> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dataerai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Import from The Cancer Imaging Archive

> Search The Cancer Imaging Archive (TCIA) for imaging collections and import them into a project as metadata-first assets — without copying the bytes into Dataerai. Each imaging series streams straight from TCIA over HTTPS.

Dataerai can pull medical imaging from [The Cancer Imaging Archive](https://www.cancerimagingarchive.net/) (TCIA), an NCI-funded archive of de-identified cancer imaging (CT, MR, PET, pathology) organized into named **collections**, into one of your projects as **metadata-first assets**. The collection's metadata — name, modality mix, and a link to the source archive — is copied into Dataerai so it's searchable and shareable alongside your own data. The **files are not copied**: clicking **Download** on an imported asset streams that imaging series (a `.zip` of DICOM images) directly from TCIA's NBIA archive over HTTPS.

<Note>
  Imported assets are [metadata-first](/data/metadata-first-assets): they describe and link to data that lives on TCIA. Nothing is stored in your Dataerai allocation, so importing doesn't consume your quota.
</Note>

One TCIA **collection** (such as `LIDC-IDRI` or `Soft-tissue-Sarcoma`) imports as a **collection of assets — one asset per imaging series** in the collection, grouped under a collection named for the TCIA collection and sub-grouped by series description and modality. Each asset's file (e.g. `<series-uid>.zip`) downloads straight from TCIA. A collection with no publicly downloadable series instead imports as a single asset that links out to the collection's [NBIA browser page](https://nbia.cancerimagingarchive.net/).

## Search The Cancer Imaging Archive

Open the **Import from The Cancer Imaging Archive** page and enter a search term, then **Search**.

* The search matches the **collection name** — try `lung`, `CPTAC`, or `breast`.
* Results show **Title**, **Type**, and **Files**, with the collection name under each title.
* Use **Per page** to trade clicks for a longer page. The page count reflects the full result set, so you can page to the last page directly.

## Select and import

<Steps>
  <Step title="Select collections">
    Tick the collections you want — click a row, or press Space or Enter on it, to toggle its checkbox. The header checkbox selects or clears everything on the current page; your selection is kept as you page through results.
  </Step>

  <Step title="Choose the destination project">
    Paste the UUID of the project that should own the new assets into the **Import into project** field.
  </Step>

  <Step title="Import">
    Click **Import**. Large selections are sent in batches automatically, and a single **Import result** panel summarizes how many collections were imported (and how many assets were created), plus any that failed and why.
  </Step>
</Steps>

<Note>
  A single collection can contain hundreds or thousands of imaging series, so importing one collection can create many assets. Very large collections import a capped subset of their series; the rest remain available directly on TCIA.
</Note>

<Warning>
  Because imported assets point at TCIA rather than your own storage, their files are only available while the TCIA series are. If you need a durable copy under your control, download the series and [upload it](/data/upload) as a regular asset instead.
</Warning>

## Next steps

<CardGroup cols={2}>
  <Card title="Metadata-first assets" icon="file-plus" href="/data/metadata-first-assets">
    How assets that describe data elsewhere work in Dataerai.
  </Card>

  <Card title="Import from Zenodo" icon="download" href="/data/import-zenodo">
    Pull public records from Zenodo the same metadata-first way.
  </Card>

  <Card title="Organize into collections" icon="folder-tree" href="/organize/collections">
    Group imported collections alongside your own data.
  </Card>

  <Card title="Search & discover" icon="search" href="/discover/search">
    Find imported collections by title, type, and metadata.
  </Card>
</CardGroup>