A free-breathing non-contrast-enhanced pulmonary magnetic resonance angiography at 3 Tesla

Transcription

1 Chinese Medical Journal 2009;122(18): Original article A free-breathing non-contrast-enhanced pulmonary magnetic resonance angiography at 3 Tesla YANG Jian, WANG Wei, WANG Ya-rong, NIU Gang, JIN Chen-wang and WU Ed Xuekui Keywords: magnetic resonance angiography; pulmonary; blood vessels C Background The breathhold contrast-enhanced three-dimensional magnetic resonance angiography (MRA) using T1-weighted gradient-echo imaging sequence is the standard technique for MRA of the thorax. However, this technique is not desirable for certain patients with respiratory insufficiency, serious renal impairment, or allergy to contrast agents. The objective of this study was to optimize and evaluate a non-contrast-enhanced free-breathing pulmonary MRA protocol at 3 Tesla. Methods The time-of-flight protocol was based on a two-dimensional T1-weighted turbo field echo sequence with slice-selective inversion recovery and magnetization transfer preparation together with respiratory navigator gating, cardiac gating, and parallel imaging. Optimal values for time of inversion delay, flip angle and slice thickness were experimentally determined and used for all subjects. Results Excellent pulmonary MRA images, in which the 7th order branches of pulmonary arteries could be reliably identified, were obtained in the 12 free-breathing healthy volunteers. TI of ~300 ms provides the best suppression of background thoracic and cardiac muscles and effective inflow enhancement. With increasing flip angle, the pulmonary vessels gradually brightened and exhibited optimal contrast at The 2 mm slice thickness and 0.5 mm slice overlap is suitable for visualization of the peripheral pulmonary vessel. Conclusions The MRA protocol at 3 Tesla may have clinical significance for pulmonary vascular imaging in patients who are not available for contrast-enhanced 3D MRA and CT angiography examination or are unable to sustain a long breath-hold. ontrast-enhanced pulmonary magnetic resonance angiography (MRA) has been widely used in clinical diagnosis, because of its high spatial and temporal resolution, and avoidance of radiation exposure. However, contrast-enhanced pulmonary MRA techniques have several drawbacks. 1-4 First, the administration of a contrast agent makes the procedure invasive, posing some restrictions on repeated use in follow-up patients after treatment or surgery. Second, most of the contrast-enhanced pulmonary MRA techniques require the breath-holding to obtain an approving image quality, but some patients with respiratory insufficiency cannot hold breath for an extended period of time. Furthermore, the allergic potential of the contrast agents and the burden to kidneys have been demonstrated and cannot be negligible. 5,6 Therefore, these techniques are not desirable for certain patients with difficulty in contrast administration, holding breath or cooperating during MRI scan. Non-contrast-enhanced MRA may be an attractive alternative approach in which data are acquired during free breathing without intravenous contrast media administration. 2-4 Reduced cost is another benefit of non-contrast-enhanced imaging. In non-contrastenhanced pulmonary MRA, spoiled gradient echo, steady-state free precession, and single-shot half-fourier fast spin echo have been developed using three-dimensional (3D) acquisition. 2-4,7-9 The use of thin and sequential slices in two-dimensional (2D) time-offlight (TOF) could potentially produce the maximum contrast between moving blood and facilitate the visualization of peripheral pulmonary vascular branches. However, 2D TOF approach has not been deployed for pulmonary MRA because of its intrinsically low signal-to-noise ratio. With increasing availability of high-field clinical scanners such as 3 Tesla systems, increase DOI: /cma.j.issn Medical Imaging Center, First Hospital, School of Medicine, Xi an Jiaotong University, Xi an, Shaanxi , China (Yang J, Niu G and Jin CW) Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong SAR, China (Yang J and Wu EX) Department of Diagnostic Radiology, Tangdu Hospital, Fourth Military Medical University, Xi an, Shaanxi , China (Wang W and Wang YR) School of Medicine, Xi an Jiaotong University, Xi an, Shaanxi , China (Wang W) Dr. YANG Jian and WANG Wei contributed equally to this work. Correspondence to: Dr. YANG Jian, Medical Imaging Center, First Hospital, School of Medicine, Xi an Jiaotong University, Xi an, Shaanxi , China (Tel: Fax: yj1118@mail.xjtu.edu.cn) This work was supported in part by grants from the National Natural Science Foundation of China (No ), Ph.D. Programs Foundation of the Ministry of Education of China (No ) and Hong Kong Research Grant Council (No. GRF7794/07M).

