The FDA Wants to Watch Your Clinical Trial Live. Like, Right Now.
The FDA is now watching clinical trial data stream in live from AstraZeneca and Amgen studies, ditching the old batch-submission model entirely. It's the biggest change to how regulators oversee drug development in decades, and a broader pilot launches this summer.
The End of "We'll Get Back to You"
Imagine running a marathon, but instead of checking your time at mile markers, your coach watches every single step on a live GPS feed. That's essentially what the FDA just did to clinical trials.
On April 28, the agency announced it has launched two proof-of-concept "real-time clinical trials" with AstraZeneca and Amgen. The concept: instead of sponsors collecting data for months (or years) and then shipping it to regulators in giant batches, the FDA now sees pre-specified safety signals and clinical endpoints as they happen, streamed through the cloud.
This isn't a press release about a future vision. AstraZeneca's trial is already live, and the FDA has already received and validated data from it. The plumbing works. The water is flowing.
What's Actually Running
The two pilot trials are both in oncology, which makes sense: cancer trials generate lots of safety data, fast.
AstraZeneca's TRAVERSE trial is a Phase 2 study in treatment-naïve mantle cell lymphoma, with sites at MD Anderson Cancer Center and the University of Pennsylvania. Data flows through a platform built by Paradigm Health, and FDA reviewers can already see signals arriving in real time.
Amgen's STREAM-SCLC trial is a Phase 1b study testing tarlatamab (marketed as IMDELLTRA), a bispecific T-cell engager targeting DLL3, in limited-stage small cell lung cancer. Final site selection is still underway, putting it slightly behind AstraZeneca's effort.
Both trials transmit curated, pre-defined signals to the FDA. Not raw patient records. Not a firehose of unstructured data. Think of it like a dashboard showing the metrics regulators agreed to watch: adverse event rates, tumor responses, dose-escalation triggers. The agency and sponsors negotiated upfront exactly which signals would flow and what thresholds matter.
Why This Is a Bigger Deal Than It Sounds
To understand why RTCT (real-time clinical trials) could reshape drug development, you need to understand how painfully slow the current system is.
Get tomorrow's biotech intelligence before your competitors.
Join thousands of biotech professionals who start their day with our free, daily briefing.
Traditionally, a sponsor collects trial data, locks the database, analyzes it, writes it up, and submits it to the FDA. The agency then reviews it. This process introduces what experts call "dead time," periods where sponsors are contextualizing unexpected signals, shaping narratives, and basically explaining what happened months after it already happened.
Real-time trials could compress or eliminate much of that gap. If the FDA can see safety problems emerging in real time, it can act faster: adjusting doses, expanding cohorts, or halting dangerous arms before more patients are exposed. Conversely, if data looks promising, development could accelerate without waiting for the traditional data-lock-analyze-submit cycle.
Jefferies analysts called the initiative "another step towards reducing structural inefficiencies and accelerating timelines." Market commentary on Amgen estimated that scaled RTCT could cut oncology regulatory review timelines by 20 to 30 percent, potentially extending effective exclusivity windows by 6 to 12 months.
The Bigger Pilot Is Coming Fast
These two trials are just the opening act. Alongside the proof-of-concept announcement, the FDA released a Request for Information titled "AI-Enabled Optimization of Early-Phase Clinical Trials Pilot Program." That RFI, published April 29 in the Federal Register, lays out the vision for a broader pilot launching this summer.
The timeline is aggressive:
Comment period: open until June 29 (extended from the original May 29 deadline)
Final selection criteria: July 2026
Pilot participant selection: August 2026
By late summer, more companies could be streaming trial data to the FDA in real time. The agency is actively soliciting input on how to design the program, what success looks like, and what guardrails are needed.
The AI Layer Nobody's Talking About Enough
Buried in the framework is something significant: the FDA explicitly couples RTCT with AI and machine learning. The pilot program's full name references "AI-Enabled Optimization," and the vision includes automated signal processing, continuous monitoring, and potentially AI-assisted summarization of emerging data.
