Chapter 38: SONOHYSTEROGRAPHY AND SONOHYSTEROSALPINGOGRAPHY

Gynecologic problems for which patients seek care include excessive or ill-timed bleeding, amenorrhea, pelvic pain, pelvic masses, and infertility. Transvaginal sonography enhances the diagnosis of problems beyond the external genitalia and cervix that are not visible on physical examination. One further refinement of this approach to gynecologic diagnosis is the use of sonographic contrast media. Clear fluids improve the detection and localization of soft tissue abnormalities in the uterine and pelvic cavities. Echogenic contrast media, with or without Doppler enhancement, allow acceptable demonstration of tubal patency, according to the limited data available. Sonohysterography (SHG) of the uterine cavity and sonohysterosalpingography (sonoHSG) of the tubes are informative variations of hysterosalpingography (HSG), a standard radiographic technique for studying the reproductive lumina outlined by transcervical infusion of iodinated contrast under fluoroscopic observation. When sonographic evaluation of the uterine lumen with contrast is combined with evaluation of the tubes, this procedure can be termed sonohysterosalpingography, or sonoHSG. This use of "double contrast" involves transcervical injection of saline to outline the uterine cavity, which then pools in the posterior cul-de-sac before instillation of positive (echogenic) contrast. The tubal lumen is thus outlined and the fimbriated end is evaluated for spill and its relationship to the ovary.

Contrast-enhanced transvaginal sonography eliminates exposure to unmetered fluoroscopic ionizing radiation and iodinated contrast, while obtaining detailed structural information. It can be performed in a small examining room. A standard gynecologic examining table with stirrups and a retractable basin are required. A single assistant is required, and the procedural equipment is readily available and inexpensive. A foot pedal for the ultrasound equipment is desirable.

Sonohysterography has also been called hysterosonography and saline infusion sonohysterography (SIS). Most importantly, the term hysterosonography has been designated a Current Procedural Terminology (CPT) code for billing. Sonohysterography can be confused with HSG, resulting in improper ordering and scheduling.

Historic Development Using Transabdominal Technique

The ease of discerning small details of the intrauterine contents in the presence of amniotic or spontaneous uterine fluid is well known. Early efforts to supply artificial fluid in the nonpregnant woman using transabdominal scanning were effective but cumbersome. In 1984, Richman et al reported the instillation of 70% dextran through a rigid uterine cannula during transabdominal sonographic observation before standard HSG in 34 patients.¹ Their interest in the uterine cavity was limited to the observation that tubal obstruction produced sustained expansion of the uterine cavity. The observation of the accumulation of peritoneal fluid in 25 of 34 patients correctly identified at least unilateral tubal patency with an accuracy of 97%.

Randolph et al described the results of a similar approach using sterile saline as the medium to predict surgical findings in anesthetized women about to undergo laparoscopy or hysteroscopy in 1986.² They accurately described 53 of 54 uteri, and the pathologies described included unicornuate uteri, septae, polyps, and intracavitary or submucous myomata. Only a unicornuate uterus with a noncommunicating horn was mistaken for a small normal uterus, providing 98% sensitivity and 100% specificity regarding intrauterine abnormalities. They also found that identification of fluid accumulation in the posterior cul-de-sac reliably indicated at least unilateral patency with a sensitivity of 100% and specificity of 91%. Tube-specific accuracy, however, was poor. They required 100 cc of fluid for visualization of the fimbria of the tube in the cul-de-sac, and they noted that isolated accumulation of fluid above the fundus indicated adhesions or a mass obliterating the cul-de-sac. Easily identified hydrosalpinges were observed to form with distal obstruction.

Other small series supported the use of abdominal ultrasound and the observation of accumulation of cul-de-sac fluid as a simple screening technique for at least unilateral tubal patency, but made little comment on the uterine cavity.^1,3

In Belgium, however, van Roessel et al scanned 30 uteri distended with dilute dextran 70 during hysteroscopy and compared the ultrasonic and endoscopic diagnoses.⁴ Despite procedural artifacts, only a 2-mm polyp was missed; 8 normal cavities were accurately assessed as normal. Three myomata (only 1 intracavitary), 3 septae, 2 cases of synechiae, and 6 polypoid lesions were predicted. Unusual "ragged" endometrium was noted in 7: 1 was due to cancer, 3 were due to hyperplasia, and 3 were normal proliferative endometrium. Ultrasound with contrast allowed them to identify both polypoid lesions and thick, irregular endometrium. Two intrauterine devices (IUDs) were located precisely, including one embedded in the myometrium that could not be seen by the hysteroscopist. Three uterine septae were diagnosed, and resection was monitored successfully by ultrasound rather than laparoscopy in 2 of the cases. Imaging of another cancer failed due to the inability to maintain uterine distention. Abnormalities were identified in 10 of 21 apparently normal uteri by the use of cavitary distention, and precision of diagnosis was enhanced in 5 of 9 uteri that were abnormal with conventional ultrasound.

Limitations of Transvaginal Sonography

The widespread use of high-frequency transvaginal sonography beginning in the late 1980s brought exciting detail to images of pelvic organs. Cyclic endometrial changes and concepts of normal ranges of thickness had already been described in transabdominal studies.^{5,6, and 7} Extreme distortion and pathologic thickening due to cancer or hyperplasia had also been identified, as had uterine anomalies and thick synechiae.^{8,9, and 10} The detailed magnified images of endometrium and ovaries that are routinely obtainable with transvaginal scanning allow accurate diagnosis of both hormonal and structural causes of abnormal bleeding,^11,12 amenorrhea,^{13,14, and 15} and infertility.¹⁶

Even with high-frequency transducers, the cause of an abnormal endometrial image is not always apparent. Fedele et al found transvaginal ultrasound comparable to hysteroscopy in detecting submucous myomata, but had difficulty distinguishing between myomata and polyps.¹⁷ Typically hyperechoic, endometrial polyps are best seen during the proliferative phase. However, they may be indistinguishable against a background of secretory endometrium, which is typically echogenic. Relatively hypoechoic myomata are well imaged when outlined by secretory endometrium, but can obscure the endometrium entirely by acoustic shadowing. Thick synechiae and cavitary anomalies are best imaged in the periovulatory or secretory endometrium. Intrauterine synechiae may be highlighted by sequestered hematometra, but often the endometrial deficiency is subtle. Distinction between an anomaly, such as a T-shaped uterus, and synechiae can be difficult.

Dodson evaluated 45 women with abnormal bleeding and imaged structural abnormalities in one-third of them.¹² He used the ovarian and endometrial appearance to confirm functional ovulatory abnormalities in the rest. Three ovulatory patients with appropriately thick hyperechoic endometrium had intermenstrual bleeding that eluded diagnosis, suggesting they might have had polyps obscured by secretory endometrium.

