Abdominal applications of fast MR imaging: A comparison of fast field echo (FFE) and spin echo (SE) pulse sequences

Magneric Resonance Printed in the USA.

Imaging, Vol. 7. pp. 297-303, All rights reserved.

1989 Copyright

0130-725X/89 $3.00 + .OO 0 1989 Pergamon Press plc

l Original Contribution

ABDOMINAL APPLICATIONS OF FAST MR IMAGING: A COMPARISON OF FAST FIELD ECHO (FFE) AND SPIN ECHO (SE) PULSE SEQUENCES L. TE STRAKE,* N.J.M. FRELING,* R.L. KAMMAN,~ E.L. MOOYAART,~ H. DOORENBOS,$ H. SCHRAFFORDT KOOPS§ Departments of *Radiology, TMagnetic Resonance, $Endocrinology, and &SurgicalOncologic. University Hospital Groningen, The Netherlands. Fast gradient echo (GE) MR imaging during breathhold was performed in 73 abdominal MR examinations in order to determine the value of GE technique in the reduction of movement artefacts and improvement of image quality. The results were compared with conventional spin-echo (SE) technique. Ti-weighted SE images consistently demonstrated normal anatomy and pathology. T,*-weighted and intermediate GE technique proved to be superior to Ts-weighted SE technique in a considerable number of cases providing not only better delineation of normal structures but also better lesion detection. At this stage GE technique should be considered as an adjunct to T,-weighted SE technique. If the artefacts presently inherent to GE imaging could be reduced or compensated for, this technique of rapid MR imaging during suspended respiration could become an important tool being one of the fastest techniques for abdominal MR imaging. Keywords:

MRI; Fast field echo; Abdominal application.

INTRODUCTION

Table 1. Methods: Protocol of SE and GE pulse sequences

The efficacy of MRI using conventional spin echo (SE) pulse sequences is hampered by physiological motion. Respiratory movements and pulsations during long acquisition times result in image blurring, ghost artefacts, inaccurate measurements of relaxation times, and spurious signal intensities.1,2,s Based on these observations it was predicted that MR image quality could be improved by shortened acquisition time during suspended respirationi In this paper we present our experience with abdominal applications of fast MRI during breathhold and compare these results with conventional spin-echo (SE) MRI. MATERIALS

TR

Flip Acquisition TE Angle NAV time

650 20 36 17 2000 50 2000 loo 86 25 Tz*-weightedGE 37 17 IntermediateGE 86 17 Ti-weighted SE T, -weighted GE T,-weightedSE

60 15 15

6 1 1 1 1 1

12 min 6.4 set 6 min 6 min 15.4 set 6.6 set

30

1

15.4 set

GE technique used has initially been described as a fast field echo (FFE) sequence.i6 Multiple-slice T,-and T2-weighted SE pulse sequences were compared with a T,-weighted GE sequence and GE sequences which display predominantly T2 contrast. GE images were acquired during breathhold with single-slice technique.

AND METHODS

Imaging was performed on a 1.5T MR scanner (Philips, Gyroscan S15) with a gradient strength of 3 mT/m. All examinations were performed according to a fixed imaging protocol including conventional SE and gradient echo (GE) pulse sequences (Table 1). The RECEIVED7/30/88; ACCEPTED1l/23/88. Address correspondence to N.J.M. Freling, Ph.D., Dept.

of Radiology, University Hospital Groningen, 59, 9713 EZ Groningen, The Netherlands. 297

Oostersingel

Magnetic Resonance Imaging 0 Volume 7, Number 3, 1989

298

For GE, T2 contrast was obtained by selection of a small flip angle and relatively long TE and TR. A larger flip angIe and short TR and TE GE pulse sequence was selected for T, contrast. Images with short TR and TE and small flip angle are referred to as intermediate, but in fact these images display mainly T2 contrast. (Acquisition and display matrix was 256 x 179, slice thickness 10 mm, and field of view 450 mm.) Five normal volunteers and 63 patients were examined. Five patients were examined twice. Patients were selected for MRI according to two criteria: 1. Known pathology demonstrated on CT or ultrasound. 2. Strong clinical suspicion despite negative CT or ultrasound. A total of 73 MRI examinations were evaluated in conference by two radiologists (LtS and NJMF) and graded as 0. 1. 2. 3.

