Relaxation time changes of the uterus during the menstrual cycle: correlation with hormonal status

European Journal of Radiology.

90

0

16( 1993) 90-94

1993 Elsevier Scientific Publishers Ireland Ltd. All rights reserved. 0720-048X/93/$06.00

EURRAD 00330

Relaxation time changes of the uterus during the menstrual cycle: correlation with hormonal status M. Varpula”, M. Komu a and K. Irjalab aDepartment of Diagnostic Radiology and b Central Laboratory of the University Central Hospital of Turku. Finland

(Received 29 March 1992; accepted after revision 23 July 1992)

Key words: Uterus, MR; Magnetic resonance, uterus

Abstract Sk women volunteers underwent pelvic MR imaging at 0.02 T four times during their menstrual cycle. The Tl and T2 relaxation times of the myometrium and endometrium were measured and correlated with the serum estradiol and progesterone levels. The magnitude of the relaxation times were highly individual but the pattern of their variation during the menstrual cycle was similar. The relaxation times were shortest at the beginning and end of the cycle. The most rapid increase occurred during the proliferative phase, followed by little or no increase through to the middle of the secretory phase. The Tl and T2 times of the endometrium correlated directly with the serum estradiol levels during the entire menstrual cycle (r = 0.5, P = 0.02) and the T2 times of the endometrium with the serum progesterone levels during the secretory phase (r = 0.6, P = 0.05). The correlation between the relaxation times of the myometrium and the serum hormonal levels was poor. The results indicate that the relaxation times of the myometrium and endometrium vary during the menstrual cycle reflecting the serum hormonal status. MR imaging of the uterus with relaxation time measurements may be clinically useful to examine the menstrual cycle and its pathology.

Introduction

Material and Methods

The cyclic changes in endometrial histology and the hormonal variations during the menstrual cycle are well documented [ 1,2]. Morphological changes of the uterus during the menstrual cycle have been previously reported with both magnetic resonance [3-71 and ultrasound imaging [ 7,8] methods. The relaxation times Tl and T2 of myometrium and endometrium have both been reported to be longer during the secretory phase than during the proliferative phase [4], but more detailed sequential changes of these parameters during the menstrual cycle have not been previously described to our knowledge. This work was undertaken to characterize the changes in the relaxation times Tl and T2 of the myometrium and endometrium during the normal menstrual cycle in relation to the serum hormonal status.

Six women volunteers of reproductive age (27 to 42 years, mean 37 years) without gynecologic problems underwent MR imaging four times during their menstrual cycle. None of them used exogenous steroid contraceptives. One of them had used intradermal contraceptive capsules up to eight months prior to this study. One had an intrauterine contraceptive device. All of the women had a history of relatively regular menstrual cycle. None were pregnant or were attempting to become pregnant during the study. Four of them had had one or two term deliveries. Informed consent was obtained from each women after the nature of the procedures had been fully explained. Immediately after MR imaging a blood sample was taken for determination of estradiol and progesterone levels. The separated sera were stored at -20°C until analyzed. Serum estradiol was measured with an RIA kit from Baxter Dade (Dudingen, France) and serum progesterone with an RIA kit from Farmos Diagnostica

Correspondence to: Matti Varpula, Dept. of Diagnostic University Central Hospital, 20520 Turku, Finland

