Ultrasound approaches

Transvaginal ultrasonography is the method of choice, as it results in the best image quality, due to the proximity of the probe to the gynaecological organs. However, a transabdominal scan may be required to assess the uterus in the presence of large fibroids or to evaluate the adnexa when these are enlarged and extending beyond the level of the pelvis. When a transvaginal scan is considered inappropriate (e.g. virginity, vaginismus or vaginal stenosis due to atrophy, surgery or pelvic irradiation) and if the transabdominal approach is inconclusive, a transrectal ultrasound examination should be considered.

During transvaginal ultrasonography the hand not holding the probe can be used to press on the abdomen and help optimize the image or assess mobility or pain [1]. Although a full bladder can be useful for transabdominal scanning, in many cases it is not required. A good technique is to identify the femoral vessels and follow these upwards to the iliac vessels and finally to the bifurcation of the aorta. An ovarian mass and any lymphadenopathy present can be assessed in this way. A gynaecologist learning to scan should be aware that an examination of the abdomen does not stop in the pelvis and the examiner should be familiar with the normal appearances of the organs of the upper abdomen and the likely sites of metastatic disease in the event of gynaecological malignancy. These might include the peritoneum, lymph nodes around the large vessels, the spleen, liver or the omentum. Ascites may be present in the upper abdomen.

In the event of suspected intra‐abdominal blood loss, the presence of fluid above the fundus of the uterus or in the utero‐vesical pouch is significant. A further marker of serious intra‐abdominal bleeding is the presence of fluid in Morrison’s pouch between the liver and the kidney [9].

The uterus should be assessed in the sagittal plane, including detailed evaluation of the cervix, the body of the uterus, the fundus and the endometrial cavity. Views obtained in the transverse plane may help in defining the location of focal pathology. The detection of congenital uterine abnormalities is best achieved using three‐dimensional ultrasound whereby the unique coronal view allows most anomalies to be characterized relatively easily [10].

The endometrial thickness, the maximal measurement across the lumen of the endometrial cavity in the sagittal plane including both endometrial layers (double endometrial thickness), is an important assessment. In premenopausal women, endometrial thickness generally varies from 4 to 8 mm in the follicular phase and from 7 to 14 mm in the secretory phase. A sonographic examination in this population should preferably be performed in the early proliferative phase (cycle day 4–6) [3]. Recent consensus documents have outlined proposed terminology to describe the endometrium and myometrium, with any pathology that may be present [3,6].

Difficulties in assessment due to variations in uterine position (particularly when axial) or uterine rotation (endometriosis or adhesions) can be overcome by pressing on the abdomen with the non‐scanning hand, or by filling the bladder [3]. The ovaries can be visualized by tracking the vessels in the broad ligament in the transverse plane from the uterus and moving laterally to the pelvic wall nearby the external iliac vessels. It might be useful to apply abdominal pressure to displace superimposed intestines or to use a transabdominal approach if necessary [1].

Sonohysterography

Sonohysterography is the instillation of fluid through a catheter into the uterine cavity to act as a negative contrast agent to outline any focal intra‐cavity pathology. Sterile saline or gel may be used [3]. In saline instillation, a neonatal suction catheter may be used, although backflow may necessitate the use of a more expensive balloon catheter to obtain a more stable filling of the uterine cavity. The higher viscosity of gel results in a smaller instillation volume, which makes the procedure better tolerated by patients compared with saline instillation [11]. Moreover, transtubal spillage is less likely to occur and this may be an additional argument for the use of gel instead of saline, as endometrial malignancy is never completely excluded with ultrasonography. A small neonatal suction catheter (2 mm) is sufficient, but it might be helpful to warm the gel to 37 °C to promote flow through the catheter.

It might be prudent in fertile women to advocate contraception during the cycle when the sonohysterography procedure is planned, and some examiners advise that prophylactic antibiotics should be administered prior to the procedure in all potentially fertile women to avoid pelvic inflammatory disease [11,12].

Possible indications for sonohysterography include a thickened or irregular endometrium, poor views of the endometrium (e.g. due to axial position of uterus) and further characterization of focal intra‐cavity pathology such as polyps, submucous fibroids and adhesions (Fig. 36.1). There is a large amount of data to suggest that sonohysterography is comparable to hysteroscopy for the assessment of most focal endometrial pathology [13]. Ultrasound without saline instillation is significantly less accurate for this purpose [14]. In the event of uncertainty regarding the endometrial findings, sonohysterography is always a good option to consider. It is also important to remember that the endometrium in premenopausal women is a dynamic structure and so simply repeating a scan after menstruation will often clarify whether there is focal intrauterine pathology or not.

