{"paper_id":"ea6d01de-b785-4bc7-a793-98c4f722c5a3","body_text":"Lorusso et al. Insights Imaging          (2021) 12:105  \nhttps://doi.org/10.1186/s13244-021-01054-x\nEDUCATIONAL REVIEW\nMagnetic resonance imaging for deep \ninfiltrating endometriosis: current concepts, \nimaging technique and key findings\nFilomenamila Lorusso1, Marco Scioscia2, Dino Rubini1, Amato Antonio Stabile Ianora1, Doriana Scardigno3, \nCarla Leuci1, Michele De Ceglie1, Angela Sardaro4, Nicola Lucarelli1 and Arnaldo Scardapane1*  \nAbstract \nEndometriosis is an estrogen-dependent chronic disease affecting about 10% of reproductive-age women with \nsymptoms like pelvic pain and infertility. Pathologically, it is defined by the presence of endometrial tissue outside the \nuterine cavity responsible for a chronic inflammatory process. For decades the diagnosis of endometriosis was based \non surgical exploration and biopsy of pelvic lesions. However, laparoscopy is not a risk-free procedure with possible \nfalse negative diagnosis due to an underestimate of retroperitoneal structures such as ureters and nerves. For these \nreasons nowadays, the diagnosis of endometriosis is based on a noninvasive approach where clinical history, response \nto therapy and imaging play a fundamental role. Trans-vaginal ultrasound and magnetic resonance imaging are suit-\nable for recognizing most of endometriotic lesions; nevertheless, their accuracy is strictly determined by operators’ \nexperience and imaging technique. This review paper aims to make radiologists aware of the diagnostic possibilities \nof pelvic MRI and familial with the MR acquisition protocols and image interpretation for women with endometriosis.\nKeywords: Endometriosis, Deep infiltrating endometriosis, Magnetic resonance imaging, Imaging protocol\n© The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which \npermits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the \noriginal author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or \nother third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line \nto the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory \nregulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this \nlicence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.\nKey points\n• Diagnostic laparoscopy is considered the gold stand -\nard for endometriosis, but it is invasive with possible \nfalse negative results.\n• Nowadays there is a paradigm shift from surgical to \nnon-invasivediagnosis based on symptoms, response \nto therapy and imaging.\n• MRI is highly accurate for the diagnosis of Deep Infil-\ntrating Endometriosis.\n• The diagnostic results of MRI depend on an accurate \nimaging technique and on the comprehension of spe-\ncific MR-findings.\nBackground\nEndometriosis, particularly deep pelvic infiltrating endo -\nmetriosis (DIE), is a clinical issue affecting premenopau -\nsal women who may experience severe pelvic pain and \ninfertility [1]. These symptoms are mainly associated \nwith the growth of endometrial tissue outside the uterine \ncavity, with consequent chronic inflammatory reactions \nand fibromuscular hyperplasia affecting the pelvic peri -\ntoneum and the pelvic wall and organs [2]. The disease \naffects approximately 10% of women of reproductive age \nand is diagnosed in approximately 20%–50% of infertile \nwomen and nearly 90% of infertile women with chronic \npelvic pain [3]. Accurate early diagnosis of DIE is cru -\ncial to provide women with early tailored treatments and \navoid inappropriate surgery [4]. Nevertheless, although \nmany diagnostic techniques have been used, early diag -\nnosis of DIE remains a major challenge [5].\nLaparoscopic exploration is considered the diagnostic \ngolden standard, because it allows for direct visualization \nOpen Access\nInsights into Imaging\n*Correspondence:  arnaldo.scardapane@uniba.it; arnaldo.scardapane@gmail.\ncom\n1University of Bari Medical School - Interdisciplinary Department \nof Medicine, Section of Diagnostic Imaging, Piazza Giulio Cesare, 11, \n70124 Bari, Italy\nFull list of author information is available at the end of the article\n\nPage 2 of 12Lorusso et al. Insights Imaging          (2021) 12:105 \nof lesions; however, it is not a risk-free surgery and may \nunderestimate retroperitoneal structures, such as nerves \nand ureters, with possible false negative procedures [6, \n7]. Because a poor correlation has been demonstrated \nbetween the symptoms and severity of lesions, some \nauthors suggest a paradigmatic shift to a more clinical \ndiagnostic approach based on the combination of symp -\ntoms, imaging findings and response to empiric treat -\nment, even before any surgical confirmation [8]. In this \nscenario, transvaginal ultrasound, magnetic resonance \nimaging (MRI) and in some cases computed tomography \n(CT) play a fundamental role [9, 10]. MRI, the imaging \ntechnique with the highest overall accuracy for assess -\ning the extent of DIE, has high specificity for endometri -\notic foci, owing to its inherent soft-tissue resolution [11, \n12]. Nevertheless, to achieve the expected accuracy, the \nexamination itself and the image interpretation should \nbe tailored to each woman’s specific issues. This review \nprovides radiologists with information on how to obtain \ngood quality MRI images, interpret and report them \ncorrectly.