Radiology,

(Turku, Finland). The day of ovulation was determined by following the basal body temperature. MR imaging was performed on an ultra low field (0.02 T) magnetic resonance imaging device (Acutscan@, Instrumentarium Corp., Helsinki, Finland). All examinations were performed using a U-shaped close coupled body coil underneath the pelvis. The imaging matrix was 256 x 256 and the field of view was 268 x 402 mm. Sagittal contiguous sections of 10 mm thickness were obtained. In the determination of Tl times two identical images of the uterus were obtained using inversion recovery (IR) sequences with two different inversion times (IR 1000/50/40 and IR 1000/250/40). Two excitations were averaged. T2 times were measured by a dual echo sequence with a repetition time of 500 ms, a gradient echo time of 30 ms and a spin echo time of 125 ms. Four excitations were averaged in this sequence. The relaxation times of the myometrium and endometrium were determined by measuring the signal intensities of the tissue with a ROI (region of interest) method from image pairs obtained with different pulse sequence parameters, and fitting the measured signals to the relaxation curves. A rectangular region of interest was drawn in the most homogeneous areas of the myometrium and endometrium, avoiding the junctional zone. Because the junctional zone was very narrow in many cases, its relaxation times were not determined. The most midsagittal sections of the uterus were used in order to avoid the partial volume effect. 6-12 (mean 9) regions of interest were used in the relaxation time determination of each myometrium and 2-7 (mean 4) regions of interest in the relaxation time determination of each endometrium (Fig. 1). The regions of interest of the myometrium consisted of 6-36 (mean 17) pixels and the regions of interest of the endometrium of 4-20 (mean 10) pixels. Because many separate regions of interest were drawn on the same tissue, many separate relaxation times of the same tissue were calculated. The mean of these separate relaxation times was regarded as the relaxation time of the given tissue. The standard deviation (SD) of these means was generally l- 10%. The SD was 13-14% in only three measurements. In these cases the image quality was suboptimal but the measurements were acceptable. In order to better demonstrate the changes of relaxation times and serum hormonal levels the menstrual cycle was divided into five sections according to the common hormonal composition: 1) the first half of the proliferative phase, 2) the latter half of the proliferative phase, 3) the beginning of the secretory phase, 4) the middle of the secretory phase during estradiol and

Fig. 1. A sag&al T2-weighted image of a normal uterus. Because one rectangular ROI covered only a small part of the myometrium or endometrium, many separate ROI areas were drawn on the same tissue for the determination of its signal intensity.

progesterone peak, and 5) the end of the secretory phase. In the statistical analysis non-parametric Wilcoxon-Pratt and Spearman tests were used. Two T 1 time measurements of the endometrium were inaccurate because of a narrow endometrium and poor image quality (volunteer III, section 2, and volunteer IV, section 3), and were excluded from the material. Two volunteers were not imaged during the first sections of the menstrual cycle in question, so the relaxation times of the first sections of the next cycle were used (volunteers I and III). Results Desctiption of the menstrual cycles

The lengths of the cycles varied from 24 to 3 1 days. Three women had a clear rise in basal body temperature after ovulation. Three women underwent no clear temperature changes. The 14th day of the cycle was selected as the day of the ovulation in these cases. The serum progesterone levels of all women during the secretory phase (over 15 nmol/l) demonstrated that ovulation had occurred. Relaxation time changes The relaxation times of the myometrium

and endometrium and the serum hormonal levels in the six volunteers during their menstrual cycles are presented

92

in Table 1. There are parameters missing from each menstrual cycle because the volunteers were imaged only four times. The magnitude of the relaxation times and the serum hormonal levels was considerably individual. There were no statistically signifkant changes during the menstrual cycle (Wilcoxon-Pratt test), but certain trends were evident. The intra-individual patterns of the Tl and T2 time variations of the myometrium and endometrium during the menstrual cycle were almost equal. The changes in the relaxation times of the myometrium were relatively small during the menstrual cycle, compared with the changes in the relaxation times of the endometrium. The relaxation times were generally shortest at the beTABLE

ginning of the menstrual cycle. There values gradually lengthened until they again decreased at the end of the cycle. The most rapid increase occurred during the proliferative phase. The relaxation times of the endometrium seemed to shorten towards the end of the cycle more quickly than the relaxation times of the myometrium. The pattern of the changes in the relaxation times of the endometrium was similar to the pattern of the changes in the serum estradiol levels (Fig 2.). The relaxation times decreased slightly following ovulation but became longer again during the middle of the secretory phase. The relaxation time changes were not as rapid as those of the serum hormonal levels. Relaxation

1

The relaxation times of the myometrium and the endometrium and the serum estradiol and progesterone levels in the six volunteers during their menstrual cycles. The cycles are divided into five sections (see text). Negative cycle days are the days before ovulation and positive days the days after ovulation. Day 0 is the day of ovulation Volunteer

