Clinical imaging sample
Record the full collection counts separately from the small review batch. Ask the buyer to confirm the number of sample studies per cohort before export.
What to collect
- Cohort table
- For a brain CT brief, confirm non-contrast acquisition, slices up to 3 mm and both smooth and sharp kernels. For plain brain MRI, confirm slices up to 5 mm. For mammography, confirm DR acquisition and the requested BI-RADS 4, 5 or 6 assessment when applicable.
- Studies and reports
- Pair the de-identified DICOM series with the corresponding reviewed report using a stable study alias. Include EMR context only if explicitly approved and requested.
- Acquisition metadata
- List modality, series, slice thickness or acquisition type as applicable, study count and report availability. Mark unknowns.
Work with your institution’s authorised data team. Keep identifiers, unreviewed images, raw clinical notes and re-identification keys out of the shared package.
Export from your systems
Pack and check
Put the README and manifest at the top of the folder, with a subfolder for each system. The manifest is a file index a reviewer can read without opening every file.
Download blank manifest (.csv) ↓Example manifest entry
- File in the package
imaging/STUDY-07/series-01/image-001.dcm- Matching key
study_alias = STUDY-07- Related record
- reports/STUDY-07.txt describes the same study
Includes all systems for this sample type and your current checks.
Share through your private workspace.
Once Maryah confirms the scope and permission to share, open the matching dataset and choose Samples. Follow the size and format limits shown there.
Open my datasetsFor a larger package or help locating your dataset, contact Maryah for a private transfer route. Include your dataset reference and file size.