TMEscore

TMEscore quantifies tumor microenvironment (TME) scores to predict response to immune checkpoint blockade (ICB) therapies and to characterize TME-associated genomic, transcriptomic, and epigenetic features.

Key Features:

• Implementation: Provided as an R package for computation of TME scores.
• Scoring methods: Calculates TME scores using principal component analysis (PCA) or z-score transformations.
• Multi-omics integration: Integrates genomic, transcriptomic, and epigenetic data to quantify TME states.
• Input data types: Operates on RNA sequencing and NanoString data across multiple cohorts.
• Predictive performance: Predicts response to ICB in metastatic gastric cancer (mGC) with reported AUCs of 0.891 in a pembrolizumab phase 2 trial (NCT02589496) and 0.877 in a multicenter NanoString cohort.
• Biomarker comparison: Outperforms programmed death-ligand 1 combined positive score, tumor mutation burden, microsatellite instability, and Epstein-Barr virus status in predictive performance for mGC treated with ICB.
• Mechanistic insights: Analyzes multi-omics from TCGA-STAD and ACRG to identify mutations, metabolic pathways, and epigenetic features associated with TME profiles.

Scientific Applications:

• ICB response prediction in mGC: Predicts responses to immune checkpoint blockade in metastatic gastric cancer using RNA sequencing and NanoString-derived TME scores.
• TME mechanism characterization: Characterizes intrinsic TME mechanisms, including mutations, metabolic pathways, and epigenetic features, via TCGA-STAD and ACRG multi-omics analyses.
• Biomarker benchmarking: Benchmarks TMEscore against PD-L1 combined positive score, tumor mutation burden, microsatellite instability, and Epstein-Barr virus status for patient stratification.
• Support for precision immunotherapy: Informs stratification strategies to guide precision immunotherapy and personalized treatment approaches.

Methodology:

Computes TME scores by integrating genomic, transcriptomic, and epigenetic datasets and applying principal component analysis (PCA) or z-score transformations; analyzes TCGA-STAD and ACRG multi-omics and uses RNA sequencing and NanoString data for cohort validation.