Search

Simple Summary Liquid biopsy offers a novel minimally invasive approach to tumor sampling and is believed to capture a comprehensive overview of the molecular tumor landscape. However, current liquid biopsy analytes in cancer are principally derived from the malignant cells without regard to the tumor microenvironment. The Stroma Liquid BiopsyTM (SLB) proteomics panel contains proteins from key stromal pathways in cancer and was designed to address the tumor microenvironment in liquid biopsy. We aimed to explore and characterize SLB panel constituents using an in-silico transcriptomics approach in colon cancer. Additionally, the association between the SLB panel constituents and histologic intratumoral stromal content, a poor prognostic tumor characteristic, was investigated. This explorative study presents an alternative workflow to gene signature development and provides a molecular characterization of the SLB panel. We believe that our findings contribute to the ever-increasing appreciation of the tumor microenvironment in cancer. Abstract Liquid biopsy has emerged as a novel approach to tumor characterization, offering advantages in sample accessibility and tissue heterogeneity. However, as mutational analysis predominates, the tumor microenvironment has largely remained unacknowledged in liquid biopsy research. The current work provides an explorative transcriptomic characterization of the Stroma Liquid BiopsyTM (SLB) proteomics panel in colon carcinoma by integrating single-cell and bulk transcriptomics data from publicly available repositories. Expression of SLB genes was significantly enriched in tumors with high histologic stromal content in comparison to tumors with low stromal content (median enrichment score 0.308 vs. 0.222, p = 0.036). In addition, we identified stromal-specific and epithelial-specific expression of the SLB genes, that was subsequently integrated into a gene signature ratio. The stromal-epithelial signature ratio was found to have prognostic significance in a discovery cohort of 359 colon adenocarcinoma patients (OS HR 2.581, 95%CI 1.567–4.251, p < 0.001) and a validation cohort of 229 patients (OS HR 2.590, 95%CI 1.659–4.043, p < 0.001). The framework described here provides transcriptomic evidence for the prognostic significance of the SLB panel constituents in colon carcinoma. Plasma protein levels of the SLB panel may reflect histologic intratumoral stromal content, a poor prognostic tumor characteristic, and hence provide valuable prognostic information in liquid biopsy.

The best current biomarker strategies for predicting response to immune checkpoint inhibitor (ICI) therapy fail to account for interpatient variability in response rates. The histologic tumor–stroma ratio (TSR) quantifies intratumoral stromal content and was recently found to be predictive of response to neoadjuvant therapy in multiple cancer types. In the current work, we predicted the likelihood of ICI therapy responsivity of 335 therapy-naive colon adenocarcinoma tumors from The Cancer Genome Atlas, using bioinformatics approaches. The TSR was scored on diagnostic tissue slides, and tumor-infiltrating immune cells (TIICs) were inferred from transcriptomic data. Tumors with high stromal content demonstrated increased T regulatory cell infiltration (p = 0.014) but failed to predict ICI therapy response. Consequently, we devised a hybrid tumor microenvironment classification of four stromal categories, based on histological stromal content and transcriptomic-deconvoluted immune cell infiltration, which was associated with previously established transcriptomic and genomic biomarkers for ICI therapy response. By integrating these biomarkers, stroma-low/immune-high tumors were predicted to be most responsive to ICI therapy. The framework described here provides evidence for expansion of current histological TIIC quantification to include the TSR as a novel, easy-to-use biomarker for the prediction of ICI therapy response.