- Email: [email protected]
Magnetic resonance imaging of the uterus at an ultra low (0.02 T) magnetic field
Mogneric Resonance Printed in the USA.
Imaging, Vol. 10, pp. 195-205. All rights reserved.
1992 Copyright
0
0730-725X192 $5.00 + .oO 1992 Pergamon PressLtd.
l Original Contribution
MAGNETIC RESONANCE IMAGING OF THE UTERUS AT AN ULTRA LOW (0.02 T) MAGNETIC FIELD M. Departments
VARPULA,*
M.
KOMU*
of *Diagnostic Radiology and tPathology
AND P. KLEMIT
of the University Central Hospital of Turku, Finland
In vivo pelvic imaging of 39 women and in vitro relaxation time measurements of four uterine specimens were performed using an ultra low field (0.02 T) MRI unit. Average T1 times measured in vitro at 37°C for the myometrium and endometrium were 206 ms (SD 47 ms) and 389 ms (SD 21 ms), respectively. Corresponding T, times were 95 ms (SD 20 ms) and 167 ms (SD 13 ms). The proton relaxation of almost all myometrial specimens proved to be biexponential, but of all endometrial specimens was monoexponential. Contrast measurements between endometrium versus myometrium and myometrium versus the junctional zone were performed after imaging 18 volunteer women using different pulse sequence parameters. Normal uterine structures were optimally demonstrated by SE 700/70. Relatively short repetition times could be used, because spin-lattice relaxation times were short at the low magnetic field. Consequently, the short repetition times allowed averaging of four excitations to create adequate images within an acceptable scanning time. In addition to Tz-weighted images a T,-weighted inversion recovery sequence with a short inversion time of 50 ms (IR 1000/50/40) adequately differentiated the three uterine zones. Although pathologic lesions of the uterus including leiomyomas, anomalies and carcinomas were well demonstrated, especially with the Tz-weighted spin echo pulse sequence, further investigations are needed to evaluate the optimal technique for ultra low field MR imaging of uterine tumors.
Keywords: Uterus; MR studies; Uterus, relaxation times; Uterine neoplasms,
MR studies.
to image the uterus, and to find optimal rameters for that purpose.
INTRODUCTION
Magnetic resonance imaging (MRI) has been reported to have a great potential for the examination of the normal and pathologic uterus. l-l4 The uterine corpus, cervix and vagina and their relationships to the bladder and rectum are clearly delineated by sagittal images. Cyclic changes and other hormonal effects can be demonstrated.‘*” Many studies have suggested that MR imaging is accurate in the detection of uterand anomalies. I1 ine leiomyomas, 2*4s adenomyosis6 The extent of the growth of malignant tumors is reported to be easy to perceive.7~8~12-‘4 A wide range of magnetic fields, commonly 0.15 1.5 T, has been applied. Only in a few experiments is imaging in ultra low field strength (ULF) (0.02-0.045 T) reported. ‘%I6 The aim of this investigation was to evaluate the capabilities of an ultra low magnetic field unit (0.02 T)
MATERIALS
AND
imaging
pa-
METHODS
In Vitro Measurements To help define the optimal pulse sequences for imaging the uterus, both spin-lattice (Tl) and spin-spin (T2) relaxation times were measured in vitro. Myometrial and endometrial specimens were excised from four uteri immediately after removal of the organ. The operations were performed for leiomyomas or bleeding problems. The patient’s ages ranged from 41 to 60 years. One of them had used hormonal substitution (estradiol) just before surgery. The volume of each myometrial specimen was about one cubic centimeter. The endometrial specimens were smaller, one of which was too small for measurements. The specimens were
RECEIVED 6/l l/91; ACCEPTED 8/21/91. Financial assistance from the Cancer Society of Southwestern Finland and the Instrumentarium Corp., Helsinki, Finland, is gratefully acknowledged.
Address all correspondence to Matti Varpula, M.D., Dept. of Diagnostic Radiology, University Central Hospital, 20520 Turku, Finland. 195
196
Magnetic Resonance Imaging 0 Volume 10, Number
immediately sealed in glass tubes with rubber stoppers and kept for 3 to 6 h (mean 4.5 h) at +6”C until measured. After relaxation time measurements all specimens were examined histologically. Serum estradiol levels of the patients were also recorded on the day of operation. The relaxation time measurements were performed with an ultra low field (ULF, 0.02 T) MRI unit (Acutscan, Instrumentarium Corp., Helsinki, Finland) operating at a frequency of 833 kHz in spectrometer mode, using a special device for in vitro measurements.28 T1 determination was based on an inversion recovery (IR) pulse sequence 180’-TI-90” with repetition time (TR) of 2000 ms. At least 10 different inversion times (TI) from 6 ms to 850 ms were used for each relaxation time measurement. T2 determination was done using a standard spin echo (SE) pulse sequence with at least 10 different echo times (TE) from 15 ms to 325 ms and with TR of 2000 ms. Relaxation times were then calculated by fitting the measured signals to the biexponential relaxation curves. The temperature of the specimens was kept constant (37 f 0.4”C) during measurements with an electronic heater. The weighted averages (Tavg) of the short (T,) and long (7;) components of the relaxation times were calculated based on the equation: l/T,,,
= w”/T, + WI/T,
where ws and w’ represent the corresponding factors of the components of T, and 7i.
(1) weight
Volunteer Examinations Eighteen female volunteers without any gynecologic problems (age 23-54 years, mean 39 years) were examined in order to determine the proper imaging parameters for the uterus. All examinations were performed using a U-shaped close coupled body coil underneath the pelvis with the patient in either the supine or prone position. The imaging matrix was 256 x 256, and the field of view was 268 x 402 mm. The slice thickness was 10 mm. The number of contiguous slices obtained simultaneously varied from one to seven depending on the selected pulse sequence. Sagittal, axial and coronal imaging planes were used. Various spin echo (SE), saturation recovery (SR, gradient echo) and inversion recovery (IR) sequences were used. TR of the spin echo sequences varied from 600 to 1500 ms and TE from 70 to 150 ms. Respectively, TR of the SR sequences varied from 160 to 1000 ms and TE from 30 to 70 ms. TR of the IR sequences was 1000 ms, TI 50 ms or 250 ms, and TE 40 ms. The shortest available TE for the spin echo sequences was
2, 1992
70 ms. Images with short TR and TE were obtained using SR sequences. One to four excitations were averaged in SE and SR sequences with long TR, up to 16 excitations in SR sequences with short TR, and one to two excitations in IR sequences. Signal intensity of the myometrium, endometrium and junctional zone was measured using the cursor to define the region of interest (ROI). The percent contrast in signal intensity of the endometrium versus myometrium and the myometrium versus junctional zone was calculated from the intensity measurements made of images obtained at different pulse sequence parameters. Calculations were based on the equation: % Contrast = y
x 100
(2)
where II and Z2are the intensities of different tissues. Clinical Examinations Twenty-one patients with different gynecologic problems were imaged in order to evaluate the capability of the ULF MRI to detect pathologic conditions (Table 1). Most cases had histopathological confirmation at operation (16 cases) or at curettage (2 cases); one uterine anomaly was verified by hysterosalpingography and two leiomyomas by clinical examination. Hysteroscopy was performed on all the patients with endometrial carcinoma, and computed tomography on all the patients with malignant tumors. RESULTS In Vitro Measurements The spin-lattice (TI) and spin-spin ( T2) relaxation times of the four myometrial and the three endometrial specimens are presented in Table 2. Three of the myometrial specimens had a biexponential spin-lattice relaxation and two of them also had a biexponential spin-spin relaxation. All relaxations of the endometrial specimens were monoexponential. In the cases of the biexponential relaxation the weighted averages of the relaxation times were calculated to help compari-
Table 1. Diagnosis
of 21 patients
imaged
Leiomyoma Uterine anomaly Invasive hydatidiform mole Endometrial carcinoma (adenocarcinoma) Cervical carcinoma (squamous cell carcinoma)
by MRI
Ultra low MR imaging
of uterus 0 M. VARPULA ET AL.
