From incoming scan to a reviewed finding.
The concept: imaging teams keep the tools they already use, and Lucentaa adds one step between upload and sign-off, an automated first pass that a radiologist reviews.
Upload → AI Analysis → Clinical Review
In this design, every study takes the same path, and a radiologist has the final say.
Upload
The plan is for users to upload DICOM files or connect a PACS, with identifying details removed and the images normalized before analysis starts.
AI Analysis
The AI models would look for candidate findings, mark the region on the image, and give each one a confidence score.
Clinical Review
A radiologist would go through each flagged finding in the viewer and confirm or dismiss it. Nothing would reach a report without that sign-off.
Where it applies.
We are looking first at the study types that fill most of a reading room's day. None of these use cases has been clinically tested yet.
Chest & Musculoskeletal X-Ray
Chest and bone X-rays. The model would mark regions a radiologist may want to look at a second time.
CT Triage Support
Incoming CT studies would get a quick first pass, so a scan with a possible urgent finding could move up the worklist.
MRI Series Review
An MRI study can have dozens of series. The goal is to list flagged findings with the slices and sequences they appear on, so there is less scrolling to find them.
Structured Reporting Assist
Once a radiologist confirmed a finding, Lucentaa could fill in the matching fields of a structured report. Every field would stay editable.
What the review screen looks like.
This is a mockup with made-up sample data. It does not show a real patient or real model output.
Modality: CT
Patient ID: ****3821 (demo)
Series: 3 / Slices: 214
Want to help shape it?
We are seeking pilot partners among healthcare providers.