I

II

III

IV

V

VI

SEstradiol

Day of cycle

Tl

1 2 3 4 5

+ 15* -

215 -

91

251

62

150 -

1

0 +4 +ll

249 239 256

85 98 105

459 318 246

111 110 99

180 630 140

11 39 2

-9 -3

217 243

83 91 -

227 553 -

82 147

120 500 -

1 1

+4 + 10

224 218

93 109

324 328

105 135

590 350

21 25

74 208

410 1700

1 1

1 2 3 4 5

Endometrium T2

Tl

sProgesterone

T2

1 2 3 4 5

+ 18** -2 -

235 248

85 84

170

+5 + 12

250 252

78 91

553 444

179 143

1100 490

54 9

1 2 3 4 5

- 12 -4 +6 + 10 -

202 229 225 245

111 108 120 132

225 457

89 172 137 154 -

110 130 240 330

1 1 12 30

1 2 3 4 5 1 2 3 4 5

The fourth day of the next cycle. ** The fifth day of the next cycle. l

Myometrium

Section of cycle

-

-

436

-

-

-

-5 +3 +9 + 16

244 246 248 212

95 105 104 88

291 282 443 201

133 161 198 80

180 190 380 150

1 20 44 12

-9 -1 -

255 248 -

133 118

211 416

92 105 -

100 540 -

4 1 -

+ 11 + 18

231 216

128 68

215 211

84 70

240 160

19 2

93

Tl

TIME

T2

(ms)

TIME

Discussion

(ms)

500

1 60

200 j

100

i

SECTION

OF

THE

CYCLE

Fig. 2. The sequential changes in the relaxation times of the endometrium and in the serum estradiol and progesterone levels during the menstrual cycle. The median values of each parameter in the five menstrual cycle sections (see text) are presented. The changes are not statistically significant. The pattern of the changes in the relaxation times is similar to the pattern of the changes in the serum estradiol levels. (The scale of the serum hormonal levels is not presented.)

times could remain long after the estradiol peak, although the serum estradiol levels were decreased (volunteers I and II), and in some cases the relaxation times were not lengthened, although the serum estradiol levels were incresed (volunteers I, II and IV). Yet, the Tl and T2 times of the endometrium correlated significantly with the serum estradiol levels during the entire menstrual cycle and T2 times with the progesterone levels during the secretory phase (Table 2). The correlation between progesterone and Tl times of the endometrium was poor. The correlation between the serum hormone levels and the myometrial relaxation times was also poor.

TABLE 2 Correlation between the serum hormonal levels and the relaxation times of the uterus during the menstrual cycle (Spearman test, R = correlation coefficient). 24 measurements made on six women. The estradiol levels are correlated with the relaxation times during the entire menstrual cycle but the progesterone levels are correlated only during the secretory phase S-Estradiol (R)

S-Progesterone (secretory phase) (R)

Myometrium Tl time T2 time Endometrium Tl time T2 time

0.3 (P=O.16) - 0.2 (I’= 0.32)

- 0.2 (P = 0.53) - 0.4 (P = 0.25)

0.5 (f = 0.02) 0.5 (P= 0.02)

0.4 (P = 0.20) 0.6 (P = 0.05)

During the hrst (proliferative) half of the menstrual cycle the estradiol levels gradually increase with a substantial rise just before ovulation followed by a precipitous fall. During the second (secretory) half the estradiol levels rise again, falling at the end of the cycle. The progesterone levels are minimal during the proliferative phase and begin to rise after ovulation reaching maximal levels at about the middle of the secretory phase and falling toward the end of the cycle [ 1,2]. The changes in uterine morphology during the menstrual cycle have been demonstrated with MR imaging [ 3-71. The thickness of the endometrium grows rapidly during the proliferative phase but more slowly during the secretory phase, decreasing abruptly at the onset of menstruation. The changes in total myometrial thickness follow the same pattern. The signal intensity of the myometrium and the width of the junctional zone both increase during the secretory phase. An increase in the Tl and T2 times between the early proliferative and the midsecretory phases has been reported [4]. A corresponding increase in the Tl times of red blood cells has also been described [ 91. Our results show that the relaxation times of the myometrium and endometrium do change during the menstrual cycle, corresponding to the changes in uterine morphology and relaxation times described by other authors [ 3-71. The relaxation times of the endometrium correlate with the serum estradiol and progesterone levels but the changes of the relaxation times are not as rapid as those of the serum hormonal levels. The slight, temporary decrease in the relaxation times of the endometrium after ovulation, or the corresponding change in the uterine morphology, have not been reported previously. The small number of subjects is the major limitation of our material. Four observation points during the menstrual cycle were insufficient to demonstrate all the intra-individual changes in both serum hormone levels and the respective relaxation times. Because the hormonal status of the volunteers was not followed daily, the timing of the imaging was less than optimal, and the rapid estradiol changes associated with ovulation were often missed. Two women (volunteers III and VI) were imaged during their preovulatory estradiol peak. One woman (volunteer I) was imaged just after the peak, and two women (volunteers II and IV) just before the peak. Four women (volunteers I-IV) showed shortening in their endometrial relaxation times after ovulation. The in vivo measurement of relaxation times using only two data points has been commonly criticized. There are several sources of error including inhomoge-

