Chapter 36. Imaging in Gynecology

C.O. is a 29-year-old white woman who presented with a history of infertility for several years, followed by a history of recurrent pregnancy losses.

Her past medical and surgical histories were negative. On gynecologic history, she was remarkable in that she reported severe dysmenorrhea for several years that was relieved by nonsteroidal anti-inflammatory drugs. Her gynecologist found a low luteal phase progesterone level and treated her with 50 mg of clomiphene citrate on days 5–9 of the cycle.

She responded well to the medication, with a subsequent conception. The pregnancy resulted in a spontaneous abortion 5 weeks later. No dilatation and curettage (D&C) was required, and the patient recovered well. She was still unable to conceive on her own and was again given clomiphene citrate therapy. Again, she conceived and had a spontaneous abortion—this time at 7 weeks' gestation. No D&C was performed.

The patient was evaluated for recurrent pregnancy losses. Karyotype was normal for both partners. Hormonal evaluation was normal with the exception of a low midluteal phase progesterone level. Immunologic and infectious screening failed to reveal a cause for the recurrent losses. The hysterosalpingogram (HSG) demonstrated a midline filling defect.

The patient was informed of the results and the potential for future miscarriages. The need for further evaluation and possible repair performed via a hysteroscopic or abdominal approach, together with its risks and benefits, was carefully explained to the patient. She elected to try clomiphene citrate therapy one more time and hoped to avoid surgery.

At 8 weeks' gestation, vaginal ultrasonography (US) revealed positive fetal cardiac activity in an ovulation induced by clomiphene citrate. While still taking micronized progesterone 100 mg 3 times daily, she was referred to her gynecologist for routine obstetric care. At 12 weeks' gestation, the patient had an incomplete abortion that required a D&C. She recovered uneventfully and later returned to the office for further evaluation and treatment.

Several months were allowed to lapse before a hysteroscopy/laparoscopy was performed, which revealed a broad-based intrauterine septum and stage I endometriosis. To evaluate the depth and width of the septum, a LaparoScan (EndoMedix, Irvine, CA) laparoscopic 7.5-Hz probe was used during the procedure. The septum was removed with a hysteroscopic resectoscope loop on a 40-W setting. After the resection, the ultrasonic probe was again used to measure the thickness of the myometrium and to verify the resection of the septum. A 30-mL 18-F Foley catheter with the distal tip resected was placed in the fundus and inflated. The patient was discharged and placed on therapy consisting of a broad-spectrum antibiotic and conjugated estrogen 2.5 mg daily.

The new millennium saw a proliferation of imaging techniques used in medical practice. Research into the development, refinement, and application of imaging in gynecology is apparent in the literature.

The HSG has been considered the gold standard for imaging the uterine corpus for benign disorders (submucous myomas, submucous polyps, localization of tubal occlusion, and evaluation of müllerian fusion defects) and malignant disease (endometrial carcinoma). In the case reported, the standard scout film was obtained, and the cervix was prepared after the following were assured: position of the uterus, absence of pelvic tenderness, and negative pregnancy test. The water-soluble contrast medium was injected into the uterine cavity, and oblique and anteroposterior films were obtained. These films showed a midline uterine filling defect of the type usually seen with septate or bicornuate uteri.

US performed on this patient during her pregnancies failed to show the filling defect. If suspected, the septum might have been encountered with more careful scanning. The scans of the last pregnancy revealed only an eccentrically placed pregnancy that might be seen on US even in normally structured uteri. Although not helpful at this point, US examination of the uterus between conceptions might have been helpful if used with a distending medium. This is especially useful in patients allergic to iodine contrast medium (Table 36–1). This technique of ultrasonic HSG is performed by occluding the cervix with a uterine injector and distending the uterus. The method can demonstrate the separate cavities as well as the possible difference between the septate and the bicornuate uterus while demonstrating tubal patency. The technique was adopted for this patient during her uterine septum resection, with the addition of ultrasonic contrast between the endometrial cavity and septum and the myometrium. Readers are referred to the many fine texts on diagnostic pelvic US for instruction and further discussion of these techniques. The development of “sonicated” contrast solutions may add greatly to the usefulness of US.

Using 2 video cameras (1 for the resectoscope and the other for the laparoscope and the LaparoScan laparoscopic US probe), all aspects of the surgery were evaluated. This setup allowed the operating surgeon adequate visualization of the uterine cavity during the resection and enabled other personnel in the operating room to follow the progress of the surgery. The laparoscopic video allowed for careful monitoring of the uterine surface and assured the surgeon that there would be less likelihood of a uterine perforation, a complication that could result in bowel injury. The laparoscopic US probe with a picture within a picture was useful because it allowed visualization of the 2 separate cavities and measurement of the length and width of the septum. It also enabled the operator to demonstrate the complete removal of the septum.

