Dynamic contrast enhanced-MRI in rectal cancer: Inter- and intraobserver reproducibility and the effect of slice selection on pharmacokinetic analysis.

PURPOSE: To assess inter- and intraobserver reproducibility of DCE-MRI measurements and possible differences between two directly adjacent slices. MATERIALS AND METHODS: DCE-MRI measurements of 30 patients with histologically proven rectal carcinoma were performed on a 1.5 Tesla (T) MR system during intravenous contrast agent application before and after neoadjuvant radiochemotherapy with two directly adjacent slices used for calculation per patient. Images were analyzed semiquantitatively (parameters TTP and MITR) and quantitatively using the Brix compartment model (parameters kep and A) by two different observers and at two different time points. The concordance correlation coefficient was calculated for every parameter in intra-/interobserver comparison and slice comparison. RESULTS: Median relative differences below 10% for all parameters and high values of the concordance correlation coefficient (CCC) were found for most pharmacokinetic parameters in inter-/intraobserver comparison and slice comparison, with the exception of the parameter A before therapy in intra-/ interobserver comparison (CCC: 0.315/0.452) and kep before therapy in intraobserver comparison (CCC: 0.362). CONCLUSION: Our results indicate good inter- and intraobserver reproducibility for most pharmacokinetic parameters and for the two adjacent slices measured. However, as there were some parameters that demonstrated poor correlation, testing for reproducibility and a multiobserver approach might be considered whenever using pharmacokinetic parameters as biomarkers.