Annals of Biomedical Engineering, Vol. 27, pp. 372–379, 1999 Printed in the USA. All rights reserved.

0090-6964/99/27~3!/372/8/$15.00 Copyright © 1999 Biomedical Engineering Society

Dynamics of the Intrauterine Fluid–Wall Interface OSNAT EYTAN,1 ARIEL J. JAFFA,2 JOSEPH HAR-TOOV,2 EITAN DALACH,1 and DAVID ELAD1 1

Department of Biomedical Engineering, Faculty of Engineering, Tel Aviv University, Tel Aviv, Israel, and 2Ultrasound Unit, Department of Obstetrics and Gynecology, Lis Maternity Hospital, Tel-Aviv Sourasky Medical Center, Tel-Aviv, Israel, and the Sackler Faculty of Medicine, Tel Aviv University, Tel-Aviv, Israel (Received 10 August 1998; accepted 12 March 1999)

Abstract—Intrauterine fluid movements, which are responsible for embryo transport to a successful implantation site at the fundus, may be induced by myometrial contractions. Myometrial contractions in nonpregnant uteri were studied from in vivo measurements of intrauterine pressures with fluid-filled catheters and by visual observations of high-speed replaying of ultrasound images of the uterus. Transvaginal ultrasound ~TVUS! images of sagittal cross sections of the nonpregnant uterus were scanned with an intravaginal ultrasound probe. Images at consecutive times ~2 s apart! were digitized and processed by employing modern techniques of image processing. The sets of images were compared to evaluate time variation of the fluid–wall interface with respect to amplitude, frequencies, and wavelength of myometrial contractions. Analysis of TVUS images from 11 volunteers during the proliferative phase revealed that myometrial contractions are fairly symmetric and are propagated from the cervix towards the fundus at a frequency of about 0.01–0.09 Hz. The wavelength, amplitude, and velocity of the fluid–wall interface during a typical contractile wave were found to be 10–30 mm, 0.05–0.2 mm, and 0.5–1.9 mm/s, respectively. Additional data acquisition from a large number of normal subjects is needed to build a data base to predict normal characteristics of myometrial contractions in a nonpregnant uterus, in order to better understand their role in the preimplantation process. © 1999 Biomedical Engineering Society. @S0090-6964~99!02003-2#

is the result of spontaneous myometrial contractions towards the fundus.15,32 Irregular uterine motility may introduce a mechanical factor of infertility,20 an event which needs to be avoided in an effort to increase the currently very low rate of success of embryo transfer after in vitro fertilization ~IVF! in the laboratory,12 or alternatively, can be utilized for development of new contraception techniques. The putative mechanism for pregnancy failure remains elusive, but intrauterine fluid flow may play an important role in human reproduction. The nonpregnant uterus is a small organ with a thin cavity ~about 1 mm! that has a triangle-like shape in the oblique anterior–posterior cross section ~a base of 3 cm and height of 5 cm! that is enclosed by relatively very thick walls. The uterine wall is composed of three layers: the inner endometrium which is 1–8 mm thick, the middle muscular layer ~called the myometrium! which is 1.5–2.5 cm thick, and the perimetrium which is the outer coating.9 The myometrium consists mainly of elongated smooth muscle cells which are embedded in extracellular material in which collagen fibers transmit the contractile forces generated by individual muscle cells.27 The uterine muscle cells, similar to those of other smooth muscles, are small and spindle shaped and are embedded in abundant connective tissue. While it is known that myometrial cells can either be excited by an action potential from a neighboring cell or generate their own impulse, the role of autonomic nerves in uterine function is poorly understood.14,25,28 Uterine motility was first studied by measuring intrauterine pressures in nonpregnant women using methods involving invasive introduction of fluid-filled catheters.6–8,18,29 The frequency of intrauterine pressure signals was shown to depend on the subject’s posture and the level of hormones: estrogen accelerates it and progesterone decelerates it. A summary of the values reported in these studies is shown in Table 1. It should be noted that since the transducers were of the same size as the gap between the uterine walls, they could have induced direct contact between the endometrium and the

Keywords—Uterine motility, TVUS, Embryo transport, Implantation.

