At BIO Europe 2025, we spoke with oncology sponsors across Europe and the U.S. about how they’re planning early-phase trials in 2026. Three themes came up repeatedly — and they reveal how successful sponsors are thinking about regulatory strategy, site selection, and the role of AI.
The debate is always framed as: ‘Europe is cheaper/faster’ vs. ‘The U.S. is essential for long-term registration.’ In practice, FDA doesn’t require that all safety data come from U.S. sites, but they do expect data that’s directly relevant to the intended U.S. label population—and they engage more deeply when you bring them into the conversation early, especially around first-in-human and dose-escalation plans.
What we see consistently working for U.S.-focused oncology programs:
Patterns that create friction with FDA and slow U.S. development:
Real talk: If your primary goal is a U.S. approval and U.S. commercial launch, your early development plan should be built with the U.S. in mind—that usually means U.S. site involvement and early FDA input, even if you also run in Europe. If budget is the constraint, a focused, well-designed Phase I with the right oncology CRO and sites is often more efficient than a larger, diffuse program that you have to retrofit for the U.S. later.
This is the question everyone asks but few have a good answer for.
The reality: Top Phase I sites don’t prioritize big pharma because of size—they prioritize programs with strong science, clean protocols, reliable execution, and clear patient value. Big pharma often meets that bar, but emerging biotechs do too when the science is compelling.
Budget is rarely the deciding factor; scientific strength and operational readiness matter far more— assuming the budget is competitive and realistic.
What top sites actually want:
What gets you pushed aside:
Medelis advantage: Our founders spent 15+ years in oncology drug development before starting a CRO. When we call a PI at MSK or Dana-Farber, we’re talking science first, logistics second.
AI came up in nearly every keynote, panel, and side discussion at BIO Europe—and for good reason. Sponsors are trying to understand what AI can realistically do today versus what remains aspirational.
Where AI is genuinely adding value today:
✓ Patient–trial matching — AI-assisted screening tools can reduce screen failures and accelerate identification of eligible patients, depending on disease area and data quality.
✓ Predictive toxicity insights — Early AI models can flag potential safety concerns in preclinical or retrospective datasets, though these tools are not yet standard for prospective decision-making.
✓ Biomarker and pathology analysis — In controlled studies, AI has matched or exceeded expert performance in tasks like HER2 scoring; in practice, it functions as an adjunct, not a replacement.
✓ Protocol optimization support — AI can analyze historical trials to highlight restrictive inclusion/exclusion criteria and operational bottlenecks, but humans still make the design decisions.
Where AI is still largely theoretical for clinical development:
✗ Autonomous trial design — Not accepted by regulators and not feasible for oncology complexity.
✗ Regulatory submission authorship — AI can assist with drafting, but agencies require human authorship, certification, and full QC.
✗ Site selection for early oncology — Analytics can help, but for Phase I studies PI relationships, scientific fit, and operational trust remain the controlling factors.
Our take:
AI is a powerful tool—but not a development strategy. Use it to augment data-heavy tasks, accelerate insights, and reduce operational friction. But in oncology, where safety, biology, and clinical nuance drive outcomes, AI complements expert judgment—it does not replace it.
If you’re planning a Phase I or II oncology study in 2026 and want to pressure-test your strategy, Medelis works with sponsors at this exact stage. We’re happy to talk through feasibility, site strategy, or FDA engagement.
Reach out to schedule a call.