Effects of T2* On the Accuracy of Two-Point Dixon Water and Fat Imaging
O Rambow1, J Ma2*, (1) Rice University, Houston, TX, 2) UT MD Anderson Cancer Center, Houston, TXWE-C-116-1 Wednesday 10:30AM - 12:30PM Room: 116
Purpose: To quantify the error caused by T2* decay in two-point Dixon imaging with flexible echo times. Specifically, we seek the error in the water values as well as the relative background phasor, which is vital for identifying pixels as water-dominant or fat-dominant, as a function of the echo times. Understanding how these errors depend on the echo times will enable us to choose echo times that will result in more accurate fat-water separation.
Methods: In simulation, a model that includes T2* decay was used to generate signals. Water and fat values and the background phasor were then reconstructed from these signals using the traditional model that neglects T2* decay, and the error in these values was calculated. These results were validated experimentally by scanning an oil-water phantom using a dual-echo GRE sequence. The first-echo signal values were taken as the "true" fat and water values, and the error was determined by comparing them to the values obtained using the traditional model.
Results: For long T2* values, the error is very small except at TE combinations for which noise performance is also bad. Error is much larger for shorter T2* values and depends strongly on the echo times. For example, error at TE1/TE2 = 1.8/3.0 ms is approximately ten times that at TE1/TE2 = 1.2/2.4 ms. Plots are presented to show dependence on both T2* and echo times.
Conclusion: Shorter T2* values result in greater error, as expected. More interesting is the dependence on echo times, which is similar to error caused by signal noise. For two-point Dixon imaging, the echo times should be chosen carefully to avoid large errors that can be caused by T2* decay.