Models for Predicting Stage in Head and Neck Squamous Cell Carcinoma Using Proteomic and Transcriptomic Data.

Area of Science:

• Oncology
• Bioinformatics
• Genomics & Proteomics

Background:

• Late diagnosis significantly impacts survival rates for head and neck squamous cell carcinoma (HNSCC), with five-year survival ranging from 40%-66%.
• Molecular distinctions between early and advanced HNSCC stages are vital for identifying novel therapeutic targets.
• Prior bioinformatics efforts to identify diagnostic markers and predictive models for HNSCC have yielded limited or inconsistent results.

Purpose of the Study:

• To develop and evaluate computational models for predicting HNSCC stage using multi-omics data.
• To identify informative gene and protein expression signatures associated with HNSCC progression.
• To enhance the understanding of molecular mechanisms underlying HNSCC staging.

Main Methods:

• Utilized RNA sequencing (RNAseq) and reverse-phase protein array (RPPA) data from The Cancer Genome Atlas (TCGA) and The Cancer Proteome Atlas (TCPA).
• Systematically assessed individual and ensemble binary classifiers for HNSCC stage prediction.
• Employed filter and wrapper methods for feature selection to identify key molecular markers.

Main Results:

• Integrated models combining both RNAseq and RPPA data demonstrated superior performance in HNSCC stage prediction compared to single-data type models.
• Identified robust sets of informative gene and protein features that distinguish between early and advanced HNSCC stages.
• Developed well-performing classification models with potential for clinical application.

Conclusions:

• Multi-omics data integration significantly improves the accuracy of HNSCC stage prediction models.
• The identified gene and protein features offer insights into HNSCC progression and may serve as potential biomarkers.
• This study provides a foundation for developing more precise diagnostic and prognostic tools for HNSCC.