Cancer Cell releases a new model to predict the response of melanoma patients to immunotherapy

A research team led by the University of Sydney has recently conducted a multi-omics analysis to study which factors may affect the efficacy and drug resistance of patients with advanced melanoma undergoing immunotherapy, and published the results in the journal Cancer Cell
.

Researchers found that tumor mutation burden (TMB), neoantigen burden, IFNɣ-related genes, and T cells in the tumor microenvironment were related to the treatment response, but no specific genetic mutations were found
.

Combining two factors (TMB and IFNɣ related gene expression), they can predict immunotherapy response with high sensitivity

Senior author Professor Georgina Long from the University of Sydney said: “There are some resistance mechanisms that are common, and many are very rare
.

Ultimately, we need to match the treatment plan with the resistance-driving mechanism of each patient

In this study, Long and colleagues performed multiple analyses on pre-treatment tumor biopsy samples and matched germ cell samples from 77 patients with advanced melanoma, including whole-genome sequencing, transcriptome sequencing, methylome analysis, and Immunohistochemical analysis
.

These patients later received PD-1 antibody treatment alone, or a combination of PD-1 antibody and CTLA-4 antibody

They flagged multiple factors related to treatment response
.

For example, if the tumor mutation burden is higher, the treatment response is better

In addition, they also found that the expression characteristics of six genes related to IFNɣ were higher in good responders
.

At the same time, the tumor microenvironment of responders contains more M1 macrophages and CD8+ T cells

However, researchers have not been able to determine the specific genetic mutations associated with treatment response, although some previous studies have shown that certain somatic mutations are related to this
.

They believe that these genes and related mechanisms may play a role in the treatment resistance of individual patients, but they are not universal as a whole, indicating that the treatment resistance mechanism is heterogeneous

Using these treatment response factors, researchers developed predictive regression models
.

The best model contains genes related to TMB and IFNɣ, with a sensitivity of 89% in predicting treatment response, but a specificity of only 53%

The researchers wrote that the model's limited ability to predict drug resistance may be due to the heterogeneity of related mechanisms
.

They analyzed in-depth samples inconsistent between theory and reality, trying to explain the biological mechanism of these outliers

Long added that she and her colleagues plan to further study patients who are inconsistent with their predictions
.
"Big tissue and analysis of a large number of patients may not be the ultimate way to overcome drug resistance
.
We need to explore drug resistance based on the specific circumstances of each patient," she pointed out
.

###

Multiomic profiling of checkpoint inhibitor-treated melanoma: Identifying predictors of response and resistance, and markers of biological discordance

DOI: https://doi.
org/10.
1016/j.
ccell.
2021.
11.
012