FDA begins accepting live AI and cloud data feeds from AstraZeneca’s TRAVERSE and Amgen’s STR trials

Clinical trials are usually slow in a very specific way — not just because patients take time to enroll, but because the data moves in batches. Sites collect it, sponsors clean it up, analysts package it, and only then does the FDA really see the important signals. On April 28, that changed a little. The FDA said it had begun two proof-of-concept “real-time clinical trials” with AstraZeneca and Amgen, where regulators can receive selected safety and endpoint signals as they happen. ### What actually changed? The concrete news is not that the FDA approved a drug or cleared a new AI model. It launched a new review setup. The agency announced two proof-of-concept studies — AstraZeneca’s Phase 2 TrAVeRse trial in treatment-naïve mantle cell lymphoma and Amgen’s Phase 1b STREAM-SCLC trial in limited-stage small cell lung carcinoma — plus a Request for Information for a broader pilot program expected this summer. ### Why is that different from a normal trial? Normally, the FDA sees important trial information after a sponsor has already aggregated and submitted it. That lag can stretch for months and sometimes longer. The new model is narrower but faster: the FDA and each sponsor predefine which signals matter, then those signals get transmitted in near real time through cloud infrastructure instead of waiting for a traditional submission package. ### Which trial is furthest along? AstraZeneca’s is. The FDA said it has already received and validated signals from TrAVeRse through Paradigm Health, which is the clearest sign that this is not just a whiteboard concept. TrAVeRse is a Phase 2 study of acalabrutinib, venetoclax, and rituximab in previously untreated mantle cell lymphoma, with participation from MD Anderson and the University of Pennsylvania. ### What about Amgen’s study? Amgen’s STREAM-SCLC trial is earlier-stage and not as operationally mature yet. The FDA described it as a Phase 1b study in limited-stage small cell lung carcinoma and said final site selection is still in process. So the headline is not that both pipelines are equally live today — it’s that AstraZeneca has already proven the feed can work, while Amgen is part of the same framework as the next test case. ### Where does AI fit in? AI is part of the broader program, but it is easy to overread that point. The April 28 announcement bundled the real-time trial proofs of concept with an FDA request for input on an “AI-enabled optimization” pilot for early-phase trials. That proposed pilot asks how AI tools might help with things like safety monitoring, dose selection, signal detection, and recruitment — not just raw data transport. ### Is the 20% to 40% speedup real? That number is coming from FDA leadership as an estimate of what this model could eventually save across overall clinical-trial time, not from completed results in these two studies. Basically, think of it as an ambition, not an outcome. The hard evidence so far is narrower: the FDA says it successfully received and validated AstraZeneca trial signals, which shows the technical plumbing works. ### Why does this matter beyond these two cancer trials? Because once regulators can watch prespecified signals in the cloud, the bottleneck shifts from “when do we submit?” to “what data can we trust right now?” That raises very unglamorous but crucial engineering problems — schemas, provenance, monitoring, audit trails, and rules for which signals are important enough to stream. If those pieces hold, the FDA gets earlier visibility into problems and promising effects. ### Bottom line The real story is not that the FDA suddenly turned drug review into a live dashboard. It is that the agency moved real-time review from idea to working pilot on April 28, 2026, with AstraZeneca already sending validated signals and Amgen lined up behind it. That is a small operational step, but in drug development, small workflow changes can reshape the whole timeline.