Multi-modality Imaging in the Prediction of Response to Systemic Treatment in Colorectal Cancer

Colorectal cancer is a frequently occurring cancer and about half of the patients develop distant metastases to the liver. Only a subset of patients respond to systemic treatment, which is potentially toxic and expensive. Therefore, predictive markers are needed to determine treatment efficacy at an early stage. Preferably, they should provide insight into the biology of colorectal cancer liver metastasis. In the past it has been shown that several biomarkers as assessed with PET, MRI and MRS can serve as predictive markers. Response to systemic therapy depends on several factors: delivery of the drug by the tumor vascular system; cellular uptake, retention and metabolism; intrinsic sensitivity to a specific drug. The relative contribution of these factors to response will be different for different drugs. Since systemic treatment of colorectal cancer involves a combination of drugs, predictive markers should be sensitive to a range of these factors. In this project we propose the integrated analysis of noninvasive functional and molecular in vivo imaging methods in order to predict the response to treatment in patients with liver metastases of colorectal cancer. Tumor vascularity can be assessed by dynamic contrast enhanced magnetic resonance imaging (DCE-MRI). The fluoropyrimidines FU and capecitabine - which are part of standard treatment regimens - contain a fluorine atom which can be measured by fluorine-19 MR spectroscopy (19F MRS). The intracellular uptake and metabolism of drugs is an energy-consuming process. 18F-Fluoro-2-deoxyglucose positron emission tomography (FDG-PET) provides information on glucose uptake and hexokinase activity. Tumor characterization in terms of grade and aggressiveness may give a relevant approximation to the intrinsic sensitivity of a tumor to a drug. With 1H MRS the total amount of choline (tCho), a precursor of cell membranes, can be measured. Tumor cellularity and extracellular matrix composition can be assessed with diffusion weighted MRI (DWI). The aim of this study is to obtain data on the biology data of colorectal cancer liver metastasis, namely tumor vascularization (DCE-MRI), tumor cellularity (DWI), tumor (choline) metabolism (1H MRS) and tumor glucose metabolism (FDG-PET). These data will be correlated with the clinical outcome of patients and with drug uptake and metabolism (19F MRS). We will study the relative contribution of each imaging method to predict the outcome of patients with colorectal cancer at an early stage. Research questions: 1. Do pre-treatment characteristics of liver metastases of colorectal cancer as assessed by dynamic contrast enhanced MRI (DCE-MRI), diffusion weighted MRI (DWI), 1H MR spectroscopy (MRS) and FDG-PET predict treatment outcome? 2. Do early changes (one week after start of treatment) in DCE-MRI, DWI, 1H MRS and FDG-PET characteristics of liver metastases of colorectal cancer predict treatment outcome? 3. Do 19F MRS parameters of fluoropyrimidine metabolism in liver metastases of colorectal cancer predict treatment outcome at an early stage (one week after start of treatment)? 4. What is the relative contribution of each above mentioned imaging method to predict treatment outcome of colorectal liver metastases? Design: 60 patients with liver metastases of colorectal cancer treated with fluoropyrimidine-based therapy will participate to the study. Baseline 1H MRS of the liver will be performed in a session at 3 Tesla followed by DCE-MRI and DWI at 1.5 Tesla. FDG-PET will be added to the standard baseline CT scan, using our clinical PET-CT scanner. DCE-MRI, DWI, 1H and 19F MRS as well as FDG-PET of the liver will be repeated one week after start of treatment. Clinical response will be evaluated after three treatment cycles by FDG-PET-CT. We will assess the relative contribution of each imaging method as well as the integrated use of these methods for the identification of predictive biomarkers for response to treatment. Relevance of this study: Since the response of a tumor to systemic drugs may be highly variable between patients, a method that predicts the sensitivity of a tumor to treatment at an early stage would enable individualization of therapy and consequently would protect patients against the toxic effects of ineffective treatment. Preferably, those predictive markers should also give us further insight into the biology of colorectal cancer. For this reason we propose to study a combination of in vivo noninvasive imaging methods which allow the monitoring of relevant biomarkers