Towards model-based personalised medicine in oncology: Using biomarkers to predict clinical outcome

Abstract

The use of personalised medicine in oncology is gaining recognition as a way of enabling individualised tailored-treatment. Circulating biomarkers have been proposed to provide early indications of treatment and thus to support individualised disease monitoring. The field of pharmacometrics is a potentially useful discipline here which focuses on obtaining quantitative mathematical and statistical models of the different physiological processes from drug administration to measurement of drug exposure, biomarker response and clinical outcome.

In this thesis we show two different examples of the use of circulating biomarkers and pharmacometric tools to facilitate personalised medicine. The first section proposes a model-based framework to identify at risk patients early supporting therefore personalised medicine in small cell lung cancer patients. This framework is based on a semi-mechanistic pharmacodynamic model integrating lactate dehydrogenase and neuron specific enolase from routinely collected data. The second section focuses on the use of pharmacometrics to support drug development. We pooled data from four clinical trials to develop a model to describe the pharmacokinetics of Lanreotide Autogel® in patients with gastroenteropancreatic neuroendocrine tumours. We then expanded that model to establish the link between exposure, the dynamics of the circulating biomarker Chromogranin A and the clinical outcome.

The use of circulating biomarkers within semi-mechanistic population models provides a better understanding of the relationship between drug exposure, biomarker dynamics and clinical outcome and allows to classify patients according to their disease and responses offering therefore the possibility to individualise therapy strategy and disease monitoring.