\nManagement of deep endometriosis: current concepts\nEndometriosis is a complex and heterogeneous dis -\nease that may manifest with three clinical patterns with \nincreasing severity. Superficial peritoneal lesions are \ncharacterized by superficial implants of the pelvic perito -\nneum; ovarian endometriomas (OMA) are hemorrhagic \ncysts arising from ectopic endometrial tissue growing \nwithin the ovaries and less frequently outside the ovaries \n[13, 14]; and DIE leads to the most severe clinical pattern \nand is characterized by ectopic endometrial tissue pen -\netrating deeper than 5 mm under the peritoneal surface, \nthus leading to local inflammation and consequently to \nfibrosis and muscular hyperplasia [7, 15, 16]. DIE is usu -\nally found as a multifocal disease simultaneously involv -\ning multiple pelvic sites such as the Douglas pouch, the \nutero-sacral ligaments (USL), pelvic nerves, the rectum, \nthe bladder, and the ureters [17, 18] (Table 1). The patho-\nphysiology of such lesions is widely unknown; the type \nof lesion may vary over the life course, with no evidence \nsupporting an ordered progression of endometriotic \nlesions [19]. Similarly, the heterogeneity of symptoms is \nhigh. Women with endometriosis may experience dys -\nmenorrhea, dyspareunia, dysuria, constipation, chronic \npelvic pain and infertility; however, a clear characteri -\nzation of the pain types and topologies of implants is \nlacking [20]. Consequently, some women with minimal \ndisease may report severe pelvic pain and infertility, \nwhereas others with diffuse pelvic lesions can be almost \nasymptomatic [6]. The heterogeneity of the disease and \nthe uncertainties about its pathogenesis make its diagno -\nsis challenging. For decades, laparoscopic visualization \nwith histologic verification of lesions was considered \nthe golden standard for diagnosis [21]. However, lapa -\nroscopy, even if diagnostic, is not a risk-free surgery. \nFurthermore, deep infiltrating endometriosis may not \nbe clearly seen in some cases during diagnosis laparos -\ncopy [22]. In fact, the diagnosis of adenomyosis and deep \nendometriosis involving retroperitoneal structures, par -\nticularly ureters and nerve roots, is extremely challeng -\ning, particularly when performed by non-experienced \ngynecologists. False negative procedures may result, \nthus significantly delaying the start of appropriate man -\nagement and potentially leading to major complications \n[7, 8]. Symptoms and clinical findings of endometriosis \ncan result in clinical diagnoses that may be strongly sup -\nported by imaging techniques even without histological \nconfirmation. Transvaginal ultrasound, because of its \nnon-invasiveness, dynamicity, ease of use, availability, \ncost-effectiveness and reproducibility, is currently con -\nsidered by many endometriosis experts as the best first-\nline method for assessment of DIE [4, 5, 23]. A systematic \nsonographic approach as defined by International Deep \nEndometriosis Analysis (IDEA) consensus was shown \nto improve detection rates of pelvic lesions [24]. MRI is \nalso considered a highly accurate imaging modality in \nthe evaluation of DIE, particularly when involvement of \nthe rectum, ureters and nerve roots is suspected; these \nareas are highly important for both the patient and the \nsurgeon and may be poorly visualized even with laparos -\ncopy [25, 26]. MRI is also increasingly used to assess the \nanatomic response to medical or surgical treatment and \nto differentiate endometriosis from adenomyosis; the lat -\nter is a specific and heterogeneous disease contributing, \nindependently of endometriosis to symptoms, defined \nas the invasion of endometrial tissue into the myome -\ntrium occurring in different forms (diffuse, focal, cystic \nor superficial). Adenomyosis may exist on its own but \nin about 30% of cases it is associated with DIE [7, 27]. \nTable 1 Most frequent pelvic localization of DIE\nReferences no. [12, 18, 37, 46, 53]\nSite Frequency (%)\nPouch of Douglas/retrocervical 55–60\nUterosacral ligaments 32–57\nRecto-vaginal septum 20–48\nBowel (overall)\nRectum\nRecto-sigmoid junction\nSigmoid\nCecum/appendix\nSmall bowel\n16–35\n30–40\n25–30\n15–20\n5\n5\nBladder 5–8\nSacral nerves 3–5\n\nPage 3 of 12\nLorusso et al. Insights Imaging          (2021) 12:105 \n \nThe shift toward clinical and imaging-based diagnosis \nshortens the time between the first consultation and the \nfinal diagnosis [28], but the use of non-invasive methods \nrequires a rigorous approach to ensure meaningful and \nconsistent results [8, 29].