Magnetic resonance imaging of pelvic floor dysfunction - joint recommendations of the ESUR and ESGAR Pelvic Floor Working Group

Objective To develop recommendations that can be used as guidance for standardized approach regarding indications, patient preparation, sequences acquisition, interpretation and reporting of magnetic resonance imaging (MRI) for diagnosis and grading of pelvic floor dysfunction (PFD). Methods The technique included critical literature between 1993 and 2013 and expert consensus about MRI protocols by the pelvic floor-imaging working group of the European Society of Urogenital Radiology (ESUR) and the European Society of Gastrointestinal and Abdominal Radiology (ESGAR) from one Egyptian and seven European institutions. Data collection and analysis were achieved in 5 consecutive steps. Eighty-two items were scored to be eligible for further analysis and scaling. Agreement of at least 80 % was defined as consensus finding. Results Consensus was reached for 88 % of 82 items. Recommended reporting template should include two main sections for measurements and grading. The pubococcygeal line (PCL) is recommended as the reference line to measure pelvic organ prolapse. The recommended grading scheme is the “Rule of three” for Pelvic Organ Prolapse (POP), while a rectocele and ARJ descent each has its specific grading system. Conclusion This literature review and expert consensus recommendations can be used as guidance for MR imaging and reporting of PFD. Key points • These recommendations highlight the most important prerequisites to obtain a diagnostic PFD-MRI. • Static, dynamic and evacuation sequences should be generally performed for PFD evaluation. • The recommendations were constructed through consensus among 13 radiologists from 8 institutions.


Introduction
Imaging of the female pelvic floor is of rising interest due to an ageing population, harboring an increasing incidence of pelvic floor disorders (PFD) and the rising need for Rania Farouk El Sayed and Celine D. Alt share first authorship * Rania Farouk El Sayed rania729.re@gmail.com comprehensive diagnosis and treatment. The Population Reference Bureau reported the percentage of the population aged 65 and older to be 13 % of the total population in the U.S. in 2010 with an expected increase to 20 % in 2050, whereas in Europe, the percentage was reported around 18 % in 2010 with an expected increase to 28 % in 2050 [1]. Women that are affected by PFD, often complain most about the impairment of their quality of life and ask for sufficient therapy, which is commonly surgical repair [2,3]. Thus, imaging techniques have been constantly developed in recent years to support therapy planning and management. Magnetic resonance imaging (MRI) of the female pelvic floor, particularly, combines high-resolution images with an excellent soft tissue contrast and provides the possibility to assess noninvasively and more objectively a spectrum of possible disorders affecting the pelvic floor in one examination [4][5][6][7]. There is general agreement that MRI of the pelvic floor should encompass static and dynamic MR images, whereas dynamic means imaging under maximum stress to the pelvic floor and MR defecography. Static MR images visualize pelvic floor anatomy and defects of the supporting structures, while dynamic MR images visualize pelvic organ mobility, pelvic floor weakness, pelvic organ prolapse (POP) and associated compartment defects [5,[8][9][10][11]. Additionally, MRI may diagnose unexpected underlying masked functional abnormalities, which might be discrepant from the dominant symptom and may influence the choice of the surgical technique in around 42 % of patients with different spectra of PFD [12,13]. Several studies and detailed reviews are published about MRI of the pelvic floor and different acronyms have been used for this examination including static and dynamic MR of the pelvic floor, MR defecography or MR proctography [4,12,[14][15][16]. However, to date, there is neither consensus on a standardized imaging protocol nor on a systematic reporting scheme for MR-imaging of PFD. This may be due to the complexity of the anatomy and the functional interaction of the organs with the supporting structures resulting in a broad spectrum of PFD. Another important factor that contributes to this lack of consensus is the fact that PFD is treated by urologists, urogynecologists or proctologists. Consequently, each clinician may manage the patients' condition from a different perspective. Therefore, MR-imaging acquisition varies according to the referring specialty and their rudiments for proper management and treatment decision. The wide range of different available MR protocols and a lack of standardization additionally increase variation between different centers. There is, therefore a necessity for recommendations from an expert panel that clearly defines the minimum prerequisites to obtain a state-of-the-art MR examination of the pelvic floor. This paper reports the recommendations of a panel of expert radiologists in pelvic floor imaging, which are joined in the pelvic floorworking group, which is under the umbrella of the European Society of Urogenital Radiology (ESUR) and the European Society of Gastrointestinal and Abdominal Radiology (ESGAR).

