Bayesian analysis

MRI is inherently a low-sensitivity technique, and thermal noise is often a confounding influence, particularly when quantifying image data. To improve the accuracy and precision of such analysis, we have developed post-processing techniques based on Bayesian analysis. These techniques take into account the Rician distribution of noise in MR images, and have been applied to the analysis of diffusion and susceptibility MRI data. We have also developed and validated a technique for optimally sharing information between neighbouring pixels, using a Markov random field, which significantly improves both the accuracy and precision of apparent diffusion coefficient (ADC) estimates and helps define boundaries between regions with significantly differing ADC values.

Associated publications:

Walker-Samuel S, Orton M, Boult JK, Robinson SP. Improving apparent diffusion coefficient estimates and elucidating tumor heterogeneity using Bayesian adaptive smoothing. Magn Reson Med 2011;65(2):438-447. [link]

Walker-Samuel S, Orton M, McPhail LD, Boult JK, Box G, Eccles SA, Robinson SP. Bayesian estimation of changes in transverse relaxation rates. Magn Reson Med 2010;64(3):914-921. [link]

Walker-Samuel S, Orton M, McPhail LD, Robinson SP. Robust estimation of the apparent diffusion coefficient (ADC) in heterogeneous solid tumors. Magn Reson Med 2009;62(2):420-429. [link]

Orton MR, Collins DJ, Walker-Samuel S, d'Arcy JA, Hawkes DJ, Atkinson D, Leach MO. Bayesian estimation of pharmacokinetic parameters for DCE-MRI with a robust treatment of enhancement onset time. Physics in medicine and biology 2007;52(9):2393-2408. [link]

Page last modified on 19 jan 12 14:19