\nMRI acquisition protocol: dos, don’ts and maybes\nMRI is a widely used technique for the diagnosis of DIE; \nhowever, an international consensus on the best imaging \nprotocol is lacking. Recent guidelines published by the \nEuropean Society of Urogenital Radiology describe con -\nsensus suggestions from a conference among nine imag -\ning centers in Europe and one in Japan [29]; however, \nthe indications and imaging protocols may vary among \ninstitutions according to local expertise. In general, when \nsuspicion of endometriosis exists, MRI should be used \nfirst to provide an adequate anatomic representation of \nthe entire pelvis and its organs and second to ensure the \nrecognition of DIE, according to the contrast between \nnormal pelvic fatty tissue and endometriotic lesions or on \nthe detection of hemorrhagic cysts and foci (Table 2).\nDOS\nMRI for endometriosis should be performed with a 1.5 T \nor 3  T scanner and high-resolution phased array coils \n(with 8–16 channels), whereas low-magnetic field or \nopen-MRI lacks sufficient image quality to image DIE. \nHigh-resolution, thin section (3 mm) TSE-T2w sequences \nin the sagittal, axial and coronal planes are crucial to \nevaluate DIE, whereas TSE T1w (with and without fat \nsaturation) should always be obtained to depict adnexal \nhemorrhagic lesions such as OMA. Oblique planes may \nbe highly useful to visualize specific anatomical struc -\ntures such as utero-sacral ligaments (Fig.  1) [30]. A sub -\nstantial improvement in image quality may be obtained \nby using rectal cleansing and anti-peristaltic agents such \nas butyl-scopolamine or glucagon which can also be \nhelpful in the evaluation of adenomyosis. Some authors \nsuggest a more reliable effect of such agents when intra -\nvenous rather than intramuscular injection is used; \nhowever, intramuscular administration ensures longer \nanti-peristaltic results, in line with an average imaging \nduration of 20–25 min [31]. The pelvis should be imaged \nregardless of the phase of the menstrual cycle, in patients \nwith a moderately full bladder [29].\nDON’Ts\nBecause the recognition of DIE is based on the contrast \nbetween the high signal intensity of fatty tissue and low \nsignal intensity of endometriotic nodules, fat-saturated \nT2w images should not be used [32, 33]. Among T1w fat-\nsaturated techniques, STIR sequences should be avoided. \nTable 2 Standard MRI protocol for endometriosis in our center\n* Performed with rectal distension\nSequence Plane Voxel mm \n(AP-RL-\nthickness)\nFOV (mm) NEX TE\nTSE T2 Axial/Obl \naxial\n0.9–0.9–3 280–350 2 100\nTSE T2 Sagittal 0.9–0.9–3 180–250 2 100\nTSE T2 Coronal 0.8–0.8–3 280–300 2 100\nTSE T1 Axial 0.9–0.9–3 280–350 1 Shortest\nTHRIVE Axial 0.75–0.75–3 280–350 3 Shortest\nTHRIVE Sagittal 0.75–0.75–3 280–350 3 Shortest\nOptional sequences\nCE-THRIVE Axial/Sagittal 0.75–0.75–3 280–350 3 Shortest\nBTFE* Axial/sagittal 1.5–1.5–4 280–350 1 Shortest\nSSFSE T2* Axial/sagittal 1–1–4 280–350 1 100\nFig. 1 Mild thickening of the right USL in a woman with DIE. a–b T2w sagittal images. c T2w axial image obtained along the red plane shown in b. \nd T2w oblique axial image obtained along the blue plane shown in b. The USL produces a better depiction of the sagittal and oblique axial plane \n(arrows)\n\nPage 4 of 12Lorusso et al. Insights Imaging          (2021) 12:105 \nThese sequences, which are based on the T1 relaxation \ntime, yield a non-specific saturation, which may sup -\npress nonfatty tissues with similar T1 values, such as the \nblood when methemoglobin is present, thus leading to \na difficult differential diagnosis between mature cystic \nteratomas and endometriomas (Fig.  2) [34, 35]. Selective \nsaturation of fatty tissue can be obtained with spectral \nsaturation (SPAIR or SPIR) used in 2D SPIR T1 images \nor in 3D interpolated sequences, such as THRIVE or \nDIXON.\nMAYBES\nThe use of intravenous contrast media is widely debated \nin the literature. Deep endometriosis is recognized by a \nlow signal intensity tissue with small hyperintense foci in \nT2w images, which may also show distortion of the pelvic \nanatomy associated with adhesions. Therefore, contrast-\nenhanced (CE) images appear to be useless in the diagno-\nsis of DIE. However, for specific indications, the injection \nof Gd-based contrast agent may be advisable. CE-images \nare mandatory in cases of complex adnexal hemorrhagic \ncysts showing mural thickening or other potentially \nmalignant features in T2w images. Similarly, the use of \ncontrast agents may aid in differentiating endometrio -\nmas from luteal ovarian cysts or tubo-ovarian abscesses \n[36, 37]. In our center, we have found that combining MR \ncolonography with CE THRIVE images may enable the \ndiagnosis of colorectal involvement by less experienced \nradiologists thanks to an easier recognition of thickened \nwall and for the possibility to distinguish enhancing nod -\nules from endoluminal fecal material or air [38]. Post-\ncontrast MR urography should be used when the ureteral \ninvolvement is suspected to define the degree of urinary \ntract dilation and the precise site of infiltration [39].