Replimune Won Accelerated Approval on 91 Evaluable Patients and a 24% Response Rate
- The FDA granted accelerated approval to Tudriqev (vusolimogene oderparepvec-wtpg) on 6 August 2026, to Replimune, in combination with nivolumab, for adults with unresectable advanced cutaneous melanoma who progressed on a PD-1 blocking regimen. It is a genetically modified oncolytic herpes simplex virus type 1.
- The evidence is a single-arm, open-label, multiregional trial that enrolled 140 adults with Stage IIIB, IIIC or IV disease following at least eight consecutive weeks of prior anti-PD-1 therapy. 91 patients were evaluated, 24% achieved an objective response, and median duration of response was 14.1 months.
- Tudriqev is injected directly into tumours every two weeks for eight doses, with the amount set by tumour size and a lower concentration used for the first dose. Nivolumab is given intravenously from week three. Warnings cover accidental herpes transmission to close contacts, herpes infection or reactivation in the patient, and injection procedure complications.
- The FDA convened its Cellular, Tissue and Gene Therapies Advisory Committee on 30 July 2026, including a public hearing from patients, advocates, clinicians and independent experts. The application held Breakthrough Therapy and Priority Review. Approval rests on response rate and duration, with confirmatory trials required to verify clinical benefit.
Clinical read
Take the numbers plainly. A single-arm trial enrolled 140 patients, 91 were evaluated, and 24% responded, which is roughly twenty-two people. That was sufficient for accelerated approval, and the advisory committee that heard it sat one day after the same committee voted nine to three against Capricor’s deramiocel. The contrast between those two files is the most useful regulatory signal available this month, and it has nothing to do with effect size. It is about how the effect was measured.
- Modest and clean beats large and contested. Capricor’s problem was never whether patients improved, it was which statistical analysis plan governed the answer. Replimune arrived with a pre-specified endpoint, a single-arm design everyone understood going in, and a response rate nobody has to argue about. For anyone filing at CBER the lesson is uncomfortable but plain: the agency will take a small number honestly derived long before it takes a large number whose derivation is in dispute. Note too that accelerated approval is CBER buying an option rather than reaching a verdict. The confirmatory trial is where the judgement actually gets made.
- The denominator is doing work and the release does not explain it. One hundred and forty enrolled, ninety-one evaluated. In a single-arm trial the evaluable population is the entire basis of the response rate, so a gap of roughly a third is not a footnote. Anyone modelling real-world performance should establish what happened to the other forty-nine before assuming a quarter of treated patients will respond. The analysis population definition in the label is the number to demand, and it will tell you more about durability of the claim than the 24% does.
- The administration profile is the access problem, and almost nobody will write about it. This is a live modified herpes virus injected into tumours across eight visits, dosed by lesion size, carrying a warning about transmitting herpes to household contacts. That constrains site of care, handling, waste disposal and household counselling in ways an infused antibody never does. Uptake here will be gated by institutional biosafety committees and injection capacity rather than by payers, so watch which centres adopt it and how fast. That curve, not the label, is what determines whether a 24% response rate reaches anyone.