Imaging Data: The Missing Layer in Real-World Evidence
Why EHR and Claims Alone Aren’t Enough
Real-world evidence (RWE) has become the backbone of modern clinical research, shaping everything from regulatory submissions to AI model development. Yet despite its importance, the foundation of most RWE strategies remains incomplete.
The majority of real-world data (RWD) pipelines are built around electronic health records (EHR) and claims data. These sources have transformed how we understand patient journeys and treatment outcomes — but they can’t tell the full story. Without imaging, researchers are missing the visual truth of disease: what’s actually happening inside the body. That missing layer limits the precision, validation, and clinical confidence of every insight drawn from RWE.
The Building Blocks of Real-World Data
For years, pharma and AI innovators have relied on two primary data types — claims and EHR data — to build their RWE strategies.
Claims data is highly standardized and structured, ideal for tracking healthcare utilization, billing, and long-term outcomes. EHR data adds valuable clinical context, capturing diagnoses, lab results, medications, and provider notes. Together, these datasets have powered everything from HEOR studies to predictive AI models, largely because they’re widely available, easier to de-identify, and supported by familiar regulatory frameworks.
But while claims and EHR data reveal what was recorded, they don’t show what was seen.
The Blind Spots in Today’s Evidence
Neither claims nor EHR data captures the visual information that physicians rely on every day — the images that guide diagnosis and treatment. And as medical imaging data continues to surge — with healthcare data volume growing nearly 47% per year and imaging representing roughly 90% of that total — this missing layer becomes even more significant.
In oncology, for example, claims might indicate chemotherapy cycles, but not whether a tumor actually shrank. In neurology, an EHR may include cognitive test results, but not the subtle changes in brain volume that precede them. In cardiology, notes might describe abnormalities, but can’t quantify how much plaque has built up or whether heart function is improving.
This gap is more than an academic problem — it affects how early disease is detected. Most conditions are far more manageable when caught early, and imaging is often the first place those changes appear. Earlier detection doesn’t just improve patient outcomes; it also enables life sciences companies to identify the right patients sooner, initiate treatment earlier, and generate clearer, faster evidence on therapeutic effectiveness.
Without imaging, researchers are left to infer outcomes rather than observe them directly. It’s a gap between what’s documented and what’s truly happening inside the patient — and that gap has real implications for research validity, AI model performance, treatment discovery, and ultimately, patient outcomes.
Imaging: The Clinical Truth Layer
Imaging fills that gap. It provides objective, visual evidence of disease — the clinical truth at the source.
By integrating imaging data, researchers gain a level of clarity and validation that text-based data alone can’t provide. Imaging can confirm disease presence, stage, or progression; measure treatment response with precision; and enhance phenotyping for AI and machine learning models. It also enables earlier detection and more personalized treatment strategies — moving research from descriptive to truly diagnostic.
For AI innovators, imaging serves as the ground truth that ensures algorithms are not only accurate, but generalizable across real-world patient populations.
Why Imaging Has Been Left Out
Despite its value, imaging data has long been one of the most underutilized assets in clinical research. The reasons are understandable.
Imaging data typically lives in PACS or VNA systems, designed for clinical operations rather than research. The files are massive — often high-resolution DICOMs that are difficult to transfer or store at scale. Each modality comes with unique acquisition parameters, making standardization a challenge. And de-identifying imaging data is far more complex than redacting text, since protected health information can be embedded at the pixel level.
These hurdles have made imaging difficult to incorporate into RWE pipelines, despite its potential to deliver a more complete and reliable view of patient health.
Unlocking Imaging for Real-World Evidence
That’s now changing. Avandra is enabling researchers and AI developers to harness imaging data directly — securely, compliantly, and without moving it from its source.
Through source-level access, Avandra connects directly to health systems’ PACS and VNAs, eliminating the need to export or duplicate files. Its federated architecture allows researchers to work within the health system’s environment, preserving data governance and HIPAA compliance while dramatically accelerating access.
Avandra’s advanced de-identification technology maintains the integrity of pixel-level data while ensuring full privacy protection, and its data linkage capabilities connect imaging with EHR and claims data to create a complete, longitudinal patient view. The result is regulatory-grade reliability, with transparent data provenance ready for submissions, audits, or AI validation.
A New Era of Multimodal Evidence
Real-world evidence is evolving beyond traditional boundaries. The next frontier is multimodal integration — bringing together imaging, EHR, claims, genomics, and other data types to form a unified, patient-level view.
Imaging adds depth, validation, and clinical richness across therapeutic areas:
In oncology, it allows researchers to monitor therapeutic response in near real time.
In neurology, it provides quantitative biomarkers of disease progression.
In cardiovascular disease, it visualizes subtle structural and functional changes over time.
In rare diseases, it can surface phenotypes invisible to text-based systems.
With imaging in the mix, RWE becomes not just a record of care — but a reflection of biology in action.
Closing the Evidence Gap
EHR and claims data tell the story. Imaging shows the truth. By closing the gap between what’s recorded and what’s visible, researchers can produce stronger, faster, and more clinically grounded real-world evidence.
With Avandra, pharma, biotech, and AI innovators can finally access source-level imaging data — securely and at scale — unlocking a more complete understanding of disease and driving better outcomes for patients everywhere.
