| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Experimental Studies |
1 From the Department of Radiology, Georgetown University Medical Center, 3800 Reservoir Rd NW, Washington, DC 20007-2197 (R.E.F., S.M.A.) and Siemens Medical Systems, MR Research and Development, Iselin, NJ (D.T.). From the 1997 RSNA scientific assembly. Received September 3, 1998; revision requested October 23; final revision received March 29, 1999; accepted July 28. S.M.A. supported in part by a grant from Siemens Medical Systems. Address reprint requests to S.M.A. (e-mail: aschers@gunet.georgetown.edu).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: A phantom simulating the uterus, fallopian tubes, and surrounding pelvic cavity was constructed. T2-weighted acquisition strategies—breath-hold fast spin-echo, rapid acquisition with relaxation enhancement (RARE), and half-Fourier RARE—were refined to acquire sequential 70-mm coronal imaging volumes. Contrast agent was injected into the introducing catheter entering the os of the simulated uterus. Interacquisition interval, type of contrast agent (eg, sterile saline solution or water), and quantity of contrast agent (eg, 1–5 mL per acquisition) were varied. Digital image subtraction was used to enhance image quality. Images were qualitatively analyzed and rated good, fair, or poor for temporal resolution, spatial resolution, fallopian tube conspicuity, and free spill conspicuity. Once the technique was refined, the phantom was reconfigured to simulate unilateral and bilateral hydrosalpinx.
RESULTS: The RARE sequence with an 8-second interacquisition interval and a 5-mL interacquisition injection of sterile water produced good images of the simulated fallopian tubes and free spill. Depiction of unilateral and bilateral hydrosalpinx was also reliably demonstrated.
CONCLUSION: This study with a phantom model demonstrates the feasibility of MR hysterosalpingography to depict normal and diseased fallopian tubes.
Index terms: Fallopian tubes, MR, 853.121411, 853.121415, 853.121416 • Fallopian tubes, stenosis or obstruction, 853.2172, 853.2175 • Magnetic resonance (MR), experimental studies, 853.121411, 853.121415, 853.121416 • Magnetic resonance (MR), half-Fourier imaging, 853.121416 • Magnetic resonance (MR), rapid imaging, 853.121416 • Phantoms, 853.121411, 853.121415, 853.121416
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Retrograde chromopertubation, the current standard, involves injecting aqueous dye into the uterus and depicting spill from the distal fallopian tube into the pelvic cavity. This technique only indicates patency of the entire hysterosalpingal complex and does not provide information concerning the location of the potential abnormality. Hysteroscopy allows examination of the uterine lumen, but the fallopian tubes cannot be depicted beyond the interstitial segment. Although combining these two techniques increases diagnostic information, these methods are invasive, costly, and carry the usual risks that accompany laparoscopic procedures (3).
Dynamic hysterosalpingography is an alternative to retrograde chromopertubation with hysteroscopy and can depict the presence and location of a uterine or fallopian tube abnormality (4). Its accuracy for determining fallopian tube patency is 63%–93% (5,6); however, dynamic hysterosalpingography relies on ionizing radiation and may be complicated by, in order of decreasing frequency, pain, bleeding, intravasation of contrast agent, pelvic infection, and/or reaction to contrast agent (2,7). Of these, infection is the most serious and occurs in 1.4% of women with dilated fallopian tubes (2). Prophylactic doxycycline limits the occurrence of infection.
Two-dimensional transvaginal US hysterosalpingography is a minimally invasive procedure whereby fluid is instilled into the uterus, and fallopian tube patency is assessed in real time (8). US hysterosalpingography does not require ionizing radiation, iodinated contrast agent, or a laparoscopic procedure. Reported accuracy rates vary from 69% to 83%, which reflects the operator dependence of the procedure and type of fluid instilled (9,10).
Magnetic resonance (MR) imaging is an attractive method for evaluating the female pelvis. It does not use ionizing radiation and has no known adverse biologic effects in women of reproductive age (11). Although MR imaging has been used to detect a wide range of conditions that affect fertility (12–26), fallopian tube patency continues to be inferred from images obtained at conventional hysterosalpingography, US hysterosalpingography, and retrograde chromopertubation with hysteroscopy (2).
To our knowledge, MR imaging has not been used in conjunction with intrauterine contrast agent to assess fallopian tube patency in women. The purpose of this study was to determine the feasibility of MR imaging to assess fallopian tube patency with intrauterine contrast agents in a phantom model of the female pelvis.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The uterus, fallopian tubes, and introducing catheter were placed in a 40-cm long x 25-cm wide x 10-cm deep plastic container that was 1-mm thick and that was modified to allow entry of the three tubes. Each entry point was reinforced with a waterproof sealant. Each simulated fallopian tube was coiled three times before exiting the container to keep the tubes exclusively on one side of the simulated uterus. The distal ends of all tubes were able to accept syringe adapters. Syringes were used to simulate fallopian tube obstruction and to inject contrast agent into the introducing catheter.