Hysterosalpingo‐contrast sonography and hysterosalpingo‐foam sonography

The gold standard test to assess tubal patency is laparoscopic chromoperturbation with methylene blue. However, this procedure is associated with the risks of laparoscopy under general anaesthesia. Therefore, as a less invasive radiography technique, X‐ray hysterosalpingography has been applied for decades in the diagnostic work‐up of the patient presenting with infertility [15]. At present, this exposure to radiation and risk of allergy to the contrast agent can be avoided by tracking the movement of a suspension of air bubbles through the fallopian tubes during hysterosalpingo‐contrast sonography [16]. In general, a balloon catheter is used. The contrast media are commercially available, but a cheaper option is to use an alternating injection of water and air in the uterine cavity and to sonographically detect air bubbles around the ovary and free fluid in the pouch of Douglas [17]. More recently, hysterosalpingo‐foam sonography was introduced, using ‘gel foam’, with advantageous physical properties compared with saline [18] (Fig. 36.2). Both hysterosalpingo‐contrast sonography and hysterosalpingo‐foam sonography are well tolerated by the patient, and show very good agreement in assessing tubal patency compared with laparoscopic chromoperturbation [18–20].

Ultrasound‐guided procedures

Ultrasound can guide different invasive procedures in gynaecology. Besides the well‐known application in reproductive medicine for oocyte retrieval, there is increasing interest in the field of gynaecological oncology. Ultrasound‐guided Tru‐cut biopsy has been shown to be an accurate and safe, minimally invasive technique to acquire tissue diagnosis in the management of advanced or recurrent gynaecological tumours or when pelvic metastatic disease from another primary origin is suspected [21,22].

In general, this procedure is performed with a core‐cut biopsy needle inserted through a needle guide mounted on the transvaginal probe. Local anaesthesia is not required in a transvaginal approach. A similar approach with an aspiration needle connected to a syringe or vacuum container allows the drainage of symptomatic pelvic cysts, on condition that there is no suspicion of malignancy. Examples are large peritoneal pseudocysts, simple ovarian cysts or tubo‐ovarian abscesses in pelvic inflammatory disease [23].

Abnormal uterine bleeding

Over recent years transvaginal ultrasonography has significantly improved our ability to accurately manage patients with abnormal uterine bleeding. Ultrasound findings are described according to international consensuses [3,6]. Other imaging modalities such as MRI are indicated as second‐stage tests in certain circumstances.

Premenopausal bleeding

In 2011, the International Federation of Gynecology and Obstetrics (FIGO) published the PALM‐COEIN system (polyp; adenomyosis; leiomyoma; malignancy and hyperplasia; coagulopathy; ovulatory dysfunction; endometrial; iatrogenic; not yet classified), which classifies the aetiology of abnormal uterine bleeding [24]. From these causes, imaging methods can only confirm or exclude anatomical abnormalities, i.e. the ‘PALM group’. However, these might or might not be the cause of the bleeding problem, as they may also be found in asymptomatic women. In the large majority of premenopausal women, abnormal uterine bleeding is associated with benign lesions such as an endometrial polyp or an intra‐cavitary fibroid, whereas endometrial cancer is an uncommon cause. In abnormal uterine bleeding, the usefulness of an ultrasound evaluation of the endometrial thickness to detect pathology is much less obvious in premenopausal women compared with postmenopausal women [25].

Endometrial polyps are a very common finding in women with abnormal uterine bleeding. On ultrasound, they are often hyperechoic and feature a ‘bright edge’ [3] (Fig. 36.3). In a large proportion of patients, the feeding blood vessels can be visualized on colour Doppler imaging, the so‐called ‘pedicle artery’ sign, which is pathognomonic for focal endometrial pathology [26] (Fig. 36.4). Fluid instillation should be considered if the endometrium is not well visualized on unenhanced ultrasonography, or if within a thickened endometrium no pedicle artery is detectable on colour Doppler examination [27]. MRI is not considered useful in the assessment of endometrial polyps [28].

