Plasma Proteome-Based Test for First-Line Treatment Selection in Metastatic Non-Small Cell Lung Cancer.

Area of Science:

• Oncology
• Proteomics
• Machine Learning

Background:

• Current metastatic non-small cell lung cancer (NSCLC) treatment relies on PD-1/PD-L1 inhibitors, often with chemotherapy, but overlooks individual patient variability and immune factors.
• This generalized approach can lead to suboptimal outcomes for patients with NSCLC lacking driver mutations.

Purpose of the Study:

• To develop and validate a machine learning algorithm utilizing pretreatment plasma proteomic profiles for personalized treatment selection in metastatic NSCLC.
• To address the limitations of current guidelines by incorporating individual patient data for more precise therapeutic decisions.

Main Methods:

• A multicenter observational trial included 540 patients receiving PD-1/PD-L1 inhibitor-based therapy and 85 receiving chemotherapy.
• Pretreatment plasma proteome profiling was conducted using the SomaScan Assay v4.1.
• A machine learning algorithm was developed and blindly validated using this proteomic data.

Main Results:

• The developed test, PROphet, showed high concordance between predicted and observed clinical benefit (R² = 0.98, P < .001) for PD-1/PD-L1 inhibitor treatments.
• PROphet categorizes patients as PROphet-positive or PROphet-negative, stratifying outcomes beyond PD-L1 expression.
• PROphet-negative patients with high PD-L1 (≥50%) had improved survival with combination therapy (chemo + immunotherapy) vs. immunotherapy alone (HR, 0.23; P = .0003).

Conclusions:

• Plasma proteome-based testing, combined with PD-L1 testing, identifies distinct patient subsets with varying responses to PD-1/PD-L1 inhibitor therapies.
• This personalized approach has the potential to enhance the precision of first-line treatment selection for metastatic NSCLC.
• The PROphet test offers a novel strategy for optimizing treatment decisions in NSCLC management.