Transvaginal ultrasound plays a pivotal role in the management of non-pregnant women with abnormal vaginal bleeding. No other imaging technique has a role in the triage of these women. In women with postmenopausal bleeding, ultrasound is used to categorise women as at low or high risk of endometrial cancer, and the result of the ultrasound examination is the basis for further management. In women with abnormal vaginal bleeding before the menopause, the role of ultrasound is less clear. This is because some common causes of abnormal vaginal bleeding before the menopause cannot be diagnosed with ultrasound, such as infection, dysfunctional bleeding, or problems with intrauterine contraceptive devices or contraceptive pills. Nonetheless, transvaginal ultrasound may also sometimes be helpful in women with abnormal vaginal bleeding before the menopause. In this chapter, I present ultrasound findings in women with endometrial cancer, endometrial polyps, endometrial hyperplasia, adenomyosis, uterine myomas, including submucous myomas and leiomyosarcoma, and describe ultrasound-based triage of women with postmenopausal bleeding.
Introduction
The causes of abnormal vaginal bleeding differs between pre- and post-menopausal women. Endometrial cancer and other endometrial malignancies are relatively common causes in postmenopausal women, but are rare before menopause. Myomas and adenomyosis may cause abnormal bleeding before menopause but rarely thereafter. Endometrial polyps, hyperplasia, and uterine leiomyosarcomas may explain abnormal vaginal bleeding before and after menopause, but leiomyosarcomas are extremely rare. Infection, dysfunctional bleeding, or problems with contraceptives are common causes of abnormal vaginal bleeding before the menopause.
Transvaginal ultrasound plays a pivotal role in the management of non-pregnant women with abnormal vaginal bleeding. No other imaging technique has a role in the triage of these women. In women with postmenopausal bleeding, ultrasound is used to categorise women at low or high risk of endometrial cancer, and the result of the ultrasound examination is the basis for further management. In women with abnormal vaginal bleeding before the menopause, the role of ultrasound is less clear. This is because some common causes of abnormal vaginal bleeding before the menopause cannot be diagnosed with ultrasound, such as infection, dysfunctional bleeding, or problems with intrauterine contraceptive devices or contraceptive pills. Nonetheless, transvaginal ultrasound may sometimes be helpful in these women.
A gynaecological ultrasound examination in a woman with abnormal vaginal bleeding must be preceded by a thorough history and a careful speculum examination and gynaecological palpation. The role of ultrasound is to detect pathology not detectable at a clinical examination (e.g. endometrial pathology, small submucous myomas, adenomyosis, cancer of the urinary bladder, or small hormone-producing ovarian tumours). The ultrasound examination is also used to confirm or refute a diagnosis suspected on the basis of abnormal findings at palpation (e.g. uterine intramural or subserous myomas, or adnexal masses). The clinician then needs to decide if an abnormal ultrasound finding is the likely cause of the abnormal bleeding or if it is an incidental finding unrelated to the woman’s symptoms.
The examination technique to be applied when scanning the uterus, and the terminology to be used when describing ultrasound images of the endometrium and the uterine cavity, have been described elsewhere . An ultrasound examination carried out because of abnormal vaginal bleeding should also include examination of the adnexa and the urinary bladder, because abnormal bleeding may be explained by a hormone-producing ovarian tumour or a tumour in the urinary bladder (i.e. the woman confusing bleeding from the urinary tract with vaginal bleeding).
Imaging techniques in the management of postmenopausal bleeding
Ultrasound plays an important role in the management of women with postmenopausal bleeding. About 10% of women with postmenopausal bleeding have endometrial cancer, but as many as 50% may not have any endometrial pathology at all . Strong scientific evidence supports transvaginal ultrasound examination and measurement of endometrial thickness to discriminate between women with postmenopausal bleeding that are at high risk of endometrial cancer and those that are at low risk. The risk of finding an endometrial cancer in a woman with postmenopausal bleeding and endometrial thickness as measured by ultrasound 4 mm or less is low. In a meta-analysis that included almost 6000 women with postmenopausal bleeding, this risk was estimated to be about 1 in 100 women not using hormone replacement therapy and about 1 in 1000 women using hormone replacement therapy . It is considered safe to refrain from endometrial sampling to obtain a histological diagnosis in women with postmenopausal bleeding and endometrial thickness 4 mm or less . This endometrial thickness cut-off is applicable to users and non-users of hormone replacement therapy . Although it has been suggested that it would be safer to use a cut-off of 3 mm to exclude endometrial cancer in women with postmenopausal bleeding , the 4-mm cut-off prevails in clinical practice.
In rare cases, a cervix cancer not detectable at speculum examination or palpation, a bladder tumour ( Fig. 1 ) or an ovarian tumour (e.g. a granulosa cell tumour) may be detected at the transvaginal ultrasound examination. Imaging of cervix cancer is described in chapter 7 of this issue.