INTRODUCTION In the normal reproductive process, the spermatozoa propel themselves through the uterine fluid towards the fallopian tube where fertilization occurs. The formed zygote is driven to the uterine cavity within four days of ovulation, and is then conveyed during another four days to an optimal implantation site in the fundal area at the upper part of the uterus.16 Fulfillment of these essential events within the time limits, first for fertilization, and then for implantation, depends on concomitant intrauterine fluid motion induced by uterine wall motility which Address for correspondence to David Elad, Department of Biomedical Engineering, Faculty of Engineering, Tel Aviv University, Tel Aviv 69978, Israel. Electronic mail: [email protected]

372

Dynamics of the Uterine Wall

373

TABLE 1. Uterine motility measurements by fluid-filled catheters and transvaginal ultrasound imaging. Intrauterine pressuresa Phase Menstruation (days 1–6)

Myometrial contractions (TVUS)b

Amplitude (mm Hg)

Frequency (Hz)

Amplitude (mm)

Frequency (Hz)

Wavelength (mm)

Direction

100–200

0.005–0.008

¯

0.008–0.04

¯

Fundus to cervix

Proliferative (days 7–14)

5–25

0.03–0.08

1.6

0.05–0.07

7.6

Cervix to fundus

Secretory (days 15–28)

5–25

0.025–0.05

1.7

0.05–0.09

7.0

Cervix to fundus

a

References 6–8, 18, and 29. References 1, 4, 5, 10, 19, 24, and 31.

b

probe, and this could have altered the normal uterine activity. The introduction of transvaginal ultrasound ~TVUS! imaging enabled minimally invasive observation of myometrial motility from sagittal images of the uterus ~Fig. 1!. The characteristics of myometrial activity were determined by visual inspection of these ultrasound recordings played at a high speed. The spontaneous contractions were observed to propagate from the cervix to the fundus at a rate of 1–5 contractions/min during the proliferative and secretory phases ~Table 1!, with the direction being reversed and the contractions’

rate slowed ~0.5–2.5 contractions/min! during menstruation.1,4,5,10,19,24,26,31 Myometrial contractions are responsible for the frequent changes in the inner fluid–wall interface, events which induce intrauterine fluid flow. However, it is unclear how they are attuned or coordinated. The objective of this work was to characterize the dynamics of the intrauterine fluid–wall interface from in vivo images of the sagittal cross sections of the uterus. Techniques of image processing were employed to TVUS images to provide an objective evaluation of intrauterine fluid–wall dynamics in nonpregnant women.

METHODS The goal of this study was achieved in two stages. First, the recorded TVUS images were processed to yield the instantaneous contour of the intrauterine cavity. Then, analysis of variation of these contours with time ~i.e., from consecutive images! provided the dynamic characteristics of the intrauterine fluid–wall interface. In Vivo Data Acquisition

FIGURE 1. „a… Orientation of transvaginal ultrasound imaging. „b… Ultrasound image of a sagittal cross section of a nonpregnant uterus.

Transvaginal ultrasound images of sagittal cross sections of a nonpregnant uterus @Fig. 1~b!# were recorded from 11 healthy volunteers in the active reproductive phase of their life ~age range 23–40 years! during the proliferative and early secretory phases of the menstrual cycle ~days 10–16!. The images were acquired with a TVUS system of Advanced Technology Laboratories ~ATL-9HDI!, which is equipped with a compound multifrequency probe ~5–9 MHz!. The specified resolution ~axial and lateral! of the system was 0.2 mm in the center of the image and the spatial intensity was 1 mW/cm2. The system had a built-in SVHS video tape that allowed on-line recording of the images.

374

EYTAN et al.

FIGURE 2. „a… TVUS image of a sagittal cross section of a uterus. „b… The reference line connecting the maximal intensities of the uterine fluid. „c… Binary image of the TVUS image „fluid, white; uterine wall, black…. „d… Detection of the uterine cavity boundaries. „e… Smoothed boundaries „fluid– wall interface… of the uterine cavity.

