Baseline gene expression profiling determines long-term benefit to programmed cell death protein 1 axis blockade

Area of Science:

• Oncology
• Immunology
• Genomics

Background:

• Immune checkpoint inhibitors, particularly anti-programmed cell death protein 1 (PD-1) therapies, have significantly improved outcomes for advanced melanoma patients.
• However, reliable biomarkers to predict individual responses to immunotherapy are currently lacking.
• Predicting long-term benefit remains a critical unmet need in managing advanced melanoma.

Purpose of the Study:

• To identify gene expression profiles that predict long-term benefit (LTB) from PD-1-based immunotherapy in advanced melanoma.
• To develop and validate a gene signature capable of forecasting sustained progression-free survival (PFS) of at least 24 months.
• To explore the dynamic changes in predictive gene signatures across different time intervals of clinical benefit.

Main Methods:

• Transcriptomic data from 105 baseline tumor samples of advanced melanoma patients treated with PD-1 inhibitors were analyzed.
• Elastic net binomial regression with cross-validation was used to correlate gene expression with PFS at 6, 12, 18, and 24 months.
• Lasso regression was employed to construct a predictive signature for LTB (24-month disease-free survival).

Main Results:

• Baseline gene expression profiles accurately predicted long-term immunotherapy outcomes.
• A distinct set of genes predicted 24-month benefit, differing from those predicting earlier outcomes.
• A 12-gene signature achieved an AUC of 0.92 in the training set and 0.74 in the validation set for predicting LTB.

Conclusions:

• A novel 12-gene signature can predict long-term benefit in advanced melanoma patients receiving anti-PD-1 therapy.
• This signature may enhance the assessment of immunotherapy response beyond objective response criteria.
• The findings suggest that baseline gene expression can characterize the potential for sustained antitumor immune responses.