By Dan Rafter, MD, SVP of Product and Partner Strategy
The FDA just took major steps toward making real-time clinical trials (RTCT) a reality. In short, it has enabled a framework in which safety signals and trial endpoints are reported to the agency continuously as a study progresses, rather than at discrete phase endpoints. For life sciences organizations, this could become the new normal, and Biopharma and MedTech companies need to explore how they could enable similar protocols within their organization. In this blog, we’ll explore what the FDA and the involved parties have accomplished as well as how other life sciences organizations can do the same.
How Is the FDA Establishing Real-Time Clinical Trials?
On April 28, 2026, the FDA unveiled two active proof-of-concept trials with AstraZeneca and Amgen, and released a Request for Information for a broader pilot program launching this summer. Final selection criteria are expected in July and pilot selections in August 2026.
“For 60 years, we’ve been conducting clinical trials in the same way, where key data signals can take years to reach the FDA,” said FDA Commissioner Marty Makary, M.D. “We are boldly advancing a modern approach whereby FDA scientists can view safety signals and endpoints in real time as a trial progresses.”
Why Should Clinical Trial Sponsors Care About the FDA Announcement?
For clinical trial sponsors running trials that rely on medical imaging, providing near-real-time metrics could soon be the standard that defines leadership in life sciences. Across oncology, neurology, pulmonology, ophthalmology and other fields of study, organizations must reconsider how imaging data is collected, processed, validated and transmitted in order to stay at the forefront of medical research.
5 Steps to Enable Near-Real-Time Clinical Trial Data
Here’s what it will likely require to stay current with the FDA’s new framework for clinical trials.
1. Establish a Continuous Data Pipeline from Sites to Sponsors
Traditional trials batch data from sites to sponsors, who then submit to the FDA at defined milestones. Real-time clinical trials eliminate that batching model. Under the RTCT framework, pre-agreed signal criteria must be transmitted to the agency on an ongoing basis, as the FDA has already validated in AstraZeneca’s TRAVERSE trial through partner Paradigm Health.
For imaging-driven endpoints, this means your pipeline must process DICOM data at a pace compatible with continuous reporting, from ingestion to processing and signal extraction. Bottlenecks like manual workflows and email-based data transfers will no longer be sufficient. Sponsors should audit their current site-to-sponsor data flows now and identify opportunities to automate and accelerate data processing and sharing.
2. Utilize Architecture that Protects Blinding while Exposing Signals Continuously
Today, the firewall between data the FDA sees and data that could unblind a study team only has to stay firm at key moments, such as interim analyses and submissions. But continuous reporting means that firewall would need to hold steady at all times, in real time, which is a much harder governance and technical problem.
To accomplish this level of data de-identification, you need role-based, minimum-necessary views so FDA reviewers see aggregate signal trajectories without exposure to arm-level or site-level detail that could unblind sponsor personnel or bias ongoing trial conduct. This would be genuinely new infrastructure, not an extension of standard access controls, and it would need to be developed specifically for theimaging-data pipeline problem, as imaging is often the modality most likely to inadvertently reveal treatment effect.
3. Enable Real-time Anomaly Detection at Ingestion, Plus Reconciliation Between Provisional and Final Data
Batch-based trials have a built-in buffer: data managers catch scanner drift, mislabeled DICOM tags and protocol deviations before anything reaches the analysis stage. Continuous streaming removes that buffer, so a bad signal (such as a recalibrated scanner or corrupted volumetric read) could get reported to the FDA as a real safety or efficacy signal before anyone catches the underlying data quality issue. Catching an issue like this requires automated, validated anomaly detection built into the ingestion layer.
Real-time-reported values may later be revised once full QC and adjudication happen, so sponsors also need a formal reconciliation workflow that can explain — with full provenance — why a value reported in real time differs from what ends up in a locked database. Neither validated automated anomaly detection nor a formal reconciliation workflow currently exists in traditional batch workflows.
4. Build Multisite Harmonization Into Your Protocol From Day One
Both RTCT proof-of-concept trials are multi-site studies — AstraZeneca’s TRAVERSE spans MD Anderson and the University of Pennsylvania — but harmonization is manageable for two academic centers running Phase II trials. Phase III trials running across hundreds of sites leveraging multiple CROs is a different story.
At that scale, the methodology of segmentation would need to be aligned across imaging service providers. Real-time signal integrity assumes a measurement means the same thing no matter who generated it. Trial sponsors would need to be able to verify that integrity across CRO boundaries. This makes harmonization a decision around infrastructure, rather than a single trial. Imaging infrastructure would need to enforce the same acquisition standards, QC logic and provenance, regardless of vendor, site or trial.
5. Prepare Your Regulatory Affairs and Data Science Teams to Work in Lockstep
The RTCT model collapses the traditional distance between data collection and regulatory submission. That means the firewall that historically separated data science teams (who process imaging data) from regulatory affairs teams (who manage FDA interactions) no longer works.
Sponsors should establish cross-functional RTCT working groups now, well before the pilot program’s August selection deadline. These groups should include biostatisticians, imaging scientists, regulatory strategists, and IT/infrastructure owners — and they should be reviewing the FDA’s RFI together. The technical framework for real-time signal sharing has been demonstrated to work. The question now is whether your organization has the internal coordination to implement it.
Preparing for What’s Next
The FDA’s RTCT framework is still in early stages, but the proof-of-concept results signal a clear directional shift. Sponsors who want to position themselves to align with the pilot program or simply stay ahead of where clinical trial standards are heading should start auditing their data infrastructure and imaging workflows while enabling cross-functional coordination. And they should start now, rather than wait for formal requirements to materialize.
Flywheel is built for exactly this kind of moment. We built our platform on the expertise of imaging research veterans, and we have helped half of the top 20 biopharma companies establish a framework for regulatory-grade research. Reach out to one of our experts to lay the groundwork for near-real-time clinical trial data delivery now.
