A modification of the dual-echo steady state (DESS) pulse sequence was recently proposed that allows the extraction of both T2 and apparent diffusion coefficient (ADC) maps from a pair of DESS acquisitions (termed quantitative DESS, or qDESS). The sequence modifies the gradients played between the first and second echoes in a DESS acquisition to vary the diffusion-weighting present in the images. It then uses a signal model to fit the four acquired echoes and extract T2 and ADC [1].

A careful validation of the accuracy of the qDESS T2 and ADC estimates has not yet been performed. Furthermore, the T2 and ADC estimates resulting from the qDESS sequence are different depending on what signal model is assumed for the reconstruction (the Buxton-Wu model [2] or the Extended Phase Graph (EPG) model [3]).

In this study, we acquired 3D qDESS data in phantoms (3T) and in the human knee in vivo (3T and 7T). The data was then reconstructed, and T2 and ADC maps were produced using the EPG method. At 3T, the T2 and ADC values estimated from qDESS were then compared to T2 values extracted using a 2D multi-echo spin echo sequence and ADC maps extracted using a traditional single-direction 2D-EPI diffusion scan.

3T Phantom: A diffusion phantom was constructed using water, 1.5% w/v agarose gel, and 2mM CuSO4 5H2O. The phantom was scanned using a qDESS on a Siemens Trio 3T whole-body MRI scanner. The parameters used in the qDESS sequence are as follows: TR = 26 ms, TE1 = 9 ms, TE2 = 17 ms, flip angle 1 = 35°, flip angle 2 = 18°, FoV = 256x256 cm, matrix size = 256 x 256, slice thickness = 2 mm, number of slices = 60, voxel size = 1 x 1 x 2 mm, gradient spoiler in slice direction (10,000 mT*s/m).

Maps of ADC and T2 values were calculated using the acquired qDESS echoes using the EPG model.

The qDESS ADC and T2 values were compared to a standard T2 mapping ME-SE sequence (TE = 12, 30, 50, 75, 100, 150, 200, 250, 300, 400, 500ms), and to a single-direction 2D-EPI sequence tested on the same diffusion phantom.

3T In Vivo, Human Knee: Images were acquired using the same qDESS sequence described above of the knee of a normal volunteer. ME-SE T2 and single-direction 2D-EPI ADC maps were also acquired for comparison.

7T In Vivo, Human Knee: A qDESS protocol was further implemented on a Siemens 7T whole-body scanner, and the knee of a normal volunteer was scanned. The 7T images followed a similar protocol as the 3T study with slightly different parameters: TR = 15.5 ms, TE1 = 3.5 ms, TE2 = 12.0 ms, flip angle 1 = 24°, flip angle 2 = 14°, FoV = 256 x 256 cm, matrix size = 256 x 256, slice thickness = 2 mm, number of slices = 60, voxel size = 0.6 x 0.6 x 2.0 mm, gradient spoiler in slice direction (10,000 and 80,000 mT*s/m).

The comparison between commercially acquired and qDESS-acquired T2 and ADC maps of the phantom is shown in Figure 1. The qDESS-acquired images of the agarose phantom provided average ADC and T2 values consistent with results from 2D-EPI and ME-SE sequences. qDESS ADC values of the phantom averaged 1.69 *10^-9 m²/s and 2D-EPI ADC values averaged 1.66 *10^-9 m²/s. qDESS T2 values averaged 72.4 ms and ME-SE T2 values averaged 74.0 ms.

As can be seen, the phantom qDESS scans are in excellent agreement with the traditional ME-SE T2 maps and the 2D-EPI ADC maps. Note, however, that the qDESS scans are at a significantly smaller voxel size, and are true 3D scans with full volume coverage.

Figure 2 shows the results of the 7T and 3T qDESS scans in vivo on a human knee. The high-resolution 3D T2 maps achievable at 7 Tesla are particularly interesting, given the difficulty of acquiring accurate T2 maps at 7 Tesla due to transmit B1 inhomogeneity.

The qDESS sequence produces accurate, high-resolution ADC and T2 maps. This is especially notable at 7T, where transmit B1 inhomogeneity makes it very difficult to produce reasonable T2 maps using typical multi-echo spin echo techniques. The qDESS technique shows great promise in providing both high-resolution morphological data and maps of T2 and ADC at 3 and 7 Tesla.