Pelvic lymph node (LN) imaging is an important tool for regional staging in different types of cancer [1]. Because these LNs are often small and assessment of their status is based on morphological characteristics, high resolution MRI could aid in improving the detection of LN metastases. Ultra high magnetic field potentially offers high NMR sensitivity, but suffers from radiofrequency (RF) transmit field inhomogeneities. This is detrimental for lymph node imaging, as it may cause lymph node metastases to be missed due to local signal intensity losses. In this abstract we propose to use the time interleaved acquisition of modes (TIAMO) technique to overcome the homogeneity issues, such that the high signal to noise ratio (SNR) of ultra high field MRI can be exploited to image LNs over the whole pelvis with a high spatial resolution. The larger chemical shift dispersion between water and lipids on 7 Tesla is used to enable short water or fat selective excitation pulses, avoiding the fat/water chemical shift displacement artifact in gradient echo (GRE) imaging. This results in T2* weighted (T2*W) and lipid (similar to T1 weighted) images to assess lymph nodes with two different types of imaging contrast [2]. In addition, a multi GRE sequence is used, to enable the calculation of images at arbitrary echo times (TEs), providing flexibility in T2* contrast [3]. All measurements were performed on a 7T whole body MR system (Siemens Magnetom, Erlangen, Germany) with an 8 channel ¹H transceiver body-array coil with meander-type microstrip elements [4]. TIAMO was used in all measurements and a phase and amplitude shimming algorithm was applied to optimize the two complementary TIAMO acquisition modes for root sum of squares homogeneity [5,6]. A 3D GRE sequence was used with a 970 μs excitation block pulse (Figure 1A) to obtain lipid images. The duration of the excitation pulse was chosen such that the zero crossing of the frequency profile of the pulse was at the water resonance frequency, so only lipid tissue was excited. The images were acquired with a TR of 5.2 ms, TE of 2.09 ms and a voxel size of 0.66x0.66x0.66 mm³ (FOV 210x210x169 mm³, matrix 320x320x256, phase oversampling 50%, GRAPPA acceleration of 2 in phase encoding and 2 in 3D direction) with 2 alternating TIAMO modes as averages and an acquisition time of 2:51 min. A 3D multi-GRE T2*W sequence was also performed with the same excitation pulse, except now centered on the water frequency (Figure 1B) while not exciting lipid resonances. The TR was 14 ms and the TEs were 2.1, 4.19, 6.21, 8.3 and 10.32 ms. The additional imaging parameters were the same as in the lipid selective sequence, resulting in an acquisition time of 8:23 min. The calculated echo time images were based on a linear least squares fit to the logarithmically transformed signal data. The imaging protocol was tested in 6 healthy male volunteers (age 25-41).Homogeneous images with high SNR were obtained in all volunteers, both in lipid-selective and T2*W water selective sequences; Figure 2 shows coronal images of 3 different volunteers. In a different volunteer the transversal, sagittal and coronal imaging of a mesorectal LN of 2 mm diameter is shown (Figure 3). The LN is clearly distinguishable with dark signal intensity in the lipid images and bright signal intensity in the T2*W images. Figure 4 shows a comparison of T2*W images acquired with TIAMO and with phase-only B1 shimming optimized for homogeneity. The image with only B1 shimming shows a dark area near the left iliac vein, conceiling the LN that is visible in the TIAMO image (red circle). In Figure 5 the measured image at a TE of 8.3 ms is compared to the computed TE image at a TE of 8.0 ms. Additionally, a computed TE=0 ms image is shown. The contrast in the computed echo time image at 8.0 ms is similar to that in the acquired image and is clearly different than the contrast in the more proton density weighted TE=0 ms image. Next to that, the SNR in the computed echo time image is increased compared to the acquired image.With the proposed measurement setup and protocol we were able to obtain high resolution homogeneous GRE images of the pelvis at 7 Tesla with reliable visualization of pelvic LNs over a large FOV. Because of the spectrally selective excitation no substantial chemical shift artifact was present in the images. Moreover, the multi-echo T2*W imaging approach enables flexible echo time imaging through calculated echo time images, while maximizing SNR efficiency.