It has taken over 20 years for imaging to develop as a mainstream diagnostic tool in the investigation of female pelvic organ prolapse, urinary and fecal incontinence, and defecation disorders. Physicians have been slow in realizing that clinical assessment alone is a very inadequate tool to assess pelvic floor function and anatomy. Our examination skills are poor, focusing on surface anatomy rather than true structural abnormalities, and recurrence after pelvic reconstructive surgery is common.¹ The uptake of pelvic floor ultrasound by clinicians has varied substantially from one specialty to another, with gynecologists having a major advantage over urologists and colorectal surgeons because an entire generation of OB/GYN specialists has now grown up with ultrasound imaging. Sonography is an accepted component of any clinical assessment in both obstetrics and gynecology, so why should it be any different in urogynecology and female urology?

In theory, clinical assessment skills could be improved to such a degree as to make imaging unnecessary in many cases. However, this clearly is not the case at present, and it is unlikely to happen unless we allow imaging techniques to demonstrate what (and where) the actual problems are. To give just one example: the missing link between vaginal childbirth and prolapse—major levator trauma in the form of avulsion of the anteromedial aspects of the puborectalis muscle off the pelvic sidewall^2,3—is palpable, but palpation of levator trauma requires considerable skill and teaching,^4-6 preferably with imaging confirmation. Certainly, diagnosis by imaging is more reproducible than diagnosis by palpation,⁶ and easier to teach. And suspected levator trauma or abnormal distensibility (“ballooning”) are by no means the only reason to perform pelvic floor imaging (Table 42-1).

Two-Dimensional Imaging

Basic requirements for translabial pelvic floor ultrasound include a B-mode capable two-dimensional (2D) ultrasound (US) system with cine loop function, a 3.5- to 6-MHz curved array transducer, and a video printer. A midsagittal view is obtained by placing a transducer (usually a curved array with frequencies between 3.5 and 8 MHz) on the perineum (Figure 42-1), after covering the transducer with a glove, condom, or thin plastic wrap. Air bubbles can cause reverberation artefacts and need to be avoided by covering the entire transducer surface with gel. This is particularly important with three-dimensional/four-dimensional (3D/4D) transducers as bubbles lateral to the main transducer plane may not be noticed during acquisition. Sterilization as for intracavitary transducers is usually considered unnecessary. We use alcoholic wipes to clean the transducer between patients, but regulations may vary between jurisdictions.

Powdered or coated gloves can impair imaging quality due to reverberations and should be avoided. It is worthwhile testing several types of probe covers for their effect on image quality and ease of application. Imaging is usually performed in dorsal lithotomy, with the hips flexed and slightly abducted, or in the standing position. Requiring the patient to place heels close to the buttocks will result in an improved pelvic tilt. Bladder filling should be specified; usually prior voiding is preferable. The presence of a full rectum can impair diagnostic accuracy and sometimes necessitates a repeat assessment after bowel emptying. Parting of the labia can improve image quality, which generally is best in pregnancy and poorest in menopausal women with marked atrophy, most likely due to varying hydration of tissues. Vaginal scar tissue can also impair visibility, but obesity virtually never seems to be a problem.

The transducer can usually be placed firmly on the perineum and the symphysis pubis without causing discomfort, unless there is marked atrophy. The resulting image includes the symphysis pubis anteriorly, the urethra and bladder neck, the vagina, cervix, rectum, and anal canal (see Figure 42-1). Posterior to the anorectal junction a hyperechogenic area indicates the central portion of the levator plate, ie, the puborectalis muscle. The cul-de-sac may also be seen, filled with a small amount of fluid, echogenic intraperitoneal fat, or peristalsing small bowel. Parasagittal or transverse views may yield additional information, eg, enabling assessment of the levator ani muscle and its insertion on the arcus tendineus of the levator ani, and for imaging of implants.

While there has been disagreement regarding image orientation in the midsagittal plane, the author prefers an orientation as on conventional transvaginal ultrasound (cranioventral aspects to the left, dorsocaudal to the right). The latter also seems more convenient when using 3D and 4D systems.

