Analysing within and between patient patient tumour heterogenity via imaging: Vemurafenib, Dabrafenib and Trametinib

In most mathematical/statistical modelling studies in Oncology researchers have focused on analysing the temporal evolution of the sum of longest diameters (SLD). SLD is used to summarise drug efficacy and plays a key role in determining which response category a patient belongs in when summarising treatment effect. Here we chose to analyse the individual longest diameters rather than the SLD.  The model used is based on historical observations and its link through to an empirical law is shown. We show this model does a reasonable job in fitting to the data.  We also show that knowing which lesion belongs to which patient improves model fits over assuming lesions are independent of each other, thus implying there is some degree of correlation in the pharmacological effect within a patient.  We then explored how the within and between patient variability in initial tumour size and decay rates across the three treatments and found that for two drugs which share the same target, Vemurafenib and Dabrafenib, can be differentiated via this analysis. This approach highlights that additional information can be gleamed from modelling individual lesions and suggests other markers for relating to drug exposure.