The sagittal cross section near the fundus was magnified ~‘‘zoomed out’’! to reveal a length of about 3 cm from the fundus towards the cervix rather than produce a full sagittal cross section ~about 5 cm long! as in regular clinical practice @Fig. 1~b!#. The images were recorded for about 5 min, while the subject was lying relaxed in a supine position and with the operator holding the probe steady. At the beginning of each recording, the operator arbitrarily marked four points on the images ~with a known distance between each pair! which were used for horizontal and vertical calibration of the image geometry. The images were replayed on a variable-speed professional videotape system ~Panasonic AG-7355! that allowed for sampling of consecutive frames which were 2 s apart from each other by a frame grabber ~Data Transmission DT-2853! directly to a personal computer. The data of each subject yielded about 150 images which were digitized into files of 5123512 pixels each. Data Analysis The data processing was composed of two stages. First, the boundary of the uterine cavity from each TVUS image was detected by an edge-detecting technique. In the second stage, the dynamics of uterine motility ~i.e., the intrauterine fluid–wall interface! was analyzed from the variation of the geometry of the uterine cavity with time, as observed from consecutive images. Edge Detection of the Uterine Cavity. Analysis of the digitized data was performed only in the region containing the uterine cavity @Fig. 2~a!# in order to isolate and study the precise contour that separates the inner wall from the fluid. The uterine fluid could be clearly ob-

served as a long and very narrow bright band which appeared to have some discontinuities @Fig. 2~a!#. In order to identify and to connect all regions of the fluid, which are expected to be continuous in the real cavity, a ‘‘reference line’’ which connected the two ends of the cavity and passed through all the points with maximal intensity within the cavity was used @Fig. 2~b!#. The processing procedure required two manual steps: ~i! reading of the four calibration points, and ~ii! marking the end points of the observed uterine cavity @A and B in Fig. 2~b!#. The rest of the procedure is fully automated and is comprised of the following steps: ~i! evaluation of the reference line between points A and B @Fig. 2~b!#, ~ii! thresholding of the image in order to transform the gray-scale values of each pixel into a binary image with white pixels for the fluid and black ones for the uterine wall @Fig. 2~c!#, ~iii! detection of the edges of the fluid– wall interface @Fig. 2~d!#, ~iv! transformation of the image plan given in pixels into dimensional values by using the four calibration points, and ~v! smoothing of the boundaries of the fluid–wall interfaces @Fig. 2~e!#. The reference line, which outlines the uterine cavity, was determined with an algorithm that allows proceeding from point A to point B @Fig. 2~b!# in a path that connects adjacent points that yield the maximal value of accumulated intensity. Transformation of the gray-scale image into a binary one ~with only black and white pixels! was done using a threshold value Th that was determined from the following empirical relationship:

Th5

h ims rl1 h rls im , s im1 s rl

where h im is the mean pixel intensity of the whole image @shown in Fig. 2~a!#, h rl is the mean pixel intensity along the reference line @shown in Fig. 2~b!#, s im is the image matrix pixel intensity standard deviation, and s rl is the standard deviation of the pixel intensities along the reference line. This formula is based on widespread experience and has proven itself to be useful for images of different size, brightness, and contrast.13 Edge detection of the contour that defines the intrauterine fluid–wall interface was obtained by proceeding from each point on the reference line in both the anterior and posterior directions @Fig. 2~d!# until reaching the first black pixel. The image geometry, which is given in pixels, was transformed into dimensional values with a calibration matrix whose coefficients were computed from the four calibration points and the known distance between each pair. The resolution of a typical image was 6–8 pixels per mm along the cavity and 10–12 pixels per mm across the cavity. The resulting curves that defined the anterior and posterior boundaries of the cavity were smoothed by using cubic spline approximation.2

Dynamics of the Uterine Wall

FIGURE 3. The uterine cavity: „a… real geometry of the uterine cavity, „b… straightened geometry of the uterine cavity, and „c… smoothed geometry of the uterine cavity.

The physical geometry of the uterine cavity appeared as a narrow channel along the curved reference line @Fig. 3~a!# and changed with time due to myometrial motility. In order to compare consecutive images and to analyze the uterine dynamic characteristics, we evaluated the instantaneous width of the cavity along the reference line and presented the results along a straightened reference line @Fig. 3~b!#. The observed contour of the uterine cavity had small surface curvatures, which may be due to the corrugated inner surface of the endometrium. However, the contractile fibers in the myometrium, which are responsible for wall motility, are expected to have much larger curvatures. Accordingly, for the analysis of the dynamics of uterine wall motility, the contour was further smoothed with a low-pass Butter filter to avoid larger gradients along the reference line @Fig. 3~c!#. We found that this smoothing procedure did not change the frequency spectrum of uterine motility ~as analyzed later!. Analysis of Uterine Motility. The dynamic characteristics of the wave displacement at the fluid–wall interface were obtained from analysis of the straightened geometry of the uterine cavity from consecutive ultrasound images ~Fig. 4!. The analysis was performed with respect to the fundal end of the cavity that can be identified in each image. Figure 5 depicts the time variation of the width of the cavity, W( j ,t), at fixed distances from the fundus ~j!, which provided information on the frequency and amplitude of uterine motility as well as on the nature of the symmetry of wall displacement. To compensate for the image resolution ~0.125 mm per pixel along the cav-

