AI-assisted imaging reads that your radiologist stays in charge of
QueroCura's imaging workspace views X-ray, CT and MRI series, ultrasound, ECG and more, and can prepare an AI draft read for a qualified clinician to review. The QueroCura Vision Engine can run on the hospital's own server, so images never leave it. It is decision support, not a diagnosis: every finding is accepted or rejected and the report signed by your clinician.
What the on-premise engine reads
| Study | What it does |
|---|---|
| Chest X-ray (frontal, adult) | Two independently trained models read for 18 findings, behind safety checks that refuse lateral views, other body parts and unusual images. Heatmap boxes show where the model looked. |
| 12-lead ECG (digital waveform) | Signal quality per lead, rhythm (including atrial fibrillation and broad-complex tachycardia), PR, QRS, QT and QTc, axis, bundle-branch blocks, and ST elevation by published criteria. |
| CT chest and head (DICOM series) | Hounsfield-unit measurements: lung nodule candidates with Fleischner follow-up, an emphysema index, and possible acute intracranial blood with midline shift. |
Built to avoid confident mistakes
- Each reader refuses rather than guesses when an image or trace is outside what it was built for, and says why.
- Can't-miss chest findings such as pneumothorax are shown as "possible" when either model flags them.
- Every model is pinned and checksummed; nothing runs from an altered file.
- Hospitals validate the models on their own labelled cases; until then every draft says it is not yet validated there.
- Clinicians' accept and reject decisions are tallied per model, with a warning if a model's findings are mostly rejected.
The radiologist decides
Patients never see an unsigned draft. AI is off until an administrator switches it on, and patients can opt out of AI help with their own scans. QueroCura's imaging AI is not a cleared medical device and can be wrong or miss things.
Frequently asked questions
Does QueroCura's AI diagnose scans?
No. It prepares a draft for a qualified clinician, who accepts or rejects every finding and signs the report. It is decision support, not a diagnosis or a cleared medical device.
Can AI imaging run without sending images to the cloud?
Yes. The QueroCura Vision Engine can run on the hospital's own server on ordinary CPUs, so images never leave the hospital.
Which scans can it read?
On-premise: frontal adult chest X-rays, digital 12-lead ECGs and CT chest or head series. Other studies are viewed and reported by your clinicians as usual.
Related
Ready to see it for your hospital?
Answer a few questions and get a setup plan sized to your facility.
See the hospital portalLast updated 03 November 2026.