Inside Virtua Health's Pipeline For Anonymizing Trial Imaging Data
By John Oncea, Chief Editor, Clinical Tech Leader
Every imaging study that leaves Virtua Health for a clinical trial – a cardiac CT, an ultrasound, an X-ray – starts as a piece of protected health information tagged with a patient’s name, birthday, medical record number, address, and phone number. Before it can go anywhere near a research vendor’s upload portal, all of that has to come off, cleanly and completely, without corrupting the image itself.
That process – not a single tool, but a full operational pipeline spanning IT, clinical operations, and a biweekly privacy committee – is what Laura Patrone, who manages imaging informatics for cardiology and neurology at Virtua Health in South Jersey, has spent years building out. It’s also a piece of the job that rarely makes it into conversations about hospital AI and imaging technology, which tend to focus on diagnostic tools rather than the research pipeline running quietly alongside them.
The Request Doesn’t Start With It
At Virtua, a clinical trial doesn’t begin with a systems decision; it begins with a clinical one. “The request initiates from an operational department,” Patrone said. An orthopedic team wants to take part in a study; a cardiology group is approached by a trial sponsor. That department puts Patrone’s team in contact with the vendor, and from there, two parallel reviews start: an IT evaluation and questionnaire, and a separate review by Virtua’s safety and privacy committee, which meets every other Monday and includes the CISO, the cybersecurity officer, and the CIO.
Nothing gets uploaded to a research system – no images, no workflow conversations with the vendor – until both approvals clear. “We vet each system before we upload into any system,” Patrone said. “We make sure we go through a secure process to make sure that our patient’s data is safe.”
What “Anonymized” Actually Requires
The technical core of the work is stripping every imaging file of the metadata it’s automatically tagged with on capture – name, address, phone number, MRN, birthday, gender – before it’s uploaded to a research platform. Patrone’s team has run this process across a range of active studies: cardiology cases tied to hypertension and cardiovascular disease research, an orthopedic study using CT scans of cadaver femurs, and imaging support for Virtua’s renal denervation program, where the health system was the first in New Jersey to offer the procedure, in a trial led by principal investigator Dr. Kintur Sanghvi.
Virtua is also currently supporting the PEERLESS II pulmonary embolism study, transplant-related research out of Virtua Our Lady of Lourdes, the only solid-organ transplant program in South Jersey, and, through a joint venture with Penn Medicine at the region’s only Proton Center, research into surgical and radiation treatment for gynecological cancer.
One failure point Patrone’s team ran into wasn’t in the IT stack at all; it was in the imaging equipment itself. “On ultrasounds, the demographics can be burned in on the images, and you can’t remove them,” she said. Because that patient information is written directly into the pixel data rather than stored as separate, strippable metadata, standard anonymization software couldn’t touch it. The fix required going back to the ultrasound machines and changing how the fields were captured in the first place, so the demographic data stayed editable instead of permanently embedded, a hardware- and workflow-level change made specifically to support research uploads, not clinical care.
The Research Team Had To Learn IT’s Discipline
The other gap wasn’t technical; it was cultural. Virtua’s research staff, Patrone said, weren’t used to the discipline anonymization requires. “We really had to do some training with them,” she said, “as well as investigate some software to really help their team have a seamless workflow so that they’re doing the same thing when they’re anonymizing images.” Multiple study types – ultrasound, CT, X-ray – can be in play depending on the trial, and each modality carries its own risk of exposed data if the process isn’t followed consistently.
That combination – standardized software tooling plus hands-on training for a clinical research staff that had never had to think like a data-privacy team – is what closed the gap, rather than any single anonymization product.
Why The Governance Structure Matters More Than The Tooling
It would be easy to read this as a story about anonymization software. It isn’t. The software is replaceable; the governance structure around it is what makes the pipeline trustworthy at scale. Every research request passes through the same two gates regardless of which department originated it or which vendor is involved – an IT security and technical review, then a privacy and safety committee that includes the organization’s top security leadership. No trial, however clinically promising, bypasses that sequence.
For sponsors and CROs working with sites, that structure is the actual story: a repeatable process that vets both the technology and the data-handling risk before a single image moves, backed by a research staff that had to be trained into a new discipline rather than assumed to already have it. That’s the kind of infrastructure that doesn’t show up in a study’s published results but determines whether a site can take on the next trial without a fire drill.
It’s also infrastructure that scales in a way ad hoc anonymization work doesn’t. Because the IT-review-then-privacy-committee sequence applies uniformly across cardiology, orthopedics, and transplant research alike, adding a new trial doesn’t mean building a new process; it means running an existing one against a new vendor and a new imaging modality. That’s part of why Virtua has been able to stack multiple concurrent studies, from renal denervation to pulmonary embolism to gynecological cancer research, without each one requiring its own bespoke data-handling solution. The consistency is the point: a site that can show sponsors a mature, repeatable anonymization pipeline is a site that’s easier to bring onto the next protocol.