Anatomy and pathology not shown. Anatomy and pathology poorly visualized. Anatomy and pathology well shown. Unique or superior demonstration of anatomy and pathology.

A variety of benign and malignant pathological conditions was studied (Table 2).

Table 2. Materials: Clinical diagnosis, confirmed on MRI (n = number of patients) Liver

cyst (n = 3), echinococcai cyst (n = 1), metastasis (n = 7), cavernous hemangioma (n = 4), cirrhosis (n = 2), hepatitis (n = I), hemochromatosis (n = l), subcapsular hematoma (n = 1).

Pancreas

carcinoma (n = 5), gastrinoma (n = 1), pancreatitis (n = 2), transplant rejection (n = 1).

Kidney

cyst (n = 1), polycystic kidneys (n = 1), renal cell carcinoma (n = 5), nonHodgkins’ lymphoma (n = I).

Adrenal gland

non-functioning adenoma (n = 1), adenoma (MEN 1, n = l), macronodular hyperplasia (primary hyperaldosteronism, n = 2), non-Hodgkins’ lymphoma (n = l), hemorrhagic cyst (n = 1).

Cl tract

carcinoma of stomach (n = 1), carcinoma of sigmoid (n = 1).

Retroperitoneal space

metastatic lymphonodes (n = 3), nonHodgkins’ lymphoma (n = 2), extraadrenal pheochromocytoma (n = 1).

lntraperitoneal space

omental metastases of ovarian carcinoma (II = I), metastatic adenocarcinoma of colon (n = 1), non-Hodgkins’ lymphoma (n = 1).

Vascular system

portal vein thrombosis (n = 3), hepatic artery aneurysm (n = 1).

RESULTS T,-weighted GE (TR = 36 msec, TE = 17 msec, flip angle = 60”) images showed good cortico-medullary demarcation in the kidney and maximum contrast between blood flow and stationary tissue. Small intrarenal and intrahepatic blood vessels are easily identified due to the high signal compared to surrounding tissue (Fig. l(A)). GE images obtained with longer TR and TE and particularly smaller flip angle (86/25/15) provide T2 contrast. On these images signal intensity of flow is less pronounced and in the kidney the cortico-medullary demarcation is less distinct (Fig. l(B)). The results of the scoring of the MRI examinations are listed in Table 3. T,-weighted SE technique most consistently demonstrated normal anatomy and pathology. Score 1 and 2 are similarly distributed for T,-weighted SE and intermediate and T;-weighted GE technique. However, unique or superior demonstration of pathology (score 3) was considered to be more frequently present on intermediate

Table 3. Results: Total scores of MRI examinations (n = 73) TR/TE/or

0

1

2

3

T,-weighted SE T,-weighted FFE

450-650/20 36/17/60

1 3

9 41

60 28

3 1

Tz-weighted SE T,-weighted SE T,*-weighted FFE

2000/50 2000/100 86/25/15

2 4 3

30 30 30

39 36 34

2 3 6

Intermediate Intermediate

37/17/15 86/17/30

2 1

36 29

34 40

2 2

FFE FFE

and T;-weighted GE images than on T,-weighted SE images. The low contrast T,-weighted GE images were considered to provide poor visualization of anatomy and pathology (score 1) most frequently and

Abdominal applications of rapid MRI 0 L.

299

TE STRAKE ET AL.

