Quantitative analysis of dynamic contrast-enhanced MRI datasets of the metacarpophalangeal joints.

Rationale And Objectives: In this article, we propose an alternative approach to voxel-by-voxel analysis, which overcomes problems associated with heuristic methods currently used for dynamic contrast-enhanced MRI (DCE-MRI) data assessment. We aim to allow fully automated extraction of various heuristic parameters via robust preprocessing methods and a new technique for classification of temporal patterns of contrast agent uptake, making full use of all available dynamic frames of the datasets. We also demonstrate that application of efficient preprocessing methods permits more accurate analysis of the dynamic data.

Material And Methods: Ten DCE-MRI datasets enhanced by gadolinium diethylene triamine pentacetic acid were acquired from patients with rheumatoid arthritis using a 1.5-T MRI scanner. Fully automated voxel-by-voxel analysis of DCE-MRI signal intensity curves from 60 temporal slices was performed using a new method. Qualitative evaluation of the degree of inflammation was done via constructing parametric maps and quantitative by computing various heuristics such as maximum rate of enhancement, initial rate of enhancement, and time of onset of enhancement.

Results: Quantitative and qualitative evaluation obtained for 10 DCE-MRI datasets is presented. We demonstrate that preprocessing techniques compensate for patient movement, contribute to data fidelity and therefore permit more robust estimation of various heuristics such as maximum rate of enhancement, initial rate of enhancement, and time of onset of enhancement. Automatically generated parametric maps of these heuristics show favorable characteristics, permitting easier differentiation of structures of interest. These results are free from the subjective input and therefore easily reproducible. Furthermore, the proposed classification scheme provides information on the pattern of contrast uptake previously unavailable.

Conclusion: Our preliminary results demonstrate the potential of the proposed method for providing objective quantitative and qualitative assessment of DCE-MRI in the metacarpophalangeal joints. Further evaluation within a clinical setting is needed to examine the method's diagnostic utility.