Limitations of skipping echoes for exponential T2 fitting.

Abstract

Exponential fitting of multiecho spin echo sequences with skipped echoes is still commonly used for quantification of transverse relaxation (T2 ). To examine the efficacy of skipped echo methods for T2 quantification against computational modeling of the exact signal decay. Prospective comparison of methods. Eight volunteers were imaged at 4.7T, six volunteers at 1.5T, and phantoms ([MnCl2 ] = 68-270 mM). 1.5T and 4.7T; multiple-echo spin echo. Exponential fitting for T2 using all echoes, skipping the first echo or skipping all odd echoes, compared with Bloch simulations. Resulting T2 values were examined over a range of T2 (10-150 msec), refocusing flip angles (90-270°), and echo train lengths (ETL = 6-32). Shapiro-Wilk tests and Q-Q plots were used to check for normality of data. Paired sample t-tests and Wilcoxon rank tests were used to compare fitting models using α = 0.05. Multiple comparisons were accounted for with Bonferroni correction. In examined regions of interest, typical incorrect estimation of T2 ranged from 23-39% for exponential fitting of all echoes, or 15-32% for skipped echo methods. In vivo, T2 estimation error was reduced to as little as 10% with skipped echo methods using 180° refocusing and ETL = 8, although error varied due to refocusing angle, T2 , and ETL. In vivo, skipped echo T2 values were significantly different than all echo exponential fitting (P < 0.004), but also were significantly different from reference values (P < 0.002, except frontal white matter). Simulations showed skipping the first echo was the most effective form of exponential fitting, in particular for T2 <50 msec and ETL = 8, with potential to reduce T2 errors to 10%, depending on refocusing angle and T2 . Skipping echoes is insufficient for avoiding stimulated echo contamination. Resulting T2 errors depend on a complicated interplay of T2 , refocusing angle, and ETL. Modeling of the multiecho sequence is recommended. 2 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2018;47:1432-1440.