{"paper_id":"a2f45c97-ee4b-4878-b045-93085fe08601","body_text":"GLOBAL RADIOLOGY (J FRENCHNER, SECTION EDITOR)\nImaging of Endometriosis: The Role of Ultrasound and Magnetic\nResonance\nValentina Testini 1,2 • Laura Eusebi 3 • Gianluca Grechi 4 • Francesco Bartelli 3 •\nGiuseppe Guglielmi 1,2,5\nAccepted: 7 February 2022 / Published online: 18 February 2022\n/C211 The Author(s) 2022\nAbstract Endometriosis is a chronic gynecological disease\ncharacterized by the growth of functional ectopic\nendometrial glands and stroma outside the uterus. It causes\npelvic pain, dysmenorrhea, dyspareunia, or infertility.\nDiagnosis requires a combination of clinical history, non-\ninvasive and invasive techniques. The aim of the present\nreview was to evaluate the contribution of imaging tech-\nniques, mainly transvaginal sonography and magnetic\nresonance imaging to diagnose different locations and for\nthe most appropriate treatment planning. Endometriosis\nrequires a multidisciplinary teamwork to manage these\npatients clinically and surgically.\nKeywords Endometriosis /C1Deep endometriosis /C1\nEndometrioma /C1Transvaginal ultrasonography /C1MRI\nIntroduction\nEndometriosis is a common gynecological inflammatory\ncondition that is defined as functional ectopic endometrial\nglands and stroma outside the uterus. This disease affects\nwomen of reproductive age, with a prevalence of approx-\nimately 10% [ 1]. Patients can be asymptomatic or present\nwith chronic pelvic pain and/or infertility.\nThe phenotypes of endometriotic lesions can be divided\ninto three groups: pelvic endometriosis, ovarian\nendometriomas (OMA) and deep infiltrating endometriosis\n(DIE) [ 2]. In particular, pelvic endometriosis is defined as\nthe presence of any endometrial tissue within the pelvic\ncavity, including the peritoneum, within any of the pelvic\norgans and inside the pouch of Douglas (POD). Ovarian\nendometriosis, an endometrioma, is defined as an ovarian\ncyst, of different size, lined by endometrial tissue. DIE is\ndefined as endometriotic tissue that penetrates the\nretroperitoneal space for a distance of 5 mm or more and\nmay be present in multiple locations, involving anterior or\nposterior pelvic compartments, or both [ 3]. Posterior DIE, a\nmultifocal disease that may affect a variety of anatomical\nsites, represents the most common type of DIE [ 4]. The\nmost typical sites of DIE include uterosacral ligaments\n(USL), rectovaginal septum (RVS), vaginal wall, POD and\nbowel, predominantly below the rectosigmoid junction.\nAnterior DIE corresponds to disease involving the anterior\npouch or bladder and is much less common. DIE is a source\nof pain and infertility [ 5].\nA frequently association with endometriosis is repre-\nsented by adenomyosis, a disease characterized by infil-\ntration of endometrial tissue into the myometrium [ 6].\nThis article is part of the Topical collection on Global Radiology .\n& Giuseppe Guglielmi\ngiuseppe.guglielmi@unifg.it\n1 Radiology Unit, ‘‘Dimiccoli’’ Hospital, Viale Ippocrate 15,\n70051 Barletta, Italy\n2 Department of Clinical and Experimental Medicine, Foggia\nUniversity School of Medicine, Viale L. Pinto 1, 71121\nFoggia, Italy\n3 Radiology Unit, ‘‘Carlo Urbani’’ Hospital, Jesi, Italy\n4 Obstetrics and Gynecology Unit, ‘‘Carlo Urbani’’ Hospital,\nJesi, Italy\n5 Radiology Unit, Hospital ‘‘Casa Sollievo Della Sofferenza’’,\nViale Cappuccini 2, 71013 San Giovanni Rotondo, Foggia,\nItaly\n123\nCurr Radiol Rep (2022) 10:21–39 (0123456789().,-volV)(0123456789().,-volV)\nhttps://doi.org/10.1007/s40134-022-00393-x\n\nEtiopathogenesis\nAlthough the pathogenesis of endometriosis has not been\nfully elucidated, it is commonly thought that endometriosis\noccurs when endometrial tissue contained within menstrual\nfluid flows retrogradely through the fallopian tubes and\nimplants at an ectopic site within the pelvic cavity [ 7]. In\nthis process, menses transports viable endometrial frag-\nments through the fallopian tubes to the peritoneal cavity,\nwhere they are able to implant, develop and sometimes\ninvade other tissues of the pelvis [ 8]. In favor of this\nhypothesis is that all known factors that increase menstrual\nflow are also risk factors for endometriosis, including early\nage at menarche, heavy and long periods as well as short\nmenstrual cycles [ 9]. The anatomical distribution of\nendometriotic lesions can also be explained by the\nhypothesis of retrograde menstruation as endometriotic\nlesions tend to have an asymmetrical distribution, which\ncould be explained by the effect of gravity on menstrual\nflow, the abdominopelvic anatomy and the peritoneal\nclockwise flow of menses [ 10]. However, this theory does\nnot explain the fact that although retrograde menstruation\nis seen in up to 90% of women, only 10% of women\ndevelop endometriosis [ 3]. Moreover retrograde menstru-\nation does not explain the mechanism of endometrial tissue\ngrafting onto the peritoneum. It is therefore evident that a\nvariety of environmental, immunological and hormonal\nfactors contribute to the onset of endometriosis, with\nmechanisms not yet known [ 11].\nGenetic factors play an important role in the genesis of\nendometriosis, with an up to six times greater risk of\ndeveloping the disease for first degree relatives of patients\nwith endometriosis [ 12]. Despite this clear inheritance, the\nidentification of the genetic factors that drive the disease is\nstill incomplete.\nDiagnosis\nThe first step in diagnosing DE is to establish the patient’s\nclinical history with particular emphasis on symptoms\n(dysmenorrhea, dyspareunia, dysuria, dyschezia, and\nchronic pelvic pain) as well as, age, height, weight, ethnic\norigin, gravidity, parity, previous surgery for endometrio-\nsis, family history of endometriosis, previous non-surgical\ntreatment for endometriosis, and infertility. No symptom is\nspecific to endometriosis [ 13].\nHowever, several authors have underlined the poor\nrelationship between symptoms exhibited by patients and\nthe severity of the lesions rendering clinical diagnosis\ndifficult. Moreover, it is thought that up to 50% of women\ncould have asymptomatic endometriosis [ 14, 15].\nThe second step is based on physical examination\nincluding a systematic analysis of the posterior vaginal\nfornix with a speculum to look for retraction and dark\nnodules. Digital examinations should be performed of the\nvagina to assess the characteristics of the uterus and\nadnexa, of the vesicouterine pouch to detect bladder\ninvasion, and of the retrocervical area to detect infiltration\nof the torus uterinus, uterosacral ligaments (USLs), pouch\nof Douglas (POD), vagina, and rectovaginal septum (RVS)\n[13]. Rectal digital examination can help in assessing the\ninvolvement of the rectum, parametrium and visceral pel-\nvic fascia. In the particular setting of DE, few data are\navailable to evaluate the accuracy of physical examination.\nOne retrospective study found that routine clinical exami-\nnation detected DE in only 36% of 140 women with DE,\nand the authors suggest the accuracy of physical exami-\nnation improves during menstruation [ 16]. To detect rec-\ntosigmoid and retrocervical DE without differentiating\nbetween the different specific DE locations, Abrao et al.