All structures in the plastic container were embedded in a gel that had MR characteristics of smooth muscle tissue. Specifically, 100% animal hide gelatin and 17% glycerol were mixed together in accordance with the protocol developed by Blechinger et al (27). The completed phantom is shown in Figure 1.
|
Several T2-weighted acquisition strategies were evaluated. Requirements for the imaging protocol included (a) good conspicuity of the fluid-filled fallopian tubes and (b) temporal resolution and patient tolerance comparable with those for standard hysterosalpingography. Gradient- echo–based acquisition strategies, including echo-planar techniques, were not studied owing to concerns about known susceptibility artifacts associated with air-filled loops of bowel in the pelvis. The three sequences evaluated were (a) breath-hold fast spin echo (3,000/138 [repetition time msec/echo time msec]) performed with a matrix of 116 x 256, a pixel size of 2.3 x 1.4 mm, and an acquisition time of 16 seconds; (b) half-Fourier rapid acquisition with relaxation enhancement ([RARE] HASTE; Siemens Medical Systems), (
/95) performed with a matrix of 240 x 256, a pixel size of 1.1 x 1.4 mm, and an acquisition time of 1 second; and (c) RARE (/1,100) performed with a matrix size of 240 x 256, a pixel size of 1.1 x 1.4 mm, and an acquisition time of 7 seconds. Each acquisition produced a 70-mm coronal imaging volume of the pelvic phantom.
To increase image quality, digital image subtraction of each image from a baseline image was performed by using the standard image subtraction algorithm provided with the MR imaging system.
Protocol for Experiment 1
Before each experiment, the intrauterine introducing catheter was filled with either saline solution (0.9% sodium chloride; Baxter Healthcare, Deerfield, Ill) or sterile water (Baxter Healthcare). After each acquisition, saline solution or water was injected into the intrauterine introducing catheter. The volume of contrast agent was 1–5 mL, in 1-mL increments, and was held consistent throughout each experiment. Interacquisition intervals were 3–10 seconds, in 1-second increments. Twenty sequential acquisitions were obtained five times per experiment. After each experiment, the phantom was emptied of contrast agent and air.
Simulated Hydrosalpinx for Experiments 2 and 3
Once the sequence and protocol were refined, the phantom was reconfigured to simulate unilateral hydrosalpinx for experiment 2 and bilateral hydrosalpinx for experiment 3. Hydrosalpinx was simulated by securing a 20-mL syringe to the adapter at the distal end of the simulated fallopian tube. To attempt to recreate the compliance properties of the uterus and fallopian tubes, the plunger of the syringe was able to respond to (ie, withdraw from) the increasing pressure caused by the injection of contrast agent into the phantom.
Sequence Evaluation
Images were interpreted by two investigators in concert: one was a radiologist (S.M.A.) experienced in pelvic MR imaging, and the other was a medical student (R.E.F.) with an engineering background who received a tutorial in MR hysterosalpingography that concentrated on normal "fill and spill" and hydrosalpinx, regardless of cause. Images obtained with each sequence were evaluated for temporal and spatial resolution, fallopian tube conspicuity, and depiction of free spill from the distal end of each fallopian tube. Qualitative ratings of good, fair, or poor were assigned for each experiment.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
Interacquisition interval.—Short interacquisition intervals resulted in poor fallopian tube conspicuity. An interacquisition interval of 8 seconds produced images with good water-tissue depiction and allowed reproducible depiction of the fallopian tubes and free spill.
Quantity of contrast agent.—A 5-mL interacquisition injection of contrast agent was required to provide depiction of both fallopian tubes and to ensure contrast agent was spilled from the distal ends of both simulated fallopian tubes in our model (Fig 2). Quantities less than 5 mL did not reproducibly opacify both tubes.
|
Experiment 2: Simulated Unilateral Hydrosalpinx
By using the refined sequence and protocol defined with experiment 1, the phantom was reconfigured to simulate unilateral hydrosalpinx. Sterile water highlighted the lumen of the patent left fallopian tube, and free spill was reproducibly depicted at its distal end. In contradistinction, the simulated obstructed fallopian tube did not opacify with contrast agent (Fig 3).
|
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
That the RARE sequence performed best is not unexpected, as this sequence has been exploited to depict other small fluid-filled structures such as the biliary tree (28). The superior performance of RARE over that of half-Fourier RARE is also not unexpected; the longer echo time of RARE improves its contrast-to-noise ratio. Still, sequence modification may be necessary for human application.