Narayan and Goswamy recommended transvaginal ultrasound for screening in infertility patients, having diagnosed abnormalities correctly in 182 of 193 women (94%).¹⁶ After weekly evaluation of 200 patients throughout their cycles followed by hysteroscopy, they found that periovulatory multilayered endometrium best revealed intracavitary lesions and the uterotubal junctions. Fourteen patients, however, had nonspecific irregularities on ultrasound that required diagnostic hysteroscopy for definition. On hysteroscopy 2 had normal cavities, 4 had polyps, 2 had synechiae, 1 had a septum, and 3 had endometritis. Furthermore, only 44 of 51 submucosal myomata seen by ultrasound were truly intracavitary. Seventy-one of seventy-six synechiae were seen by ultrasound, as were 43 of 46 polyps, and 5 of 6 septae. There was a 5.5% false-positive rate. These are the cases that will typically benefit from saline infusion for better definition before either excluding or undertaking therapeutic surgery.

The atrophy of the endometrium in postmenopausal women makes evaluation of their endometrium straightforward in the absence of distorting leiomyomas or adenomyosis. Measurement of endometrial thickness is an efficient way to exclude endometrial neoplasia. A meta-analysis of endometrial thickness in postmenopausal women by Smith-Bindman et al included 35 studies and over 3500 women. Use of the 5-mm threshold for endometrial thickness resulted in a 96% sensitivity for detection of endometrial cancer.¹⁸ Symmetrical endometrium of 4 to 5 mm decreases the a priori risk of endometrial cancer in bleeding postmenopausal women by 90%, from 10% to 1% in untreated women and from 1% to 0.1% in women with hormone replacement therapy (HRT).^18,19

A single report of 3 cancers that were not detected by 7-MHz vaginal ultrasound in 3 uteri with sonographic double-wall measurements of 3.3 and 2 mm, respectively, did not include gross measurement, sonographic images, or histologic description of the cancers.²⁰ Hence, one must suspect an error in technique or inadequately imaged endometrium due to distortion or uterine position, rather than invisible endometrial cancers in atrophic endometrium. The endometrial cavity is a three-dimensional (3D) structure. Pathologic changes may be focal. A single "frozen" long-axis view may miss the abnormality. Care must be taken to recreate the 3D anatomy from multiple views. Efficient acquisition of this complete cavitary assessment is an advantage of 3D sonography.

A scanty or atrophic biopsy specimen from a thickened endometrium may be due to inadequate sampling of focal areas. This may be due to inadequate sampling of focal areas of hyperplasia, for instance, using a 2-mm piston-type suction device²¹; to compression of a fibrous polyp in an atrophic endometrium²²; to distortion by myomata; or to the appearance of myometrial changes surrounding an otherwise normal cavity as has been seen in selected patients on tamoxifen therapy.²³

Hence, conventional transvaginal ultrasound is quite sensitive, but is not as specific as direct examination of the endometrial cavity with hysteroscopy, the gold standard for the detection of endometrial abnormality.²⁴ SHG aids in the sonographic identification, localization, and description of intracavitary lesions.

Transvaginal Sonohysterography

Investigators around the world have independently developed methods to enhance transvaginal images of the cavity with intrauterine fluid, usually sterile saline, instilled through a fine catheter during scanning. In 1988, Deichert et al reported their pioneering observations in the German literature on the effects of fluid in the uterine cavity, but then turned to the study of positive contrast agents for tubal patency.^25,26

In 1991, Mitri et al in South Africa, using an 8 French Foley catheter in the cervix, demonstrated that SHG was more informative than conventional HSG.²⁷ Fifty normal cavities were demonstrated by both methods, but 9 women had otherwise undetectable extracavitary myomata visible by ultrasound. All 4 submucous myomata were seen by both methods. Four bifid uterine cavities were tentatively diagnosed as septate by HSG, and were definitively diagnosed as harboring 3 septae and 1 intramural myoma by ultrasound. The 2 methods agreed in one case of intrauterine synechiae. One HSG failed due to intravasation of contrast, whereas SHG and laparoscopy demonstrated a normal uterus and bilateral hydrosalpinges.

Bonilla-Musoles et al in Spain found SHG to have a sensitivity of 96%, specificity of 97%, positive predictive value (PPV) of 96%, and negative predictive value (NPV) of 97%.²⁸ Twenty-two normal women, 16 women with menometrorrhagia, and 16 infertile women were studied. Two cases of focal hyperplasia were missed, and 1 normal uterus was thought to contain hyperplasia, in a total of 74 patients. There were 2 failures due to pain and stenosis, respectively. Submucosal myomata, hyperplasia, synechiae, and a septum were correctly diagnosed, and they considered SHG equal to hysteroscopy except for the detection of tiny focal hyperplasia. They suggested it be used for preoperative and posttreatment evaluation and monitoring.

Syrop and Sahakian diagnosed, measured, and correctly located polyps in 14 infertility patients with abnormal images noted during sonographic screening before in vitro fertilization, using a rigid Rubin cannula and Ringer lactate or human tubal fluid.²⁹

Parsons and Lense reported 100% detection of intracavitary abnormalities in 39 patients with abnormal bleeding and abnormal endometrial images, which was confirmed by hysteroscopy or hysterectomy.³⁰ Polyps, myomata, synechiae, and irregular thickening that proved to be hyperplasia or cancer were accurately identified. Two small, raised areas were seen that proved to be polypoid islands of normal proliferative endometrium on hysteroscopic biopsy. Cancer could not be distinguished from hyperplasia, because both produced irregularly thickened endometrial surfaces. Twenty infertile patients judged to have normal cavities by SHG demonstrated normal cavities with hysteroscopy, HSG, or both.

Goldstein described accurate results of SHG in 21 postmenstrual women with abnormal bleeding. He found 11 focal lesions and confirmed them with hysteroscopy: 8 polyps and 3 submucous myomata.³¹ Nine of the other ten patients had early proliferative changes on biopsy, dilation and curettage (D&C), or hysteroscopy. Hyperplasia was confirmed in the remaining patient with an inappropriately but symmetrically thick endometrium, and atrophy was confirmed in the rest. Sonohysterography was also used to demonstrate normal cavities in 5 patients undergoing sonographic monitoring during tamoxifen treatment.²³ Despite what appeared to be thickened, honeycombed endometrium, curettage and hysteroscopy yielded scant tissue in one. Infusion confirmed that the abnormality appeared to be myometrial with symmetrical unstimulated endometrium confirmed by biopsy in every case.

Farquar et al performed a systematic review of transvaginal sonography, sonohysterography, and hysteroscopy for investigation of abnormal uterine bleeding.³² The pooled likelihood ratio for diagnosis of submucous fibroids by SHG was 29.7. This was similar to the likelihood ratio for hysteroscopy.