Materials and methods
The study went through five basic steps that are displayed in Fig. 1.
Step 1 Member recruitment and data sheet creation Participants for the working group were recruited among ESUR and ESGAR members between 2010 and 2011. The final working group consisted of 13 radiologists from one Egyptian and seven European institutions, all with known expertise in pelvic floor imaging. One member (RFE) created a data sheet to collect technical protocol details of the members' institution. This sheet focused on information about the clinical referrer, patient population, patient preparation, and MR technique (hardware, imaging sequence and imaging parameters).
Step 2 Review of imaging protocols of the participating institutions and data sheet creation for literature review Data collection, review and discussion of all imaging protocols of the participating institutions took place between 2012 and 2013. During this period modifications on the data sheet were implemented by (RFE) in which full details about both the geometry and the contrast of the static and dynamic MRI during straining as well as those of MRI defecography were added to the original data sheet. The results were presented and discussed in a face-to-face meeting during ECR 2014 during which a consensus was reached to finalize the data sheet for literature research (Appendix 1).
Step 3 Literature search, data collection and analysis Literature search was conducted in the Medline database for articles published between 1993 and 2013 using the following keywords: BMRI AND Pelvic FloorB, BMRI defecographyB, BMRI pelvic organ prolapseB, BMRI anal incontinenceB, BMRI stress urinary incontinenceB, BMRI AND defecographyB, BPelvic obstruction syndrome and MRIB, BPelvic outlet obstruction and MRIB, BMRI and fecal incontinenceB, BPelvic floor and MRIB, BMRI and urinary incontinenceB and BPelvic organ prolapse and MRIB.
Inclusion criteria were original data with full information about the parameters and the protocol of the examination that matched with our final data collection sheet for literature review.
Articles that were not written in English, did not deal with a human study population or lack of information about the performance of the examination were excluded. The papers concerning MRI of PFD were divided by (RFE) into the following subchapters: urinary incontinence (160 articles), pelvic organ prolapse (182 articles) and MR-defecography (172 articles). Paper revision and data extraction was divided among participating members into three subspecialty groups (urology, gynaecology and proctology) with one leader for each group (GM, CDA, DW). Each subspecialty leader wrote a final report summarizing the data that was agreed upon. The collected evidence by this literature analysis was used to extract the relevant topics, which should be addressed by the working group panelists in order to construct a questionnaire.
Step 4 Creation and analysis of a questionnaire From October 2014 to March 2015, one author (CDA) developed a questionnaire to define the most important information and requisites needed to perform MRI of PFD with standardized imaging protocol and reporting scheme. It was finalized in consensus with one author of ESGAR (DW). Since all panelists are using MR systems with a conventional closed-magnet design where the patient can only be examined in supine (lying) body position, procedural and technical aspects of pelvic floor imaging was focused to this type of magnet design. The questionnaire included binomial, multiple choice, numerical and open questions, in total 89 items (Appendix 2). This questionnaire was mailed to all panelists. In total, 82 of 89 questions were answered by all experts and were scaled according to the individual item in question for further analysis. The data obtained were analyzed using descriptive statistics. Agreement of at least 80 % was defined as consensus finding.
Step 5 Discussion and voting for the final consensus recommendations The second face-to-face meeting took place during ECR congress in 2015. For those questions that did not reach consensus at the first round of the questionnaire analysis, wording was modified to obtain better-defined statements subjected for voting by the experts in a face-to-face meeting. During that meeting the panelists discussed those items and were asked to vote. However, there were items that did not reach consensus but were reported by number of panelist to be important and warrants being included in the recommendations. These items were re-analyzed, and those that were found to be supported by case control or cohort studies from the literature, in particular level of evidence 2 according to the sign criteria, whereas expert opinion is level of evidence 4 (www.sign.ac.uk), were also included in the final recommendation.

Results
Consensus was reached for 88 % of 82 items and the recommendations regarding indication, patient preparation, imaging protocol, criteria for MRI assessment and reporting were constructed from these.

