Immunogram defines four cancer-immunity cycle phenotypes with distinct clonal selection patterns across solid tumors.

Area of Science:

• Oncology
• Immunology
• Genomics

Background:

• The cancer-immunity cycle (CI cycle) framework illustrates anticancer immune responses.
• Tumor immunological phenotype is crucial, yet pan-cancer phenotypes based on CI cycle are lacking.

Purpose of the Study:

• To visualize and classify pan-cancer immunological phenotypes using the CI cycle theory.
• To investigate the relationship between tumor immune phenotypes and evolutionary genomics.

Main Methods:

• Applied a 'cancer immunogram' visualization method to 8460 solid tumors from TCGA.
• Utilized nonnegative matrix factorization (NMF) for unsupervised clustering of cancer immunograms.
• Employed evolutionary genomics (dN/dS ratio) to assess clonal selection in distinct immunogram subtypes.

Main Results:

• Defined four major CI cycle patterns (immunograms) across 32 cancer types.
• Immunogram-I: 'hot' and 'exhausted' (favorable prognosis).
• Immunograms II-IV: distinct 'cold' immunosuppressive patterns with varying features and prognoses, with Immunogram-II showing the worst prognosis.
• Immunogram-I and -IV showed favorable survival in metastatic urothelial cancer post-immune checkpoint inhibitor (ICI) therapy.
• Immunogram-I correlated with higher response rates to ICI therapy in metastatic gastric cancer.

Conclusions:

• The study provides novel insights into cancer-immunity interactions and tumor evolution.
• Identified distinct tumor immune phenotypes that can inform the optimization of immunotherapy strategies.
• Cancer immunograms offer a valuable tool for understanding tumor immunobiology and predicting treatment response.