Carlos Garrido PhD, Presenter: Nothing to Disclose

Jorge Elias MD, PhD, Abstract Co-Author: Nothing to Disclose

Fernanda Domenici, Abstract Co-Author: Nothing to Disclose

Eduardo Rocha Arruda, Abstract Co-Author: Nothing to Disclose

Henrique S Trad MD, Abstract Co-Author: Nothing to Disclose

Ana de Lourdes Candolo Martinelli, Abstract Co-Author: Nothing to Disclose

Helio Vannucchi, Abstract Co-Author: Nothing to Disclose

Valdair Francisco Muglia MD, PhD, Abstract Co-Author: Nothing to Disclose

To evaluated three fat quantification MR methods (IO, 3P-Dixon, MRS) in different experimental conditions in order to define an adequate methodology for clinical routine.

19 NASH patients (48±10 years, 9 female) underwent MR procedure several times. Imaging was performed on a 3.0T clinical MRI scanner (Achieva, Philips) using a conventional 2D spoiled gradient echo for IO (TE=1.15 and 2.3 ms) and 3P-Dixon methods (TE=1.15, 2.3 and 3.45 ms).  MRS data was acquired using single voxel (3x3x3 cm) PRESS acquisition (TR=1500 ms, 1024 points, 2 KHz bandwidth). The PRESS acquisition was performed in a single TE (60 ms, 32 averages) and multiple TE (32, 45, 80 and 135 ms; 16 averages). The software jMRUI was used for peak quantification in the MRS data. The fat fraction was estimated from magnitude image in the IO and automatically by the equipment in 3P-Dixon. The MRS results was T1 and T2 corrected using generic values (T2fat=55ms and T2water=25 ms) for single TE data and individual values estimated from multi-TE data before fat fraction calculation.

The figure illustrates the comparison between methods using the fat fraction as main variable. The continuous lines indicate linear fits between 3P-Dixon in a breath-hold condition and the others strategies. The T2 values estimated were 69±14 and 27±3 ms for fat and water, respectively.

The IO method had the best correlation with 3P Dixon breath-hold. 3P Dixon using a respiratory trigger present more disperse results. The results from MRS data using a generic T2 value was more correlated to Dixon BH than the data using individual values, suggesting that is more important a good signal-to-noise ratio (from more averages) than a very precise T2 value for correction. The high standard deviation of T2fat value suggests that this parameter can be used as a monitor disease progression. The reduced range of T2water values doesn’t reflect any considerable iron deposition in NASH patients. In our experimental conditions IO is more convenient method for fat quantification, however an additional correction must be include in this fat fraction specially in low values (<0.1).

In order to use noninvasive MR fat quantification to evaluate and follow NASH patients it is important to understand the accuracy of different MR methods in clinical routine.

Garrido, C, Elias, J, Domenici, F, Arruda, E, Trad, H, Martinelli, A, Vannucchi, H, Muglia, V, MR Fat Quantification Methods in NASH Patients.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11015935.html