\nNo consensus exists in the literature regarding the use -\nfulness of vaginal and rectal opacification for the diagno -\nsis of DIE; some authors find them extremely useful and \nhave proposed the use double contrast barium enema \nor cross-sectional colonography with either CT or MRI \nFig. 2 OMA mimicking a mature cystic teratoma in a STIR sequence. (a) TSE T2w axial image, (b) TSE T1w axial image, (c) STIR axial image and (d) \nTHRIVE axial image. The loss of T1w high signal intensity in the STIR image is not specific to fat (c), because endometriomas and fatty tissue may \nhave similar T1 relaxation times. In the THRIVE sequence, on the basis of a spectral saturation of fat, the endometrioma remains hyperintense (d)\nFig. 3 Midsagittal T2w image of female pelvis with main anatomic \nlandmarks to be considered in the evaluation of DIE\n\nPage 5 of 12\nLorusso et al. Insights Imaging          (2021) 12:105 \n \n[40–42], whereas others have reported no diagnostic \nimprovement from these procedures [43]. In our center, \nwe use rectal distension in patients showing an endome -\ntriotic lesion infiltrating the rectum in standard TSE T2w \nimages to quantify the stenosis, which according to our \nexperience is predictive of the need for bowel resection \n[25]. However, several alternative methods based on T2w \nimages without rectal distension have been described to \npredict the need for bowel resection [44, 45].\nMR anatomic landmarks\nThe imaging evaluation of endometriosis should be \nguided by the statistical frequency of involvement of \nthe pelvic anatomy [11] and be consistently accurate. \nTo achieve a uniform evaluation of women with sus -\npected endometriosis, the IDEA group in 2016 and the \nsociety of Abdominal Radiology have proposed a con -\nsensus lexicon for reporting US and MRI, respectively \n[24, 46]. In both experiences, it is suggested to report \nfindings by pelvic compartments (anterior, middle and \nposterior) and using consistent anatomic landmarks \n(Fig. 1, 3).\nAnterior compartment is the space limited anteriorly \nby the pubic symphysis and posteriorly by the uterus \nand contains the urinary bladder, the vesico-uterine \nfold and the round ligaments. The middle compartment \ncontains the uterus and the ovaries, while the posterior \ncompartment can be divided into the recto-uterine, \nFig. 4 Bilateral OMA. TSE T2 coronal (a) and axial (b) images and TSE T1w axial image (c). Bilateral endometriomas; the left-sided endometrioma \nshows a stratified aspect in b (shading sign). Of note, the ovaries are prolapsed in the pouch of Douglas, touching each other at the midline (kissing \novary sign)\nFig. 5 Large OMA with irregular mural vegetation and dark spots. TSE T2w sagittal image (a), TSE T2w axial image (b), TSE T1w axial image (c), axial \nTHRIVE image (d), contrast-enhanced axial THRIVE image (e) and ADC map (f). A large right multiloculated OMA is shown with an irregular mural \nvegetation (arrow) and small dark spots (arrowhead). The vegetation shows no significant contrast enhancement (e) but restricted diffusion (f) and \nshould be considered suspected for malignancy. A smaller left-sided OMA with shading sign is also visible (*)\n\nPage 6 of 12Lorusso et al. Insights Imaging          (2021) 12:105 \nrecto-cervical spaces and the recto-vaginal septum and \ncontains the serosal surface of the uterus, the pouch of \nDouglas, the torus uterinus, the USL well as the rectum \nand the sigmoid colon.\nMRI findings\nEndometriosis is a multifocal disease that may involve \nmultiple pelvic structures with possible extra-pelvic \nextension. OMA, superficial peritoneal lesions, and DIE \nhave been reported in surgical series studies to affect \nthe ovaries in 65–80%, 45–50%, and 63–70% of women, \nrespectively [12, 47, 48]. DIE is usually more frequent \nin the posterior pelvic compartment (95% of cases) \nincluding the torus uterinus, the recto-vaginal septum, \nUSL, pouch of Douglas and anterior wall of the rectum \nthan the anterior pelvic compartment (including the \nFig. 6 Retro-uterine DIE nodule. TSE T1w image and TSE T2w image. DIE nodules are characterized by intermediate signal intensity in T1w images \n(* in a) and low signal intensity in T2w images (* in b), with high intensity foci in both sequences (arrows). Adenomyosis is also shown with a similar \nMR aspect within the anterior wall of the uterus (arrowheads)\nFig. 7 Severe DIE with ureteral involvement. TSE T2w axial image (a), TSE T2w sagittal image (b) and CE MR urography (c). Large DIE nodule of the \npouch of Douglas extending to the right USL (*) and the anterior rectal wall (arrowhead). MR urography shows the involvement of the right pelvic \nureter (arrow), with consequent moderate hydronephrosis\n\nPage 7 of 12\nLorusso et al. Insights Imaging          (2021) 12:105 \n \nvesico-uterine pouch and bladder; 16% of cases). Both \ncompartments may be involved in approximately 10% of \ncases, whereas ureter and nerve lesions are seen in 5% of \npatients [47, 49].