375

FIGURE 4. Example of a set of straightened cavities that resulted from the processing of consecutive ultrasound images.

ity!, the ultrasound resolution ~0.2 mm!, and measurement errors, we examined the motility of a thin slice perpendicular to the cavity axis ~about 0.625 mm thick! that is represented by the averaged data of five pixels in the vicinity of a fixed distance from the fundal end. Symmetric uterine motility ~i.e., opposite walls move either outwardly or inwardly in a simultaneous manner!

FIGURE 5. Time variation of the width of the uterine cavity at different distances from the fundus.

376

EYTAN et al.

speed was computed from the time lapse between consecutive peaks from each distance from the fundus. The wavelength was computed by dividing the wave speed by the leading frequency obtained from the analysis. RESULTS

FIGURE 6. Time variation of the width of the uterine cavity at two distances from the fundus. The time lapse between two peaks „D t … is used for calculation of the wave speed.

is defined as 21.0, while full asymmetry is 1.0. The level of symmetry of the fluid–wall interface displacement was determined by a correlation function that describes the extent of symmetry between displacements of the anterior and posterior curves with respect to gradients ~e.g., ‘‘CORREL’’ of EXCEL 7.0!. The geometry of the passive intrauterine cavity ~i.e., in the absence of contractions! in the sagittal cross section was obtained from the time variations of the width of the uterine cavity at a given fixed distance from the fundus. Assuming wall displacement is with respect to the passive state, we computed the arithmetic average of the distance between the anterior and posterior walls at any given distance from the fundus over the time of measurement. The frequency of uterine wall motility at fixed distances from the fundus was derived by means of a fast Fourier transform routine. Ultrasound images were sampled at a rate of 0.5 Hz (Dt52 s); thus, the frequency spectrum was analyzed in the range 0, f ,0.25 Hz. The speed of the uterine contractile waves was derived from the time variation of the cavity width at two fixed distances from the fundus ~Fig. 6!. The

Visual inspection of the recorded ultrasound images from all the subjects confirmed that the general direction of spontaneous uterine contractions was from the cervix towards the fundus.10 These waves seemed to decay as they approached the fundus. The instantaneous geometry of the uterine cavity changed from image to image, providing a visual display of wall motility ~Fig. 4!. Analysis of the induced wave propagation along the intrauterine fluid–wall interface revealed a periodic oscillatory motion at fixed distances from the fundus ~Fig. 5! which was fairly symmetric ~→21.0! but then became progressively more asymmetric as it neared the fundus ~Table 2!. The passive width (W p ) of the uterine cavity of one subject is shown in Fig. 7. Measurements from all the subjects of this study showed a maximal width of 0.88 mm at the fundus end that decreased as the distance from the fundus increased, as shown in Table 2. The amplitudes of anterior and posterior wall displacements varied between 0.03 and 0.26 mm in all subjects, with the highest values being near the fundus. The averaged amplitudes at various locations from the fundus are given in Table 2. The frequency of wall displacement for all subjects and at different distances from the fundus was found to be within the two frequency ranges of 0.01– 0.09 and 0.13–0.25 Hz. The group of low frequencies is in the range of previous measurements with different experimental modalities ~Table 1! while the group of the

TABLE 2. Averaged characteristics of intrauterine motility from all subjects at various distances from the fundus „standard deviations are given in parenthesis…. Distance from fundus, j (mm) 1

4

7

10

13

16

19

22

Passive width W p (mm)

0.593 (0.225)

0.564 (0.111)

0.463 (0.113)

0.406 (0.131)

0.346 (0.127)

0.357 (0.154)

0.362 (0.133)

0.461 (0.206)

Anterior wall displacement amplitude (mm)

0.152 (0.068)

0.097 (0.034)

0.077 (0.023)

0.071 (0.024)

0.077 (0.030)

0.073 (0.028)

0.070 (0.028)

0.063 (0.039)

Posterior wall displacement amplitude (mm)

0.144 (0.054)

0.092 (0.023)

0.081 (0.022)