197
Table 2. Spin-lattice (T,) and spin-spin (T2) relaxation times (ms) of the myometrium and endometrium of four uterus specimens measured in vitro at body temperature Myometrium Endometrium Uterus 1
2 3 4 Average SD
Tl short 140 (44%) 28 (13%) 130 (37%) 99 51
478 253 420 238 347 104
Tl long
Tl avg
G short
T, long
T2 avg
Tl
r,
(56%) (87%) (63%) (100%)
232 124 230 238
123 (100%) 85 (100%) 80 (76‘70) 59 (78%)
455 (24%) 221 (22%)
123 85 100 70
400 359 408 -
175 149 177 -
95 20
389 21
167 13
206 47
87 23
338 117
The short and long components of the relaxation times and their calculated of the short and long components of the relaxation time are in parentheses. presented.
Estradiol (pmol/L) 420 50 220 50
weighted averages (avg) are presented. The relative weights (070) The serum estradiol levels on the day of measurement are also
Demonstration of Normal Structures Uterus, cervix and vagina were well demonstrated on T,-weighted images, and the three uterine zones with high-intensity endometrium, intermediate-intensity myometrium and low-intensity junctional zone were clearly seen (Fig. 1A). SE 700/70 proved to be ideal pulse sequence for the T2-weighted images of the uterus. The scanning time (12 min) was not too long for the patient, despite the fact that four excitations were needed for diagnostic images. The contrast between myometrium, junctional zone and endometrium were optimal using SE 70170 (Figs. 2A, 2B). Contrast between endometrium and myometrium was greatest with pulse sequence SE lOOO/lOO, however. Longer TR and TE had no advantages. Shorter TR (500 ms) did not diminish the con-
son with previous in vitro and in vivo measurements (Table 3). Mean T’, relaxation times of the myometrium and endometrium were 206 ms (SD 47 ms) and 389 ms (SD 21 ms), respectively. Corresponding T2 relaxation times were 95 ms (SD 20 ms) and 167 ms (SD 13 ms). The relaxation times of the second myometrial and endometrial specimens and the T2 time of the fourth myometrial specimen were shorter than those of the other specimens. This may have been due to the low serum estradiol levels of the corresponding patients. The relaxation times of the fourth myometrial specimen may have been affected by the hormonal substitution the patient had taken just before surgery. The nature of each specimen was confirmed histopathologically. The three endometrial specimens were in the secretory phase.
Table 3. Relaxation times of the uterus according to the literature and to our measurements Myometrium Magnetic field (Tesla/MHz) In vitro Olszewski et a1.26 Fruchter et al.” McCarthy et al. I9 Our measurements
90 MHz 22.5 MHz 20 MHz 0.02 T/0.833 MHz
In vivo McCarthy et al. I9 Hricak et al.’ Schmidth et al. ” Demas et al. I0 Hamlin et al.’ Moore et al.16
1.5 T/64 MHz 0.35 T 0.35 T 0.35 T 0.15 T 0.045 T
T, (ms)
1365 553 709 206
i f + +
142 42 71 47
650 700 Z!Z150 814 -t 112 579 f 5 -
Endometrium T2 (ms)
70 + 66t67 -t 95 f 63 60 50 55 52
6 10 8 20
T, (ms)
801 836 + 172 389 * 21
k 12
600
+ 10 * 5 + 8 -
1077 -t 115 264 f 2 175
Tz (ms)
95 87 * 23 167 + 13 85 f 25 70 61 * 5 -
198
Magnetic Resonance Imaging 0 Volume 10, Number 2, 1992
(B)
03 Fig. 1. Normal uterus. (A) 7”-weighted sagittal image (SR 500170). T‘he three uterine zones with high-intensity endometri um, intei rmediate-intensity myometrium and low-intensity junctional zone are clearly demonstrated. The cervix with low-inter isity cervical canal are well seen. (B) T,-weighted axial image of the normal uterus (SE 1000/70). (C) The stro: ma and high-intensity con1:rast between uterine zones dimihishes with a shorter TE (SR 500, ‘30). (p) When both TR and TE are short (SR 160/ 30), no zonal contrast can be seen.
trast markedly. The difference between intensities of the myometrium and junctional zone was greatest with an echo time of 125-150 ms. TR had no real effect on the contrast between the myometrium and the junctional zone. With TR under 1000 ms the myometrium had a higher signal intensity than urine but with TR 1500 ms the signal from urine was more intense than that of myometrium. With echo times over 100 ms the signal from the uterus diminished and with echo times shorter than 70 ms the contrast between the uterine zones diminished. Still, on SR 500/30 images the endometrium could be distinguished from the myometrium (Fig. 1C). When TR was shortened to 160 ms,
this contrast
also disappeared and the uterus was seen as a homogeneous organ (Fig. 1D). On T,-weighted images obtained with inversion recovery sequence the three uterine zones were seen when the inversion time was 50 ms (Figs. 3A, 3B; Fig. 4A). Contrary to the T,-weighted image the endometrium had a very low signal intensity, the junctional zone was most intense and the myometrium had a slightly lower intensity than the junctional zone. If the inversion time was 250 ms the myometrium and junctional zone were not usually discernible from each other and only a small area of endometrium could be seen (Fig. 4B). The uterus was best demonstrated on sagittal im-
Ultra low MR imaging of uterus 0 M. %
VARPULA
ET AL.
199
Contrast
60
Pulse Sequence
Parameters
(TRITE)
(4 % Contrast
60
30
0 160130
250/30
SOOf
30
700/70
ages, but was also well seen on the axial plane (Fig. 1B). On 7”-weighted images the uterine cervix had a low signal intensity with a central high-intensity stripe representing the endocervical canal (Fig. 1A). On images
1000170
with short TR and TE (SR 160/30) the cervix had a homogeneous appearance with an intermediate signal intensity. The slice thickness was 10 mm, which is relatively
200
Magnetic Resonance Imaging 0 Volume 10, Number 2, 1992 % Contrast
’
-50
I
I
1000/60/40
1000/250/40
Pulse Sequence
Parameters (A)
(TRAFVTE)
9%Contrast
-40
-50
-
I
I
1000150/40
10001250/40
Pulsb’~~~quen~e
ParaFeters
(TR/IR/f
E)
0% Fig. 3. (A) The signal intensity of endometrium relative to myometrium at two different inversion recovery pulse sequences (see text).
thick for imaging of uterus, so interleaved scans at 5-mm increments were used. Because a U-shaped surface coil was used, the signal intensity decreased considerably with increasing distance from the coil. The
(070)and (B) myometrium relative to junctional
zone
supine position of the patient was preferred, although the signal intensity of the anteflexed uterine corpus was higher in the prone p,Qsition. Low signal intensity structures such as the vagina and the cervix were bet-
201
Ultra low MR imaging of uterus 0 M. VARPULA ET AL.