Indications for MR imaging of pelvic floor dysfunction
The indications for MR imaging of the pelvic floor that scored the highest number of agreement among the group members and the literature review are rectal outlet obstruction (92 % agreed upon), rectocele (92 % agreed upon), recurrent pelvic organ prolapse (POP) (85 % agreed upon), enterocele (85 % agreed upon) and dyssynergic defecation (anismus)(85 % agreed upon) ( Table 1).

Patients' preparation and hardware requirements
Full patients' history of pelvic floor disorder should be taken prior to scanning (consensus 100 %). The patient should be examined at least in a 1.5 T MRI unit with a phased array coil, as this is the most agreed-upon field strength (consensus 100 %). The patient is examined in the supine position with the knees elevated (e.g. on a pillow with firm consistency) as this was found to facilitate straining and evacuation (consensus 100 %). The coil should be centered low on the pelvis to ensure complete visualization of prolapsed organs [4,15]. The bladder should be moderately filled, therefore voiding 2 hours before the examination is recommended (consensus 100 %).
Prior to the examination the patient should be trained on how to correctly perform the dynamic phases of the examination and the evacuation phase (consensus 100 %). The patient is instructed to squeeze as if trying to prevent the escape of urine or feces and hold this position for the duration of the sequence. For maximum straining, the patient is instructed to bear down as much as she/he could, as though she/he is constipated and is trying to defecate [15]. For the evacuation phase, the patient should be instructed to repeat the evacuation process until the rectum is emptied.
To decrease possible patient's discomfort, a protective pad or a diaper pant should be offered to the patient, which helps to increase patients' compliance during dynamic and evacuation phases (consensus 100 %). No oral or intravenous contrast is necessary [15].
The rectum should be distended in order to visualize the anorectal junction (ARJ), rectoceles and intussusceptions, and to evaluate the efficacy of rectal evacuation (consensus 100 %). Ultrasound gel is the recommended medium to distend the rectum, however, the amount varies between 120 to 250 cc (consensus 100 %). For rectal distension a large amount of gel (180-200 cc) likely improves the capacity of the patient to defecate. A checklist for the recommended patients' preparation is listed in (Table 2).
A rectal cleansing enema prior to the examination is helpful but reached no consensus to be generally performed. Vaginal filling with 20 cc ultrasound gel is helpful for better demarcation, however, it reached no consensus for general performance and its application may be limited due to social or religious backgrounds.

MR-imaging protocol
The recommended MR-imaging protocol is summarized in (Table 3). The protocols consists of static MR sequences and dynamic sequences, whereas dynamic means imaging during straining, squeezing and during evacuation or defecation.
According to the concordance of experts and level of evidence, high resolution T2-weighted images (T2WI) (e.g. Turbo Spin Echo, TSE ; Fast Spin Echo, FSE; Rapid Acquisition with Relaxation Enhancement, RARE) Technique was reported by 3/8 experts and is supported by reference [15,21] b Technique was reported by 3/8 experts and is supported by reference [15,21] c Technique was reported by 3/8 experts and is supported by reference [22] in three planes are recommended for static images, whereas steady state (e.g. FISP, GRASS, FFE, PSIF, SSFP, T2-FFE) or balanced state free precession sequence (e.g. trueFISP, FIESTA, B-FFE) in sagittal plane is recommended for dynamic sequences (squeezing and straining) and evacuation sequence (consensus 100 %). The dynamic sequence should not exceed 20 seconds each, as breath holding is required (consensus 100 %). The evacuation sequence  should be repeated until the rectum is emptied to exclude rectal intussusception (total time duration around 2-3 minutes)(consensus 100 %). Dynamic MR imaging during evacuation is mandatory, because certain abnormalities Determine organ-specific reference points 100% 2 [25] Measure the descent of reference points below the PCL 100% 2 [15,26] 2 Measurements for posterior compartment Measure the bulging of the anterior rectal wall at evacuation phase/straining phase 100% 2 [15,20] Measure the ARA at rest -squeezing phase -straining phase/evacuation phase 100% 2 [16,27] B Reporting

