European Society of Urogenital Radiology (ESUR) Guidelines: MR Imaging of Leiomyomas

Objective The aim of the Female Pelvic Imaging Working Group of the European Society of Urogenital Radiology (ESUR) was to develop imaging guidelines for MR work-up in patients with known or suspected uterine leiomyomas. Methods Guidelines for imaging uterine leiomyomas were defined based on a survey distributed to all members of the working group, an expert consensus meeting at European Congress of Radiology (ECR) 2017 and a critical review of the literature. Results The 25 returned questionnaires as well as the expert consensus meeting have shown reasonable homogeneity of practice among institutions. Expert consensus and literature review lead to an optimized MRI protocol to image uterine leiomyomas. Recommendations include indications for imaging, patient preparation, MR protocols and reporting criteria. The incremental value of functional imaging (DWI, DCE) is highlighted and the role of MR angiography discussed. Conclusions MRI offers an outstanding and reproducible map of the size, site and distribution of leiomyomas. A standardised imaging protocol and method of reporting ensures that the salient features are recognised. These imaging guidelines are based on the current practice among expert radiologists in the field of female pelvic imaging and also incorporate essentials of the current published MR literature of uterine leiomyomas. Key Points • MRI allows comprehensive mapping of size and distribution of leiomyomas. • Basic MRI comprise T2W and T1W sequences centered to the uterus. • Standardized reporting ensures pivotal information on leiomyomas, the uterus and differential diagnosis. • MRI aids in differentiation of leiomyomas from other benign and malignant entities, including leiomyosarcoma. Electronic supplementary material The online version of this article (10.1007/s00330-017-5157-5) contains supplementary material, which is available to authorized users.


Introduction
Leiomyomas (syn: myomas or fibroids) constitute the most common benign uterine tumours. They are typically multiple and are classified as submucosal, intramural and subserosal. It is estimated that 20-50% become symptomatic, with symptoms including abnormal uterine bleeding, pelvic pain, bowel or urinary obstruction, or pregnancy related complications [1]. Due to their hormonal sensitivity, they shrink in menopause and occasionally may rapidly grow and bleed during pregnancy. Torsion or infarction can cause an acute abdomen [2].
Sonography is the first line imaging modality to assess leiomyomas. MRI, due to its better soft tissue contrast, larger field-of-view and multiplanar imaging capabilities, assists clinicians in complementary pre-treatment mapping (exact origin, number and size) and in differential diagnostic issues [3]. MRI is better able than ultrasound to distinguish malignant ovarian lesions that might otherwise mimic leiomyomas. MRI is also better able to detect features of malignant change in a leiomyoma, which although rare, are of vital importance if the leiomyoma is being considered for morcellation during laparoscopic removal. Morcellation carries the risk of disseminating malignant tissue. MRI has been shown to impact on treatment planning in about 20 % of patients with symptomatic leiomyomas [4][5][6]. Although not mandatory with highquality ultrasonography, MRI has become the mainstay of diagnosis and has an important role for treatment planning and monitoring of uterine embolization.
Clinical guidelines for diagnostic work-up and treatment of patients with leiomyomas are well established (e.g. NICE clinical knowledge summary, Fibroids 2013. cks.nice.org.uk and the Society of Obstetricians and Gynaecologists of Canada, clinical practice guideline 318, February 2015). However, despite its wide utilisation, a lack of standardised recommendations/guidelines for MR imaging and reporting of leiomyomas is notable.
These recommendations prepared by the Female Pelvic Imaging Working Group of the European Society of Urogenital Radiology (ESUR) are based on the current practice among specialist radiologists in the field of female pelvic imaging, and also incorporate essentials in MRI in the light of the current published literature of uterine leiomyomas.

Materials and methods
Questionnaire A questionnaire was designed (RKH) and then amended by two other authors (MW, RF). Indications, clinical practice, technical details including patient preparation, the MRI protocol and details of reporting were assessed. The questionnaire is added in the supplementary material. For some questions more than one answer was possible. Not all answers were ticked by all respondents.
Institutional Review Board approval was not required because this manuscript presents guidelines only, without the use of patient-sensitive data.
The questionnaire was distributed to all members (55 members from 46 different institutions) of the Female Pelvic Imaging Working Group of the ESUR. All surveys returned by January 15 th 2017 were included in the analysis.

