Development of a Novel Immunohistochemistry Assay to Characterize EDB+FN Protein Expression in Solid Tumors

Tumor stroma, the cellular and non-cellular support system surrounding cancer cells, is a promising source of novel therapeutic targets. Extradomain-B of fibronectin (EDB+FN), a tumor-specific antigen expressed in the extracellular matrix (ECM) of tumor stroma and minimally in healthy tissues, is the target of the first-in-concept antibody drug conjugate (ADC) micvotabart pelidotin (MICVO) which is currently in clinical trials (NCT05720117 and NCT06795412) for the treatment of solid tumors. A specific and reproducible immunohistochemistry (IHC) assay and scoring approach are needed to detect EDB+FN protein expression for indication selection, preclinical research, and potential future evaluation of correlations between target expression and therapeutic efficacy.

Here, an EDB+FN IHC assay was developed to detect EDB+FN protein in human solid tumors. A novel scoring system was developed adapting H-score methodology to evaluate intensity and localization of EDB+FN extracellularly in tumor stroma and cellularly in cancer cells in ten solid tumor types. Moreover, a custom, AI-powered digital pathology algorithm was developed to streamline scoring of EDB+FN protein expression for research and future analysis of spatial distribution.

The developed IHC assay reproducibly detected EDB+FN in the tumor stroma across FFPE samples from ten tumor types and minimally in normal tissue. Percent stroma was not predictive of EDB+FN expression within a tumor, suggesting heterogeneous EDB+FN protein expression in tumor stroma. The modified H scoring approach reproducibly quantified EDB+FN expression extracellularly in tumor stroma by pathologist evaluation and a digital algorithm built with open-source tools demonstrated proof-of-concept for digital quantification of EDB+FN protein expression.

This work demonstrates technical characterization of a research-use IHC assay for detection of EDB+FN in FFPE tumor sections, reveals broad expression of EDB+FN across cancer types, and establishes a novel approach for scoring expression of a non-cell associated protein in the tumor ECM, both by pathologists and using digital pathology.