How to measure endometrial thickness at transvaginal ultrasound
Endometrial thickness is measured on a sagittal scan through the uterus. The uterus is scanned from one side to the other, and the endometrial thickness is measured where it seems to be at its thickest from its outermost border on one side to that on the other . The endometrium must not be measured on a transverse scan, because a transverse scan may be an oblique scan and, if so, will yield too large a measurement. If spontaneous fluid fills the uterine cavity, each endometrial layer is measured separately, and the two measurements are added . The measurement technique is shown in Fig. 2 .
In about 6–7% of women with postmenopausal bleeding, the endometrium is not clearly visible and so is not measurable . In this situation, saline-contrast sonohysterography should be carried out (see below).
Can the estimation of risk of endometrial malignancy be refined in women with postmenopausal bleeding and endometrial thickness 5 mm or more?
A differentiation of risk in women with endometrial thickness 5 mm or more allows individualised management. A woman at relatively low risk of endometrial cancer, despite her endometrium being thick, will be managed differently from a woman at extremely high risk. Clinical information, the grey-scale ultrasound morphology of the endometrium, and the vascularisation of the endometrium (as assessed by colour Doppler or power Doppler ultrasound) add information to endometrial thickness when estimating the risk of endometrial malignancy in women with postmenopausal bleeding and endometrial thickness 5 mm or over . Irregular echogenicity of the endometrium ( Fig. 3 ), irregularly branching vessels, densely packed vessels, or colour splashes in the endometrium at power Doppler examination ( Fig. 3 ) increase the risk of malignancy . High colour content in the endometrial scan at power or colour Doppler examination is also a sign of malignancy ( Fig. 3 ). The older the woman, the thicker the endometrium, and the higher the colour content of the endometrial scan, the higher the risk of malignancy; however, if the woman uses hormone replacement therapy, the risk decreases .
Mathematical formulas to calculate the individual risk of malignancy in women with postmenopausal bleeding and endometrial thickness 5 mm or over have been published . No studies have been published describing their diagnostic performance on prospective validation. Therefore, it is too early to introduce these models into clinical practice.
Saline-contrast sonohysterography
Infusion of saline into the uterine cavity during transvaginal scanning (saline-contrast sonohysterography, saline infusion sonography or hydrosonography) clarifies whether or not focal lesions are present in the uterine cavity . A focal lesion is anything that protrudes into the uterine cavity above the baseline endometrial surface ( Fig. 4 ). Unless the cervical canal is stenotic, saline-contrast sonohysterography is easy to carry out. A thin plastic catheter (without a balloon) with a sterile 20 ml syringe filled with sterile saline attached to it is introduced into the uterine cavity through the cervical canal. Before introduction, the catheter must be flushed with saline to expel all air (air reflects the ultrasound beams making the ultrasound image difficult to interpret). Then the vaginal ultrasound transducer is introduced into the vagina and a few millilitres of saline is infused into the uterine cavity during scanning. If the cervical canal is stenotic, it may be necessary to use both a tenaculum and a small uterine sound before the catheter can be introduced into the uterus. Saline-contrast sonohysterography fails in 10–20% of all women with postmenopausal bleeding .
Virtually all endometrial pathology grows focally in the uterine cavity . If no focal lesions are present in the uterine cavity, the odds of malignancy decrease 20-fold, and the odds of any endometrial pathology decrease 30-fold . Thus, a smooth endometrium outlining the uterine cavity at saline-contrast sonohysterography is a strong sign of normality.
Irregular focal lesions in the uterine cavity at saline-contrast sonohysterography in women with postmenopausal bleeding and endometrial thickness 5 mm or over is a strong sign of endometrial malignancy ( Fig. 4 ).
Because most focal lesions cannot be removed at all, or can only be partially removed if a blind endometrial sampling technique is used, such as Pipelle ® (Prodimed, Neuilly en Thelle, France), Endorette ® (MedScand AB, Malmö, Sweden), or dilatation and curettage, focal lesions should be hysteroscopically resected under direct visual control . This is to ensure that a representative sample is obtained.
Staging of endometrial cancer and discrimination between high-risk and low-risk endometrial cancer
If endometrial cancer is suspected at ultrasound examination when the woman first consults with her bleeding, the spread of the cancer can be assessed at that primary consultation using a combination of vaginal and abdominal ultrasound. Moreover, the likelihood of a specific histological type of cancer (high risk or low risk) can be estimated. Computed tomography and magnetic resonance imaging can also be used for staging of endometrial cancer. Staging of endometrial cancer and discrimination between high-risk and low-risk endometrial cancer is discussed in chapter 6 in this issue.
