Metabolic and proteomic signatures differentiate inflammatory phenotypes from cancer and predict treatment response in patient sera

Abstract

Tumors shift their metabolic needs to enable uncontrolled proliferation. Therefore, metabolic assessment of cancer patient sera provides a significant opportunity to noninvasively monitor disease progression and enable mechanistic understanding of the pathways that lead to response. Here, we show Raman spectroscopy (RS), a highly sensitive and label‐free analytical tool, is effective in metabolic profiling across diverse cancer types in patient sera from both pancreatic ductal adenocarcinoma (PDAC) and locally advanced rectal cancer (LARC). We also combine metabolic data with proteomic signatures to predict treatment response. Our data show RS peaks successfully differentiate PDAC patients from healthy controls. Peaks associated with sugars, tyrosine, and DNA/RNA distinguish PDAC patients from chronic pancreatitis, an inflammatory condition that is notoriously difficult to discern from PDAC via current clinical approaches. Furthermore, our study is expanded to investigate response to chemoradiation therapy in LARC patient sera where at pre‐treatment multiple metabolites including glycine, carotenoids, and sugars are jointly correlated to the neoadjuvant rectal (NAR) score indicative of poor prognosis. Via classical univariate AUC–ROC analysis, several RS peaks were found to have an AUC>0.7, highlighting the potential of RS in identifying key metabolites for differentiating complete and poor responders of treatment. Gene set enrichment analysis revealed enrichment of metabolic, immune, and DDR‐related pathways associated with CRT response. Notably, RS‐derived metabolites were significantly correlated with multiple immune signaling proteins and DDR markers, suggesting these distinct analytes converge to reflect systemic changes within the tumor microenvironment. By integrating metabolic, proteomic, and DDR data, we identified pre‐treatment activation of galactose and glycerolipid metabolism, and post‐treatment engagement of cell cycle and p53 signaling pathways. Our findings show that RS, when integrated with complementary protein marker analysis, holds the potential to bridge the translational divide enabling a clinically relevant approach for both diagnosis and predicting response in patient samples.