Ultrasound Imaging of the Pelvic Floor



Ultrasound Imaging of the Pelvic Floor


Hans Peter Dietz



Introduction

Ultrasound is the most-used medical imaging technology in general and by far the most popular technique used in obstetrics and gynecology. This is primarily due to easy availability, limited cost, and innate physical properties such as the lack of ionizing radiation and high temporal resolution, resulting from an absence of heavy moving parts. The main downside of diagnostic ultrasound is the rapid attenuation of signals in tissues and the reflection of ultrasound waves by bone and gas. As a result, indications are limited to situations where an external or intracavitary transducer can be brought close to the tissues in question. The pelvis tends to fulfill those criteria. The deeper pelvic organs are reached by endovaginal or endoanal transducers, and the pelvic floor is within the 5 to 7 cm depth covered by the frequency range of modern abdominal and obstetric transducers. The development of three-dimensional (3D) capable transducers in the 1990s allowed access to the axial plane, which previously was only possible with endocavitary probes, and four-dimensional (4D) technology using either mechanical oscillating sectors or, more recently, solid-state matrix probes, have greatly increased temporal resolution.

It is not surprising that there is increasingly widespread uptake of this technology in pelvic floor medicine, even if its speed varies with the availability of equipment and teaching resources, with the adequacy of remuneration by health care funders, and with the extent of regulatory interference. In some jurisdictions with heavy-handed state regulation of medical technology, ultrasound imaging can only be provided in a research setting; in others, professional jealousies have impeded progress. Fortunately, after a decade of efforts, there finally has been a modicum of international standardization,1 and there is standardized online teaching under the auspices of the International Urogynecological Association.2

In this chapter, the author intends to give an overview of what is possible with current technology both in research and in clinical practice, summarize the available literature, and provide an outlook of what is to be expected in this field in the near to medium term.


INSTRUMENTATION AND METHODOLOGY


Two-Dimensional Imaging

Translabial pelvic floor ultrasound at a basic level requires a B-mode capable two-dimensional (2D) ultrasound system with cine loop function, a 3.5- to 6-MHz curved array transducer, and a video printer; all technology that has been available since the mid-1980s. A midsagittal view of the pelvis can be obtained by placing a transducer on the perineum (Fig. 14.1), after covering the transducer with a probe cover such as a glove, condom, or plastic wrap for hygienic reasons. Powdered or coated gloves may impair imaging quality; hence, it is recommended to test probe covers before use to make sure they do not affect image quality. Air between probe and probe cover will cause acoustic artefacts, shadowing, and reverberations and need to be avoided especially when using 3D/4D transducers as bubbles lateral to the main transducer plane may be invisible during acquisition. Sterilization is usually considered unnecessary, with mechanical cleaning and alcoholic wipes used between patients.

Imaging is generally undertaken with the patient in dorsal lithotomy, hips flexed and abducted, or in the standing position. Heels placed close to the buttocks result in an improved pelvic tilt due to reduced lordosis. Scanning is usually done after bladder emptying to allow the determination of residual urine volume and make the patient more relaxed about the risk of leakage of urine on Valsalva. A full rectum sometimes necessitates a repeat assessment after defecation because a full rectum can obscure other structures. Parting of the labia often improves image quality, especially if the labia are hirsute. Imaging conditions are best in pregnancy and poorest in the senium due to tissue hydration. Vaginal scar tissue and mesh implants may also impair visibility, but obesity virtually never is a problem.

