To investigate the T2* values within the junctional zone and outer uterine myometrium and their changes during the menstrual cycle, and thus to evaluate their physiologic changes on blood oxygenation level-dependent (BOLD) magnetic resonance (MR) imaging.
Materials and Methods
Single-shot echo-planar imaging (EPI) was used to acquire T2*-weighted images (TR/TE = 1000 msec/23–150 msec) from 15 healthy females with a 1.5-T magnet. T2* values of both junctional zone and outer uterine myometrium were measured within a single breathhold and during three menstrual cycle phases (menstrual, periovulatory, and luteal phase). Signal intensities of uterine myometrium on T2-weighted images were also evaluated.
T2* could successfully be calculated in 13 subjects. T2* values for the junctional zone were significantly lower than those of the outer myometrium at every phase(P < 0.001), and T2* values of both junctional zone (P < 0.05) and outer (P < 0.01). Myometrium in the menstrual phase was significantly lower than those in the other phases. On T2-weighted images, the signal intensity of the junctional zone was significantly lower than outer myometrium in every phase (P < 0.01), but there was no significant difference among menstrual cycle phases in both layers (P > 0.05).
THE UTERINE myometrium in reproductive women consists of layers with two different signal intensities on T2-weighted in magnetic resonance (MR) images: the junctional zone appears as a band of distinct low intensity, and the outer myometrium of intermediate intensity. This morphologic structure of the myometrium exhibits dynamic change according to phase in the menstrual cycle (1). The myometrial thickness is minimum in the menstrual phase, and it gradually increases during the proliferative phase. The contrast between these two layers is most prominent in the proliferative phase, while it becomes indistinct in the luteal and menstrual phases (1).
These morphologic changes of the myometrium on T2-weighted images may be associated with and explained by hemodynamic changes during the menstrual cycle. The relation between uterine blood flow and several uterine factors, such as endometrial thickness and uterine volume, was previously examined using Doppler transvaginal ultrasound (2). For example, the uterine myometrium strongly contracts and leads to decreased local uterine blood flow during the menstrual phase (2, 3). The strong contraction in the uterine myometrium may lead to decreased perfusion or even hypoxia in this muscle tissue.
Recent developments in MR imaging techniques have enabled the noninvasive measurement of tissue blood oxygenation using blood oxygenation level-dependent (BOLD) MR imaging. This technique is based on the principle that deoxyhemoglobin and oxyhemoglobin differ in magnetic properties. This affects the T2* relaxation time of the neighboring water molecules and thus influences the MRI signal in T2*-weighted images (4). The ratio of oxyhemoglobin to deoxyhemoglobin is related to the oxygen saturation (pO2) of blood, and the pO2 of capillary blood is thought to be in equilibrium with surrounding tissue. Therefore, changes of T2* value measured by BOLD MRI can be considered as changes in tissue pO2 (4–6).
This technique was originally developed for functional MR imaging of the human brain (7). Although this technique does not measure blood flow directly, many studies on brain functional MRI have demonstrated a strong correlation between BOLD measurements and a mismatch between blood flow and oxygen uptake (6, 8–10). Thus it is recognized that blood oxygenation within various tissues and organs can be estimated by BOLD measurements. In experimental studies with human or mice, BOLD MR imaging has been widely applied to evaluate perfusion or hypoxia in other body organs, including the heart, kidney, and muscle (5, 11, 12). Because the uterus consists of smooth muscle, measurement of its oxygenation may provide new noninvasive information for determining the causes of infertility or dysmenorrhea.
The purpose of this study is to investigate T2* values of the junctional zone and outer myometrium of the uterus and their changes during the menstrual cycle on BOLD MR imaging, and thus to evaluate physiologic changes of the uterus on BOLD MR imaging.
MATERIALS AND METHODS
The protocol of this study was approved by the Ethical Committee of our institute. Written informed consent was obtained from all participants.
The study was carried out in 15 healthy female volunteers of reproductive age (age range: 21–37 years, mean: 28.6 years). The volunteers were recruited from staff, medical students, and graduate school students at our institute. Exclusion criteria were history of gynecologic disease, abnormality on MR imaging, and use of contraceptive medication or methods at the time of this study. Any volunteers showing some abnormality on a screening MR scan were excluded from further study protocols.
MR study was scheduled at three phases of the menstrual cycle (menstrual, periovulatory, and luteal phases) for each volunteer. The date of ovulation was estimated as either 14 days prior to the anticipated first day of the next cycle or based on the elevation of basal body temperature. MR scans in the periovulatory phase were performed within two days before or after the expected ovulation date. MR images in the luteal phase were obtained approximately in the middle of this phase. MR images in the menstrual phase were obtained within the first two days of menstruation. In each volunteer, estimated cycle dates were confirmed by the onset of next menstruation.