(A)
@I
Fig. 4. Sagittal images of the normal uterus at two different inversion recovery sequences. (A) IR lC00/50/40. All three uterine zones are well seen. In contrast to the T,-weighted image (Fig. 1A) the junctional zone has the highest and the endometrium the lowest signal intensity. (B) IR 1000/250/40. The junctional zone cannot be discerned. A narrow low-intensity endometrial stripe can be seen.
ter seen when the coil was underneath since the signal intensity sufficiently high.
of the uterine
the buttocks, corpus was still
Pathologic Findings As well, the normal anatomy
uterine pathology was also well demonstrated. Five patients had a total of six uterine leiomyomas. Their diameter varied from 10 to 60 mm. The signal intensity of the leiomyomas was lower than that of the myometrium on T2-weighted images (Fig. 5). A septate uterus was well demonstrated on Tz-weighted axial image (Fig. 6).
Fig. 5. Sagittal T,-weighted image (SR 500/70). An intramural leiomyoma (arrow) has a signal intensity lower than that of the myometrium.
Two young women (aged 26 and 30 years old) with invasive hydatidiform moles were imaged. The diagnosis was verified by curettage. In the one case a highintensity focus was found within the myometrium on T2-weighted images (Fig. 7). The patient was not operated on but ultrasonically a sponge-like area typical of an invasive mole was seen in the myometrium. In the other case the entire uterine body was inhomogeneous with relatively high signal intensity, corresponding to the pathologic vasculature demonstrated by angiography.
Fig. 6. 7”-weighted (SE 1000/70) axial image of a uterine anomaly (septate uterus). Two separate uterine cavities (asterisks) and two separate cervical canals (small arrows) are well seen in the undivided uterine body (arrowheads).
202
Magnetic Resonance Imaging 0 Volume 10, Number
2, 1992
lated endometrium of a high signal intensity (Fig. 8A). In two cases also small areas of low signal intensity were seen in the dilated high-intensity endometrium (Fig. 8B). At hysteroscopy they proved to be exophytic growth of carcinoma. If carcinoma was entirely confined to the endometrium (three cases), the junctional zone was well demarcated (Fig. 8A). However, in one of the two cases with superficial myometrial involvement the junctional zone was also intact. If carcinoma involved at least half of the myometrial thickness (four cases) the junctional zone was not visible, and the border between the endometrium and myometrium was indistinct (Fig. 8B). In all cases the uterus was well demarcated on MR images and the tumor growth outside the uterus was easily excluded. MRI findings of two cervical carcinomas were normal. One of them was in situ type (IA) and in the other case the carcinoma was apparently removed by conisation before imaging. Two stage IB carcinomas were seen on T2-weighted scans as central high-intensity tumors surrounded by a normal low signal intensity cervical stroma (Fig. 9). Parametrial tumor growth was therefore easily excluded.
Nine patients with endometrial carcinoma and four patients with cervical carcinoma were imaged. All of them had stage I disease, and they were operated on after radiation therapy. Histopathologically the endometrial carcinomas were adenocarcinomas and the cervical carcinomas were squamous cell carcinomas. The most common finding of the endometrial carcinoma on T2-weighted images (seven cases) was a di-
ation time measurements of the uterus are described in
(4
(W
DISCUSSION SOme in vivo1,5,10,16,17,19 and in vitro’8,19,26 relax_
Fig. 8. (A) &weighted sagittal image (SE 700/70) of a Stage IA endometrial carcinoma. The tumor was confined histopathologically to the endometrium. The high-intensity endometrium is slightly dilated. Because the low-intensity junctional zone is intact, myometrial invasion of carcinoma can be excluded by MRI. (B) T2-weighted sagittal image (SE 700/70) of a Stage IA endometrial carcinoma. This tumor invaded histopathologically halfway through the myometrium. Exophytic carcinoma tissue is seen as areas with low signal intensity in dilated endometrium with high signal intensity. The junctional zone is not detectable and the tumor can be interpreted as invading to the middle of the myometrium. The uterus is well circumscribed excluding the growth of carcinoma outside the uterus.
Ultra low MR imaging of uterus 0 M. VARPULA ET AL.
(A)
203
09
Fig. 9. Sagittal (A) and axial (B) T,-weighted images (SE 700170) of a Stage IB cervical carcinoma confining to the cervix. The tumor with high signal intensity (asterisk) is entirely surrounded by uninvolved cervical stroma with low signal intensity. Tumor growth outside the cervix is easily excluded by MRI. (u: involuted uterine body)
the literature (Table 3). Most of them have been done at magnetic fields from 0.15 to 1.5 T. Only one author has measured T, of the myometrium.and the cervix at an ultra low magnetic field (0.045 T).“j As expected, the spin-lattice relaxation times (T,) at our ultra low magnetic field proved to be shorter than at higher field strengths. This field dependence of spin-lattice relaxation time is well known.20*21 The spin-spin relaxation time does not depend as much on field strength. The T2 relaxation times in our material, especially those of endometrial specimens, were longer than those described at higher magnetic fields. Individual variations may have been responsible for this finding. The relaxation times of the myometrium and endometrium seem to correlate with serum estradiol levels, although our material is not sufficiently large to confirm such a correlation. Both spin-lattice and spin-spin relaxations of almost all myometrial specimens were biexponential. Multiexponential T, relaxation of the myometrium has been reported by Adamski et al.,** but the corresponding phenomenon of T2 relaxation of the myometrium, to our knowledge, has not been described previously. Multiexponential relaxation seems to be typical for both smooth and striated muscles. The multiexponentiality of both longitudinal and transverse relaxations of skeletal muscles is well known.23-25 Different water phases of tissue are considered to be responsible for this phenomenon.24*25 Several studies have suggested that T2-weighted images are most useful for imaging the uterus and its pathology.‘-‘4 Relatively short spin-lattice relaxation times at our ultra low magnetic field made it possible to use shorter repetition times (500-1000 ms) than at
higher fields (commonly 2000 ms or longer). Our results suggest that the contrast between different uterine zones will not be suppressed by a shorter TR. One report concerning high magnetic field strength supports this result.27 Because the spatial resolution of our 0.02 T unit is lower than that of higher fields, averaging of at least four excitations was needed for an acceptable signal to noise ratio. If longer (1000-1500 ms) repetition times had to be used, the scanning time would have become impractical (17-25 min), and motion of the uterus, bowel or patient would have muddled the images. Using SE 700/70 with averaging of four excitations, four consecutive 10 mm thick slices were available during a scanning time of 12 min. Because the slice thickness was 10 mm, interleaved scans at 5-mm increments were used. If the uterus was imaged in both sagittal and axial directions and some images with short TR and TE were obtained, the entire examination time was about 1 h. The scanning time could be further reduced by using a shorter repetition time. The zonal contrast of the uterus would still be sufficient for some purposes, that is, for imaging of uterine anomalies. In the present material the normal uterus with its zonal structures was well demonstrated both on Tland T2-weighted images. On Tl-weighted inversion recovery (IR) images the zones were well seen when short inversion time (50 ms) was used. On images with short TR and TE (SR 160/30) the zones were not discernible. The T,contrast of the images obtained with the latter kind of sequences is not as T,-specific as that of inversion recovery sequences. Pathologic changes of the uterus, such as leiomyomas, anomalies and malignant tumors, were well dem-
204
Magnetic Resonance Imaging 0 Volume 10, Number 2, 1992
onstrated by our ultra low field strength magnetic unit. T2-weighted SE images were most useful. SE 700/70 is a proper pulse sequence as well for imaging of normal structures of uterus as for imaging of uterine pathology. Endometrial carcinoma has no specific MRI findings. Differentiation of adenomatous hyperplasia and blood clots from carcinoma is reported to be impossible. l4 Cervical carcinoma has higher signal intensity than that of the normal cervical tissue on T2-weighted images, but if tumor is small it is not discernible even at high magnetic fields. ‘2,‘3 The value of MRI is in the staging of the carcinomas. Vaginal and parametrial infiltration of the cervical carcinoma and myometrial invasion by endometrial carcinoma have been evaluated relatively precisely on images obtained using high magnetic field units. 73,12-14 Our material is very restricted and more advanced conclusions are not justified. However, our results indicate that it is possible to evaluate the myometrial involvement of the endometrial carcinoma and the confinement of the cervical carcinoma to the uterine cervix using MRI at 0.02 T. The benefits of an ultra low magnetic field imaging of the pelvis are its low cost, greater convenience, and relatively short scanning times. Image quality is fairly good and relatively small pathologic processes are discernible. Although the spatial resolution is inferior to that of higher field strengths, the diagnostic information is satisfactory for clinical practice. Because the surface coil is U-shaped the signal intensity diminishes considerably with increasing distance from the coil. This may be a problem, particularly with obese patients. Developing a ring-like surface coil would resolve this problem. Further investigations are also needed to evaluate other problems, especially those.concerning the staging of carcinomas. The results of the present material are promising. REFERENCES
’
Hricak, H.; Alpers, C.; Crooks, L.E.; Sheldon, P.E. Magnetic resonance imaging of the female pelvis: Initial experience. AJR 141:1119-1128; 1983. Hricak, H. MRI of the female pelvis: A review. AJR 146:1115-1122; 1986. Lukas, P.; Schrock, R.; Rupp, N.; Reiser, M.; Allgayer, B.; Fenerbach, St.; Hofrichter, A.; Heller, H.J. Die MR-tomographie bei gynakologischen erkrankungen im kleinen becken. Fortschr. RGntgenstr. 144(2):159-165; 1986. ‘. Butler, H.; Bryan, P.J.; LiPuma, J.P.; Cohen, A.M.; El Yousef, S.; Andriole, J.G.; Lieberman, J. Magnetic resonance imaging of the abnormal female pelvis. AJR 143:1259-1266; 1984.