Dueholm et al evaluated 189 women with transvaginal sonography and sonohysterography, comparing the sonographic findings to that found at hysteroscopy or hysterectomy.³³ The sensitivity for the diagnosis of polyps in submucous myomas in women with abnormal bleeding was 99%, with a specificity of 72%. The authors concluded that sonohysterography was a sensitive tool, superior to unenhanced transvaginal sonography for the evaluation of abnormal uterine bleeding. Similarly, Krampl et al compared transvaginal sonography, SHG, and operative hysteroscopy in the evaluation of abnormal uterine bleeding in 100 consecutive patients.³⁴ They found that the detection rate of focal intrauterine pathology using SHG was significantly higher than using transvaginal sonography alone. Visual examination by hysteroscopy yielded no additional information to the detection of focal lesions than that obtained at SHG. Even in women with abnormal uterine bleeding with endometrial thickness of less than 5 mm, up to 20% will have focal intracavitary lesions seen by SHG but not visualized by unenhanced transvaginal sonography.³⁵

Indications for Sonohysterography

The most common indication for SHG is to determine the cause of abnormal uterine bleeding. The main contraindication for SHG is the presence of a known or suspected viable intrauterine pregnancy. In most cases conventional sonography will provide a rapid diagnosis when performed in the periovulatory period: Dysfunctional bleeding due to disorganized hormone production is confirmed by the lack of evidence of impending ovulation and lack of a periovulatory multilayer endometrium. When the endometrium is asymmetrical, unexpectedly thickened, or poorly imaged, SHG will clarify the anatomy. Although clots should be evacuated to avoid confusion, SHG often provides an interpretable image despite uterine bleeding.

Although the diagnosis of endometrial masses requires histologic examination, SHG allows one to make the distinction between global and focal processes, an important consideration before office biopsy or D&C. Small, flexible Pipelle-type biopsy instruments are very popular with physicians and patients despite their tendency to sample less than 5% of the endometrial cavity.²¹ When a global process is identified, this device is adequate. When there is a focal lesion or asymmetry, direct visualization of the endometrial cavity with resection of masses via hysteroscopy guidance is indicated. An inexplicably abnormal endometrial image (by computed tomography, magnetic resonance imaging [MRI], HSG, or ultrasound) due to thickness or distortion, particularly in an asymptomatic individual, can be conveniently assessed to plan further intervention only where it is warranted.

Secondary amenorrhea or hypomenorrhea is infrequently caused by endometrial synechiae due to trauma or infection in a poorly estrogenized endometrium. These are easily demonstrated by SHG, enabling hysteroscopic resection if pregnancy is desired or observation if it is not.

Evaluation of the endometrial cavity is a standard part of the workup for infertility, usually by HSG. The use of hysteroscopy has increased as it has become clear that there is a high incidence of intracavitary abnormalities in infertile patients.^36,37 HSG, however, produces a high rate of false-positive uterine findings, reported as 30% by Valle, prompting many unnecessary procedures.³⁸ Sonographic screening before undertaking in vitro fertilization is increasingly practiced. With the use of SHG to further define suspected abnormalities, hysteroscopy can be reserved for those requiring surgical repair or directed biopsy, rather than as an expensive diagnostic tool. Anomalies, synechiae, polyps, submucosal myomata, and hyperplasia due to prolonged anovulation have all been demonstrated in reported series and in our patients using sonographic contrast, often after years of infertility treatment. As widespread experience in the use of sonographic contrast for tubal patency increases, unnecessary exposure of the gonads to radiation may become limited to special cases.

The diagnosis of müllerian anomalies is aided by sonohysterography. Alborzi et al studied 20 patients with a history of recurrent pregnancy loss with the HSG diagnosis of a septate or bicornuate uterus.³⁹ Surgical confirmation of the diagnosis was available in 10 cases. SHG was able to differentiate between septate and bicornuate uteri in all cases. The authors concluded that sonohysterography alone was sufficiently accurate to proceed directly to hysteroscopic septum resection without performing a laparoscopic examination. Valenzano et al examined 54 patients with infertility or recurrent miscarriage, and with a clinically or sonographically suspected abnormal uterus.⁴⁰ All subjects had an HSG and SHG, and in those with suspected anomalies, hysteroscopy or laparoscopy was performed. SHG was able to detect all anomalies accurately. The authors utilized defined criteria for diagnosis of uterine malformations by SHG as defined in Table 38-1.

The location and measurement of intracavitary versus submucosal myomata should be mapped preoperatively before undertaking hysteroscopic resection. As described by Narayan and Goswamy, a truly submucosal myoma may appear to distort the cavity on conventional sonography, but it may not be evident through the hysteroscope.¹⁶ The depth of the intramural component of a submucosal myoma can also be appreciated preoperatively. Occasionally, synechiae are difficult to distinguish intraoperatively, especially in the cornua and along the walls. The postoperative effects of all hysteroscopic procedures, including endometrial ablation, are evaluable with SHG. It is an important method with which to evaluate the effects of new procedures and treatments.

The now common use of transvaginal ultrasound for screening for endometrial neoplasia or hyperplasia in postmenopausal women uncovers some women with thin but asymmetrical or poorly imaged endometrium. Delineation of the cavity with SHG eliminates or demonstrates the need for biopsy in these few unclear instances. SHG can demonstrate a smooth endometrial surface and symmetrical thickness that are as reassuring in postmenopausal women as they are in younger cycling women. The Society for Radiologists in Ultrasound consensus statement on use of ultrasound for evaluation of abnormal bleeding indicates that transvaginal sonography and sonohysterography may be used in the initial evaluation of postmenopausal bleeding (Figure 38-1).⁴¹

Women undergoing treatment with tamoxifen constitute a special category. The unpredictably agonistic effects of this estrogen receptor blocker stimulate endometrial proliferation, polyp formation, and possible cancer in some women. Sonohysterography is a useful technique for evaluation of patients on tamoxifen. The endometrium may have a bizarre cystic appearance on unenhanced transvaginal sonography examination. However, it is actually the subendometrial region that appears abnormal and the endometrium itself is often thin and atrophic. In a study of 48 consecutive tamoxifen-treated women, the authors found that sonohysterography aids in the diagnosis of endometrial abnormalities even if prior endometrial biopsies were negative.⁴² Surgery was avoided in 14% of the subjects when the SHG revealed a normal endometrium or subendometrial cysts. Markovitch et al studied the value of sonohysterography in screening asymptomatic postmenopausal tamoxifen-treated breast cancer patients.⁴³ In 85 patients with an endometrial thickness of greater than 8 mm the authors found that sonohysterography improves the accuracy of diagnosis of an intrauterine mass. SHG was found to be a highly specific technique for detection of abnormalities in tamoxifen-treated patients.

The location of an unextractable IUD among myomata or embedded in the uterine wall can be verified, and operative hysteroscopy can be planned using SHG.

Contraindications to Sonohysterography

There are few contraindications to SHG.⁴⁴ These include the following:

1. Known or suspected pregnancy

2. Infection or pelvic pain suggestive of pelvic infection

3. Intractable cervical stenosis

Equipment

A few common items are needed in addition to a gynecologic exam table equipped with stirrups, a basin, and adjustable head height, and an ultrasound machine with a vaginal probe. Also needed are a speculum and cervical cleanser with swabs, a sterile long packing or ring forceps, a catheter, injectable-grade sterile saline, a 10-cc sterile syringe for small uteri or a 40-cc sterile syringe for lar-ger ones.

