A live tumor fragment platform to assess immunotherapy response in core needle biopsies while addressing challenges of tumor heterogeneity

Background: Immune checkpoint inhibitors (ICIs) have revolutionized cancer treatment, providing durable and even curative responses. However, most patients do not respond and current biomarkers (eg, programmed death 1 ligand 1 [PD-L1]), mismatch repair deficiency [dMMR]/high microsatellite instability [MSI] and tumor mutational burden) lack predictive accuracy. Ex vivo profiling of patient-derived tumor fragments shows promise as a predictive biomarker but relies on substantial surgical tissue to mitigate intra-specimen heterogeneity. Innovations are needed that address these challenges, particularly where limited tissue is available such as in core needle biopsies (CNBs). Methods: Live tumor fragments (LTFs) were generated from 59 human tumor resections and 31 CNBs from patients enrolled in observational clinical trials (ClinicalTrials.gov identifiers: NCT05478538, NCT05520099, NCT06349642) to assess cytokine induction following ICI treatment. LTFs were encapsulated in hydrogel and cultured ex vivo for up to 72 hours. A sequential treatment strategy that applies control and treatment within the same well was used with response to ICI or αCD3/αCD28 assessed using a multiplex secretome assay. Viability was assessed using established metabolic assays and dynamic optical coherence microscopy. Results: LTFs maintained viability and retained T cells responsive to stimulation throughout ex vivo culture. Multiplex immunofluorescence and immunohistochemistry showed key components of the tumor microenvironment, including relative proportions of CD4⁺ and CD8⁺ immune cell populations, were preserved. Specimens positive for PD-L1 or dMMR/MSI-high were enriched for cytokine upregulation, including T-cell response cytokines IFNγ and CXCL10, after αPD-1 treatment. To demonstrate clinical applicability of the sequential treatment strategy, CNBs from patients with lung, gastrointestinal or kidney cancer were profiled and differential cytokine induction in response to ICI treatment was observed. Conclusions: The novel ex vivo platform presented is capable of detecting T-cell response to ICI treatment by using a sequential treatment strategy. This approach addresses challenges associated with cross-well heterogeneity in tissue composition and requires half as much tissue as a cross-well comparison, mitigating tissue limitations typically associated with non-surgical biopsies. Importantly, the platform is compatible with established functional assays as well as non-destructive spatial imaging, enabling researchers to characterize response to ICI longitudinally. Ongoing trials will enable clinicians to assess platform performance in predicting response to immunotherapy.