\nOMA may manifest as solitary or multiple thick-walled \ncysts showing homogeneous high signal intensity in T1w \nand fat-saturated T1w images regardless of the intensity \nin T2w images. According to the age at bleeding onset, \nOMA may be either hyperintense or hypointense in T2w \nimages or may show a typical stratified appearance (shad-\ning sign), as a result of cyclic bleeding with blood prod -\nucts accumulating over the course of months (Fig. 4) [37]. \nFig. 8 DIE with rectal infiltration. TSE T2w axial (a) and sagittal (b) images. A large DIE nodule (*) infiltrates the anterior wall of the rectum. The \nnodule has a mushroom-cap shape with a bright peripheral rim (arrowhead) corresponding to a normal mucosa layer\nFig. 9 DIE with cecal infiltration. TSE T2w axial (a) and coronal (b) images, SSFSE T2 MR-colonography coronal (c) and sagittal (d) images. A large DIE \nnodule with a mushroom shape (arrowheads) infiltrates the cecum which has a pelvic position in the pouch of Douglas\n\nPage 8 of 12Lorusso et al. Insights Imaging          (2021) 12:105 \nIn some cases, dark spots (low-intensity, well-defined \nimages in T2w sequences) may be visible within cysts \n(Fig. 5) [50]. Irregular mural thickenings or mural vegeta-\ntions should be studied after the intravenous injection of \ncontrast agents and DWI sequences to exclude malignant \ntransformation (Fig. 5).\nDIE can manifest as pelvic nodules or plaque-like \nlesions and adhesions [51, 52]. Nodules and plaque-like \nlesions are composed of endometrial glands and stroma \nsurrounded by a thick fibro-muscular and inflammatory \nreaction, and usually have an irregular, spiculated shape \nand a signal intensity similar to that of pelvic muscles, \nwith intermediate signal intensity in T1w sequences and \nlow signal intensity in T2w images. Small hyperintense \nfoci corresponding to endometrial glands are almost \nalways recognized within the endometriotic nodules \nin both T1w and T2w images (Fig.  6). The most com -\nmon site of DIE nodules is the posterior pelvic compart -\nment, where all anatomic structures bordering the pouch \nof Douglas can be involved (the posterior border of the \ncervix, the torus uterinus, the uterosacral ligaments, \nthe vaginal wall, the anterior wall of the rectum and the \nrecto-sigmoid junction; Fig. 7a, b) [53].\nDiagnosis of bowel involvement is based on the pres -\nence of a nodular or plaque-like endometriotic bowel \nwall thickening and loss of the fat tissue plane between \nthe intestinal loop and the uterus or other adjacent \norgans. The most frequent sites of bowel endometriosis \nFig. 10 DIE with vesical infiltration. TSE T2w sagittal (a, b) and coronal (c) images. The vesico-uterine fold is occupied by a DIE nodule (arrowhead) \nanteriorly tethering the uterine body (curved arrow)\nFig. 11 DIE with right ureter infiltration. TSE T2w axial (a, b) and sagittal (c) images. A right para-uterine DIE nodule (arrowheads) infiltrates the distal \ntract of the right ureter, which is dilated (arrows)\n\nPage 9 of 12\nLorusso et al. Insights Imaging          (2021) 12:105 \n \nare the rectum and the sigmoid colon, while the involve -\nment of the cecum or the ileum can be found in about \n5% cases (Table  1). The diagnosis may be facilitated by \nthe presence of ancillary findings such as a “mushroom \ncap” sign (Figs. 8, 9) [54]. This sign can be visible in any of \nthe plane of the space and represents the endometriotic \nnodule growing into a mushroom-like shape in the bowel \nwall, covered by a high intensity signal rim representing \nthe normal mucosa and submucosal layer (Figs. 7, 8, 9).\nEndometriotic nodules of the anterior or lateral pel -\nvic compartment are less frequently observed and usu -\nally involve the urinary system, particularly the vesical \ndome for nodules of the vesico-uterine fold (Fig.  10) \nand the ureters for lesions extending in the para-vesical \nspace (Fig. 11). Axial and sagittal TSE T2w images are the \nmost sensitive in identifying ureteral nodules; however, \nFig. 12 Adhesive obliteration of the pouch of Douglas. TSE T2w axial (a) and sagittal (b) images. The anterior rectal wall is tethered (arrow) to the \nposterior surface of the uterus, where a DIE nodule is seen (arrowhead)\nFig. 13 DIE of the pouch of Douglas. TSE T2w axial image. Teardrop \ndeformation of the rectum for DIE adhesions (arrowheads) is seen\nFig. 14 Adhesive endometriosis of the vesico-uterine pouch. TSE T2w (a) and THRIVE (b) sagittal image. Small hyperintense spot-like images of the \nvesico-uterine pouch are seen. An endometriotic plaque was found through laparoscopy\n\nPage 10 of 12Lorusso et al. Insights Imaging          (2021) 12:105 \nin these cases, the examination should be completed \nwith post-contrast MR urography to demonstrate even \nmild urinary dilatation and the exact position of ureteral \ninvolvement (Fig. 7c).\nIn many cases, MRI may depict pelvic changes consist -\nent with the presence of adhesions, which indirectly sug -\ngest DIE. In general, adhesions are suspected when fatty \ninterfaces between adjacent structures are not clearly \nvisible in any orthogonal planes. The most reliable find -\ning to diagnose endometriotic adhesions is tethering \nand angulation of normal pelvic structures and bowel \nloops (Fig.  12). Adhesion between the anterior wall of \nthe rectum and the posterior surface of the uterus, with \na consequent “teardrop” deformation of the rectum and \nretroversion of the uterine body, is frequently seen in pel-\nvic MRI and is specific for DIE (Fig. 13). Similarly, ovaries \nmay prolapse in the Douglas pouch and create adhesions \nbetween each other and the uterine wall on the midline, \nthus producing a so-called kissing ovary sign, which is a \ncommon finding in DIE of the posterior pelvis (Fig.  4). \nDouglas obliteration should be suspected when nodules \nextend from the retro-cervical space to the anterior wall \nof the rectum or when adhesions are seen at this level \n(Figs. 7, 8, 12). In contrast, if small bowel loops are seen \nbetween the uterus and the rectum, the obliteration of \nthe pouch of Douglas can be ruled out [53].