To allow atraumatic pelvic manipulation by the ultrasound probe, catheters must be flexible, narrow, and at least 25-cm long. Any sterile straight catheter with a diameter of 2 mm can be used, for instance, the Tampa catheter, which has an introducer (Ackrad Labs, Inc, Cranford, NJ, USA); or a Soules intrauterine insemination catheter (Cook Co, Spencer, IN, USA); or a 25-cm, 5.3 French atraumatic catheter. There is a movable silastic marker collar 7 cm from the tip. This is for small uteri with normal internal ostia and can be used in about 90% of cases. A plastic premature infant feeding tube (Davol Inc, Cranston, RI, USA); 5 French catheter, 38-cm long; or any other narrow, long, flexible catheter that will accept a syringe may be used. The Davol catheter has a mark 8 inches from the tip. A 2-mm semiflexible plastic biopsy curet, such as a Milex Exploracuret (Milex Products, Inc, Chicago, IL, USA), is a little more uncomfortable but will work as well if biopsy is anticipated. Straight catheters allow discharge of the fluid through the cervix and into the basin, which prevents overdistention, tubal reflux, and cramping.

In women with patulous cervices, uteri longer than 8 cm, or suspected synechiae, a balloon catheter is required for adequate distention. An H/S catheter (Ackrad Co, Cranford, NJ, USA) is a 5.3 French flexible hysterosalpingogram catheter with a 3-cc latex or spindle-shaped plastic balloon. This catheter is also recommended for infusion for evaluation of tubal patency and intraperitoneal disease. An 8 French pediatric Foley catheter with a 3-cc balloon can also be used successfully in lieu of the H/S catheter. If Asherman syndrome is suspected, it will allow adequate separation of the tethered uterine surfaces. These balloons can also be used to gently dilate the cervical canal so that larger biopsy instruments can be passed.

The Goldstein Sonobiopsy Catheter (Cook Medical Inc, Bloomington, IN, USA) is 7.2 French, and 26-cm long (Figure 38-2). Instead of a balloon tip, it has an acorn-shaped silicone positioner that is placed against the cervix to prevent fluid backflow.

Dessole et al compared 6 different types of catheters for SHG (Figure 38-3).⁴⁵ They assessed reliability, physician's ease of use, time required for insertion of the catheter, volume of contrast medium used, tolerability for the patient, and the cost. No statistically significant differences were found among the catheters. The Foley catheter was the most difficult to use and required the most time to position correctly. The Goldstein catheter was best tolerated by the patients. The Foley catheter was the least expensive.

Technique

As with hysteroscopy, lesions arising from the endometrium are best defined against minimally stimulated or postmenstrual endometrium. A history is taken, and the procedure is explained, at which time a rapid urine human chorionic gonadotropin assay is done if there is a strong likelihood of pregnancy. The patient is examined in the dorsal lithotomy position. A baseline transvaginal ultrasound exam is done, during which there is careful evaluation of the uterus, ovaries, and tubes for morphology, mobility, and pain. The posterior cul-de-sac is inspected for free fluid.

The speculum is then introduced in the vagina and the external cervical os is visually identified. The speculum can be manipulated to optimally expose and immobilize the cervix.

The vagina is then inspected for sources of bleeding, and the cervix for yellow mucus that would suggest Chlamydia infection. If pain or cervical discharge suggests pelvic infection, the procedure is delayed until diagnosis and therapy have been accomplished.

The cervix is cleaned with iodine solution or other antiseptic. The catheter is then slowly advanced into the cervix. The catheter is threaded through the cervix with a sterile ring or packing forceps. The catheter should be flushed with saline before insertion to decrease the echogenic artifact caused when air is initially injected into the cavity. A straight catheter should be advanced until its tip is in the fundus; a balloon catheter should have the balloon placed in the midcervix and gently inflated with fluid (not air) until meeting resistance. While the catheter is held in place with forceps, the speculum is carefully removed and the vaginal probe is reintroduced above the catheter anterior to the cervix with an anteverted uterus, or below it and posterior to the cervix with a retroverted uterus. Video recording allows review for more careful analysis of details.

Optimal balloon catheter placement during SHG has been studied. Spieldoch et al randomized women undergoing SHG to intracervical or intrauterine balloon catheter placement. There was significantly less pain with intracervical balloon placement.⁴⁶ Significantly, less saline was required with intracervical catheter placement.

A small (50- to 250-cc) bag of intravenous fluid can be hung in the lab and used as needed if sterile technique is used for saline withdrawal into a 10- to 40-cc syringe. The syringe is attached to the catheter, and the saline is slowly infused while scanning the uterus longitudinally, fanning from cornu to cornu. The transducer is then turned 90 degrees, and the uterus is scanned in a transverse fashion from the external os of the cervix to the fundus during infusion, while the cavity is examined for asymmetry or unusual shape.

Each cornu is identified and the uterotubal junction and interstitial part of the tube are sought to confirm a normally shaped cavity. A three-dimensional mental image of the cavity is thus assembled, and lesions are precisely located. Rapid cervical expulsion of the saline can be partly prevented by pressure with the probe on the internal os while scanning longitudinally. A cervical balloon may be required if distention cannot be maintained. Two to 8 cc in the cavity is enough to outline most lesions; unlike the case in hysteroscopy, overdistention of the uterus is not required. Even a small rim of fluid will serve as an interface to allow cavitary assessment. Repetitive injection of small amounts of fluid using a 40- to 60-cc syringe is not difficult if one prefers not to use a balloon, and for this an assistant is very helpful. It is important to deflate the balloon under direct visualization to evaluate the lower uterine cavity and cervix at the end of the procedure so as not to miss pathology in these regions potentially obscured by the balloon.

Significant imaging artifacts preventing accurate interpretation may arise if a substantial amount of air is inadvertently introduced into the lumen prior to instillation of saline. If the lumen is filled with air, the procedure may need to be aborted and repeated the next day, allowing for air to escape from within the lumen. Occasionally, the injected air can be aspirated successfully.

Directed Biopsies

If biopsy is being considered, 10 cc of 1% lidocaine can be infused as the distending medium in nonallergic individuals. Guney et al randomized patients to receive intrauterine lidocaine or saline at the time of SHG. They found modest reductions in pain scores both immediately after the procedure and 20 minutes later.⁴⁷ Residual intrauterine fluid should be aspirated, if possible, before withdrawing the catheter if a Pipelle curet with limited volume will be used for biopsy. Nicoletti et al developed a new technique to obtain directed biopsy during sonohysterography.⁴⁸ They developed the NiGo forceps, which has a length of 35 cm and an outer diameter of 7 French (Figures 38-4 and 38-5). An irrigation channel allows for expansion of the uterine cavity with distention media. The biopsy forceps permits taking tissue samples of approximately 6 to 8 mm maximum diameter. They performed SHG and hysteroscopy on 18 patients. SHG was successful using the NiGo device in 15 of the 18 women and an endometrial sample was obtained in 14 of these patients. In one patient endometrial biopsy provided too little tissue for pathologic evaluation. All adequate samples obtained with the device were identical to those obtained in hysteroscopy. It also should be noted that the office-based hysteroscopy procedure was not successful in 2 of the patients.

