A new mathematical model finds that tumor neoantigen quality trumps quantity when it comes to predicting response to immunotherapy and the likelihood of long-term survival among patients with cancer.

It's the quality, not quantity, of tumor neoantigens that may best predict response to immunotherapy and the likelihood of long-term survival among patients with cancer.

A team led by Marta Łuksza, PhD, of the Institute for Advanced Study in Princeton, NJ, and Benjamin Greenbaum, PhD, of the Icahn School of Medicine at Mount Sinai in New York, NY, combined concepts from immunology, evolutionary biology, physics, and computer science to study how the immune system recognizes tumors, and how tumors mutate and evolve in response, especially in the face of checkpoint inhibition.

The researchers developed a mathematical model and tested it on three data sets—two cohorts of patients with melanoma given anti-CTLA4 therapy, and a group with non–small cell lung cancer given anti–PD-1 therapy (Nature 2017;551:517–20). They found that two main factors determine the importance of any tumor neoantigen in shaping responses to immunotherapy. First, the mutated peptide must have a greater binding affinity than its wild-type counterpart to a class I MHC molecule. T-cell receptors must then recognize the neoantigen as foreign, much as it might a pathogen, and mount an immune attack.

The likelihood of both these events happening is at the heart of the researchers’ “neoantigen fitness” model. When they compared their model against one that simply tallied up the number of mutated peptides present on the surface of tumor cells, they showed that the fitness-based analysis, by capturing both neoantigen and tumor heterogeneity, better predicted survival outcomes in patients receiving immunotherapy.

“Our approach has consistent predictive value across the patient cohorts we studied, but we think its strength comes from its universality,” Łuksza says. “It can be extended to include other factors as we learn more about how the immune system recognizes tumors under therapy.”

A companion study, led by Vinod Balachandran, MD, of Memorial Sloan Kettering Cancer Center in New York, NY, further validated this model in long-term survivors with pancreatic ductal adenocarcinoma (Nature 2017;551:512–16). All had undergone surgery and some had received adjuvant chemotherapy, but immunotherapy was not part of the treatment regimen. Initially, the team found that tumors with the highest neoantigen number and the most abundant cytotoxic T-cell infiltrates—but neither alone—stratified patients with the longest survival. Digging into possible reasons, they showed that long-term survivors displayed lasting circulating T-cell reactivity to high-quality neoantigens, as defined by the fitness model.

Because this approach worked for three different tumor types, two flavors of checkpoint inhibitor, and two different clinical settings—with or without immunotherapy—“our data may be identifying some common principles on how the immune system recognizes mutations,” says Balachandran.

Elizabeth Jaffee, MD, of the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University in Baltimore, MD, isn't yet convinced. “It's an interesting model,” she says, but points out that, in the context of pancreatic cancer, it's so far been tested only in a unique subset of patients that may not represent the population as a whole. How relevant is this going to be, she asks, for the 93% of patients who don't survive beyond 5 years with surgery alone, but may do so with newer therapeutic interventions?

Balachandran and his colleagues next plan to see if the model predicts response rates to immunotherapy among participants in the Pancreatic Cancer Action Network's Precision Promise trial. Additionally, his team is engaging with Genentech and Mainz, Germany–based BioNTech to determine how insights gleaned from this work can be applied to trials of personalized mRNA-based neoantigen vaccines.

“Pancreatic cancer is a challenge, given the relatively low number of mutations harbored by these tumors,” says Ugur Sahin, MD, BioNTech's cofounder and CEO. It's also considered a “cold” tumor, with very few infiltrating T cells. The new findings upend this conventional wisdom, Sahin says, and “strongly suggest” that even pancreatic cancer might be responsive to a neoantigen-based therapeutic strategy. –Elie Dolgin

For more news on cancer research, visit Cancer Discovery online at http://cancerdiscovery.aacrjournals.org/content/early/by/section.