Full-Study Imaging Application & AIR


AIR is a self-service platform which enables Automated Image Retrieval (AIR) from UCSF's clinical and research picture archiving and communication system (PACS). Investigators requiring direct access to image data files (i.e. DICOM files and services formerly proved by the Radiology Film Library) must request AIR access.

  1. Please review the AIR Overview prior to requesting AIR access.
  2. Individuals performing image tansfers for clinical trials (e.g. HDFCCC study coordinators) are highly encouraged to review Imaging Data 101.
  3. To request and/or update access to AIR, please complete the AIR Access Request Form.


MD.ai annotator is a secure web-based collaborative annotation application, that can be utilized for machine learning and other projects. Groups of any size can work together separately or in teams to annotate datasets. It allows for textual (global) and pixel level annotations. 

Images can be directly uploaded to MD.ai platform via a web browser as well as AIR (Automated Image Retrieval), a self-service platform for retrieving imaging studies. If you use AIR to upload images, you will need to have AIR access.  

If your research absolutely requires working with identified data and this use is approved by the IRB, our MD.ai instance is HIPAA-compliant.  However, we strongly recommend working with de-identified images if at all possible to minimize risk and consequences of PHI-disclosure. The PI is responsible for limiting access to the MD.ai project to the people listed on the IRB. 

  1. For MD.ai account as well as to get your project started, please complete the MD.ai access request form.
  2. Here is a MD.ai presentation to familiarize users of the platform
  3. Useful links:
    1. MD.ai Documentation (under API, Python Client Library)
    2. MD.ai GitHub


mPower is a search engine for Radiology (and soon, Pathology) reports. It supports searching based on patients and study metadata as well as the content of the report. Advanced search options include Boolean logic in searches, excluding negated terms and identifying patients who have results in two different studies. there are limited facilities for analytics of search results, and search results can be exported in excel format.

mPower is useful particularly for cohort selection in retrospective research studies, quality improvement projects and identification of teaching cases. Research use requires appropriate IRB/CHR approval, and is the responsibility of the researcher.

  1. To request and/or update access to mPower, please complete the mPower Request Form.
  2. Please review mPower Help for instructions on how to use mPower.

Feasibility Analysis (Optional)

Individuals lacking advanced training in radiologic or biomedical imaging sciences are encouraged to submit an Imaging Research Proposal Feasibility Analysis Request (~2 minutes) prior to submitting a full study application. This request will be routed to relevant senior investigators who will provide feedback to the submitter. 

Full Study Application

The Department of Radiology and Biomedical Imaging requires investigators to submit a Full-Study Imaging Application for review and approval before an imaging study may begin:

  1. Application to Perform Imaging Research
  2. Cancer Center Application to Perform Imaging Research (This is a dedicated online application cycle for Cancer Center requests for imaging (scan time) or imaging-related services)