The instillation of fluid into the uterus, when coupled with high resolution endovaginal probes, can allow tremendous diagnostic enhancement in an inexpensive, simple, and well-tolerated office procedure.

Sonography was initially a tool of the obstetrician. Early linear array ultrasound equipment had barely enough resolution to localize placenta, find fetal lie, and measure BPD. High resolution endovaginal probes were developed to provide a degree of image magnification that is on par with ultrasound imaging through a low power microscope (sonomicroscopy).1 Structures that could not be appreciated with the naked eye can now be discerned using these probes. Endovaginal ultrasound is being used increasingly in a variety of clinical situations.

Numerous attempts have been made to apply this method of high resolution ultrasound to the diagnosis of patients with abnormal uterine bleeding. If no organic pathology is found in these patients, their bleeding is either anovulatory (premenopausal) or atrophic (menopausal). Studies indicate that biopsy can be avoided in postmenopausal patients with abnormal bleeding when the ultrasound assessment of their endometrial thickness and texture are suggestive of a lack of significant tissue (24 to 5 mm).2,3,4,5

Although the use of fluid enhancement in abdominal ultrasound was described both for uterine and tubal observations,6,7 it was not widely used until more recently.

Evolution of a concept

All those who perform ultrasound intuitively know that "fluid is your friend," although few of us may have stopped to consciously realize it. Take, for example, the imaging of a baby who has polyhydramnios; fetal structure is seen quite clearly. In contrast, in a baby with oligohydramnios, fetal anatomy is poorly visualized.

In a normal, unsuppressed ovary at midcycle, the fluid of the dominant follicle allows for easy recognition of its structure. This is in contradistinction to the recognition of an ovary in a postmenopausal patient, or even that of a patient on oral contraceptive pills, in whom the ovary is seen less distinctly because of the lack of fluid-filled follicles to serve as a sonic marker. Furthermore, this is the reason why we see normal early pregnancy detail so nicely-the normal gestation sac is fluid-filled. It is this realization that has prompted recent investigations into the use of endovaginal probes with fluid instillation for enhanced endometrial assessment, known as saline infusion sonohysterography (SIS).8,9

Technique

The first step in sonohysterography is to perform a palpatory bimanual examination. As with any type of uterine instrumentation, absolute knowledge of whether the uterus is anteverted or retroverted (and if so, how sharply) will only serve to enhance the success and safety of the exam. Often, uterine version can be appreciated by the skilled sonographer. The clinician in obstetrics and gynecology also should be concerned about the presence of uterine tenderness or decreased mobility.

After insertion of a standard speculum, the cervix is cleansed with an antiseptic solution (10% iodine-based solution). A sonohysterography catheter (Cook ObGyn, Spencer, Indiana) is then inserted by grasping it with a ring forceps and gently feeding it through the cervical os. The catheter should be flushed with sterile saline to rid it of small amounts of air which, when first injected, will cause an echogenic artifactual appearance. The speculum is then removed carefully so as not to dislodge the catheter.

The sonohysterography catheter is 25 cm long and, thus, will come out through the introitus despite its remaining in the proper place relative to the uterus. The vaginal probe is then reinserted. A 10-cc syringe, cleared of any air bubbles, is then attached to the catheter.

Scanning is done in a long axis projection. While watching the video monitor, slowly instill sterile saline. The amount of fluid instilled will depend on the image seen on the ultrasound screen. In that long axis projection, the transducer is moved from side to side (i.e. from cornua to cornua). When the uterus has been completely surveyed from cornua to cornua, the transducer is then rotated 90 degrees into a coronal plane. Infuse additional fluid while fanning down towards the endocervical canal and up towards the uterine fundus. In this way, three-dimensional anatomy is recreated, and great care can be taken not to miss any portion of the uterine cavity, as some polyps or hyperplasia/carcinomas may be focal.

Videotaping the procedure may be helpful for review of the examination after the patient has left the office. A detailed report with representative hard copy images also should be produced.

Indications for sonohysterography

Unscheduled uterine bleeding-Diagnosis and treatment of unscheduled uterine bleeding in perimenopausal women, as well as postmenopausal women on hormone replacement therapy, accounts for a great deal of medical care. Invasive procedures for diagnosis have become commonplace, though they are only occasionally therapeutic. For example, dysfunctional uterine bleeding in perimenopausal patients caused by a lack of ovulation is best treated hormonally, not surgically.

Diagnostic procedures that can be performed in the office include biopsy, suction aspiration, suction curettage, and diagnostic hysteroscopy. Patients with abnormal uterine bleeding also may have curettage performed, with or without diagnostic or operative hysteroscopy, in an operating room setting. Keep in mind that the amount of time required, the equipment used, and the skills that the operator must possess will vary with each diagnostic conclusion, as it does for evaluation of endometrial polyps, submucous myomas, hyperplasia, and carcinoma. Treatment options also vary per patient; some patients with submucous myomas are inappropriate candidates for resectoscopic surgery, whereas in others it would be the treatment of choice.

