Transperineal and endovaginal ultrasound for evaluating suburethral masses: comparison with magnetic resonance imaging

To evaluate the utility of pelvic floor ultrasound (US) in the detection and evaluation of suburethral masses, using magnetic resonance imaging (MRI) as the reference standard.


INTRODUCTION
Suburethral masses are either congenital or acquired in etiology. It is difficult to estimate their incidence due to the low number of cases reported, however, reported rates are less than 1% in population-based studies 1,2 . Suburethral masses occur in all ages and most frequently in women aged 30-60 years 3 . Differential diagnoses include urethral diverticulum, Gartner duct cysts, vaginal inclusion cysts, Skene gland abscess and urethral caruncle 1,[4][5][6] . Urethral diverticulum is the most common, accounting for approximately 80% of cases 1 .
Congenital suburethral masses arise from various embryological components and remnants of the genitourinary system, particularly Müllerian. They may display columnar, ciliated-columnar or stratified squamous epithelium 3,7 . Inclusion cysts are lined with stratified squamous epithelium and may contain purulent material 3,7 . Urethral diverticula are lined typically with transitional epithelium, but may undergo squamous or glandular metaplasia secondary to chronic inflammation 8 . Suburethral masses have a smooth external surface and may be filled with caseous, purulent or thin mucoid material 7 . They are usually small but vary in size, ranging from a few mm to over 5 cm in diameter 9 .
Signs and symptoms of suburethral masses include presence of a painful palpable mass, dyspareunia, vaginal discharge, dysuria and voiding dysfunction 4 . Most suburethral masses are benign and can be successfully excised surgically. However, operation of an undiagnosed urethral diverticulum can result in complications including urethral stricture and urethrovaginal fistula 4 . Therefore, accurate diagnosis with imaging and preoperative planning is important.
Although magnetic resonance imaging (MRI) has not been established formally as the imaging reference standard in this context, it is the recommended modality for the detection of suburethral masses 4 . In recent years, pelvic floor ultrasound (US) has gained increasing popularity among urogynecologists for the diagnosis of pelvic floor disorders 10,11 . Compared with MRI, US offers better availability of equipment and healthcare professionals to report the imaging results, reduced cost, less patient restriction as well as better patient tolerance 12 . Moreover, MRI failed to detect a urethral diverticulum in 7.3% of cases in a previous report 13 .
The aim of this study was to evaluate the utility of pelvic floor US in the detection and evaluation of suburethral masses, using MRI as the reference standard.

METHODS
This was a retrospective analysis of US and MRI scans of all patients diagnosed on clinical examination with a suburethral mass in a single urogynecology clinic (Urogynaecology and Pelvic Floor Reconstruction Unit, Croydon University Hospital, London, UK) over a 13-year period (February 2007 to March 2020). Women with a vaginal lump or lower urinary tract symptoms suggestive of a suburethral mass were referred to the clinic by either their general practitioner or a different specialty team within the hospital.
All women were examined using two-dimensional transperineal US (2D-TPUS) with or without threedimensional endovaginal US (3D-EVUS), using the Flex Focus 500 US system or Pro Focus 2202 system (BK Medical, Herlev, Denmark). 2D-TPUS was performed using a convex transducer applied to the perineum. Images were obtained in the coronal, axial and sagittal planes. 3D-EVUS was performed by inserting the endocavitary probe into the vagina in a neutral position and obtaining images in the midsagittal view and a 3D volume of the surrounding structures. If the mass was too large or distal, assessment using 3D-EVUS was not possible. US was performed with the woman in the supine position and the bladder partially filled. The urethra was seen as a hypoechoic structure extending from the bladder base to adjacent to the vaginal opening. Cystic structures were seen as hypoechoic areas and solid masses of mixed echogenicity. Images were analyzed by reporting on the presence of a suburethral mass, its size, location and whether there was communication with the urethral lumen. The three perpendicular diameters of the mass were measured in mm in the coronal, axial and sagittal planes and the largest diameter of the mass was used in the analysis. The location of the mass was determined by measuring the distance between the most proximal edge of the mass and the bladder neck in the sagittal plane. Communication with the urethral lumen was diagnosed when there was disruption of the hyperechoic rhabdosphincter, with a hypoechoic connection between the mass and the urethral lumen.
MRI was considered the reference standard in this study. Unenhanced T1-weighted and T2-weighted MRI images were obtained. A suburethral mass was diagnosed if a mass lesion was present and located posterolateral to the urethra 14 . The two perpendicular diameters of the mass were measured in mm in the axial and sagittal planes, and the largest diameter of the mass was used in the analysis. The location of the mass was measured as the distance between the most proximal point of the mass and the bladder neck in the sagittal plane. Communication with the urethral lumen was diagnosed when there was a hypointense disruption within the hyperintense rhabdosphincter, connecting the mass and the urethral lumen.