B Fig. 1. Normal anatomy. (A) T,-weighted GE 36/17, flip angle 60”. Note corticomedullary demarcation in kidneys and bright signal of intra renal and hepatic blood vessels. (B) T;-weighted GE 37/15, flip angle 15”. Decreased cortico-medullary demarcation in kidneys and decreased flow signals.

were certainly inferior to conventional ( T, -weighted) SE imaging. Liver metastases were visualized on Tiweighted SE images in all 7 patients. Both T2-weighted SE and T;-weighted GE sequences failed to demonstrate liver metastases in two cases. In one case, T;weighted GE images provided superior demonstration of liver metastases (Fig. 2). Two cavernous hemangiomas showed the typical appearance of a high signal intensity lesion on T2weighted SE and TT GE technique. A third hemangioma remained isointense on TT-weighted GE images and showed high signal intensity on the T2-weighted SE sequence. A fourth small hemangioma, 1 cm in diameter could not be demonstrated on any sequence. Portal vein thrombosis was equally well-diagnosed on

c Fig. 2. Liver metastases of bronchogenic carcinoma. (A) SE 450/20. Irregular liver contour and inhomogenous signal intensity of the liver. (B) SE 2000/100. Poor visualization of metastases due to extensive blurring. (C) GE 86/25, flip angle 15”. Superior demonstration of metastases. The smallest lesion measures 1 cm in diameter.

300

Magnetic Resonance Imaging 0 Volume 7, Number 3, 1989

SE and GE technique as an area of increased respectively decreased signal intensity within the lumen of the portal vein. In all cases in which GE technique was superior to T,-weighted SE technique, the SE images showed blurring and loss of spatial resolution. Blurring could be subtle as in the case of a small adrenal nodule, which was only clearly demarcated on GE images (Fig. 3). Complete loss of organ visualization on SE Tz-weighted images was observed in more severe cases of movement artefacts (Fig. 4). Pancreatic carcinoma and also a small gastrinoma were welldemonstrated on TT -weighted GE images. We have observed, that with intermediate GE technique using a flip angle of 25”-30” fat becomes isointense with muscle. On these GE images, demarcation of the high signal intensity retroperitoneal tumor from adjacent hypointense psoas muscle and retroperitoneal fat is possible on one image, whereas SE technique requires both T, - and T2-weighted sequences for determination of the extension of retroperitoneal disease (Fig. 5). Despite these examples of improved image quality with GE technique and superior demonstration of pathology, there are disadvantages to the GE technique used in this study. Motion artefacts due to flow are very pronounced on particularly T, -weighted images due to the large flip angles and may mask pathology in case of inadequate selection of the preparation gradient (Fig. 6(A),(b)). Signal cancellation shown as black lines occurs at fat-organ interfaces, if the fat and water protons are 180” out of phase (Fig. 6(C)). Furthermore, signal intensity differences between air and tissue may lead to signal loss and nonvisualization of structures surrounding the air. This susceptibility artefact may mask intraperitoneal pathology (Fig. 6(D)). A general observation in all cases was the similar high or low signal intensity of different lesions including hemorrhagic adrenal or renal cysts on T, -weighted SE and GE images and on T2-weighted SE and TT GE images. DISCUSSION Motion artefacts severely degrade the quality of abdominal MR images. Several techniques have been evaluated to reduce motion artefacts. Respiratory gating can improve image quality, but never became clinically useful because of the prolongation of the examination time.9 Respiratory ordered phase encoding (ROPE) did not have the disadvantage of prolongation of exami-

Fig. 3. Adrenal macronodular hyperplasia in patient with primary hyperaldosteronism. (A) SE ZOOO/lOO.Enlargement of left adrenal demonstrated. (B) GE 86/17, flip angle 30”. Superior demonstration of hypo-intense nodule on posterior aspect of adrenal gland. (C) CT image confirms findings on GE image.

Abdominal applications of rapid MRI 0 L. TE STRAKEar AL.