Transducer placement between the clitoris and anus can be quite firm, unless there is vulval dermatitis or marked atrophy. The field of view includes the symphysis pubis anteriorly, the urethra and bladder neck, trigone, vagina, cervix, rectum, and anal canal (see Fig. 14.1). The anorectal angle indicates the location of
the levator plate in the midline. The cul-de-sac may also be visible, filled with anechoic intraperitoneal fluid, echogenic intraperitoneal fat, and peristalsing small bowel or sigmoid.3







Three-Dimensional/Four-Dimensional Imaging

The introduction of 3D imaging has given access to the axial plane which is particularly important for pelvic floor medicine as this plane previously was only accessible to sonography through the use of intracavitary transducers. The physical dimensions of transducers designed for prenatal imaging are, rather fortuitously, perfect for pelvic floor imaging. A volume data set obtained with field of vision and acquisition angles of ≥70° will include the entire levator hiatus at rest. Four-dimensional imaging, that is, the continuous acquisition of dozens of volumes at 0.5 to 4 Hz during a Valsalva or Kegel (pelvic floor muscle contraction [PFMC]) maneuver literally adds another dimension to our diagnostic capabilities.4

Optimally, the system should allow volume data imaging with field of vision and acquisition angles of ≥80° and store ≥5 seconds of data at 0.5 to 5 Hz. The result are “cine loops” of blocks of imaging data that can be manipulated to allow the examination of any arbitrarily defined plane and changes in that plane during maneuvers.


Display Modes

Figure 14.2 shows both main display modes used on modern ultrasound systems, i.e., sectional orthogonal planes (Fig. 14.2A-C) and rendering (Fig. 14.2D). For pelvic floor imaging, the orthogonal planes are defined as midsagittal (top left), coronal (top right), and axial (bottom left) planes, although the location and angulation of these planes relative to the main transducer axis can be varied arbitrarily to enhance the visibility of a given structure, either during acquisition or for offline analysis at a later time. The levator ani muscle, for example, requires an axial plane that is tilted, with the direction and degree of this tilt varying between the resting state and maneuvers such as Valsalva or pelvic floor contraction.

These three conventional orthogonal planes are complemented by a “rendered volume,” that is, a semitransparent representation of all volume pixels (“voxels”) in a “region of interest” (ROI) that again can be varied arbitrarily. Such an ROI is shown in Figure 14.2D with the render direction from caudally to cranially, showing the levator hiatus and the puborectalis component of the levator ani. The ROI thickness is set to 1 to 2 cm for optimal imaging of the muscle and hiatus.

The real-time acquisition of volume ultrasound data at a temporal resolution of several Hertz, especially during maneuvers such as Valsalva and PFMC, makes this technology clearly superior to magnetic resonance imaging (MRI). Assessment of maneuvers by MRI requires ultrafast acquisition and the reliance on predetermined planes,5 which will not allow optimal resolutions and often result in suboptimal slice location. In addition, it is very difficult to ensure proper performance of maneuvers because observation in real time is virtually impossible. Most women will not perform a proper pelvic floor contraction when asked without biofeedback teaching.6 A Valsalva is frequently confounded by concomitant levator activation, or a full bladder or rectal gas stops the patient from performing a proper push.7 Without real-time imaging, confounders are very difficult to control for, explaining the dearth of MR studies
in prolapse and urinary incontinence. Another advantage is the availability of offline analysis software that is much more powerful than what is available with Digital Imaging and Communications in Medicine (DICOM) viewer software on a set of single-plane MRI images.







Functional Assessment


Valsalva

The Valsalva maneuver, that is, forced expiration against closed glottis and contracted abdominal muscles and diaphragm, is useful in revealing signs of pelvic organ prolapse and to demonstrate distensibility (“ballooning”) of the levator hiatus. During Valsalva, one observes a dorsocaudal displacement of urethra and bladder neck that can be quantified against the inferoposterior symphyseal margin (Fig. 14.3).5 At the same time, there is caudad movement of the bladder, uterus, rectal ampulla, and abdominal contents as well as lateral and caudad distension of the levator hiatus, the largest potential hernia portal in the human body. It is important for the operator to avoid impeding downward movement of pelvic organs and perineum; there should be only enough pressure exerted by the transducer to avoid loss of contact, visible by the appearance of acoustic artifact. A Valsalva should last at least 6 seconds.8 Intra-abdominal pressure, on the other hand, may not have to be standardized.9 In women with a strong levator muscle, a Valsalva may be confounded by levator activation.7 Three consecutive coughs seem sufficient for the assessment of anterior and posterior compartments.10 At times, it is necessary to repeat imaging in the standing position in order to avoid false-negative findings, which are most common in the central compartment.11


