Considering Both Tumor and Immune Features May Improve Prediction of Melanoma Outcomes After Checkpoint Inhibition
Key findings
- In this study, DNA and RNA were sequenced from melanoma samples obtained before and after checkpoint blockade (CPB) in an attempt to identify factors that predict response to CPB and overall survival
- Some factors that predicted those outcomes were (a) tumor purity and tumor mutational burden, (b) T- and B-cell infiltration combined with tumor mutational burden from DNA and (c) gene expression subtyping of tumors
- Expression of MAP4K1 + TBX3, as well as some other gene pairs involving an immune-associated and a tumor-associated gene, also improved the ability to predict outcomes
- Including T-cell receptor and immunoglobulin sequences in targeted sequencing panels, along with genes that allow estimation of tumor mutational burden, may be useful for predicting outcomes after CPB for melanoma using a single DNA assay
Integrative models, such as those that consider both tumor mutational burden and immune expression signatures, have improved the ability to predict the outcome of checkpoint blockade (CPB) for melanoma. However, these efforts require large samples, multiple assays and multiple nucleic acid isolation techniques.
Subscribe to the latest updates from Cancer Center Advances in Motion
To examine how prediction methods might be improved, Nir Hacohen, PhD, of the Center for Cancer Immunology at the Mass General Cancer Center, and colleagues sequenced DNA and RNA from samples of metastatic melanoma before and after patients received CPB. In Cell Reports Medicine, they report several pre-treatment features that were associated with treatment response and overall survival (OS).
Tumor Mutational Burden and DNA Measures
Single-gene assessments weren't worthwhile, but analyses of tumor purity and tumor mutational burden improved predictions of CPB outcomes. In addition, by using rearranged T-cell receptor and immunoglobulin sequences, it was possible to quantify T- and B-cell infiltration, which was associated with OS.
Considering total mutational burden and DNA-based tumor cell burden together also identified patients with a higher chance of benefiting from CPB.
Melanoma Subtype and Survival
Based on RNA sequencing of 469 other melanoma specimens, five tumor subtypes were identified: Immune (high levels of immune infiltrate), Keratin-high, MITF-high (high melanocyte differentiation), MITF-low and MITF-intermediate.
Tumor subtypes were not associated with response, but they were significantly associated with OS. The Immune subtype was linked to the longest OS.
Gene Pairs
The researchers distinguished 55 genes associated with long OS, mostly expressed in immune cells, and 28 genes associated with short OS, mostly expressed in melanoma cell lines. They created predictive models that paired immunity genes with tumor-associated genes:
- A metagene-pair model, which averaged the expression of the 55 long-OS genes as one metagene and the 28 short-OS genes as the other metagene, was highly predictive of response and OS
- The three gene pairs most significantly associated with OS and response were MAP4K1 + TBX3, MAP4K1 + AGER and the metagene pair
- Additionally considering tumor mutational burden significantly improved the ability of MAP4K1 + TBX3 and MAP4K1 + AGER to predict OS
- The predictive accuracy of the top gene pairs was validated in two independent cohorts of patients receiving CPB (n=180)
The Path Forward
Including T-cell receptor and immunoglobulin sequences in targeted sequencing panels, along with genes that allow estimation of tumor mutational burden, may be useful for predicting outcomes after CPB for melanoma using a single DNA assay.
Clinical trials should examine whether patients with multiple positive prognostic factors might be best suited to PD-1/PD-L1 monotherapy, whereas those with negative factors may need more aggressive combination therapy.
view original journal article Subscription may be required
Learn about the Center for Cancer Immunology
Refer a patient to the Mass General Cancer Center