\nreported that digital vaginal examination had a sensitivity\nof 72% and 68%, a specificity of 54% and 46%, respec-\ntively [ 17].\nLaboratory tests are limited in the diagnosis of\nendometriosis (CA-125 has a detection rate of only 54% in\npatients with severe endometriosis 5 and is neither sensi-\ntive nor specific for the diagnosis) [ 18].\nThe gold standard for diagnosis of endometriosis is\nbased on laparoscopy or surgery with histological verifi-\ncation of endometrial glands and/or stroma.\nImaging is needed to diagnose endometriosis and to plan\ntreatment. The techniques used are transvaginal ultrasound\n(TVUS) and magnetic resonance imaging (MRI), the latter\nshould be considered as a second-line technique after\nultrasound [ 1].\nTreatment\nTreatment of endometriosis is complicated and involves\nconservative approaches combined with medical therapies\nor surgery. Imaging is crucial to guide the type of treat-\nment. The American Society for Reproductive Medicine\nPractice Committee states that ‘ ‘endometriosis should be\nviewed as a chronic disease that requires a lifelong man-\nagement plan with the goal of maximizing the use of\nmedical treatment and avoiding repeat surgical proce-\ndures’ ’ [ 18]. Current treatment is essentially surgical,\nmedical, or a combination of both approaches. To this end,\nmany patients are stratified for medical treatment with or\nwithout surgical treatment based on symptom severity or\nimaging results and desire to have children, with medical\ntherapy typically including non-steroidal anti-inflammatory\ndrugs, oral contraceptives, androgens, progestogens and\n22 Curr Radiol Rep (2022) 10:21–39\n123\n\ngonadotropin-releasing hormone (GnRH) and/or surgery\n[19].\nLaparoscopy is an effective surgical approach with the\ngoal of excision of visible endometriosis in a hemostatic\nfashion. Radical surgery is reserved for those patients with\nsevere symptoms where there is no desired fertility\npotential and especially when other forms of treatment\nhave failed. Total abdominal hysterectomy and bilateral\nsalpingo-oophorectomy are performed along with resection\nof any endometriotic lesions as completely as possible [ 4].\nUltrasound\nTransvaginal ultrasound (TVUS) is typically the initial\nimaging evaluation performed in patients with pelvic pain\nand infertility or when there is clinical suspicion for\nendometriosis. This examination is widely available with\nlow relative cost and the sensitivity and specificity for\ndetection of ovarian endometriomas and lesions in the\nrectal wall are high [ 18].\nAll patients should be examined systematically and\ncarefully using an endocavitary sonography, with a\nmicroconvex array probe inserted transvaginally or tran-\nsrectally. Both techniques are optimal approaches for\nexamining uterus (including the different uterine zones:\ncervix, endometrium, junctional zone, and myometrium),\nadnexa, paracolpium, parametrium, vesicocervical, vesi-\ncovaginal, and rectovaginal spaces as well as urinary\nbladder, ureters, and rectum [ 20].\nLimitations of sonographic technique include anterior\ncompartment detection of endometriosis (bladder and\nvesicouterine pouch detection) and detection within the\nmiddle compartment (torus uterinus and round ligaments)\n[18].\nThe pelvic localization of endometriosis can be descri-\nbed according to three compartments (central, anterior, and\nposterior) according to functional and clinical relevance\n[21]. The anterior compartment includes the insertion site\nof the ureters, the bladder, the vesicouterine pouch, and the\nvesicovaginal pouch. The middle compartment contains the\nuterine body, fallopian tube, and uterine ligaments. The\nposterior compartment contains the uterosacral ligaments,\nrectovaginal septum, anterior rectal wall, and sigmoid\ncolon [ 21].\nRecently, the IDEA (International Deep Endometriosis\nAnalysis group) published a consensus report 18 on the\nappropriate terms, definitions, and measurements that may\nbe used to describe the sonographic features of the different\nphenotypes of endometriosis. A standardized pelvic TVS\napproach is proposed by this IDEA report consisting in four\nstep pelvic evaluation [ 22]:\n(1) Evaluation of uterus and ovaries:\n(a) Evaluation of uterus: 2D–3D sonographic signs\nof adenomyosis\n(b) Evaluation of the adnexa: presence or absence\nof endometrioma or tubal pathology\n(2) Evaluation of TVS organ mobility (adhesions):\nadnexa and uterine mobility site-specific tenderness\n(3) Pouch of Douglas (POD) assessment using real-time\nultrasound-based ‘ ‘sliding sign’ ’\n(4) Assessment for DIE nodules in anterior lateral and\nposterior compartments\nAnterior Compartment\nThe anterior compartment is comprised of the urinary\nbladder, the vesicouterine pouch, round ligaments and\nureters. Involvement of the urinary tract occurs in\napproximately 1–2% of patients with endometriosis and\ninvolves bladder in 85% of these cases [ 23]. Ureteric\ninvolvement is found in 4% of patients with rectovaginal\nendometriosis [ 24]. The prevalence of round ligament\nendometriosis is estimated between 4.3% and 13.8% [ 25].\nBladder endometriosis is considered only in case of\ninfiltration of the bladder wall and not in case of adhesions\nor superficial peritoneal implants on the bladder serosa.\nBefore TVS scan, patients are asked not to empty\ncompletely the bladder, because the slightly filled bladder\npermits to better evaluate the structure of the walls.\nOn ultrasound, bladder endometriosis appears as\nhypoechoic lesion, either containing cystic lesions or not,\nwith regular/irregular margins of the bladder wall, bulging\ntoward the lumen, involving the serosa, muscularis (most\ncommon), or (sub)mucosa of the bladder [ 26].\nIn the assessment of the bladder DIE localization, the\nbladder wall can be divided into three zones: the trigonal\nzone and vesical base; the vesical dome (which lies supe-\nrior to the trigone and is intra-abdominal) and the anterior\nretroperitoneal bladder. Most frequently bladder\nendometriosis is located in the vesical dome on the pos-\nterior bladder wall close to the vesicouterine pouch [ 27].\nBladder adhesions of the vesicouterine pouch are eval-\nuated by the presence or absence of the ‘ ‘sliding sign’ ’\nbetween the uterus and the bladder [ 28].\nDuring examination, from a longitudinal section through\nthe cervix and moving the probe toward the lateral pelvic\nwall, it is possible to assess the distal part of the ureter\nadjacent to the bladder trigone, in order to evaluate the\npresence of stenosis and subsequent cephalad dilatation of\nthe pelvic ureters. This finding can suggest direct invasion\nCurr Radiol Rep (2022) 10:21–39 23\n123\n\nor compression of the ureter by endometriotic nodules,\novarian endometriomas, or adhesions [ 28].