Replimune 的 RP1 黑色素瘤疗法获 FDA 顾问委员会 10 比 3 支持

Following a pair of FDA rejections for its melanoma drug, Replimune’s fortunes flipped today after an advisory committee voted 10 to 3 in favor of the candidate’s effectiveness.
The panel was asked to determine whether or not Replimune’s trial of oncolytic virus RP1, also called vusolimogene oderparepvec and Tudriqev, produced data that were evaluable and clinically meaningful in advanced melanoma in combination with Bristol Myers Squibb’s Opdivo.
While the experts all agreed that the trial had significant issues, and that much hinges on a still-ongoing phase 3 trial of the candidate, the majority felt that the data showed enough signs of efficacy for RP1 to be approved.
“The way the study was conducted was maybe not perfect,” Hussein Tawbi, M.D., a melanoma expert at MD Anderson Cancer Center, said when explaining his ‘yes’ vote. “But at the end, there were tumors that shrank, both the injected ones and the non-injected ones, that resulted in clinical benefit for patients that otherwise don’t have any other options available to them.”
The three adcomm members who voted no did so because they could not get past the methodological issues with Replimune’s trial outlined by the FDA.
“There’s just so much uncertainty,” Karla Ballman, Ph.D., a cancer biostatistician at the Mayo Clinic, said when justifying her ‘no’ decision. “I just couldn’t vote yes.”
While the FDA takes views from adcomms into account, it is still the agency’s decision whether or not to approve RP1. A decision is slated to come by Aug. 2.
RP1 was initially rejected by the FDA last July, with the regulator deeming Replimune’s phase 1/2 Ignyte trial neither adequate nor well-controlled. The trial’s results, the FDA added at the time, “cannot be adequately interpreted due to the heterogeneity of the patient population.”
Similar issues plagued Replimune’s first resubmission, which was also spurned by the FDA in April after the agency deemed the data “insufficient” to conclude that RP1 works. This resubmission included Ignyte data as well as some data from an ongoing phase 3 trial, Ignyte-3. This decision, which the biotech’s CEO called “deeply disappointing,” led the company to shed 60% of its workforce.
The third submission, which was considered by the adcomm, again centered on Ignyte data with additional information now added on overall survival. Ignyte focused on patients whose disease progressed even while actively taking an anti-PD-1 drug like Opdivo, with 140 patients enrolled.
Replimune, the FDA, the adcomm members, and many who spoke during the public comment period all agreed that the unmet need for refractory melanoma patients is huge. Once PD-1 drugs no longer work, options are exceedingly limited or nonexistent.
The standard treatment for these refractory patients is tumor-infiltrating lymphocyte (TIL) therapy, a complex procedure that can take months to complete, comes with a chance of serious side effects and isn’t easy to access for many patients.
RP1, in contrast, has a better safety profile than TIL, MD Anderson’s Tawbi argued, and has already had its mechanism somewhat validated by Amgen’s Imlygic, which is also an oncolytic virus but is only approved to locally treat easily accessible melanoma lesions rather than systemic disease.
Replimune prevailed the day after Capricor faced a disappointing result in an adcomm of its own, which saw the advisors shoot down the biotech’s Duchenne muscular dystrophy cell therapy in a 9-to-3 vote.
Messy data
The core issues raised by the FDA centered on Ignyte’s methodology, which the agency said made the overall response rate unreliable and impossible to interpret.
The FDA took issue with the trial’s lack of a control arm, as well as certain methodological decisions. Replimune contends that a control arm of anti-PD-1 therapy alone is unethical, as the Ignyte study only enrolled patients for whom these PD-1 drugs had already failed.
Oncologists who spoke during the public comment period agreed that such a control arm would have been unethical. For patients already not responding to PD-1 drugs, retrying Opdivo leads to an ORR of just 5% to 10%, the advisors noted.
RP1, according to Replimune, produced an ORR of 33.6%.
Ignyte was marked by many deviations from the standard RECIST v.1.1 guidelines for evaluating solid tumors, which the FDA had explicitly told Replimune to adhere to. Chief among these was the allowance for investigators in the trial to reinject RP1 into tumors, even after a patient’s disease had progressed. If the patient’s tumor then shrank after reinjection, they were classified as a responder rather than as having disease progression.
Another big issue was the number of patients who had all of their target lesions, as identified under RECIST protocol, injected with RP1. Though RP1 is directly administered into tumors, it is meant to have a systemic effect by stimulating the immune system, which Replimune outlined in its presentation this morning.
If there were no non-injected tumors measured, the agency argued, it is difficult to determine whether or not RP1 has a systemic effect rather than just a local one.
“When excluding these patients with no non-injected target lesions to assess systemic response, the resulting ORR decreases markedly to 15.7%,” Katherine Barnett, M.D., an acting branch chief in the FDA’s Oncology Center of Excellence, said during the agency’s presentation.
“The response rate is unreliable due to assessment methodology that artifactually inflates ORR and [durability of response],” Barnett summarized. “FDA has identified multiple study design and conduct issues that impacted the interpretation of the efficacy results. The systemic effect of RP-1 is uncertain.”
In exploratory analyses, though, Replimune did find that 10 patients who had some target lesions not injected still had complete responses, suggesting a potential systemic effect. However, this was too few patients to sway advisor Paul Chapman, M.D., who placed the most importance on complete response rates.
“I talked about the CR rates. There are very few of them that are attributable to the treatment effect, 10 out of 140,” said Chapman, a melanoma specialist at Weill Cornell Medical Center. “I just think it’s a high chance of this being wrong.”
For Tawbi, the issues identified by the FDA largely comport with standard practice when treating melanoma, though he recognized they aren’t ideal for a clinical trial.
“What I see with the protocol here is it was actually copying what we do in clinical practice,” Tawbi said. For example, he said, tumor biopsies—which the FDA said could artificially inflate response rates—are routine to confirm that melanoma cells have died.
“I don't think Replimune did everything perfectly,” Tawbi added. “But overall, I see a population of patients that have very low expectations of response to any immunotherapy responding to the rate of 33%.”
And even if RP1’s ORR is ultimately lower than the 33.6% claimed by Replimune, most of the advisors thought there was still enough evidence that the virus is producing a benefit for patients.
“We can debate whether the response rate was truly 33% or if it was 16% or 25%,” David Miller, M.D., Ph.D., a melanoma specialist at Massachusetts General Hospital, said during the discussion. But as evidenced by patient testimony during the public comment period, he said, RP1 “is worthy of consideration and worthy of further discourse.”
During the advisors’ discussion, several strands of criticism for RECIST v1.1 emerged, with many calling the protocol inappropriate for therapies like RP1 that are injected directly into tumors.
“We have a moral obligation, and this is a call to the FDA as well, to help us generate a guidance for intra-tumoral therapy,” Tawbi said. “This is not the first agent. This will not be the last agent.”