AI System for Prechecking Medical Image Quality

The problem

In day-to-day operations, technical issues may not be noticed until the reporting stage. When DICOM metadata, series lists, device records and quality signals are scattered across different screens, documents or team records, problems do not become visible in time.

Fragmented data delays preparation of the technical quality precheck report and forces staff to search manually for supporting evidence again. Before making a decision, the radiographer and radiologist have to complete missing fields and verify inconsistencies one by one.

A common exception is that a protocol may differ from the standard series for a clinical reason. Incomplete or incorrect automation that overlooks this context could lead to an unnecessary repeat scan or rejection of a suitable image, so the output cannot be applied directly.

What we set out to improve

• Rate of technical quality alerts confirmed by the specialist
• Rate of outputs corrected or rejected by the radiographer and radiologist
• Time between the source record and the review output

How the system works

1. 01 Define the scope of the review and the authorised radiographer and radiologist roles
2. 02 Collect the inputs: DICOM metadata, series lists, device records and quality signals
3. 03 Check source identity, date, version and data freshness
4. 04 Check file and series integrity
5. 05 Compare quality signals with the protocol
6. 06 Check the exception: a protocol may differ from the standard series for a clinical reason
7. 07 Place low-confidence or inconsistent records in a separate human review queue
8. 08 Write the source identity, rule, output and time to the audit trail
9. 09 Have the radiographer and radiologist approve, correct, defer or reject the output

Methods we used

• DICOM rule checks and preliminary technical quality classification
• Source, format, version and required-field checks for DICOM metadata, series lists, device records and quality signals
• A missing-data warning instead of an estimate when inputs are missing or inconsistent
• Checking exceptions separately from the main rule
• Storing the source identity, rule result, technical quality precheck report and human decision together

Where people stay involved

Data and security

Access to DICOM metadata, series lists, device records and quality signals is limited to the minimum permissions required for each role. Identifiers and sensitive fields are masked where possible, and only the necessary section is sent to the model rather than the entire raw file. Access, outputs and decisions by the radiographer and radiologist are recorded; the system does not enter production until the organisation has approved the retention period and data location.

Who this suits

Frequently asked questions