301

Fig. 4. Pancreatic carcinoma. (A) SE 650/20. Clear demarcation of slightly enlarged body of pancreas. (B) GE 86/17, flip angle 30”. Excellent demarcation of pancreas and necrosis, which was confirmed on CT. (C) SE 2000/50. Poor visualization of contour and structure of pancreas. (D) 1SE 2000/100. Complete non visualization of pancreas due to extensive blurring.

nation time, but this technique was too dependent on the co-operation of a regularly breathing patient.3 The short TI inversion recovery sequence (STIR) reduces ghost artefacts from subcutaneous fat due to respiratory movements.6 However, this technique requires long acquisition times and does not control other motion artefacts. GE pulse sequences, however, using short TR, allow for data acquisition during suspended respiration. As with SE pulse sequences, the use of short TR and TE with GE technique will result in T, contrast; long TR and TE will provide T2 contrast. GE technique offers the flip angle as a third option to manipulate contrast.5 In fact, in order to create T, contrast the effect of the selection of a small flip angle is more important than the choice of a long TE.4 Since T2 contrast in GE imaging is not only dependent on the spin interactions but also on the inhomogeneities of the main magnetic field, for GE

technique the term T,t is used.‘,15 The results of our study confirm that T, -weighted SE images consistently show normal anatomy whereas the pathology often is readily shown by the other pulse sequences and are therefore overall superior to T2-weighted SE and GE pulse sequences.15 The optimal technique for detection of liver metastases has to be determined yet. Also in our limited series of 7 patients, some liver metastases were better demonstrated on T2+-weighted GE images than on the T2-weighted SE images7 A proper choice of the flip angle is critical for optimal contrast, particularly in the diagnosis of retroperitoneal lymphadenopathy. The maximum intensity for tissue with long T, will be at small flip angles, whereas for tissue with short T, the maximum intensity will occur at a larger flip angle. ” A peak intensity for fat has been observed at 70 0 . 12

302

Magnetic Resonance Imaging 0 Volume 7, Number 3, 1989

Fig. 5. Retroperitoneal non Hodgkin’s lymphoma infiltrating teft psoas muscle. (A) SE 6.50/20. Tumor is well demarcated from retroperitoneal fat. (B) SE 2000/50. Poor delineation of tumor from retroperitoneal fat, infiltration of psoas muscle is obvious. (C) GE 86/15, flip angle 25”. Note that fat and muscles are iso intense. Excellent demonstration of tumor extension into psoas muscle and retroperitoneal space on single GE image.

In our experience retroperitoneal fat and psoas muscle are of the same Iow signal intensity at flip angles of 25”-30”. Retroperitoneal lymphnode metastases appear bright with a flip angle of 30”, which results in optimal contrast between tumor and both fat and muscle. Because of the absence of respiratory movement artefacts GE TT-weighted images were superior to SE T2-weighted images in a considerable number of cases, providing improved image quality and superior lesion detection. Since we did not have the software available to compensate for flow, flow artefacts due to pulsatile blood flow are a serious problem requiring proper selection of the preparation gradient. The application of presaturation pulses appears to be most useful to reduce flow artefacts.“,” Motion artefact suppression technique (MAST) is another technique which reduces motion artefacts by refocussing the signal from flow.13 However, this technique appears to be less useful in GE technique, since MAST requires relatively long TE. Long TE values are not desirable in GE imaging, since they tend to enhance the field inhomogeneity effects.4 Another limitation of our study was the gradient strength. The low gradient strength of 3 mT/m did not allow the selection of a shorter combination of TR and TE than 36/17 for a field of view of 450 mm and an acquisition matrix of 256 * 179. This means that possibly no optimal T,-weighted GE images could be acquired and that our results of T,-weighted GE technique may be improved at higher gradient strength. The effects of field inhomogeneity due to the absence of 180” refocussing pulses with GE imaging, the magnetic susceptibility artefacts and fat-water cancellation artefacts are not typical for our study and need to be minimized for optimal image quality. At this stage we agree that GE imaging should be considered as an adjunct to 7’,-weighted SE technique.‘5s’7 The results of our study have also shown that T,*weighted GE technique is superior to T2-weighted conventional SE technique in a considerable number of cases. These results need confirmation in a larger number of patients and more experience is needed with pulse sequence optimization in GE imaging. However, with technical improvements, GE imaging could become an important tool being one of the fastest MR imaging techniques for the evaluation of the abdomen. REFERENCES I. Aisen, A.M.;

Glazer, G.M.; Carson, P.L.; Heashen, D.O. Mugn. Reson. Imaging 4:207-213; 1986.

