DICOM Sample Catalog
Free Sample DICOM Files
Curated public DICOM studies covering 6 modalities — CT, MR, mammography, X-ray, PET, ultrasound. All CC-BY licensed (commercial reuse OK with attribution), sourced from NCI Imaging Data Commons. No signup. Click Open in viewer on any tile to load the study in our browser-based viewer.
Browse the catalog
6 representative studies across 6 modalities. Every tile deep-links into the viewer with the study pre-selected on the worklist — one click and you’re reviewing real DICOM data.
Chest CT — NLST Lung Cancer Screening
Chest CT of Chest / Lungs.Single-volume thin-slice chest CT from the NLST low-dose screening cohort. Demonstrates volumetric viewing, MPR, and window/level controls.
Breast MR — I-SPY 1 Trial (with segmentations)
Breast MR of Breast.Dynamic-contrast breast MRI from the I-SPY 1 trial, including pre- and post-contrast volumes plus tissue + tumor segmentation overlays. Demonstrates multi-series viewing and SEG rendering.
Source: ISPY1
Mammography — CBIS-DDSM Mass Test Set
Mammography of Breast.Standardized digital mammography (CC view) from the CBIS-DDSM benchmark set. Demonstrates 2D high-resolution viewing, zoom, and pan.
Chest X-ray — MIDRC RICORD COVID-19
Chest X-ray of Chest.Digital chest radiograph from the MIDRC RICORD COVID-19 imaging dataset, expert-annotated for COVID-19 findings. Demonstrates single-frame viewing on chest plain film.
Source: MIDRC RICORD-1C
Breast PET — QIN-Breast Quantitative Imaging
PET of Breast / Whole Body.FDG-PET imaging from the Quantitative Imaging Network breast cohort. Demonstrates SUV viewing and PET fusion with co-registered anatomical MR.
Source: QIN-Breast
Prostate Ultrasound — MRI-Targeted Biopsy
Ultrasound of Prostate.Trans-rectal ultrasound from a multi-modal MR/US fusion biopsy cohort. Demonstrates ultrasound rendering and the multi-modality workflow of correlating MR + US imaging.
Source: Prostate-MRI-US-Biopsy
About this catalog
We host this catalog as a discovery layer over NCI Imaging Data Commons, not as a re-publisher. The actual DICOM bytes stream from IDC’s public DICOMweb proxy at request-time — we don’t cache or rehost any imaging data on saga-it.com. The catalog metadata in this page (study UIDs, descriptions, attributions) is curated by hand to surface clinically interesting samples that demonstrate the viewer’s capabilities across modalities.
The IDC proxy has a per-IP daily quota and reduced performance vs. direct Google Cloud Healthcare API access (see IDC’s proxy policy). For research workloads that exceed casual viewing, query IDC via their authenticated DICOMweb endpoint directly.
Common Questions
Every sample in the catalog is sourced from NCI Imaging Data Commons (IDC), a publicly funded cancer-imaging archive that aggregates collections from The Cancer Imaging Archive (TCIA) and other research consortia. We curate a representative subset and surface them through Saga IT’s browser-based DICOM viewer.
Yes — every sample in this catalog is CC-BY 4.0 licensed, which permits commercial use with attribution. We deliberately exclude CC-NC (non-commercial) collections from the curated catalog to avoid licensing ambiguity. If you use a sample in a commercial product, cite the original collection per the attribution shown on each tile.
Each tile shows the canonical citation for its source collection (DOI + author list when available). For research papers, also cite the IDC platform itself: Fedorov et al., NCI Imaging Data Commons. Cancer Research, 2021. DOI: 10.1158/0008-5472.CAN-21-0950.
TCIA (The Cancer Imaging Archive) is the underlying repository that hosts most cancer-imaging collections. IDC (Imaging Data Commons) is NCI’s cloud-native interface that ingests TCIA collections and exposes them via a unified DICOMweb API. Our viewer queries IDC’s public proxy, which means you get TCIA content without managing TCIA-specific access workflows.
Yes — every collection in IDC is de-identified per HIPAA Safe Harbor before it enters the archive. Patient names and identifiers visible in the metadata are randomized study IDs (e.g. 172205^LSS, UPENN-GBM-00603), not real patient data. Our viewer also redacts any DICOM-tag values that match PHI patterns before they leave the browser via observability beacons (see the viewer page for the PHI-redaction pipeline detail).
Single-frame studies (X-ray, mammography) are typically 1–15 MB. Multi-frame volumetric studies (CT, MR, PET) range from 30 MB to several hundred MB depending on slice count and resolution. The viewer streams data lazily via WADO-RS, so you don’t download the full study upfront — only what’s needed to render the visible slice.
No — everything runs in your browser. The Saga DICOM viewer is a static SPA built on OHIF v3 with cornerstone3D rendering. WebGL handles the imaging math; WebAssembly handles DICOM transfer-syntax decoding. Works on Chrome, Firefox, Safari, and Edge on desktop and mobile.
Beyond IDC, notable open repositories include TCIA directly (NBIA REST, no DICOMweb), NBIA (TCIA’s data portal), the NIH Chest X-ray dataset, Stanford MURA (musculoskeletal X-ray), and MIMIC-CXR. For the OHIF viewer specifically, you can also point it at the OHIF demo server — we ship that as a secondary data source in our viewer too.