Improved Head and neck contrast enhanced imaging using high resolution isotropic 3D T1 SPACE: A feasibility study

Introduction Post Gadolinium MR head and neck examinations remain challenging due to the need for a large anatomic coverage in minimum acquisition time. For cranial nerves and skull base investigations, 3D acquisitions are very useful not only to better visualize and analyze the cranial nerves but also to provide large head and neck coverage of wide-spread or multi-focal pathology. Until recently, 3D acquisitions implemented to image the head and neck area were based on a gradient echo imaging kernel (T1 3D Vibe FS, T1 MP-RAGE [1-3]. Unfortunately, fast 3D T1w gradient echo imaging is limited by the presence of air-tissue interfaces and inherent susceptibility artifacts. Nevertheless, to study the entire course of cranial nerves and localize focal or global contrast enhancement, a fat saturated spin-echo 3D T1 sequence seems more adequate. The purpose of our study was to investigate the utility of a 3D T1 FS Space (Sampling Perfection with Application optimized Contrasts using different flip angle Evolutions) [4] for head and neck imaging at 3T and its clinical relevance in various pathologies of this region. Materials and Method 30 patients with a variety of benign and malignant pathologic conditions affecting the head and neck area were included in a prospective imaging protocol on a 3T Trio MRI scanner (Siemens, Erlangen, Germany). The protocol consisted of standard 2D sequences (contrast enhanced axial T1 weighted TSE FS and/or T1 3D vibe FS sequence) followed by an additional 3D T1 FS SPACE sequence. For the head and neck protocol, a 12 elements head-neck coil was used, with an additional 4 elements surface coil (Machnet B.V., Eelde, Netherlands) to enforce signal covering the area of the mandible and inner ear, respectively. For head imaging a 32 elements head-coil was generally preferred provided patient head-size and/or patient compliance are compatible with the coil size. The 3D T1 FS SPACE sequence [4] was modified to allow maximized T1 weighting with fat saturation contrast (spectral saturation) at 3T. Main MR parameters for our T1 weighted sequences were: TE/TR/BW=9.4ms/993-1050/849Hz/pixel, voxel=0.9mm. Parallel acquisition (iPAT factor=2) and variable flip angles along the echo train (DP var mode) were used. T1 weighting of the designed sequence was compared to existing T1 weighted clinical and already published sequences (T1 TSE FS, T1 SE FS and T1 3D VIBE FS) by imaging a phantom consisting in 25 tubes filled with agar gel of varying Gadolinium concentrations. Two values of TR for the T1 FS 3D SPACE sequence were tested: TR1=993ms and TR2=1150ms. Based on our clinical experience, despite using a variable flip angle as a refocusing RF pulse the specific absorption rate (SAR) at 3T imposed very often an increase of TR between 993ms and 1150ms. Each 3D T1 FS SPACE acquisition consisted of a large sagittal stack with an anterior-posterior phase encoding direction and around 192 partitions. For each patient, multiplanar reconstruction reformatted coronal, oblique and/or curvilinear 3D datasets were obtained. The 3D datasets were analyzed using simple MPR, MIP or MINIP reconstruction. Thin MIP or thin MINIP were additionally obtained in any orientation depending on the region of interest. Results Fig1. show that the SNR measurements: trends of the T1 FS 3D SPACE sequence are similar to those of all reference tested sequences. As expected the two T1 FS 3D SPACE offer a much better signal to noise ratio (SNR), the highest SNR is obtained for longest TR (TR2=1150ms), then for shortest TR (TR1=993ms), then TSE, SE and VIBE.