Objective
Delivery-related levator avulsion can cause pelvic floor dysfunction. We compared agreement between tomographic ultrasound and magnetic resonance–based models for the detection of levator defects.
Study Design
Sixty-nine Chinese women with pelvic organ prolapse were assessed prospectively by 3-dimensional ultrasound scans and magnetic resonance imaging. Levator-urethra gap (LUG), levator-symphysis gap (LSG), and puborectalis attachment width were measured offline with state-of-the-art software. Interobserver variability and agreement between the 2 methods were determined.
Results
Interobserver repeatability was moderate-to-excellent for all parameters that were measured with both methods and agreement between methods in diagnosing levator avulsion. LUG and LSG measurements were significantly higher in women with a levator avulsion. With a diagnosis of complete levator avulsion, receiver operating characteristics analysis suggested a cutoff of 23.65 mm for LUG and 28.7 mm for LSG.
Conclusion
Levator avulsion can be diagnosed reliably by tomographic ultrasound scanning and magnetic resonance imaging evaluation, and linear measures, such as LSG and LUG, can be proxy measurements for avulsion.
The levator ani muscle plays a vital role in the maintenance of pelvic organ support and function. This muscle has 2 major components, the pubovisceral (including the pubococcygeus and puborectalis muscles) and the iliococcygeal muscles. Delivery-related trauma to the pubovisceral muscle is common and obviously associated with female pelvic organ prolapse, especially anterior and central compartment prolapse. It can be diagnosed by palpation, magnetic resonance imaging (MRI), and 3-dimensional pelvic floor ultrasound scanning. Palpation for the diagnosis of such trauma requires substantial teaching, and the results from palpation seem less repeatable than imaging results. Recently, MRI has become the primary imaging method of choice because of its superior spatial resolution capabilities and ability to identify the different muscle groups of the levator ani. Given the complicated 3-dimensional arrangement of the levator ani, multiple 2-dimensional MRIs are not sufficient to demonstrate the relationship of the pubovisceral muscles in a complex fashion; 3-dimensional imaging can give good results with some reconstruction software.
Currently, technologic advances in 3-dimentional ultrasonography allow access to the arbitrarily defined planes anywhere within ultrasound volume data and permit direct imaging of the entire levator hiatus. Tomographic ultrasound imaging (TUI) allows us to process imaging information into parallel slices of a predetermined number and spacing, similar to computed tomography or MRI. The depth and width of levator ani defects can also be quantified by tomographic ultrasound scanning. Dietz et al proposed an assessment of complete avulsion and provided results that were reproducible and associated strongly with vaginal palpation. This raises the issue of a potentially inferior method, palpation, being used as the gold standard for assessment of the performance of a potentially superior technology, ultrasound imaging. The true gold standard is live dissection with microscopic correlation, but it is impractical at present. Strohbehn et al reported excellent correlation between cadaveric structures and MRI anatomy, so the only other option for the determination of optimal assessment of levator defects would be to perform MRI on the same population. There are few data about agreements between MRI and 3-dimensional ultrasound scans, especially to evaluate levator defects. The purpose of this study was to determine the agreement in the detection of abnormalities in the pubovisceral portion of the levator ani muscle that are obtained with an MRI-based 3-dimensional model and by ultrasound scanning in a group of Chinese women with pelvic organ prolapse.
Materials and Methods
In this prospective study, 3-dimensional pelvic floor ultrasound scanning and an MRI-based 3-dimensional model were used to assess the abnormal morphologic condition of the pubovisceral muscle in women who had significant pelvic organ prolapse. It was conducted between December 2008 and February 2010 at the Fuzhou General Hospital, Fuzhou, Fujian Province, China. Over the course of 1 year, 69 women who were to undergo surgery for prolapse repair were assessed with the use of pelvic floor 3-dimensional ultrasound scanning and MRI. Before the imaging procedure, all participants were questioned regarding symptoms of urinary incontinence, prolapse, or fecal incontinence, with the use of a standardized questionnaire. They were interviewed and examined in the clinic for prolapse with prolapse staging according to the grading system of the International Continence Society.
The exclusion criteria were an inability to understand the instructions that were given in Mandarin, the presence of an intrauterine device made of metal, or claustrophobic tendencies that would preclude undergoing MRI. The time interval between the MRI and pelvic floor ultrasound examination was 1-3 days. Written informed consent was obtained from all participants; the Ethics Committee of the hospital granted approval for this study.