Ultrasound-based triage of women with postmenopausal bleeding
On the basis of the information provided above, women with postmenopausal bleeding can be managed as described in Fig. 5 . After a thorough history and clinical examination, a vaginal smear is taken to try to rule out cervical cancer (because small cervical cancers are unlikely to be detectable with transvaginal ultrasound). Then, a transvaginal ultrasound examination is carried out with measurement of endometrial thickness. If the endometrial thickness is 4 mm or less, the woman is dismissed without any endometrial sample being taken. If the endometrium measures 5 mm or more, saline-contrast sonohysterography is carried out. If it reveals focal lesions, the woman is scheduled for operative hysterosocopy with removal of the focal lesion(s) under direct visual control. If no focal lesions are present, an endometrial sample can be taken using an outpatient endometrial sampling device. If this fails, the woman should be scheduled for dilatation and curettage in anaesthesia or analgesia.
If the endometrium is not seen well, and so cannot be reliably measured, saline-contrast sonohysterography should be conducted to clarify the situation. If saline-contrast sonohysterography fails, the woman should undergo diagnostic hysteroscopy in anaesthesia (or analgesia).
Can pathology other than endometrial cancer be diagnosed with transvaginal ultrasound in women with postmenopausal bleeding?
Benign endometrial polyps
Benign endometrial polyps are often found in women with postmenopausal bleeding . They are then supposed to be the cause of the abnormal bleeding, even though this is not necessarily the case . The typical ultrasound appearance of a benign endometrial polyp is thick hyperechogenic endometrium with or without regular small cysts (cysts are common in atrophic polyps where the glands are cystically dilated ) and the presence of a ‘bright edge’ ( Fig. 6 ). The bright edge is explained by the interface between the polyp (or any other focal lesion in the uterine cavity) and the endometrium . When these ultrasound signs of endometrial polyp were prospectively validated in women with postmenopausal bleeding and endometrial thickness 5 mm or over, however, they did not perform very well: sensitivity 49% (21/43), specificity 81% (50/62) .
At Doppler ultrasound examination, an endometrial polyp is characterised by the presence of a ‘pedicle artery’ (also called ‘feeding vessel’), which is one big vessel seen to enter the endometrium from the surrounding myometrium ( Fig. 6 ). In women with postmenopausal bleeding, Timmerman et al. found that the pedicle artery sign had a sensitivity for detecting endometrial polyps of 78% (47/60) and the specificity 88% (88/100). Therefore, the pedicle artery sign had at most moderate ability to correctly identify polyps in this patient group (positive likelihood ratio 6.5 and negative likelihood ratio 0.25). In women with postmenopausal bleeding, Alcazar et al. reported the presence of a single vessel penetrating into the endometrium from the myometrium (corresponding to the pedicle artery sign) to have a sensitivity of 97% (33/34) for endometrial polyps and a specificity of 86% (42/49). Therefore, in the study by Alcazar et al. , the pedicle artery sign performed better than in the study by Timmerman et al. . To the best of my knowledge, the ability of the pedicle artery sign to correctly identify polyps in postmenopausal women with vaginal bleeding has not been prospectively validated. On prospective external validation in women with abnormal vaginal bleeding, either before or after menopause, however, the pedicle artery as a sign of endometrial polyp had a sensitivity of 67% (26/39) and a specificity of 98% (57/58) . This corresponds to moderate diagnostic performance.
The typical appearance of an endometrial polyp at saline-contrast sonohysterography is a polypoid focal lesion with regular hyperechogenic echotexture, with or without regular small cysts, and with a smooth surface ( Fig. 6 ). When these criteria of endometrial polyp were prospectively validated in women with postmenopausal bleeding and endometrial thickness 5 mm or more, they did not perform well: sensitivity 79% (26/33), specificity 76% (34/45) .
Some polyps may contain foci of malignancy , but it is not known if such polyps manifest other ultrasound features than benign polyps, even though it has been suggested that, in asymptomatic women, polyps with malignant changes are larger than benign polyps . Therefore, in women with postmenopausal bleeding and endometrial thickness 5 mm or over, all focal lesions, irrespective of their ultrasound appearance at saline-contrast sonohysterography, should be hysteroscopically resected under direct visual control to ascertain their complete removal.
Endometrial hyperplasia
The ultrasound characteristics of endometrial hyperplasia have been described for women of any age not separating pre- from post-menopausal women and not separating asymptomatic from symptomatic women . In the study by Davis et al. , the ultrasound characteristics of endometrial hyperplasia were described as thick, hyperechogenic endometrium (sometimes containing small cysts), with a polypoid surface at saline-contrast sonohysterography . To the best of my knowledge, these ultrasound criteria of endometrial hyperplasia have not been prospectively validated. Ultrasound images of endometrial hyperplasia from my own clinical practice are shown in Fig. 7 .