0.069 (0.022)

0.074 (0.026)

0.069 (0.024)

0.065 (0.023)

0.068 (0.042)

¯

0.03 (0.019)

0.035 (0.015)

0.030 (0.017)

0.036 (0.020)

0.026 (0.020)

0.038 (0.019)

¯

Parameter

Frequency (Hz) Motility mode (symmetry[21.0)

20.571

20.739

20.788

20.806

20.881

20.885

20.919

20.945

Dynamics of the Uterine Wall

FIGURE 7. The passive width of the sagittal cross section from one subject.

faster ones ~0.13–0.25 Hz! covers the range of normal breathing frequencies. Based on the published data, and assuming that breathing of the subject may contribute wall displacement information due to slight out-of-plan imaging, the slower frequencies ~0.01–0.09 Hz! were considered as the ones typical of intrauterine wall displacement of the subjects of this study. The averaged frequencies at fixed distances from the fundus for all the subjects are summarized in Table 2. The displacement wave along the intrauterine fluid– wall interface was propagated in a variable velocity that ranged from 0.5 to 1.9 mm/s, with an averaged value of 0.99 mm/s. The velocities were smaller near the fundus, but changed randomly as the distance increased from the fundus. The wavelengths were relatively long ~10–30 mm! compared to the length of the uterine cavity when measured in a sagittal cross section ~50 mm!. DISCUSSION Motility of the uterine cavity resulting from myometrial contractions during the proliferative and early secretory phases ~days 10–16! was characterized from ultrasound images by a new technique of image processing. Generally, the contractions were propagated from the cervix to the fundus, in a fairly symmetric pattern, as had been observed previously.1,5,10,19,24,26,31 However, their symmetry deteriorated at locations close to the fundus. Uterine activity is a contractile wave composed of a range of frequencies ~0.01–0.09 Hz! which are in a good agreement with frequencies that were extracted from high-speed replay of the TVUS recordings.5,10,24,26 In some subjects, the frequency of wall displacement varied along the axis of the uterine cavity. This may be attributed to the morphology of myometrial smooth muscle and their electrohpysiological activity.25,27 Uterine contraction probably involves several groups of muscle fibers of different lengths which may have different dynamic characteristics. This may explain why the results obtained for the wave speeds varied between 0.5 and 1.9 mm/s. The averaged wave speed found here, 0.99 mm/s, was slightly smaller than those of IJland et al.21 for the

377

mid- and late-proliferative phases ~1.4 and 1.7 mm/s, respectively!, which were interpreted from visual observations. The oblique cross section of the uterine cavity had been previously investigated since its size is important for the design of intrauterine contraceptive devices;17 information on the cavity width, however, is not available. In the present study, it was found that during the period of 10–16 days of the menstrual cycle, the averaged passive width of the uterine cavity was about 0.6 mm at the fundus and that it decreased to about 0.4 mm as the distance from the fundus increased ~Table 2!. The averaged amplitudes of anterior and posterior wall displacements varied between 0.06 and 0.15 mm ~which represent local reduction of the cavity width in the order of 30%–50%!, and were the highest near the fundus ~Table 2!. It should be borne in mind that the accuracy of the measured geometry of the fluid–wall interface is limited by the resolution of the system for ultrasound imaging ~0.2 mm!. The results for the amplitudes of wall displacement had the same magnitude as the resolution. However, reproducible results were obtained in cases in which we repeated the edge-detection procedure @starting with marking of the end points A and B of the reference line in Fig. 2~b!#.

Application to Embryo Transport Successful implantation of the embryo involves hormonal, biochemical, and mechanical aspects,3 but very little attention has been devoted to the role of myometrial contractions. These contractions change directions during the menstrual cycle, aid the transport of spermatozoa towards the fallopian tube following sexual intercourse, and direct the embryo towards the implantation site.24 How they are attuned or coordinated, however, is unclear. Better understanding of uterine activity may have decisive practical applications in the process of IVF, in which successful embryo transfer to implantation in the endometrium is still so woefully disappointing. IVF in the laboratory is successfully achieved in 70% of the cases, but embryo transfer to the endometrium fails in 80% of them.12 Embryos suspended in liquid are transferred into the uterus with the aid of a thin catheter which is inserted through the cervix. Introduction of the catheter into the uterine cavity interferes with the regular uterine motility and may induce chaotic contractions22 that may drive the embryo in chaotic directions and prevent implantation. Recently, we developed a theoretical simulation of wall-induced peristaltic fluid flow in a model of the uterine cavity that demonstrated the strong dependency of uterine fluid transport patterns on the mode of wall displacement ~symmetric versus asymmet-

EYTAN et al.

