How to Automate Visage Imaging

Visage 7 puts every prior on the screen at once

Visage Imaging builds Visage 7, a cloud-engineered enterprise imaging platform. The diagnostic viewer sits on top of Visage 7 Open Archive, a standards-based VNA, with Visage Ease Pro for iPad and iPhone and Visage 7 Reporting for the dictation. Academic medical centres, cancer centres, health systems and veterans' hospitals run on it, on their own servers or as CloudPACS.

The thing the platform is built around is speed. Subsecond access to the current study and every relevant prior, with reconstructions rendered in real time rather than pulled from a cache, so a radiologist opens a case and the whole imaging jacket is already assembled.

Not everyone touching it is a radiologist. A technologist runs the exam. A PACS administrator keeps the routing and the protocols honest. An oncologist, an orthopaedic surgeon and a referring clinician all open the same One Viewer with role-based access, and increasingly they open it on an iPad in a corridor rather than at a workstation.

Half the studies a reading list needs are at another hospital

A patient is transferred in overnight with a CT already done somewhere else. That study is the one the radiologist needs, and it is on the sending hospital's system, on an image exchange network, or on a disc in a transfer envelope.

So somebody in the department spends the morning asking for it. Then somebody watches for the report going back out to the referring practice, and takes the call when it did not arrive. Somebody keeps the folder of accreditation evidence for each modality, with a submission window that opens once and closes on a date nobody put in a calendar. Somebody logs into a scanner vendor's service site to find out whether the fault reported on Friday has an engineer against it, because Tuesday's schedule depends on the answer.

None of that is reading. All of it decides whether the reading happens on time, and none of it sits in Visage 7.

Visage Imaging

Outside priors asked for before the appointment Tomorrow's transfers checked against the sending hospital's exchange portal overnight, so a study is requested before the patient arrives rather than after.

Accreditation evidence gathered as it accrues What each modality's submission still needs, named while the window is open rather than found in the last week before it closes.

Scanner faults with an engineer against them Vendor service sites read each morning, so a fault with nobody booked is visible while tomorrow's list can still be moved.

Speed was the hard problem, and Visage solved it

Visage has automated most of what happens between a study arriving and a report going out, and a department should exhaust that before adding anything.

One Viewer gives every role the same client on Windows and macOS, and hanging protocols display multi-dimensional, multi-modality imaging without anyone rearranging panes. Auto Advance moves the reader down the list, and Link Related Orders joins the separate foot, knee and leg orders that came off a single trauma scan. Visage 7 Reporting pulls measurements, annotations, lesion tracking and AI inference results into the report itself, with Draft From Prior, Revise and Auto Impression cutting the dictation and a Lesion tool card for oncology follow-up. Open Archive sits underneath as the VNA, handling Imaging Object Change Management, encounter-based imaging capture and managed migration onto read-only storage.

Every one of those acts on imaging the platform already holds, over its own standards-based interfaces.

The other hospital's exchange portal is not one of those. Nor is the accrediting body's submission site, the scanner vendor's service portal, or the referring practice's own system. Each has a login, a layout, and a person who reads it.

Outside studies that arrive before the patient does

Give the same reading to the sites the department already has accounts on, and the morning stops beginning with a phone call.

Tomorrow's transfers checked overnight against the sending hospital's exchange portal, with the studies not yet released listed first, so the request goes in before the patient is on the table rather than after. A disc posted three days ago traced to whether it actually arrived.

Accreditation evidence for each modality gathered as the submission window opens, with what is still missing named rather than discovered in the final week. Scanner faults read off the vendor's service site each morning, and the ones with no engineer booked flagged while Tuesday's list can still be moved. Reports confirmed as received at the referring practice, so the ones that silently failed are a list at nine rather than a complaint at four. Follow-up recommendations from last month's reports matched against whether anything is actually booked.

No new interface has to exist for any of that. The pages carry the answer already, in plain words, to anyone who opens them.

The boundary in imaging is not a preference. This gathers, compares and flags. It makes no clinical judgement and reads no image diagnostically, and anything that would change a study record or reach a patient is prepared for a qualified person to check and release.

Chasing an outside study is not a clinical skill

It does need somebody who can sign in, read carefully and notice a date. That describes a browser as well as it describes a coordinator.

WebRun is an AI agent that works a real Chrome browser, signed in the way your staff sign in. It opens the exchange portal, the vendor's service site or the submission page, reads what is on the screen, and brings the answer back with the dates attached.

It runs on your schedule in your own private environment, sessions are not shared between tools, and a workflow can be locked to an explicit list of sites. You can watch a run and stop it part way through. Anything irreversible drafts and waits.

Each workflow below names the sites it opens and what it brings back.