APEC Vox QCA
Fast, transparent, reproducible
coronary quantification.
AI-assisted QCA on a standard hospital computer.
Research and validation platform. Not a certified medical device; not for clinical use. Intended use
How it works
From angiogram to measurement in four steps.
- 01
Open the study
A DICOM cine, a CD export or a single image. The end-diastolic frame is picked automatically.
- 02
Run the analysis
Vessel and stenosis segmentation in seconds, on the local CPU. No GPU.
- 03
Two-point measurement
Mark the start and the end. Diameters, MLD, lesion length and %stenosis follow.
- 04
Export
One CSV per session, with the masks and how every number was produced.
Numbers
Performance, measured.
- Dice, vessel segmentation
- 0.80 Dice, vessel segmentation ARCADE test set, n=300. Inter-observer: 0.81.
- clDice, vessel centreline
- 0.81 clDice, vessel centreline Centreline topology.
- Per frame, CPU only
- ~2.4s Per frame, CPU only Full cascade, no GPU.
- Images sent off the computer
- 0 Images sent off the computer All inference is local.
Support and credentials
Supported by Scottish Enterprise – SMART: SCOTLAND 18-month feasibility study, awarded April 2025.- Recognition Women's Enterprise Scotland Awards 2026 Finalist, Innovation Business of the Year.
- ESC Digital & AI Summit Abstracts in 2025 and 2026 Presented in Berlin, 2025. Accepted for Basel, November 2026.
- Intellectual property Two UK patent applications GB2416340.4 (November 2024) and GB2613265.4 (June 2026).
- Clinical collaboration Barts Health NHS Trust London.
Pilot programme and partnerships
We are looking for pilot sites.
Run APEC Vox QCA on one or two lab PCs, alongside your normal workflow, under a research-use agreement.