Although plain film radiographs are one of the most common forms of imaging in radiology, they are rarely the test of choice for identifying gynecologic pathology. However, they can be used to detect calcified leiomyomas as well as an intrauterine device (IUD). Such films can help determine if an IUD has been expelled from the uterine cavity or has penetrated the uterine wall and migrated to an ectopic location.

Pelvic US, magnetic resonance imaging (MRI), and computed tomography (CT) imaging are more common modalities in assessment of the uterus and the cervix. Pelvic US is the most common initial imaging approach in diagnosis of uterine disease. Modalities for pelvic US include a transvaginal approach, transabdominal approach, and saline infusion sonohysterosalpingogram. Pelvic US plays a significant role in the diagnosis of uterine leiomyomas (submucosal, intramural, and subserosal) and polyps and in the monitoring of follicular development in assisted reproduction. In more recent years, 3-dimensional (3D) US has been investigated in comparison to 2-dimensional (2D) US and hysteroscopy. Salim and colleagues showed 3D US to be superior over 2D US in the measurement of intramural versus submucosal involvement of leiomyoma. Additionally, US is often the initial test to suggest other pathology such as müllerian anomalies and adenomyosis. Occasionally, the detection and localization of myomas, assessment of their size, and their differential diagnosis are difficult. In the circumstances of adenomyosis, müllerian anomalies, or additional information needed on myomas, it can be useful to perform MRI of the pelvis. MRI produces images with excellent soft tissue resolution and is useful for evaluation of congenital abnormalities of the uterus, leiomyomas, adenomyosis, gestational trophoblastic disease, and endometrial carcinoma diagnosis and staging. MRI can accurately measure the volume of the myoma, which aids in determining whether medical management of myomas has resulted in shrinkage or whether conservatively treated myomas are growing. Malignant degeneration of myomas visualized by MRI, as described by some authors, allows for early and appropriate intervention.

MRI can effectively discern between the septate and the bicornuate uterus, thus avoiding the more costly laparoscopy. MRI may provide a clear anatomic picture of complicated müllerian fusion defects (didelphys with transverse vaginal septum or noncommunicating uterine segment) and allow for proper planning of surgical repair. If pelvic MRI had been performed on the patient in the opening case report, the image probably would have appeared the same as the MRI shown in Figure 36–1. (See review of MRI findings of müllerian fusion defects in Table 36–2.)

Currently, histopathologic evaluation of colposcopic biopsies is required to diagnose cervical cancer and its precursor lesions. However, the technique is expensive and often requires a waiting period before histopathologic results are available and necessary treatment can be scheduled. Several new imaging techniques that evaluate the cervical epithelium are under investigation. Optical techniques, such as elastic backscattering and fluorescence and Raman spectroscopies, have been used to noninvasively examine tissue morphology and the biochemical composition of the cervix. Optical coherence tomography (OCT) is a noninvasive imaging technique that uses coherent light to form images of subsurface tissue structures with 10- to 20-μm resolution and up to 1-mm depth. A study by Zuluaga and colleagues showed that simple quantitative analysis of images obtained with an OCT system can be used for noninvasive evaluation of normal and abnormal cervical tissue in vivo. OCT imaging could have broad applications for screening and detection of cervical malignancies and their precursors. It may also aid in surgical planning by allowing surgeons to identify margins in vivo without obtaining frozen sections.

Pelvic US has been used to evaluate the uterine cavity, and endometrial thickness has been used as a marker for endometrial pathology. The following guidelines should be used to obtain interobserver consistency in the evaluation of the endometrium. Obtain measurements from the midfundal region in the sagittal plane. Obtain the maximal double-thickness dimension, remembering to exclude the hypoechoic area between the myometrium and the endometrium. Any fluid between the anterior and posterior walls should be subtracted from the total measurement. Endometrial thickness ranges from 4–8 mm during the follicular phase and 7–14 mm during the luteal phase with a uniform echogenic appearance.

Premenopausal women should be evaluated during the early follicular phase, immediately following the menses when the endometrium has a uniform linear appearance. Menopausal women usually have an endometrial stripe <4 mm. Menopausal women on hormone replacement therapy (HRT) may have an endometrial thickness exceeding 8 mm and a small amount of fluid (<1 mm) (Table 36–2).

Approximately one-fifth of patients with abnormal uterine bleeding have submucous myomas or polyps. A study by Tur-Kaspa and colleagues found that 20% of infertile patients had some abnormality on SIS: arcuate uterus (15%), polyps (13%), submucosal fibroids (3%), and adhesions (<1%). These lesions may be detected by irregularities in the endometrial stripe or by saline infusion sonohysterography (SIS), a technique that involves saline infusion to distend the uterine cavity. Using SIS, a polyp appears as a smoothly marginated focal lesion that protrudes into the endometrial cavity. Kelekci and colleagues showed the sensitivities and specificities for detecting cavitary lesions with transvaginal US, SIS, and hysteroscopy to be 56.3% and 72%, 81.3% and 100%, and 87.5% and 100%, respectively. The sensitivity and specificity rates of US in detecting endometrial pathology reportedly increase when color flow and power Doppler imaging are used. However, tissue sampling is required to make a definitive diagnosis and to rule out malignancy in any patient not on HRT with a hyperechogenic endometrial stripe >4 mm.

3D US has been studied for evaluation of the endometrium. The ability of 3D US to produce coronal images of the cornua may increase slightly the sensitivity of SIS for detecting lesions in this location that otherwise might be difficult to evaluate. Endometrial abnormalities that can be seen in women with congenital malformations of the uterus may be imaged to greater advantage with 3D US techniques. 3D US has also been demonstrated to be a valid measurement technique for assessing volume. However, hysteroscopy likely will become the new gold standard for evaluating the endometrium because of the ability of hysteroscopy to directly visualize the endometrium and perform biopsies as indicated. The technique may become more cost effective as use of office hysteroscopy becomes more widespread. Evaluation of the endometrium using transvaginal sonography as the initial screening tool, followed by endometrial biopsy or possibly hysteroscopy, is likely to become the standard of care (Fig. 36–2).

Newer imaging modalities are under investigation to decrease the number of imaging tests often ordered in the evaluation of infertility. These tests seek to evaluate tubal patency, tubal architecture, the uterine cavity, and the myometrium in 1 study. Unterweger and colleagues showed that 3D dynamic magnetic resonance hysterosalpingography (3D dMR-HSG) can evaluate tubal patency in addition to MRI evaluation of the uterus and pelvic structures. Virtual hysterosalpingogram (VHSG) and multislice CT hysterosalpingography (MSCT-HSG) are similar tests but use CT technology in place of MRI for evaluation. Although these tests are more expensive than traditional procedures, they can provide the information previously gained from pelvic US, HSG, and MRI/CT in a single radiologic modality.

The flat plate of the abdomen may still be useful in the diagnosis of dermoid cysts of the ovary, which are identified by the presence of calcified teeth. However, cystic and solid structures of the ovary now are better evaluated by transabdominal US, transvaginal US (TVUS), CT, and MRI.

Transvaginal assessment of the ovary is frequently used in assisted reproduction cycles, both in management and prediction of success. Antral follicles measure 2–10 mm in size and appear as small hypoechoic structures within the ovary and are typically measured in a basal state. Hendriks and colleagues have shown that the number of basal antral follicles is superior to follicle-stimulating hormone in predicting poor ovarian response. The ovarian follicles also change size in response to ovarian stimulation, whether exogenous or endogenous. These changes are monitored during assisted reproductive technology (ART) cycles to help in the timing of human chorionic gonadotropin (hCG) injections, oocyte retrieval, management of gonadotropin stimulation, and determining the need for cycle cancellation.

US is frequently used in the evaluation of ovarian pathology. TVUS combined with color flow and Doppler can be used for evaluation of blood flow to the adnexal structures and for diagnosis of ovarian torsion. Venous and lymphatic flow is occluded in early torsion, but arterial flow may be present. Arterial flow ceases completely later in the torsion. Torsion is diagnosed by Doppler study showing no venous or arterial flow, but a study showing arterial blood flow does not necessarily rule out the diagnosis.

Approximately 12,000 women in the United States die annually of ovarian cancer. Unfortunately, the ability of the pelvic examination to detect early ovarian malignancy is poor. CA-125 monoclonal marker for ovarian cancer also is a poor predictor of early cases. In an attempt to discriminate between malignant and benign adnexal masses, morphologic criteria have been assigned to increase suspicion concerning US findings when ovarian cancer is suspected. Cysts larger than 4 cm, solid and cystic components, septa, and papillary nodules have all been described (Fig. 36–3). In addition, Doppler flow studies have been used to distinguish between benign and malignant masses (Table 36–3).

CT may be useful for preoperative staging of ovarian cancer or for planning second-look procedures. CT may also be useful for biopsy and drainage in patients with benign-appearing adnexal masses (ovarian cysts or tubo-ovarian abscesses). Contraindications to needle biopsy and drainage include lack of a safe unobstructed path for the needle, bleeding disorders, and lack of a motivated patient. 3D US can be useful in the evaluation of gynecologic diseases. It can reconstruct any plane of interest and is particularly valuable in visualizing abnormalities in the coronal plane. In addition, 3D US is better able to measure volumes than is 2D US and, therefore, is helpful when evaluating patients with conditions ranging from fibroids to infertility. CT can be used in conjunction with pelvic US to diagnose and manage several conditions, such as pelvic inflammatory disease, adnexal torsion, ovarian vein thrombosis, and hemorrhagic ovarian cysts. In addition, MRI has been shown to be particularly useful in the evaluation of ovarian vein thrombosis.

Endoscopic techniques provide the best direct evaluation of the patency and architecture of the fallopian tubes. HSG provides the best indirect evaluation of tubal function. HSG allows demonstration of tubal patency and visualization of tubal rugations while avoiding the more costly laparoscopic surgery. Some disadvantages of HSG are pelvic infection, dye allergies, failure to detect adnexal adhesions, and false-positive results for tubal occlusion. Salpingitis isthmica nodosa (SIN) is suggested when a honeycombing of the contrast material appears in the tubes during HSG. Hydrosalpinges are suggested when a hypoechoic “sausaging” of the tube is seen on US. Hysterosalpingo-contrast sonography, 3D dMR-HSG, VHSG, and MSCT-HSG are alternatives to laparoscopy, as women with normal findings probably have a normal pelvis.

Adnexal sonography is a valuable tool in assessing women with suspected ectopic pregnancy (Table 36–4). When hCG levels reach 6500 mIU/mL, most normal intrauterine pregnancies can be detected as a gestational sac by transabdominal US. However, the sonographic appearance of a pseudogestational sac should not be confused with the gestational sac. In the latter, a double ring sign resulting from the decidua parietalis is seen abutting the decidua capsularis.

TVUS, on the other hand, has the advantage of earlier and better localization of the pregnancy, with less pelvic discomfort because the bladder is not painfully distended. An hCG level of 1000–2000 mIU/mL is the discriminatory zone in which an intrauterine pregnancy can be detected by TVUS. The double-ring sign and the yolk sac must be identified to ensure that the pregnancy is intrauterine. When an intrauterine pregnancy is not visualized on TVUS and the hCG level exceeds 1000–2000 mIU/mL, suspicion for an ectopic pregnancy should be high. However, multiple gestations and heterotopic pregnancies may take several more days to be identified. Both of these conditions are more common in patients undergoing ART. Because it is not well established at what hCG levels multiple gestation or heterotopic pregnancies should be visualized on US, particular caution is needed in evaluating for ectopic pregnancies in ART patients.

In recent years, newer imaging modalities have been evaluated for their ability to detect peritoneal pathology, primarily malignant processes and endometriosis. US has limited ability to detect peritoneal pathology, especially smaller processes. Shaw and colleagues concluded that, although MRI may have a role, CT imaging remains the test of choice for the diagnosis of pelvic malignancy. For deep and surface pelvic endometriosis, Bazot and colleagues showed that MRI has a high sensitivity and specificity for diagnosis. Despite these recent advances in peritoneal imaging, laparoscopy remains the gold standard for the diagnosis of peritoneal pathology.

Although the pituitary gland is located outside the pelvis, imaging of the gland is indicated in gynecologic patients with hyperprolactinemia. Patients with hyperprolactinemia may present with infertility, galactorrhea, decreased libido, and oligomenorrhea. Imaging of the pituitary has been recommended with prolactin levels over 100 ng/mL or in any patient with persistently elevated prolactin levels. A recent study by Bayrak and colleagues showed that while the size of the pituitary prolactinoma does correlate with serum prolactin levels, patients with macroadenomas may present with only moderate prolactin elevations, and therefore, any patient with persistently elevated prolactin levels should undergo pituitary imaging.

Historically, pituitary imaging was performed with a coned-down x-ray film of the sella turcica. MRI has replaced this approach with a greater sensitivity for detecting microadenomas as compared to both x-ray and CT scans. On MRI examination, microadenomas are commonly identified as areas of low-intensity T1 signaling.

The imaging techniques prevalent today have proved to be valuable tools in the diagnosis and treatment of benign and malignant gynecologic disorders. To provide the patient with the highest level of medical care, the contemporary practicing gynecologist must constantly keep abreast of the new developments and applications of diagnostic imaging. No matter what technology is used today and in the future, the goal will always be the same: to provide quick, low-risk, accurate diagnosis of gynecologic conditions while keeping in mind the cost effectiveness of the care delivered.