Clinical Lecturer in Medical Oncology University of Cambridge Cambridge, England, United Kingdom
Background: The presence of venous tumour thrombus (VTT) provides a particular challenge in the management of locally advanced kidney cancer. Extensive surgery is needed to achieve complete resection, but this is associated with high morbidity and significant relapse rates. The NAXIVA trial (NCT03494816) demonstrated the feasibility and safety of using eight weeks of pre-operative Axitinib to downstage VTT (Stewart et al. Br J Cancer 2022). Here we present updated results of multiparametric assessment of the tumour microenvironment and peripheral blood in NAXIVA trial patients.
Methods: Pre- and post-treatment MRI scans were assessed for percentage change in VTT length. Response was defined as a ≥30% reduction in VTT length following treatment. Samples were analysed by multiplex immunofluorescence, automated image analysis, flow cytometry, cytokine array, RNA-seq and DNA hotspot panel.
Results: 7/20 patients achieved a response of ≥30% reduction in VTT length. Responders had a significantly higher baseline microvessel density, which reduced to the levels seen in non-responders following treatment. Responders had lower baseline levels of angiogenic factors, with significant rises in pro-angiogenic cytokines upon treatment. In non-responders there were trends to higher numbers of exhausted T-cells in the tumour and T-effectors in blood. Non-responders had significantly higher blood levels of cytokines associated with cytotoxic T cell response.
Conclusions: Responders had a pro-angiogenic phenotype. Treatment induced increases in angiogenic factors, potentially as an escape mechanism from VEGF pathway blockade. Non-responders had a CD8+ T-cell high phenotype observed in the tumour and blood. Considering measures together may give better understanding of treatment response; for example, responders had low T-cell numbers in both the tumour AND blood, whereas non-responders had high T-cell numbers in either the tumour OR blood. We are performing multiparametric data integration and modelling to identify the key predictors of treatment response.