Leone et al performed endometrial sampling on 128 patients with a diffusely thickened or inhomogeneous endometrium at the time of sonohysterography.⁴⁹ A 14 French urinary catheter was used for the SHG and also to perform directed sampling of the endometrium. They compared the results from the SHG-directed sampling to that obtained at office hysteroscopy. The median area of endometrial specimens obtained by SHG was not significantly different than that obtained at hysteroscopy. There was good correlation in pathologic diagnoses between the samples obtained using the 2 different techniques. Another study examined the use of the Uterine Explora curette (Milex, Chicago, IL, USA) to guide endometrial biopsy at the time of SHG or to remove intraluminal masses. In 20 patients this technique was successful in biopsying endometrial filling defects without the need for hysteroscopy.⁵⁰

Use of 3D Imaging in Sonohysterography

Three-dimensional rendering of volume-acquired images has been used to further refine the technique of SHG. Three-dimensional imaging has the advantage of providing a true coronal plane, which is not always possible using two-dimensional (2D) imaging. Surface rendering, while useful in obstetrical imaging, does not play a major role for SHG. Lev-Toaff et al compared 3D SHG to 2D SHG and HSG.⁵¹ They found that in 9 of 13 comparisons between 3D SHG and 2D SHG, and in 11 of 12 comparisons between 3D SHG and HSG, the 3D technique was advantageous. It was easier to identify müllerian anomalies and to localize pathology. In some cases small intracavitary masses were identified with the 3D technique, yet not with the 2D technique.

An additional advantage of 3D imaging is the ability to interpret the images after the study is completed. A 3D volume set may be acquired quickly and then the catheter and probe removed. The volume set may then be viewed and interpreted at a later time. This may minimize discomfort for the patient. It also may increase the efficiency of the scanning laboratory.⁵²

Sylvestre et al examined 209 infertile patients with 2D and 3D SHG and compared the findings with those found at surgery.⁵³ They found that 3D SHG allowed more precise recognition and localization of lesions. They found 2 more polyps with the 3D technique. The sensitivity of 2D SHG was 98% and the sensitivity of 3D SHG was 100%.

Traditional Hysterosalpingography

The HSG, an indirect test of the reproductive canal, requires skill and experience for optimum performance and interpretation. A finding of a normal uterine cavity by HSG has been highly predictive of the same finding at hysteroscopy in many^54,55 but not all³⁸ series; however, up to 30% false-positive rates have been described in these series for uterine abnormalities. Failure to demonstrate peritubal adhesions despite tubal patency, inability to distinguish between bicornuate and septate uteri, and inability to demonstrate noncommunicating hydrosalpinges are frequently described shortcomings of this technique. Sensitivity for tubal patency in series is reportedly between 73% and 100%, and specificity ranges from 87% to 94%.^3,55

The advantages of HSG include the production of a global image of the reproductive canal with exquisite detail of the endothelial architecture from cervix to fimbria when using water-soluble contrast. Lipid-soluble contrast seems to promote fertility when compared with highly irritating water-soluble contrasts.⁵⁶ There is also a widespread familiarity with this technique that will perpetuate its use.

Sonographic Hysterosalpingography

Sonographic hysterosalpingography decreases exposure to allergenic contrast. The additional anatomic information regarding the tubes, uterus, and ovaries provided by transvaginal sonography is unparalleled in its detail, and three-dimensional information is simpler to obtain. Preovulatory status can also be confirmed before the procedure.

The published world experience with ultrasound-detected tubal patency demonstrates an evolution from nonspecific observation of accumulation of cul-de-sac fluid to positive tubal identification and clear detection of hydrosalpinges equal to or better than with standard HSG. The most effective techniques require positive contrast and experience for optimal performance, similar to traditional HSGs. Most series report significant extratubal findings that are unavailable with HSG.

A variety of sonographic contrast agents for the tubes have been tested, including:

1. Saline alone or mixed with unstabilized air bubbles

2. Saline with Doppler assessment of the cornual region

3. Bubbles stabilized in a galactose matrix (Echovist, Schering AG, Berlin, Germany)

4. Flourocarbon contrast (Optison-Mallinckodt Medical, Inc, St. Louis, MO, USA)

As of April 2008, 2 agents are approved in the United States for intravenous use as ultrasound contrast agents for echocardiography: Definity and Optison. There are no agents approved by the US Food and Drug Administration (FDA) specifically for use in sonoHSG. Outside the United States, Echovist is approved for use in sonoHSG. Smooth, flat, early proliferative postmenstrual endometrium provides a clear definition of the endometrial cavity, but tubal patency may be best studied later in the cycle. The more fluid there is in the pelvis, the easier it is to identify turbulence or contrast emanating from the distal tube. Because pelvic fluid accumulates just before ovulation, the preovulatory period (days 11 to 14) is an ideal time to test tubal patency. Endometrial development can also be evaluated, as can follicular development. One expects a dominant follicle of 14 or more mm and no corpus luteum. Confirmation of this also allows one to proceed without fear of disturbing an oocyte in the tube. The size of the largest pocket of ambient fluid is measured to confirm the subsequent addition of fluid through patent tubes. Figures 38-6, and 38-7 depict a sonoHSG using Echovist.

Using anechoic fluids alone during vaginal ultrasound has produced variable results. Because turbulence in free fluid at the fimbria is the most reliable evidence of patency, when this is not seen, patency is assumed by indirect methods: noting flow at the interstitial tube without formation of a hydrosalpinx and accumulation of fluid in the cul-de-sac. Using sterile saline, Bonilla-Musoles et al identified turbulence at only 24 of 54 distal tubes, and abandoned scrutiny of the tubes in favor of the endometrial cavity.²⁸

Tufekci et al in Turkey tested saline infusion through an 8 French Foley.⁵⁷ Patency was defined as the sonographic demonstration of flow in the interstitial and isthmic tube observed for at least 5 seconds, or turbulence at the fimbria. Turbulence was seen in only 11 out of 35 cases with at least 1 patent tube. In 29 out of a total of 38 patent cases (76%), there was complete agreement between sonoHSG and laparoscopy. There was partial agreement in 6 additional cases (15.8%), discordance in 1 case, and in 2 patients fluid in the cul-de-sac was correctly assumed to indicate bilateral patency. Using these criteria for specific tubal diagnosis in 36 patients (excluding the 2 indirect assessments based on peritoneal pooling), the sensitivity was 76%, the specificity was 94%, the PPV was 84%, and the NPV was 90%. Mitri et al used an 8 French Foley in the cervix and 10 to 20 cc of saline in 60 patients.²⁷ Their criteria included only the formation of hydrosalpinges and the appearance of fluid in the cul-de-sac. SonoHSG agreed with HSG in 72%, and the 14 women in whom there was a discrepancy underwent laparoscopy. In 11 of these (79%) the sonographic diagnosis was confirmed; in 3, the HSG was correct. Sonography was more accurate than HSG, because seven noncommunicating hydrosalpinges were detected that were thought to be cornual obstruction by the latter method.