Literature search
In addition, published literature between 2007 and 2017 from Medline (Ovid), EMBASE (embase.com) and Cochrane Library was reviewed. The literature was searched with the concepts of MRI and Uterine Leiomyoma and its synonyms (angioleiomyoma* or angiomyoma* or elastomyofibroma* or fibromyoma* or hemangioleiomyoma* or hemangiomyoma* or leiomyoma* or myofibroma* or myofibromatos* or fibroid* or fibromyoma* or myoma* in combination with uterine or uterus), using a combination of MeSHrespectively EMTREE-Terms and freetext search in the title and abstract. The search includes original research articles as well as review articles. The results were limited to English language and to publications from important radiology and gynaecology journals (impact factor above 1).

Consensus meeting
Results were discussed at the Female Imaging Working Group meeting at the ECR 2017 on March 2, 2017 in Vienna and recommendations amended based on the inputs of the members present. A draft manuscript proposing an imaging algorithm for the work-up of uterine leiomyoma was then developed by the three leading authors (RKH, MW, RF) and subsequently finalized with contributions from the other co-authors. The final version of the manuscript was circulated among members of the Female Pelvic Imaging Subcommittee for approval.

Participating institutions
Most of the 26  Previous ultrasound examination of the uterus is performed by gynaecologists (n=21) and/or radiologists (n=16) and/or sonographers (n=2). When reporting MR scans, ultrasound images are available always or mostly in 13 centres, while images can be used for comparison only sometimes or not at all in the other institutions.

Role of computed tomography
Twenty respondents did not see any role for computed tomography (CT) in assessing uterine leiomyomas. Five believe CT plays a role in selected (n=4) or all (n=1) indications (e.g. acute pain, to assess calcifications or CT angiography in the follow-up of uterine artery embolization with incomplete response).

Patient preparation
In most institutions (n=20/25), the patient is not scheduled according to her menstrual cycle, whereas in 4 institutions the menstrual cycle is considered. No answer was given by one institution.
About two-thirds of the participants (n=17/25) ask for clinical symptoms or provide a questionnaire including questions such as time of last menstruation, clinical symptoms or hormonal medication.
Seventeen of 25 sites ask the patient to fast before the examination. Fasting ranges between 2-6 hours, mostly (n=9) of 4 hours.
Only 4 institutions never use antispasmolytic agents. Glucagon® (n= 9) and/or Buscopan® (n= 15) is given in the other 21 institutions. Whereas Glucagon® is administered intravenously, for Buscopan® intravenous and intramuscular applications are used.
There was agreement at the ECR consensus meeting that a Bhalf-full^, moderately distended urinary bladder is optimal for pelvic MR assessment, with the instructions given to the patients to achieve this varying among the institutions.

MR equipment & MR protocol
Different machines (GE, Philipps, Siemens) are used. Nine institutions performed MRI at 1.5 Tesla only, 3 institutions at 3 Tesla only, in 11 institutions scanners with both field strengths were used; no answer regarding MR equipment was given by 1 institution.
A variety of coils are used for imaging with a preference for pelvic phased array coils and multi-channel body coils.
Patient positioning (head first or feet first) is not uniform; some institutions remarked that this depended on the clinical question.
Imaging of the upper abdomen and kidneys is included by 10/25 sites. All perform a sagittal T2W (weighted) series of the female pelvis and the uterus. Additional oblique T2W images are performed in 17, whereas two centers perform axial and coronal T2W scans without angulations to the uterus.
Twenty-one centers perform axial, native T1W of the pelvis, one group prefers coronal T1W imaging and two others sagittal T1W pre-contrast planes. T1WFS depends on institutional preference and on clinical questions. For distinction of adenomyosis from subserosal endometriosis, precontrast FS T1W sequences are considered most helpful.
Diffusion-weighed imaging was routinely used by 21 members and sometimes by 4 others. The rationale given is that DWI is part of the routine uterine protocol but not used specifically for leiomyomas.
Two institutions never use intravenous Gadolinium for imaging of leiomyoma. All the other institutions use Gadolinium either routinely administered (n=12) or depending on the clinical question (n=11). Indications for intravenous administration include treatment planning, differential diagnosis of leiomyosarcoma, and differential diagnosis from an adnexal mass as well as follow up of uterine artery embolization.
Dynamic contrast enhanced imaging (DCE) is routinely performed in 5 sites and in selected cases in 11 sites, whereas 9 institutions never use dynamic imaging. Indications for DCE include the differential diagnosis from an adnexal mass, differential diagnosis of leiomyosarcoma, or vascularisation of the uterus and the leiomyoma before embolization.
Late contrast enhanced scans are obtained at 14 institutions.
None of the institutions participating in the survey performs MR angiography as part of their routine protocol. Twelve sites perform MR angiography in selected cases, i.e. if embolization is a treatment option.

