Martin R. Prince, MD, PhD Gadolinium-enhanced MR Aortography’ PURPOSE: To evaluate the potential VJAGNETIC resonance (MR) angiog- ies directly with excellent contrast for preferential arterial enhancement raphy is increasingly per- and high resolution in a three-dimen- at magnetic resonance (MR) aortogra- formed as a noninvasive method of sional Fourier-transform acquisition phy with an intravenous infusion of evaluating patients with suspected optimized for the shortened Ti of the paramagnetic contrast material. vascular disease (1-3). MR angiogra- blood. The use of gadolinium chebates MATERIALS AND METHODS: phy is particularly well suited to the for enhancing MR angiognaphy has Gadolinium chebates were adminis- evaluation of low-resistance vessels been described in the intracranial cm- tered intravenously (0.2 mmol/kg) such as the carotid and intnacerebnal culation (4-9), aorta (10-13), and pom- during three-dimensional MR imag- arteries. In healthy subjects, vessels of tal vein (14) and in popliteal-tibiab ing (1.5 T) in 125 patients (77 male normal caliber have strong, continu- vessels (15). The previously reported ous laminar flow, which is readily de- problems of background tissue and and 48 female patients, aged 4-86 years [mean, 66 years]) with sus- tected with standard time-of-flight venous enhancement were avoided in pected aorta or aortic branch vessel techniques. But in patients with vas- this study by imaging during infusion disease. cuban disease that disturbs the normal of contrast material. In addition to laminar flow, time-of-flight images are demonstrating the technique, this an- RESULTS: Infusion for the duration degraded. There is loss of signal when tide evaluates the effects of infusion of the MR acquisition resulted in sig- inflow is reduced in ectatic vessels on time, dose, and differences between nificant preferential arterial enhance- in patients with low cardiac output. three currently available gadolinium ment without the confounding effects Tortuous vessels may have segments compounds. of excessive venous or background- that are not oriented perpendicular to tissue enhancement (P < 10). An- the plane of imaging, resulting in loss MATERIALS AND METHODS giographic or surgical correlation in of signal from in-plane saturation. 48 patients revealed an 88% sensitiv- Loss of signal also results from turbu- Theory ity and a 97% specificity for detection bent dephasing in stenoses and aneu- Gadolinium chelates are paramagnetic of stenoses or occlusions and a 100% rysms. sensitivity and a 100% specificity for agents that shorten the spin-lattice relax- These problems are exacerbated in ation time, Ti, of blood according to the detection of aortic or iliac artery an- the abdomen and pelvis, where it is equation eurysms. necessary to use a larger body coil, 1 1 CONCLUSION: Preferential arterial which results in a reduced signal-to- (1) enhancement is possible at MR aor- noise ratio. Peripheral arteries also 1,200 R < [GdJ, tography with an intravenous infu- have higher resistance with pubsatile, sion of paramagnetic contrast material. sometimes retrograde flow; this me- where 1/1,200 is the Ti of blood without sults in decreased blood signal and gadolinium, R is the relaxivity, and [Gd] is Index terms: Aorta, MR, 981.12943 #{149}Magnetic pulsatility artifacts. In addition, the the blood concentration of a gadolinium resonance (MR), contrast enhancement, considerable length of many periph- chelate. 981.12943, 9*129432 #{149}Magnetic resonance crab vessels necessitates long imaging During dynamic imaging at times (MR), vascular studies, 981.12943, 9*129432 shorter than the recirculation time, the times with time-of-flight techniques blood concentration of a gadolinium che- Radiology 1994; 191:155-164 because imaging must be performed late is determined by the intravenous infu- in a plane perpendicular to the dinec- sion rate and the cardiac output as tion of flow. I From the Department of Radiology, Massa- This article describes a method of Gd infusion rate chusetts General Hospital, Harvard Medical [Gdlarteriai = cardiac output (2) School, Boston, Mass, and the Department of MR aontography that does not de- Radiology, University of Michigan, University pend on blood inflow or blood mo- The three gadolinium chelates currently Hospitals B1D530, Ann Arbor, MI 48109-0030. tion and thereby circumvents many of available for clinical use were investigated Received August 31, 1993; revision requested the problems of conventional time-of- in this study: gadopentetate dimeglumine October 12; revision received November 4; ac- flight MR angiography. By adminis- cepted December 10. Supported in part by the (Magnevist; Berlex Laboratories, Wayne, RSNA Research and Education Fund as an tering a paramagnetic contrast agent NJ), gadoteridol (Prohance; Squibb Diag- RSNA Research Resident and by grant no. intravenously, the Ti of blood is HL46384 from the National Institutes of Health. made short compared with that of fat Address reprint requests to the author. 