Fleischer et al found that tubal patency could be successfully determined in 79% of women with saline alone and 92% if combined with the use of a positive contrast.⁵⁸

Enhanced Saline

Allahbadia increased turbulence by rapidly injecting 20 cc of air and saline per tube through an intrauterine 8 French Foley catheter.⁵⁹ He sought visualization of proximal flow for 5 seconds and distal turbulence. The patency status of the tubes was 91% concordant between ultrasound, HSG, and laparoscopy in 67 patients. Volpi et al, using a half-filled 30-cc syringe of normal saline, obtained an accuracy of 89% per tube compared with laparoscopy.⁶⁰

Peters and Coulam, and then Stern et al reported the results of color Doppler ultrasonic hysterosalpingography in 238 women.^61,62 They used an intrauterine H/S catheter and an unspecified amount of saline with color Doppler detection of interstitial flow or fimbrial expulsion, finding 81% agreement with laparoscopy in 121 women. Hysterosalpingography in 62 women correlated with laparoscopy in only 60%—an unusually low number. The tube-specific sensitivity and specificity of this new technique were 87%, the false-positive rate was 9%, and the false-negative rate was 20%. They found that repetition of the procedure when unilateral patency was diagnosed by color ultrasound (US)-HSG lowered their false-negative rate to 16%.

Yarali et al evaluated 12 women after laparoscopy and chromopertubation, and 5 after laparoscopic tubal ligation with the operative findings unknown to the sonologist.⁶³ They used transabdominal color-flow Doppler during intermittent infusion of 5 to 30 cc of saline through an intrauterine pediatric Foley catheter. They required color signal in the interstitial tube for at least 10 seconds and "emergence of saline" into the posterior cul-de-sac. The tube-specific sensitivity was 93%, and the specificity was 83%. They calculated a kappa value, indicating clinical agreement beyond chance, of 73%. The pitfalls of color-flow evaluation were reported to include confusing the turbulence in a hydrosalpinx with fimbrial flow, recording continual flow in the interstitial tube as a hydrosalpinx is filled, and tubal spasm or unimaged flow causing false positives.

Stabilized Bubble Contrast Agents

Small particles or microbubbles in fluid provide highly reflective sonographic contrast.^8,63 Observation of washed sperm preparations after intrauterine insemination reveals their rapid propulsion toward and into the interstitial tubes, easily tracked by the strong echogenicity afforded them by bubbles in their albumin-based medium as well as their motile contents. Echovist (Schering AG, Berlin, Germany) is a sonographic contrast composed of microbubbles stabilized by a galactose matrix. It is widely used in Europe and became available in Canada (Berlex) in 1999.

Echovist was tested extensively in Germany by Deichert et al in the late 1980s.²⁶ They were the first to systematically incorporate contrast infusions into transvaginal sonography. Terming the procedure hysterosalpingo-contrast sonography, or HyCoSy, they initially tested it in 120 women and found it to be 100% predictive of patency with a sensitivity of 88% per tube, but to have a PPV of only 50% for tubal occlusion compared with laparoscopy. Criteria for tubal patency included sonographic flow to the isthmus lasting at least 10 seconds, visualization of contrast expulsion from the tube and around the ovary, and lack of a hydrosalpinx. The addition of color-flow Doppler enhanced the contrast but did not improve accuracy. The demonstration of isthmic flow by pulsed (duplex) Doppler improved the accuracy to 100% in an additional 20 women, which was confirmed by laparoscopy. The technique consisted of a baseline exam and then injection of 20 cc of sterile normal saline or Ringer lactate solution into the uterus for SHG, with careful evaluation for intrauterine abnormalities using a small Foley catheter or HSG balloon cannula in the cervix or uterine cavity. They then pulsed incremental doses of Echovist into the uterus and tubes during sonographic observation and looked for periovarian and isthmic flow.⁶⁴

Echovist is mixed immediately before injection because dilution and heat cause dissolution of the bubbles and galactose particles. Galactose has not caused allergic reactions, but it should not be used in individuals with galactosemia.

A review of the use of Echovist in more than 600 patients reports an 84% to 91% concordance with HSG, and an 81% to 93% concordance with laparoscopy.⁶⁵ The extraluminal information, convenience, decrease in discomfort, and lack of irradiation may outweigh the failure to improve on the variable accuracy of standard HSGs.

Hamilton et al reported that the rate of unassessable tubes dropped from 7.4% to 4.8% after they had examined 100 patients using saline followed by Echovist, in a series of 500 patients. One-third (185) were compared with independent laparoscopy, with a per tube concordance of 86%, sensitivity for patency of 90.4%, and specificity of 70%. Noting that patency was only one aspect of normalcy, they noted tortuous slow spilling tubes in 5% of technically patent tubes. HyCoSy also provided important additional information: a third of patients had ovarian abnormalities.⁶⁶

Stacey et al described a series of 118 patients who underwent tubal patency examinations by ultrasound with Echovist.⁶⁷ In 15 patients where tubal occlusion was documented by sonography, HSG was also performed. The ultrasound findings were only confirmed in 8 of these 15 patients. Twenty-three of the 118 patients had severe protracted pain or vasovagal reactions with bradycardia and hypotension. The authors felt that the sonographic technique for evaluating tubal patency compared unfavorably with HSG.

Boudghene et al compared an investigational contrast agent consisting of air-filled albumin microspheres with saline solution in the sonographic detection of tubal patency.⁶⁸ Tubal patency was defined by visualization of the contrast agent from the tube into the abdominal cavity. Twenty-three infertile women were studied with both techniques. Comparison was made to HSG. The number of fallopian tubes correctly classified with ultrasound was greater using the contrast agent than with saline alone. The success of use of the ultrasound contrast agent led the investigators to terminate the trial early. The use of color Doppler did not have an impact on the final diagnosis, although it appeared to increase the confidence as the tube was easier to identify and color than in B-mode.

Radic et al studied 68 patients and compared saline and Echovist as ultrasound contrast agents for determining tubal patency.⁶⁹ Comparison was made to laparoscopic and dye infusion as the gold standard. They also found that Echovist was superior to saline as a contrast media. The sensitivity for Echovist was 100% and the specificity was 77%. As in other studies, sonoHSG falsely identifies tubal occlusion at a higher rate than HSG. They also found that patients tolerated the procedures well without vagal complications. Others have reported less pain with sonoHSG than with HSG.⁷⁰

Three-Dimensional SonoHSG

Several investigators have studied the use of 3D power Doppler imaging in sonoHSG. The technique differs from 2D in that the contrast is injected while a 3D power Doppler volume acquisition is obtained. The rendered image is then stored and can be reviewed after completion of the study. Chan et al performed sonoHSG with 3D power Doppler on 21 consecutive patients undergoing laparoscopy.⁷¹ Findings included a sensitivity for detection of tubal patency of 100% with a specificity of 67%. Kiyokawa et al compared sonoHSG using 3D power Doppler to HSG and found a sensitivity of 84% and specificity of 100%.⁷² The lone report comparing 2D to 3D sonoHSG found that the mean duration of the imaging procedure was shorter with the 3D technique and a lower volume of contrast material was needed.⁷³ The 3D technique allowed for visualization of spill of contrast from the distal end of the tube. No studies comparing 2D to 3D sonoHSG in diagnostic accuracy were found in the English language literature.