2 2112 in signal and T1 value may offset this disadvantage. 10 This study aims to optimize and evaluate a free-breathing 2D TOF pulmonary MRA protocol at 3 Tesla using the slice-selective inversion recovery (SIR) in conjunction with magnetization transfer (MT) preparation for effective inflow enhancement and static tissue suppression, respiratory navigator gating for the acquisition of images during free breathing, and parallel imaging for the decrease of scan time. In addition, we hypothesize that such protocol will benefit from the SNR increase at 3 Tesla, and yield significantly improved pulmonary vascular details without requiring breath-holding and contrast agent. METHODS The proposed 2D TOF pulmonary MRA sequence is shown in Figure 1. The diagram represents one cardiac cycle. After triggering delay (TD) adjusted to the individual heart rate, a SIR 180 pulse is applied to invert the magnetization of the stationary tissue within the slice. The time of inversion delay (TI) is chosen to null the background muscle tissue signal approximately at the start of turbo field echo (TFE) acquisition. Prior to the SIR pulse, an off-resonance MT irradiation is performed. Because MT reduces the effective T1 of the thoracic and cardiac muscle, 11 relatively short TI is needed for inversion recovery nulling of the background muscles. This inversion delay also allows adequate TOF inflow of non-inverted fresh arterial and venous blood into the imaging slice. Then a 2D coronal T1-weighted segmented TFE acquisition is performed during the relatively stable diastole phase to minimize blood spin dephasing and motion-induced discontinuities between consecutive 2D slices. To suppress fat tissue, a frequency selective adiabatic inversion pulse (SPAIR) is applied prior to each TFE shot. In addition, the pencil-beam navigator gating is employed to synchronize the acquisition with respiratory motion by detecting the spatial position of locomotory diaphragm. Subjects Twelve healthy volunteers (8 men, 4 women) were enrolled in this study with approval from the institutional review board and given informed consent. The volunteers were years of age, and had the heart rate of beats per minute (bpm). Methods All experiments were performed on a 3 T whole body MRI scanner (Gyroscan Achieva, Philips Medical Systems, Netherlands) with the peak gradient amplitude of 80 mt/m and the gradient switch rate of 200 mt m -1 s -1. A 6-element radio frequency coil was used for sensitivity encoding parallel acquisition. The 2D coronal T1-weighted segmented TFE sequence had the shortest TR/TE = 3.4 ms/1.45 ms and partial echo to minimize the strong T * 2 effect in lung; TFE factor = 30, field Figure 1. Timing diagram of the proposed 2D TOF pulmonary MRA sequence. It is based on the segmented 2D TFE sequence with static tissue suppression and blood inflow enhancement using SIR and MT preparation, SPAIR, VCG gating, and respiratory navigator gating. of view (FOV) to 370 mm with 80% rectangular FOV; acquisition matrix of ; in-plane reconstruction resolution = 1.45 mm 1.84 mm; number of slices = 60 80; slice thickness 2 to 4 mm; slice overlap 0.5 to 1 mm; number of acquisitions (NSA) = 2 to 4; and sensitivity encoding factor of 2. Data were collected with centric reordering and 3/5th partial Fourier phase encoding. The flip angle sweep, i.e., successive excitations with increasing flip angles, was used in TFE sequence to achieve progressive spin saturation of fresh inflow blood, thus avoiding the ghost artifacts associated with the constant flip angles. Optimal values for TI and flip angle were experimentally determined and used for all subjects. The 5 out of 12 volunteers were also performed 2D sagittal acquisition for unilateral pulmonary vessels. FOV = 370 mm 200 mm, number of slices = 60; slice thickness = 2 mm; slice overlap = 0.5 mm. The other preferences were the same as above. For cardiac compensation, the sequence was cardiactriggered to avoid potential artifact and motion during systole. Cardiac triggering was accomplished using vectorcardiogram (VCG)-gating, and the trigger delay was at least 150 ms to ensure diastolic acquisition. For respiratory compensation, the leading type of navigator echo technique was implemented. As illustrated in Figure 2, the navigator column was oriented to capture the diaphragm-lung border by positioning