Comparative study into the robustness of compartmental modelling and model-free analysis in DCE-MRI studies

Purpose: To evaluate and compare the reproducibility ofthe preferred phenomenological parameter IAUC60 (initialarea under the time-concentration curve [IAUC] definedover the first 60 seconds postenhancement) with the preferredmodeling parameter (Ktrans), as derived using twosimple models, in abdominal and cerebral data collected intypical Phase I clinical trial conditions.Materials and Methods: Dynamic contrast enhanced MRI(DCE-MRI) time series were acquired at two imaging centersfrom a group of patients with abdominal tumors and agroup with gliomas. At both imaging centers, precontrast T1was calculated using a variable flip angle three-dimensionalspoiled gradient echo acquisition that was used toquantify tissue contrast agent concentration, allowing voxelwisedefinition of summary DCE-MRI parameters.Results: A comparison of reproducibility showed that therewas no statistically significant difference in reproducibilitybetween IAUC60 and Ktrans, although there was a trendtowards better reproducibility for Ktrans (P 0.0782). The95% confidence intervals (CIs) for individual changesshowed that for IAUC60 and Ktrans, changes in excess of47% and 31%, respectively, are outside the range of normalvariability.Conclusion: Although modeling is more complex and morecomputationally intensive than an IAUC parameterization,our data suggest this approach to be preferable to a modelfreeapproach since it provides greater physiological insightwithout a reduction in statistical power for Phase I/II clinicaldrug trials.Key Words: contrast-enhanced MRI; reproducibility; pharmacokineticmodeling; microvascular endothelial permeability;quantitative analysis; semiquantitative analysis