SonoHSG and Transcervical Sterilization

In 2002, the US FDA approved the use of the Essure device (Conceptus, Menlo Park, CA, USA) for transcervical sterilization. The Essure procedure involves placement of micro-inserts into the fallopian tubal ostia by hysteroscopy. The devices contain Dacron fibers, which cause a tissue reaction that leads to occlusion of the tube over the course of 3 months. In the United States the FDA requires that an HSG be performed prior to relying on the method for contraception. In other countries ultrasound or abdominal x-ray is used to localize the devices. Connor has investigated the use of sonoHSG to evaluate tubal occlusion after the Essure procedure.^74,75 Sonographic evaluation of tubal status was compared with subsequent HSG in 17 women who had successful bilateral Essure placements. In all patients, with a least 1 tube patent on HSG, patency was also demonstrated by sonographic technique. Patients preferred to have the sonographic examination to the HSG. In Figure 38-8, the proximal tip of an Essure device is seen on SHG.

Possible Complications of Uterine Infusion

There are a number of theoretic complications possible with transcervical infusion, but most are avoidable. They include the following:

1. Postprocedure exacerbation of pelvic inflammatory disease (PID)

2. Severe transient pain

3. Perforation of the uterus

4. Vasovagal reaction

5. Inability to enter the cavity

6. Failure to image the endometrium

7. Retrograde spillage of neoplastic epithelium into the peritoneal cavity

Schlief and Deichert reported 8 complications in 140 patients undergoing HyCoSy: 1 case of hypotension, 1 episode of PID, 1 episode of intolerable pain due to catheter insertion, and 5 episodes of moderate pain during infusion.⁷⁶

Gentle, slow insertion of the catheter into the uterus and slow initial infusion are necessary to avoid cramping. Hitting the fundus with the catheter is painful to the patient. Having the patient cough during inflation of a cervical balloon can distract her from cramping.

The catheters described are too soft to perforate a normal uterus. Infusion of the uterus can detect perforation with other instruments, however. Vasovagal reactions are usually accompanied by complaints of feeling hot, nauseated, or faint. The procedure should then be immediately halted.

Dealing with severe cervical stenosis requires a good light, a paracervical block, and a gall bladder or lachrymal duct probe. Tapered Teflon dilators are also useful for overcoming stenosis. The first maneuver is to use the toe of the speculum or transabdominal fundal pressure to decrease uterine flexion. A tenaculum may be necessary to stabilize the cervix in more difficult cases. Fine forceps can be used to nuzzle the external os open. The H/S catheter can provide gentle dilation by inflation of its balloon during incremental advances up the canal. Lidocaine can be infused a few drops at a time during this maneuver. Concurrent abdominal scanning may be necessary to guide a probe or 2-mm curet to avoid perforation. Cervical ripening with vaginal or oral misoprostol 4 to 6 hours prior to the procedure may also be useful in overcoming cervical stenosis.

Transrectal or transabdominal scanning is helpful when a large or myomatous fixed uterus is poorly imaged. If the cervix will not hold a balloon, one can be inflated in the uterine cavity, pulled to the internal os, and deflated during sonographic observation until it lodges in the proximal cervix, then reinflated.

Severe cervicovaginal atrophy in a postmenopausal woman can cause pain and difficulty with the insertion of a speculum, much less a cervical catheter. Two weeks of 1 g of estrogen cream in the vagina nightly will make the procedure possible and less uncomfortable, without significant hormonal effects on breast or endometrium.

Normal

SHG is most effective in identifying excluding significant intrauterine pathology (Figure 38-9). The endometrium of a normal cavity appears uniform, without focal thickening or masses. Hyperplasia may be focal or global, but the endometrium is almost always asymmetric. The increased size and tortuosity of the glands produce hyperechogenicity reminiscent of secretory endometrium, but the echo pattern is usually less uniform. The surface may be undulating, but it is most often ragged or overtly polypoid. One cannot distinguish among simple and complex hyperplasia, hyperplasia with atypia, and cancer with ultrasound. One can, however, distinguish isolated polyps, locate focal lesions, and follow response to treatment.⁷⁷

Polyps

Polyps are thought to be focal areas of hyperplasia, often of the basalis layer, and as such are resistant to sloughing during progesterone withdrawal (Figures 38-10,38-11,38-12, and 38-13).⁷⁸ They are sessile as they erupt out of the endometrium and become progressively pedunculated. They have a predilection for the fundus and the tubal recesses. Although they are a common cause of abnormal bleeding, many do not bleed but can cause expulsive contractions of the uterus, resulting in cramping. They occur at all ages and are found most often in women exposed to prolonged unopposed estrogen or tamoxifen and in infertile women. They appeared in 24% of 1305 endometrial biopsies in symptomatic women, and only 0.06% were precancerous or cancer.⁷⁸ Those identified in curettages represent only a fraction of those present. Dilation and curettage and certainly office biopsy have often missed them, resulting in repeated attempts at diagnosis of abnormal bleeding, hysterectomy, or abandonment of hormone replacement therapy. In postmenopausal women with persistent bleeding, 38% were found to have polyps.⁷⁹

Polyps are always hyperechoic relative to normal myometrium, but they often have tiny cystic areas. The feeding vessel may be identified using color or power Doppler imaging as seen in Figure 38-13. Some have described a hyperechoic line that separates the myometrium from the polyps from mucus or fluid. This finding can help distinguish diffuse endometrial thickening from a polyp.⁸⁰

Because many, but not all, cancers are polypoid, all bleeding polyps should be sampled.⁸¹ It is not clear that removal is required when small, smooth, asymptomatic polyps surrounded by atrophic (1- to 2-mm) regular endometrium are identified sonographically.

Tamoxifen Changes

Tamoxifen is a mixed estrogen agonist/antagonist with occasional profoundly estrogenic effects (Figures 38-14 and 38-15). Whereas it has anti-estrogen effects in breast tissue, it can act paradoxically as an agonist on endometrial stroma. The unpredictably proliferative endometrial and even myometrial effects are reflected in their unusual sonographic images.⁸²

If the unenhanced vaginal ultrasound shows a thin atrophic endometrium, the patient may be followed up. If a bizarre central uterine appearance is seen, fluid enhancement should be carried out. If the changes are microcysts in the inner myometrium with thin atrophic endometrial changes, the patient may be followed up.²³ If there are focal intracavitary changes, hysteroscopically directed removal is suggested. Endometrial biopsy is required to elucidate the histology of all except clearly atrophic endometria.

Leiomyomas

Myomata are common, occurring in 20% to 30% of women older than 30 (Figures 38-16 and 38-17). Hysteroscopy and SHG have made it clear that submucous myomata are distinct from intracavitary myomata.

Intracavitary myomata are more likely to bleed because they have exposed surfaces often traversed by large, fragile veins. Truly submucous myomata are covered with endometrium and may not be visible hysteroscopically, although they distort the cavity on HSG and ultrasound. To remove them transcervically entails the removal of a fair amount of endometrial surface area, which may produce subsequent synechiae. They may have a substantial intramural component and often cannot be completely removed at once without risking uterine perforation.

The precise location and relationship to the endometrium are clear during SHG, even when the myomata are dense and cause acoustic shadowing. They are almost always distinguished from polyps by their typical whorled, hypoechoic sonographic pattern and their tendency to reflect and deflect ultrasound due to their dense homogeneous structure. However, tiny intracavitary myomata smaller than 1 cm in diameter often do appear hyperechoic. Fleischer and Shappell performed color Doppler SHG on 25 women with abnormal bleeding.⁸³ They identified distinctive vascular patterns, which helped in differentiating between fibroids and polyps. In particular, polyps contained typically a single feeding vessel while fibroids had several vessels that arose from the inner myometrium.