Reporting
Standardised proforma reporting is only provided by 4/ 24 institutions. In the written report all institutions include the size, localisation, signal characteristics on T1W and T2W and differential diagnosis.
Consensus was reached that the presence of commonly coexisting adenomyosis needs to be clearly stated. In those cases, the thickness of the junctional zone should also be reported.
Except for 4 institutions, the size of the uterus is reported. Consensus was reached that the length of the cervix and corpus uteri, respectively, should be stated in the written report. Several institutions include all 3 dimensions of the uterine corpus.
The number of leiomyomas in the uterus is reported by all except 3 centres. Almost all (n=24/25) institutions report necrosis or degeneration as well as the perfusion pattern. In most cases (n=23/25), other findings in the abdomen will be reported. Lymph node enlargement or kidney obstruction as well as the distance of the fundus uteri to the umbilicus is reported for laparoscopic surgery.
Eleven institutions incorporate the FIGO classification of leiomyoma in their report.

Imaging the pregnant patient
Almost half of the institutions (n=12/25) image patients with leiomyomas during pregnancy. Questions to answer are the relation to the placenta and the internal cervical os, assessment of acute abdominal pain and indeterminate sonography. During pregnancy, fast imaging sequences are used and Gadolinium is avoided.

Discussion
The variety of responses to the questionnaire emphasises the need for this consensus statement. Practice clearly varies between institutions, sometimes on very basic issues. The authors intend that this article will help institutions understand the rationale behind currently considered best practice and this, in turn, will improve and unify practice.

Patient preparation
Neither optimal scheduling according to the menstrual cycle nor a clinical questionnaire before MRI of uterine benign disease is covered in the recent literature. It is recommended that institutions do not schedule their patients according to their menstrual cycle, but obtain clinical information at the time of the scan (e.g., time of menstruation, clinical symptoms, hormonal medication, prior surgical procedures). This regimen obtained consensus among the ESUR Female Pelvic Imaging Working Group (Table 1).
Despite lack of substantial evidence to support fasting for pelvic MRI examinations, there was agreement that patients should fast (3-6 hours), especially when intravenous (iv) gadolinium injection is considered. This is also supported by previous ESUR guidelines [3,7,8]. The rationale behind fasting is to reduce the risks from vomiting occurring as a reaction to the contrast medium injection as well as to reduce bowel motion artifacts.
There are no new data regarding the use of antispasmodic agents for the reduction of bowel artefacts. Its supporting evidence has also been extensively discussed in previous ESUR guidelines [3,[7][8][9][10]. The committee recommends the use of antiperistaltic agents for the improvement of image quality of pelvic MR (20 mg butyl-scopolamine im/iv or 1mg of glucagon iv), unless it is contraindicated.
There is consensus that a moderately filled bladder is optimal for pelvic MRI. The literature only sparsely addresses this issue. Previous ESUR guidelines recommend patients should empty their bladder 1h prior to examination to achieve a moderately filled bladder [8,[11][12][13][14][15].

