Standard Operating Procedure ( SOP) Summary Steps:
This Standard Operating Procedure (SOP) defines a standardized workflow for AI-integrated breast POCUS in
outpatient settings. It is designed for trained clinicians and serves as an adjunct to existing diagnostic
pathways, not a substitute for formal radiologic breast screening.
Workflow Steps:
1.
a.
b.
2.
a.
b.
3.
a.
b.
4.
Patient arrival and verification
Confirm patient identity using institutional standards.
Review clinical indications, relevant history, and any prior breast imaging.
Explain the scope and obtain consent
Clearly explain the role and limitations of breast POCUS relative to diagnostic radiology.
Answer patient questions and obtain informed consent in accordance with institutional policy.
Patient positioning and preparation
Position the patient supine or in a slight oblique position with the ipsilateral arm raised.
Drape appropriately and expose the breast region from the clavicle to the inframammary fold.
Device setup
Select the breast preset (or equivalent) on the POCUS device and confirm appropriate baseline settings prior to
scanning.
5.
Optimize image settings Adjust depth, focus, and gain so that subcutaneous fat and fibroglandular tissue
are clearly differentiated.
6.
Identify normal anatomy Confirm the expected anatomic layers and overall breast architecture.
7. Perform a systematic breast survey Scan the entire breast in a radial or clock-face pattern, including the
retro-areolar region and axillary tail.
8.
Assess axilla as indicated Perform targeted axillary scanning when clinically appropriate.
9. Document focal findings For any focal abnormality, obtain orthogonal views, measure in three dimensions,
and record clock-face position and depth, using structured documentation.
10. Capture and label images
a. Acquire representative images and cine loops of normal survey views and any abnormalities.
b. Label them according to the standardized protocol.
11. Secure transmission and storage of data to the AI platform
a.
Store images in the designated system using the approved file structure to support downstream AI training
and analysis.
b.
De-identify image sets before transmission, then transmit de-identified image sets to the approved AI
integration platform using secure, encrypted channels.
Page 667