\nThe prevalence of ureteral endometriosis ranges from\n0.01 to 1% of all patients with the disease and most often\naffect the distal segment of the ureter [ 29]. There are two\ntypes of endometriosis involvement of ureters: (1) extrinsic\nthat represents 75–80% of the cases and is defined as the\npresence of endometrial tissue in the outer adventitia of the\nureter that occurs as a nodule encasing the ureter by\nextension from pelvic foci; (2) intrinsic that represents\n20–25% of cases and is defined as the presence of\nendometrial tissue in the mucosal and/or muscular layer of\nthe ureter. Imaging signs are nodule or mass occurring in\nthe ureter along its course, dilatation of the pelvic ureteral\ntract, or ureteropelvic hydronephrosis superior to the sus-\npected lesion [ 22]. Pelvic ureteral dilation can be easily\nseen by TVS as a tubular anechoic image, very similar to a\nblood vessel but with negative color/power Doppler signs.\nAn extrinsic compression, also without ureteral dilatation,\nis suspected in cases where a DIE lesion is located close to\nthe ureter. The observation of a possible ureteral involve-\nment requires transabdominal ultrasound to evaluate the\nrenal pelvis. In all women with DIE, a transabdominal scan\nof the kidney to search for ureteral stenosis is necessary\nbecause the prevalence of endometriotic lesions in the\nurinary tract may be underestimated and women with DIE\ninvolving the ureter may be asymptomatic [ 30].\nA review of the literature for bladder endometriosis\nreveals a reported mean US sensitivity of 55% and speci-\nficity of 93.5% [ 31].\nFor vesicouterine pouch endometriosis, two series\nreported US sensitivities and specificities of 16.7% and\n33% and 99% and 100% [ 32]. These discrepant results\ncould be partly explained by selection bias of inclusion\namong studies.\nCentral Compartment\nThe central compartment includes uterus and adnexa.\nUterus\nAdenomyosis is characterized by the migration and pro-\nliferation of endometrial glands and stroma from the basal\nlayer of endometrium into the myometrium. It is associated\nwith smooth muscle hyperplasia leading to an ultrasound\nimage of ill-defined lesions within the myometrium [ 22].\nRecently, the ultrasound features have been systemati-\ncally described by the international Morphological Uterus\nSonographic Assessment (MUSA) group [ 33].\nAccording to the MUSA consensus, adenomyosis should\nbe described as localized and diffuse.\nAdenomyosis is classified as diffuse, if the total\ninvolvement of myometrium exceeds 50% of the corpus\nuteri (when the findings are present in only one part of the\nmyometrium on one or more sites within the uterine wall),\nand localized (or focal) when less than 50% of myome-\ntrium is involved (one or more lesions) [ 33]. An adeno-\nmyoma is defined as a focal consolidation of endometrial\nglands and or endometrial stroma located within the\nmyometrium with additional compensatory hypertrophy of\nthe surrounding myometrium [ 27]. In rare cases it may\npresent as a large cyst (adenomyotic cyst or cystic adeno-\nmyoma, with largest diameter 2 mm and echogenic rim)\n[34].\nTo evaluate adenomyosis, the size of the lesions should\nbe measured (in particular the largest diameter of each\nfocal lesion or the myometrial wall thickness in cases of a\ndiffuse lesion), the involvement of the uterine layers and of\nthe extent of the disease, based on the estimated volume of\nthe uterine corpus affected by adenomyosis (mild \\ 25%,\nmoderate 25–50%, and severe [ 50%) [ 27, 35].\nAccording to several studies, there are different features\nassociated with adenomyosis visible on 2D transvaginal\nsonographic [ 33].\nOn ultrasound, an adenomyotic uterus appears with a\nglobular shape, enlarged dimensions, and uterine wall\nasymmetry. The myometrium typically appears inhomo-\ngeneous on gray-scale, characterized by the presence of an\nindistinctly defined area with either decreased or increased\nechogenicity with myometrial hypoechoic linear striations\n[36]. Round anechoic areas of 1-mm to 7-mm diameter,\nnamed myometrial cysts, also could be present within the\nmyometrium [ 36]. In cases of focal adenomyosis, the\nadenomyotic lesion appears as a heterogeneous and\nhypoechogenic area within the myometrium, usually with\nanechoic lacunae or cysts with ill-defined contours and fan-\nshaped shadowing. These hypoechogenic areas reflect\nmuscular hypertrophy of the myometrial tissue [ 27].\nIrregularities of the endometrial-myometrial junctional\nzone is another common ultrasound marker in the diagnosis\nof adenomyosis [ 27]. This endomyometrial interface is\nnormally visualized as hypoechoic tissue layer seen beyond\nthe endometrial basal layer. In women with adenomyosis,\nthe diffuse or focal hyperplasia and hypertrophy of myo-\ncytes determine whether diffuse or focal thickening of this\nzone is seen [ 37].\nA characteristic sign is the ‘ ‘question mark sign’ ’ defined\nwhen the corpus uterus was flexed backward, the fundus of\nuteri was facing the posterior pelvic compartment, and the\ncervix was directed frontally toward the urinary bladder\n[38]. Investigators found 93% specificity and 75% sensi-\ntivity of this sign in detecting adenomyosis [ 22].\nPower Doppler can be used to distinguish myometrial\ncysts from blood vessels and discriminate between\n24 Curr Radiol Rep (2022) 10:21–39\n123\n\nleiomyomas and focal adenomyosis. Uterine leiomyomas\nmanifest a circular flow along the myoma pseudocapsule,\nwhile localized adenomyosis and adenomyomas are char-\nacterized by diffusely spread vessels inside the lesions [ 39].\n2D ultrasound can yield equivocal result in the case of\nfocal adenomyosis especially if there are coexistent\nfibroids. A meta-analysis of 14 trials and 1985 participants\nreported the sensitivity and specificity of ultrasound in the\ndiagnosis of adenomyosis to be as high as 82.5 and 84.6%,\nrespectively, values in line with MRI values [ 40].\nThe use of 3D vaginal ultrasound for the diagnosis of\nadenomyotic pathology allows a more complete evaluation\nin the sagittal, transverse and coronal planes, evaluating the\nultrasound signs on the acquired 3D volume of the uterus\n[41].\nOvaries\nOvarian endometriomas occur when ectopic endometrial\ntissue in the ovary hemorrhages, forms a hematoma,\nenveloped by ovarian parenchyma (Fig. 1).\nAn ovarian endometrioma has different imaging\nappearances on US, with the classic appearance being a\ncyst unilocular or multilocular (less than five locules) with\nhomogeneous low-level echogenicity (ground glass\nechogenicity) of the cyst fluid, with increased posterior\nthrough transmission and no vascularization on color\nDoppler [ 22] (Fig. 2).\nAnother feature is the presence of peripheral echogenic\nfoci (thought to reflect cholesterol deposits) seen in up to\n36% of endometriomas. Endometriomas tend to be multi-\nlocular and bilateral (up to 50%) [ 42] (Fig. 3).\nHowever, endometriomas may have a variable appear-\nance because of the range of appearance of the internal\nblood products within them, which can cause fluid–fluid\nlevels, echogenic regions, and papillary projections. In\nthese cases, additional evaluation with MR imaging may be\nwarranted to better evaluate and to exclude malignancy\n[43]. There is evidence that ovarian endometriomas origi-\nnate from ovulatory events and it is likely that the number\nof endometriomas may increase with age, and multiple\nendometriomas in the same ovary may assume a multi-\nlocular morphology [ 44]. Guerriero and colleagues repor-\nted that ultrasound appearance of endometriomas differed\nbetween premenopausal and postmenopausal patients [ 44].