Method for quantitation of dynamic MRI contrast agent uptake in colorectal liver metastases

Purpose: To investigate the reproducibility of dynamic contrast-enhanced MRI (DCE-MRI) in colorectal liver metastases using a vascular normalization function (VNF) from pixels in the spleen and to compare this with a technique using an arterial input function (AIF) from pixels in the aorta. Materials and Methods: DCE-MRI with gadolinium-DTPA (Gd-DTPA) was performed in patients with colorectal liver metastases. The VNF and AIF were determined using an automated algorithm. The average Gd-DTPA uptake rate (k(ep)) was calculated for the metastases using a physiological pharmacokinetic model. The protocol was repeated on a second day to calculate the repeatability coefficient of the measurement of k(ep). Results: Using the VNF from the spleen the overall mean k(ep) of the two sessions for 11 patients was 0.033 per second and the repeatability coefficient was 0.009 per second. Using the AIF from the aorta these values were 0.031 per second and 0.028 per second, respectively. Conclusion: The mean Gd-DTPA uptake rate using a VNF taken from the spleen can be determined with adequate reproducibility in colorectal liver metastases. The use of a VNF from pixels in the spleen is better than in AIF from pixels in the aorta in terms of reproducibility, and is recommended when the DCE MRI technique is used for prediction and monitoring of therapy outcome in colorectal liver metastases.