Robust pharmacokinetic analysis for population studies in Breast Cancer detection using the Mohan-Shinagawa model

The pharmacokinetic (PK) analysis of breast MRI data using prior methods like the Tofts model-based approaches involved the estimation of the amount of contrast agent (CA) fed to the tissue, called the Arterial Input Function (AIF). The Mohan-Shinagawa model (henceforth referred to as the M-S model), is a novel expanded model (derived from the Tofts model) proposed in (1). It analytically eliminated the AIF from the analysis but required the robust selection of suitable reference regions across images. In this paper, the authors propose a novel framework for Tofts model estimation, using the M-S model as an intermediate stage. The advantages are that the AIF estimation is eliminated, and the final estimated PK parameters are independent of the reference region selected. This highly simplifies the overall analysis and improves the robustness in population studies by reducing the bias introduced by the reference region selection while keeping the advantages of the M-S framework including a reduction in scattered false positives. Also, as compared to the M-S model, the physical interpretation of the Tofts model parameters is well documented (2). This framework could potentially also be used for analysing DCE-MRI of other anatomical structures.