An expectation-maximisation approach for simultaneous pixel classification and tracer kinetic modelling in dynamic contrast enhanced-magnetic resonance imaging

Abstract

Traditionally, tracer kinetic modelling and pixel classification of DCE-MRI studies are accomplished separately, although they could greatly benefit from each other. In this article, we propose an expectation-maximisation scheme for simultaneous pixel classification and compartmental modelling of DCE-MRI studies. The key point in the proposed scheme is the estimation of the kinetic parameters (K(trans) and K(ep)) of the two-compartmental model. Typically, they are estimated via nonlinear least-squares fitting. In our scheme, by exploiting the iterative nature of the EM algorithm, we use instead a Taylor expansion of the modelling equation. We developed the theoretical framework for the particular case of two classes and evaluated the performances of the algorithm by means of simulations. Results indicate that the accuracy of the proposed method supersedes the traditional pixel-by-pixel scheme and approaches the theoretical lower bound imposed by the Cramer-Rao theorem. Preliminary results on real data were also reported.

A geometrical perspective on the 3TP method in DCE-MRI

“…In fact, the curvature of the model manifold is made of two contributes: the intrinsic and the parameters-effect curvature [17]. In this study, the intrinsic curvature was always zero (the model manifold is a plane embedded in the data-space), while the parameters-effect curvature was negligible for the most part of interest within the parameters space [3][4][5]9]; however, for many-parameters models this curvature effect could become important and should be considered.…”

Can semi-quantitative evaluation of uncertain (type II) time-intensity curves improve diagnosis in breast DCE-MRI?

“…In the last two decades, several experimental studies have demonstrated that quantitative methods (based on tracer kinetics modelling) can be more specific in distinguishing benign from malignant breast disease, because of the capability to derive parameters strictly related to tissue microvasculature without any operator dependency [19,[20][21][22][23][24][28][29][30][31][32][33][34][35][36][37][38][39][40][41]. However, as there is not yet sufficient standardisation of quantitative methods, semi-quantitative approaches have been used because they could represent a compromise between qualitative and quantitative approaches.…”

“…Kinetic curve analysis can be performed qualitatively (visual inspection of the curve shape), semi-quantitatively (by means of empirical parameters of signal intensity changes as gradient of the upslope of enhancement curves, maximum signal intensity and wash-out gradient) or quantitatively through pharmacokinetic modelling techniques [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24].…”

Dynamic Contrast Enhanced Magnetic Resonance Imaging in Rectal Cancer

“…This approach can potentially benefit from both ROI-based and pixel-by-pixel processing (Sourbron, 2010). Also, semi-automatic approaches for model-based segmentation of DCE-MRI images are currently being developed (Buonaccorsi et al, 2007;Kelm et al, 2009;Sansone et al, 2011;Schmid et al, 2006;Xiaohua et al, 2005) …”