Evaluating a data-driven approach to biomarker discovery for tumor-targeted imaging in epithelial ovarian cancer

Abstract

In epithelial ovarian cancer (EOC), surgical outcome is the strongest prognostic factor for survival. Yet, the estimation of intra-abdominal tumor load to determine the best treatment strategy for obtaining complete cytoreduction remains challenging. Additionally, visual, and tactile inspection during surgery may leave metastases undetected. Tumor-targeted molecular imaging might improve the identification of tumor cells both pre- and intra-operatively. Targeting folate receptor alpha (FRα) showed promise, however a more specific tracer is warranted. Here, we present a novel, data-driven method using RNA expression data for the identification of new target proteins for tumor-targeted imaging of EOC. A knowledge-platform was used to search through omics-databases to find membrane proteins present on EOC cells and absent or lowly expressed in surrounding tumor-negative and inflammatory cells. Differential gene expression analysis was performed to determine which genes were highly expressed in EOC tissues compared to surrounding healthy tissues. Two new genes were identified: VTCN1 and AQP5, encoding for proteins B7-H4 and AQP5, respectively. Immunohistochemical validation showed that B7-H4 expression aligned with RNA levels, indicating its potential as a new target. In contrast, there was a discrepancy in AQP5 expression at the protein level compared to its gene counterpart. While this innovative approach was valuable in identifying novel targets for tumor-targeted imaging of EOC, immunohistochemistry or cell studies remain imperative for validation of RNA expression results.