Pharmacokinetic modelling of breast tumour physiology by dynamic contrast enhanced MRI

• Main Author: Di Giovanni, Pierluigi
• Published: University of Aberdeen 2010
• Subjects: 615.84; Breast : Cancer cells : Magnetic resonance imaging
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.540469

This work is focussed on the analysis of breast tumour physiology by pharmacokinetic modelling of dynamic contrast enhanced MRI (DCE-MRI) data. DCEMRI consists of the intravenous bolus injection of a small molecular weight contrast agent into the patient followed by the rapid acquisition of MR images across both breasts. Due to the leaky nature of the lesion microvasculature there is a greater uptake of contrast agent within the tumour than in the surrounding tissues. The dynamic contrast enhanced MR signal curve can be fitted by compartmental analysis providing information linked to the tumour’s permeability and flow. The effect of the DCE-MRI acquisition parameters on the accuracy of the estimated pharmacokinetic quantities was investigated together with the assumptions lying behind the pharmacokinetic model used for the fitting. Contrast enhanced MRI data were also examined using a fractal measure of tumour heterogeneity with the aim of assessing whether this could be a potential predictor of the tumour response to chemotherapy. Among the factors believed to play an important role in terms of tumour treatment is an increased interstitial fluid pressure (IFP) in the central areas of some large tumours. Here DCE-MRI data were analyzed in a way to see whether it could provide any information related to IFP distribution across tumour volumes. Finally, when performing quantitative DCE-MRI, particular care needs to be taken in the choice of an arterial input function (AIF) which accurately describes the passage of the contrast agent bolus at the lesion location. Here a new approach was proposed and demonstrated for the estimation of a tumour capillary input function together with lesion pharmacokinetic parameters. This was achieved by optimizing the capillary input function with a measure of the patient’s cardiac output, a parameter which is expected to vary depending on the patient’s pathology/physiology.