the 2/3 of beam within the right diaphragm dome and 1/3 within the lung. Navigator length of 50 mm with the acceptance window of 5 mm was chosen to correspond to end expiration. Respiratory waveforms can be scouted by a navigator display tool. Based on the position of the diaphragm-lung border and cardiac cycle, the data were collected for the furthest decrease of the motion artifact. Image analysis Pulmonary MRA images were produced by maximum intensity projection (MIP) of the source 2D TOF images. The 3D-MIP images were generated in 15 projective orientations at each 12 interval. All images were independently manipulated and evaluated on a commercially available workstation (ViewForm, Philips Medical Systems) by an experienced chest radiologist in a blinded review. For image quality analysis, the sharpness, artifact and the highest branch order of arteries in peripheral lung that could be visualized were respectively

3 Chinese Medical Journal 2009;122(18): pulsation. Moreover, the venetian blind artifacts, which often appear in 2D TOF MRA, were designedly observed. Presence of artifact was rated on a 4-point scale (0, no artifact; 1, mild artifact, not interfering with diagnostic content; 2, moderate artifact, degrading diagnostic content; 3, severe artifact, resulting in non-diagnostic images). The branch order of pulmonary artery was identified as a conventional means, as follows: the first order for the main pulmonary artery, the second order for the left or right pulmonary artery, the third order for the lobar branches, the fourth order for the segmental branches, etc. The assessment of image quality was based on the multi-view MIP images. Figure 2. Demonstration of the position of navigator echo and imaging volume in 2D coronal acquisition. The typical imaging volume is showed in coronal scout in A (orange rectangular box) and in axial and sagittal scout in B. The navigator column is placed over the diaphragm (a small white box in A), which generates a spin echo and is used to track the diaphragm (liver-lung boundary). The reference position of the diaphragm is noted on the navigator scout in C. The distance of central two horizontal blue lines in C indicates the 5-mm gating window, and the red points in C tracks the end-expiratory position of the diaphragm. The Y axis inc shows absolute distance scale of the diaphragm and beam length of 50 mm. The data falling within the 5-mm gating window were accepted. assessed. Evaluation of sharpness was based on the clarity of vessel margins. The sharpness of the vascular branches was graded using a 4-point scale (0, not visualized; 1, poorly defined with substantial blurring; 2, well defined with mild blurring; 3, excellent definition without blurring with high confidence for diagnosis of the vascular structure). Artifact was evaluated based on appearance of artifact anywhere in the images, arising mainly from causes such as respiratory motion or Statistical methods A t test was used to evaluate the significance of differences in image quality (the scale of sharpness, artifact and the highest branch order of pulmonary arteries) between the images acquisition using slice thickness of 4 mm and 2 mm. A 2-sided value of P <0.05 was used as the criterion to indicate a statistically significant difference. RESULTS The proposed pulmonary MRA protocol was performed with free-breathing comfort in all twelve volunteers. The typical acquisition time ranged from 4 to 10 minutes. The data acceptance rate in respiratory navigator was ~50% on average. TI, flip angle, and other acquisition parameters were experimentally optimized. Figure 3 shows the pulmonary MRA images acquired with various TIs in one volunteer. It indicates that TI of ~300 ms provides the best suppression of background thoracic and cardiac muscles. Such inversion recovery nulling at relatively short TI arises from the reduction of effective T1 in muscle caused by the MT preparation prior to the SIR pulse. 10,11 At TI of 300 ms or longer, there is also Figure 3. A 26-year-old female healthy volunteer. TI of ~300 ms is found to be optimal to produce effective static tissue suppression and adequate blood inflow effect. The heart rate was 75 beats/min. Figure 4. A 56-year-old female healthy volunteer. Flip angle of provides the optimal contrast for pulmonary vessels. The heart rate was 85 beats/min.