MRI technical considerations (Table 2)
Although there are no studies comparing 3 Tesla versus 1.5 Tesla or pelvic array coils versus multi channel body coils in benign uterine lesions, all these options seem valuable.
It is recommended that institutions routinely include at least one fast T2WI sequence of the upper abdomen. This strategy allows exclusion of complications in patients with large leiomyoma, e.g. signs of compression such as hydronephrosis, or to rule out lymphadenopathy or metastases in patient with suspected malignancy.
We recommend a basic MRI examination protocol with at least two T2WI orthogonal planes of the uterus, including a sagittal sequence of the uterine corpus. Doing the sagittal sequence first allows subsequent scan planes to be aligned to the axis of the uterus rather than just aligned to the body. An oblique sequence aligned along the long axis of the uterus is optional, but provides useful information for anatomic localization of leiomyomas, particularly in the work-up of infertility, as it best displays the relationship of the leiomyomas to the endometrial cavity. A 3D T2W sequence may be an alternative in selected cases, e.g. mapping in multiple myomas. In the work-up of an adnexal mass of indeterminate origin, a T2W plane selected across the maximum point of contact between the mass and the uterus is advised [10,13,16,17] (Tables 2 & 3).
Besides the T2W sequences, the basic protocol should include an axial T1W sequence in order to assess other pelvic pathologies and to depict high-signal intensity lesions. In this scenario, a fat-supressed T1W sequence in the same plane allows characterization of haemorrhagic from fatty uterine pathologies (lipoleiomyomas vs leiomyomas with haemorrhagic degeneration; haematometra; endometriosis and adenomyosis), adnexal masses (teratomas or endometriomas) or other pelvic fat or blood-containing masses [10,13,16,17] (Table 3).
Major indications for iv gadolinium include: characterization of a rapidly growing leiomyoma; characterization of a leiomyoma with areas of high signal intensity on T2W images; differentiation from an adnexal mass; pre-embolization assessment including vascularization and viability of leiomyomas; post-embolization assessment to define treatment response in persistence of clinical symptoms or postembolization complications [12,[17][18][19][20][21][22][23] (Table 3).  (3-6h) • The use of antiperistaltic agents is recommended (20 mg butyl scopolamine im/iv or 1 mg of glucagon iv), unless their use is contraindicated due to patient medical background • Patients should empty their bladder 1h prior to examination in order to achieve a moderately filled bladder -Add an oblique coronal T2W sequence along the long axis of the uterus to the basic protocol -Add dynamic contrast-enhanced study to the basic protocol in the oblique coronal plane along the long axis of the uterus. Alternatively select a plane across the maximum point of contact between the mass and the uterus -If the mass shows to be ovarian follow the ESUR guidelines for characterization of indeterminate adnexal masses. • Work-up of a rapid growing uterine mass -Add dynamic contrast-enhanced study to the basic protocol in the plane that best depicts the morphology of the lesion to be characterized -DWI study is optional • Work-up of an intermediate to high T2W leiomyoma -Add dynamic contrast-enhanced study to the basic protocol in the plane that best depicts the morphology of the lesion to be characterized -DWI study is optional • Pre-and post-embolization assessment -Add MR angiography or dynamic contrast-enhanced study to the basic protocol -Add DWI sequence to the basic protocol -MRA pre-embolization: evaluate the number, site of origin and size of uterine and ovarian arteries -MRA post-embolization is optional: useful in demonstrating eventual collateral supply to LM that where not fully devascularized Gadolinium is optimally administered in a dynamic acquisition in the plane that best depicts the morphology of the lesion to be characterized, as determined based on the T2W sequence.
Especially for radiologic interpretation of therapeutic response, image subtraction can be helpful in distinguishing enhancing areas from T1W hyperintense material such as haemorrhage or debris.
In a T2W hypointense solid mass of indeterminate origin, studies have shown that ovarian fibromas enhance late and less compared to the myometrium, whereas typical subserosal leiomyomas tend to have a similar enhancement pattern [18,19].
3D-reconstructed contrast-enhanced MR angiography, has proven to be valuable in determining details of the pelvic vasculature and the optimal image intensifier Carm obliquity before embolization, thus reducing the fluoroscopy time, the radiation dose and the contrast medium volume (Fig. 2) [20][21][22]. Currently, MR angiography is not part of the routine protocol. Twelve of the participating institutions perform MR angiography in selected cases.
Based on the current literature DWI is not mandatory for assessing leiomyomas in a clinical routine setting. However it might have a potential for characterization of T2W hyperintense leiomyomas or in rapid growth (Fig. 3). Both leiomyosarcomas and benign cellular leiomyomas display restricted diffusion. Thus, restricted diffusion in a leiomyoma is a pitfall of which radiologists should be aware [11,[27][28][29]. Two publications found significantly lower mean ADCs in leiomyosarcomas than in degenerated leiomyomas (Fig. 3) [30,31]. As reported previously by our working group, for gynaecological imaging, the optimal b value is usually 800-1000 s/mm 2 , but may be increased up to 1200 or 1400 s/mm 2 [10].
DWI has also been shown to be useful in pre-and post-embolization evaluation of uterine leiomyomas. Persistent decrease of the ADC values of the dominant leiomyoma 6 months following embolization is reported, which also correlates with its volume reduction [32][33][34][35][36][37]. Therefore, a high pre-embolization ADC value seems related with a greater volume reduction and to be useful in predicting the response of a dominant leiomyoma to embolization [33][34][35].  Fig. 2 Ovarian artery parasitation. 3D reconstructed MRA image depicts an enlarged left ovarian artery (arrow) extending from the aorta to the pelvic midline, where a bulky leiomyoma is located, apparently contributing to the leiomyoma supply. Note the typical corkscrew appearance of the ovarian artery. The right ovarian artery is not seen as the normal caliber ovarian arteries are too small to be depicted by this imaging modality

Reporting
Although standardized pro forma reporting was used in only a few institutions, there was consensus that this will become an issue in the future and that the written report has to include mandatory findings (Table 5). Leiomyomas are typically multiple and are classified according to their location as submucosal, intramural and subserosal (Figs. 4 and 5). To standardize the nomenclature in patients with abnormal bleeding, regardless of their number and size, they can be further subdivided into 8 categories according to the FIGO classification (PALM-COEIN) ( Table 5) [38,39].
Size and MR features, e.g. signs of necrosis or degeneration as well as enhancement patterns, must be described (Table 5) [1].
Unusual variants, such as diffuse leiomyomatosis characterized by symmetrical enlargement of the uterus due to innumerable small ill-defined leiomyomas, lipoleiomyomas, and parasitic leiomyomas should be mentioned in the report [1,2]. Inclusion of any differential diagnosis is mandatory in every report. These include myometrial contractions, adenomyosis (Fig. 6), and, most importantly, uterine malignancy or adnexal fibromas.
Depending on the planned procedure other imaging information may be required (Table 5).

