A 19-protein signature to predict outcomes in a cohort of stage II and III CRC.

Abstract

3129 Background: Colorectal cancer (CRC) is the second leading cause of cancer-related deaths worldwide and remains an area of unmet clinical need. Despite advances in treating stage II and III disease, 30-50% of patients relapse. Additionally, variable survival outcomes are seen in patients with similar disease stages, making it difficult to predict relapses. Here we outline some of the proteomic pathways of early-stage CRC along with the identification of a proteomic-based prognostic signature that predicts relapse in stage II/III CRC. Methods: Archival formalin fixed paraffin embedded tumour (TUM) specimens and matched normal adjacent tissues (NAT) from a cohort of 495 CRC patients (with upfront resected stage II and III CRC) were prepared for proteomic analyses using data-independent acquisition mass spectrometry (DIA-MS). TUM samples with < 20% tumour were excluded from the analysis. Differential expression analyses were performed to identify the top differentially expressed proteins (DEP) within TUM. DEP were subjected to crosstalk and pathway enrichment analysis (PEA). Survival analysis based on initial univariate and subsequent 100 runs of multivariate Cox regression with Least Absolute Shrinkage and Selection Operator (LASSO) was performed to obtain a reduced list of candidate proteins associated with recurrence-free survival (RFS). The proteins that appeared in greater than 95% of the LASSO runs were used in a multivariate Cox model with recursive feature selection to yield the final list of 19 proteins. A risk score was calculated as the sum of the intensities of each of the 19 proteins multiplied by the corresponding regression coefficients. The median of the risk scores was used as the threshold to dichotomise patients into either a low- or a high-risk group. Results: 399 TUM and 381 NAT were analysed, showing 799 DEP, with 529 upregulated and 270 downregulated in TUM vs NAT, respectively. PEA of the upregulated proteins showed enrichment of pathways related to extracellular matrix organisation, tRNA processing and collagen formation. PEA of the downregulated DEPs revealed pathways related to smooth muscle contraction, focal adhesion and oxidation. The 19-protein signature dichotomised patients into low- and high-risk groups with a C-index of 0.829. The signature was prognostic for RFS (hazard ratio (HR) 2.7, 95% confidence interval (CI): [2.24 – 3.18]) and for overall survival (OS) (HR 1.6, 95%CI [1.40, 1.76]). The prognostic significance of the signature was maintained after adjusting for other clinical variables for both RFS and OS (both p<0.001) using multivariate Cox regression. Conclusions: Our findings show that CRC disrupts the extracellular matrix and upregulates collagen formation, making these potential targets for novel therapies. We have also identified a 19-protein signature that is associated with survival in our cohort independent of other clinical variables.