\nAnterior pelvic adhesions usually occur between the \nuterus and the bladder for plaque-like or linear implants \nin the vesico-uterine pouch, which may be visible as small \nspots with high signal intensity in the sagittal T1w fat-\nsaturated images (Fig.  14) but can nonetheless be easily \nmissed by pelvic MRI, whereas endo-vaginal US, owing \nto its ability to show an absence of sliding of the uterus \nalong the bladder surface, is by far more sensitive.\nNeural endometriosis is a rare condition character -\nized by perimenstrual radicular pain with no evidence \nof any alteration of the lumbar spine. The most affected \nnerves are the sacral plexus (57% of cases) and the sci -\natic nerve (39% of cases) [49, 55]. MRI is the method \nof choice for the diagnosis of neural endometriosis, \nbecause transvaginal ultrasound cannot depict this \nanatomic area. The diagnosis relies on the recognition \nof endometriotic nodules along pelvic nerves and on \nindirect signs such as denervation muscular atrophy of \nthe affected site (Fig. 15).\nAdenomyosis in 30% of cases is associated with DIE; \nthe presence of ill-defined nodules with hyperintense foci \nwithin uterine wall in T1w and T2w images or a thicken -\ning of the junctional zone > 12 mm, is the most common \nfindings of this specific condition (Fig.  6); however, the \ndescription of adenomyosis is beyond the purpose of this \npaper and is detailed elsewhere [27, 56].\nConclusion\nBecause of its extreme clinical heterogeneity, pel -\nvic endometriosis remains challenging to diagnose. \nCurrent evidence demonstrates that the disease \nshould be diagnosed non-invasively by combining the \nFig. 15 Neural endometriosis. TSE T2w axial image (a), THRIVE axial image (b) and TSE T2w coronal images (c, d). A DIE nodule of the ischiatic \nforamen (arrowhead) surrounds the ischiatic nerve (arrows). Note the atrophy of the right piriformis muscle (black star) and normal left piriformis \nmuscle (white star)\n\nPage 11 of 12\nLorusso et al. Insights Imaging          (2021) 12:105 \n \ninformation from patient history, clinical examina -\ntion, imaging and response to medical treatment [7 ]. \nBecause the diagnostic accuracy of MRI may differ \ndepending on radiologist experience, this review arti -\ncle aims to help radiologists obtain meaningful images \nwith a tailored MR-acquisition protocol and recognize \na wide range of pelvic changes that may result from \nendometriosis.\nAbbreviations\nBTFE: Balanced turbo field echo; CE: Contrast enhanced; CT: Computed \ntomography; DIE: Deep pelvic infiltrating endometriosis; ESUR: European \nsociety of urogenital radiology; MRI: Magnetic resonance imaging; NE: Neural \nendometriosis; OMA: Ovarian endometrioma; RVS: Recto-vaginal septum; \nSSFSE: Single shot fast spin echo; SUP: Superficial peritoneal lesions; THRIVE: T1 \nhigh-resolution volume; TSE: Turbo spin echo; TVUS: Transvaginal ultrasound; \nUSL: Utero-sacral ligaments.\nAuthors’ contributions\nFL, AS, MS contributed to conceptualization and drafting. NL, CL, AASI helped \nin drafting and revising. DS, DR, MDC, AS helped in image collection. All \nauthors read and approved the final manuscript.\nFunding\nNot applicable.\nAvailability of data and materials\nThe datasets used and/or analyses during the current study are available from \nthe corresponding author on reasonable request.\nDeclarations\nEthics approval and consent to participate\nNot applicable.\nConsent for publication\nNot applicable.\nCompeting interests\nThe authors declare that they have no competing interests.\nAuthor details\n1University of Bari Medical School - Interdisciplinary Department of Medicine, \nSection of Diagnostic Imaging, Piazza Giulio Cesare, 11, 70124 Bari, Italy. 2 Unit \nof Gynecology, Mater Dei Hospital, Bari, Italy. 3 Unit of Gynecology and Obstet-\nrics, Di Venere Hospital, Bari, Italy. 