Fibroids are the most common uterine tumours. Their location is described according to the FIGO leiomyoma classification, where the submucosal lesions (i.e. types 0–2) are the most likely to result in abnormal uterine bleeding [24]. Besides the number of fibroids and their size, the relationship to the endometrium is an important factor in choosing the proper surgical strategy should one be needed, and in particular for hysteroscopic or laparoscopic myomectomy. At ultrasonography, fibroids are generally well‐circumscribed round lesions within the myometrium or attached to it, often showing shadows at the edge of the lesion and/or internal fan‐shaped shadowing due to calcifications. On colour or power Doppler imaging, circumferential flow around the lesion is often visible (Fig. 36.5). However, some fibroids do not exhibit such typical features [6,29]. Three‐dimensional ultrasound may help in localizing the fibroid with respect to the uterine cavity. MRI and ultrasound seem to have similar ability to diagnose uterine fibroids, but MRI is superior to ultrasound in determining the exact location of a fibroid, particularly in a large uterus with multiple fibroids [30].

It is often very challenging to discriminate between fibroids and malignant tumours of mesenchymal origin, the so‐called leiomyosarcoma. Although several features at ultrasonography and MRI can raise suspicion of a uterine sarcoma, there are no pathognomonic features on any imaging technique [31]. On ultrasound, a sarcoma is generally solitary, large and oval‐shaped with heterogeneous echogenicity. There might be irregular anechoic areas due to internal necrosis or haemorrhage. On colour Doppler imaging, a sarcoma is more often highly vascularized, with prominent irregular intralesional vascularity more often seen than in fibroids. The vessels are of unequal size and exhibit irregular branching. Because of their rapid growth, most sarcomas are large at diagnosis. Myometrial lesions with a large diameter (e.g. >8 cm) should be managed with circumspection [6,29] (Fig. 36.6).

Findings in uterine leiomyosarcomas on MRI vary and include a lobulated mass of high‐signal intensity on T2‐weighted images, a sharply marginated mass of low signal intensity that closely resembles a leiomyoma, or a mass with focally infiltrative margins. Detection of scattered foci of haemorrhage or necrosis can suggest the diagnosis of uterine leiomyosarcoma. A consistent finding in uterine leiomyosarcomas is the absence of calcifications [32]. Presence of these features at imaging should discourage the clinician to select the patient for minimally invasive surgery with tissue morcellation in order to avoid the fragmentation and intra‐abdominal spread of malignant disease [29].

Note that computed tomography (CT) is unable to differentiate between different types of uterine pathology. In particular, it has been shown that CT is not able to differentiate between fibroids and sarcomas [31].

Another differential diagnostic difficulty with fibroids relates to their discrimination from adenomyosis. Adenomyosis may be present at one or more sites within the uterine wall or involve most of the myometrium, and may often be dispersed within the myometrium rather than forming a confined lesion, i.e. diffuse adenomyosis. On the other hand, it may be present in only one part of the myometrium, i.e. focal adenomyosis. In rare cases it may present as a large cyst (an adenomyotic cyst or cystic adenomyoma) [6].

The ultrasound features of adenomyosis are myometrial asymmetry, cystic areas within the myometrium, hyperechoic islands, fan‐shaped shadowing, echogenic subendometrial lines and buds, and blood vessels passing through the abnormal‐looking area as opposed to around the lesion as is seen with fibroids. More recently, the presence of an irregular or interrupted endometrial–myometrial junctional zone imaged using three‐dimensional ultrasound in the coronal plane has been reported to have a high diagnostic accuracy for adenomyosis [6,8]. Comparable with ultrasound, MRI can confidently diagnose adenomyosis, and can be used when transvaginal ultrasonography provides indefinite findings or when dealing with difficult cases with coexistence of other abnormalities (e.g. multiple fibroids) [33].

Pedunculated and broad ligament fibroids present their own problems. These principally relate to how confident the examiner is that he or she is not looking at a solid ovarian lesion. This is important given the higher likelihood of malignancy in solid masses thought to be ovarian fibromas [34]. The demonstration of two normal ovaries is the obvious solution to this problem. The use of Doppler to demonstrate that the blood supply originates from the uterus may identify the lesion as uterine in origin and acoustic shadowing is a reassuring sign.