Preoperative assessment of submucous myomata is an important role for SHG. Fibroids that penetrate more deeply into the myometrium are more difficult to remove via hysteroscopic resection. Preoperative planning for the more difficult deeply penetrating myomas is helpful to the gynecologic surgeon. The European Society for Hysteroscopy developed a classification scheme for submucous myomas. Type 0 myomas are completely intracavitary, while type 1 myomas penetrate less than 50% into the myometrium. Type 2 myomas penetrate greater than 50% into the myometrium. The depth of penetration of the myoma into the myometrium may be calculated using SHG. Leone et al found that using strict SHG criteria for preoperative grading of submucous myomas was as accurate as hysteroscopy.⁸⁴ In addition to depth of penetration they also used the angle between the endometrium and the circumference of the fibroid for classification purposes (Figure 38-18). Salim et al found good agreement between 3D SHG and diagnostic hysteroscopy.⁸⁵ The correlation was better when a greater proportion of the fibroid was contained within the endometrial cavity.

Synechiae

Uterine synechiae range from a few filmy synechiae that have no reproductive consequences to complete obliteration of the uterine cavity (Figure 38-19). Hysterosalpingography is the traditional method of diagnosis; however, nothing will be seen if there is a supracervical obstruction. Conventional scanning demonstrates poorly defined or absent endometrium best appreciated in the secretory phase. Sequestered hematometra or odd, tiny fluid collections can highlight thick or thin adhesions, respectively. Sonohysterography, however, allows one to make a definitive "map" of the defects and to confirm correction of the cavity postoperatively. A balloon is required for severe adhesions for adequate expansion. The patient should be premedicated in this case for analgesia. Because of the theoretic possibility of intravasation due to damaged endometrium and the higher pressures generated, lidocaine is not used.

Uterine Anomalies

Ultrasound efficiently distinguishes a unicornuate uterus from one with an obstructed noncommunicating horn and from the uterus bicornis bicollis with unilateral hematocolpos due to a unilateral vaginal septum, an important distinction that will determine the reparative procedure (Figures 38-20 and 38-21). Rapid sonographic evaluation of the kidneys is also possible: More than 80% of women with these asymmetric anomalies have renal agenesis on the vestigial side, and virtually 100% have no kidney on the side of a unilateral vaginal septum.⁸⁶

The distinction between a septate uterus and a bicornuate uterus cannot be made by HSG, but it can accurately be made using MRI or transvaginal ultrasound with abdominal manipulation of the uterus and saline infusion if necessary.⁸⁷ The septate uterus is associated with repetitive second trimester loss, whereas a bicornuate uterus generally requires no repair to maintain pregnancy until viability; hence, the woman in whom the latter is identified can be spared diagnostic surgery. Hysteroscopic metroplasty has been performed for many years and is an outpatient procedure.

Subtle cavitary abnormalities such as those in diethylstilbestrol (DES)-exposed women are clarified by saline infusion, as are unicornuate uteri with vestigial horns. Both are associated with pregnancy wastage.

The highest value for 3D imaging in gynecology is in evaluation of müllerian anomalies. Primarily, the value of 3D imaging is in the accurate delineation of the fundal contour of the uterus. When used concomitantly with saline infusion sonohysterography, sonography is likely a better imaging modality for evaluation of müllerian anomalies than hysterosalpingography.⁸⁸ If transvaginal sonography is nondiagnostic, MRI is the method of choice for evaluating possible uterine anomalies.

Cesarean Scars

In patients who have had prior cesarean deliveries, a niche or filling defect can be identified anteriorly at the site of the uterine scar (Figures 38-22 and 38-23). The clinical significance of this finding is uncertain. Monteagudo et al studied 44 patients with histories of cesarean delivery with SHG.⁸⁹ They found a niche in all patients. There was no correlation between the number of cesarean deliveries and the depth of the niche.

Endometrial Carcinoma and Hyperplasia

Distinguishing endometrial cancer or hyperplasia from other conditions is not possible based solely on unenhanced measurements of the endometrial echo (Figures 38-24,38-25, and 38-26). Sonohysterography may demonstrate asymmetric thickening of the endometrium. However, no specific cutoff values have been developed to define asymmetry for endometrial stripe measurements during sonohysterography. Inability to distend the endometrial cavity during attempted sonohysterography has been reported as a finding consistent with endometrial carcinoma.⁹⁰

The use of infusion in women with endometrial cancer may occur before diagnosis or may be used to help determine the depth of invasion or cervical involvement. A very large polyp extending to the endocervical canal and distending the myometrium can appear to be invasive. When outlined with fluid, however, the superficial nature of the process is apparent and surgical planning can be done accordingly.

Will saline infusion cause tubal reflux that could change the woman's prognosis? The opportunity for tubal reflux is present during hysteroscopy, chromopertubation, D&C, and spontaneous bleeding in any woman with patent tubes.⁹¹ Hysterosalpingography was used in the past for preoperative staging of endometrial cancer, and women demonstrating tubal patency fared no worse than those with tubal occlusion when matched by stage of disease and followed up for 5 years.⁹² Alcazar et al attempted to assess the risk of malignant cell dissemination in patients with proven endometrial cancer during SHG.⁹³ At the time of surgical treatment for their endometrial cancer 14 women underwent infusion of 10 to 20 mL of saline via a transcervical catheter. The saline infusion was performed at the time of laparotomy and any fluid that spilled from the fallopian tubes was collected and analyzed for the presence or absence of malignant endometrial cells. The infusion of saline was not done under ultrasound guidance. There was no spillage of fluid from either tube in 9 patients. Cytologic analysis revealed the presence of malignant cells in the spilled fluid in 1 case. Dessole et al prospectively performed SHG at the time of laparotomy for treatment of known endometrial cancer.⁹⁴ Of the 32 patients studied, malignant cells were found in the fluid that spilled from the fallopian tubes in 2 patients. Also of interest was that sonohysterography correctly identified the depth of myometrial invasion in 84% of cases.

Takac retrospectively studied 173 patients who had either dilation and curettage alone or dilation and curettage followed by SHG.⁹⁵ All patients had confirmed endometrial cancer. There was no difference between the groups in terms of positive peritoneal cytology, adnexal metastases, or distant metastases within the abdomen. It was concluded that SHG does not increase the risk of malignant cell dissemination in patients with endometrial cancer.

Development of improved contrast agents for determination of tubal patency by sonoHSG may improve the diagnostic accuracy of the technique and overcome some of the technical challenges. The continued use of volume imaging (3D/4D) may improve the efficiency and accuracy of SHG. As mentioned above, development of newer catheters may allow for directed biopsy or operative interventions such as polypectomy during SHG. Recently, investigators have tried using gel instead of fluid during SHG.⁹⁶ The use of a more viscous solution may be useful in cases where the uterus is difficult to distend due to a patulous cervix.

Highlighted References