BioNTech's Cancer Vaccine Missed the Endpoint Pembrolizumab Itself Never Won
AHEAD-MERIT missed its progression-free survival endpoint. In this disease, pembrolizumab itself never won PFS either. The trial now turns to overall survival, the endpoint where this cancer actually rewards immunotherapy.
On October 13, 2026, BioNTech announced first topline results from AHEAD-MERIT (NCT04534205), a Phase 2/3 trial of BNT113 (tenumomeran/raplemomeran) added to pembrolizumab in first-line unresectable recurrent or metastatic HPV16-positive, PD-L1-expressing head and neck squamous cell carcinoma (CPS ≥1). The combination did not show a statistically significant improvement in progression-free survival over pembrolizumab alone. The Independent Data Monitoring Committee nonetheless recommended the trial continue to its second dual primary endpoint, overall survival, which under the trial's pre-specified design would establish a positive outcome on its own. Shares closed down about 7% that day at $89.01, according to data on BNTX.
A fixed-target vaccine, not a personalized one
BNT113 is an mRNA-lipoplex immunotherapy encoding two fixed HPV16-derived antigens, the E6 and E7 oncoproteins, built to drive HPV16-specific anti-tumor T-cell responses that strengthen checkpoint inhibitor treatment. Because the targets never vary between patients, the product is off-the-shelf, and it earned FDA Fast Track designation in January 2026. The same release describes AHEAD-MERIT as open-label and two-arm, with a Part A safety run-in and a randomized Part B, treatment up to roughly 24 months, and dual primary endpoints of PFS and OS. The commercial logic is broad: about one third of HNSCC cases are HPV-associated, roughly 90% of oropharyngeal cancers are driven by HPV16, and HNSCC ranks seventh worldwide in incidence with HPV16-related oropharyngeal tumors pushing rates higher.
The scoreboard problem: pembrolizumab never won PFS either
The uncomfortable arithmetic sits in the control arm's own history. In KEYNOTE-048, pembrolizumab alone failed to improve progression-free survival versus cetuximab-chemotherapy in any population tested, yet it prolonged overall survival: median OS of 14.9 versus 10.8 months in the PD-L1 CPS ≥20 population and 12.3 versus 10.4 months in CPS ≥1, with a plateau of approximately 20% of patients alive at 48 months. Objective response rates ran lower on pembrolizumab alone (23.3% versus 36.1% in CPS ≥20), but responses lasted substantially longer. A drug that loses PFS while winning OS is the defining pattern of this disease, which makes the design choice less like spin and more like precedent. The safety run-in, presented as abstract 877P in Annals of Oncology, points the same direction on a small scale: in 15 patients, unconfirmed overall response rate reached 40% (four complete, two partial responses), disease control 53%, median PFS just 3.9 months, but median overall survival 22.6 months, with BNT113-related events mainly grade 1-3 flu-like symptoms. A 22.6-month median on a 15-patient cohort is not evidence of efficacy, but a 3.9-month PFS paired with that survival figure is exactly the dissociation that immunotherapy produces.
Özlem Türeci, BioNTech's co-founder and chief medical officer, said the PFS result was not statistically significant at this first analysis, that the data support continued evaluation, and that they "show that it is engaging its intended therapeutic mechanism." She added that as immune-mediated anti-tumor effects build and deepen over time, the OS analysis will be critical.
Why the stakes run past this one readout
Pressure on BioNTech's oncology pivot is not hypothetical. On August 28, 2026, BioNTech and Genentech terminated the Phase 2 trial of autogene cevumeran (BNT122) in resected colorectal cancer after an internal safety review, and trade press reported the scrapped test recorded more deaths than the control arm. The leadership layer is shifting too: on March 10, 2026 the company said co-founders Uğur Şahin and Özlem Türeci plan to leave at the end of 2026 for a new co-founders-led venture, and on August 3, 2026 BioNTech appointed Guido Oelkers to the management board as CEO. As pharmaphorum notes, BNT113 leads the off-the-shelf mRNA cancer immunotherapy track ahead of BNT111 in advanced melanoma and BNT116 in advanced NSCLC, while BNT122 anchors the individualized vaccine track partnered with Roche/Genentech. With COVID revenue long faded, a miss on the lead off-the-shelf asset compounds quickly.
What would actually settle it
The OS readout, whenever it arrives, carries more weight than a typical endpoint. If BNT113 follows the KEYNOTE-048 pattern and adds overall survival where PFS stays flat, it would vindicate a scoring framework oncology already accepts for checkpoint monotherapy in this disease. If OS also misses, the fixed-antigen mRNA approach in HPV16-driven cancer loses its lead proof point, and the trial's statistical escape hatch will have proven to be exactly what a skeptical read already suspected: a second chance rather than a second mechanism.