Abstract
2103
Objectives This study concerns 3D ultra short-echo time (UTE) sequence testing and comparison to LAVA sequence for MR based attenuation correction for PET lung imaging [1]. 3D UTE is advantageous over LAVA, being able to capture bone structures by its very short TE (~0.03 ms). Another advantage that is examined here, could be the detection and visualisation of lung parenchyma density, as studies have already shown [2]. Both these advantages could lead to more precise attribution of the linear attenuation coefficients of bone and lung structure.
Methods The study was conducted on a GE Discovery MR750w 3T scanner. Gradient echo sequences used are dual-echo Liver Acquisition with Volume Acquisition (LAVA) and 3D UTE Cones. A healthy volunteer was scanned with both sequences. For LAVA the parameters are: 15 sec breath-hold acquisition, TR/TE/FA 4.4ms/2.6ms/12 degrees, FOV 30*30*30 cm, res/slice 2mm/4mm. For Cones: 1 min prospective gated acquisition, TR/TE/FA 4.2ms/0.03ms/7 degrees, FOV 30*30*26 cm, res/slice 2mm/4mm. The images are set to same contrast window width and a gaussian blur filter is applied.
Results Figure 1 shows slices of both LAVA and Cones acquisitions. ROIs were drawn towards the anterior and posterior parts of the lung assessing mean signals. The mean signal from Cones are 50% higher that that of LAVA (6.3/9.3 a.u) for the anterior ROI, and 271% higher (6.9/25.9 a.u) for the posterior ROI. These quantified results suggest that this higher signal is due to lung density detection and could be used for attenuation correction of the lung.
Conclusions Preliminary results show that 3D UTE captures more lung signal than LAVA and could be used for lung parenchyma density visualisation. Further clinical studies are needed in order to ascertain this perspective that might lead to use of continuous lung structures attenuation coefficients instead of only one that it is the standard method up to date. Along with better bone signal, Cones could outperform LAVA for PET AC.