A Food and Drug Administration advisory panel voted 10-3 Thursday for Replimune Group's experimental melanoma therapy, after a contentious meeting in which experts raised doubts about whether the company's data and trial design could be trusted to show the skin-cancer drug works.

Thursday's recommendation is the latest turn in a regulatory saga that has stretched nearly two years. Earlier this week, a critical FDA analysis of the company's results triggered a more than 30% drop in its stock price.

Replimune first submitted its application for the therapy, called RP1, in November 2024. The FDA rejected it in July of last year, rejected a resubmission again in April, then allowed a third submission in June -- this time setting an Aug. 2 decision date and scheduling the public advisory to be held days before.

In May, The Wall Street Journal reported that the White House had pushed health officials to revisit the agency's earlier rejections of the therapy.

Several members of Thursday's independent panel described Replimune's data as "messy" and difficult to parse, but said they voted in favor because the drug showed promise among patients whose cancer had been resistant to prior treatment. Others suggested their vote hinged on the anecdotes of tumor shrinkage presented by experts and patients who spoke during the meeting.

More than 30 people spoke in a public comment period during the meeting, held at the FDA's campus in Silver Spring, Md., including several who had taken the experimental medication. All spoke in favor of Replimune's medicine.

Panelists said that an additional, Phase 3 study would further clarify whether the drug is effective. The agency's own scientists don't have to follow the panel's recommendation, but they often do.

Earlier this week the agency released briefing documents warning that some results from Replimune's single-arm study were "not interpretable," prompting shares in the company to lose nearly a third of their value. The FDA reviewers said they couldn't tell whether tumor responses came from RP1 itself or from a drug manufactured by Bristol-Myers Squibb it was paired with.

The FDA wrote that Replimune's study "does not include an adequate and well-controlled investigation that demonstrates substantial evidence of effectiveness."

Trading of Replimune stock was halted on Thursday, a regular practice around these panels, where outside experts make recommendations to the FDA about drugs they are considering for approval.

Representatives from Replimune made the case that the company had been aligned with the FDA on its review, and that a previous internal review panel had recommended the approval of the drug. That recommendation was later overruled.

RP1 is engineered to kill tumor cells directly and then trigger a broader immune response. Replimune has said that about one-third of patients responded to the treatment, and that the median patient response stretched for nearly two years in patients whose melanoma had already progressed on standard immunotherapy.

The FDA's own reanalysis, excluding responses the agency said couldn't be attributed to a systemic effect, found a considerably lower rate.

Thursday's meeting came a day after another closely watched advisory session, in which FDA advisers voted against Capricor Therapeutics' cell therapy for Duchenne muscular dystrophy -- a rejection that, like Replimune's, followed a prior FDA denial. Capricor stock fell 36% on Thursday.