Pelvic floor muscle contraction

A PFMC can be quantified on imaging as a cranioventral shift of pelvic organs and/or a reduction in hiatal diameters.12 Co-contraction of rectus abdominis and other muscles of the abdominal wall commonly results in an unwanted increase in intra-abdominal pressure which is visible as a dorsocaudal shift of the bladder neck as during a Valsalva maneuver. This is prevented by having the patient place one hand on the abdomen during a PFMC and asking her to keep the abdominal muscles soft while trying to act as if she wanted to stop the escape of gas or urine. Asking the patient to observe the effect of maneuvers on the monitor may provide for visual biofeedback.13 Reflex pelvic floor activity can be observed on coughing.14


CLINICAL CONDITIONS


Urinary Incontinence

Imaging of pelvic floor structures is increasingly employed in the investigation of both stress and urge urinary incontinence. In stress urinary incontinence (SUI), there usually is a moderate degree of cystocele in the sense of a rotatory descent of the bladder and urethra. The commonest observations include opening of the retrovesical angle, rotation of the urethra by 60° or more, and funneling of the internal urethral meatus as shown
in Figure 14.3. These findings are not truly diagnostic of urodynamic stress incontinence and can most certainly not be used to posit what is a urodynamic diagnosis.15 However, such finding will be of help in women with exercise-related urodynamic stress incontinence (USI) in whom urodynamic testing frequently yields a false-negative result, especially in younger women with good urethral closure. Urge urinary incontinence is associated with detrusor hypertrophy, that is, a detrusor wall thickness (DWT) of 5 mm or more.16 DWT is measured in the midsagittal plane, on the bladder dome, and is unreliable at bladder volumes over 50 mL due to distension of the detrusor. Unfortunately, the association between DWT and detrusor overactivity is too weak to serve as a diagnostic criterion.17 A markedly thickened DWT should raise a suspicion of obstruction, either mechanical (Fig. 14.4) or due to neuropathic bladder dysfunction. Such findings are, however, quite uncommon and rarely found in older women with clinical voiding dysfunction.













Pelvic Organ Prolapse


Cystocele

Clinical examination by the Pelvic Organ Prolapse Quantification System (POP-Q)18 is necessarily limited to quantifying changes in surface anatomy. A view of deeper structures with imaging is highly useful because it will identify urethral diverticula or Gartner cysts that can mimic cystocele. In addition, imaging allows to distinguish between cystourethrocele (usually with good urine flow rates and urodynamic stress incontinence, as in Fig. 14.3), and a cystocele with intact retrovesical angle (associated with poor voiding due to urethral kinking).19 A cutoff of descent to 10 mm or more below the symphysis pubis has been suggested for the definition of “significant cystocele,” that is, a degree of bladder descent that is likely to cause prolapse symptoms,20 equivalent to a Ba of -0.5 cm on POP-Q.21 In women after colposuspension, an anterior enterocele may occasionally mimic cystocele, easily diagnosed due to peristalsis and the iso- to hyperechoic nature of small bowel.


Uterine descent

Clinical assessment for prolapse sometimes produces false-negative results, and these seem to be most common in the central compartment.11 Clearly, uterine supports need more time to distend than those of the anterior and central compartments. The uterus can also be harder to identify on imaging, especially if retroverted and/or atrophic, as the isoechoic echotexture of the cervix is similar to vaginal wall. Provided it is not high or obscured by stool or bowel gas in a rectocele, the leading edge of the cervix often shows as a fine hyperechogenic (specular) line, and Nabothian follicles can also act as a marker (Fig. 14.5). An anteriorized cervix and an enlarged, retroverted uterus may impair bladder emptying as occasionally occurs in pregnancy. On the other hand, descent of an anteverted uterus can result in symptoms of obstructed defecation due to compression of the anorectum and/or intussusception. The descending uterus seems to cause prolapse symptoms at a higher station than anterior or posterior vaginal wall, suggesting that clinical staging used in conjunction with the POP-Q needs revision.22 Consistent with such findings, a sonographically determined position of 15 mm above the symphysis on maximal Valsalva seems to be the optimal cutoff for the definition of “significant uterine descent.”23