378

ric motility, ratio of amplitude-to-cavity width and wave speed! which are the results of myometrial contractions.11 The other side of the fertility coin is contraception. Alteration of the normal pattern of uterine contractions may induce intrauterine fluid motions which could act as a contraceptive mechanism in two ways: prevent sperm transport to the ampulla where fertilization takes place, and prevent embryo implantation by impeding its approach to the endometrium. Since sperm transport through the cervical mucus is too rapid ~50 mm/s! to be accounted for only by flagellar propulsion by its tail,30 it was hypothesized that myometrial contractions from cervix to fundus provide the sperm the additional propulsion force needed to reach the ampulla.23,24 By interfering with the natural and synchronized uterine contractions, sperm will not reach the ampulla to fertilize the ovum 24 h after ovulation, thus obviating pregnancy.

CONCLUSIONS Intrauterine fluid movements induced by myometrial contractions were studied from in vivo measurements of TVUS images of the sagittal cross section of the uterine cavity by utilizing techniques of image processing. Analysis of the time variation of the contours that define the uterine cavity provided objective measures of wave propagation along the fluid–wall interface in terms of amplitudes, frequencies, speed, and wavelength. The width of the sagittal cross section of the uterine cavity increases towards the fundus, and the cavity is not fully occluded during the contractions. Characterization of nonpregnant uterine contractions is complicated, because they are composed of a range of frequencies, variable amplitudes, and different wavelengths. The general direction of uterine contraction in the proliferative and early secretory phases is from cervix to fundus and supports the assumption that embryos are transported by uterine contractions. This study is the first objective effort to quantify spontaneous uterine contractions. The value of some results were of the same order of magnitude as the resolution of the ultrasound; however, repeated processing of the same image resulted in the same values. Since the phenomenon is of a complex spatial pattern, additional interesting and important information will be made available when time–sequence three-dimensional ultrasound imaging will be possible for gynecological applications. Future studies will also need improved resolution of the ultrasound instrumentation as well as data acquisition from a large number of normal subjects to construct a data base of healthy subjects.

ACKNOWLEDGMENTS The authors are thankful to Esther Eshkol for editorial assistance. This work was partially supported by the Ela Kodesz Institute for Medical Engineering and Physical Sciences and the Basic Research Fund of Tel Aviv University.

REFERENCES 1

Abramowicz, J. S., and D. F. Archer. Uterine endometrial peristalsis: A transvaginal ultrasound study. Fertil. Steril. 54:451–454, 1990. 2 Ahlberg, J. H., E. N. Nilson, and J. L. Walsh. The Theory of Splines and their Applications. New York: Academic, 1967. 3 Baggott, L. M. Human Reproduction. Cambridge, UK: Cambridge University Press, 1997. 4 Birnholz, J. C., Ultrasonic visualization of endometrial movements. Fertil. Steril. 41:157–158, 1984. 5 Chalubinski, K., J. Deutinger, and G. Bernaschek. Vaginosonography for recording of cycle-related myometrial contractions. Fertil. Steril. 59:225–228, 1993. 6 Cibils, L. A. Contractility of the nonpregnant human uterus. Obstet. Gynecol. 30:441–461, 1967. 7 Coutinho, E. M., L. Nascimento, H. D. S. Maia, and M. T. R. Goncalves. Motility of human myometrium in vivo: Intramural versus transcervical recording. Am. J Obstet. Gynecol. 134:801–812, 1979. 8 Csapo, A. I., and C. R. Pinto-Dantas. The cyclic activity of the nonpregnant human uterus. A new method for recording intrauterine pressure. Fertil. Steril. 17:34–38, 1966. 9 Demopoulos, R. I., and K. R. Mittal. Anatomy, histology and physiology. In: The Uterus Pathology, Diagnosis and Management, edited by A. Altchek and L. Deligdisch. New York: Springer, 1991. 10 DeVries, K., E. A. Lyons, J. Ballard, C. S. Levi, and D. J. Lindsay. Contractions of the inner third of the myometrium. Am. J. Obstet. Gynecol. 162:679–682, 1990. 11 Eytan, O., and D. Elad. Analysis of intra-uterine fluid motion induced by uterine contractions. Bull. Math. Biol. 61:221– 238, 1999. 12 Fahy, U. M., D. J. Cahill, P. G. Wardle, and M. G. Hull. In vitro fertilization in completely natural cycles. Hum. Reprod. 10:572–575, 1995. 13 Fisher, R., S. Perkins, A. Walker, and E. Wolfart. Hypermedia Image Processing Reference. Chichester, UK: Wiley, 1996. 14 Garfield, R. E., Structural studies of innervation on nonpregnant rat uterus. Am. J. Physiol. 251:C41–C54, 1986. 15 Guyton, A. C., and J. E. Hall. Textbook of Medical Physiology, 9th ed., Philadelphia, PA: Saunders, 1996, p. 1033. 16 Hafez, E. S. E. Human Reproduction Conception and Contraception. New York: Harper and Row, 1980. 17 Hasson, H. M., Uterine geometry and IUCD design. Br. J. Obstet. Gynecol. (Suppl). 89:1–10, 1982. 18 Hendricks, C. H., Inherent motility patterns and response characteristics of the non-pregnant human uterus. Am. J. Obstet. Gynecol. 96:824–843, 1966. 19 IJland, M. M., L. H. Evers, G. A. J. Dunselman, C. van Katwijk, C. R. Lo, and H. J. Hoogland. Endometrial wavelike movements during the menstrual cycle. Fertil. Steril. 65:746–749, 1996.

