Why pelvic floor surgeons should utilize ultrasound imaging

There have been attempts in the past to integrate imaging methods into the clinical practice of pelvic floor surgeons. From the 1920s onwards, radiological techniques were introduced1-3, first to describe bladder descent, and later for central and posterior compartment prolapse4. Ultrasound has found more widespread acceptance since the late 1980s, whether by the transvaginal5 or the introital/transperineal/translabial6 route. Even magnetic resonance imaging (MRI) has been used in a research context7, but none of those methods has become a generally accepted part of clinical practice as yet. The uptake of new diagnostic technology depends on a number of factors. Some potential motivators have nothing to do with medicine, such as the level of remuneration for a certain test, or whether equipment suppliers are interested in publicizing the benefits of a given new technique—that is, whether a new market is recognized and developed. Other factors relate to logistic barriers, such as equipment availability in a clinical area or specialty, or the need for further training of practitioners. To give an example, generally gynecologists have been more ready to use ultrasound imaging than have their urological colleagues, mainly because ultrasound systems are more commonly available in gynecology units. This is particularly true in central and northern Europe, where sonographic imaging has been firmly in the hands of gynecologists ever since its introduction. Another main motivator, clinical utility, is the focus of this Editorial. I am convinced that clinical examination alone, in particular without the examiner being aware of its shortcomings, is a woefully inadequate tool with which to assess pelvic floor function and anatomy. This is not a particularly original thought. In 1943, Howard Gainey, an obstetrician from Kansas City, first described levator trauma in parous women8—and in editorial comments two reviewers stated ‘I am convinced that there is more to the examination of the postnatal patient than I have been practising’ and ‘None of us has learned to examine the pelvis completely’. This is just as true today as it was in 1943. Our examination skills are poor, focusing on surface anatomy, rather than true structural abnormalities. This is why there is such a multitude of surgical procedures for female organ prolapse, and such constant searching for improvement. We know that we are not doing very well in repairing prolapse, and that this is not least because eyeballs and fingers are often insufficient for accurate diagnosis. One excellent example is vaginal hysterectomy for the treatment of prolapse. Imaging helps us to understand that the uterus is usually the ‘messenger’, not the underlying problem itself—and this messenger is still being shot on a very regular basis. The results are predictable, and imaging can go a long way towards explaining why so many women come back with recurrent prolapse9. The introduction of prolapse quantification systems such as the POP-Q10 has not really changed the situation. While quantification of prolapse beyond the four stages of the Baden–Walker classification has to be regarded as progress, it still involves only surface anatomy, without appreciation of function. Clinical examination often results in false-negative findings due to levator coactivation: we ask the patient to bear down, and she reflexively contracts the pelvic floor, trying to stop loss of urine, flatus or stool, or simply because of the stressfulness and embarrassment of the situation, perhaps even involving a cold metal instrument. Levator activation is particularly common in nulliparous women11, is part of a generalized defensive reflex12, and has to be overcome by visual or tactile biofeedback. False-negative findings due to levator coactivation have been a constant source of surprise and frustration for trainees in my unit over the last few years. Often they present a patient with good pelvic organ support, only to see all kinds of prolapse develop on imaging once we've used biofeedback techniques to ensure that the patient understands what we want from her. Ultrasound will allow you to see levator coactivation and impending loss of urine or stool, alerting you to the need for bladder or bowel emptying and/or biofeedback teaching. Also, while there have been attempts at quantifying levator function13, these have been largely ignored by pelvic floor surgeons, and do not contain any reference to either muscle elasticity or distensibility, or morphological integrity, two of the main predictors of pelvic organ descent14, 15. While it may well be possible to improve our clinical assessment skills, this will not happen unless we allow imaging techniques to show us what the real problems are. The missing link between vaginal childbirth and prolapse—major levator trauma in the form of avulsion of the anteromedial aspects of the pubovisceral muscle off the pelvic sidewall16, 17—is palpable, but palpation of levator trauma requires considerable skill and teaching18, 19, preferably with confirmation by imaging. I will now try to outline the main benefits of pelvic floor ultrasound to the pelvic floor surgeon. Clinical examination is limited to grading anterior compartment prolapse, which we call ‘cystocele’. In fact, imaging can identify a number of entities that are difficult, if not impossible, to tell apart clinically. Pelvic floor ultrasound enables us to distinguish between two types of cystocele with very different functional implications (Figure 1). A cystocele with intact retrovesical angle (first described on X-ray cystourethrography as Green Type III in the 