In one prospective pilot study, saline infusion sonohysterography was performed in 21 women who had unexplained perimenopausal uterine bleeding.10 Obvious polypoid lesions were found in 8 of the 21 patients; all underwent triage for operative hysteroscopic removal. The pathology report confirmed that the polyps were benign in all 8. Submucosal myomas were found in 3 patients, two of which were treated with wire loop resectoscopic excision. The third, who had a submucous myoma that extended to the serosal edge of the uterus, received expectant management. Nine patients had no obvious anatomic lesion, and their endometrial wall thickness (either anterior or posterior) was a maximum of 3.2 mm. In this study, sonohysterography was purposely performed on day 4 to 6 of the bleeding cycle, which is when one would expect early proliferative change if no anatomic abnormality existed. Biopsy in all 9 of these patients did reveal early proliferative endometrium.

A diagnosis of dysfunctional (i.e. anovulatory) uterine bleeding was made, and all 9 women were successfully treated with progestational agents. The final patient in the study had an endometrial thickness of 7.6 mm along the anterior wall, although the posterior wall was thin (2.3 mm). Curettage with hysteroscopy revealed simple hyperplasia without atypia, and this patient also was treated with progestational agents.

As a result of these findings, it was concluded that endometrial fluid instillation (sonohysterogram) to enhance vaginal ultrasonography in perimeno-

pausal women could reliably distinguish between patients with minimal tissue (24 mm single layer) whose bleeding is anovulatory and best treated hormonally from those patients with significant tissue (>4 mm single layer) who are in need of formal curettage and hysteroscopy. Furthermore, it was found that polyps can be distinguished from submucous myomas, a discovery that will allow appropriate triage for operative hysteroscopy in terms of skill, length of time, and equipment required. The need for diagnostic hysteroscopy in patients whose bleeding is dysfunctional is, therefore, eliminated.

Based on these results, it seems apparent that any "blind" endometrial sampling should be preceded by a fluid-instillation sonohysterogram. To justify such a blind procedure, a process must be shown to be symmetrically "pan uterine." When changes are focal, as is the case in polyps, some hyperplasias, and some carcinomas, they can be appreciated as such with fluid-instillation sonohysterography, and then directed biopsies can be carried out (figure 1).

Other considerations in SIS

Myomas co-existing with dysfunctional bleeding-Although 9 of 21 patients in the pilot study10 had obvious sonographic and clinical evidence of fibroids, only 3 had a submucous component. Six of 21 had intramural subserosal myomas co-existing with dysfunctional uterine bleeding.

Myoma versus polyp-Usually polyps are clearly discernable on sonohysterogram, as are submucous myomas. However, sometimes a broad-based polyp will be difficult to distinguish from a submucosal myoma. This distinction is important for preoperative triage: A truly pedunculated submucous myoma will behave more like a polyp in terms of the skill and equipment required for its removal in the operating room, while a broad-based polyp may behave more like a myoma and require resectoscopic capabilities.

In addition, a non-fluid-enhanced ultrasound picture may demonstrate a definite myoma that, at first, appears to be submucosal; when it is viewed with sonohysterography, however, it will be seen to be merely distorting the endometrial cavity, though it is clearly more intramural than submucosal.

Endocervical polyps-Usually endocervical polyps can be visualized on speculum examination. However, occasionally they may distort the endocervical canal when imaged with vaginal probe ultrasound. Saline infusion sonohysterography often can accurately diagnose a polyp's presence and size, and in some instances the attaching stalk also can be visualized.

Three-dimensional reconstruction-All frozen images are two-dimensional pictures. As the uterine cavity is a three -dimensional structure, any one view in one projection may not be representative of the total picture and in fact may actually be misleading. It is, therefore, imperative to reconstruct three-dimensional anatomy with multiple images at right angles to each other.

Unusual uterine findings-Another example of the advantage of saline infusion sonohysterography is its ability to exquisitely detail the endometrium

and its relationship to the proximal myometrium. In some patients who are on tamoxifen, traditional imaging will display centrally-thickened, bizarre uterine echoes which should not be assumed to be endometrial. SIS has demonstrated that some such changes are associated with microcystic changes in the proximal myometrium while the endometrium remains thin and atrophic.11 It was determined that these changes represent reactivation of the foci of adenomyosis due to the paradoxical estrogen effect of tamoxifen in the genital tract of some patients.12 This also will be the case in situations where one wants to characterize the endometrium that appears to have an irregular texture (figure 2).

Potential pitfalls and concerns

Anesthesia analgesia-The use of anesthesia or analgesia in SIS is not generally required, as the procedure has been found to be well tolerated; the overwhelming majority of patients have reported no evidence of pain, and minimal cramping has been reported only in a small minority of cases. The catheter used in the infusion of sterile saline is small-only 1.8 mm in diameter-and is remarkably painless upon insertion. I have seen only one case of a vasovagal response similar to a plastic IUD insertion in a nulliparous patient. This case involved a patient with submucous fibroids and an endocervical polyp.

