A T1-weighted 2D spoiled gradient-recalled acquisition in the steady state (2D SPGR) image is one of the basic contrast images, and commonly acquired in abdominal MRI. In abdominal MRI, a respiratory navigation technique is widely used to reduce respiratory motion artifacts when a patient has difficulty in breath-holding for an adequate duration. The purpose of this study was to compare navigator-gating techniques for free-breathing 2D SPGR images of the liver using pencil-beam (PB) excitation¹ and self-navigation² (SN) techniques in 3 T MRI.

MR imaging: Twelve healthy volunteers underwent axial 2D SPGR MRI of the liver using PB navigator and SN techniques. Figure 1 outlines the four navigator techniques used in this study. In PB navigator, single-check (PB-SC) and double-check (PB-DC) modes were examined; the former accepts the imaging data when the following navigator value falls within the acceptance window,³ whereas the latter accepts only when both the precedent and following values are in the window. Self-navigator signals were acquired at the k-space origin and signals from the superior half of the imaging slices were combined for reducing the spin saturation effects,⁴ and self-navigator echo was acquired before or after the imaging read-out. We refer to these two self-navigation acquisition scheme as SN-Pre and SN-Post, respectively. Free-breathing (FB) and breath-held (BH) 2D SPGR scans were also performed without any respiratory navigation. All scans were performed on a 3 T clinical scanner (Discovery 750w; GE Healthcare, Waukesha, WI) with floating anterior and fixed posterior coil arrays. Sequence parameters included: parallel imaging using ASSET with an acceleration factor of 2.5, TR = 245 ms, TEs (SN-Pre) = 2.3/5.8 ms, TEs (others) = 1.1/2.3 ms, rBW = ±83.3 (SN-Pre)/±166.7 (others) kHz, FA = 60°, FOV = 36 cm, slice thickness/gap = 6.0/1.5 mm, matrix = 288 ×192, number of slices = 22, NEX = 1.

Waveform/Image analysis: Visual waveform quality was graded using a three-point scoring system: 1 (poor) = respiratory movement not apparent, 2 (fair) = apparent respiratory movement with substantial fluctuation, 3 (good) = smooth and well representative of respiratory movement. Image analysis was performed quantitatively and qualitatively. The ghost signals in the SPGR images were measured using a rectangular ROI (20 mm × 50 mm) placed anteriorly outside the body. The final ghost levels were calculated with the averaged ghost signals from the upper, middle, lower slices divided by the liver signals in the middle slices. Visual motion artifacts were graded using a four-point scoring system: 1 (poor) = unacceptable for clinical use, 2 (fair) = distorted diagnostic capability, 3 (good) = evident artifacts without distorting diagnostic capability, 4 (excellent) = negligible artifacts. Statistical analysis was conducted by the Friedman test, followed by the Wilcoxon signed-rank test with corrections for multiple comparisons.

Table 1 shows the scan time and the waveform analysis results. SN scans took longer than PB; especially the scan time of SN-Post was almost doubled compared to those of PB methods. The PB techniques generated better waveform quality than the self-navigator, and SN-Pre was better than SN-Post. Ghost levels of all the navigator scores were higher than BH and lower than FB (Table 2). Among navigator techniques, PB-DC and SN-Post scores were lower than PB-SC and SN-Pre in several conditions. Visual analysis showed that all the navigator scores were significantly lower than BH and higher than FB (Table 3). Among navigator techniques, PB-DC and SN-Post scores were significantly better than PB-SC and SN-Pre in several comparisons. Representative images were shown in Fig. 2.

PB generated better waveform quality than SN, showing that the PB navigator reflected pure respiratory motion whereas the SN signals are influenced by cardiac motion and spin saturation effects.² In comparison between the two self-navigator techniques, SN-Post was worse than the SN-Pre regarding waveform quality, though ghost level and image quality scores were better. One possible reason may be larger cardiac motion effects are involved in SN-Post signals, causing fluctuations in waveforms and consequently allowing both monitoring of respiratory and cardiac motions. Decomposition of respiratory and cardiac motions can improve scan efficiency and/or image quality. Among navigator techniques, PB-DC and SN-Post had better image quality than the others, and PB-DC should be the best considering scan time among the techniques tested in this study. Further investigation should be conducted in the future to reduce the difference between navigator and BH techniques.