94

nities in the static magnetic field and in the applied radio frequency fields, imperfections in slice selection and molecular diffusion processes [ 10,111. Measurement of transverse (T2) relaxation times is especially sensitive to these errors. The selected pulse sequence parameters may seriously affect the results, especially in the case of multiexponential relaxation [ 12,131. In the present study a relatively short repetition time (500 ms) was used in the determination of T2 times to shorten the scanning time and to minimize the problems caused by uterine movement during the imaging. Although the measured Tl and T2 times are only approximations of the true values, all the measurements have been done by the same method and with the same apparatus and should be internally consistent. Since our study consisted of only the events of the normal menstrual cycle, e.g. no anovulatory cycle was included, further investigations are needed to evaluate the usefulness of magnetic resonance imaging in menstrual cycle pathology. Sequential MR imaging during the menstrual cycle may be helpful in the examination of dysmenorrhea, hormonal disturbances and infertility problems. Acknowledgements

Financial assistance from the Science Foundation of Instrumentarium Corp., Helsinki, Finland is gratefully acknowledged.

References 1 Mishell DR, Nakamura RM, Crosignani PG, Stones S, Kharma K, Nakata Y, Thomeycroft IH. Serum gonadotropin and steroid patterns during the normal menstrual cycle. Am J Obstet Gynecol 1971; 111: 60-65.

2 Punnonen R, Nummi S, Yhkorkala 0, Alapiessa U, Karvonen P, Viinikka L. A composite picture of the normal menstrual cycle. Acta Obstet Gynecol Stand 1975; Suppl 51: 64-70. 3 McCarthy S, Tauber C, Gore J. Female pelvic anatomy: MR assessment of variations during the menstrual cycle and with use of oral contraceptives. Radiology 1986; 160: 119-123. Demas BE, Hricak H, JatTe RB. Uterine MR imaging: Effects of hormonal stimulation. Radiology 1986; 159: 123-126. Haynor DR, Mack LA, Soules MR, Shuman WP, Montana MA, Moss AA. Changing appearance of the normal uterus during the menstrual cycle: MR studies. Radiology 1986; 161: 459-462. 6 Janus CL, Wiczyk HP, Laufer N. Magnetic resonance imaging of the menstrual cycle. Magn Reson Imaging 1988; 6: 669-674. I Wiczyk HP, Janus CL, Richards CJ, Graf MJ, Gendal ES, Rabinowitz JG, Laufer N. Comparison of magnetic resonance imaging and ultrasound in evaluating follicular and endometrial development throughout the normal cycle. Fertil Steril 1988; 49: 969-972. 8 Forrest TS, Morteza KE, Muilenburg MI, Bewtra C, Kable WT, Sullivan P. Cyclic endometrial changes: US assessment with histologic correlation. Radiology 1988; 167: 233-237. 9 Rosenthal JS, Strauss A, Minkoff LA, Winston A. Variations in red blood cell proton Tl relaxation times that correspond to menstrual cycle changes. Am J Obstet Gynecol 1985; 153: 812813. 10 Kjaer L, Thomsen C, Hem&en 0, Ring P, Stubgaard M, Pedersen El. Evaluation of relaxation time measurements by magnetic resonance imaging. A phantom study. Acta Radio1 1987; 28: 345-351. 11 Wong STS, Roos MS. Effects of slice selection and diffusion on T2 measurement. Magn Reson Medic 1987; 5: 358-365. 12 Komu M, Alanen A, Maatmnen H, Kormano M. Method dependence of proton spin-lattice relaxation analysis in biologic tissues. Acta Radio1 1989; 30: 97-100. 13 Kjaer L, Hemiksen 0. Comparisonof different pulse sequences for in vivo determination of Tl relaxation times in the human brain. Acta Radio1 1988; 29: 231-236.