\nThe endometriomas in the postmenopausal patients were\nless often unilocular cysts and less likely to exhibit ground\nglass echogenicity [ 27].\nThe primary differential diagnosis of an endometrioma\nis a hemorrhagic cyst. On US, a hemorrhagic cyst classi-\ncally has internal reticular strands with retractile clot [ 18].\nHowever, these features may not be seen, and instead,\nhomogeneous low-level echoes mimicking that of an\nendometrioma may be present. Hemorrhagic cysts are\nunlikely to have the peripheral echogenic foci occasionally\nseen in endometriomas, and they are less likely to be\nbilateral or multifocal. Sonographic follow-up demon-\nstrating resolution at 6–12 weeks is diagnostic of a hem-\norrhagic cyst [ 18].\nAnother differential diagnosis of an endometrioma is an\novarian epithelial neoplasm, which may contain low-level\ninternal homogeneous echoes similar to an endometrioma.\nThis imaging appearance was seen in up to 6% of ovarian\nserous cystadenomas in the study by Patel et al. [ 42] and in\nup to 20% of mucinous cystadenomas in the study by Van\nHolsbeke [ 45]. To better evaluate for the presence of\nmalignancy (cystadenocarcinomas) in these cases, careful\ninterrogation of the cyst should be performed to assess for\ninternal solid components, such as papillary projections,\nmural nodules, and thickened septations.\nDoppler helps avoid classifying malignancies as\nendometriomas, especially when evaluating a papillary\nFig. 1 Endometriotic cyst, with regular parenchyma at the periphery,\ncalled ‘ ‘crescent sign’ ’ characteristic of benign lesions. Absent\nintralesional vascularization\nFig. 2 Color Doppler of multiple endometriomas of both ovaries,\nwhich appear enlarged and sharpened (ovarian kissing)\nCurr Radiol Rep (2022) 10:21–39 25\n123\n\nprojection. Generally, these different ultrasound criteria\nproposed have a sensitivity ranging from 62 to 73%, a\nspecificity of 94–98% [ 46].\nMasses in postmenopausal women whose cystic con-\ntents have a ground glass appearance have a high risk of\nmalignancy. Borderline tumors and carcinomas arising\nfrom endometrioid cysts show a vascularized solid com-\nponent on ultrasound examination [ 47].\nTubes\nThe uterine tubes can be involved with endometriosis\neither with adhesions occluding the tube up to 6%) or by\nDIE foci affecting the tubal walls (up to 26% of the time)\n[18].\nIn case of endometriosis of the tube, we can observe a\ndilated tube with thick walls and incomplete septa with a\nfluid dense content similar to an endometrioma (hematos-\nalpinx) [48]. A ‘ ‘cog-wheel’ ’ appearance of the longitudinal\nfolds can be seen when the tube is imaged in cross-sec-\ntion. The presence of a hematosalpinx may be the only sign\non imaging of endometriosis in the pelvis [ 49].\nIn case of occlusion of the tube due to adhesion or DIE\nthat involved the distal part and the fimbriae, a hydros-\nalpinx is seen with the typical ‘ ‘beads-on-a-string’ ’ sign,\ndefined as hyperechoic mural nodules measuring approxi-\nmately 2–3 mm as seen on the cross-section of the fluid-\nfilled distended structure [ 48].\nThe differential diagnosis of a hematosalpinx includes\npelvic inflammatory disease (PID) or fallopian tube\nmalignancies. Pyosalpinx of PID can be differentiated\nclinically by the presence of extreme tenderness on\nexamination as well as the clinical signs of infection (fever,\nwhite count). On imaging, hyperemia surrounding the fal-\nlopian tube with fatty proliferation/edema in the adjacent\nfat suggests a pyosalpinx. Fallopian tube carcinoma\npresents sonographically with solid, vascular internal\nnodules within the fallopian tube and tends to occur in an\nolder demographic group [ 43].\nPosterior Compartment\nRecently, the ultrasound features of the deep infiltrating\nendometriosis nodules have been systematically defined by\nthe International Deep Endometriosis Analysis group [ 34].\nThe most common sites of the posterior compartment are\nposterior vaginal fornix/rectovaginal septum, uterosacral\nligaments, anterior rectum/anterior rectosigmoid junction,\nand sigmoid colon [ 10] (Fig. 4).\nDeep endometriosis on sonography is subtle and pre-\nsents as hypoechoic nodular or infiltrating regions. Occa-\nsionally, the infiltrative regions of DIE may have internal\nhyperechoic foci or complex internal cysts [ 50]. The dif-\nferential diagnosis for DIE includes peritoneal implants; in\nthese cases, to help differential diagnosis, an additional\nevaluation with MR is recommended [ 18]. Three-dimen-\nsional (3D) TVS has been also proposed in the evaluation\nof posterior locations of DIE without intestinal involve-\nment, improving the diagnostic accuracy of 2D ultra-\nsonography [ 51].\nIn the cases of DIE, it is necessary to describe of the\nanatomical localizations, the size and number of DIE\nnodules, the depth of infiltration of the nodules, and the\ndegree of stenosis of the bowel lumen which is important to\nplan the surgical procedures [ 22].\nUterosacral Ligaments\nThe uterosacral ligaments (USL) are usually not visible on\nultrasound. The uterosacral ligaments affected by deep\nFig. 3 Endometrioma with evidence of the characteristic double\nfluid–fluid level. Inside the formation, hyperechoic spots can be\nhighlighted, a symptom of hemosiderin accumulation\nFig. 4 Hypoechoic nodule of the rectosigmoid portion that obliterates\nthe Douglas. At this level, the ‘ ‘sliding sign’ ’ can be seen, the sign of\nthe sliding structures on each other which does not occur in the case\nof endometriosis\n26 Curr Radiol Rep (2022) 10:21–39\n123\n\ninfiltrating endometriosis can be seen in the longitudinal\nview of the uterus at the insertion on the posterior lateral\ncervix wall, as hypoechoic tissue, with regular/irregular\nmargins within the peritoneal fat surrounding the uter-\nosacral ligaments [ 13]. On the transverse cervical section,\nthese hypoechoic nodules appear on the posterior lateral\npart of the cervix and interrupt the hyperechoic external\ncervical fascia. Sometimes, the uterosacral ligaments\nappear thickened and hyperechoic, probably as the mor-\nphologic expression of fibrosis, due to the chronic process\nof inflammation [ 27].\nUSLs lesions may be isolated or may be part of a larger\nnodule extending into the vagina or into other surrounding\nstructures. In some cases, the DIE lesion involving the USL\nis located at the torus uterinus as a central thickening of the\nretrocervical area between USLs [ 52]. Two recent meta-\nanalyses of USL endometriosis have reported pooled sen-\nsitivities and specificities of 53–64% and 93–97%,\nrespectively [ 53].