C Grading 1 Anterior compartment
Use the "rule-of-three' grading for cystocele 100% 2 [32,33] Report cystocele as pathological starting from °II 88% 4 [33] 2 Middle compartment Use the "rule-of-three' grading for uterine prolapse and enteroceles 100% 2 [34,35] Report POP as pathological starting from °II 88% 4 [35] 3 Posterior compartment Use the grading for Anorectal Junction descent (ARJ) starting at 3 cm below the PCL 100% 2 [19,36] Report a rectocele as pathological starting from °II 100% 2 [19,20] Use the "rule-of-two" grading for rectoceles 88% 2 [16,19] PCL pubococcygeal line, ARA anorectal angle, POP pelvic organ prolapse, ARJ anorectal junction * Level of evidence 2 = based on systematic reviews, case control or cohort studies; Level of evidence 4 = based on expert opinion (www.sign.ac.uk) and the full extent of POP is only visible during evacuation. Optional MRI sequences can be added and acquired for further assessment of pelvic floor relaxation. These include axial and coronal dynamic sequences during maximum straining. Illustration of all the recommended imaging sequences and patients' maneuvers is summarized in (Fig. 2).
Since the performance of adequate pelvic stress during the dynamic sequences is important in order to assess the full extent of PFD, quality control of the study is essential. The study can only be considered diagnostic if a clear movement of the abdominal wall is seen during squeezing and straining. If no evacuation of rectal content at all or a delayed evacuation time (more than 30 seconds to evacuate 2/3 of the rectal content) is present, anismus should be considered (consensus 88 %) [23].

Image analysis
A clear consensus was reached that the assessment of a MR study of the pelvic floor should include analysis of static images for detection and classification of structural abnormalities. The dynamic images are analyzed with regard to functional abnormalities that are assessed by metric measurements of the three compartments of the pelvic floor (consensus 100 %) (Fig. 3). The measurements help to recognize and grade the extent of POP and pelvic floor relaxation (PFR), as well as they are used to grade anterior rectoceles and enteroceles (consensus 100 %). Both static and dynamic MRI findings as well as the results of the metric measurements should be reported in a structured MR reporting scheme (consensus 100 %) ( Table 4).
Due to the different views of the clinical specialists involved in the treatment of PFD, it is suggested to consider adapting the MRI reporting scheme according to the specialty of the referring physician. A proposal for a specialty-based MRI report is given in (Table 5).

Measurements
The pubococcygeal line (PCL), drawn on sagittal plane from the inferior aspect of the pubic symphysis to the last coccygeal joint, is recommended as reference line to measure POP (consensus 100 %). It shows the highest inter-and intraobserver reliability of MRI measurements in women with POP of the anterior and middle compartment compared to all proposed reference lines in the literature with an intercorrelation coefficient (ICC) between 0.70-0.99 (Fig. 3a) [14,37,38].
After defining the PCL, the distance from each reference point is measured perpendicularly to the PCL at rest and at maximum strain (consensus 100 %) [26,29]. In the anterior compartment, the organ-specific reference point is the most inferior aspect of the bladder base (B), in the middle compartment, the organ-specific reference point is the anterior cervical lip (most distal edge of the cervix)(C), or the vaginal vault in case of previous hysterectomy (V), and in the posterior compartment, the organ-specific reference point is the anorectal junction (ARJ) (consensus 100 %) (Fig. 3a) [15,16,20,25,29,39]. Measured values above the reference line have a minus sign, values below a plus sign (consensus 100 %) [25].
Reporting of the movement of the organs compared to their location at rest is stated to give more valuable information for the referrer than a grading system alone [8,25]. We therefore recommend giving the difference of the values at rest and during straining for each organ-specific reference point (pelvic organ mobility)(consensus 100 %) (Fig 3a, b).
A rectocele is diagnosed as an anterior rectal wall bulge and it is measured during maximum straining and evacuation (Fig 4). Typically, a line drawn through the anterior wall of the anal canal is extended upward, and a rectal bulge of greater than 2 cm anterior to this line is described as a rectocele (consensus 100 %) [28,34]. The anorectal angle (ARA) should be drawn along the posterior border of the rectum and a line along the central axis of the anal canal on sagittal plane (Fig. 4b) at rest, squeezing and maximum straining (consensus 100 %) [20,27].
Pelvic floor relaxation (PFR) often coexists with POP, but it is a different pathologic entity. For quantification of the (c) Shows how to measure and diagnose a pathological rectocele: a line drawn through the anterior wall of the anal canal is extended upward, and a rectal bulge of greater than 2 cm anterior to this line is described as a rectocele (R). The levator plate angle (LPA) is enclosed between the levator plate and the PCL. d,e. Dynamic Balanced Fast Field Echo (BFFE) sequence in axial (d) and coronal (e) plane at rest and during maximum straining. In the axial plane the width of the levator hiatus is enclosed between the puborectalis muscle slings. On the coronal plane, the iliococcygeus angle is measured between the iliococcygeus muscle and the transverse plane of the pelvis in posterior coronal images at the level of the anal canal weakness of the levator ani and to reflect pelvic floor laxity, five measurements can be performed [15] , however, it reached no consensus to measure it routinely. The length of the hiatus (H-line), the descent of the levator plate (M-line) and the levator plate angle are evaluated in the sagittal plane (Fig 4a, c), whereas the transverse width of the levator hiatus and the iliococcygeus angle are assessed in the axial and coronal plane during maximum straining (Fig. 4e,d) [15]. Table 6 provides an overview of the entire spectrum of the published reference values for quantitative MR-measurements of the pelvic floor.