4 2114 sufficient inflow enhancement in both proximal and peripheral vessels. Figure 4 shows the effect of various flip angles. With increasing flip angle, the pulmonary vessels gradually brightened and exhibited optimal contrast at Satisfactory pulmonary MRA quality was achieved in all subjects, permitting the routine visualization of high order pulmonary vascular branches. Up to 4th order pulmonary artery branches could be visualized in all 12 subjects using the coronal 2D TOF sequence with 4 mm slice thickness and 1 mm slice overlap. Ten subjects were also scanned with 2 mm slice thickness and 0.5 mm slice overlap. Moreover, five out of the ten subjects were performed with coronal acquisition, the rest were performed using sagittal acquisition for unilateral pulmonary vessels. The 7th order branches could be reliably identified in all those ten subjects. The results of the radiologist evaluation are given in Table. Significant differences were found in the scale of artifact (P <0.05) and the highest branch order of pulmonary arteries (P < ) between the images acquisition using slice thickness of 4 mm and 2 mm. Figures 5 and 6 illustrate the typical MIP view of the pulmonary MRA respectively acquired from coronal and sagittal 2D TOF images in two health volunteers. Table. Image evaluation ratings of pulmonary MRA in multi-view MIP images Volunteer Sharpness Artifact The highest branch order 1 3 (3) 1 (0) 5 (7) 2 2 (3) 1 (0) 5 (7) 3 3 (3) 1 (1) 5 (7) 4 2 (3) 1 (0) 5 (7) 5 2 (2) 1 (1) 4 (6) 6 3 (3) 1 (1) 5 (7) 7 2 (2) 0 (0) 5 (7) 8 2 (3) 1 (0) 4 (6) 9 3 (3) 1 (0) 5 (7) 10 3 (3) 1 (1) 5 (6) Average 2.5 (2.8) 0.83 (0.4) 4.75 (6.7) t test P >0.05 P <0.05 P < The data out of parentheses represent the images acquisition using slice thickness of 4 mm, and the data in the parentheses represent the images acquisition using slice thickness of 2 mm. The coronal 2D TOF was performed in volunteers of No.1 5, and the sagittal 2D TOF was achieved in volunteers of No DISCUSSION Nephrogenic systemic fibrosis caused by contrastenhanced 3D MRA with a high dose of intravenous administration of gadolinium-based contrast agents was recently reported in patients with advanced renal disease. 5,6 Patients with an allergy to iodine and with bronchial asthma also tend to show adverse responses to gadolinium-based contrast agents. Moreover, intravenous administration of gadolinium-based contrast is relatively contraindicated in pregnancy, and CT angiography is related to the adverse effects of the radiation exposure and the hypersusceptibility to iodine. Therefore, non-contrast-enhanced MRA may have important clinical applications in the near future and be preferable for some of these patients. In the present study, a 2D TOF pulmonary MRA with SIR plus MT preparation was proposed and demonstrated at 3 Tesla. This new protocol is robust, requires no breath-holding and contrast administration, and has the potential to depict high quality images for the assessment of anatomic information of pulmonary vessels and aorta, so much as the hepatic vessels (Figure 6). This is especially important in patients who are unable to provide long breath-holds during MRI scans or have difficulty with contrast administration. Non-contrast-enhanced pulmonary MRA has been performed with TOF, steady-state free precession, and single-shot half-fourier fast spin echo sequence by 3D data acquirement. 2-4,7-9 In 3D TOF technique, the problems frequently occur in the decrease of signal difference between blood flow and static tissue due to too many and too strong radiofrequency pulses. Despite the saturation problem can be overcome by technical variations, such as multiple overlapping thin slab acquisition (MOTSA), tilted optimized non-saturating excitation (TONE), and variable-angle uniform signal excitation (VUSE). 7 The further problems arise from the multiple-orientation and tortuous pulmonary vessels reducing signal enhancement due to the saturated partial blood, and the insensitivity to slow flow resulting in unreliable exhibition of the pulmonary peripheral vascular branches. The steady-state free precession sequence has been demonstrated the potential for excellent evaluation of blood vessels. 