Credit: Macmillan Publishers Limited

The majority of patients with pancreatic ductal adenocarcinoma (PDAC) die within 5 years of diagnosis, but around 7% of them live longer. Two studies now published in Nature show that predictive models based on cancer neoantigens can be developed; the use of such models revealed that similar immune-response-related mechanisms underlie long-term survival in patients with PDAC, as well as in patients with metastatic melanoma or lung cancer treated with immune-checkpoint inhibitors.

Marta Łuksza explains that, in the first study, researchers “adopted an evolutionary approach to predict how a tumour will evolve in the presence of an activated immune system. First, we reconstructed the genealogical tree of the tumour. We then predicted how the tumour, composed of different clones, would be affected by immune-checkpoint inhibition. For that, we attempted to quantify the factors that determine the fitness of different clones.” In the mathematical model developed by Łuksza and colleagues, the size of a cancer-cell population in a tumour depends on the immune interactions that define the fitness of that clone, and on its initial frequency. Of note, the fitness of each clone was defined by the neoantigen with the highest immune-recognition potential. This full-clonal model was compared with another one, in which uniform immune recognition was assigned to all neoantigens, in two cohorts of patients with melanoma and one cohort of patients with non-small-cell lung cancer. For all data sets, the full-clonal model enabled better prediction of long-term and short-term survival.

“The approach developed with this model could be expanded to other tumour types as we learn more about how the immune system recognizes tumours under therapy,” Łuksza comments. Indeed, a similar approach was adopted in the second study. As Vinod Balachandran describes, “tumour samples from 82 patients with PDAC who survived an average of 6 years were compared with those from 68 patients with PDAC who survived <1 year. Importantly, all patients in both groups had received the same treatments, namely surgery and chemotherapy.”

our results point to common, quantifiable principles of immune recognition of cancer mutations

In comparison to tumours from patients with short survival durations, tumours from long-term survivors had higher densities of CD8+ T cells, dendritic cells, Treg cells and macrophages, lower densities of CD4+ T cells, and upregulation of immune-response-related markers in the microenvironment (such as PD-1 and TIGIT). The longest survival durations were associated with the presence of both the highest neoantigen burden and degree of CD8+ T-cell infiltration, indicating that neoantigen quality — that is, an elevated immune-recognition potential — is associated with survival. Balachandran and colleagues postulated that neoantigen homology to infectious-disease-derived peptides is a partial surrogate of neoantigen 'non-selfness' because it reflects a greater likelihood of immune recognition. The analysis of peripheral blood mononuclear cells from seven patients with a median overall survival of 10.5 years revealed the expansion of CD8+ T cells reactive to cancer-specific neopeptides and their homologous infectious-disease-derived peptides, but not to their wild-type peptides.

Balachandran explains that “T cells were considered largely ineffective at controlling pancreatic cancer. These studies suggest, however, that T cells might have a key role in the exceptional survival outcomes of a subgroup of patients with PDAC and other cancer types.” For Benjamin Greenbaum, who was involved in both studies, “our results point to common, quantifiable principles of immune recognition of cancer mutations. We need a deeper understanding of how the immune system interacts with cancers.”