Cancer cell rendering [Adobe Stock]
Under the leadership of Timothy Chan, MD, PhD, Chair of Cleveland Clinic’s Global Center for Immunotherapy, the researchers launched an eight-year study focused on neoantigens, small peptides created when cancer cells mutate. They are a primary marker for the immune system to identify cancer cells.
Nature Medicine published the research.
The immune system and cancer cells are locked in a constant interplay. Immunotherapy treatments must navigate this complex web of influence, aiming to bolster our immune cells to target the cancer. While researchers, including Chan, have made progress in understanding these relationships, human data has traditionally remained limited.
Timothy Chan, MD, PhD, Cleveland Clinic
Bristol Myers Squibb sponsored the CheckMate-153 trial and Chan’s team oversaw the analysis. As part of the primary trial, the researchers conducted a biomarker sub-study to investigate the role of neoantigens in nivolumab’s response. This involved collecting tumor samples from patients before and three weeks after therapy and then sequencing them to identify mutations that generate neoantigens.
The study found that patients who responded well to nivolumab within three weeks of treatment experienced a significant reduction in clonal neoantigens. While those who did not achieve remission still exhibited an immune response, it was directed towards smaller, sub-clonal populations. This finding challenges the prevailing notion that non-responders are incapable of recognizing and activating against the tumor. It suggests instead that they may be mounting an immune response to neoantigens that is insufficient to eradicate all tumor clones.
Current neoantigen prediction tools largely rely on HLA-binding neoantigens, but they often miss the T cell recognition aspect of immunogenicity, according to Tyler Alban, PhD, co-first author of the study and Project Staff in the Chan Lab. To address this shortcoming, Alban, along with data scientist Prerana Parthasarathy and other team members, developed a machine-learning program. The model uses the new screening data to improve immunogenic neoantigen prediction. This program has also uncovered novel characteristics of these cancer-derived neoantigens.
Through this analysis, Alban challenged the prevailing notion in immunotherapy that a tumor needs only a single lucky mutation to become recognizable to the immune system. The findings suggest that a robust response to treatment requires a diverse population of T cells, each recognizing a different cancer-causing mutation. Chan emphasizes the importance of this type of observational research for advancing immuno-oncology.
The researchers are now working with IBM at the Cleveland Clinic-IBM Discovery Accelerator to develop more advanced AI models for predicting new cancer treatments and vaccines.
“Learning why our immune systems respond to some cancerous mutations but not others are like the holy grail for immunotherapy researchers,” he said in a press release. “Our findings are one of the closest things we have to figuring these things out.”
Filed Under: machine learning and AI, Oncology