Grading
The BRule of three' is the recommended grading system in the anterior and middle compartment starting at 1 cm below the PCL (Table 4) [15,16,32,34,40]. This is based on the fact that the pelvic floor may descend and widen up to 2 cm during abdominal pressure. Consequently, the pelvic organs follow the movement of the pelvic floor inferiorly but without protrusion through their respective hiatuses [4]. The bladder base, particularly, may descend up to 1 cm below the PCL during straining in continent women and should not be stated as a cystocele (consensus 100 %) [24,34].
Small intussusceptions of the rectal wall are considered to be normal findings during defecation, observed in nearly 80 % of healthy subjects [41].

Functional abnormalities on dynamic MR images
Loss of urine through the urethra during maximum straining records urinary incontinence (UI) and should be reported if present (consensus 88 %) [15]. Urethral hypermobility as a predictor for UI should be reported if present (consensus 88 %) [29]. If a cystocele is present, the differentiation of a distention or a displacement cystocele can be made, which is helpful for therapy planning, however it reached no consensus for general reporting [42].
If an enterocele is present, the report should include the content of the peritoneal sac, as clinical examination alone  [15] may have shortcomings in identifying the content (consensus 100 %) [5,20,22,31,43]. The end of evacuation phase is important to identify intussuception (Fig. 3c) [30].
The change of the ARA during dynamic and evacuation sequence compared to the ARA at rest expresses the functioning of the puborectal muscle. In particular, the ARA should sharpen during squeezing and should become more obtuse during straining and evacuation [16,27,39]. We recommend to report the individual function, as the literature presents with a widespread of normal reference values (consensus 100 %).

Structural defects on static MR images
Description of structural defects and anatomical abnormalities, that are assessed in static T2WI are more likely specialty-based PFD-related questions from the referrer ( Table 5). The functional three-part pelvic supporting system (Fig. 5) includes the urethral support system, which maintains urinary continence; the vaginal support system, which prevents prolapse; and the anal sphincter complex that maintains anal continence. Urethral support system defects may include urethral ligament defect and / or distortion, level III endopelvic fascial defects, or puborectalis muscle detachment (Fig. 5b), disruption, atrophy or avulsion [15,18,21,33,[44][45][46]. The spectrum of vaginal support system abnormalities includes level I and II paravaginal fascial defects and/or iliococcygeus diffuse or focal muscle abnormality [35].

Limitations of the study
The study has few limitations. Four panelists who participated in Step 1 and 2 of the study were from the same institution. Therefore, only 1 out of their 4 completed questionnaire was included in the final analysis to avoid biased results. Nevertheless, since all 8 panelists who have completed the questionnaire were from different institutions these recommendations can be considered to represent the entire spectrum of expert opinions in the field of pelvic floor MRI. Second, the recommendations given in this study with regard to technical aspects of MRI of the pelvic floor relate to conventional closedconfiguration magnets for MR imaging allowing patient positioning in lying body position only. However, this is the most agreed upon scanner, in addition several studies have shown that patient positioning does not significantly influence diagnostic performance of MR imaging of the pelvic floor [17,19,47,48].

Conclusion
Based on an extensive literature review and analysis and of expert consensus, these proposed recommendations can be used as guidance for standardized MR imaging and reporting of PFD. Nevertheless, our joint ESUR-ESGAR pelvic floorworking group is aware about the complexity of the topic and that further studies are mandatory to achieve additional refinements of guidelines for MR imaging, diagnosing and reporting of PFD.