4University of Bari Medical School - Interdisci-\nplinary Department of Medicine, Section of Radiation Therapy, Bari, Italy. \nReceived: 23 May 2021   Accepted: 5 July 2021\nReferences\n 1. Stepniewska A, Pomini P , Scioscia M, Mereu L, Ruffo G, Minelli L (2010) \nFertility and clinical outcome after bowel resection in infertile women \nwith endometriosis. Reprod Biomed Online 20:602–609\n 2. Bulun SE (2009) Endometriosis. N Engl J Med 360:268–279\n 3. Giudice LC, Kao LC (2004) Endometriosis. Lancet 364:1789–1799\n 4. Bazot M, Lafont C, Rouzier R, Roseau G, Thomassin-Naggara I, Darai E \n(2009) Diagnostic accuracy of physical examination, transvaginal sonog-\nraphy, rectal endoscopic sonography, and magnetic resonance imaging \nto diagnose deep infiltrating endometriosis. Fertil Steril 92:1825–1833\n 5. Noventa M, Scioscia M, Schincariol M et al (2019) Imaging modali-\nties for diagnosis of deep pelvic endometriosis: comparison between \ntrans-vaginal Sonography, rectal endoscopy Sonography and magnetic \nresonance imaging. A head-to-head meta-analysis. Diagnostics (Basel) \n9:225\n 6. Zondervan KT, Becker CM, Missmer SA (2020) Endometriosis. N Engl J \nMed 382:1244–1256\n 7. Chapron C, Marcellin L, Borghese B, Santulli P (2019) Rethinking mecha-\nnisms, diagnosis and management of endometriosis. Nat Rev Endocrinol \n15:666–682\n 8. Agarwal SK, Chapron C, Giudice LC et al (2019) Clinical diagnosis of endo-\nmetriosis: a call to action. Am J Obstet Gynecol 220:354 e351-354 e312\n 9. Scardapane A, Lorusso F, Scioscia M, Ferrante A, Stabile Ianora AA, \nAngelelli G (2014) Standard high-resolution pelvic MRI vs. low-resolution \npelvic MRI in the evaluation of deep infiltrating endometriosis. Eur Radiol \n24:2590–2596\n 10. Stabile Ianora AA, Moschetta M, Lorusso F et al (2013) Rectosigmoid \nendometriosis: comparison between CT water enema and video laparos-\ncopy. Clin Radiol 68:895–901\n 11. Chamie LP , Blasbalg R, Pereira RM, Warmbrand G, Serafini PC (2011) \nFindings of pelvic endometriosis at transvaginal US, MR imaging, and \nlaparoscopy. Radiographics 31:E77-100\n 12. Scardapane A, Lorusso F, Bettocchi S et al (2013) Deep pelvic endome-\ntriosis: accuracy of pelvic MRI completed by MR colonography. Radiol \nMed 118:323–338\n 13. Athey PA, Diment DD (1989) The spectrum of sonographic findings in \nendometriomas. J Ultrasound Med 8:487–491\n 14. Carbognin G, Guarise A, Minelli L et al (2004) Pelvic endometriosis: US \nand MRI features. Abdom Imaging 29:609–618\n 15. Koninckx PR, Meuleman C, Demeyere S, Lesaffre E, Cornillie FJ (1991) \nSuggestive evidence that pelvic endometriosis is a progressive disease, \nwhereas deeply infiltrating endometriosis is associated with pelvic pain. \nFertil Steril 55:759–765\n 16. Laux-Biehlmann A, d’Hooghe T, Zollner TM (2015) Menstruation pulls the \ntrigger for inflammation and pain in endometriosis. Trends Pharmacol Sci \n36:270–276\n 17. Vercellini P , Chapron C, Fedele L, Frontino G, Zaina B, Crosignani PG (2003) \nEvidence for asymmetric distribution of sciatic nerve endometriosis. \nObstet Gynecol 102:383–387\n 18. Chapron C, Fauconnier A, Vieira M et al (2003) Anatomical distribution of \ndeeply infiltrating endometriosis: surgical implications and proposition \nfor a classification. Hum Reprod 18:157–161\n 19. Laufer MR, Goitein L, Bush M, Cramer DW, Emans SJ (1997) Prevalence of \nendometriosis in adolescent girls with chronic pelvic pain not respond-\ning to conventional therapy. J Pediatr Adolesc Gynecol 10:199–202\n 20. Schliep KC, Mumford SL, Peterson CM et al (2015) Pain typology and \nincident endometriosis. Hum Reprod 30:2427–2438\n 21. Dunselman GA, Vermeulen N, Becker C et al (2014) ESHRE guideline: \nmanagement of women with endometriosis. Hum Reprod 29:400–412\n 22. Goncalves MO, Siufi Neto J, Andres MP , Siufi D, de Mattos LA, Abrao MS \n(2021) Systematic evaluation of endometriosis by transvaginal ultrasound \ncan accurately replace diagnostic laparoscopy, mainly for deep and ovar-\nian endometriosis. Hum Reprod 36:1492–1500\n 23. Guerriero S, Saba L, Pascual MA et al (2018) Transvaginal ultrasound vs \nmagnetic resonance imaging for diagnosing deep infiltrating endome-\ntriosis: systematic review and meta-analysis. Ultrasound Obstet Gynecol \n51:586–595\n 24. Guerriero S, Condous G, van den Bosch T et al (2016) Systematic \napproach to sonographic evaluation of the pelvis in women with sus-\npected endometriosis, including terms, definitions and measurements: a \nconsensus opinion from the International Deep Endometriosis Analysis \n(IDEA) group. Ultrasound Obstet Gynecol 48:318–332\n 25. Scardapane A, Lorusso F, Francavilla M et al (2017) Magnetic resonance \ncolonography may predict the need for bowel resection in colorectal \nendometriosis. Biomed Res Int 2017:5981217\n 26. Chiantera V, Petrillo M, Abesadze E et al (2018) Laparoscopic neuronavi-\ngation for deep lateral pelvic endometriosis: clinical and surgical implica-\ntions. J Minim Invasive Gynecol 25:1217–1223\n 27. Bazot M, Darai E (2018) Role of transvaginal sonography and magnetic \nresonance imaging in the diagnosis of uterine adenomyosis. Fertil Steril \n109:389–397\n\nPage 12 of 12Lorusso et al. Insights Imaging          (2021) 12:105 \n 28. Soliman AM, Fuldeore M, Snabes MC (2017) Factors associated with \ntime to endometriosis diagnosis in the United States. J Womens Health \n(Larchmt) 26:788–797\n 29. Bazot M, Bharwani N, Huchon C et al (2017) European society of urogeni-\ntal radiology (ESUR) guidelines: MR imaging of pelvic endometriosis. Eur \nRadiol 27:2765–2775\n 30. Bazot M, Gasner A, Ballester M, Darai E (2011) Value of thin-section \noblique axial T2-weighted magnetic resonance images to assess uterosa-\ncral ligament endometriosis. Hum Reprod 26:346–353\n 31. Gutzeit A, Binkert CA, Koh DM et al (2012) Evaluation of the anti-peri-\nstaltic