2 9* indicates generalized vein and artery in- (Ti = 270 msec), muscle (Ti = 600 Abbreviations: MIP = maximum intensity msec), and other background tissues. volvement. projection, Si = signal intensity, TE = echo . RSNA, 1994 It is then possible to image the arten- time, TR = repetition time. 155 1200 0.5 Single Double 1000 Dose Dose 0.4 E 800 U) 0.3 C = 100 600 a 0.2 = 150 C 400 0) Ti =270(fat) = 270 (fat) .t0 200 ____________________________________________________ 0 - L1_J200 (blood) 0 0 0,1 0.2 0.3 0.4 0.5 0 20 40 60 80 100 Gadopentetate Dimeglumine Dose (mmol/kg) Flip Angle (degrees) 1. 2. Figures 1, 2. (1) Graph depicts predicted blood Ti versus total dose of gadopentetate dimeglumine for infusion times of i, 3, 5, and iO mm- utes, assuming a negligible venous accumulation and a relaxivity of 4.5 msec1 mmol’. Note that for infusion times of 3-5 minutes, a double dose is required to make the Ti of arterial blood less than the Ti of fat. (2) Graph depicts predicted relative Sb versus flip angle for a spoiled gradient-echo sequence with TR of 25 msec and TE much less than T2*, for tissues of various Ti. A flip angle of about 40#{176}is optimal for main- taming a high SI and for maximizing the SI difference between arterial blood (predicted Ti, iOO-150 msec) and fat (Ti, 270 msec). nostics, Princeton, NJ), and gadodiamide Patients xiphoid. Then a three-dimensional Fou- injection (Omniscan; Sanofl Winthrop rier-transform spoiled gradient-echo vol- Pharmaceuticals, New York, NY); the re- Gadolinium-enhanced MR arteriogra- ume was acquired that was centered on laxivities are about 4.5 msec’ mmoM. phy was performed in 125 patients re- the arteries of interest. For groups 1 Assuming a constant infusion rate, corn- ferred for routine MR angiography of the and 2, the imaging parameters were 9-12- bining Equations (1) and (2) predicts the aorta or branch vessels. The patients in- cm-thick sections with 60 partitions (1.5- arterial blood Ti as a function of infusion cbuded 77 male and 48 female patients, 2.0-mm partition thickness) oriented with time and total gadolinium administered aged 4-86 years (mean, 66 years). The the arteries in plane (ie, coronal for aor- (Fig i). For contrast material infusion main criteria for enrollment in the study toibiac disease and sagittal for the thoracic times that are longer than the recircubation included aortic aneurysm (n = 33), hyper- aorta), TR of 25 msec, TE of 6.9 msec time, the arterial blood Ti will be even tension and/or renal failure (n = 67), and (where fat and water are out of phase), shorter, since a large fraction of the in- claudication (n = 19). The patients were flip angle of 40#{176},first-order flow compen- jected gadolinium chelate will recirculate. divided into four groups, reflecting the sation, 32-40-cm field of view, and 256 x From Figure i, it is apparent that the chronological development of the tech- i92 matrix. No saturation pulses were em- shortest Ti occurs with the shortest infu- nique. ployed. The imaging time was 5 minutes 8 sion time and the largest dose of gadobmn- Group i comprised i2 patients who Un- seconds. In groups 3 and 4, the number of ium chebate. For the typical imaging times derwent dynamic imaging after bolus in- partitions was 28 with 2-mm partition of 3-5 minutes, Figure i shows that the jection of 0.2 mmob/kg gadopentetate thickness, the matrix was 256 x 256 with a dose must be at least 0.2 mmol/kg to dimeglummne during the first i-2 minutes 36-cm field of view, the TE remained the achieve a blood Ti substantially shorter of a 5 minute 8 second acquisition. Group same (6.9 msec), and the TR was 24 msec, than that of the brightest background tis- 2 comprised 25 patients who underwent resulting in a higher-resolution image and dynamic imaging during continuous infu- sue, fat (Ti = 270 msec). a shorter imaging time of 3 minutes 18 5cc- With a gadolinium chelate dose of 0.2 sion of 0.2 mmol/kg gadopentetate dimeg- onds. With this three-dimensional pulse mmob/kg and a 3-5-minute infusion dur- lumme over a 5 minute 8 second acquisi- sequence, k space was acquired in the ing imaging time, the arterial blood Ti is tion. A subset of this data (16 patients) has standard linear fashion, in which the par- predicted to be i50-200 msec. It will actu- been reported previously (iO). Group 3 titions are encoded sequentially and, for ally be shorter since the recirculation time comprised 19 patients who underwent each partition, the lines of k space are ac- is less than 3-5 minutes. The relative sig- dynamic imaging during continuous infu- quired from bottom to top (negative to nab intensity (SI) in a three-dimensional sion of 0.2 mmob/kg gadopentetate dimeg- positive).