The BNT122-01 Phase II trial of autogene cevumeran in resected colorectal cancer has been terminated after an independent data safety monitoring board identified a numerical imbalance in overall survival favoring the control arm — a finding that goes beyond the futility signal the board had flagged in October 2025 and raises more substantive questions about the vaccine's efficacy in this setting. BioNTech SE (Nasdaq: BNTX) announced the decision in consultation with Genentech, a member of the Roche Group, which co-develops the drug.
The trial enrolled patients with circulating tumor DNA (ctDNA)-positive, surgically resected Stage II (high risk) or Stage III colorectal cancer (CRC) and compared autogene cevumeran as adjuvant monotherapy against watchful waiting — the current standard of care. When the data safety monitoring board (DSMB) crossed the futility boundary in October 2025, it concluded the data were too immature to support reliable efficacy conclusions and did not recommend stopping the trial. At its most recent review, the DSMB found the numerical OS imbalance and determined that further continuation was unlikely to change the outcome. No new safety signals were identified. BioNTech said full trial data will be shared with the scientific community at a future date; the magnitude, statistical significance, and causes of the OS imbalance have not been disclosed.
Autogene cevumeran is a uridine mRNA–lipoplex nanoparticle vaccine encoding patient-specific tumor neoantigens identified through tumor sequencing, designed to train T cells to recognize and attack residual cancer cells. The monotherapy design in BNT122-01 was a deliberate test of whether the vaccine alone, without checkpoint inhibitor support, could prevent recurrence in a molecularly selected population.
The failure reflects a well-established immunological obstacle. CRC is predominantly microsatellite stable (MSS), a subtype characterized by low mutational burden and an immunosuppressive tumor microenvironment that has consistently resisted checkpoint inhibition. Using ctDNA positivity as a selection biomarker was intended to enrich for patients at highest recurrence risk, but that enrichment strategy could not overcome the fundamental immunological hostility of MSS CRC to T cell-based approaches. The OS imbalance, if confirmed in the final analysis, would represent a more concerning outcome than futility alone.
The termination is the second discontinuation for autogene cevumeran in a solid tumor indication outside pancreatic cancer. Earlier in 2026, BioNTech and Roche discontinued the Phase II IMcode004 trial evaluating autogene cevumeran plus nivolumab in adjuvant muscle-invasive urothelial carcinoma, citing a rapidly shifting treatment landscape.