2-4 However, the sensitivity to magnetic field inhomogeneity and out of slice contributions are considerable limitations of the steady-state free precession sequence when applied to large FOV imaging of the lung, and can result in signal loss and image artifacts, respectively. 3 The fresh blood imaging using coronal in-plane 3D single-shot half-fourier fast spin echo sequence synchronized with electrocardiography gating at every slice encoding is another non-contrast-enhanced MRA technique for lung. However, this technique requires a weighted subtraction of two images in different phases of blood flow to producing contrast enhancement, which results in the more spatial mismatch artifacts, the repeated scans and a difficult implementation in patients with arrhythmia. 8 In this new 2D TOF MRA protocol, the vascular enhancement depends on the flow velocity, slice thickness and inversion time. 12 For small pulmonary arteries with slow blood flow, besides adequately prolonging TI, another controllable parameter is slice thickness ( Z). Nishimura et al 13 showed that the average velocity of blood flow perpendicular to slice, is in direct proportion to Z/TI. Therefore, slice thickness can be made smaller to ensure sufficient wash-in. This effect was demonstrated in this study, that the selection of

5 Chinese Medical Journal 2009;122(18): Figure 5. A 28-year-old male healthy volunteer. Typical coronal pulmonary MRA acquired with the proposed protocol is shown. The parameters were slice thickness 2 mm, slice overlap 0.5 mm, 80 slices, TI 300 ms, TD 250ms, flip angle 30, and NSA = 4. Navigator beam acceptance window was 5 mm, yielding ~80% acceptance rate. The heart rate was 80 beats/min. The total acquisition time was ~8 minutes. For image quality analysis, the scales of sharpness, artifact and the highest branch order of pulmonary arteries that could be visualized were 3, 0, 7, respectively. The 7th branching order identified by white arrow can be visualized in an MIP image. Figure 6. A 32-years-old male healthy volunteer. Typical sagittal pulmonary MRA acquired with the proposed protocol is shown. The parameters were slice thickness 2 mm, slice overlap 0.5 mm, 60 slices, TI 300 ms, TD 300 ms, flip angle 30, and NSA = 2. Navigator beam acceptance window was 5 mm, yielding ~50% acceptance rate. The heart rate was 78 beats/min. The total acquisition time was ~6 minutes. For image quality analysis, the scales of sharpness, artifact and the highest branch order of pulmonary arteries that could be visualized were 3, 1, 7, respectively. The 7th branching order identified by white arrow can be visualized. The blood vessels in the liver, inter ribs and upper arm were also respectively identified. 2 mm slice thickness for pulmonary MRA can obtain the better visualization of the pulmonary vessels than using 4 mm slice thickness. Previous studies have employed SIR pulse to enhance incoming blood in 2D TOF due to its long T1 relaxation time and inflow effect. 12,13 In the present study, the SIR pulse also serves to suppress the background static tissues, especially on muscle, with a relatively short TI in conjunction with the off-resonance MT irradiation. Because MT effect leads to a reduction of M 0 and markedly shortening T1 of muscle tissue and yet signals from blood flow only accept little weakening. 10 Therefore, the null point of muscle alters close to 300 ms at 3 Tesla after using the SIR following the MT preparative pulses. Despite long TI is desired for strong blood inflow effect, this will reduce the upper limit on the patient heart rate that can be accommodated by this 2D TOF MRA protocol for the given turbo factor in TFE sequence. At 3 Tesla, we found that TI of approximately 300 ms can yield an effective background tissue suppression and adequate blood inflow enhancement. The resulting protocol provides satisfactory pulmonary MRA contrast and can accommodate patients with heart rate up to 140 beats/min. There were a few disadvantages associated with this non-contrast-enhanced 2D TOF MRA for the thoracic blood vessels. One disadvantage was the relatively long scan time, which was mainly due to breathing navigator-gating and VCG-triggering. In this study, the 12 out of 17 examinations were successfully achieved within 10 minutes. However, the contrast-enhanced 3D MRA protocol frequently spends above 10 minutes on the full process in clinic, which are composed of some preparative periods of time for the setting up of contrast agents, intravenous cannulation, and additional imaging techniques, such as bolus timing and fluoroscopic triggering. 