Adnexal masses

When ultrasound evaluation of a pelvic mass has excluded non‐adnexal pathology (e.g. pedunculated fibroids, see section on premenopausal bleeding), it is very important to discriminate between pathological and functional findings. Since fibroids are only found in premenopausal women, they can be expected to resolve spontaneously and rarely require surgical intervention.

Follicular cysts originate from anovulatory follicles. Usually they appear as unilocular and thin‐walled lesions with anechoic contents. They rarely exceed 8–10 cm in diameter and typically resolve spontaneously within 6 weeks [39]. Corpus luteum cysts are formed following ovulation. They are thick‐walled cysts that often show circumferential blood flow, described on ultrasound as the ‘ring of fire’. The cyst content may have a ‘spider‐web’ appearance or contain blood clots, resembling solid components (Fig. 36.7). In these cases, Doppler examination (a clot will have no blood flow) and ‘pushing’ the lesion with the probe (a clot will have typical jelly‐like movement) can be used to help differentiate between clots and solid parts [1]. In most cases, haemorrhagic cysts resolve within 6–12 weeks without intervention [39]. If a lesion does not meet the clinical and ultrasound features of functional changes or persists during follow‐up, further investigation is recommended.

Distinguishing between benign and malignant pathology is important both to lessen unnecessary anxiety and to select the optimal patient‐tailored management. Depending on the clinical presentation of the patient, benign pathology may be best treated conservatively or in a general gynaecology unit using a minimally invasive approach. On the other hand, masses suspicious for malignancy should be referred to specialized units for appropriate staging and treatment, for example in the case of ovarian cancer where this is known to improve survival [40].

Certain types of ovarian tumours exhibit characteristic ultrasound features which make them very easy to recognize. Mature teratomas or dermoid cysts are the most frequently encountered non‐functional ovarian masses in premenopausal women. Usually they appear as unilocular cysts with mixed echogenicity, due to the presence of different tissue components such as fat, bone, hair and fluid. Acoustic shadowing is typically present and often prevents the cyst being completely visualized (‘tip of the iceberg’ phenomenon). Different hyperechoic tissues often pack together into a Rokitansky nodule (Fig. 36.8) and the presence of hair is often seen as multiple stripy hyperechoic interfaces (‘dermoid mesh’). In general, vascularity is minimal [39]. Endometriomas are typically unilocular tumours with echogenic content representing old blood (‘ground glass’ appearance) and limited vascularity (Fig. 36.9). However, atypical features may be present. Debris within the cyst may give the impression of solid components. It is important to consider the age of the patient when a lesion with the features of an endometrioma is observed, as in postmenopausal women the risk of malignancy in such lesions is significant [41].

Subjective assessment of the morphology and vascularity, also called ‘pattern recognition’, by an experienced operator is known to be the best‐performing strategy for characterizing adnexal masses. Besides predicting whether a mass is likely to be benign or malignant, in most cases a reasonably accurate evaluation of the likely histological outcome is also possible [42,43]. However, as pattern recognition is very user‐dependent [44], more objective methods are needed to allow accurate ultrasound evaluation by less‐experienced examiners. The risk of malignancy index (RMI) is one such method that has been in use since the 1990s. It combines ultrasound information with menopausal status and serum CA‐125 to produce a numerical score, where a result of 200 or more is usually considered to indicate malignancy [45]. As the RMI score is largely influenced by the serum CA‐125 level, it suffers from the inherent problems associated with that marker: a relatively low sensitivity for early‐stage and borderline disease, especially in premenopausal women [46]. Despite accumulating evidence in favour of more recent ultrasound‐based models, many national guidelines for the management of ovarian masses still advocate the use of RMI.

In 2000, the International Ovarian Tumour Analysis (IOTA) group published a consensus paper in order to standardize the terms, definitions and measurements used to characterize ovarian pathology [5]. With more than 20 000 patients included, this prospective multicentre project aims to develop diagnostic algorithms to classify adnexal masses. In 2008, the IOTA group published the Simple Rules, which are based on five ultrasound features suggestive of a benign lesion (B‐features) and five features suggestive of a malignant lesion (M‐features) [4]. The Simple Rules have been validated extensively; even in the hands of less‐experienced examiners, they retain excellent performance in discriminating between benign and malignant pathology [47].

