This work describes the development of a quality control protocol, which can be implemented to assess the accuracy, precision and reproducibility of the apparent diffusion coefficient (ADC) measurement on a clinical magnetic resonance imaging (MRI) system. The precision and accuracy of the ADC measurement are analysed with regard to MRI system noise, signal reproducibility and differences between nominal and effective b values. Two aqueous test-solutions of CuSO4 and sucrose are prepared for the quality control protocol. ADC measurement with the CuSO4 solution is more sensitive to differences between nominal and effective b values, on account of the solution's high ADC. ADC measurement with the sucrose solution is more sensitive to signal reproducibility due to the solution's low baseline signal intensity. The ADC of the test-solutions is measured on an MRI system at our centre with a sequence used for clinical studies using diffusion imaging. Two parameters, Q and R, are defined for the analysis of the quality control ADC values. The Q parameter is the ratio of the standard deviation of the quality control mean ADC values over time to the optimal standard deviation, as derived from the effect of thermal noise on the ADC measurement uncertainty. Analysis with the Q parameter indicates that signal reproducibility errors contribute to ADC variations on our MRI system when imaging with high b values (b > 500 mm s(-2)), whereas differences between nominal and effective b values have a greater impact on the ADC measurement when imaging with low b values (b < 500 mm s(-2)). The R parameter is defined as the ratio of the directional variation of the ADC quality control values to the uncertainty of the ADC measurement. Analysis with the R parameter shows that the effect of directional variation of the ADC measurement on our MRI system is more pronounced when imaging with low b values. The quality control protocol identified a systematic error, which introduced a small system-induced anisotropy in the ADC measurement. This error is currently taken into account in the analysis of clinical studies employing the diffusion imaging sequence used in this quality control protocol.

Wavelet-based de-noising has been shown to improve image signal-to-noise ratio in magnetic resonance imaging (MRI) while maintaining spatial resolution. Wavelet-based de-noising techniques typically implemented in MRI require that noise displays uniform spatial distribution. However, images acquired with parallel MRI have spatially varying noise levels. In this work, a new algorithm for filtering images with parallel MRI is presented. The proposed algorithm extracts the edges from the original image and then generates a noise map from the wavelet coefficients at finer scales. The noise map is zeroed at locations where edges have been detected and directional analysis is also used to calculate noise in regions of low-contrast edges that may not have been detected. The new methodology was applied on phantom and brain images and compared with other applicable de-noising techniques. The performance of the proposed algorithm was shown to be comparable with other techniques in central areas of the images, where noise levels are high. In addition, finer details and edges were maintained in peripheral areas, where noise levels are low. The proposed methodology is fully automated and can be applied on final reconstructed images without requiring sensitivity profiles or noise matrices of the receiver coils, therefore making it suitable for implementation in a clinical MRI setting.

Purpose:To study the frequency response characteristic of the MRI signal receiver system as a contributing factor to the formation of Nyquist ghosting in echo-planar imaging (EPI). Materials and Methods:Experimental work was undertaken on a 1.5 T system. A cylindrical test object filled with water was imaged axially with EPI in the center of the quadrature, transmit-receive head coil. In the first set of experiments, the water conductivity was increased progressively with the addition of salt between EPI acquisitions. In the second set of experiments, the conductivity of the water in the test object was kept constant and EPI images were acquired at several different bandwidths. A computer simulation was also implemented to demonstrate the impact of changes in the frequency response characteristic of the signal receiver system on EPI Nyquist ghosting.Results: Experimental and simulation results showed that Nyquist ghosting increased with the variation of the frequency response characteristic within the effective frequency range determined by the image bandwidth. One can increase the variation in the frequency response characteristic by increasing its steepness over the image's bandwidth window when coil loading is decreased, or by increasing the effective frequency range when image bandwidth is increased. Conclusions:The results of this research may help reduce Nyquist ghosting in EPI studies when the imaging coil is not sufficiently loaded, such as in pediatric and phantom studies. ECHO-PLANAR IMAGING (EPI) is susceptible toNyquist ghosting because of the bipolar frequency-encode gradients implemented in the EPI sequence. The MRI signal is reversed in k-space when it is recorded under a negative frequency-encode gradient. As a consequence, any modulation or discrepancy in the MRI signal becomes antisymmetric with respect to the center of the acquisition window, between successive kspace lines. The resulting discontinuity between even and odd k-space lines gives rise to Nyquist ghosting. Several sources of MRI signal interference have been identified as contributing factors in EPI Nyquist ghosting, such as eddy currents, concomitant fields, oblique imaging, image filters, imperfect pulse sequence timing, B 0 field inhomogeneity, susceptibility differences, and chemical shift (1-4). In this work, the influence of the frequency response characteristic of the MRI signal receiver system on the formation of Nyquist ghosting in EPI is demonstrated using theoretical analysis and experimental data. We showed that Nyquist ghosting depends on the frequency response characteristic by changing the receiver coil loading and the image bandwidth. To confirm these results, a computer simulation was implemented to demonstrate the effect of changes in the frequency response characteristic of the MRI signal receiver system.Previous studies have identified the role of the receiver system in image quality with regard to signal intensity and uniformity (5,6). In this work we analyzed the effect of the receiver system on an additional...

Limiting spatial resolution is a key metric of the quality of magnetic resonance (MR) images, which can provide an indication of the smallest region that can effectively be imaged. In this paper a novel methodology for measuring the limiting spatial resolution of MR images is mathematically analyzed and successfully implemented on phantom images. The methodology presented in this paper is based on a direct fit of a mathematical expression of the edge spread function (ESF) profile to the ESF data acquired at the interface between different materials. The mathematical expression of ESF was determined by approximating the line spread function (LSF) of the system with a sinc function. The proposed methodology can be applied using signal data from magnitude MRI spin echo images and is not sensitive to noise amplification introduced by differentiating the ESF to produce the LSF, as performed in previous studies. In addition, the proposed methodology provides a quantitative, representative measurement of the limiting spatial resolution of MR images.

No abstract