High-throughput proteomics profiling-derived signature associated with chemotherapy response and survival for stage II/III colorectal cancer

Abstract

Adjuvant chemotherapy (ACT) is usually used to reduce the risk of disease relapse and improve survival for stage II/III colorectal cancer (CRC). However, only a subset of patients could benefit from ACT. Thus, there is an urgent need to identify improved biomarkers to predict survival and stratify patients to refine the selection of ACT. We used high-throughput proteomics to analyze tumor and adjacent normal tissues of stage II/III CRC patients with /without relapse to identify potential markers for predicting prognosis and benefit from ACT. The machine learning approach was applied to identify relapse-specific markers. Then the artificial intelligence (AI)-assisted multiplex IHC was performed to validate the prognostic value of the relapse-specific markers and construct a proteomic-derived classifier for stage II/III CRC using 3 markers, including FHL3, GGA1, TGFBI. The proteomics profiling-derived signature for stage II/III CRC (PS) not only shows good accuracy to classify patients into high and low risk of relapse and mortality in all three cohorts, but also works independently of clinicopathologic features. ACT was associated with improved disease-free survival (DFS) and overall survival (OS) in stage II (pN0) patients with high PS and pN2 patients with high PS. This study demonstrated the clinical significance of proteomic features, which serve as a valuable source for potential biomarkers. The PS classifier provides prognostic value for identifying patients at high risk of relapse and mortality and optimizes individualized treatment strategy by detecting patients who may benefit from ACT for survival.