\nIn case of endometriotic lesions involving the uter-\nosacral ligaments, special attention must be paid to the\nparametrium. Parametria are examined lateral to the uterine\ncervix first on the sagittal planes moving the probe from the\nlateral sites where the parametrium is attached to the cer-\nvix, to the uterine vessels bifurcation, to the lateral pelvic\nwall, and then on the transverse planes moving the probe\nfrom the uterine isthmus to the external cervical os. The\nparametrial involvement is seen as an infiltrating hypoe-\nchogenic irregular tissue, and it can be medially delimi-\ntated from the cervical vascular plexuses using color or\npower Doppler [ 22].\nRectovaginal Septum\nInvolvement of the rectovaginal septum should be sus-\npected when an endometriotic nodule, which appears as\nhypoechoic solid nodule with smooth or irregular contours,\nthat replaces the normal hyperechoic aspect of this layer\nbetween the vagina and the rectum, seen in the rectovaginal\nspace below the horizontal plane passing through the lower\nborder of the posterior lip of the cervix (under the peri-\ntoneum) [ 34].\nIsolated rectovaginal septum nodule is rare, and it is\nusually an extension of posterior vaginal wall, anterior\nrectal wall, or both posterior vaginal wall and anterior\nrectal wall involvement. Hourglass-shaped or diabolo-like\nnodules can occur when endometriosis lesions from the\nposterior vaginal fornix extend to the anterior rectal wall\n[54].\nDIE of the RVS may extend into the rectum and/or in\nthe posterior vaginal fornix [ 22].\nMajor discrepancies exist between the pooled sensitiv-\nities and specificities provided by meta-analyses reporting\nvalues from 49% to 88% and 98% to 100%, respectively\n[34].\nVagina\nVaginal endometriosis is diagnosed when the posterior or a\nlateral vaginal fornix shows a nodular wall thickened\n([ 5 mm) (mean normal vaginal thickness ranges from 3 to\n5 mm), with or without round cystic anechoic areas, that\ndoes not get thinner with probe compression [ 13].\nThe nodule may be hypoechoic, homogeneous, or\ninhomogeneous with or without cystic areas and there may\nalso be some vascularization at power Doppler. More fre-\nquently, the lesions are localized in the posterior vaginal\nfornix [ 22].\nThe insertion of saline solution in the vagina (sono-\nvaginography) could improve the visualization of these\nlesions [ 55].\nIn a meta-analysis including ten studies, the pooled\nsensitivity and specificity of TVS was 57% and 99%,\nrespectively [ 56]. Among the various TVS techniques,\nSVG provided the highest sensitivity and specificity\nreaching 91% and 89%, respectively [ 56].\nBowel\nThe endometriosis affecting the bowel can appear as a\nthickening of the muscularis propria or as a hypoechoic\nnodule penetrating the intestinal wall with blurred margins,\nwith or without hypoechoic or hyperechoic foci, usually\nassociated with retraction and adhesion (the so-called\nIndian headdress sign), and few vessels at power Doppler\nevaluation [ 57].\nThe rectum and the rectosigmoid segment is the most\nfrequent site of bowel involvement accounting for 70–88%\nof cases of bowel involvement with endometriosis, fol-\nlowed by the sigmoid colon, rectum, ileum, appendix, and\ncecum [ 10]. Intestinal nodules located below the peri-\ntoneum of the POD (or the level of the insertion of the\nUSLs on the cervix in case the cul-de-sac is obliterated) are\nconsidered low rectal lesions, while the ones above this\nlevel are considered upper rectal or the rectosigmoid\njunction lesions. This virtual line should delineate the plane\nunder the peritoneum of the POD and correspond laterally\nto the parametria and medially to the RVS. The lowest\nlimit of the nodule on the bowel wall should be determined,\nbecause the lower rectal lesions are more difficult to\nremove surgically by shaving or segmental resection and\nhave higher complication rate [ 22].\nEndometriotic nodules of the rectum can be evaluated if\nnecessary also by transrectal examination as well with the\nsame transvaginal convex probe. This has the advantage of\nvisualizing better the vagina, the rectovaginal septum\nCurr Radiol Rep (2022) 10:21–39 27\n123\n\n(RVS), and the low rectal walls. Moreover, during the\ntransrectal or transvaginal examination, a fluid contrast\nmedium can be inserted in the vagina to visualize better the\nRVS (sonovaginography) [ 55]. It has been reported that\nadding water contrast in the rectum during transvaginal\nultrasonography (RWC-TVS) improves the diagnosis of\nrectal infiltration in women with rectovaginal\nendometriosis. RWC-TVS is performed by injecting saline\nsolution into the rectal lumen under ultrasonographic\ncontrol through a catheter [ 58].\nDuring the evaluation of posterior compartment, a\nnegative ‘ ‘sliding sign’ ’ between the rectosigmoid and\nuterus could indicate an obliteration of the pouch of Dou-\nglas (POD), frequently associated with severe DE. Using\nthis new technique, Reid et al. found sensitivity, specificity\nof 83.3%, 97.1% [ 59].\nMultifocal lesions are defined as the presence of deep\nlesions within 2-cm area of the main lesions or multiple\nendometriotic lesions affecting the same intestinal segment.\nMulticentric lesions are defined as a satellite deep nodule\nfound more than 2 cm from the main lesions or endometri-\notic lesions affecting several digestive segments [ 60].\nThe use of volume acquisition with 3D TVS permits a\nmore accurate measurement and evaluation of the DIE\nlesion in different planes.\nThe pooled sensitivity and specificity of TVS for rec-\ntosigmoid endometriosis are reported as 90% and 96%,\nrespectively, with similar results being provided by RES [56].\nGuerriero et al., in a one-paired study, suggested that 2D-TVS\nwas more sensitive but less specific than 3D TVS [ 61].\nMRI\nProtocols\nMR imaging of the pelvis is frequently performed for the\ndetection of endometriosis, either as the second-line\nimaging examination (after US) for the detection/confir-\nmation of endometriosis, in particular in deep infiltrating\nendometriosis, or as the initial examination in a patient for\nwhom there is a high clinical suspicion for endometriosis\nand not ultrasound confirm [ 1].\nMR imaging can be performed with either a 1.5-T or 3-T\nmagnet, using a high-resolution phased-array surface coil\nfor improved resolution. There is no consensus regarding\nwhether to perform the examination around the timing of\nthe patient’s menstrual cycle [ 13].\nPatients are positioned supine on the scanner, with\nabdominal strapping after phased coil array placement. Fasting\nbefore the examination for 4 h is typically recommended in\norder to empty the upper gastrointestinal tract; the use of\nantiperistaltic agents is recommended to reduce motion\nartifacts caused by intestinal peristalsis. However, the type of\nagent (oral agents, nonoral agents), dose, and route (intra-\nmuscular, subcutaneous, or IV) is debated [18].\nPatient preparation required a moderately filled bladder;\nthis is required to change the angle of uterine anteversion,\nleading to better detection of implants in the anterior\ncompartment. Moreover, a moderately filled bladder dis-\nplaces the bowel superiorly, by reducing the artifacts from\nbowel motion [ 62].\nFor the evaluation of deep endometriosis it was sug-\ngested to introduce intra-vaginal aqueous gel to distend the\nvaginal cavity and better explore the vaginal fornices and\nthe retrocervical area [ 62].