Submucous myomas
Submucous myomas are sometimes detected at transvaginal ultrasound examination of women with postmenopausal bleeding . The typical ultrasound appearance of a submucous myoma is a solid tumour protruding into the uterine cavity from the surrounding myometrium and with the same echogenicity as the surrounding myometrium ( Fig. 8 ). When these ultrasound criteria of submucous myoma were applied in women with postmenopausal bleeding, and endometrial thickness was 5 mm or more, they did not perform well at unenhanced ultrasound examination. They were very specific (specificity 97% [96/99]) but not sensitive (sensitivity 33% [2/6]). This means that a submucous myoma could be present even if the typical ultrasound signs of submucous myoma were absent . On the other hand, saline-contrast sonohysterography was a good method for diagnosing submucous myomas: sensitivity 80% (4/5), specificity 99% (72/73) . In some cases, a submucous myoma can be seen to be covered by endometrium at saline-contrast sonohysterography ( Fig. 8 ).
It has been suggested that submucuous myomas are surrounded by a ring of colour at colour or power Doppler ultrasound. On prospective validation in women with abnormal vaginal bleeding (proportion of pre- and post-menopausal women not reported), the colour ring sign had a sensitivity for diagnosing submucous myoma of 67% (26/39) and a specificity of 98% (57/58) (i.e. it manifested moderate diagnostic performance).
Uterine leiomyosarcoma
Although leiomyosarcoma is a rare disease , in some cases the cause of postmenopausal bleeding is a uterine leiomyosarcoma. The largest study published comparing the ultrasound appearance of uterine leiomyosarcomas with that of benign leiomyomas includes eight leiomyosarcomas and 225 benign leiomyomas . The menopausal status of the women in the study cited was not reported. The results showed that leiomyosarcomas were more often solitary lesions than benign leiomyomas (100% v 53%) and that the leiomyosarcomas more often contained cystically degenerated areas (50% v 14%), and more often manifested marked central vascularisation at power Doppler examination (87.5% v 3%). Leiomyosarcomas were generally more richly vascularised than benign leiomyomas, and they were larger, seven of eight being 8 cm or more in diameter. The combination of solid lesion 8 cm or more in diameter with ultrasound signs of cystic degeneration and marked central vascularisation was a specific (but not sensitive) ultrasound sign of leiomyosarcoma (sensitivity 50%, specificity 99%). These ultrasound signs of leiomyosarcoma need to be prospectively validated to better estimate their ability to distinguish leiomyosarcomas from benign leiomyomas. Ultrasound images of a benign uterine leiomyoma and of a leiomyosarcoma are shown in Fig. 9 .
The role of magnetic resonance imaging in the differential diagnosis between benign leiomyoma and malignant leiomyosarcoma is unclear. In one published study that included four uterine leiomyosarcomas and 41 benign leiomyomas, magnetic resonance imaging correctly diagnosed all four leiomyosarcomas, with no false – positive result, the criterion of leiomyosarcoma being ill defined margins of the tumour . In a more recently published study that included five women with leiomyosarcoma and 76 women with benign leiomyomas, diffusion-weighted magnetic resonance imaging was reported to discriminate between leiomyosarcoma and leiomyoma, with a sensitivity of 100% and a specificity of 94% . The diffusion-weighted magnetic resonance imaging criteria for classifying a uterine nodule as being at high or low risk of leiomyosarcoma remain to be prospectively validated.
Imaging techniques in the management of postmenopausal bleeding
Ultrasound plays an important role in the management of women with postmenopausal bleeding. About 10% of women with postmenopausal bleeding have endometrial cancer, but as many as 50% may not have any endometrial pathology at all . Strong scientific evidence supports transvaginal ultrasound examination and measurement of endometrial thickness to discriminate between women with postmenopausal bleeding that are at high risk of endometrial cancer and those that are at low risk. The risk of finding an endometrial cancer in a woman with postmenopausal bleeding and endometrial thickness as measured by ultrasound 4 mm or less is low. In a meta-analysis that included almost 6000 women with postmenopausal bleeding, this risk was estimated to be about 1 in 100 women not using hormone replacement therapy and about 1 in 1000 women using hormone replacement therapy . It is considered safe to refrain from endometrial sampling to obtain a histological diagnosis in women with postmenopausal bleeding and endometrial thickness 4 mm or less . This endometrial thickness cut-off is applicable to users and non-users of hormone replacement therapy . Although it has been suggested that it would be safer to use a cut-off of 3 mm to exclude endometrial cancer in women with postmenopausal bleeding , the 4-mm cut-off prevails in clinical practice.
In rare cases, a cervix cancer not detectable at speculum examination or palpation, a bladder tumour ( Fig. 1 ) or an ovarian tumour (e.g. a granulosa cell tumour) may be detected at the transvaginal ultrasound examination. Imaging of cervix cancer is described in chapter 7 of this issue.
