variable spacing. Images scanned at 20x | Colorectal cancer sample | 1 (47 serial histological sections) | 2.1 TB |
| 07/15/26 | PNAS(2022) |
3d virtual histology of the human hippocampus [Paper] |
Three-dimensional virtual histology of the human hippocampus based on phase-contrast computed-tomography | Phase-contrast microCT | 160 nm/px and 50 nm/px | Human hippocampus tissue | N/A | 64.8 GB |
| 07/15/26 | eLife(2021) |
3D virtual Histopathology [Paper] |
3D virtual histopathology of cardiac tissue from Covid-19 patients based on phase-contrast X-ray tomography | Phase-contrast microCT (Histomography scanner) | 1.3µm | Cardiac Tissue from Covid-19 Patients | 6 | 175.4 GB |
| 07/15/26 | eLife(2020) |
3D virtual pathohistology [Paper] |
3D virtual pathohistology of lung tissue from Covid-19 patients | Phase-contrast microCT (Histomography scanner) | 1.3µm, 650nm, and 167nm | Human lung tissue | 6 Covid-19 patients and one control tissue (healthy lung) | 568.4GB |
| 08/25/26 | Bioinformatics (2018) |
3D histology reconstruction benchmark [Paper] |
Comparative analysis of tissue reconstruction algorithms for 3D histology | Serial H&E whole-slide images, tissue masks and manually annotated landmarks | 0.46 µm/pixel, scanned at 20×; 5 µm section spacing | Murine prostate and murine liver | 2 volumes: prostate = 260 sections; liver = 47 sections | 59.41 GB |
| Added | Journal / Date | Dataset | Title | Modality | Resolution | Organ | Volumes | Z-levels | Num. spots/cells | Size |
|---|---|---|---|---|---|---|---|---|---|---|
| 07/15/26 | Nature Methods (November 2025) |
Deep-STARmap [Paper] |
Scalable spatial single-cell transcriptomics and translatomics in 3D thick tissue blocks | 3D Spatial Transcriptomics (Deep-STARmap) and proteomics (Deep-RIBOmap) | Single-cell 3D ST/SP data | Mouse brain and human cutaneous squamous cell carcinoma (cSCC) tissue | 2 | cSCC: 60µm depth, mouse brain: 150µm depth | cSCC: N/A cells,254 genes; Mouse brain: 198,675 cells, 1,017 genes | 2.1GB |
| 07/15/26 | Nature Methods (September 2025) |
3Dniches [Paper] |
Highly multiplexed 3D profiling of cell states and immune niches in human tumors | 3D CyCIF | 140 nm × 140 nm × 280 nm voxels (200–500 voxels per cell) - 20–54-plex | Melanoma specimens: (1) preinvasive cutaneous melanoma, (2) invasive VGP primary melanoma, (3) Metastatic melanoma to the skin | 3 | 5–50-µm-thick sections | N/A | 3.17TB |
| 07/15/26 | Nature Methods (December 2025) |
SpatialZ [Paper] |
Bridging the dimensional gap from planar spatial transcriptomics to 3D cell atlases | 3D Spatial Transcriptomics with StarMAP | Single-cell 3D ST data | Mouse visual cortex | 1 | 89 | 32,845 cells | 30.5MB |
| 07/15/26 | Nature Metabolism (March 2025) |
MetaVision3D [Paper] |
AI-driven framework to map the brain metabolome in three dimensions | 3D metabolomics (MALDI) | 50 µm | Brain mice samples: two wild-type, 5xFAD, and GAA | 4 | 79 serial brain sections (10-µm-thick sections). Sections are 50 µm apart (z axis), corresponding to the MALDI imaging spatial resolution of 50 µm (x and y axes). | N/A | 3.3GB |
| 07/15/26 | Nature Cancer (January 2022) |
3D IMC [Paper] |
Three-dimensional imaging mass cytometry for highly multiplexed molecular and cellular mapping of tissues and the tumor microenvironment | 3D proteomics (IMC) | Single-cell resolution (40 markers) | HER2+ ductal breast carcinoma sample | 4 (Size of the main sample: 652×488×304µm3) | 152 consecutive slices (2-µm-thick sections) | N/A | 6.6GB |