\nIn the presence of symptoms that may be related to\nrectal involvement, gel may also be useful to distend the\nrectal\\sigmoid bowel wall.\nThe typical imaging protocol [ 21, 63] includes three T2\nturbo-spin-echo (TSE)—weighted sequences (T2W) in\ndifferent slice orientations (sagittal, coronal, and axial\nplanes), followed by three T1-weighted (T1W) sequences\nin an identical imaging plane (TR 500 ms, TE 14 ms)\nwithout fat suppression and fat-suppressed T1W before and\nafter intravenous injection of contrast media because the fat\nsuppression is useful for the detection of subtle foci of\nhemorrhage, which may be obscured on non–fat-saturated\nimages [ 64].\nUse of contrast-enhanced imaging is primarily required\nto identify solid enhancing nodules within endometriotic\ncysts when malignant transformation is suspected.\nDixon technique or conventional in- and out-of-phase\nT1-weighted images are useful for the differentiation of\nfat-containing lesions, such as dermoid cysts from\nendometriomas, both of which have high signal on non–fat-\nsaturated T1-weighted images [ 18].\nNo recommendation can be achieved for the use of DWI\nand SWI sequences.\nHalf-Fourier acquisition single shot turbo-spin-echo\n(HASTE) is recommended for the evaluation of uterine\nperistalsis because it enables multiphase and multislice\nimage acquisition producing kinematic images for the\nevaluation of pelvic adhesions [ 1]. During the peri-ovula-\ntory phase, uterine peristalsis is significantly reduced in\nsubjects with endometriosis when compared to normal\ncontrols that may be due to increased, sustained contrac-\ntions in endometriosis patients [ 65].\nMRI Evaluation\nAt MR imaging the signal intensity of endometriotic\nlesions is a function of the quantity and age of the hem-\norrhage on the one hand and the proportion of endometrial\ncells and stroma on the other [ 66].\n28 Curr Radiol Rep (2022) 10:21–39\n123\n\nThe lesions have a micronodular or microcystic\nappearance; however, cysts do not enlarge except in the\novary. Only pigmented lesions can be detected at non-\ncontrast-enhanced MR imaging because of the presence of\nhemorrhage [ 67]. At MR imaging these small implants\nmanifest as multiple round (cystic or nodular) lesions\nhomogeneously hyperintense on fat-suppressed T1-weigh-\nted images, due to old hemorrhagic content, regardless of\ntheir signal intensity on T2-weighted images [ 67].\nInvolvement of peritoneal reflections over the cul-de-sac\nand the uterus may also manifest on contrast-enhanced fat-\nsaturated T1-weighted images as diffuse peritoneal\nenhancement secondary to the inflammatory reaction\ninduced by endometrial implants. Over time a fibrotic\nreaction may occur, thus leading to adhesions formation\nbetween pelvic structures [ 67].\nT2W sequences are used for the evaluation of fibrotic\nlesions, notably those that involve the pelvic ligaments,\nretrocervical space, or prevesical recess [ 21].\nThey appear as speculated hypointense peritoneal\nstrands arranged in confluent angles. Posterior displace-\nment of uterus and ovaries, angulation of rectosigmoid\ncolon and bowel loops, elevation of the posterior vaginal\nfornix, loculated fluid collections, and a hydrosalpinx may\nbe indirect signs of adhesions [ 68].\nAt MR imaging both fibrous tissue and smooth muscle\nshow intermediate signal intensity on T1-weighted images\nand low-signal intensity on T2-weighted images. There-\nfore, on T2-weighted images, solid endometriotic lesions\nappear as hypointense nodular structures with irregular or\nstellate margins due to fibrous tissue and smooth muscle\nproliferation. In certain cases, deep endometriotic lesions\nmay also appear as irregular and hypointense soft-tissue\nnodular thickening on T2-weighted sequences, as it occurs\nwhen the disease involves the USL or the vaginal or rectal\nwall [ 66].\nIn some cases the glandular component can be pre-\ndominant, compared to the fibrous tissue and in this case\nthe MRI appearance will show high signal intensity of\nT2W images; in this case the use of contrast material can\nbe useful because this will show enhancement, thus dis-\ntinguishing it from intramural hemorrhage or necrosis [ 63].\nUsually, endometrial glands without hemorrhage are not\ndetectable on fat-suppressed T1-weighted images; so deep\nlesions may show homogeneous intermediate signal\nintensity on T1-weighted images. When red cell extrava-\nsation outside the glandular ducts into the surrounding\nstroma occurs, these small hemorrhages become visible as\nsmall hyperintense spots on fat-saturated T1-weighted\nimages. After the intravenous administration of gadolin-\nium, lesion enhancement may occur due to inflammatory\nreaction, glandular and fibrous tissue.\nOwing to the possibility to perform a complete assess-\nment of all pelvic compartments at one time, MRI repre-\nsents the best imaging technique for preoperative staging of\nendometriosis [ 66].\nAnterior Compartment\nEndometriotic lesions may affect the urinary tract in up to\n20% of cases.\nBladder\nBladder lesions appear as small masses of round or lobu-\nlated hypertrophic tissue covered by normal mucosa [ 67].\nBladder involvement is often multifocal, the trigone and\nthe dome being the most frequently affected sites [ 23].\nAccording to the degree of wall infiltration, bladder\ninvolvement may be classified as extrinsic or intrinsic. In\nextrinsic involvement, the most common form, implants\nare confined to the serosal surface or the surrounding\nconnective tissue; in intrinsic involvement lesions infiltrate\nthe muscular layer manifesting as mural masses [ 66].\nAt MR imaging bladder endometriosis may manifest as\nlocalized or diffuse wall thickening and signal intensity\nabnormalities [ 63]. The appearance is of low-signal\nintensity on T2-weighted and intermediate signal intensity\non T1-weighted images, with or without spots of high\nsignal intensity on T1-weighted images, representing\nhemorrhagic content. Sometimes, hyperintense foci on\nT2W images corresponding to the dilated endometrial\nglands may be found [ 66]. Implants minimally enhance\nafter injection of a gadolinium-based contrast material\n[64].\nThe maximum lesion diameter varies between 1 and\n5 cm. MRI reaches sensitivity up to 88%, specificity up to\n99% and diagnostic accuracy of about 98% for the diag-\nnosis of bladder endometriosis [ 23]. The differential\ndiagnosis of bladder endometriosis includes urachal rem-\nnant, epithelial tumors (bladder carcinoma) and mes-\nenchymal tumors (angiomas, leiomyoma) [ 23].\nUrethra\nUrethral endometriosis is uncommon; direct implantation\nof endometrial tissues during procedures is hypothesized to\nbe the etiologic mechanism [ 69]. In some cases,\nendometriosis may be mistaken for a urethral diverticulum\nand therefore a precise diagnosis is essential to facilitate\noptimal management. Usually, urethral involvement is\nobserved as a contiguous extension from the bladder and\nthe MRI findings are the same as those described for\nbladder endometriosis [ 64].