Three-Dimensional/Four-Dimensional Imaging

The introduction of 4D ultrasound has had a major impact on pelvic floor imaging. One could argue that 4D enhances our diagnostic capabilities to a greater degree than in any other area of OB/GYN. This is mainly because 4D ultrasound gives access to the axial plane to a degree and with an ease that far surpasses what was possible using intracavitary transducers in the past. A single volume obtained at rest with aperture and acquisition angles of 70 degrees or higher will include the entire levator hiatus with symphysis pubis, urethra, paravaginal tissues, the vagina, anorectum, and levator ani muscle from the pelvic sidewall to the posterior aspect of the anorectal junction. A cine loop of volumes obtained during Valsalva or pelvic floor contraction enhances capabilities even further.

Basic requirements for 3D/4D pelvic floor ultrasound would include a system that allows acquisition, reconstruction, and analysis of volume datasets, including the ability to measure distances and areas in this volume. Currently, the most common 3D probes are those that combine an electronic curved array of 3 to 8 MHz with mechanical sector technology. Matrix transducers are becoming available, allowing much higher acquisition speeds, but so far it is difficult to see practical benefits.

In essence, any system that allows satisfactory 3D imaging using an abdominal obstetric probe will be suitable provided the acquisition angle is sufficient to include the entire levator hiatus (ie, between 70 and 85 degrees). Optimally, one should be able to obtain volumes at an acquisition angle of 80 to 85 degrees and store at least 5 seconds’ worth of sequential volumes on the system’s hard disk for later evaluation.

Display Modes

Figure 42-2 demonstrates the two basic display modes currently in use on 3D ultrasound systems. The multiplanar or orthogonal display mode shows cross-sectional planes through the volume in question. For pelvic floor imaging, this most conveniently means the midsagittal (top left), the coronal (top right), and the axial (bottom left) planes. Imaging planes on 3D ultrasound can be varied in a completely arbitrary fashion in order to enhance the visibility of a given anatomical structure, either at the time of acquisition or offline at a later time. The levator ani, for example, usually requires an axial plane that is tilted, and the direction of the tilt can vary greatly between maneuvers such as Valsalva or pelvic floor contraction and the resting state.

The three orthogonal images are complemented by a “rendered image,” ie, a semitransparent representation of all voxels in an arbitrarily definable “box,” the “region of interest.” Figure 42-2D shows a standard rendered image of the levator hiatus, with the rendering direction set from caudally to cranially, which is the most appropriate for imaging the hiatus. Usually a rendered volume of 1 to 2 cm is optimal for imaging of the hiatus. The possibilities for postprocessing are restricted only by the software used for this purpose.

Four-Dimensional Imaging

4D imaging implies the real-time acquisition of volume ultrasound data, which can then be represented in orthogonal planes or rendered volumes. Systems are now capable of storing cine loops of dozens of volumes, which is of major importance in pelvic floor imaging as it allows enhanced documentation of functional anatomy. Whether on 2D or 3D imaging, a static assessment at rest gives little information compared with the evaluation of maneuvers such as a levator contraction and Valsalva. Their observation will allow assessment of levator function and delineate levator or fascial trauma more clearly.

The ability to perform a real-time 3D (or 4D) assessment of pelvic floor structures makes the technology clearly superior to magnetic resonance imaging (MRI). Prolapse assessment by MR requires ultrafast acquisition,⁷ which is of limited availability and will not allow optimal resolutions. Alternatively, some systems allow imaging of the sitting or erect patient, but again accessibility will be limited for the foreseeable future. The sheer physical characteristics of MRI systems make it much harder for the operator to ensure efficient maneuvers as over 50% of all women will not perform a proper pelvic floor contraction when asked,⁸ and a Valsalva is often confounded by concomitant levator activation.⁹ Without real-time imaging, these confounders are impossible to control for, which is why there is no research to date on dynamic MRI in the axial pane.

As a result, ultrasound has major potential advantages when it comes to describing prolapse, especially when associated with fascial or muscular defects, and in terms of defining functional anatomy. Offline analysis packages allow distance, area, and volume measurements in any user-defined plane (oblique or orthogonal), which is much superior to what is currently possible with Digital Imaging and Communications in Medicine (DICOM) viewer software on a standard set of single-plane MRI images.

