Oxygen-enhanced MRI can accurately identify, quantify and map tumour hypoxia in preclinical models

Abstract

Aim: There is need for non-invasive methods to identify, quantify and map tumour hypoxia. In this study we used an emerging technology - R1 oxygen enhanced MRI (OE-MRI) - to distinguish those tumour sub-regions that respond to hyperoxic gas challenge from refractory sub-regions. We hypothesised that the proportion of refractory tumour tissue (Oxy-R) would be a robust biomarker of tumour hypoxia across multiple models with different vascular and hypoxic phenotypes. Methods: OE-MRI signal precision, stability and relationship to tissue pO2 were evaluated in well vascularised renal cancer 786-O xenografts. Dynamic sensitivity of proportional Oxy-R to acute changes in hypoxia was evaluated using hydralazine challenge. Relationship of proportional Oxy-R to tissue immunohistochemistry and gadolinium DCE-MRI were explored in parental and drug-resistant 786-O models and in SW620 xenografts. Results: Phantom and in vivo experiments demonstrated the accuracy, precision and stability of R1measurement. The proportion of tumour Oxy- R increased significantly following hydralazine challenge (p=0.045) relative to control. The proportion of tumour with perfused Oxy-R voxels was correlated to chronic hypoxia in well perfused 786-O-R xenografts(rho 0.810, p=0.028) and in relatively necrotic SW620 xenografts (rho 0.929, p=0.002). Conclusion: The proportion of tumour perfused Oxy-R is a robust biomarker of tumour hypoxia. Voxel-wise analysis of dual oxygen and gadolinium challenge has potential to quantify and map tumour hypoxia as prognostic, predictive and pharmacodynamic biomarkers that could facilitate personalised healthcare.