5. Hamlin, D.J.; Pettersson, H.; Fitzsimmons, J.; Morgan, S. MR imaging of uterine leiomyomas and their complications. J. Comput. Assist. Tomogr. 9(5):902-907; 1985. K.; Itoh, K.; Fujisawa, I.; 6. Togashi, K.; Nishimura, Noma, S.; Kanaoka, M.; Nakamo, Y.; Itoh, H.; Ozasa, H.; Fujii, S.; Mori, T. Adenomyosis: Diagnosis with MR imaging. Radiology 166:111-114; 1988. 7. Bies, J.R.; Ellis, J.H.; Kopecky, K.K.; Sutton, G.P.; Klatte, E.C.; Stehman, F.B.; Ehrlich, C.E. Assessment of primary gynecologic malignancies: Comparison of 0.15-T resistive MRI with CT. AJR 143:1249-1257; 1984. J.L.; Balfe, D.M.; Lee, J.K.T.; Gersell, 8. Worthington, D.J.; Heiken, J.P.; Ling, D.; Glazer, H.S.; Jacobs, A.J.; Kao, M.S.; McGlennan, B.L. Uterine neoplasms: MR imaging. Radiology 159:725-730; 1986. 9. McCarthy, S.; Tauber, C.; Gore, J. Female pelvic anatomy: MR assessment of variations during the menstrual cycle and with use of oral contraceptives. Radiology 160: 119-123; 1986. 10. Demas, B.E.; Hricak, H.; Jaffe, R.B. Uterine MR imaging: Effects of hormonal stimulation. Radiology 159: 123-126; 1986. D.I.; Gussman, D.; Kressel, 11. Mintz, M.C.; Thickman, H.Y. MR evaluation of uterine anomalies. AJR 148: 287-290; 1987. K.; Itoh, K.; Fujisawa, I.; 12. Togashi, K.; Nishimura, Asato, R.; Nakano, Y.; ltoh, H.; Torizuka, K.; Ozasa, H.; Mori, T. Uterine cervical cancer: Assessment with high-field MR imaging. Radiology 160:431-435; 1986. 13. Hricak, H.; Lacey, C.G.; Sandles, L.C.; Chang, Y.C.F.; Winkler, M.L.; Stern, J.L. Invasive cervical carcinoma: Comparison of MR imaging and surgical findings. RQdiology 166:623-631; 1988. 14. Hricak, H.; Stern, J.L.; Fisher, M.R.; Shapeero, L.C.; Winkler, M.L.; Lacey, C.G. Endometrial carcinoma staging by MR imaging. Radiology 162:297-305; 1987. 15. Lotz, H.; Ekelund, L.; Hietala, S.O.; Wickman, G. Low field (0.02 T) MR imaging of the whole body. J. Comput. Assist. Tomogr. 12(6):1006-1013; 1988. R.L. Tl relax16. Moore, F.A.; Todd, L.E.; DeJimenez, ation times in vivo of human normal pelvic tissues and gynecologic malignancies. Physiol. Chem. Phys. Med. NMR. 15(1):23-25; 1983. D.; Hricak, H.; Higgins, 17. Schmidt, H.C.; Tscholakoff, C.B. MR image contrast and relaxation times of solid tumors in the chest, abdomen, and pelvis. J. Comput. Assist. Tomogr. 9(4):738-748; 1985. 18. Fruchter, R.G.; Goldsmith, M.; Boyce, J.G.; Nicastri, A.D.; Koutcher, J.; Damadian, R. Nuclear magnetic resonance properties of gynecological tissues. Gynecol. Oncoi. 6:243-255; 1978. S.; Shapiro, B.; 19. McCarthy, S.; Scott, G.; Majumdar, Thompson, S.; Langer, R.; Gore, J. Uterine junctional zone: MR study of water content and relaxation properties. Radiology 171:241-243; 1989. P.A.; Foster, T.H.; Argersinger, R.E.; 20. Bottomley, Pfeifer, L.M. A review of normal tissue hydrogen NMR relaxation times and relaxation mechanisms from I- 100
Ultra low MR imaging
21.
22.
23.
24.
of uterus 0 M. VARPULA ET AL.
MHz: Dependence on tissue type, NMR frequency, temperature, species, excision, and age. Med. Phys. 1 l(4): 425-448; 1984. Johnson, G.A.; He&ens, R.J.; Brown, M.A. Tissue relaxation time: In vivo field dependence. Radiology 156: 805-810; 1985. Adamski, J.; Olszewski, K.J.; Bucko, J. A study of the freezing of water in human uterine muscle by proton magnetic resonance. Magn. Reson. Imaging 3:125-130; 1985. Bakker, C.J.G.; Vried, J. Multi-exponential water proton spin-lattice relaxation in biological tissues and its implications for quantitative NMR imaging. Phys. Med. Biol. 29:509-518; 1984. Hazlewood, C.F.; Chang, D.C.; Nichols, B.L.; Woessner, D.E. Nuclear magnetic resonance transverse relaxation times of water protons in skeletal muscle. Biophys. J. 14:583-606; 1974.
205
25. Le Rumeur, E.; De Certaines, J.; Toulouse, P.; Rochcongar, P. Water phases in rat striated muscles as determined by T2 proton NMR relaxation times. Mugn. Reson. Imaging 5:267-272; 1987. 26. Olszewski, K.J.; Adamski, J.; Bucko, J.; Pislewski, N.; Fischer, Z. NMR study of the malignancy index and water content in a tumor uterus muscle. Acta Physicu Polonicu A58(6):859-864; 1980. 27. Duberg, A.C.; McCarthy, S.; Lange, R.C. Determination of optimal contrast within the normal and pathologic uterus via synthetic MRI. Mugn. Reson. Zmuging 5(3):1; 1987. 28. Komu, M.; Alanen, A.; Maattanen, H.; Kormano, M. Method dependence of proton spin-lattice relaxation analysis in biological tissues. Actu Rndiol. 30(1):97-100; 1989.