Valsalva

The Valsalva maneuver, ie, a forced expiration against a closed glottis and contracted diaphragm and abdominal wall, is routinely used to effect downward displacement of pelvic organs, reveal the symptoms and signs of female pelvic organ prolapse, and demonstrate distensibility of the levator hiatus. The result is a dorsocaudal displacement of urethra and bladder neck that can be quantified using a system of coordinates based on the inferoposterior symphyseal margin (Figure 42-3) or the central axis of the symphysis pubis.¹⁰ There also is downward movement of the uterine cervix and the rectal ampulla, and frequently the development of a rectocele, ie, a sacculation of the anterior wall of the rectal ampulla, toward the vaginal introitus or beyond. In the axial plane the hiatus is distended, and the posterior aspect of the levator plate is displaced caudally, resulting in a varying degree of perineal descent. All this can be observed on pelvic floor ultrasound, but it is important to let the transducer move with the tissues, avoiding undue pressure on the perineum, which would prevent full development of a prolapse. To achieve maximal or near maximal organ descent, it is necessary to obtain Valsalva pressures of at least 60 cm H2O for at least 5 to 6 seconds,¹¹ and on clinical examination this is frequently not achieved.

In addition, especially in young, nulliparous women, a Valsalva is frequently confounded by levator activation,⁹ as any obstetrician is able to observe on a daily basis. Levator coactivation during Valsalva is highly inconvenient—not just in the labor ward during the expulsive phase of a woman’s first vaginal delivery, but also when we assess women for pelvic floor dysfunction and prolapse. It is visible as a reduction in the anteroposterior diameter of the levator hiatus on Valsalva (Figure 42-4) and has to be avoided in order to obtain an accurate assessment of pelvic organ descent. Any imaging assessment of organ descent requires real-time observation of the effect of a Valsalva maneuver to correct suboptimal efforts, especially if leakage from bladder or bowel is likely. At times, levator coactivation can prevent adequate assessment in the supine position, in particular in women with a strong, intact levator shelf. Sometimes it is necessary to repeat imaging in the standing position, which seems to increase the likelihood of an adequate bearing-down effort.

Pelvic Floor Muscle Contraction

Ultrasound is a highly useful tool in the assessment of the pelvic floor musculature, both in purely anatomical terms and for function. A levator contraction will reduce the size of the levator hiatus in the sagittal plane and elevate the anorectum, changing the angle between the levator plate and symphysis pubis. As an indirect effect, other pelvic organs such as the uterus, bladder, and urethra are displaced cranially (Figure 42-5), and there is compression of the urethra, vagina, and anorectal junction, explaining the importance of the levator ani for urinary and fecal continence as well as for sexual function.

In its most basic form, transabdominal B-mode imaging can demonstrate elevation of the bladder base on pelvic floor muscle contraction (PFMC), but quantification is difficult and repeatability is lower than for translabial ultrasound.¹² The latter has been employed for the quantification of pelvic floor muscle function, both in women with stress incontinence and continent controls¹³ as well as before and after childbirth.^14,15 A cranioventral shift of pelvic organs imaged in a sagittal midline orientation is taken as evidence of a levator contraction. The resulting displacement of the internal urethral meatus is measured relative to the inferoposterior symphyseal margin (see Figure 42-5). Care has to be taken to avoid concomitant activation of the abdominal muscles, especially the rectus abdominis or the diaphragm, as this would tend to cause caudal displacement of the bladder neck.