There appears to be substantial interethnic variations regarding uterine descent which may be more common in East Asians.24 In Caucasians, uterine prolapse is associated with cystocele and levator trauma, but uterine retroversion also seems to be a risk factor. The known high rates of uterine prolapse in Nepalese, for example, may largely be due to retroversion, which seems very common in Nepalese women.25


Vault prolapse

In women after hysterectomy, the vault may be identified as the cranial aspect of hypoechoic vaginal walls, sometimes separated by iso- or hyperechogenic echoes that will be more evident after a vaginal examination. Identification of the vault can be helped by depositing a small amount of ultrasound gel in the vagina. Often, however, the vault is obscured by rectocele or enterocele causing acoustic shadowing. As an aside, vault descent to a given position seems to be just as likely to cause symptoms of prolapse than equivalent descent of the cervix.26


Posterior compartment

Imaging is most useful in the posterior compartment, not least due to the high prevalence of obstructed defecation in women, a symptom group strongly associated with several distinct forms of vaginal prolapse. In Caucasians, the clinical observation of a “rectocele,” that is, descent of the posterior vaginal wall, seems to be most commonly due to a true or “radiologic” rectocele, that is, a defect of the rectovaginal septum or Denonvilliers fascia (associated with symptoms of prolapse, incomplete bowel emptying, and straining at stool).27 However, an abnormally distensible, intact rectovaginal septum (associated only with prolapse symptoms); a combined rectoenterocele (less common); an
isolated enterocele (uncommon); a rectal intussusception (uncommon); or just a deficient perineum may also be diagnosed as rectocele on clinical examination if the latter is limited to the observation of surface anatomy.28 Hence, this term should probably be avoided and replaced with posterior compartment descent unless a true rectocele is diagnosed by imaging or rectal examination on Valsalva.29 Figure 14.6 shows four of the most common abnormalities associated with clinical prolapse of the posterior vaginal compartment, in descending order of prevalence: true rectocele (Fig. 14.6A), rectoenterocele (Fig. 14.6B), isolated enterocele (Fig. 14.6C), and rectal intussusception (Fig. 14.6D).


Rectocele

A “true” rectocele is due to an anterior diverticulum of the rectal ampulla, which develops into the vagina, more obvious on Valsalva and easily observed on imaging (see Fig. 14.6A). Posterior and lateral rectoceles are rare and associated with more severe anatomical abnormalities such as intussusception or herniation through the iliococcygeus muscle.

The pocket or diverticulum formed by a rectocele mostly contains iso- to hyperechoic feces and bowel gas, causing acoustic artefacts such as shadowing, specular (mirror-like) echoes, and reverberations. Large rectoceles can necessitate repeat assessment after bowel emptying as they will obscure a large part of the field of vision and create a conical area of acoustic shadowing that may even impede imaging of the levator hiatus. If there is no stool in the ampulla, a rectocele may appear much smaller. Appearances can vary considerably from one day to the other making posterior compartment descent less reproducible than other imaging findings.30






A rectocele is quantified relative to the symphyseal margin for descent as usual for any other form of prolapse and by measuring the depth of the diverticulum against a baseline drawn through the anterior aspect of the internal anal sphincter (see Fig. 14.6A). “Significant descent of the posterior compartment” is diagnosed if the rectal ampulla or a diverticulum of the ampulla (“true rectocele”) reaches to at least 15 mm below the symphysis pubis regardless of the presence of a diverticulum or true rectocele.20 Attempts at producing a cutoff for significant true rectocele have been unsuccessful likely due to multiple confounding factors.31 Findings on translabial ultrasound are comparable to those obtained by defecation proctography, provided the evaluation is produced using similar diagnostic criteria (see Dietz and Cartmill32 for an overview). Finally, it should be pointed out that rectocele can very likely be congenital,33 that it is less strongly associated with childbirth than just about any other form of prolapse,34 and that it is the one form of prolapse that is clearly linked to obesity.35