\nCurr Radiol Rep (2022) 10:21–39 29\n123\n\nUreter\nUreteral endometriosis may be defined as any situation\nwhere endometriosis or surrounding associated fibrosis\ncauses compression or distortion of the normal ureteral\nanatomy, even when hydroureteronephrosis is not yet\npresent [ 70]. Ureteral endometriosis is quite uncommon,\nmost often unilateral, with a left predisposition; bilateral\ninvolvement is present in approximately 10–20% of cases\n[71]. Ureteral involvement is often associated with an\nipsilateral endometrioma or with a rectovaginal nodule\nlarger than 3 cm. The distal ureter, 3–4 cm above the\nvesicoureteral junction, is the most common ureteral seg-\nment involved [ 71].\nThe symptoms can vary according to the type of nodule\ninfiltration: indeed endometrial tissue can directly infiltrate\nthe muscularis propria, lamina propria, or ureteral lumen\ncausing symptoms that may be related to the pelvic\nendometriosis itself (dysmenorrhea, dyspareunia) or sec-\nondary to urinary tract involvement (flank pain, obstruc-\ntion, and in some cases decline of renal function) [ 64]. Also\nin the ureteral endometriosis there are two major patho-\nlogical types: extrinsic and intrinsic. Extrinsic\nendometriosis is the most common form (80%) of ureteral\ninvolvement and it represents endometrial glandular and\nstromal tissue within the submucosa and adventitia of the\nureter [ 21]. In addition, scar tissue or fibrosis without true\nendometriotic invasion of the ureter may also be classified\nas extrinsic disease. In contrast, intrinsic endometriosis\n(about 20%) involves the uroepithelial and muscular layer.\nAt MRI, ureteral endometriosis usually appears as\nhypointense solid nodules on T2W images with spiculated\nmargins, that envelop the ureter, causing dilatation of the\nureter upstream [ 72]. Extrinsic disease may be hypothe-\nsized when the interface of fat between the nodule and\nureter is no longer visible. MR urographic techniques can\nbe used to obtain three-dimensional reconstructed images\nfrom coronal volumetric excretory phase T1W data. The\ndifferential diagnosis of ureteral endometriosis includes\nureteral invasion by cervical cancer [ 64].\nVesicouterine Pouch\nThe vesicouterine pouch or anterior cul-de-sac is a com-\nmon site of endometriotic involvement [ 21]. These lesions\nare associated with anteflexion of the uterus and oblitera-\ntion of the anterior cul-de-sac due to extensive adhesions\nbetween the peritoneum of the bladder fold and the uterus.\nAt MR imaging deep endometriotic implants involving the\nanterior uterine serosa demonstrate infiltrative pattern with\nindistinct margins and show hypointense nodules on T2-\nweighted images, with small cystic areas that typically\nadhere to the anterior uterine surface, forming an obtuse\nangle with the vesical wall [ 72].\nCentral Compartment\nOvaries\nThe ovaries are the most common site of endometriosis\n(20–40% of cases) [ 72]. They may be affected in two ways:\n(i) the endometriomas or chocolate cysts that are associated\nwith ovarian enlargement, and which are caused by repe-\nated episodes of hemorrhage; and (ii) small nodular\nsuperficial implants which may cause paraovarian fibrous\nscarring and adhesions [ 64] (Fig. 5A–C).\nPeritoneal implants confined to the ovarian surface are\noften underdiagnosed at imaging due to their small size\n(\\ 5 mm) [ 66].\nEndometriomas are frequently multilocular and bilateral\n(50% of cases). In MRI an endometrioma appears as a\nhomogeneously hyperintense mass on T1W MR images; on\nT2W MR images, it appears as a low-signal intensity mass\nwith areas of high signal intensity [ 67] (Fig. 6A–F).\nEndometriomas have a wall with a variable appearance\n(from thin to thick and fibrotic) and they usually contain\ndark/brown semi-solid material that represents the degen-\nerated blood products (the so-called ‘‘chocolate cyst’’) [ 67]\n(Fig. 7A–E). Because the endometriomas contain blood\nproducts of different ages and concentrations, they may\nshow a variable appearance, in fact as free water in the cyst\nis resorbed, the iron concentration increases along with the\nviscosity of the contents of the cyst: this condition deter-\nmines the ‘‘shading effect’’. Shading is present when a\ncyst, hyperintense on a T1W image, shows a gradient from\nhypointense to hyperintense on a T2W image. Shading can\nrange from faint, dependent layering to complete signal\nvoid, according to the concentration of blood products [ 64].\nIt reflects the chronic nature of endometriomas and is the\nresult of cyclic bleeding occurring over time. Old blood\nproducts contain high iron and protein concentrations\nwhich determine a decrease in T2 relaxation time. There-\nfore, on T2-weighted images endometriomas will show a\ngradual loss of signal within the lesion with low-signal\nintensity till complete signal void in the declivous portion.\nThe most specific pathologic feature of endometrioma is\nthe thick fibrous capsule containing a cluster of hemosi-\nderin-laden macrophages due to repeated hemorrhage [ 66].\nIn certain cases, the ovaries may be joined together\nbehind the uterus in the pouch of Douglas due to adhesion\nformation between the adjacent peritoneal surfaces, a sign\ndescribed at US as B kissing ovaries ^ and suggestive of\nsevere pelvic endometriosis [ 73].\n30 Curr Radiol Rep (2022) 10:21–39\n123\n\nTogashi et al. found that an extremely sensitive sign for\nthe presence of an endometrioma was the presence of a cyst\nhyperintense on T1W images and the presence of shading\non T2W images [ 74]. Another criterion is to observe\nmultiple hyperintense cysts on T1W images (and T1W fat-\nsuppressed images) regardless of their signal intensity on\nT2W images (Fig. 8A–F).\nThe differential diagnosis of endometriomas includes\nlesions with high signal intensity on T1-weighted images:\ndermoids, mucinous cystic neoplasms, and hemorrhagic\nmasses. Fat-saturated T1-weighted sequences are helpful to\nrule out a fat-containing lesion (such as dermoids) and to\nconfirm the presence of blood [ 75]. Mucinous lesions may\nshow hyperintensity on T1-weighted images, but signal\nintensity is lower than that of blood.\nIt is possible to recognize and differentiate dermoids\nfrom endometriomas by the presence of chemical shift\nartifact and signal drop-out on the fat suppression image.\nThus, on a T1-weighted frequency selective fat saturation\nsequence, the mature cystic teratoma will be low in signal,\nwhereas an endometrioma will have high signal [ 64].\nHence, the most challenging differential diagnosis is\nwith other hemorrhagic masses. To differentiate\nendometriomas from functional hemorrhagic cysts is\nimportant in order to prevent unnecessary surgical inter-\nventions. Functional hemorrhagic cysts (i.e., hemorrhagic\nfollicular cysts and hemorrhagic corpus luteum cysts) are\nusually unilocular and unilateral, do not display shading on\nT2-weighted images, and mostly disappear on follow-up\nexaminations (generally in 4–6 weeks), while endometri-\nomas are usually multilocular and bilateral [ 68].