This raises questions the FDA hasn't fully answered yet. How will AI outputs interact with human reviewer judgment? What validation standards will govern AI-assisted analysis? Who's accountable when an algorithm flags (or misses) a safety signal?
FDA's Chief AI Officer Jeremy Walsh has emphasized that the agency receives aggregated metrics, not raw patient records. That helps with privacy. But the governance framework for AI-driven regulatory decisions is still very much under construction.
Not Everyone's Celebrating
Amgen's CMO Paul Burton struck a measured tone, stressing that "the fundamentals of good science do not change" and that protocol oversight, informed consent, monitoring, and governance remain unchanged pillars.
But critics have sharper concerns.
Some experts worry about "dirty data": streaming large volumes of poorly curated information could add noise rather than clarity. Others point to a potential equity problem. AstraZeneca and Amgen have massive data infrastructure teams, cloud architecture, and regulatory affairs armies. They can afford the upfront investment in EHR integration, clean data pipelines, and intensive early regulatory engagement.
Smaller biotechs? Academic sponsors? They might not have the resources to play this game. Regulatory science expert Aman Khera warned that early participants in the voluntary pilot will help set the standards and expectations. If smaller sponsors don't participate, they may later face rules they had no hand in shaping.
There's also a philosophical concern. A Policy Watch commentary suggested RTCT could represent a shift toward prioritizing industry speed over patient safety, especially when combined with the FDA's recent move toward accepting single pivotal trials for some approvals. The worry: faster data plus fewer trials equals less certainty.
The Statistical Headache
Continuous monitoring of accumulating data introduces a classic statistical problem: the more often you peek at results, the more likely you are to see false signals. It's like checking your retirement account every hour; the daily fluctuations look terrifying, but they don't mean anything long-term.
In clinical trials, repeated looks at interim data without proper statistical controls inflate type I error (false positives). Adaptive trial designs already grapple with this through pre-specified decision rules and alpha-spending functions that "budget" how much statistical confidence you burn at each interim look.
RTCT adds another dimension: regulators watching the same data in real time. Who decides when a signal is real versus noise? How does FDA's emerging view interact with the independent Data Monitoring Committee's conclusions? What happens when they disagree?
These questions don't have clear answers yet. The ICH E20 guidance on adaptive designs (endorsed June 2025) and an FDA Bayesian methodology guidance (January 2026) provide some framework, but neither was written with continuous regulator streaming in mind.
What This Means for the Industry
If you're a large pharma company with oncology assets, the calculus is simple: get in early, help write the rules, and potentially shave years off your development timeline. Evercore ISI called RTCT a "logical next step" in regulatory evolution.
If you're a smaller biotech, the situation is more complicated. You'll need to invest in data infrastructure, engage with the FDA earlier and more intensively, and demonstrate operational readiness for continuous signal sharing. That's expensive and resource-intensive, especially for a company running its first clinical program.
For patients, the promise is real: faster detection of safety problems, quicker identification of effective therapies, and potentially years shaved off the journey from lab to pharmacy shelf.
But promises and execution are different things. The FDA is essentially beta-testing a new operating system for drug development. AstraZeneca and Amgen are the first two users. By August, more will join. And within a few years, we'll know whether real-time trials are the future of regulation or an ambitious experiment that couldn't scale.
The marathon coach is watching. Every step, in real time. The question is whether that makes the runner faster, or just more nervous.
Clinical & Regulatory4 min read
Exelixis Built a Backup Plan for Its $2B Drug. It Just Hit a Wall.
Exelixis' next-generation drug zanzalintinib missed a key survival endpoint in colorectal cancer, sending shares down 10-12%. With Cabometyx still driving 90% of revenue, the pressure to prove the successor can carry the franchise just got a lot more intense.