MR Scanning Protocol
MR images were obtained with a 1.5-T magnet unit (Symphony; Siemens Medical Systems, Erlangen, Germany) using a phased-array coil. At the beginning of each MR session, axial half-Fourier acquisition single-shot turbo spin-echo (HASTE) images was obtained for the purpose of defining anatomy. Sagittal fast spin-echo T2-weighted images (TR/TE = 5470 msec/122 msec, slice thickness = 5 mm, flip angle = 150°, matrix size = 512 × 192) and spin-echo sagittal T1-weighted images (TR/TE = 592 msec/15 msec, slice thickness = 5 mm, flip angle = 90°, matrix size = 512 × 192) were scanned as a screening for gynecologic abnormalities. No premedication agents were given in any participants.
Since the zonal anatomy of the uterine myometrium is most clearly recognized in the midsagittal plane of the uterus, one slice in this plane was selected for obtaining echo planar imaging (EPI) images, using the axial HASTE images as reference (Fig. 1a). For BOLD MR imaging, a single-shot EPI sequence (TR = 1000 msec, TE = 23, 28, 33, 38, 43, 100, 150 msec, flip angle = 90°, bandwidth = 2894 Hz/pixel) was used to acquire a series of T2*-weighted images with seven different TEs within a single breathhold of less than 20 seconds. Other scan parameters included a field of view of 200 mm, matrix size of 64 × 64 pixels, voxel size of 3.1 mm × 3.1 mm × 5 mm and slice thickness of 5 mm with the use of fat saturation. From the series of seven images a T2* map was created by performing a voxel-by-voxel exponential fit. Each set of BOLD MR images was obtained twice at each examination, the number of acquisitions being constrained by scanner hardware limitations associated with the severe demands of EPI. The interval between the two BOLD scans was more than six minutes, separated by the acquisition of sagittal T1-weighted images and T2-weighted images.2
BOLD MR imaging was obtained twice at every examination. One of the two T2* maps was randomly selected for the main computer analysis (SPSS Japan Inc., Tokyo, Japan). In addition, both of the two T2* maps were examined regarding the repeatability of T2* value on BOLD MR imaging.
Measurement of the T2* value was performed on T2* map images, which were created as mentioned above using a satellite console of the MR unit. The T2* values of both junctional zone and outer myometrium were calculated by averaging the fitted T2*s over two appropriately-located regions of interest (ROIs) (Fig. 1b). The junctional zone was defined as the low signal intensity area outside the high signal intensity endometrium. Outer myometrium was defined as the region external to the junctional zone with moderately high signal intensity. ROIs were made as large as possible in the myometrium, but excluded the areas on the T2* map affected by susceptibility artifact caused by intestinal air. Such areas suffering severe artifact were identified on the images obtained with TE = 150 msec, which was most sensitive for the artifacts (Fig. 1c).
Additionally, signal intensities of the inner and outer myometrium were measured on T2-weighted image using ExaVision-LITE (Ziosoft Inc., Tokyo, Japan). Two ROIs, as large as possible, were placed at the anterior and posterior wall of each of the inner and outer myometrium. The signal intensities of the two ROIs were averaged.
Random selection for one of the two T2* maps was done for the main computer analysis (SPSS version 11; SPSS Japan Inc.). Both data for T2* values and signal intensity of the two myometrial layers on T2-weighted image are presented as means ± standard deviation (SD).
Analysis of variance (ANOVA) was performed with JMP 5.0 software (SAS Institute, Cary, NC, USA) to determine whether significant differences existed among the three menstrual cycle phases of T2* values and signal intensity of the T2-weighted image. If statistical significance was observed among the menstrual phases on ANOVA, a paired t-test (Bonferroni correction for multiple comparison) was provided for all pair-wise comparisons. P-values of less than 0.017 were considered as significant after Bonferroni correction.
Repeatability (95% confidence interval [CI] for the difference between the two subsequent) of the two T2* values for BOLD MR imaging was evaluated according to the methods of Bland and Altman (13). It was shown using a Bland and Altman plot and within-subject SD.
One subject was excluded due to multiple fibroids. One female was excluded from data analysis due to severe susceptibility artifact caused by intestinal air at MR imaging during one of her scans. In the other participants, the T2* value could successfully be calculated. Consequently, a total of 39 studies (13 subjects with three phases) were included for data analysis. Figure 1 shows sample BOLD images for three menstrual cycle phases of one subject.