1960s3) is generally associated with voiding dysfunction and a low likelihood of stress incontinence, while a cystourethrocele (Green Type II) is associated with above average flow rates and urodynamic stress incontinence. On clinical examination these two very different entities are grouped together, which is why studies of voiding dysfunction and prolapse have yielded such varying results. In addition, occasionally a cystocele will turn out to be due to a urethral diverticulum, a Gartner duct cyst or an anterior enterocele, all rather likely to be missed on clinical examination. The two main types of cystocele as imaged on maximal Valsalva in the midsagittal plane: (a) cystourethrocele (Green Type II), associated with urinary stress incontinence and good voiding function; (b) an ‘isolated cystocele’ (Green Type III), associated with prolapse and voiding dysfunction rather than stress incontinence. RVA, retrovesical angle. Recently, synthetic suburethral slings have become very popular. These new procedures may result in complications whose diagnosis is greatly helped by imaging. Ultrasound can confirm the presence of such a sling, distinguish between transobturator and transretzius slings20, and even allow an educated guess regarding the exact type and material of the sling21 (Figure 2). A tight c-shaped appearance at rest and a gap of less than 1 cm between tape and symphysis pubis makes functional obstruction very likely and suggests that tape division would be beneficial in a patient with worsened symptoms of bladder irritability or clinically significant voiding dysfunction. Suburethral slings as seen on translabial ultrasound: transretzius TVT (tension-free vaginal tape) on midsagittal (a) and axial (c) plane imaging and transobturator Monarc in midsagittal (b) and axial (d) planes. On a more general note, translabial ultrasound may detect foreign bodies or bladder tumors22, 23 and can be used to determine residual urine, using a formula developed originally for transvaginal ultrasound24. While detrusor wall thickness (DWT) has probably been overrated as a diagnostic tool in the context of detrusor overactivity25, 26, increased DWT is associated with symptoms of the overactive bladder26, 27, and may be a predictor of postoperative de novo urge incontinence and/or detrusor overactivity after anti-incontinence procedures28. Of all three compartments, translabial ultrasound is probably least useful in the assessment of central compartment prolapse. Generally, uterine prolapse is obvious clinically, as is the distinction between uterine and vault descent. Having said that, translabial ultrasound may show graphically 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 someone with an incarcerated 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’ by colorectal surgeons and is virtually unknown in gynecology. Finally, as mentioned above, marked levator ballooning on Valsalva, or major childbirth-related trauma to the levator ani, may affect surgical decisions in women with uterine prolapse. As regards the posterior compartment, clinically we diagnose ‘rectocele’ without being able to distinguish the several different conditions leading to downwards displacement of the posterior vaginal wall. A second-degree rectocele could be due to a true rectocele, i.e. a defect of the rectovaginal septum (most common, and associated with symptoms of prolapse and obstructed defecation)29, or it could be due to an abnormally distensible, intact rectovaginal septum (common, and associated only with prolapse symptoms), a combined rectoenterocele (less common), an isolated enterocele (uncommon) or just a deficient perineum giving the impression of a ‘bulge’30; Figure 3 compares the first two conditions. The distinction between a ‘true rectocele’ (a), i.e. a defect of the rectovaginal septum, and perineal hypermobility (b), i.e. descent of the rectal ampulla without fascial defect. In addition, there is scope for functional imaging of the anorectum, with rectal intussusception and prolapse being visible on Valsalva (Figure 4). It is highly likely that women with a very wide hiatus on Valsalva are at an increased risk of developing rectal prolapse after successful vaginal prolapse repair. If an asymptomatic or even mildly symptomatic rectal intussusception is picked up on translabial imaging, one may want to provide the patient with clear visual biofeedback in order to effect behavioral modification. If we can show a patient that straining at stool is obviously counterproductive (whether due to rectocele, colpocele or rectal intussusception), she will hopefully be more likely to modify her behavior. In some this is all that is needed to cure symptoms of obstructed defecation and dyschezia. Rectal intussusception as evident in the mid-sagittal plane. Images are at rest (a), on mild Valsalva (b) and at maximum Valsalva (c). From Pelvic Floor Ultrasound, by H. P. Dietz. Reproduced with permission from Springer Verlag. Finally, there is sphincter assessment. The anal sphincter is generally imaged by endoanal ultrasound, using high-resolution probes with a field of vision of 360°. This method is firmly established as one of the cornerstones of a colorectal diagnostic workup for anal incontinence. Due to the limited availability of such probes in gynecology, obstetricians and gynecologists have taken to using high-frequency curved array or endovaginal probes placed exoanally, i.e. transperineally, in the coronal rather than the mid-sagittal plane31-33. There are advantages to this approach, and not just from the patient's point of view. Exoanal imaging reduces distortion of the anal canal and allows dynamic evaluation of the anal sphincter and