In all cases, 3-dimensional pelvic floor ultrasound scanning was performed after voiding, with the patient in the lithotomy position, with ankles close to the buttocks and about 20-30 cm apart to reduce discomfort and fatigue. A GE Kretz Voluson 730 expert system (GE Kretz Ultrasound GmbH, Zipf, Austria) was used to acquire volume imaging data with the patient at rest, on maximum Valsalva (best of ≥3 attempts), and on maximum pelvic floor contraction after visual biofeedback teaching. The most effective of at least 3 maneuvers was used for evaluation at a later date with the software 4D View V (version 7.0; GE Medical Kretztechnik, Zipf, Austria). Volumes that were obtained on maximal pelvic floor contraction were used for the assessment of muscle integrity; in those who were unable to contract, volumes were obtained at rest. The plane of the levator hiatus in the midsagittal view was identified by a line between the hyperechogenic posterior aspect of the symphysis pubis and the hyperechogenic anterior border of the pubovisceral muscle just posterior to the anorectal muscularis. This line was then used to identify the plane of minimal dimensions in the axial plane, which is the true plane of the levator hiatus. Using TUI, a set of 8 parallel tomographic slices was obtained in the axial or C-plane at intervals of 2.5 mm from 5.0 mm caudal to 12.5 mm cephalad of the plane of minimum dimensions. The appearance of the subpubic arch in the lowermost slices confirmed appropriate placement of slices. Figure 1 , A, demonstrates the identification of the reference plane, which is the true plane of the levator hiatus, in both midsagittal and oblique axial planes in a patient with an intact pubovisceral muscle. Figure 1 , B, shows the corresponding TUI. Only the reference plane of minimal dimensions and 2 parallel slices above it were considered ( Figure 2 ).
A full levator avulsion is determined if the puborectalis-to-ipsilateral sidewall attachment is not seen on any of the 3 slices. A partial avulsion is diagnosed when the puborectalis attachment to the ipsilateral sidewall is not seen on at least 1 slice. Measurements of the levator-urethra gap (LUG) were undertaken by placing calipers in the center of the hypoechogenic structure that indicates the urethral mucosa and smooth muscle and on the most medial aspect of the muscle insertion ( Figure 2 ). Data acquisition and analysis of the ultrasound images was undertaken by 2 of the authors, both fellowship-trained and experienced in ultrasound imaging, who were asked to independently perform measurements 3 times from the same volume recording. The mean of these 3 values was used for statistical analysis. Interobserver variability was also determined.
MRI examination was carried out with the patients in the supine position with a magnet (Trio 3.0T; Siemens, Berlin, Germany) and a surface coil. The following source imaging parameters were used: 3-dimensional_t2_spc_tra sequence over 8 minutes with TR 1600 msec, TE 97 msec, field of view 400 cm, slice thickness 1.0 mm/interleaved, no gap. Only static images were analyzed in this study because there are difficulties in holding the Valsalva strain or contraction effort during an MRI scan. The senior author and 1 other author, who were blinded to the subject’s symptoms and ultrasound data, reviewed each scan. On initial review, scans that appeared to have abnormalities were identified presumptively on the basis of a comparison with normal anatomy of the levator ani muscle that had been acquired by MRI data from 5 nulliparous volunteers in a preliminary study according to the methods described by Margulies et al and Kearney et al. Final classification of levator avulsion was made only when an abnormality was found in both the axial and the coronal planes and was agreed on by 2 examiners. When 2 examiners disagreed as to the presence of a disruption in the levator ani, reexamination of the scans was made to discriminate between avulsion and asymmetry to reach a consensus.
A 3-dimensional rendering was accomplished by following manual segmentation, that is, serial outlining of each anatomic structure to be displayed in the 3-dimensional rendering. This would include the levator ani, pubic symphysis, ischial spines, and the sacrococcygeal bone. Subsequently, a series of advanced imaging processing techniques that were based on triangle decimation and the marching cubes algorithm were applied to form 3-dimensional objects that can be rotated and manipulated in space. Figure 3 shows the processing of a 3-dimensional model in 1 nulliparous volunteer. With the help of a software tool, 3-dimensional Slicer 3.6 ( www.slicer.org ), computer models provide a topographic overview to aid in the visualization of the levator avulsion that is depicted in the MRI cross-sectional anatomy of the subdivisions in the axial, coronal, and sagittal scan planes. An avulsion was diagnosed on analysis of rendered volumes if a disconnection of the muscle from its insertion on the inner surface of the pubic bone was found. To quantify the levator avulsion on an magnetic resonance–based 3-dimensional model, the levator-symphysis gap (LSG) was measured as the distance from the middle of the inferior symphysis to the nearest aspect of the pubovisceral muscles on the right and left, which is similar to the LUG on ultrasound imaging. The puborectalis attachment width (PAW) was defined as the depth of origin on the inner surface of pubic bone of each arm of the levator sling ( Figure 4 ), by which we can identify “complete” or “partial” avulsion. Two authors were asked independently to perform measurements 3 times on each of the 3-dimensional datasets on 2 separate occasions using the Slicer software; the mean of these 3 values was recorded for statistical analysis. Interobserver variability was also determined for the 2 authors using intraclass correlation coefficients in the analysis of the magnetic resonance–based 3-dimensional model.