Dynamics of the Uterine Wall 20

IJland, M. M., L. H. Evers, G. A. J. Dunselman, L. Volovics, and H. J. Hoogland. Relation between endometrial wavelike activity and fecundability in spontaneous cycles. Fertil. Steril. 67:492–496, 1997. 21 IJland, M. M., L. H. Evers, and H. J. Hoogland. Velocity of endometrial wavelike activity in spontaneous cycles. Fertil. Steril. 68:72–75, 1997. 22 Knutzen, V., C. J. Stratton, G. Sher, P. I. McNamee, T. T. Huang, and C. Soto-Albors. Mock embryo transport in early luteal phase, the cycle before in vitro fertilization and embryo transfer: A descriptive study. Fertil. Steril. 57:156–162, 1992. 23 Knuz, G., D. Beil, H. Deininger, L. Wildt, and G. Leyendecker. Endocrine control of rapid sperm transport through the female genital tract as documented by hysterosalpingoscintigraphy: Physiological and pathophysiological aspects. Hum. Reprod. 10:149–150, 1995. 24 Knuz, G., D. Beil, H. Deininger, L. Wildt, and G. Leyendecker. The dynamics of rapid sperm transport through the female genital tract: Evidence from vaginal sonography of uterine peristaltic and hysterosalpingoscintigraphy. Hum. Reprod. 3:627–632, 1996. 25 Kuriyama, H., K. Kitamura, T. Itoh, and R. Inoue. Physiological features of visceral smooth muscle cells, with special

379

reference to receptors and ion channels. Physiol. Rev. 78:811–920, 1998. 26 Lyons, E. A., P. J. Taylor, and X. H. Zheng. Characterization of subendometrial myometrial contractions throughout the menstrual cycle in normal fertile women. Fertil. Steril. 55:771–774, 1991. 27 Marshall, J. M. The physiology of the myometrium. In: The Uterus, edited by H. J. Norris. Baltimore, MD: Williams and Wilkins, 1973, pp. 89–109. 28 Marshall, J. M., Effects of ovarian steroids and pregnancy on adrenergic nerves of uterus and oviduct. Am. J. Physiol. 240:C165–C174, 1981. 29 Martinez-Gaudio, M., T. Yoshida, and L. P. Bengtsson. Propagated and non-propagated myometrial contractions in a normal menstrual cycle. Am. J. Obstet. Gynecol. 115:107– 111, 1973. 30 Odeblad, E., Undulations of macromolecules in cervical mucus. Int. J. Fertil. 7:313–319, 1962. 31 Oike, K., K. Ishihara, and S. K. Kuchi. A study on the endometrial movements and serum hormonal level in connection with uterine contraction. Acta Obstet. Gynecol. Jpn. 42:86–92, 1990. 32 Seibel, M. M. Infertility: A Comprehensive Text, 2nd ed. Stanford, CT: Appleton and Lange, 1997, p. 462.