Skip to content
Open Imaging Index
CT MRI PET Pathology slides Ultrasound Pancreas

CPTAC-PDA

The Clinical Proteomic Tumor Analysis Consortium Pancreatic Ductal Adenocarcinoma Collection (CPTAC-PDA)

CT, MRI, PET, ultrasound and pathology slide images of patients with pancreatic ductal adenocarcinoma, linked to proteomic, genomic and clinical data of the same people, so that imaging can be studied next to molecular profiles. 168 patients from the NCI CPTAC cohort.

Go to dataset Edit Report

Overview

CPTAC-PDA is the imaging arm of the pancreatic ductal adenocarcinoma cohort of the National Cancer Institute's Clinical Proteomic Tumor Analysis Consortium. The Cancer Imaging Archive gathers radiology and pathology images of CPTAC patients so that image phenotypes can be studied alongside the proteomic, genomic and clinical data of the same people, which are held in the Proteomic Data Commons and Genomic Data Commons under matching patient IDs. The collection was first released in January 2018 and reached version 15 in February 2025.

Composition

The collection covers 168 subjects. All 168 have pathology, with 557 whole-slide images in SVS format. Radiology is available for 110 of them: 134 studies, 1,133 DICOM series and 132,852 images. In the radiology metadata, 99 subjects have CT, 17 MRI, 3 ultrasound and 2 PET. The radiology subjects are 59 male, 50 female and 1 recorded as other. A "Discovery Cohort" subset, tied to the consortium's proteogenomic study of the disease, can be downloaded on its own. Total download size is 155.24 GB.

Acquisition

Radiology comes from routine clinical care shortly before the pathological diagnosis, with follow-up scans where available, so scanners and protocols vary. The radiology metadata lists GE, Siemens, Toshiba and Philips systems; body regions range from the pancreas alone to chest, abdomen and pelvis. Slides come from the CPTAC tissue qualification workflow. Contributing sites include hospitals and biobanks in the United States, Canada and Poland. DICOM dates are shifted, and DICOM tag (0012,0050) stores each scan's offset in days from the pathological diagnosis.

Annotations

The collection has no image labels. Tumor annotations are published separately as the TCIA analysis result CPTAC-PDA-Tumor-Annotations, and the collection is also used by the Crowds-Cure-2018 and SAROS analysis results.

Known limitations

  • About a third of the subjects have pathology slides only and no radiology.
  • Imaging is heterogeneous in modality, vendor, protocol and number of follow-up scans per patient.
  • Clinical and molecular data are not part of the TCIA download and must be exported from the data commons.
  • The collection page labels version 15 CC BY 4.0, while earlier versions and the radiology metadata file name CC BY 3.0.

Cohort

Aggregate numbers from the sources below. Bars are relative to the 168 subjects.

Sex

  • Female 50 45%
  • Male 59 54%
  • Other 1 1%

Covers 110 of 168 subjects.

Patient ID, Patient Sex columns in TCIA CPTAC-PDA radiology metadata digest (TCIA-CPTAC-PDA_v15_20250226-nbia-digest-1.xlsx, open download, CC BY 3.0 per its License Name column), counted per unique Patient ID

Scanner vendor

  • GE HealthCare 63 38%
  • Siemens Healthineers 20 12%
  • Canon Medical 7 4%
  • Philips 3 2%

Patient ID, Manufacturer columns in TCIA CPTAC-PDA radiology metadata digest (TCIA-CPTAC-PDA_v15_20250226-nbia-digest-1.xlsx, open download, CC BY 3.0 per its License Name column), counted per unique Patient ID

Modality

  • Pathology slides 168 100%
  • CT 99 59%
  • MRI 17 10%
  • Ultrasound 3 2%
  • PET 2 1%

Data Access (Version 15) in TCIA CPTAC-PDA collection page (version 15, updated 2025/02/26); Patient ID, Modality columns in TCIA CPTAC-PDA radiology metadata digest (TCIA-CPTAC-PDA_v15_20250226-nbia-digest-1.xlsx, open download, CC BY 3.0 per its License Name column), counted per unique Patient ID

License and access

Our reading of the license, not legal advice. Before you use the data, read the original license and confirm that your use is allowed. We take no responsibility for how you use a dataset. Full disclaimer

Access
Open download

Download without an account

DICOM radiology is downloaded with the TCIA Data Retriever from a manifest file, slides through IBM Aspera. No account is needed. Clinical, proteomic and genomic data are in the Proteomic and Genomic Data Commons.

Access page

Creative Commons Attribution 4.0 International

Use, share and adapt the data for any purpose, including commercial use, as long as you credit the creators.

Original license text Version read: 4.0 Checked 2026-10-07

What you can do

  • Yes
  • Yes
  • Yes
  • Yes
  • Yes

What you can share

  • Yes
  • Yes
  • Yes

What you must do

  • Yes
  • Share alike No
  • No
  • No
  • Manuscript review No
  • Release code No
  • Return results No
  • Delete after use No

Limits

  • No
  • Location limits No

Citation

National Cancer Institute Clinical Proteomic Tumor Analysis Consortium (CPTAC). (2018). The Clinical Proteomic Tumor Analysis Consortium Pancreatic Ductal Adenocarcinoma Collection (CPTAC-PDA) (Version 15) [dataset]. The Cancer Imaging Archive. https://doi.org/10.7937/k9/tcia.2018.sc20fo18. Acknowledgement: "Data used in this publication were generated by the National Cancer Institute Clinical Proteomic Tumor Analysis Consortium (CPTAC)."

Sources

Every number on this page comes from one of these documents. Each chart names the table or page it is taken from. The raw numbers are in stats.csv.