Risk of infection-In most respects, saline infusion sonohysterography should be handled similarly to traditional hysterosalpingography. Thus, the decision about whether to obtain gonorrhea or chlamydia cultures, as well as whether to use antibiotics, will depend very much on the patient population that you normally deal with. In my experience in performing the procedure, I have not needed to routinely obtain cultures for sexually transmitted diseases, nor have I routinely used prophylactic antibiotics. In the more than six hundred cases I've overseen, no patients have experienced any infectious morbidity.

Inability to thread the catheter-The currently available catheter is very flexible and has a single opening at the tip. Because of its flexibility, it may occasionally be difficult to thread the catheter into the desired position. It also comes in a less-flexible material, depending on operator preference. If a problem occurs, use your free hand to change the position of the speculum; this will often sufficiently modify the angle of the cervix with the fundus to allow successful completion of the insertion. As a last resort, a tenaculum may be used. Along these lines, some patients may have better results using a catheter that has a stylet.

Inadequate distension of the cavity-As in hysteroscopy, some cavities are more difficult to distend then others. Patients with a patulous cervix may experience a great deal of transcervical fluid loss. Other patients may have fluid leaking out through the fallopian tubes, even when slow injection and minimal pressure are used. To account for these anomalies, first check the position of the catheter. Look to see that its acoustic shadow extends most of the way to the uterine fundus. Keep in mind, though, that unlike hysteroscopy (which requires distension for visualization), this procedure requires very little fluid to outline the cavity. Even a small ribbon of fluid should act as a sufficient interface in order to distinguish anterior and posterior endometrial surfaces and outline endometrial pathology.

Spreading adenocarcinoma into the peritoneal cavity-This is a question of the benefit outweighing any theoretical risk. It is no longer standard practice to tie the fallopian tubes with silk prior to a total abdominal hysterectomy bilateral salpingo-oophorectomy for endometrial carcinoma. Furthermore, as hysteroscopy with saline or other distending media would have the same theoretical concern, use of the SIS procedure is justified. Also, survival rates of patients with endometrial carcinoma who underwent standard hysterosalpingography were not different between patients who demonstrated intraperitoneal spill of the contrast medium and those who did not.13

Saline infusion sonohysterography enhances endovaginal ultrasound examination of the uterine cavity. It is easily and rapidly performed at minimal cost. The procedure is extremely well tolerated by patients and virtually devoid of complications. Furthermore, it may eliminate the need to use invasive diagnostic procedures in some patients, and can optimize the preoperative triage process for those patients who will require therapeutic intervention. AR

References

1. Goldstein SR: Incorporating endovaginal ultrasonography into the overall gynecologic examination. Am J Obstet Gynecol 162:625-632, 1990.

2. Goldstein SR, Nachtigall M, Snyder JR, Nachtigall L: Endometrial assessment by vaginal ultrasonography before endometrial sampling in patients with postmenopausal bleeding. Am J Obstet Gynecol 163:119-123, 1990.

3. Granberg S, Wikland M, Karlsson B, et al: Endometrial thickness as measured by endovaginal ultrasound ultrasonography for identifying endometrial abnormality. Am J Obstet Gynecol 164:47-52, 1991.

4. Osmers R, Volksen M, Schauer A: Vaginosonography for early detection of endometrial carcinoma? Lancet 335:1569-1571, 1990.

5. Nasri MN, Coast GJ: Correlation of ultrasound findings and endometrial histopathology in postmenopausal women. Br Med J 96:1333-1338, 1989.

6. Randolph JR, Ying YK, Maier DB, et al: Comparison of real-time ultrasonography, hysterosalpingography, and laparoscopy/hysteroscopy in evaluation of uterine abnormalities and tubal patency. Fertil Steril 46:828-832, 1986.

7. Lewit N, Thaler I, Rottem S: The uterus: A new look with transvaginal sonography. J Clin Ultrasound 18:331-336, 1990.

8. Parson AK, Lense JJ: Sonohysterography for endometrial abnormalities: Preliminary results. J Clin Ultrasound 21:87-95, 1993.

9. Syrop C, Sahakian V: Transvaginal sonographic detection of endometrial polyps with fluid contrast augmentation. Obstet Gynecol 79:1041-1043, 1992.

10. Goldstein SR: Use of ultrasonohysterography for triage of perimenopausal patients with unexplained uterine bleeding. Am J Obstet Gynecol 170:565-570, 1994.

11. Goldstein SR: Ultrasonographic appearance of the uterus in patients receiving tamoxifen. Am J Obstet Gynecol 170:447, 1994

12. Perrot N, Guyot B, Antoine M, Uzan S: The effects of tamoxifen on the endometrium. Ultrasound Obstet Gynecol 4:83-84, 1994.

13. DeVore GR, Schwartz PE, Morris J: Hysterography: A 5-year follow-up in patients with endometrial carcinoma. Obstet Gynecol 60:369-372, 1982.