Pelvic floor US images were analyzed blinded to MRI results. In addition to imaging results, outcome management (conservative or surgical) and histology results were noted. All clinical and imaging information was entered into a dedicated Microsoft Excel database.
Data were analyzed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA). Categorical variables are presented as n (%) and continuous variables as median (range) or median (interquartile range (IQR)). The normality of distribution of continuous variables was tested using the Shapiro-Wilk test. Differences between the measurements obtained by US and MRI were analyzed using the Wilcoxon signed-rank test. To assess the agreement between US imaging and MRI, intraclass correlation coefficients (ICC; 3,1) were used. A value between 0.75 and 0.9 indicated good agreement and > 0.90 indicated excellent agreement 15 . Standard error of measurement (SEM) was then calculated to measure the range of error for each measurement.
MRI detected a suburethral mass in 34 women. 2D-TPUS detected correctly a suburethral mass in all 34 women, 27 of whom also had 3D-EVUS. However, pelvic floor US also detected a suburethral mass in the remaining six women. Therefore, the agreement between US and MRI for detecting a suburethral mass was 85.0% (95% CI, 70.2-94.3%).
The US and MRI characteristics of the suburethral masses detected on both modalities are shown in Table 2. On US, most suburethral masses were classified as simple and unilocular (67.6%) with hypoechoic contents (82.4%), while 32.4% were complex due to septations leading to a number of fluid-filled locules within the mass. In addition, 20.6% of masses encircled the urethra, described as horseshoe in appearance. Figures 1 and 2 demonstrate different appearances of suburethral masses on 3D-EVUS and 2D-TPUS.
The six suburethral masses detected on US but not on MRI were distal (median distance from the cyst to the bladder neck of 30.9 (range, 28.3-34.0) mm) with a median diameter of 7.8 (range, 5.0-15.0) mm. All six masses were unilocular and two were hyperechoic, two had mixed echogenicity and the other two were hypoechoic. Connection to the urethra (Figure 2b,d) was detected on 2D-TPUS in the sagittal plane in seven women. In total, Data are given as median (range) or n (%). *Largest measurement obtained on either two-dimensional transperineal US (2D-TPUS) or three-dimensional endovaginal US (3D-EVUS) was used. †Smallest measurement obtained on either 2D-TPUS or 3D-EVUS was used. ‡Urethral connection was seen only on 2D-TPUS in the midsagittal plane.
17/34 women underwent cystoscopy (used traditionally to diagnose a urethral connection) 4 and a urethral connection was seen in five of them. Of the seven women diagnosed with urethral connection on 2D-TPUS, two underwent cystoscopy and the finding was confirmed in both. Connection to the urethra was seen in one patient on MRI, and this finding was confirmed on cystoscopy. Agreement between the measurements obtained on 2D-TPUS and 3D-EVUS in comparison to those taken on MRI is presented in Table 3. There was no significant difference between MRI and 2D-TPUS and 3D-EVUS with regards to the measured distance between the suburethral mass and the bladder neck or the largest diameter of the mass. The ICC analysis showed good agreement between MRI and 2D-TPUS for the measured distance from the suburethral mass to the bladder neck (ICC, 0.89; SEM, 3.64 mm) and excellent agreement for measurement of the largest diameter of the suburethral mass (ICC, 0.93; SEM, 4.31 mm). Good agreement between MRI and 3D-EVUS was observed for the measured distance between the suburethral mass and the bladder neck (ICC, 0.88; SEM, 3.48 mm) and excellent agreement for measurement of the largest diameter of the suburethral mass (ICC, 0.94; SEM, 4.68 mm).
Eighteen (45%) women underwent surgical excision of their suburethral mass. The masses were detected in all 18 women on both US and MRI. Of the remaining patients, 12 (30%) declined surgical intervention, five (12.5%) did not attend further follow-up and in five (12.5%) the suburethral mass resolved spontaneously. The most common histological finding described by pathology was urethral diverticulum in 11 patients. Of these, 36.4% had evidence of nephrogenic metaplasia and 36.4% had evidence of chronic inflammation and ulceration. Histological diagnoses in the remaining seven women included a benign suburethral cyst in two (with ciliated and squamous epithelium, respectively, and so likely of congenital origin), a large Bartholin's cyst in one, a leiomyoma in two, a malignant mesonephric adenocarcinoma in one and a malignant melanoma in one 16 . The only US feature that allowed distinction between suburethral masses of differing etiology was the presence of a urethral connection in urethral diverticula, which was seen on US in three (27.3%) cases with a urethral diverticulum. On MRI, urethral connection was seen only in one (9.1%) case with urethral diverticulum. Table 4 further describes the US and MRI findings of the different histological diagnoses, highlighting that both modalities identified similar features.

DISCUSSION
The findings of this study show that pelvic floor US has good correlation with MRI in the detection and measurement of suburethral masses. All masses identified on MRI were also seen on US. Moreover, the two modalities identified similar features and characteristics of the suburethral masses.
