Normalized averaged range (nAR), a robust quantification method for MPIO-content

Micron-sized paramagnetic iron oxide particles (MPIO) are commonly used as contrast agents in magnetic resonance imaging (MRI) that produce negative contrast enhancement, i.e. darkening, on T2*-weighted images. However, estimation and quantification of MPIO in vivo is still challenging. This limitation mainly arises from smearing and displacement of the negative contrast of the MPIO, so-called blooming, potentially leading to false-positive detection. Further, the bias field induced by the MR coils also hinders visualization and quantification of the MPIO. To mitigate these drawbacks, a positive contrast image can be generated, for example by using a frequency offset technique, which can significantly improve the accuracy of quantification methods. In this research, we introduce the normalized average range (nAR) as a new way to quantify the relative MPIO content within a study. The method compares the average value of test ROIs to that of a control ROI in range filtered images. The nAR can be used on both positive and negative contrast images. The nAR was tested on agar phantoms containing various MPIO concentrations, and on a rostral migration model for MPIO labeled stem cells in mice. The amount of MPIO was quantified for biased and unbiased data, and both for positive and negative contrast images. In addition, the presence of MPIOs in the olfactory bulb was verified by histology. The results show the nAR can indicate the presence and relative content of MPIO for both negative and positive images. However, the nAR showed slightly higher sensitivity in optimized positive contrast images compared to negative contrast images. In all cases, the bias field played a minor role in the quantification, making debiasing less of a concern. The dependency of the nAR values on the MPIO content in the ROI was further validated histologically. Thus, the nAR provides a robust and reliable tool for quantification of MPIO in mice. (C) 2019 Elsevier Inc. All rights reserved.