MRI for treatment planning and therapy monitoring
While (as corroborated by our survey) currently laparoscopic or open surgery is still the mainstay of operative intervention, minimal invasive interventions, such as Uterine Leiomyoma/  Although not mandatory for pre-UFE evaluation using high-quality ultrasonography, MRI may change diagnosis and treatment plan in about 20% of patients with symptomatic leiomyomas [4][5][6]. Patients with leiomyomas and coexisting adenomyosis are potential candidates for uterine artery embolization. The coexistence of adenomyosis should be reported, as it may require changes in the embolization protocol and might have a negative impact on long term results [4,[40][41][42]. Further, the size, number and location of leiomyomas are important for choosing treatment options for the individual patient. Though not contraindicated for UFE, it is valuable to visualize pedunculated leiomyomas for pre-treatment planning [4,43,44]. Cervical leiomyomas may not have a beneficial outcome after UFE [45]. Finally, it is crucial for the outcome of UFE that leiomyomas are vascularized, since a treatment response cannot be expected in a degenerated leiomyoma. MR-angiography (MRA) identifies major arteries supplying the leiomyomas. Usually the uterine arteries are the dominant suppliers, but blood supply by ovarian arteries may be a concern [4]. In these cases, the ovarian arteries are visualized as enlarged winding vessels on the MR angiogram (Fig. 2), while they are not clearly visualized under normal circumstances [46].
Post UFE problems may be identified by MRI. Persistent vaginal discharge may indicate a necrotic remnant or, rarely, expulsion and transvaginal sloughing of a necrotic leiomyoma may occur [4,49].
MRgFUS induces focal heating and coagulative necrosis within a leiomyoma under real-time MRI monitoring using thermal maps [50,51]. Ideally, leiomyomas should be limited in number and size (<10cm in largest diameter), show homogenously low-signal on T2W and no calcifications. Pre-treatment MRI in prone position allows targeting of a safe pathway during MRgFUS. Abdominal scars or pedunculated leiomyomas are relative contraindications, and intestine along the acoustic pathway is an absolute contraindication. Post treatment predictor of success is the non-perfused volume of the leiomyoma. Treatment typically achieves around 50-55% non-perfused volume, based on post contrast imaging after the completion of treatment [52].

Conclusion
MRI offers an outstanding and comprehensive map of the size, site and distribution of leiomyomas. A standardised imaging protocol and method of reporting ensures that the salient features are recognised and recorded. All radiologists involved in the assessment of leiomyomas need to know the varied imaging features found on MR and their significance. It should go without saying that a good working relationship between the radiologist and the gynaecologist is essential in maintaining the quality of an imaging service. Such a relationship means images can be discussed and clinical feedback given on outcomes.
Our questionnaire has shown a reasonable homogeneity of practice across institutions. It has allowed the ESUR working party to draw together a consensus of best imaging practice and formulate these guidelines.
Statistics and biometry No complex statistical methods were necessary for this paper.
Informed consent Written informed consent was not required for this study because no patient sensitive data were used Ethical approval Institutional Review Board approval was not required because the results are based on a survey only.

Methodology
• prospective • observational • multicentre study Open Access This article is distributed under the terms of the Creative Comm ons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Fig. 6 Differences between leiomyomas and adenomyoma/ adenomyosis. (a) Sagittal T2W image shows a poorly defined border, oval-shaped, low-signal mass (white arrow) with hyperintense T2 foci embedded in the lesion (black arrow), consistent with an adenomyoma. (b) Axial oblique T2W image demonstrates an ill-defined thickening of the junctional zone with hyperintense T2 foci, consistent with adenomyosis (white arrow). Nearby there is a T2 hypointense lesion, with a well-defined margin in keeping with an intramural leiomyoma (black arrow). Notice the thin T2 hyperintense rim surrounding the lesion, indicating a pseudocapsule of edema secondary to some degree of venous or lymphatic obstruction, typical for leiomyomas (arrowhead). Coexistence of adenomyosis and leiomyomas is not rare