The mean T2* values of both junctional zone and outer myometrium for each menstrual phases are summarized in Table 1, 2 and Fig. 3. In the outer myometrium the differences of T2* values between three menstrual phases were statistically significant at ANOVA test (P < 0.0001). At paired t-test (Bonferroni correction), statistical significance was revealed between the menstrual and periovulatory phase (P = 0.001) and between the menstrual and luteal phase (P < 0.0001), but was not seen between the periovulatory and luteal phase (P = 0.449). T2*values for the junctional zone were significantly lower than those of the outer myometrium at every phase (P < 0.001).
Table 1. T2* Value of Uterine Myometrium
57.0 ± 12.7
77.5 ± 10.2
83.0 ± 21.0
44.1 ± 4.3
51.2 ± 6.9
58.1 ± 9.1
Table 2. T2* Values of Uterine Myometrium of Two BOLD Scans for Evaluating Reproducibility of BOLD MR Imaging
56.7 ± 11.6
74.6 ± 11.6
87.4 ± 14.3
59.0 ± 13.4
78.7 ± 13.1
80.8 ± 21.5
45.7 ± 7.7
50.6 ± 6.4
58.2 ± 6.1
44.8 ± 5.4
50.2 ± 7.1
58.6 ± 9.7
In the junctional zone, the differences of T2* values among all three menstrual phases were statistically significant at ANOVA test (P < 0.0001). At paired t-test (Bonferroni correction), statistical significance was revealed between the menstrual and periovulatory phase (P = 0.011), between the menstrual and luteal phase (P < 0.0001), and between the periovulatory and luteal phase (P = 0.007).
Table 3 shows T2* values of two BOLD scans. Regarding repeatability across two BOLD scans, Bland and Altman (13) plots for T2* values of junctional zone and outer myometrium are shown at each menstrual cycle phases on Fig. 4a–f. Within-subject SD of outer myometrium was 6.19 in menstrual phase, 6.82 in periovulatory phase, and 8.11 in luteal phase. Within-subject SD of the junctional zone was 5.15, 4.73, and 4.20, respectively.
Table 3. Signal Intensities of Uterine Myometrium on T2-Weighted Image
271.6 ± 121.0
297.0 ± 109.3
321.7 ± 103.4
128.6 ± 52.2
136.0 ± 54.6
157.5 ± 40.7
Signal intensities on T2-weighted images of the junctional zone and outer myometrium in three menstrual cycle phases are shown in Table 3. Within each myometrial area, the difference of T2 signal intensities across three menstrual cycle phases was not statistically significant at ANOVA test (outer myometrium P = 0.30, junctional zone P = 0.12). The signal intensity of the junctional zone was significantly lower than outer myometrium in all menstrual cycle phases (P < 0.01 in three menstrual cycle phases).
The result from our study indicates that: 1) the mean T2* value of the junctional zone was significantly lower than that of outer myometrium at every menstrual phase; 2) the mean T2* values of both the junctional zone and outer myometrium during the menstrual phase were significantly lower than those in the other menstrual cycle phases.
Several reasons are hypothesized for the lower value of T2* in the junctional zone as compared with the outer myometrium. First, it should be noted that signal intensity of T2-weighted images has been observed to be consistently lower for the junctional zone, thus giving lower signal intensity than the outer myometrium in T2-weighted images. In experimental study of skeletal muscles in mice, Jordan et al (12) concluded that the level of BOLD signal intensity in moderately exercising skeletal muscle depends mainly on changes in pO2, rather than changes in blood flow or T2 effects. In their study, the blood oxygenation in the muscle was consistently decreased after contraction while blood flow change was variable (12).
In the human uterus, the decreased signal intensity in the junctional zone on T2-weighted MR images is considered to relate to subtle and constant myometrial contraction. This movement is known as uterine peristalsis, and has been previously demonstrated on transvaginal ultrasonography (TVUS) and MR imaging (14–17). In human skeletal muscle, muscle oxygenation is also reported to decrease according to the strength of muscle contraction (18, 19). Although it may be difficult to extrapolate from skeletal muscle data to the smooth muscle of the uterus, it can be argued that the decreased T2* value in the junctional zone could be caused by decreased blood perfusion due to smooth muscle contraction, leading to a consequent decrease in blood oxygenation. Although uterine peristalsis is reported to be the least during luteal phase, the T2* value of the junctional zone was still significantly lower than outer myometrium (14, 15, 17). From this fact, there may be other factors giving lower T2* value in the junctional zone than outer myometrium of the uterus, besides the effect of T2 signal intensity. More rigorous analysis, such as histological examination, will be required for fully investigating the reason for the changes of T2* signal intensity.