Our imaging findings concur with sonographic and gross macroscopic features described by others 5,7,10,17-20 . The suburethral mass size ranged between 0.1 and 7.2 cm in our cohort, with most being unilocular Table 3 Agreement between measurements obtained on magnetic resonance imaging (MRI) and those obtained on two-dimensional transperineal ultrasound (2D-TPUS) and three-dimensional endovaginal ultrasound (3D-EVUS) to determine the size and location of suburethral masses   with low-level internal echoes 5,10,17 . Mass contents may include debris, purulent, caseous or mucoid fluid, which accounts for the differing internal echogenicity reported 7 . Other studies have reported that suburethral masses may also be complex and multilocular due to internal septa 17,18 , particularly urethral diverticula, which have been described to be complex and encircle the urethra 20 , as shown in this study. Urethral leiomyomas are reported to be well-circumscribed with varying consistency, therefore, may be hypoechoic, particularly when cystic degeneration has occurred 21,22 . This happens typically in larger masses, such as the leiomyoma described in one case in this study, which measured 72 mm in its largest diameter on US 22 . Apart from urethral connection, which was seen in some of the cases with urethral diverticulum, we did not observe any distinguishing US characteristics in the other histological types of suburethral mass. There is even greater difficulty in differentiating reliably between suburethral masses secondary to carcinoma or inflammation 10 . In a recent case report we demonstrated that suburethral masses secondary to malignant melanoma and mesonephric adenocarcinoma possess similar US features to those of different etiology 16 .
MRI is the recommended imaging modality for the detection of suburethral masses 4 , as it provides good delineation of the urethral anatomy due to its soft tissue contrast. T2-weighted images typically demonstrate suburethral masses well, as their fluid contents appear bright 23,24 . MRI has been reported to be able to delineate the ostium of a urethral diverticulum in 85% of cases, which other modalities, such as cystoscopy and US, have been described to often miss 24 . However, Gillor and Dietz 20 recently showed, in a large retrospective study of 4121 women, that multiplanar TPUS can be used to identify urethral diverticulum. In this study, a diverticulum ostium was demonstrated on US and confirmed on cystoscopy in 40% of patients, compared to 20% with MRI.
We showed that US is comparable to MRI in its ability to assess the structure and consistency of suburethral masses. Agreement between US and MRI was 85% for the detection of suburethral masses and good to excellent for the measurement of the largest mass diameter and the distance between the mass and bladder neck.  for the investigation of other pelvic floor pathologies, such as pelvic organ prolapse and levator ani defects, showing moderate-to-excellent correlation between the two modalities [25][26][27][28] . However no study to date has compared the correlation of these two methods in the evaluation of suburethral masses.
Pelvic floor US did not miss any suburethral mass detected on MRI. However, a possible suburethral mass was seen in an additional 15% (6/40) of women in comparison to MRI. These masses were smaller in size and more distal in comparison to those detected on both modalities. It is important to note that these women did not have diagnostic confirmation with histology, so we were unable to confirm whether this was a true abnormality or artifact; however, as these women presented with symptoms, we believe it is unlikely that the mass seen on US was an artifact. MRI errors in detecting suburethral masses have previously been described in the literature. In a cohort of 41 patients who underwent urethral diverticulum surgery, Chung et al. 13 showed that preoperative MRI failed to identify the presence of a urethral diverticulum in 7.3% of cases. A potential reason for this is that MRI has limited sensitivity in identifying masses with little fluid content and those of small size. Also, MRI is not dynamic and captures a sequence of images at a single timepoint 13 . These image slices can be relatively thick (3-6 mm) with spacing between them, which could result in areas being missed and lesions not being identified 29 .
3D-US offers advantages, including its real-time imaging and superior temporal and spatial resolution 30 . Furthermore, the ability to manipulate the high-resolution 3D volumes in all planes, including the axial plane, previously only visible on MRI, allows for effective visualization of the pelvic floor anatomy and pathology 30 . Although the 3D-EVUS modality can obtain anatomical views that 2D-TPUS cannot, essential images can still be obtained using solely 2D imaging. The limitation of 3D-EVUS is that it may cause discomfort, particularly in case of large and inflammatory masses, and can compress and displace the urethra 31 . This may explain the inability to identify a diverticulum ostium on 3D-EVUS in this series. Therefore, in this situation 2D-TPUS would suffice. However, multiplanar transperineal systems are also available which allow real-time 3D (or four-dimensional) assessment of the pelvic floor 30 .
US has the advantage that the examination and assessment of history and US findings can be performed in the same clinic appointment, which allows correlation of the findings with patient symptoms, and subsequent discussion and counseling for further management if required. Furthermore, US can be a valuable tool in preoperative planning and can be used intraoperatively to anticipate possible complications, such as bladder or urethral injury based on anatomical proximity, therefore potentially improving surgical outcome. Pelvic floor US is not readily available in all units, it is highly operator dependent and requires expertise for interpretation. However, it is rapidly becoming a popular tool amongst pelvic floor clinicians.