Chris James Rose1, James Patrick O'Connor1, Yvonne Watson1, Caleb Roberts1, Giovannni A. Buonaccorsi1, Susan Cheung1, Brandon Whitcher2, Geoffrey J. Parker1
1Imaging Science and Biomedical Engineering, School of Cancer and Imaging Sciences, The University of Manchester, Manchester, UK; 2MRI Modelling, Clinical Imaging Centre, GlaxoSmithKline, London, UK
Dynamic contrast-enhanced MRI and tracer kinetic modelling is increasingly common in clinical trials of anti-cancer therapies, and yields voxel-wise estimates of the parameters Ktrans, ve and vp. Most methods of quantifying the heterogeneity in DCE-MRI parameter mapssuch as the commonly-applied histogram analysisanalyse parameters individually, which may neglect important relationships between the parameters. We propose a method to allow heterogeneity to be quantified by summarising the joint distribution (histogram) of model parameter values. We apply the method to clinical trial data, showing that it is highly repeatable, very quick to compute and sensitive to known drug effect.