UPDATE: Replimune (REPL) - Approved, With an Asterisk in the Denominator
TUDRIQEV is the first oncolytic virus approved in a decade, but the FDA wrote the label around 91 patients, not 140; the efficacy numbers on the package insert are not the numbers I’ve been quoting
The letter got signed. RP1 — now TUDRIQEV (vusolimogene oderparepvec-wtpg) — has accelerated approval in combination with nivolumab for adults with unresectable advanced cutaneous melanoma who progressed on a PD-1-blocking antibody regimen. That is the indication management asked for, and the label is broad: no lesion-accessibility restriction, no disease-stage carve-out, no prior-line ceiling beyond PD-1 progression.
That is the end of the regulatory saga I’ve been chronicling since the April 10 CRL, when I called this Dead Money and was wrong. Two Complete Response Letters, a public indictment of the agency, a 10-3 advisory committee vote, and now an approval letter.
But read the efficacy section of the label carefully, because the FDA did something in it that nobody in the +107% session on July 31 was pricing.
The Receipts
The approval (August 6). TUDRIQEV + nivolumab, accelerated approval, based on objective response rate and duration of response. Continued approval contingent on IGNYTE-3 (NCT06264180), the ~400-patient randomized confirmatory trial whose interim overall survival still sits in 2H 2027. This is the accelerated-approval bargain in its textbook form — convincing surrogate now, randomized proof later — exactly as I framed it on July 30.
The denominator got recut. Here is the number that matters and that the press release states plainly: of the 140 patients in the IGNYTE registrational cohort, 91 patients with at least one non-injected lesion were included in the efficacy-evaluable population. In that population, ORR was 24.2% with a median duration of response of 14.1 months.
Compare that to what I have been quoting in every update since June 1: 33.6% ORR (47/140) and 24.8-month median DOR. Those are not the label numbers. The FDA split the difference on the RECIST 1.1 denominator fight I decoded last week — it did not accept the agency reviewers’ harsh 15.7% (22/140) read, and it did not accept the company’s pre-specified 33.6%. It built the label out of the patients who had a tumor the needle never touched.
Why that recut is intellectually defensible. The whole scientific question with an intratumoral drug is whether shrinking the tumor you injected proves anything. RECIST 1.1 — the standard tumor-measurement ruleset — generally treats locally-treated lesions as non-measurable for exactly that reason. By restricting the efficacy population to patients with at least one non-injected lesion, the FDA is measuring the thing that actually distinguishes RP1 from a very expensive way to ablate a skin nodule: the systemic effect. 66.0% (35/53) of non-injected visceral lesions shrank by more than 30% in the AdCom deck. The label is essentially the agency saying we believe the systemic signal, and we’ll grade you on it.
The population is not cherry-picked easy disease. In the efficacy-evaluable group: 80% Stage 4, 54% PD-L1 negative, 45% with lung lesions, 24% with liver lesions, 13% with prior adjuvant anti-PD-1. That is a sick, visceral-disease-heavy cohort, which makes the 24.2% harder to wave off as selection.
Safety carries three boxed-adjacent warnings, not a boxed warning. Accidental exposure (healthcare providers, caregivers, pregnant women, newborns must avoid contact with injected tumors and dressings), herpetic infection or reactivation, and visceral injury — the last being the direct consequence of image-guided injection into liver and lung. Serious adverse reactions in 35% of 140; permanent discontinuation in only 2.9%. No Grade 4 or 5 common adverse events. The on-target flu-like profile I called a strength in February held.
The administration burden is now explicit in the label. Dosing is by tumor size, delivered by direct intratumoral injection, and for deep/visceral lesions is imaging-guided. Antivirals require a 72-hour delay. This is the friction I flagged as the reason I put Commercial Viability at Low/Medium in the deep dive — it is now written into the prescribing information.
A new 7.9% holder surfaced. Montanova Capital / Averill Master Fund (Aaron Cowen) filed a 13G-A today disclosing 6,609,000 shares, 7.9%, with an event date of 07/30/2026 — the day of the AdCom. Someone with real size was positioned into the panel. That’s a different signal than the Baker Bros resale shelf, and worth putting next to it.