Enterocele

An enterocele is imaged as descent of small bowel or any other abdominal contents dorsal to the (anechoic) bladder and ventral to the (hyperechogenic) rectal ampulla and anal canal (see Fig. 14.6B). There often is peristalsis, and intraperitoneal fluid may outline the most caudal point of an enterocele. Small bowel has a ground-grass-like or irregularly iso- to hyperechogenic appearance without acoustic shadowing and can easily be distinguished from stool in the ampulla or a rectocele. Enteroceles often occur in combination with rectocele (see Fig. 14.6B), commonly after hysterectomy
and in the form of vault prolapse where they can be isolated (see Fig. 14.6C), but occasionally, an enterocele may develop anteriorly, that is, between bladder and vault. Enterocele is often only visible after bladder emptying regardless of its type.

In a minority of women with a clinical rectocele, imaging will show a rectal intussusception, a condition that is found in about 4% of our patients and rarely diagnosed clinically. It is clearly associated with symptoms of obstructed defecation.36 The main diagnostic feature of this underdiagnosed condition is splaying of the (normally tubular) anal canal, while the anterior rectal muscularis (and sometimes the posterior as well) is inverted into the anal canal (see Fig. 14.6D). This inversion is propelled by abdominal contents such as small bowel or sigmoid, omentum, or uterus, the latter termed a “colpocele.” Interestingly, rectal intussusception is strongly associated with hiatal ballooning.37

Visual biofeedback, that is, demonstrating the effect of straining on anorectal anatomy to the patient, can help in modifying defecatory behavior. One may also be able to demonstrate the mechanism of anismus which is evident as an inability to relax the levator ani, producing a reduction in anteroposterior hiatal diameters on Valsalva. Translabial ultrasound seems to be a simple, inexpensive alternative to radiologic imaging methods.32


POSTOPERATIVE FINDINGS


Slings

Due to the high echogenicity of polypropylene threads, synthetic suburethral slings are easily visible on translabial ultrasound as linear hyperechogenic structures between the urethral rhabdosphincter and vaginal muscularis.38 They act by dynamic compression of the urethra under load, which is easily demonstrated on imaging, with the sling changing from a linear to a c shape.39 This is likely to be most effective if compression occurs at the locus of maximal urethral pressures, that is, at the midurethra. The type of sling, that is, whether it is a retropubic or a transobturator tape, can be determined in the axial plane, but with some experience, this distinction can also be made on 2D imaging either by placing the transducer in a transverse plane or by following the sling in parasagittal planes. Figure 14.7 demonstrates typical appearances of a transobturator tape in midsagittal, axial, and parasagittal planes. A transobturator tape will reach the levator ani and sometimes perforate its most caudal aspects,40 a retropubic sling curves ventrally to enter the space of Retzius.

Complications such as urinary retention or worsened/de novo urgency and/or urge incontinence often turn out to be due to a tightly curled, excessively tensioned sling which leaves too small a gap between the implant and the symphysis pubis, especially on Valsalva. This “sling-pubis gap” (Fig. 14.8) seems to be the most useful measure of “sling tightness,” with a gap of 8 to 14 mm on maximal Valsalva being rated as “normal.”41 An implant that is close to the urethra, with a low sling-pubis gap of less than 8 mm, will suggest either dilatation/stretching of the sling if identified within the first week or 10 days or a sling division.

Some slings may be detected close to the external meatus or the bladder neck, and there clearly is a large margin of safety regarding placement.42 The larger the incision, the higher the sling will be found on imaging; if a sling is inserted through an incision contiguous with a colpotomy for anterior repair, it may well come to rest under the trigone, creating appearances similar to after a colposuspension. In this situation, as evident in Figure 14.9, the sling-pubis gap is often very wide, and recurrence of stress incontinence is very likely due to an absence of dynamic compression.