Transparency is foundational to everything we do. Here we address the most important questions about our mission, our model, and how we protect patients, facilities, and data.
AI researchers need massive, diverse datasets — not just data from large academic medical centers. When a patient has an MRI, that facility gets paid once. That single payment covers the radiologist's read, and then the facility must store that study for up to 10 years plus a backup — and they are never compensated again, even when that same study is pulled for a comparison read or reinterpretation.
Small and mid-sized hospitals also lack both the budget and the staff to de-identify radiology, cardiology, and pathology studies at scale. And researchers are not knocking on those doors. That data — which has real scientific value — simply sits unused and uncompensated forever.
That is not just ethical — it is a win for patients, facilities, and science alike.
The originating facility does — always. We handle complete de-identification at near-zero cost to them, and for every study used in research, we return a portion of what we receive directly back to the hospital, imaging center, or pathology lab that generated it.
Facilities own the relationship and opt in entirely on their own terms. They are never locked into categories of research they are not comfortable with, and they retain full institutional control over how their data is used — not just whether it is used.
We agree those questions come first — which is why our entire architecture is built around HIPAA-compliant de-identification, institutional consent, and transparent revenue sharing.
Each facility we partner with has the ability to opt out of specific categories of research. For example, a hospital that did not want its data used in DNA modification research — or any other research type that conflicts with their values or their community's expectations — can simply exclude that category at the time they sign with us.
Under HIPAA, individual patient opt-in is not required once data has been properly de-identified — the information is no longer considered Protected Health Information (PHI) under the law.
That said, many hospitals already address this within their standard HIPAA consent agreements, and it is reasonable to assume that a patient who had concerns could work with their facility to opt out. Please note: we are not attorneys and this does not constitute legal advice.
More importantly, we have built institutional-level controls directly into our model. Each partner facility can opt out of specific categories of research at the time of signing — ensuring the facility is always in control of how their data is used, not just whether it is used.
De-identification under HIPAA is not simply removing a patient's name from a file. It requires the removal of all 18 identifiers defined by HIPAA — including dates, geographic data, device identifiers, and any other information that could reasonably be used to re-identify an individual. There are two accepted methods:
For medical imaging — radiology, cardiology, and pathology — de-identification also involves scrubbing data embedded within the files themselves, such as DICOM headers, which can contain a surprising amount of patient information beyond what is visible on screen.
And it goes even deeper. Some modalities, such as ultrasound, can have PHI physically burnt into the image itself — rendered directly into the pixel data. This cannot be removed through header scrubbing alone. It requires specialized image processing to detect and redact that information without compromising the diagnostic value of the study.
And because auditability matters, our process creates a traceable, documentable record of how each dataset was de-identified and used.
No. This is one of the most important aspects of our model. De-identification happens entirely on-site — inside the hospital, imaging center, or pathology lab — before any data ever moves.
We deploy our Radiant AI Health Data, Inc. technology at zero cost to each facility. We bring the technical infrastructure and process expertise so that small and mid-sized facilities — who do not have the internal resources to do this properly — can participate confidently and compliantly.
Yes. Our software can also be used by facilities to migrate studies between PACS and VNA systems — the enterprise platforms that hospitals and imaging centers rely on to archive and manage patient studies and data. This is a notoriously complex and costly process, and our solution addresses it directly as an added benefit to our partners.
We also have several other exciting capabilities in development that we are not quite ready to share just yet. We genuinely believe they will be a game changer for how medical data is processed and exchanged between health systems and the AI research programs being built to analyze it.