Our approach
Scanvero is built around the question a follow-up read actually asks: what changed since last time, lesion by lesion, and can you trust the answer?
Follow-up first
The unit of work is the patient's sequence of scans, not a single image. Each lesion keeps its identity from one scan to the next.
Refuse rather than guess
An unverified result is shown as "not available" with its reason. A missing number is safer than a plausible wrong one.
The radiologist decides
Scanvero proposes outlines and links. A person confirms each link before any change is reported.
Traceable to the image
Every measurement points back to the slices it came from, and every decision is written to a tamper-evident audit trail.
Evidence held to its word
Pass marks are written down before a test is run, and held-out data stays sealed until then. We publish what failed as well as what passed.
Fits the reading room
Standard DICOM CT in a web browser, with no workstation to install.
Follow-up today, and what we are building.
The right-hand column describes what Scanvero is designed to do. It is research software and has not yet been validated in clinical use.
| Step | By hand today | Scanvero, as designed |
|---|---|---|
| Find lesions | Scroll through each new scan and look again. | Liver and lesions outlined on every scan. |
| Match across scans | Work out by eye which lesion is which. | Links proposed when the two scans align; the radiologist confirms each one. |
| Measure | Pick a slice and an axis and measure by hand. | Size and volume measured from the outline. |
| Report change | Calculate per lesion and in total. | Change shown only for confirmed links; otherwise "not available" and why. |
| Record | Free-text report. | Every confirmation and correction in an audit trail. |
Where we are strong, and where we are not yet.
- Liver outlines that match reference outlines closely on external colorectal CT.
- Reliable detection of lesions over 20 mm.
- Lesion matching that refuses when two scans do not align, rather than guessing.
- Smaller lesions and false positives.
- Agreement with radiologists' own lesion measurements.
- Validation on real paired NHS follow-up scans.
Work with us
We are looking for hospitals and research groups with baseline and follow-up CT of colorectal liver metastases, radiologists for a reader study, and academic and NHS partners for funded validation.
Please do not email patient images or patient information.