Another means of quantifying levator activity is to measure reduction of the levator hiatus in the midsagittal plane, or to determine the changing angle of the hiatal plane relative to the symphyseal axis (see Figure 42-5). The method can also be utilized for pelvic floor muscle exercise teaching by providing visual biofeedback¹⁶ and has helped validate the concept of “the knack,” ie, of a reflex levator contraction immediately prior to increases in intraabdominal pressure such as those resulting from coughing.¹⁷ Correlations between cranioventral shift of the bladder neck on the one hand and palpation/perineometry on the other hand have been shown to be good.¹⁸ However, there are some limitations that affect sonographic evaluation of pelvic floor muscle function. The most obvious lies in the fact that ultrasound exclusively documents displacement of tissues which is crucially dependent on tissue elasticity or compliance. If there is substantial prolapse in a patient with highly elastic tissues, even a relatively weak contraction will produce substantial displacement. On the other hand, a woman with stiff fascial structures and high resting tone of the levator ani may produce a very strong contraction that results in much less displacement.¹⁹ Finally, one should mention that it is possible to observe reflex pelvic floor activity, but the clinical utility of this finding seems to be limited.^20,21

The original indication for pelvic floor ultrasound was (and still is) the assessment of bladder neck mobility and funneling of the bladder neck, both of which are considered important in women with urinary incontinence. Figure 42-3 shows the standard orientation used to describe bladder neck mobility. The position of the bladder neck is determined relative to the inferoposterior margin of the symphysis pubis, and comparative studies have shown good correlations with radiological methods previously used for this purpose (for an overview see reference 22). The one remaining advantage of x-ray fluoroscopy may be the ease with which the voiding phase can be observed, although some investigators have used specially constructed equipment to document voiding with ultrasound.²³

Residual Urine

Translabial ultrasound provides an easy means of determining residual urine (Figure 42-6), using the formula [(X * Y * 5.6)] = residual in mL.²⁴

Bladder Neck Mobility

Bladder neck position and mobility can be assessed with a high degree of reliability. Points of reference are the central axis of the symphysis pubis²⁵ or its inferoposterior margin²⁶; see Figure 42-3). The full bladder is less mobile²⁷ and may prevent complete development of pelvic organ prolapse. It is essential not to exert undue pressure on the perineum so as to allow full development of pelvic organ descent. Measurements of bladder neck position are performed at rest and on maximal Valsalva, and the difference yields a numerical value for bladder neck descent. On Valsalva, the proximal urethra may be seen to rotate in a posteroinferior direction. The extent of rotation can be measured by comparing the angle of inclination between the proximal urethra and any other fixed axis. Some investigators measure the retrovesical angle (RVA, or posterior urethrovesical angle [PUV]) between proximal urethra and trigone; others determine the angle between the central axis of the symphysis pubis and a line from the inferior symphyseal margin to the bladder neck.²⁸ The reproducibility of bladder neck descent seems as good as that of other sonographic measures of pelvic floor biometry.²⁹

A cutoff of 25 mm for the definition of “normal” for bladder neck descent has been proposed,³⁰ although the wide range of measurements obtained in young nulliparae³¹ and the limited association between bladder neck descent and symptoms³⁰ limit the utility of such a cut-off.

Bladder filling, patient position, and catheterization all have been shown to influence measurements (see reference 10 for an overview), and it can occasionally be quite difficult to obtain an effective Valsalva maneuver, especially in nulliparous women who routinely coactivate the levator muscle.⁹ The etiology of increased bladder neck descent is likely to be multifactorial. The wide range of values obtained in young nulliparous women suggests a major congenital component.³²

Vaginal childbirth is probably the most significant environmental factor,^33-35 with a long second stage of labor and vaginal operative delivery associated with increased postpartum descent. It is not clear what pathophysiological changes are responsible for increased pelvic organ descent. Both fascial disruption and damage to the levator ani may play a role.³⁶

Funneling and Stress Incontinence

In patients with stress incontinence, but also in asymptomatic women, funneling of the internal urethral meatus may be observed on Valsalva (see Figure 42-3b) and sometimes even at rest.³⁷ Funneling is often (but not necessarily) associated with leakage. Other indirect signs of urine leakage on B-mode real-time imaging are weak grayscale echoes (“streaming”) and the appearance of two linear (“specular”) echoes defining the lumen of a fluid-filled urethra. However, funneling may occasionally also be observed in urge incontinence and is of limited use as a predictor of urodynamic stress incontinence (USI).³⁸ Its anatomical basis is unclear. Marked funneling has been shown to be associated with poor urethral closure pressures.^39,40

Color Doppler ultrasound^41-43 can demonstrate urine leakage on Valsalva maneuver or coughing, although one may argue that urine leakage and leak point pressures can be determined much more easily by direct observation.