Imaging, Vol. 10, pp. 195-205. All rights reserved.
1992 Copyright
0
0730-725X192 $5.00 + .oO 1992 Pergamon PressLtd.
l Original Contribution
MAGNETIC RESONANCE IMAGING OF THE UTERUS AT AN ULTRA LOW (0.02 T) MAGNETIC FIELD M. Departments
VARPULA,*
M.
KOMU*
of *Diagnostic Radiology and tPathology
AND P. KLEMIT
of the University Central Hospital of Turku, Finland
In vivo pelvic imaging of 39 women and in vitro relaxation time measurements of four uterine specimens were performed using an ultra low field (0.02 T) MRI unit. Average T1 times measured in vitro at 37°C for the myometrium and endometrium were 206 ms (SD 47 ms) and 389 ms (SD 21 ms), respectively. Corresponding T, times were 95 ms (SD 20 ms) and 167 ms (SD 13 ms). The proton relaxation of almost all myometrial specimens proved to be biexponential, but of all endometrial specimens was monoexponential. Contrast measurements between endometrium versus myometrium and myometrium versus the junctional zone were performed after imaging 18 volunteer women using different pulse sequence parameters. Normal uterine structures were optimally demonstrated by SE 700/70. Relatively short repetition times could be used, because spin-lattice relaxation times were short at the low magnetic field. Consequently, the short repetition times allowed averaging of four excitations to create adequate images within an acceptable scanning time. In addition to Tz-weighted images a T,-weighted inversion recovery sequence with a short inversion time of 50 ms (IR 1000/50/40) adequately differentiated the three uterine zones. Although pathologic lesions of the uterus including leiomyomas, anomalies and carcinomas were well demonstrated, especially with the Tz-weighted spin echo pulse sequence, further investigations are needed to evaluate the optimal technique for ultra low field MR imaging of uterine tumors.
Keywords: Uterus; MR studies; Uterus, relaxation times; Uterine neoplasms,
MR studies.
to image the uterus, and to find optimal rameters for that purpose.
INTRODUCTION
Magnetic resonance imaging (MRI) has been reported to have a great potential for the examination of the normal and pathologic uterus. l-l4 The uterine corpus, cervix and vagina and their relationships to the bladder and rectum are clearly delineated by sagittal images. Cyclic changes and other hormonal effects can be demonstrated.‘*” Many studies have suggested that MR imaging is accurate in the detection of uterand anomalies. I1 ine leiomyomas, 2*4s adenomyosis6 The extent of the growth of malignant tumors is reported to be easy to perceive.7~8~12-‘4 A wide range of magnetic fields, commonly 0.15 1.5 T, has been applied. Only in a few experiments is imaging in ultra low field strength (ULF) (0.02-0.045 T) reported. ‘%I6 The aim of this investigation was to evaluate the capabilities of an ultra low magnetic field unit (0.02 T)
MATERIALS
AND
imaging
pa-
METHODS
In Vitro Measurements To help define the optimal pulse sequences for imaging the uterus, both spin-lattice (Tl) and spin-spin (T2) relaxation times were measured in vitro. Myometrial and endometrial specimens were excised from four uteri immediately after removal of the organ. The operations were performed for leiomyomas or bleeding problems. The patient’s ages ranged from 41 to 60 years. One of them had used hormonal substitution (estradiol) just before surgery. The volume of each myometrial specimen was about one cubic centimeter. The endometrial specimens were smaller, one of which was too small for measurements. The specimens were
RECEIVED 6/l l/91; ACCEPTED 8/21/91. Financial assistance from the Cancer Society of Southwestern Finland and the Instrumentarium Corp., Helsinki, Finland, is gratefully acknowledged.
Address all correspondence to Matti Varpula, M.D., Dept. of Diagnostic Radiology, University Central Hospital, 20520 Turku, Finland. 195
196
Magnetic Resonance Imaging 0 Volume 10, Number
immediately sealed in glass tubes with rubber stoppers and kept for 3 to 6 h (mean 4.5 h) at +6”C until measured. After relaxation time measurements all specimens were examined histologically. Serum estradiol levels of the patients were also recorded on the day of operation. The relaxation time measurements were performed with an ultra low field (ULF, 0.02 T) MRI unit (Acutscan, Instrumentarium Corp., Helsinki, Finland) operating at a frequency of 833 kHz in spectrometer mode, using a special device for in vitro measurements.28 T1 determination was based on an inversion recovery (IR) pulse sequence 180’-TI-90” with repetition time (TR) of 2000 ms. At least 10 different inversion times (TI) from 6 ms to 850 ms were used for each relaxation time measurement. T2 determination was done using a standard spin echo (SE) pulse sequence with at least 10 different echo times (TE) from 15 ms to 325 ms and with TR of 2000 ms. Relaxation times were then calculated by fitting the measured signals to the biexponential relaxation curves. The temperature of the specimens was kept constant (37 f 0.4”C) during measurements with an electronic heater. The weighted averages (Tavg) of the short (T,) and long (7;) components of the relaxation times were calculated based on the equation: l/T,,,
= w”/T, + WI/T,
where ws and w’ represent the corresponding factors of the components of T, and 7i.
(1) weight
Volunteer Examinations Eighteen female volunteers without any gynecologic problems (age 23-54 years, mean 39 years) were examined in order to determine the proper imaging parameters for the uterus. All examinations were performed using a U-shaped close coupled body coil underneath the pelvis with the patient in either the supine or prone position. The imaging matrix was 256 x 256, and the field of view was 268 x 402 mm. The slice thickness was 10 mm. The number of contiguous slices obtained simultaneously varied from one to seven depending on the selected pulse sequence. Sagittal, axial and coronal imaging planes were used. Various spin echo (SE), saturation recovery (SR, gradient echo) and inversion recovery (IR) sequences were used. TR of the spin echo sequences varied from 600 to 1500 ms and TE from 70 to 150 ms. Respectively, TR of the SR sequences varied from 160 to 1000 ms and TE from 30 to 70 ms. TR of the IR sequences was 1000 ms, TI 50 ms or 250 ms, and TE 40 ms. The shortest available TE for the spin echo sequences was
2, 1992
70 ms. Images with short TR and TE were obtained using SR sequences. One to four excitations were averaged in SE and SR sequences with long TR, up to 16 excitations in SR sequences with short TR, and one to two excitations in IR sequences. Signal intensity of the myometrium, endometrium and junctional zone was measured using the cursor to define the region of interest (ROI). The percent contrast in signal intensity of the endometrium versus myometrium and the myometrium versus junctional zone was calculated from the intensity measurements made of images obtained at different pulse sequence parameters. Calculations were based on the equation: % Contrast = y
x 100
(2)
where II and Z2are the intensities of different tissues. Clinical Examinations Twenty-one patients with different gynecologic problems were imaged in order to evaluate the capability of the ULF MRI to detect pathologic conditions (Table 1). Most cases had histopathological confirmation at operation (16 cases) or at curettage (2 cases); one uterine anomaly was verified by hysterosalpingography and two leiomyomas by clinical examination. Hysteroscopy was performed on all the patients with endometrial carcinoma, and computed tomography on all the patients with malignant tumors. RESULTS In Vitro Measurements The spin-lattice (TI) and spin-spin ( T2) relaxation times of the four myometrial and the three endometrial specimens are presented in Table 2. Three of the myometrial specimens had a biexponential spin-lattice relaxation and two of them also had a biexponential spin-spin relaxation. All relaxations of the endometrial specimens were monoexponential. In the cases of the biexponential relaxation the weighted averages of the relaxation times were calculated to help compari-
Table 1. Diagnosis
of 21 patients
imaged
Leiomyoma Uterine anomaly Invasive hydatidiform mole Endometrial carcinoma (adenocarcinoma) Cervical carcinoma (squamous cell carcinoma)
by MRI
Ultra low MR imaging
of uterus 0 M. VARPULA ET AL.
