AI Drives Development of Personalized Cancer Vaccines, Reducing Melanoma Recurrence Risk by 42%

·by Henderson·AI
AI Drives Development of Personalized Cancer Vaccines, Reducing Melanoma Recurrence Risk by 42%

The rapid integration of AI into daily life is evident, and its transformative impact in the medical field is particularly noteworthy. Researchers can now leverage AI to understand cancer at a level of detail unimaginable a decade ago: rather than applying the same treatment to all patients with the same type of cancer, the focus is shifting towards therapies based on the unique genetic makeup of each individual's tumor.

In October of this year, Moderna and Merck announced the results of a Phase 3 clinical trial for a personalized cancer vaccine. When used in combination with the immunotherapy drug Keytruda, the vaccine reduced the relative risk of melanoma recurrence or death by 42% compared to Keytruda alone. It is important to note that this vaccine has not yet received FDA approval, and the trial itself cannot prove that generative AI invented this therapy, but the results do demonstrate the potential of personalized cancer treatment.

AI is also addressing another half of the challenge: two individuals with the same diagnosis may respond very differently to the same drug, as cancer is not a single disease, and seemingly similar tumors may carry different mutations. A study conducted over the past six months has shown that researchers are increasingly able to examine the biological characteristics of specific patients' tumors, rather than relying solely on patterns derived from thousands of patients. The PERCEPTION tool from the National Cancer Institute is one example, analyzing gene activity within individual cancer cells to predict whether certain drugs might be effective.

AI's Potential in Cancer Treatment Continues to Grow

This is crucial—within tumors, there often exist different cell populations, and drugs may only kill a portion of them, leaving the remaining cells to cause cancer recurrence or develop resistance. In drug discovery, Google DeepMind's AlphaFold offers another pathway. Proteins determine how cells function and are involved in the onset and spread of cancer; understanding their three-dimensional structure helps predict how potential drugs will interact with them. Previously, this work could take years in the lab, but AlphaFold can provide numerous structural predictions in just minutes, and DeepMind has made over 200 million protein structures available to researchers worldwide.

This does not mean it can automatically create safe and effective anti-cancer drugs; predictions still require validation through experiments and patient-level verification, but it can narrow down the candidate pool and save time in early development. For example, the melanoma vaccine intismeran autogene is designed based on genetic mutations in the patient's tumor, which help identify cancer-specific proteins that the immune system can recognize and remember. In a Phase 3 trial involving 1,137 patients who had undergone surgical removal of melanoma, data from the Memorial Sloan Kettering Cancer Center indicated that 82% of those receiving the vaccine in combination with Keytruda did not experience recurrence at two years, compared to 70% in the group receiving Keytruda alone.

There is still a distance between discovering a promising therapy and patients actually using it. The targets identified by AI or the efficacy predictions made need to be validated; clinical trials are irreplaceable. Additionally, the costs of genetic testing, accessibility of specialized institutions, and whether AI systems are stable and reliable across different patient populations are all practical barriers. However, the direction is becoming clearer: AI is providing scientists with new tools to research cancer, identify drug targets, and determine which treatments are best suited for which patients. The goal is not to replace oncologists, but to equip them with more comprehensive information and more options.

H
About the author
Henderson