effect of glucagon and hyoscine on the small bowel: compari-\nson of intravenous and intramuscular drug administration. Eur Radiol \n22:1186–1194\n 32. Gollub MJ, Arya S, Beets-Tan RG et al (2018) Use of magnetic resonance \nimaging in rectal cancer patients: Society of Abdominal Radiology (SAR) \nrectal cancer disease-focused panel (DFP) recommendations 2017. \nAbdom Radiol (NY) 43:2893–2902\n 33. Brown G, Daniels IR, Richardson C, Revell P , Peppercorn D, Bourne M \n(2005) Techniques and trouble-shooting in high spatial resolution thin \nslice MRI for rectal cancer. Br J Radiol 78:245–251\n 34. Krinsky G, Rofsky NM, Weinreb JC (1996) Nonspecificity of short inversion \ntime inversion recovery (STIR) as a technique of fat suppression: pitfalls in \nimage interpretation. AJR Am J Roentgenol 166:523–526\n 35. Siegelman ES, Oliver ER (2012) MR imaging of endometriosis: ten imaging \npearls. Radiographics 32:1675–1691\n 36. Outwater E, Schiebler ML, Owen RS, Schnall MD (1993) Characterization \nof hemorrhagic adnexal lesions with MR imaging: blinded reader study. \nRadiology 186:489–494\n 37. Dias JL, Veloso Gomes F, Lucas R, Cunha TM (2015) The shading sign: is it \nexclusive of endometriomas? Abdom Imaging 40:2566–2572\n 38. Scardapane A, Bettocchi S, Lorusso F et al (2011) Diagnosis of colorectal \nendometriosis: contribution of contrast enhanced MR-colonography. Eur \nRadiol 21:1553–1563\n 39. Bielen D, Tomassetti C, Van Schoubroeck D et al (2020) IDEAL study: mag-\nnetic resonance imaging for suspected deep endometriosis assessment \nprior to laparoscopy is as reliable as radiological imaging as a comple-\nment to transvaginal ultrasonography. Ultrasound Obstet Gynecol \n56:255–266\n 40. Faccioli N, Foti G, Manfredi R et al (2010) Evaluation of colonic involve-\nment in endometriosis: double-contrast barium enema vs. magnetic \nresonance imaging. Abdom Imaging 35:414–421\n 41. Biscaldi E, Ferrero S, Fulcheri E, Ragni N, Remorgida V, Rollandi GA (2007) \nMultislice CT enteroclysis in the diagnosis of bowel endometriosis. Eur \nRadiol 17:211–219\n 42. Kikuchi I, Kuwatsuru R, Yamazaki K, Kumakiri J, Aoki Y, Takeda S (2014) Eval-\nuation of the usefulness of the MRI jelly method for diagnosing complete \ncul-de-sac obliteration. Biomed Res Int 2014:437962\n 43. Bazot M, Gasner A, Lafont C, Ballester M, Darai E (2011) Deep pelvic \nendometriosis: limited additional diagnostic value of postcontrast in \ncomparison with conventional MR images. Eur J Radiol 80:e331-339\n 44. Rousset P , Buisson G, Lega JC et al (2021) Rectal endometriosis: predictive \nMRI signs for segmental bowel resection. Eur Radiol 31:884–894\n 45. Brusic A, Esler S, Churilov L et al (2019) Deep infiltrating endometriosis: \nCan magnetic resonance imaging anticipate the need for colorectal \nsurgeon intervention? Eur J Radiol 121:108717\n 46. Jha P , Sakala M, Chamie LP et al (2020) Endometriosis MRI lexicon: con-\nsensus statement from the society of abdominal radiology endometriosis \ndisease-focused panel. Abdom Radiol (NY) 45:1552–1568\n 47. Bazot M, Darai E, Hourani R et al (2004) Deep pelvic endometriosis: MR \nimaging for diagnosis and prediction of extension of disease. Radiology \n232:379–389\n 48. Vimercati A, Achilarre MT, Scardapane A et al (2012) Accuracy of trans-\nvaginal sonography and contrast-enhanced magnetic resonance-colo-\nnography for the presurgical staging of deep infiltrating endometriosis. \nUltrasound Obstet Gynecol 40:592–603\n 49. Siquara De Sousa AC, Capek S, Amrami KK, Spinner RJ (2015) Neural \ninvolvement in endometriosis: review of anatomic distribution and \nmechanisms. Clin Anat 28:1029–1038\n 50. Corwin MT, Gerscovich EO, Lamba R, Wilson M, McGahan JP (2014) Dif-\nferentiation of ovarian endometriomas from hemorrhagic cysts at MR \nimaging: utility of the T2 dark spot sign. Radiology 271:126–132\n 51. Kataoka ML, Togashi K, Yamaoka T et al (2005) Posterior cul-de-sac oblit-\neration associated with endometriosis: MR imaging evaluation. Radiology \n234:815–823\n 52. Siegelman ES, Outwater E, Wang T, Mitchell DG (1994) Solid pelvic masses \ncaused by endometriosis: MR imaging features. AJR Am J Roentgenol \n163:357–361\n 53. Loubeyre P , Petignat P , Jacob S, Egger JF, Dubuisson JB, Wenger JM (2009) \nAnatomic distribution of posterior deeply infiltrating endometriosis \non MRI after vaginal and rectal gel opacification. AJR Am J Roentgenol \n192:1625–1631\n 54. Yoon JH, Choi D, Jang KT et al (2010) Deep rectosigmoid endometriosis: \n“mushroom cap” sign on T2-weighted MR imaging. Abdom Imaging \n35:726–731\n 55. Ceccaroni M, Clarizia R, Cosma S, Pesci A, Pontrelli G, Minelli L (2011) \nCyclic sciatica in a patient with deep monolateral endometriosis infiltrat-\ning the right sciatic nerve. J Spinal Disord Tech 24:474–478\n 56. Takeuchi M, Matsuzaki K (2011) Adenomyosis: usual and unusual imaging \nmanifestations, pitfalls, and problem-solving MR imaging techniques. \nRadiographics 31:99–115\nPublisher’s Note\nSpringer Nature remains neutral with regard to jurisdictional claims in pub-\nlished maps and institutional affiliations.","source_license":"CC0","license_restricted":false}