197
Table 2. Spin-lattice (T,) and spin-spin (T2) relaxation times (ms) of the myometrium and endometrium of four uterus specimens measured in vitro at body temperature Myometrium Endometrium Uterus 1
2 3 4 Average SD
Tl short 140 (44%) 28 (13%) 130 (37%) 99 51
478 253 420 238 347 104
Tl long
Tl avg
G short
T, long
T2 avg
Tl
r,
(56%) (87%) (63%) (100%)
232 124 230 238
123 (100%) 85 (100%) 80 (76‘70) 59 (78%)
455 (24%) 221 (22%)
123 85 100 70
400 359 408 -
175 149 177 -
95 20
389 21
167 13
206 47
87 23
338 117
The short and long components of the relaxation times and their calculated of the short and long components of the relaxation time are in parentheses. presented.
Estradiol (pmol/L) 420 50 220 50
weighted averages (avg) are presented. The relative weights (070) The serum estradiol levels on the day of measurement are also
Demonstration of Normal Structures Uterus, cervix and vagina were well demonstrated on T,-weighted images, and the three uterine zones with high-intensity endometrium, intermediate-intensity myometrium and low-intensity junctional zone were clearly seen (Fig. 1A). SE 700/70 proved to be ideal pulse sequence for the T2-weighted images of the uterus. The scanning time (12 min) was not too long for the patient, despite the fact that four excitations were needed for diagnostic images. The contrast between myometrium, junctional zone and endometrium were optimal using SE 70170 (Figs. 2A, 2B). Contrast between endometrium and myometrium was greatest with pulse sequence SE lOOO/lOO, however. Longer TR and TE had no advantages. Shorter TR (500 ms) did not diminish the con-
son with previous in vitro and in vivo measurements (Table 3). Mean T’, relaxation times of the myometrium and endometrium were 206 ms (SD 47 ms) and 389 ms (SD 21 ms), respectively. Corresponding T2 relaxation times were 95 ms (SD 20 ms) and 167 ms (SD 13 ms). The relaxation times of the second myometrial and endometrial specimens and the T2 time of the fourth myometrial specimen were shorter than those of the other specimens. This may have been due to the low serum estradiol levels of the corresponding patients. The relaxation times of the fourth myometrial specimen may have been affected by the hormonal substitution the patient had taken just before surgery. The nature of each specimen was confirmed histopathologically. The three endometrial specimens were in the secretory phase.
Table 3. Relaxation times of the uterus according to the literature and to our measurements Myometrium Magnetic field (Tesla/MHz) In vitro Olszewski et a1.26 Fruchter et al.” McCarthy et al. I9 Our measurements
90 MHz 22.5 MHz 20 MHz 0.02 T/0.833 MHz
In vivo McCarthy et al. I9 Hricak et al.’ Schmidth et al. ” Demas et al. I0 Hamlin et al.’ Moore et al.16
1.5 T/64 MHz 0.35 T 0.35 T 0.35 T 0.15 T 0.045 T
T, (ms)
1365 553 709 206
i f + +
142 42 71 47
650 700 Z!Z150 814 -t 112 579 f 5 -
Endometrium T2 (ms)
70 + 66t67 -t 95 f 63 60 50 55 52
6 10 8 20
T, (ms)
801 836 + 172 389 * 21
k 12
600
+ 10 * 5 + 8 -
1077 -t 115 264 f 2 175
Tz (ms)
95 87 * 23 167 + 13 85 f 25 70 61 * 5 -
198
Magnetic Resonance Imaging 0 Volume 10, Number 2, 1992
(B)
03 Fig. 1. Normal uterus. (A) 7”-weighted sagittal image (SR 500170). T‘he three uterine zones with high-intensity endometri um, intei rmediate-intensity myometrium and low-intensity junctional zone are clearly demonstrated. The cervix with low-inter isity cervical canal are well seen. (B) T,-weighted axial image of the normal uterus (SE 1000/70). (C) The stro: ma and high-intensity con1:rast between uterine zones dimihishes with a shorter TE (SR 500, ‘30). (p) When both TR and TE are short (SR 160/ 30), no zonal contrast can be seen.
trast markedly. The difference between intensities of the myometrium and junctional zone was greatest with an echo time of 125-150 ms. TR had no real effect on the contrast between the myometrium and the junctional zone. With TR under 1000 ms the myometrium had a higher signal intensity than urine but with TR 1500 ms the signal from urine was more intense than that of myometrium. With echo times over 100 ms the signal from the uterus diminished and with echo times shorter than 70 ms the contrast between the uterine zones diminished. Still, on SR 500/30 images the endometrium could be distinguished from the myometrium (Fig. 1C). When TR was shortened to 160 ms,
this contrast
also disappeared and the uterus was seen as a homogeneous organ (Fig. 1D). On T,-weighted images obtained with inversion recovery sequence the three uterine zones were seen when the inversion time was 50 ms (Figs. 3A, 3B; Fig. 4A). Contrary to the T,-weighted image the endometrium had a very low signal intensity, the junctional zone was most intense and the myometrium had a slightly lower intensity than the junctional zone. If the inversion time was 250 ms the myometrium and junctional zone were not usually discernible from each other and only a small area of endometrium could be seen (Fig. 4B). The uterus was best demonstrated on sagittal im-
Ultra low MR imaging of uterus 0 M. %
VARPULA
ET AL.
199
Contrast
60
Pulse Sequence
Parameters
(TRITE)
(4 % Contrast
60
30
0 160130
250/30
SOOf
30
700/70
ages, but was also well seen on the axial plane (Fig. 1B). On 7”-weighted images the uterine cervix had a low signal intensity with a central high-intensity stripe representing the endocervical canal (Fig. 1A). On images
1000170
with short TR and TE (SR 160/30) the cervix had a homogeneous appearance with an intermediate signal intensity. The slice thickness was 10 mm, which is relatively
200
Magnetic Resonance Imaging 0 Volume 10, Number 2, 1992 % Contrast
’
-50
I
I
1000/60/40
1000/250/40
Pulse Sequence
Parameters (A)
(TRAFVTE)
9%Contrast
-40
-50
-
I
I
1000150/40
10001250/40
Pulsb’~~~quen~e
ParaFeters
(TR/IR/f
E)
0% Fig. 3. (A) The signal intensity of endometrium relative to myometrium at two different inversion recovery pulse sequences (see text).
thick for imaging of uterus, so interleaved scans at 5-mm increments were used. Because a U-shaped surface coil was used, the signal intensity decreased considerably with increasing distance from the coil. The
(070)and (B) myometrium relative to junctional
zone
supine position of the patient was preferred, although the signal intensity of the anteflexed uterine corpus was higher in the prone p,Qsition. Low signal intensity structures such as the vagina and the cervix were bet-
201
Ultra low MR imaging of uterus 0 M. VARPULA ET AL.
(A)
@I
Fig. 4. Sagittal images of the normal uterus at two different inversion recovery sequences. (A) IR lC00/50/40. All three uterine zones are well seen. In contrast to the T,-weighted image (Fig. 1A) the junctional zone has the highest and the endometrium the lowest signal intensity. (B) IR 1000/250/40. The junctional zone cannot be discerned. A narrow low-intensity endometrial stripe can be seen.
ter seen when the coil was underneath since the signal intensity sufficiently high.
of the uterine
the buttocks, corpus was still
Pathologic Findings As well, the normal anatomy
uterine pathology was also well demonstrated. Five patients had a total of six uterine leiomyomas. Their diameter varied from 10 to 60 mm. The signal intensity of the leiomyomas was lower than that of the myometrium on T2-weighted images (Fig. 5). A septate uterus was well demonstrated on Tz-weighted axial image (Fig. 6).