The IOTA Simple Rules have become very popular because they do not require access to a computer and are perceived to be easy to use. They are incorporated in the guidelines of the Royal College of Obstetricians and Gynaecologists (RCOG) on how to manage premenopausal women with adnexal masses [48]. A weakness of the IOTA Simple Rules is that they cannot classify all adnexal masses as benign or malignant. This means that in about 20% of cases another diagnostic method is needed to classify the ‘inconclusive’ masses. A reasonable alternative is to consider this group as malignant, as in 40% of these cases this will be proven by histopathology [49]. Table 36.1 shows the ultrasound features of the Simple Rules as a simple tick box system to usefully assess a mass.

In their meta‐analysis in 2013, comprising 96 validation studies reporting on 19 different diagnostic ultrasound methods in 26 438 adnexal masses, Kaijser et al. [47] showed that IOTA strategies such as the Simple Rules had the best performance in differentiating between the benign and malignant nature of adnexal masses in a preoperative setting. The Simple Rules outperformed the RMI, with a sensitivity and specificity of 93% and 81%, respectively, for Simple Rules versus 72% and 92%, respectively, for RMI. In 2016, Meys et al. [46] performed a meta‐analysis assessing IOTA methods, RMI and subjective assessment by an expert investigator, investigating 47 articles that included 19 674 tumours. They concluded that a two‐step approach with Simple Rules as a first step and subjective assessment for inconclusive tumours yielded the best results and matched the test performance of expert ultrasound examiners.

In general, individual features of ovarian masses have not proved useful as predictors of the likely histology. Accordingly, attention turned to mathematical models using multiple variables. The IOTA group has created different prediction models based on logistic regression analysis that give an individual risk of malignancy for the patient [50–52]. Unlike the Simple Rules, risk prediction models require the use of a computer, an application for smartphone or a calculator integrated in the ultrasound device to acquire the result. As a multiclass prediction model, the recent Assessment of Different NEoplasias in the adneXa (ADNEX) model incorporates clinical and ultrasound information not only to calculate the likelihood of malignancy in adnexal masses but also the likelihood that the mass is borderline malignant, stage I primary invasive ovarian cancer, stage II–IV primary invasive ovarian cancer or a metastasis to the ovary from another primary tumour [52]. This is highly relevant given the differences in the clinical management of women with these different tumours.

Adding information on serum CA‐125 levels to ultrasound information does not seem to improve earlier mathematical models in discriminating between benign and malignant adnexal masses [53]. Neither does CA‐125 in ADNEX, where it is an optional variable. However, in ADNEX the inclusion of CA‐125 allows more accurate risk predictions for the different subgroups of malignancy [52]. A two‐step approach could be adopted to make clinical use of the predicted risks from ADNEX. First, the risk calculation can be used to discriminate between benign and malignant masses based on the specific risk cut‐off value used by individual centres to define malignancy, where the adopted cut‐off may depend on local healthcare policy. Second, we can differentiate between the four subgroups of malignant tumours using the predicted risks for these subgroups. In this step, absolute predicted risks as well as the relative change of these risks versus the baseline risks provide clinically useful information to select an appropriate patient‐specific management strategy. ADNEX has only recently been developed, but has shown excellent discriminative ability when prospectively validated on IOTA phase 3 data [52] as well as in the first external validation studies [54,55]. ADNEX is freely available online (http://www.iotagroup.org/adnexmodel/) or can be downloaded as an application for smartphone. Figure 36.10 shows the clinical and ultrasound features used in the ADNEX model with the risk calculations for a case example.

Research is ongoing on how to manage adnexal masses that are unclassifiable by ultrasound, whether using the Simple Rules or subjective assessment by an expert investigator, as evidence shows that 8% of tumours will still be indeterminate even with these methods. In a systematic review and meta‐analysis, Anthoulakis et al. [56] concluded that pelvic MRI with intravenous contrast administration should be considered the most useful of all diagnostic approaches when investigating adnexal masses that remain unclassified after assessment by ultrasound. This strategy outperformed other imaging modalities including CT.

An extensive description of staging in malignant disease would be beyond the scope of this chapter. Although ultrasound has been shown to be accurate in the assessment of the abdominal spread of gynaecological malignancies in the hands of experienced examiners [57], most centres will use CT for preoperative staging and follow‐up [58]. There is also increasing interest in whole‐body diffusion‐weighted MRI for this purpose, as this technique shows higher accuracy for the evaluation of peritoneal and distant disease compared with CT and PET‐CT [59].