Simultaneous pixel classification and compartmental modelling of DCE-MRI in breast cancer

Purpose: Tracer kinetic modelling and pixel classification of breast DCE-MRI studies are accomplished separately, although they could greatly benefit from each other. We propose an expectation-maximisation scheme for simultaneous pixel classification and compartmental modelling of DCE-MRI studies. Methods and Materials: The key point in the proposed scheme is the estimation of the kinetic parameters (Ktrans and Kep) 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 applied a Taylor expansion of the modelling equation. The theoretical framework for the particular case of two classes was developed and the algorithm performances were evaluated by means of simulations. Preliminary analysis on one patient data were reported. Results: Simulation results furnished greater accuracy than the traditional pixel-by-pixel scheme, approaching the theoretical lower bound imposed by the Cramer-Rao theorem. Conclusion: The proposed method could improve accuracy on breast DCE-MRI data evaluation, being advantageous than other approaches and providing overall parameters without loosing information on tumour heterogeneity. Moreover the proposed algorithm could be translated in clinical routine through a dedicated software, helping clinicians in detecting and evaluating breast suspicious lesions.