| Added | Journal / Date | Dataset | Title | Modality | Resolution | Organ | Volumes | Serial sections | Num. spots/cells | Size |
|---|---|---|---|---|---|---|---|---|---|---|
| 08/25/26 | Cell systems (May 2025) |
3D NSCLC CosMx + ECM/SHG [Paper] |
Combining spatial transcriptomics and ECM imaging in 3D for mapping cellular interactions in the tumor microenvironment | Serial sections; CosMx spatial transcriptomics, SHG, H&E and IF | 5 µm section thickness; CosMx sections spaced 30 µm; single-cell ST | Human NSCLC lung tumor | 1 patient / 1 tumor | 34 total; 6 profiled with CosMx |
340,644 cells | 9GB |
| 07/15/26 | Nature (October 2024) |
2D3DTumorEvolution [Paper] |
Tumour evolution and microenvironment interactions in 2D and 3D space | Serial H&E sections and Visium ST | 20x,55μm-diameter ST spots | BRCA,CHOL,CRC,PDAC,RCC,UCEC | 78 patients (combined 2D and 3D) - Detailed description | 131 | 433,000 spots | N/A |
| 07/15/26 | Cell (July 2024) |
OpenST [Paper] |
Open-ST: High-resolution spatial transcriptomics in 3D | Serial H&E sections and OpenST transcriptomics | 0.345 μm/pixel, single-cell ST data | Head-and-neck metastatic lymph node | 1 (1 patient) | 19 | 1,097,769 cells | 2.2GB |
| 07/15/26 | Nature Communications (February 2022) |
3DST_RA [Paper] |
Three-dimensional spatial transcriptomics uncovers cell type localizations in the human rheumatoid arthritis synovium | Serial H&E sections and Spatial Transcriptomics | 20x, Spatial Transcriptomics technology (~100μm-diameter spots) | Synovial joints (Rheumatoid arthritis) | 6 (6 patients) | 27 (RA1: 4, RA2: 7, RA3: 4, RA4: 3, RA5: 5, RA6: 4). RA1 spacing between sections -> 21μm ; RA2~RA6: consecutive sections | 17,117 spots | 2.2GB (.gz format) |
| 07/15/26 | Nature Neuroscience (March 2021) |
Human prefrontal cortex [Paper] |
Transcriptome-scale spatial gene expression in the human dorsolateral prefrontal cortex | Serial H&E sections and Visium ST | Visium ST technology (55μm-diameter spots) | Human dorsolateral prefrontal cortex | 3 patient | 12 (4 sections for each patient; The distance between sections is 10μm - 300μm - 10μm) | 47,681 spots | N/A |
| 07/15/26 | Nature Communications (February 2021) |
HER2+ [Paper] |
Spatial deconvolution of HER2-positive breast cancer delineates tumor-associated cell type interactions | Serial H&E sections and Spatial Transcriptomics | 20x, Spatial Transcriptomics technology (~100μm-diameter spots) | HER2+ breast cancer | 8 (8 patients) | 36 (A~D:6, E~H:3). Spacing between sections: 36μm. Section thickness: 16μm | 13,299 spots | 629.6 MB |
| 07/15/26 | Cell (December 2019) |
Human Heart [Paper] |
A Spatiotemporal Organ-Wide Gene Expression and Cell Atlas of the Developing Human Heart | Serial H&E sections and Spatial Transcriptomics | 20x, Spatial Transcriptomics technology (~100μm-diameter spots) | Developing Human Heart | Four developmental heart tissues | 19 (4,9,6 sections for 4.5–5, 6.5, and 9 PCW heart tissues). Spacing between sections: 36μm. Section thickness: 5μm | 3,115 spots | 211 MB |
If you would like to add a 3D dataset to the current compilation, please create a pull request with details for each field in the tables above, or contact calmagro@mit.edu .
