A new startup emerged from stealth on Tuesday at the HLTH conference in Las Vegas. The Irvine, California-based company, named Avandra Imaging, is a federated network for de-identified imaging data.
Avandra’s main goal is to make it easier for health systems, biopharma companies and AI developers to access imaging data that will help advance their research, explained CEO and founder Ryan Tarzy.
“The challenge we have right now is a last mile problem. We have spent the last 20 years and God knows how many dollars on digitizing medical records and moving into the EHR. But right now, the clinical point of truth of imaging is what’s arguably the most valuable data. And it is just not widely available — specifically in a large enough, diverse enough dataset for research,” he remarked.
Tarzy said he often hears medical researchers and AI developers lament the fact that they can’t get their hands on imaging data.
For example, he pointed out that there were some companies attending HLTH seeking to build tools for earlier identification of Alzheimer’s or to better manage the effectiveness of Alzheimer’s therapy. To build these kinds of tools, researchers need brain imaging, not just EHR data, Tarzy noted.
“The problem is that nobody knows where those images are, even though there are millions of patients. It’s a needle in a haystack problem, and they don’t even know what haystack to look in to find the needle,” he declared.
To address this problem, Avandra has built a de-identified data network in which researchers can more easily locate the imaging they’re seeking. Having access to this data will help researchers with things like clinical trial design and algorithms training, Tarzy said.
To him, Avandra has two main competitors: Gradient Health and Segmed.
“They don’t have a federated model, and they don’t have the network that we have,” he stated.
A federated network ensures that healthcare data stays with its original owners — allowing hospitals and other providers to maintain control and comply with privacy laws while still facilitating collaborative research, Tarzy explained.
As for Avandra’s business model, the startup operates on a marketplace model.
“It’s license plus professional services and we revenue share back to the contributor,” Tarzy said.
Avandra already has some customers, but it is waiting to disclose how many and to reveal their names, he noted. He did say that the company is profitable and has 40 employees.
In addition to emerging from stealth mode, Avandra also announced a partnership with Datavant, a health data platform that connects and de-identifies patient data across organizations. Datavant’s network includes more than 70,000 hospitals and clinics.
Datavant uses tokenization software, which converts patient identifiers, such as names or medical record numbers, into unique, encrypted tokens. This enables researchers to match records across datasets without revealing personal information.
For the new partnership, Datavant built custom technology to tokenize Avandra’s imaging data, Tarzy pointed out. This will help researchers match up a patient’s imaging data with their other medical records without ever revealing the person’s identity, he noted.
Photo: Darunechka, Getty Images
