Xerna™ TME Panel is a machine learning-based transcriptomic biomarker designed to predict therapeutic response in...

Area of Science:

• Oncology
• Biomarker Development
• Computational Biology

Background:

• Current predictive biomarkers often focus on single analytes, limiting their clinical utility.
• There is a need for broad, multi-therapeutic predictive biomarkers.
• The tumor microenvironment (TME) plays a critical role in therapeutic response.

Purpose of the Study:

• To develop and validate a novel, pan-tumor RNA expression-based biomarker, the Xerna™ TME Panel.
• To predict patient response to multiple TME-targeted therapies, including immunotherapies and anti-angiogenic agents.
• To classify tumors into distinct TME subtypes.

Main Methods:

• An artificial neural network (ANN) algorithm was trained on a 124-gene signature from 298-patient solid tumor data.
• The model identified four TME subtypes: Angiogenic (A), Immune Active (IA), Immune Desert (ID), and Immune Suppressed (IS).
• The classifier's performance was evaluated in four independent clinical cohorts across gastric, ovarian, and melanoma datasets.

Main Results:

• The TME Panel successfully discriminated four distinct stromal phenotypes based on angiogenesis and immune axes.
• The classifier demonstrated significant enrichment (1.6-to-7-fold) of clinical benefit across multiple therapeutic hypotheses.
• The Panel outperformed PD-L1 and MSI-H biomarkers in specific metrics for gastric cancer immunotherapy response prediction.

Conclusions:

• The Xerna™ TME Panel exhibits strong performance across diverse datasets and therapeutic modalities.
• The biomarker's ability to predict response to multiple TME-targeted therapies suggests broad clinical utility.
• The findings support the potential use of the TME Panel as a clinical diagnostic tool for various cancer types.