Fig. 5. Sagittal T,-weighted image (SR 500/70). An intramural leiomyoma (arrow) has a signal intensity lower than that of the myometrium.
Two young women (aged 26 and 30 years old) with invasive hydatidiform moles were imaged. The diagnosis was verified by curettage. In the one case a highintensity focus was found within the myometrium on T2-weighted images (Fig. 7). The patient was not operated on but ultrasonically a sponge-like area typical of an invasive mole was seen in the myometrium. In the other case the entire uterine body was inhomogeneous with relatively high signal intensity, corresponding to the pathologic vasculature demonstrated by angiography.
Fig. 6. 7”-weighted (SE 1000/70) axial image of a uterine anomaly (septate uterus). Two separate uterine cavities (asterisks) and two separate cervical canals (small arrows) are well seen in the undivided uterine body (arrowheads).
202
Magnetic Resonance Imaging 0 Volume 10, Number
2, 1992
lated endometrium of a high signal intensity (Fig. 8A). In two cases also small areas of low signal intensity were seen in the dilated high-intensity endometrium (Fig. 8B). At hysteroscopy they proved to be exophytic growth of carcinoma. If carcinoma was entirely confined to the endometrium (three cases), the junctional zone was well demarcated (Fig. 8A). However, in one of the two cases with superficial myometrial involvement the junctional zone was also intact. If carcinoma involved at least half of the myometrial thickness (four cases) the junctional zone was not visible, and the border between the endometrium and myometrium was indistinct (Fig. 8B). In all cases the uterus was well demarcated on MR images and the tumor growth outside the uterus was easily excluded. MRI findings of two cervical carcinomas were normal. One of them was in situ type (IA) and in the other case the carcinoma was apparently removed by conisation before imaging. Two stage IB carcinomas were seen on T2-weighted scans as central high-intensity tumors surrounded by a normal low signal intensity cervical stroma (Fig. 9). Parametrial tumor growth was therefore easily excluded.
Nine patients with endometrial carcinoma and four patients with cervical carcinoma were imaged. All of them had stage I disease, and they were operated on after radiation therapy. Histopathologically the endometrial carcinomas were adenocarcinomas and the cervical carcinomas were squamous cell carcinomas. The most common finding of the endometrial carcinoma on T2-weighted images (seven cases) was a di-
ation time measurements of the uterus are described in
(4
(W
DISCUSSION SOme in vivo1,5,10,16,17,19 and in vitro’8,19,26 relax_
Fig. 8. (A) &weighted sagittal image (SE 700/70) of a Stage IA endometrial carcinoma. The tumor was confined histopathologically to the endometrium. The high-intensity endometrium is slightly dilated. Because the low-intensity junctional zone is intact, myometrial invasion of carcinoma can be excluded by MRI. (B) T2-weighted sagittal image (SE 700/70) of a Stage IA endometrial carcinoma. This tumor invaded histopathologically halfway through the myometrium. Exophytic carcinoma tissue is seen as areas with low signal intensity in dilated endometrium with high signal intensity. The junctional zone is not detectable and the tumor can be interpreted as invading to the middle of the myometrium. The uterus is well circumscribed excluding the growth of carcinoma outside the uterus.
Ultra low MR imaging of uterus 0 M. VARPULA ET AL.
(A)
203
09
Fig. 9. Sagittal (A) and axial (B) T,-weighted images (SE 700170) of a Stage IB cervical carcinoma confining to the cervix. The tumor with high signal intensity (asterisk) is entirely surrounded by uninvolved cervical stroma with low signal intensity. Tumor growth outside the cervix is easily excluded by MRI. (u: involuted uterine body)
the literature (Table 3). Most of them have been done at magnetic fields from 0.15 to 1.5 T. Only one author has measured T, of the myometrium.and the cervix at an ultra low magnetic field (0.045 T).“j As expected, the spin-lattice relaxation times (T,) at our ultra low magnetic field proved to be shorter than at higher field strengths. This field dependence of spin-lattice relaxation time is well known.20*21 The spin-spin relaxation time does not depend as much on field strength. The T2 relaxation times in our material, especially those of endometrial specimens, were longer than those described at higher magnetic fields. Individual variations may have been responsible for this finding. The relaxation times of the myometrium and endometrium seem to correlate with serum estradiol levels, although our material is not sufficiently large to confirm such a correlation. Both spin-lattice and spin-spin relaxations of almost all myometrial specimens were biexponential. Multiexponential T, relaxation of the myometrium has been reported by Adamski et al.,** but the corresponding phenomenon of T2 relaxation of the myometrium, to our knowledge, has not been described previously. Multiexponential relaxation seems to be typical for both smooth and striated muscles. The multiexponentiality of both longitudinal and transverse relaxations of skeletal muscles is well known.23-25 Different water phases of tissue are considered to be responsible for this phenomenon.24*25 Several studies have suggested that T2-weighted images are most useful for imaging the uterus and its pathology.‘-‘4 Relatively short spin-lattice relaxation times at our ultra low magnetic field made it possible to use shorter repetition times (500-1000 ms) than at
higher fields (commonly 2000 ms or longer). Our results suggest that the contrast between different uterine zones will not be suppressed by a shorter TR. One report concerning high magnetic field strength supports this result.27 Because the spatial resolution of our 0.02 T unit is lower than that of higher fields, averaging of at least four excitations was needed for an acceptable signal to noise ratio. If longer (1000-1500 ms) repetition times had to be used, the scanning time would have become impractical (17-25 min), and motion of the uterus, bowel or patient would have muddled the images. Using SE 700/70 with averaging of four excitations, four consecutive 10 mm thick slices were available during a scanning time of 12 min. Because the slice thickness was 10 mm, interleaved scans at 5-mm increments were used. If the uterus was imaged in both sagittal and axial directions and some images with short TR and TE were obtained, the entire examination time was about 1 h. The scanning time could be further reduced by using a shorter repetition time. The zonal contrast of the uterus would still be sufficient for some purposes, that is, for imaging of uterine anomalies. In the present material the normal uterus with its zonal structures was well demonstrated both on Tland T2-weighted images. On Tl-weighted inversion recovery (IR) images the zones were well seen when short inversion time (50 ms) was used. On images with short TR and TE (SR 160/30) the zones were not discernible. The T,contrast of the images obtained with the latter kind of sequences is not as T,-specific as that of inversion recovery sequences. Pathologic changes of the uterus, such as leiomyomas, anomalies and malignant tumors, were well dem-
204
Magnetic Resonance Imaging 0 Volume 10, Number 2, 1992
onstrated by our ultra low field strength magnetic unit. T2-weighted SE images were most useful. SE 700/70 is a proper pulse sequence as well for imaging of normal structures of uterus as for imaging of uterine pathology. Endometrial carcinoma has no specific MRI findings. Differentiation of adenomatous hyperplasia and blood clots from carcinoma is reported to be impossible. l4 Cervical carcinoma has higher signal intensity than that of the normal cervical tissue on T2-weighted images, but if tumor is small it is not discernible even at high magnetic fields. ‘2,‘3 The value of MRI is in the staging of the carcinomas. Vaginal and parametrial infiltration of the cervical carcinoma and myometrial invasion by endometrial carcinoma have been evaluated relatively precisely on images obtained using high magnetic field units. 73,12-14 Our material is very restricted and more advanced conclusions are not justified. However, our results indicate that it is possible to evaluate the myometrial involvement of the endometrial carcinoma and the confinement of the cervical carcinoma to the uterine cervix using MRI at 0.02 T. The benefits of an ultra low magnetic field imaging of the pelvis are its low cost, greater convenience, and relatively short scanning times. Image quality is fairly good and relatively small pathologic processes are discernible. Although the spatial resolution is inferior to that of higher field strengths, the diagnostic information is satisfactory for clinical practice. Because the surface coil is U-shaped the signal intensity diminishes considerably with increasing distance from the coil. This may be a problem, particularly with obese patients. Developing a ring-like surface coil would resolve this problem. Further investigations are also needed to evaluate other problems, especially those.concerning the staging of carcinomas. The results of the present material are promising. REFERENCES
’
Hricak, H.; Alpers, C.; Crooks, L.E.; Sheldon, P.E. Magnetic resonance imaging of the female pelvis: Initial experience. AJR 141:1119-1128; 1983. Hricak, H. MRI of the female pelvis: A review. AJR 146:1115-1122; 1986. Lukas, P.; Schrock, R.; Rupp, N.; Reiser, M.; Allgayer, B.; Fenerbach, St.; Hofrichter, A.; Heller, H.J. Die MR-tomographie bei gynakologischen erkrankungen im kleinen becken. Fortschr. RGntgenstr. 144(2):159-165; 1986. ‘. Butler, H.; Bryan, P.J.; LiPuma, J.P.; Cohen, A.M.; El Yousef, S.; Andriole, J.G.; Lieberman, J. Magnetic resonance imaging of the abnormal female pelvis. AJR 143:1259-1266; 1984.
