AI Platform Identifies New Vaccine Targets for Glioblastoma
Evaxion A/S announced data showing its AI-Immunology platform can design personalized vaccines for glioblastoma, a deadly brain cancer with few treatment options. The company will present findings at the American Association for Cancer Research Annual Meeting on April 22, 2026.
Working with Duke University School of Medicine, Evaxion identified endogenous retroviruses (ERVs) in tumor samples from multiple glioblastoma patients. These viral sequences, dormant in the human genome, appear in tumors but not in healthy tissue, making them attractive vaccine targets.
Why This Matters for Product Development
Glioblastoma poses a specific challenge for vaccine makers: patients typically have low mutational burden, meaning fewer cancer-specific mutations to target. Traditional personalized cancer vaccines rely heavily on neoantigens derived from these mutations. Fewer mutations limit the number of targets available.
ERVs solve this problem by opening a second source of tumor-specific antigens. The AI platform can identify which ERVs are active in individual tumors, then combine those targets with whatever neoantigens are present in each patient's cancer.
Birgitte Rønø, chief scientific officer at Evaxion, said: "We are excited by these new findings confirming that AI-Immunology is a unique platform for designing vaccines for many different cancer types, combining neoantigens with other types of antigens."
The Data
Evaxion's analysis profiled ERV expression across patient samples and assessed vaccine design feasibility. Experimental work showed the discovered ERV antigens trigger immune responses. The company identified ERVs as viable antigen sources in multiple patients and found neoantigens of sufficient quality in most patients as well.
Mustafa Khasraw, who studies tumor immunobiology at Duke University School of Medicine, said: "There is a dire need for better treatment options for glioblastoma, which is refractory to immunotherapy due to low mutational burden and paucity of canonical neoantigens. We are encouraged to have found a new tumor-specific antigen source in ERVs that can be targeted by vaccines alongside mutation-derived epitopes."
What's Next
The abstract will be presented as a poster during the immunology session at the AACR meeting in San Diego. Evaxion plans to discuss the data with scientific and business partners at the conference.
The findings extend the platform's application beyond the company's current pipeline, which includes vaccines for melanoma and other cancers. For product teams evaluating AI-driven drug discovery platforms, the result demonstrates how machine learning can identify non-obvious antigen sources that human analysis might miss.
Your membership also unlocks:
