Chapter 34: TRANSVAGINAL SONOGRAPHY OF ENDOMETRIAL DISORDERS

Over the past several years, transvaginal sonography (TVS) of the endometrium has assumed an integral role in the evaluation of women with possible endometrial disorders.¹ Transvaginal sonography allows detailed delineation of endometrial thickness and texture in most patients. The sonographic findings have important implications in determining which patients need endometrial biopsy or dilatation and curettage, observation, or alteration of medication. The more extensive use of sonohysterography (SHG) (fluid instillation into the endometrial lumen) has also greatly enhanced sonographic evaluation of patients with endometrial disorders. This chapter discusses and illustrates the clinically important applications of TVS in the evaluation of the endometrium. It also discusses some of the technical factors and limitations of the technique.

One of the most common disorders evaluated by gynecologists is unexpected or excessive uterine bleeding. It should be emphasized that only 10% to 15% of postmenopausal women who present with postmenopausal bleeding have endometrial cancer; the vast majority have bleeding secondary to benign disorders such as atrophy or hyperplasia.² The role of sonography is to determine which patients can forgo endometrial biopsy and to monitor endometrial changes in patients receiving tamoxifen or other forms of hormone replacement therapy (HRT). Approximately 30% of women receiving combined HRT experience bleeding. It is the role of TVS to determine which patients need biopsy, dilatation and curettage, alteration in their medicines, or observation.

Although endometrial cancer is more common than ovarian cancer, this disease is associated with far less mortality. This is clearly related to the fact that one of the earliest signs of endometrial cancer is vaginal bleeding. In addition, endometrial cancer is usually still confined to the uterus when the patient presents, whereas two-thirds of patients with ovarian cancer have extensive disease at time of presentation (see Chapter 35 on early detection of ovarian and endometrial cancer).

It is important conceptually to differentiate causes of uterine bleeding in the premenopausal from the peri- or postmenopausal woman. In the woman of child-bearing age, bleeding is usually associated with anovulatory cycles. This is typically related to poor corpus luteum function and support of endometrium with progesterone. Fibroids are also a very common cause of bleeding in this age group. It is thought that fibroids may denude the endometrial surface, resulting in excessive bleeding. Postmenopausal women have bleeding due to a variety of disorders. These include atrophy, hyperplasia, polyps, cancer, and fibroids. Atrophy is the most common cause of bleeding and is related to excessive thinning of the endometrium, making it friable and prone to bleeding. Sonographic detection of a polyp by TVS with or without SHG enables its hysteroscopic removal and cure of bleeding. Conversely, endometrial biopsy is sufficient if the endometrium appears diffusely thickened on SHG.

Approximately 30% of patients who receive HRT may experience bleeding during the first few months of treatment. It is also not uncommon for some patients who receive cyclical HRT to have bleeding in days 12 to 14 of the pseudocycle related to decreasing amounts of progesterone. Bleeding outside of these times can be associated with endometrial pathology. Transvaginal sonography can help assess the endometrium for abnormal thickness and/or texture in these patients.

Endometrial evaluation by TVS is helpful in patients undergoing tamoxifen treatment because this medication is associated with an increased risk of hyperplasia and cancer. However, not all patients with thickened endometrium on TVS will have positive biopsies. This may be attributed to a sampling error or the fact that cystic atrophy of the endometrium can produce multiple interfaces that appear to thicken the sonographic depiction of the endometrial "stripe." In addition, tamoxifen exposure may be associated with the development of punctuate cystic areas in the inner myometrium, thought to represent reactivated adenomyosis.³

Recently, three-dimensional (3D) sonography has provided a means to better visualize the entire endometrium due to its ability to depict it in a coronal plane. Three-dimensional ultrasonography (US) can be performed using a freehand sweep or, better, with an automated sweep transducer. With either technique, multiplanar images are obtained including one in the coronal plan. The angle of the scan plane can be altered to best depict the entire endometrial width. For a more detailed discussion of 3D US in gynecology, please refer to Chapter 47.

The endometrium has a nongeometric shape. As such, measurement errors can occur if the endometrium is not evaluated in proper plane. In addition, up to a 2-mm interobserver error involved in its measurement has been reported.^4,5 Overmeasurement may occur if the endometrium is measured obliquely on a short-axis image. Inclusion of the hypoechoic inner myometrium may also erroneously enlarge measurement of the endometrial thickness. Therefore, every effort should be made to obtain the endometrial bilayer in its true long axis.

The texture and thickness of the endometrium must be evaluated. Significant changes in texture of the endometrium can also be observed in women of child-bearing age from relatively hypoechoic in the early follicular phase to echogenic in the midsecretory phase. Enhanced through transmission can be seen in the secretory or luteal phase endometrium due to increased fluid content (edema) of the stroma. The sonographic texture of the endometrium of normal postmenopausal women is typically mildly echogenic compared with the spiral (middle) layer of the myometrium.

Evaluation of the endometrium begins after sonographic identification of the long axis of the uterus. Assessment of the arrangement of the myometrial muscle bundles into the inner, spiral or middle, and outer muscle bundle layers may be helpful to determine that the endometrium is imaged in its true long axis (Figure 34-1). The optimal measurement of the endometrium is obtained in its long axis in anteroposterior (AP) dimension, or "height." This measurement excludes the hypoechoic inner myometrium and represents the bilayer thickness. Once the endometrium is imaged in its thickest sagittal plane, the probe should be turned 90 degrees and endometrium can be imaged in short axis. This is a more variable measurement and may image the endometrium obliquely, resulting in overestimation of endometrial thickness. One can observe the endometrium extending into the area of the tubal ostia, particularly during the secretory phase. This provides an anatomic landmark for assessment of the endometrium. The endometrial "width" describes the thickness of the endometrium in the coronal plane at the level of the tubal ostia. If taken as the "height," it can lead to overestimation of the endometrial thickness. However, this measurement may be helpful if an endometrial volume is calculated by using the length, AP dimension (height), and width of the endometrium.

The sonographer must be skilled at empirically adjusting the scan plane according to the uterine flexion (angle of the fundus relative to the corpus) and version (angle of the cervix relative to the vagina). For very anteflexed uteri, the probe must be placed in the posterior fornix with a sharp anterior angulation; for a retroflexed uterus, the probe needs to be directed sharply posteriorly. Assessment of the endometrium may be difficult in patients with fibroids that displace the endometrium.

Our study showed that 3D US is most useful if an abnormality is suspected in two-dimensional (2D) US; therefore, 3D US is not required in studies that appear normal on 2D. Three-dimensional US is particularly helpful in depicting polyps, submucosal fibroids, and adhesions.⁴

The endometrial thickness and texture seen in women of child-bearing age is quite different from that seen in postmenopausal women. (Figures 34-1,34-2, and 34-3) In a woman of child-bearing age, the endometrium consists of an outer basal layer that is not shed and an inner functional layer that has stromal and glandular elements. Refluxed mucus from the cervix typically produces a thin and echogenic median echo. In contrast, the endometrium of most postmenopausal women is comprised mostly of the basal layer and should be thin (<5 mm) and regular in texture.

In the patient of child-bearing age, the endometrium is shed during the first 3 to 5 days of the cycle. During this time, the endometrium may appear as an irregular, broken, echogenic interface, with focal hypoechoic intraluminal collections representing unclotted blood and sheets of sloughing endometrium. Organized clots may appear as echogenic intraluminal tissue, but are not attached to the basal layer like polyps. It is interesting to note that the subendometrial contractions directed from the fundus to the cervix can be observed during menses on real-time scanning if the videotaped segment is played back in fast-forward mode. After menses, the endometrium gradually thickens and becomes isoechoic to the surrounding myometrium. At this stage, the endometrium measures approximately 4 to 6 mm in bilayer thickness. In the periovulatory period, the endometrium may demonstrate a multilayered appearance consisting of the outer echogenic basal layer in the inner hypoechoic functionalis layers. The innermost echo represents refluxed mucus and has a variable echogenicity depending on the amount of refluxed mucus. The periovulatory endometrium usually measures between 5 and 8 mm in bilayer thickness and may contain a sliver of intraluminal fluid. At this time, subendometrial contractions can be seen directed from the cervix to the fundus. These peristaltic waves have a role in sperm transport. Under the influence of progesterone, the endometrium becomes echogenic relative to the basal layer. The increased echogenicity is probably related to stored mucus and glycogen within the more tortuous glands, and to stroma edema that produces more interfaces and is enhanced through transmission. The luteal phase endometrium can measure up to 14 mm in bilayer thickness.

Normal endometria in postmenopausal women are thin (<6 mm) and regular. In fact, the thickness of the endometrium can be related to the years postmenopausal in that the endometrium becomes much thinner as the postmenopausal patient ages.^5,6 In some instances, the echogenic interface in postmenopausal patients may be related to trapped mucus or secretions rather than to the endometrium itself.

In the patient undergoing HRT, the endometrium reflects the type of therapy administered to the patient. The patient who receives unopposed estrogen may have a thicker endometrium than those who take combined estrogen and progesterone.⁷ The sonologist should be aware that the endometrium in women taking cyclic estrogen and progesterone may differ up to 3 mm, depending on which part of the pseudocycle is being examined. The thinnest endometrium is usually observed after the progesterone phase (day 12); the thickest is later in the cycle after maximal estrogen exposure. Initial experience with a selective estrogen receptor modulator (SERM), such as raloxifene or Evista®, or idoxifene, indicate that the endometrium typically maintains a thin appearance. However, in some cases the endometrium may appear thickened, related probably to cystic atrophy. This is discussed in greater detail in the section on serial monitoring. Women on tamoxifen should demonstrate a thin (<6 mm) and regular endometrium. Changes in texture may require further evaluation by SHG and/or biopsy.

On 3D US obtained in the coronal plane, the entire surface of the endometrium is depicted including the endometrium that invaginates into the cornua.

There are a variety of causes of postmenopausal bleeding including atrophy, hyperplasia, fibroids, and cancer. (Figures 34-4,34-5,34-6,34-7, and 34-8) Atrophy is the most common cause, producing a thin (<5 mm) and regular endometrium.⁸ Many studies have indicated that an endometrium measuring less than 5 mm is usually due to atrophy or, on biopsy, tissue insufficient for diagnosis.⁹ When thicker than 5 mm, a variety of pathologies may be present including hyperplasia and cancer. Small polyps demonstrate increased echogenicity, particularly when contrasted to the surrounding endometrium. It may be helpful to rescan pre- or perimenopausal patients who have equivocal sonographic endometrial findings during the follicular phase when the endometrium should be thin. In these cases, the echogenic polyp may stand out against the hypoechoic surrounding endometrium. Polyps may contain punctuate cystic areas representing distended glands.

Hyperplasia cannot be consistently differentiated from endometrial cancer based on its sonographic characteristics. However, cancer is usually associated with markedly (approximately 10 mm) thickened and irregular endometrium; some cancers can be shown to invade and/or disrupt the inner myometrium hypoechoic layer and are more commonly associated with more advanced endometrial cancer. Hyperplasia typically produces diffuse thickening of the endometrium or appears as numerous echogenic masses within the endometrial lumen representing polyps. It should be remembered that rarely foci of cancer may also be present in polyps.

On TVS, endometrial cancer typically appears as thickened (>10 mm) and irregular endometrium. Disruption of the hypoechoic layer, representing the inner myometrium, may be observed in invasive cancer.

The most important role of sonography in evaluating women with postmenopausal bleeding is distinguishing those patients who may forgo endometrial biopsy. Specifically, women with postmenopausal bleeding in endometria thinner than 5 mm are very unlikely to have a pathologic cause for their bleeding. If a question concerning the endometrial thickness or texture remains, SHG should be performed.

Three-dimensional US affords depiction of polyps, submucosal fibroids, and adhesions. As on 2D US, polyps are typically echogenic when compared to surrounding endometrium and may contain punctuate cystic areas, which represent dilated glands within the polyp. Color Doppler sonography (CDS) frequently shows a feeding vessel into the base of the polyp. When compared to polyps, submucosal fibroids tend to be hypoechoic and may have a broader base than the typically thin pedicle of a polyp.¹⁰ Fibrous adhesions are typically hypoechoic interfaces that transverse the more echogenic endometrium.

Dysfunctional uterine bleeding is typically the result of anovulation. (Figure 34-5) These patients may have luteal phase defects associated with decreased progesterone production and withdrawal bleeding. Bleeding may also be associated with fibroids due to denuding of the endometrial surface. The intramural or intraluminal extent of fibroids is best determined with SHG or 3D ultrasound. See Chapter 38 on SHG or Chapter 47 on 3D ultrasound for further discussion.

Transvaginal sonography provides a means for the assessment of the endometrium and patients receiving HRT, SERM treatment, and/or tamoxifen, particularly in women who experience bleeding (Figures 34-4,34-5, and 34-6).¹¹ In patients on tamoxifen, the endometrium may appear thickened (>6 mm). This finding is not always associated with histologic abnormality because tamoxifen may be associated with cystic atrophy, producing increased echogenicity and apparent increased thickness on TVS.¹² In women on HRT, excessive thickening may be related to hyperplasia with or without polyps. In this patient group, it is important to perform SHG if there is a question of abnormal endometrial pathology based on the initial TVS. One should be aware that the endometrium in women taking cyclical HRT may vary up to 3 mm depending on which phase of the pseudocycle that they are imaged.⁷ Therefore, the best time to perform TVS for endometrial assessment in these patients is immediately after the progesterone portion of their pseudocycle.

We were recently involved in a multicenter study that assessed the endometrial effects of a newly developed SERM. Before selecting subjects for this study, 1733 postmenopausal women were scanned and had endometrial biopsy (Figure 34-9). Figure 34-10 shows the distribution of histology and thickness found. A 6-mm cutoff for inclusion was chosen. Of the subjects with abnormal histology, Figure 34-11 shows the thickness in each group and Figure 34-12 shows interval changes in endometrial thicknesses.^13,14 Transvaginal sonography was shown to have very high negative predictive value (98%) but relatively low (32%) positive predictive value. This emphasizes the use of SHG in equivocal or nondiagnostic cases. With the use of SHG, one can differentiate cysts contained within the endometrium that probably arise from cystic atrophy¹⁵ from cysts that are present in the inner myometrium, probably representing reactivated adenomyosis. Neither of these findings was associated with significant endometrial abnormalities on biopsy.

Determination of the extent or presence of invasion has a role in distinguishing those patients with cancer who can be treated by hysterectomy versus those who require more extensive surgery (Figure 34-8). Even though the role of TVS for detection of endometrial cancer is limited, TVS has a role in distinguishing deeply invasive from superficially invasive endometrial cancers.^16,17 Noninvasive tumors can be differentiated from invasive ones by close inspection of the subendometrial hypoechoic layer, which represents the inner myometrium. Some have used SHG to determine the extent of myometrial invasion.¹⁸ However, there is a theoretic risk of reflexing cancer cells retrograde into the tubes and peritoneum, although it is unlikely that these cells could metastasize. Some have suggested that the hypotonic saline used for SHG could lyse cells anyway. There are reports that CDS may show more vascularity within endometrial tumors than in large polyps; however, this finding would be difficult to objectify. Fibroids and/or adenomyosis may distort the endometrium, making its determination of invasion difficult. In these cases, magnetic resonance imaging with contrast should be used.

Transvaginal color Doppler sonography has a limited role in the evaluation of endometrial disorders (Figure 34-13). Initial experience suggests that both hyperplasia and cancer result in increased vascularity. Transvaginal color Doppler sonography may be used to identify the vascular pedicle of a polyp.^19,20

Abnormalities in endometrial texture can be observed in a variety of conditions (Figures 34-6 and 34-7). For example, depo provera may cause echogenic endometrial foci, probably representing areas of endometrial necrosis. Echogenic foci in the inner myometrium may be secondary to previous curettage. It is important to distinguish cystic textural abnormalities that arise from the endometrium from those in the inner myometrium. Ones contained within the endometrium may be within a polyp, representing distended glandular elements, whereas those in the inner myometrium may represent reactivated adenomyosis, especially in women taking tamoxifen.

Transvaginal sonography provides an excellent means for the assessment of the endometrium. By assessment of its texture and thickness, TVS provides more information about the endometrium than is obtained by merely measuring its thickness.²¹ It can determine which patients would benefit from endometrial biopsy versus those who do not require this procedure. Transvaginal sonography has a negative predictive value (approximately 98%) but lower positive predictive value (approximately 30%) because of the occasional lack of correlation between thickness and textural abnormalities with abnormal histology. It has an important role in evaluating patients on tamoxifen and certain forms of HRT who experience unexpected vaginal bleeding. Measurement of endometrial volume (3D) may be better than assessment of thickness (2D) for the detection of endometrial cancer in symptomatic postmenopausal women.^22,23 Sonohysterography is an important adjunct to the initial evaluation of the endometrium with TVS and is discussed in detail in a subsequent chapter. Three-dimensional US also affords global depiction of the endometrium due to the ability to provide images in the coronal plane.⁴ This technique's reproducibility indicates that it can be used reliably as a means to determine endometrial volume and vascularity.²⁴

Highlighted References