Global Oncology Trials, Local Relevance: Why the FDA Just Raised the Bar

The FDA’s recent Complete Response Letter to Genentech’s BLA for Columvi (glofitamab) + GemOx in second‑line diffuse large B-cell lymphoma (DLBCL) has sent a crystal-clear signal: global trial data must be demonstrably relevant to U.S. patients.

What Went Wrong?

The phase III STARGLO trial enrolled 274 patients across 62 sites in 13 countries, but only 9% in the U.S., compared to a striking 48% in Asia. FDA review showed that survival benefit was driven by outcomes in the Asian cohort (OS HR ≈ 0.39), whereas non‑Asian regions, including North America, showed essentially no benefit or even trends against treatment (OS HR ≈ 1.06).

FDA’s New North Star: Domestic Relevance

This isn’t just about numbers—it’s about therapeutic precision. In today’s regulatory environment, trials must reflect data relevant to the U.S. As recent FDA guidance makes clear, multiregional studies now require robust U.S. representation or pre-specified bridging strategies.

During the ODAC panel (May 2025), the vote was 8‑1 against applying STARGLO results to U.S. patients, primarily due to underrepresentation of U.S. patients and inconsistent efficacy by region.

Lessons for Sponsors & CROs

If global enrollment leans heavily on one region—especially Asia, which is often younger on average and may have different disease subtype prevalence (e.g., higher ABC‑DLBCL rates)—you risk misleading efficacy signals for U.S. regulators.

Recommendations:
1. Aim for ≥ 20% U.S. enrollment in pivotal oncology trials.
2. If that’s not feasible, build in a statistically powered U.S. subgroup analysis or a dedicated bridging study.
3. Prespecify U.S.-specific endpoints and subgroup analyses in the protocol.
4. Be ready to present real-world or supplemental trial data to justify global extrapolation.

Why This Matters

Under-enrolling in the U.S. or over-relying on regions like Asia may initially speed up timelines—or meet other strategic priorities—but could backfire when regulatory scrutiny surfaces. Genentech now faces an estimated 1–2 year delay, potentially millions in lost revenue, and a costly reactivation of U.S. sites.

The Takeaway

The FDA is no longer rubber-stamping global oncology data unless it’s anchored in relevant evidence. This isn’t protectionism—it’s precision medicine. Collection of global data is vital, but greater emphasis must now be placed upon balanced enrollment and bridging strategies to ensure that findings truly apply to respective patients.