The FDA rejected Replimune’s advanced melanoma drug in April 2026, and the fallout reached Capitol Hill within weeks.

Health and Human Services Secretary Robert F. Kennedy Jr. appeared before a Senate hearing Wednesday and said he’d had no hand in the decision. He placed the call squarely on FDA Commissioner Marty Makary. “This decision comes out of FDA, and we trust the process there,” Kennedy said. He went further: “And I’ve been told by Marty Makary that every panel that looked at that drug unanimously voted against it because it does not appear to work.”

That account didn’t hold for long.

An op-ed in The Wall Street Journal pushed back directly. Cancer physicians who’d worked on the Replimune trials said the drug did demonstrate clinical effectiveness, contradicting Kennedy’s summary of the panel findings. The disconnect between Kennedy’s Senate testimony and the oncologists’ account in the Wall Street Journal is now the axis around which the broader dispute turns.

The FDA’s rejection letter, issued earlier in April 2026, pointed to a fundamental methodological problem with Replimune’s application: the company submitted a single-arm study. No control group. No randomized comparator. Regulators said that design prevented them from drawing defensible conclusions about the drug’s efficacy in advanced skin cancer. The complete response letter was direct, telling Replimune it must produce data from a well-controlled trial before the agency would reconsider the application.

Single-arm trials don’t establish causation. That’s not a controversial claim.

When a trial lacks a comparator arm, it can’t distinguish whether the observed tumor responses came from the drug’s mechanism or from patient selection bias, regression to the mean, or the natural fluctuation of disease. The FDA has applied this reasoning consistently across oncology and other therapeutic categories, and the FDA’s longstanding evidentiary framework for drug approval has codified that standard for decades. The agency’s position is that a single-arm study showing a meaningful objective response rate still can’t reliably isolate drug effect from confounding variables. Replimune’s submission, by that measure, didn’t clear the bar.

That said, the bar itself is contested in oncology. Physicians treating patients with advanced skin malignancies, where approved options are often exhausted, have long argued that the agency’s evidentiary threshold creates a practical access problem. When a patient population is small and the disease is aggressive, mounting a randomized controlled trial is logistically difficult and, in some circumstances, ethically fraught. Randomizing patients to a placebo arm in a disease with high short-term mortality carries real human costs that regulators and trialists weigh differently.

The FDA has tried to thread this needle through accelerated approval mechanisms, allowing certain oncology drugs to reach market based on surrogate endpoints, particularly objective response rate and duration of response, while requiring sponsors to confirm clinical benefit through post-market trials. Whether Replimune’s submission could have qualified even under that more permissive standard is one of the central questions the oncologists writing in the Wall Street Journal raised, and it’s one the agency hasn’t fully answered publicly.

Oncologists who participated in the Replimune trials are contesting Kennedy’s characterization of how reviewing panels voted. If Kennedy’s statement to the Senate, that every panel unanimously voted against the drug, is inaccurate, it creates a significant evidentiary problem for HHS leadership, which is now publicly associated with that claim. The FDA hasn’t issued a correction, and the agency’s communications office hasn’t clarified the record.

The case has broader implications for how regulatory disputes get narrated at the political level. Kennedy’s appearance before the Senate on Wednesday conflated two separate questions: whether the FDA followed its own process (which Kennedy appeared to endorse), and whether the outcome of that process was scientifically sound (which the oncologists dispute). Those aren’t the same question. The agency can follow procedure and still reach a result that practicing clinicians find inadequately calibrated to patient need.

It’s worth examining what each side is actually claiming. The FDA said Replimune’s single-arm study design was insufficient to establish efficacy under established evidentiary standards. That’s a methodological objection, not a clinical one. It says nothing, technically, about whether the drug works in the biological sense. The oncologists writing in the Wall Street Journal said the drug demonstrated clinical effectiveness, a clinical claim based on their direct observation of trial data and patient outcomes. Kennedy, speaking at a Senate hearing Wednesday, offered a third framing, a vote-counting claim about panel unanimity, that’s now disputed by the physicians who were in the room.

Regulatory gatekeeping in oncology is genuinely hard. It’s not a domain where the right answer is obvious or where evidentiary standards are politically neutral. The FDA’s requirement for randomized controlled trials protects against drugs that look effective in uncontrolled settings but don’t deliver population-level benefit. That protection matters. Drugs that appear promising in single-arm studies have failed randomized trials before, sometimes badly, and patients who received them outside of trials paid real costs.

At the same time, accelerated approval exists precisely because the agency recognized that strict RCT requirements can impose delays that are indefensible in diseases with few alternatives and short survival windows. Advanced cutaneous malignancy, the category Replimune’s drug targets, often fits that description.

What Replimune must do now is concrete: produce data from a well-controlled trial. The STAT News report from April 23, 2026, situates this case within a pattern of pharmaceutical access tensions across regulatory contexts, including in Europe, where companies have separately been accused of withholding new medications from markets pending pricing negotiations. That comparison isn’t exact, but it points to a recurring structural problem: the gap between drug development timelines and patient timelines doesn’t close easily regardless of regulatory system.

The HHS Office of Inspector General and the Centers for Medicare and Medicaid Services aren’t direct parties to the Replimune dispute, but both agencies’ frameworks for coverage and reimbursement will matter significantly if the drug eventually secures approval. Oncology drugs approved through accelerated pathways frequently face post-approval coverage disputes, particularly when confirmatory trial data arrive late or with mixed results. The reimbursement question follows the regulatory question, often at a considerable lag.

Whether Replimune can execute a randomized trial in the relevant patient population within a timeframe that keeps the company’s program viable is an open question. The FDA’s complete response letter in April 2026 didn’t set a deadline. It established a condition.

Kennedy’s Senate testimony and the Wall Street Journal op-ed are now part of the public record, pulling in opposite directions on the same factual question.