Abdominal applications of rapid MRI 0 L.

303

TE STRAKE ET AL.

Fig. 6. Artefacts. (A) SE 600/20 and (B) GE 36117, flip angle 60”. Macronodular hyperplasia adrenal glands. Nodule in right adrenal masked on GE by flow artefacts. (C) GE 86117, flip angle 30”. Non functioning low signal intensity adenoma of the right adrenal. Single cancellation artefact at fat-organ interfaces, shown as black lines. (D) GE 37/17, flip angle 15”. Omental metastases of ovarian carcinoma adjacent to air-filled ascending colon are masked by the susceptibility artefact.

2. Axel, L.; Summers, R.M.; Kressel, H.Y.; Charles, C. Radiology 160:795-801;

1986.

3. Bailes, D.R.; Gilderdale, D.J.; Bydder, G.M.; Collins, A.G.; Firmin, D.N. J. Comput. Assist. Tomogr. 9:835838; 1985.

4. Buxton, R.B.; Edelman, R.R.; Rosen, B.R.; Wismer, G.L.; Brady, T.J. J. Comput. Assist. Tomogr. 11:716, 1987. 5. Bydder, G.M.; Payne, J.A.; Collins, A.G.; Thomas, D.G.T.; Davis, C.H., Cox, I.J.; Ross, B.D.; Young, I.R. J. Comput. Assist. Tomogr. 11:17-23; 1987. 6. Bydder, G.M.; Pennock, J.M.; Steiner, R.E.; Khenia, S.; Payne, J.A.; Young, I.R. Magn. Reson. Imaging 3: 251-254;

1985.

Edelman, R.R.; Hahn, P.F.; Buxton, R.; Wittenberg, J.; Ferrucci, J.T.; Saini, S.; Brady, T. J. Radiology 161:125-131;

1986.

Ehman, R.L.; Kjos, B.O.; Hricak, H.; Brasch, R.C.; Higgens, C.B. J. Comput. Assist. Tomogr. 9:315-319; 1985.

Ehman, R.L.; McNamara, M.T.; Pallack, M.; Hricak, H.; Higgens, C.B. AJR 143:1175-1182; 1984.

R.L. Radiology 164:559-564;

10. Felmlee, J.P.; Ehman, 1987.

11. Frahm, J.; Merboldt, K.D.; Hanicke, W.; Haase, A. Mag. Res. Med. 4~372-377; 1987.

12. Mills, T.C.; Ortendahl, D.A.; Hylton, N.M.; Crooks, L.E.; Carlson, J.W.; Kaufman, L. Radiology 162:531539; 1987.

13. Pattany, P.M.; Philips, J. J.; Chiu, L.C.; Lipcamon, J.D.; Duerk, J.L.; McNally, J.M.; Mohapatra, S.N. J. Comput. Assist. Tomogr.

11:369-377;

1987.

14. Schultz, C.L.; Alfidi, R. J.; Nelson, A.D.; Kopiwoda, S.Y.; Clampitt, M.E. Radiology 152:117-121; 1984. 15. Utz, J.A.; He&ens, R. J.; Johnson, C.D.; Shimakawa, A.; Pelt, N. ; Glover, G.; Johnson, G. A.; Spritzer, C.E. AJR 148:629-633;

16. van der Meulen,

1987.

P.; Groen,

J.P.;

Magn. Reson. Imaging 3:297-299:

Cuppen,

J.J.M.

1985.

17. Winkler, M.L.; Ortendahl, D.A.; Mills, T.C.; Crooks, L.E.; Sheldon, P.E.; Kaufman, L.; Kramer, D.M. Radiology 166:17-26;

1988.