The lower limit of the interacquisition interval likely reflects contrast agent saturation. Since the RARE sequence refined in this study is heavily T2 weighted, depiction of the fluid-filled fallopian tube depends on the interaction of the spins of the contrast agent (water) and the surrounding tissue. Although we have attempted to simulate the MR property of the fibromuscular tissue that composes the fallopian tubes and surrounding uterine pelvic musculature, adjustments may be needed when using this technique in women (29).
A minimum quantity of contrast agent was necessary to reproducibly depict both fallopian tubes and bilateral spill. Injection of less than 5 mL of contrast agent per acquisition into a bilaterally patent phantom resulted in opacification of only one fallopian tube with subsequent unilateral spill. In the clinical setting, this could lead to an incorrect diagnosis of unilateral hydrosalpinx. This observation may be a function of differences in relative resistance of the two simulated fallopian tubes as they enter the uterus.
In addition, whereas the phantom's gel exerted a spatially homogeneous force, trapped air and contrast agent from previous experiments, despite rigorous attempts to clear the phantom of air and water after each experiment, may have created selective low-resistance pathways through the uterus that were insurmountable when injecting small amounts of contrast agent. This circumstance may be different in vivo. Specifically, the fallopian tubes and uterus in women are very distensible and may have different patterns of flow and resistance compared with those of our model.
The unilateral hydrosalpinx simulation in experiment 2 demonstrated lack of opacification of the entire obstructed tube, despite the fact that the occlusion was at the distal end. This may reflect limited distensibility of the simulated fallopian tube and the preferential egress of water via the patent fallopian tube. In contrast, the uterus and both fallopian tubes opacified in the bilateral hydrosalpinx simulation in experiment 3. We believe that uterine opacification in our phantom is a function of the inherent distensibility of the plastic reservoir; whereas the opacification of the fallopian tubes, to include the syringes, likely reflects overcoming the resistance provided by the obstructing syringe plungers.
Opacification of the uterus and fallopian tubes to the level of obstruction in experiment 3 parallels the experience with conventional dynamic hysterosalpingography and would be expected during in vivo application. Because in the MR imaging experiment we were able to reproducibly image a water-filled 1-mm fallopian tubal lumen, nonoccluding stenoses on the order of 1 mm may be amenable to detection at MR imaging. Higher spatial resolution may be necessary in vivo to image stenoses smaller than 1 mm.
Extension of our experimental in vitro findings to in vivo human studies has several potential hurdles. The sequence may have to be modified. The use of sterile water may render distinction of free spill from bowel contents and/or free or loculated pelvic fluid problematic. Although the signal intensities of these other entities may differ from that of infused water (eg, higher protein concentration in bowel contents), the dynamic nature of the image acquisition to include sequential image analysis coupled with the subtraction algorithm should limit this problem. In addition, because MR hysterosalpingography likely will be an adjunct to standard pelvic MR imaging, MR imaging sequences (eg, breathing-independent half-Fourier RARE) that are performed before MR hysterosalpingography and that reproducibly image bowel and free fluid well may be critical.
Another potential limitation of our study is that we did not duplicate the complex morphology (eg, cilia) and distensibility of the fallopian tubes. Instead, at this preliminary phase of our investigation, we chose to focus on assessing our ability to detect a simulated fallopian tube that approximated the lumen diameter of human fallopian tubes, since lumen size has been a limiting factor in animal models (30,31).
In addition, the pressure and flow characteristics in vivo will change depending on the cause of the obstruction; every possible cause for fallopian tube obstruction is not easily reproducible with a model. For example, a pyosalpinx would be expected to be more viscous and as such may require higher pressures to opacify the fallopian tube. This is beyond the scope of our investigation.
Another limitation is that the 70-mm section thickness, although it provides adequate coverage for the phantom, may not be sufficiently large for clinical applications.Practical applications: Our results for the performance of MR hysterosalpingography in a phantom model are encouraging. We have developed a protocol that reproducibly demonstrated patent fallopian tubes and unilateral and bilateral hydrosalpinges. The limitations of our system notwithstanding, in the future MR hysterosalpingography as an adjunct to conventional pelvic MR imaging may provide comprehensive infertility evaluation without the need for ionizing radiation, anesthesia, or laparoscopic intervention. Further in vivo investigation is warranted.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, S.M.A.; study concepts, S.M.A., D.T.; study design, S.M.A., R.E.F., D.T.; definition of intellectual content, S.M.A.; literature research, R.E.F.; experimental studies, D.T., R.E.F.; data acquisition, D.T., R.E.F.; data analysis, S.M.A., D.T., R.E.F.; manuscript preparation, R.E.F., S.M.A.; manuscript editing, S.M.A., D.T.; manuscript review, D.T.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Unterweger, C. De Geyter, J.M. Frohlich, G. Bongartz, and W. Wiesner Three-dimensional dynamic MR-hysterosalpingography; a new, low invasive, radiation-free and less painful radiological approach to female infertility Hum. Reprod., December 1, 2002; 17(12): 3138 - 3141. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