5. Hamlin, D.J.; Pettersson, H.; Fitzsimmons, J.; Morgan, S. MR imaging of uterine leiomyomas and their complications. J. Comput. Assist. Tomogr. 9(5):902-907; 1985. K.; Itoh, K.; Fujisawa, I.; 6. Togashi, K.; Nishimura, Noma, S.; Kanaoka, M.; Nakamo, Y.; Itoh, H.; Ozasa, H.; Fujii, S.; Mori, T. Adenomyosis: Diagnosis with MR imaging. Radiology 166:111-114; 1988. 7. Bies, J.R.; Ellis, J.H.; Kopecky, K.K.; Sutton, G.P.; Klatte, E.C.; Stehman, F.B.; Ehrlich, C.E. Assessment of primary gynecologic malignancies: Comparison of 0.15-T resistive MRI with CT. AJR 143:1249-1257; 1984. J.L.; Balfe, D.M.; Lee, J.K.T.; Gersell, 8. Worthington, D.J.; Heiken, J.P.; Ling, D.; Glazer, H.S.; Jacobs, A.J.; Kao, M.S.; McGlennan, B.L. Uterine neoplasms: MR imaging. Radiology 159:725-730; 1986. 9. McCarthy, S.; Tauber, C.; Gore, J. Female pelvic anatomy: MR assessment of variations during the menstrual cycle and with use of oral contraceptives. Radiology 160: 119-123; 1986. 10. Demas, B.E.; Hricak, H.; Jaffe, R.B. Uterine MR imaging: Effects of hormonal stimulation. Radiology 159: 123-126; 1986. D.I.; Gussman, D.; Kressel, 11. Mintz, M.C.; Thickman, H.Y. MR evaluation of uterine anomalies. AJR 148: 287-290; 1987. K.; Itoh, K.; Fujisawa, I.; 12. Togashi, K.; Nishimura, Asato, R.; Nakano, Y.; ltoh, H.; Torizuka, K.; Ozasa, H.; Mori, T. Uterine cervical cancer: Assessment with high-field MR imaging. Radiology 160:431-435; 1986. 13. Hricak, H.; Lacey, C.G.; Sandles, L.C.; Chang, Y.C.F.; Winkler, M.L.; Stern, J.L. Invasive cervical carcinoma: Comparison of MR imaging and surgical findings. RQdiology 166:623-631; 1988. 14. Hricak, H.; Stern, J.L.; Fisher, M.R.; Shapeero, L.C.; Winkler, M.L.; Lacey, C.G. Endometrial carcinoma staging by MR imaging. Radiology 162:297-305; 1987. 15. Lotz, H.; Ekelund, L.; Hietala, S.O.; Wickman, G. Low field (0.02 T) MR imaging of the whole body. J. Comput. Assist. Tomogr. 12(6):1006-1013; 1988. R.L. Tl relax16. Moore, F.A.; Todd, L.E.; DeJimenez, ation times in vivo of human normal pelvic tissues and gynecologic malignancies. Physiol. Chem. Phys. Med. NMR. 15(1):23-25; 1983. D.; Hricak, H.; Higgins, 17. Schmidt, H.C.; Tscholakoff, C.B. MR image contrast and relaxation times of solid tumors in the chest, abdomen, and pelvis. J. Comput. Assist. Tomogr. 9(4):738-748; 1985. 18. Fruchter, R.G.; Goldsmith, M.; Boyce, J.G.; Nicastri, A.D.; Koutcher, J.; Damadian, R. Nuclear magnetic resonance properties of gynecological tissues. Gynecol. Oncoi. 6:243-255; 1978. S.; Shapiro, B.; 19. McCarthy, S.; Scott, G.; Majumdar, Thompson, S.; Langer, R.; Gore, J. Uterine junctional zone: MR study of water content and relaxation properties. Radiology 171:241-243; 1989. P.A.; Foster, T.H.; Argersinger, R.E.; 20. Bottomley, Pfeifer, L.M. A review of normal tissue hydrogen NMR relaxation times and relaxation mechanisms from I- 100
Ultra low MR imaging
21.
22.
23.
24.
of uterus 0 M. VARPULA ET AL.
MHz: Dependence on tissue type, NMR frequency, temperature, species, excision, and age. Med. Phys. 1 l(4): 425-448; 1984. Johnson, G.A.; He&ens, R.J.; Brown, M.A. Tissue relaxation time: In vivo field dependence. Radiology 156: 805-810; 1985. Adamski, J.; Olszewski, K.J.; Bucko, J. A study of the freezing of water in human uterine muscle by proton magnetic resonance. Magn. Reson. Imaging 3:125-130; 1985. Bakker, C.J.G.; Vried, J. Multi-exponential water proton spin-lattice relaxation in biological tissues and its implications for quantitative NMR imaging. Phys. Med. Biol. 29:509-518; 1984. Hazlewood, C.F.; Chang, D.C.; Nichols, B.L.; Woessner, D.E. Nuclear magnetic resonance transverse relaxation times of water protons in skeletal muscle. Biophys. J. 14:583-606; 1974.
205
25. Le Rumeur, E.; De Certaines, J.; Toulouse, P.; Rochcongar, P. Water phases in rat striated muscles as determined by T2 proton NMR relaxation times. Mugn. Reson. Imaging 5:267-272; 1987. 26. Olszewski, K.J.; Adamski, J.; Bucko, J.; Pislewski, N.; Fischer, Z. NMR study of the malignancy index and water content in a tumor uterus muscle. Acta Physicu Polonicu A58(6):859-864; 1980. 27. Duberg, A.C.; McCarthy, S.; Lange, R.C. Determination of optimal contrast within the normal and pathologic uterus via synthetic MRI. Mugn. Reson. Zmuging 5(3):1; 1987. 28. Komu, M.; Alanen, A.; Maattanen, H.; Kormano, M. Method dependence of proton spin-lattice relaxation analysis in biological tissues. Actu Rndiol. 30(1):97-100; 1989.