Cystocele

Clinical examination is limited to grading anterior compartment prolapse, which we call “cystocele.” In fact, imaging will identify a number of anatomical situations that are difficult, if not impossible, to distinguish clinically. There are at least two types of cystoceles with very different functional implications (Figure 42-7). A cystourethrocele is associated with above average flow rates and urodynamic stress incontinence (A), while a cystocele with intact RVA (B) is generally associated with voiding dysfunction and a low likelihood of stress incontinence.⁴⁴ Either form will cause symptoms of prolapse if it descends far enough, and a cutoff of 10 mm below the symphysis pubis has been proposed for the definition of “significant prolapse,”⁴⁵ equivalent to a measurement of −0.5 cm for descent of the anterior vaginal wall on clinical examination (termed “Ba” on POPQ prolapse quantification).⁴⁶ Rarely, an anterior enterocele may mimic a cystocele in women after Burch colposuspension.

Urethral Diverticula

Occasionally a cystocele will turn out to be due to a urethral diverticulum. Figure 42-8 shows a 3D representation of a typical posterior urethral diverticulum, and Figure 42-9 shows a much less common anterior urethral diverticulum, which is more difficult to treat since it requires access through the space of Retzius. The main differential diagnosis is a Gartner duct cyst (Figure 42-10), which may appear similar but does not affect or disrupt the urethral rhabdosphincter.

Urethral diverticula are often overlooked for years in women with recurrent bladder infections and symptoms of frequency, urgency, and pain or burning on voiding, until imaging is undertaken. Urethral structure and spatial relationships are much better appreciated in the axial plane (see Figures 42-8, 42-9, 42-10), which is particularly useful in the differential diagnosis of Gartner cyst and urethral diverticulum. Often a Valsalva maneuver will help in improving visibility, allowing insonation of the structure from varying angles. Standard textbooks suggest that MR is the investigation of choice in women suspected of having a urethral diverticulum, but it is difficult to see what advantages MR should have over ultrasound for this indication. Unfortunately, the condition is not common enough for studies of diagnostic efficacy.

Detrusor Wall Thickness

The thickness of the bladder wall (bladder wall thickness [BWT] or detrusor wall thickness [DWT]) can easily be determined on translabial ultrasound (Figure 42-11). As increasing bladder filling reduces BWT due to distension, measurements should only be undertaken at bladder volumes of 50 mL or less.⁴⁷ While DWT has probably been overrated as a diagnostic tool in the context of detrusor overactivity,^48,49 increased DWT is associated with symptoms of the overactive bladder^50,51 and may be a predictor of postoperative de novo urge incontinence and/or detrusor overactivity after antiincontinence procedures.⁵¹ As opposed to the situation in the male, DWT in women is not predictive of voiding dysfunction.⁵²

Other Bladder Pathology

Occasionally a foreign body, eroded mesh, or even a bladder tumor (Figure 42-12) may be picked up on translabial ultrasound,⁵³ and a careful examination using parasagittal planes may show bladder diverticula. A cystic structure that varies markedly over seconds or minutes and is located 1 to 3 cm lateral and posterior to the bladder neck is likely to be a ureterocele, a generally harmless sacculation of the distal ureter due to a minor stenosis of the ureterovesical junction (Figure 42-13). If the respective upper tract appears normal on renal ultrasound, then no further action is required.

