PhaseX: Patient tumor avatars for evaluating anticancer therapeutics.

Abstract

2657 Background: Current preclinical tumor platforms, such as in vitro 2D cell cultures and organoid models, fail to fully recapitulate the complexity of the tumor microenvironment, including critical components like the extracellular matrix and immune interactions. While humanized patient-derived xenograft models address some of these limitations, they are costly, low-throughput, artificial, and technically challenging to establish. PhaseX (Patient-derived hydrogel-assisted eXplants) presents a robust alternative, preserving the native TME, including cellular diversity, gene expression, and immune landscapes, for at least seven days. This study utilizes PhaseX to assess patient-specific responses to immune checkpoint blockade (ICB), chemotherapeutics, and targeted therapies while providing insights into their mechanisms of action. Methods: Fifteen fresh patient-derived tumor explants (PDTEs) from HNSCC patients were embedded in bioengineered hydrogel and treated ex vivo with pembrolizumab. Supernatants were collected at 2 and 4 days post-treatment for immunoassay analysis. Tumor explants were either dissociated for high-dimensional flow cytometry or processed into FFPE sections for immunofluorescence analysis at 2 and 5 days. Additionally, ten PDTEs representing various cancer types (peritoneal, colorectal, sarcoma, lung and ovarian) were used to evaluate dose-dependent responses to commonly used chemotherapeutics, including cisplatin, doxorubicin, and erlotinib. Six PDTEs were treated with plasminogen activator inhibitor-1 (PAI-1), and their post-treatment metabolic activity was assessed using the resazurin cell viability assay. Results: This study highlights the importance of capturing temporal dynamics in ex vivo tumor models to accurately predict pembrolizumab responses, achieving 100% sensitivity and specificity in HNSCC patients. Increased IFN-γ secretion and upregulation of chemokines (CXCL9, CXCL10, CXCL11) distinguished responders, alongside elevated cytotoxicity markers (perforin, granulysin, sFasL, sFas) contributing to cancer cell death. Responders exhibited reduced terminally exhausted CD8+ T cells (PD-1 + TIM3 + ), allowing reinvigoration of functional CD8+ T cells, while non-responders showed elevated Tox + CD38 + levels, indicating resistance to PD-1 blockade. Spatial analysis revealed greater T cell infiltration in responders, facilitating tumor-cell interactions. Additionally, the PhaseX platform demonstrated its utility in evaluating dose-dependent responses across multiple tumor types (sarcoma, colorectal, lung, cervical) and identified patient-specific responses to PAI-1 inhibition. Conclusions: The PhaseX platform accurately predicts patient-specific responses for ICB, chemotherapeutics and targeted therapy across multiple tumor types. These findings establish PhaseX as a valuable tumor platform to evaluate anticancer therapeutics.