mucosa at rest and on sphincter contraction, which seems to enhance the definition of muscular defects. However, resolution may be inferior34, and good comparative studies are still lacking. At the moment, axial plane imaging is limited to assessment of the levator ani muscle, although technological improvements may soon allow us to evaluate paravaginal and paraurethral tissues as well. Translabial ultrasound has confirmed 60-year-old clinical data8 and MRI studies35-37 showing that major morphological abnormalities of levator structure and function are common in vaginally parous women15, and very recently it has been proven that such morphological abnormalities are due to vaginal delivery16; Figure 5 compares MRI, ultrasound and clinical findings in patients with unilateral levator avulsion. Such major delivery-related levator trauma, affecting the inferomedial aspects of the pubovisceral muscle, may well be the missing link between prolapse and childbirth. Levator trauma seems to enlarge the hiatus38 and results in anterior and central compartment prolapse15. The larger the defect, the higher is the likelihood of prolapse38. Levator defects seem to be associated with cystocele recurrence after anterior repair39. Delivery-related levator trauma as imaged by magnetic resonance (a) and translabial 4D ultrasound (b), and as seen on exploration of a large vaginal tear after vaginal delivery (c) (Magnetic resonance image courtesy of Dr Lennox Hoyte, Boston.). Another factor readily apparent on axial plane imaging is the degree of hiatal distension on Valsalva. If the hiatus enlarges to over 30 cm2 on Valsalva, we speak of ‘ballooning’, and the degree of distension is strongly associated with prolapse40. It seems that ballooning is associated with prolapse recurrence after rectocele repair41, and the same probably holds for other forms of prolapse surgery. If excessive distensibility of the levator is indeed a risk factor for recurrence, then of course we should know about it preoperatively and adjust our surgical approach accordingly. In fact, we may even want to develop surgical methods that reduce the size and distensibility of the hiatus in an attempt to prevent recurrence. Our means of effecting such change, such as vaginal or abdominoperineal levatorplasty, are currently very limited and associated with significant short- and long-term morbidity. Having said that, vaginal mesh techniques such as anterior vaginal wall mesh repair with transobturator anchoring42 clearly reduce the effective size of the hiatus by providing a hammock-like bar across the anterior aspect of the hiatus (Figure 6), and the same may be true for posterior mesh techniques. ‘Perigee’ anterior vaginal wall mesh repair as imaged in the midsagittal plane (a) and a rendered volume in the axial plane (b). It was quite apparent even in the late 1980s that ultrasound was particularly useful for clinical audit. In a study that was never published, we found that vaginal hysterectomy and anterior repair did little to correct cystocele, and frequently made women incontinent who had been continent preoperatively. Ultrasound was pivotal in showing that colposuspension resulted in a much more durable elevation of the bladder neck, and, later, that laparoscopic colposuspension did just the same, although with slightly more ‘give’ over time. When bone anchor slings were introduced, ultrasound rapidly showed that these procedures were very poor surrogates for colposuspension, although it took years for this conclusion to be accepted on clinical grounds. And then there was tension-free vaginal tape (TVT). Imaging helped elucidate its mode of action—plain, simple mechanical compression of the urethra. There is nothing fancy involved, like ‘activation of muscular forces’, as claimed by some. By 1998 or 1999 it was also clear that results regarding stress continence should be very durable, very likely better than those for colposuspension procedures. Other implants such as the Sparc and the Monarc were assessed the same way, showing an identical mode of action, but less compressive effect. Procedural modifications were studied, showing that the original TVT technique (using an intraoperative cough test for sling adjustment) was unnecessarily obstructive. Again and again imaging helped us to adjust practice several years before clinical data in the literature told us what we already knew. All this was prior to the widespread introduction of three-dimensional/4D ultrasound, using two-dimensional imaging equipment that was already obsolete at the time. Even in 2001 it was obvious that imaging could make a real difference to the way in which we investigate and treat women with pelvic floor disorders; now it is difficult to see how anybody could possibly come to any other conclusion. Current trends, i.e. the near-universal introduction of 4D ultrasound, new software options, and increasing availability of training, will likely lead to more general acceptance of ultrasound as a standard diagnostic option in pelvic floor medicine. Its most significant impact may well be on the evaluation of new surgical procedures, i.e. in clinical audit. Ultrasound will enable us to develop improved surgical procedures, individualize reconstructive surgery, and ensure that complications or failure are detected early and dealt with appropriately. It is hard to see how pelvic floor surgeons could remain blind to the obvious benefits, both for their patients and themselves, arising from the new diagnostic method. The crucial issue, as always, is teaching and the provision of up-to-date resources, and it may well be several decades before the potential inherent in the technology is fully realized.