Improved MRIR₂* relaxometry of iron‐loaded liver with noise correction

Abstract

Accurate and reproducible MRI R2 * relaxometry for tissue iron quantification is important in managing transfusion-dependent patients. MRI data are often acquired using array coils and reconstructed by the root-sum-square algorithm, and as such, measured signals follow the noncentral chi distribution. In this study, two noise-corrected models were proposed for the liver R2 * quantification: fitting the signal to the first moment and fitting the squared signal to the second moment in the presence of the noncentral chi noise. These two models were compared with the widely implemented offset and truncation models on both simulation and in vivo data. The results demonstrated that the "slow decay component" of the liver R2 * was mainly caused by the noise. The offset model considerably overestimated R2 * values by incorrectly adding a constant to account for the slow decay component. The truncation model generally produced accurate R2 * measurements by only fitting the initial data well above the noise level to remove the major source of errors, but underestimated very high R2 * values due to the sequence limit of obtaining very short echo time images. Both the first and second-moment noise-corrected models constantly produced accurate and precise R2 * measurements by correctly addressing the noise problem.

Biopsy-based calibration of T2* magnetic resonance for estimation of liver iron concentration and comparison with R2 Ferriscan

“…Furthermore, the offset model used by Wood for R2* measurement [17] is different from our truncation model used in this study, whereas truncation model used by Hankins is really an offset model with direct noise floor subtraction based on Anderson’s early model. We have demonstrated [39] that, compared with the truncation model, the offset model tends to overestimate R2* due to its technique for noise compensation. For this particular reason, if R2* is measured using our truncation model, lower LIC values are expected than if calculated by using the calibration equations from Wood or Hankins.…”

An online open-source tool for automated quantification of liver and myocardial iron concentrations by T2* magnetic resonance imaging

“…Lack of application of this principle accounts for most of the errors in the interpretation of the examination in less experienced centres. 3 Therefore, the development of an automated algorithm by He et al 4 reducing the need for user manipulation of the data with correction for imperfections in the original T 2 * signal was greatly sought. This algorithm incorporates the following rules in order to correct for the apparent offset in the decay curve: if the original correlation coefficient is .0.995 for all data sets, it accepts the original T 2 *; if not, it automatically eliminates the last data points until the correlation coefficient exceeds that threshold; the elimination proceeds until the T 2 * values drop to ,2.5%.…”

“…1 Part of the limitation in widespread use of the technique occurs owing to restricted access to quantification software and difficulty in obtaining accurate numbers, especially in situations of severe iron overload where truncation or an offset model has to be applied. 2,3 In order to facilitate the calculation of T 2 * values in these situations, He et al 4 published an accurate algorithm for automated truncation and correction of the analysis of T 2 * decay curves, simplifying the method while maintaining excellent accuracy with a coefficient of variation of only 1.6%. Despite the significant results, the technique was implemented only in MATLAB® (MathWorks®, Natick, MA), limiting its access in most clinical centres worldwide.…”

Validation of a new t2* algorithm and its uncertainty value for cardiac and liver iron load determination from MRI magnitude images

“…[2]), $\sigma$ is the noise standard deviation and $L$ the number of receiver coils in use. As previously proposed 16, the right term of Eq. [3] is estimated as a free parameter, resulting in a three‐parameter model.…”

“…To refrain from extensive truncation which may lead to loss of precision, ADAPTS requires a minimum number of available TE images, a second constant $P2$, to proceed with the truncation method. If the number of valid TEs is below $P2$, ADAPTS assumes the number of remaining data points is insufficient for robust T2* estimation and switches to a noise‐correction approach, similar to the M2NCM method (Second‐Moment Noise‐Corrected Model), proposed by Feng et al 16. This method fits the observed signal in all available TE images to the expected value of the noncentral chi distribution in the presence of an underlying monoexponential decay: $E\left(M^2\right)=S^2+2L{\sigma }^2.$ …”