Uterine Descent

Excessive mobility of the uterus is commonly missed on clinical examination as the organ requires more time to descend than bladder or rectal ampulla.⁵⁴ In Caucasians it is commonly associated with cystocele, and this is the typical presentation of prolapse in a patient with major pelvic floor damage. In East Asians and more generally in women with retroversion one may also find isolated descent of the uterus, a condition that seems to follow a different pattern and is more likely to be treated successfully. It seems that the uterus does not need to descend nearly as far as bladder or rectum to cause symptoms of prolapse, with a position of 15 mm above the symphysis defined as the optimal cutoff for the definition of “significant uterine descent.”⁵⁵

An unusually low cervix is isoechoic, its distal margin is evident as a specular line, and it may well contain the cystic structures of nabothian follicles (Figure 42-14). A Valsalva maneuver will result in relative movement, distinguishing a nabothian follicle from Gartner duct cysts or urethral diverticula. Translabial ultrasound can graphically show the effect of an anteriorized cervix in women with an enlarged, retroverted uterus, explaining symptoms of voiding dysfunction, and supporting surgical intervention in order to improve voiding in a patient with a retroverted fibroid uterus. On the other hand, mild descent of an anteverted uterus may result in compression of the anorectum, explaining symptoms of obstructed defecation—a situation that is termed a “colpocele” on defecation proctogram and translabial US (Figure 42-15).

Vault Prolapse

It is frequently possible to image vault descent in women after hysterectomy, but just as often the thin iso- to hypoechoic structure of the vaginal wall is obscured by a descending rectocele or enterocele, which limits the usefulness of assessing the position of the vault. If imaging of the vault is required, depositing 20 to 50 mL of ultrasound gel in the vagina will help with delineation of the upper vagina.

Pelvic floor ultrasound is particularly useful in the posterior compartment, given the high prevalence of symptoms of obstructed defecation, and given that this symptom complex is often explained by anatomical abnormalities of the posterior compartment. Clinically we diagnose a “rectocele” when we really mean “descent of the posterior vaginal wall,” ignoring that several different conditions can lead to such appearances.

A second-degree “rectocele” could be due to a true or “radiological’ rectocele,” that is, a defect of the rectovaginal septum (which is most common and associated with symptoms of prolapse, incomplete bowel emptying, and straining at stool)⁵⁶; an abnormally distensible, intact rectovaginal septum (which is common and is associated only with prolapse symptoms); a combined rectoenterocele (which is less common); an isolated enterocele (which is uncommon); a rectal intussusception (uncommon) or just a deficient perineum giving the impression of a “bulge.”^57,58

Rectocele

An anterior rectocele is a diverticulum of the anterior wall of the rectal ampulla into the vagina, which generally is much more obvious on Valsalva than at rest. Posterior rectoceles are uncommon in adult women and may rather be a form of intussusception than an actual rectocele. Occasionally, a rectocele may also form laterally in the sense of a herniation through the iliococcygeus muscle; such abnormalities are rarely properly diagnosed and difficult to treat. Fortunately, most are anterior, ie, due to defects in the rectovaginal septum or “Denonvillier’s fascia,” which are relatively easy to correct surgically.

A rectocele usually contains iso- to hyperechoic feces, and often there is bowel gas as well, resulting in specular echoes and reverberations. Occasionally there is no stool in the ampulla that could be propelled into the rectocele, and as a result it remains smaller and filled only with rectal mucosa. Since distension of a rectocele will depend on the presence and quality of stool, appearances may vary considerably from one day to the next. The severity of a rectocele can be quantified by measuring maximal descent relative to the inferior symphyseal margin, and by determining the maximal depth of the sacculation as seen in Figure 42-16. Figure 42-17 shows a comparison of ultrasound and radiological findings in a patient with a simple rectocele. Significant descent of the posterior compartment has been defined as an ampulla 15 mm or less below the symphysis pubis,⁴⁵ and a rectocele may be considered ‘significant’ if its depth is measured at or above 10 or, alternatively, 15 mm.⁵⁹

These measurements are less repeatable than other comparable parameters of pelvic floor biometry,²⁹ likely due to differences in stool quality and quantity. Surgery is best undertaken using the “defect specific” technique of Cullen Richardson,⁶⁰ which is highly successful in closing defects of the rectovaginal septum.⁵⁹ Posterior compartment mesh may well block the development of a rectocele into the vagina; however, if the septal defect is not addressed the herniation may well develop into the perineum instead, as in Figure 42-18. Findings on translabial ultrasound seem to be comparable to those of defecation proctography, provided the same criteria are used in the evaluation of sonographic and radiological imaging.^61-65