1 Another disadvantage of this non-contrastenhanced 2D TOF MRA might be the pulmonary arteries and veins overlapping each other in MIP images. Nonetheless, the diagnostic capability of this non-contrast-enhanced MRA for the thoracic blood vessels was sufficient, because of the excellent image quality of source 2D and multi-view MIP images at 3 Tesla. Moreover, there are a few limitations in our study. The subjects were limited to healthy volunteers and the number of subjects was relatively small. Therefore, further studies with a large patient population are warranted. In addition, we did not correlate the results to standard contrast-enhanced 3D MRA or CT angiography, etc, which is waiting for accomplishment in a further clinical study. In conclusion, our study indicates that free-breathing noncontrast 2D TOF MRA with SIR and MT prepulse can provide sufficient contrast and high image quality for a confident evaluation of pulmonary vessels. It may be of clinical importance for pulmonary vascular imaging in patients who are not available for contrast-enhanced 3D MRA and CT angiography examination: those with serious renal impairment, respiratory insufficiency, and an allergy to gadolinium and iodine. REFERENCES 1. Kauczor HU. Contrast-enhanced magnetic resonance angiography of the pulmonary vasculature. A review. Invest Radiol 1998; 33: Krishnam MS, Tomasian A, Deshpande V, Tran L, Laub G, Finn JP, et al. Noncontrast 3D steady-state free-precession magnetic resonance angiography of the whole chest using nonselective radiofrequency excitation over a large field of view: comparison with single-phase 3D contrast-enhanced magnetic resonance angiography. Invest Radiol 2008; 43: Amano Y, Takahama K, Kumita S. Non-contrast-enhanced MR angiography of the thoracic aorta using cardiac and navigator-gated magnetization-prepared three-dimensional steady-state free precession. J Magn Reson Imaging 2008; 27:

6 Hui BK, Noga ML, Gan KD, Wilman AH. Navigator-gated three-dimensional MR angiography of the pulmonary arteries using steady-state free precession. J Magn Reson Imaging 2005; 21: Broome DR, Girguis MS, Baron PW, Cottrell AC, Kjellin I, Kirk GA. Gadodiamide-associated nephrogenic systemic fibrosis: why radiologists should be concerned. AJR Am J Roentgenol 2007; 188: Sadowski EA, Bennett LK, Chan MR, Wentland AL, Garrett AL, Garrett RW, et al. Nephrogenic systemic fibrosis: risk factors and incidence estimation. Radiology 2007; 243: Friedli JL, Paschal CB, Loyd JE, Halliburton SS. Quantitative 3D VUSE pulmonary MRA. Magn Reson Imaging 1999; 17: Miyazaki M, Sugiura S, Tateishi F, Wada H, Kassai Y, Abe H. Non-contrast-enhanced MR angiography using 3D ECG-synchronized half-fourier fast spin echo. J Magn Reson Imaging 2000; 12: Wang Y, Rossman PJ, Grimm RC, Wilman AH, Riederer SJ, Ehman RL. 3D MR angiography of pulmonary arteries using real-time navigator gating and magnetization preparation. Magn Reson Med 1996; 36: Stanisz GJ, Odrobina EE, Pun J, Escaravage M, Graham SJ, Bronskill MJ, et al. T1, T2 relaxation and magnetization transfer in tissue at 3T. Magn Reson Med 2005; 54: Jakob PM, Wang T, Schultz G, Hebestreit H, Hebestreit A, Elfeber M, et al. Magnetization transfer short inversion time inversion recovery enhanced 1H MRI of the human lung. MAGMA 2002; 15: Braendli M, Bongartz G. Combining two single-shot imaging techniques with slice-selective and non-slice-selective inversion recovery pulses: new strategy for native MR angiography based on the long T1 relaxation time and inflow properties of blood. AJR Am J Roentgenol 2003; 180: Nishimura DG, Macovski A, Pauly JM, Conolly SM. MR angiography by selective inversion recovery. Magn Reson Med 1987; 4: (Received December 30, 2008) Edited by HAO Xiu-yuan