The decrease of T2* value in both the junctional zone and outer myometrium during the menstrual phase may also be related to decreased blood perfusion and subsequently decreased pO2. Since there was no significant change in signal intensity of the T2-weighted image both of the junctional zone and outer myometrium between menstrual cycle phases, the change in T2* value does not apparently depend directly on T2 signal intensity. Menstruation is considered to be induced by vascular spasm of the uterine spiral artery, which leads to ischemia in the endometrium and decreased oxygen availability in the myometrium (3). The relationship between blood perfusion in the myometrium and uterine contraction has been investigated utilizing intrauterine pressure recording (thermistor method) (20). With this technique, the myometrial blood flow has been shown to decrease at each myometrial contraction. A combination of BOLD MR imaging and arterial spin labeling (ASL) could clarify whether the changes we have observed are due to such blood flow changes (21, 22). EPI images of the pelvic area are prone to susceptibility and movement artifacts, caused by bowel air, bowel peristalsis, and closely bunched vessels. Therefore, a version of the flow-sensitive alternating inversion-recovery (FAIR) technique, which uses true fast imaging with steady-state precession (FISP), reported in renal imaging might provide good quality perfusion information for the uterus, if there is no serious motion artifact (22). However, this FAIR technique was not currently available. Alternatively, use of contrast agent to evaluate uterine perfusion in normal volunteers might be considered. It is clearly desirable to evaluate the relationship between perfusion and BOLD images for the uterus, using appropriate MRI sequences.
The EPI technique used in this study, which had limited spatial resolution but complete freedom from motion artifact, was highly sensitive to susceptibility artifact, leading to signal loss in some areas. However, it provided excellent signal contrast for distinguishing the junctional zone and outer myometrium, as also recently shown in BOLD images of the fetus in sheep (23). Multigradient recalled echo (mGRE) techniques have been more popular for obtaining BOLD MR images in abdominal studies (24–26). The modified multiecho data imaging combination (MEDIC) sequence has been especially recommended (27). According to these reports, the defects of EPI, such as susceptibility artifact and inferior resolution, are remedied in mGRE, but with such images we found it difficult to make a distinction even between uterine endometrium and myometrium, let alone junctional zone and outer myometrium (Fig. 5). Nonetheless, it might be possible to evaluate T2* values in the junctional zone of the uterus by combining images of the T2*-weighted map obtained by mGRE with those from a T2-mapping sequence.
Repeatability was evaluated in this study by performing two BOLD scans with only six to 10 minutes intervals. T2* values of both junctional zone and outer myometrium showed little difference at any menstrual cycle phase. It would be useful to assess BOLD repeatability with a longer interscan interval and a larger number of subjects. This could be extended to a measure of reproducibility of T2* at the same phase across menstrual cycles.
There are some limitations in this study. First, there is no gold standard for this study at present, since neither the pO2 level nor blood perfusion of the uterine myometrium can yet be directly examined. Although some experimental studies with striated muscle in animal models exhibit good correlation between T2* values and oxygenation, these results may not be applicable to the smooth muscle of human uterus. Second, we did not evaluate in detail the effects of inhaling oxygen or breathholding during the scan. As an initial exploration of these effects, since most kidney BOLD MR imaging studies (26–28) were done with breathholding, we obtained images in a pilot study with two volunteers, with breathholding and free breathing, during ovulatory and menstrual phases. The T2* values in the two conditions did not differ much (the differences were within 4.4% in outer myometrium and 8.4% in the junctional zone for menstrual phase, 9.5% in outer myometrium, and 1.1% in the junctional zone for periovulatory phase). Yet, the comparative effect of oxygen on BOLD images, (breathholding with/without inhaling oxygen) should be investigated in conjunction with oxygen saturation monitoring in the subsequent studies.
When the problems described above have been resolved, this BOLD technique can be applied to both physiological and pathological conditions of the uterus as with brain and kidney. Since uterine neoplasms have been often treated by radiation therapy and arterial embolization, evaluating oxygenation before and after treatment would provide important prognostic and treatment outcome information for both patients and clinicians. Our results may provide basic data for going on to this next stage.
In conclusion, T2* values of the junctional zone were significantly lower than those of outer myometrium at every menstrual phase, and T2* values of both junctional zone and outer myometrium in the menstrual phase were significantly lower than those in the other menstrual cycle phases. Although this study is preliminary, BOLD MR imaging may be a potential modality to investigate physiologic changes of the uterine myometrium during the menstrual cycle.
We acknowledge the support of the Wellcome Trust (to R.T.), and thank Katsuya Maruyama (Siemens-Asahi Medical Technologies Ltd.) and Ken Tamai (Department of Diagnostic Radiology and Nuclear Medicine, Kyoto University) for technical assistance.