\nThe role of DWI sequences in differentiating\nendometriomas from functional hemorrhagic ovarian cysts\nis still debated. Balaban et al. found significantly lower\nADC values in endometriomas compared with functional\nhemorrhagic ovarian cysts in all b values [ 76]. Large\nlesions with wall nodularity, thick septations and enhanc-\ning solid components may be suggestive of malignancy.\nFallopian Tubes\nEndometriotic involvement of the fallopian tubes is\nstrongly associated with infertility. Serosal or subserosal\nimplants involves the peritoneal surface of the fallopian\ntubes, where repeated hemorrhages lead to fibrosis and\nretraction of the tube with hydrosalpinx. Intraluminal\nimplants determine cyclic hemorrhage thus causing\nhematosalpinx [ 66].\nMoreover, an association has been described between\nthe endometriosis in the fallopian tubes and a predisposi-\ntion to endometrial malignancies such as clear cell carci-\nnoma and endometrioid carcinoma [ 67].\nAt MR imaging hematosalpinx appears as a tortuous\nenlarged tubular adnexal structure filled with hemorrhagic\nfluid. Endoluminal content shows high signal intensity on\nfat- suppressed T1-weighted images and intermediate sig-\nnal intensity, (40% of distended tubes in endometriosis\nFig. 5 An Example of multiple endometriomas of both ovaries,\nwhich appear enlarged and sharpened (ovarian kissing): A sagittal\nT1w fat sat imaging; B, C axial and sagittal T2w imaging\nCurr Radiol Rep (2022) 10:21–39 31\n123\n\nhave hyperintense contents) with or without internal fluid–\nfluid level, on T2-weighted images [ 77]. According to\nSiegelman the presence of T1-weighted hyperintensity\nwithin a dilated fallopian tube is suggestive of\nendometriosis [ 78].\nIn addition, it is atypical to see T2 shading within the\nlumen of the distended fallopian tube even when there is\nhigh signal on T1-weighted images [ 78]. T2 shading is not\nseen because of the fact that the endometriotic implants are\nmostly along the surface of the tube and not within the\nlumen of the tube, such that the chronic bleeding within the\nimplants leads to adhesions along the tubal surface, but not\nwithin the lumen. The differential diagnosis of a\nhematosalpinx on MR imaging includes PID or fallopian\ntube malignancies. Pyosalpinx of PID can be differentiated\nby the clinical signs of infection (fever, white count) [ 18].\nOn MR imaging, hyperemia surrounding the fallopian\ntube with stranding in the adjacent fat would suggest a\npyosalpinx. Fallopian tube carcinoma demonstrates solid,\nenhancing internal nodules within the fallopian tube and\ntends to occur in an older demographic group [ 43].\nUterus and Vagina\nThe most common localization of ectopic endometrial\ntissue within the uterus is the adenomyosis. Whereas it is\nimportant to remember that uterine involvement by\nendometriosis is usually subserosal, sometimes it is possi-\nble to find nodules of endometriosis in the serosal surface\nof the uterus [ 64].\nVaginal endometriosis is usually associated with\nimplants in other pelvic locations, mostly retrocervical and\nFig. 6 An example of ovarian endometriosis: A T1w imaging; B, C axial and sagittal T2w imaging; D, E axial and coronal T2 fat sat imaging;\nF T1w imaging after contrast administration. Hyperintensity in T1w is typical of a recent bleeding\n32 Curr Radiol Rep (2022) 10:21–39\n123\n\nrectal lesions; seldom isolated involvement of the vagina\nmay occur. The upper one-third of the vagina and the\nposterior fornix are the most commonly affected sites.\nGenerally, the vaginal wall implants show a thickened or\nnodular appearance [ 21], but may also have a polypoid\nstructure. At MR imaging vaginal endometriotic implants\nshow low-signal intensity on T2-weighted images. They\noften have a multiloculated internal appearance because of\nthe presence of cystic areas [ 66]. These locules can show\nhyperintense content on T1-weighted images due to suba-\ncute blood products. Polypoid variant may have a T2\nhypointense rim corresponding to surrounding fibrous\ntissue associated with endometriosis [ 79]. Rectovaginal\nfistulation represents a complication of vaginal\nendometriosis. Differential diagnosis includes epithelial\nneoplasms arising from the uterine cervix or vaginal wall\n[66].\nUterine Ligaments\nAt MRI the round ligaments can be identified as thin\nstructures with hypointense fibrous signal on T1- and T2-\nweighted images, extending from the uterine horns to the\npelvic side- wall, passing anteriorly to the external iliac\nFig. 7 An example of endometrioma: A T1w imaging; B, C axial and coronal T2w imaging; D axial T2 fat sat imaging; E axial T1w imaging\nafter contrast administration. It’s visible a double fluid–fluid level, indicating different bleedings\nCurr Radiol Rep (2022) 10:21–39 33\n123\n\nvessels. They have an intra- and an extra-pelvic portion, the\nlatter being the distal part of the ligament in the canal of\nNuck [ 80]. When involved by endometriosis, round liga-\nments appear thickened (more than 1 cm), nodular, short-\nened and irregular. Usually endometriotic implants are a\nmixture of fibrous tissue and hemorrhage. Fibrous tissue\nshows hypointense signal on T1- and T2-weighted images;\nsmall hemorrhagic foci displays hyperintense signal on fat-\nsuppressed T1-weighted images [ 66]. The presence of free\nfluid around the intra-pelvic portion of the round ligaments\nmay represent an indirect sign of endometriosis [ 80].\nEndometriosis of the broad ligaments usually manifests\nas thickening and nodularity of these peritoneal folds\nextending between the uterus and the lateral walls of the\npelvis. These nodules are visible as hypo-intensity signal in\nT2 sequences and after administration of contrast material\nit is possible to observe a diffuse peritoneal enhancement\nsecondary to the inflammatory reaction incited by micro-\nscopic endometrial implants on contrast-enhanced fat-sat-\nurated MRI [ 64].\nPosterior Compartment\nUterosacral Ligaments\nUterosacral ligaments (USL) are the most frequent location\nof deep endometriosis. Bilateral USL involvement is often\nassociated with other posterior deep endometriotic loca-\ntions, mostly the rectosigmoid colon [ 81]. At MR imaging\nnormal USL are depicted as thin, regular, semicircular\nhypointense cords that originate from the lateral aspect of\nthe uterine cervix and the vaginal vault and course dor-\nsocranially toward the sacrum [ 21]. USL endometriosis is\nFig. 8 An example of endometriotic cyst: A axial T1w imaging; B, C axial and coronal T2w imaging; D axial T2 fat sat imaging; E, F axial and\ncoronal T1w imaging after contrast administration\n34 Curr Radiol Rep (2022) 10:21–39\n123\n\ndepicted as nodularity within the ligament or as unilateral\nor bilateral hypointense thickening of the ligament, with\nregular or irregular margins [ 21]. The proximal medial\nportion of the USL is most commonly affected by\nendometriosis.\nAccording to Bazot et al. thin-section oblique axial T2-\nweighted sequences (3 mm thick, perpendicular to the long\naxis of the cervix) can improve the capability of conven-\ntional MRI to assess USL endometriosis [ 81].\nSaba et al. suggested that the diagnosis of endometriosis\nof the USL is simple when ligaments are involved together\nwith the torus uterinus, whereas, when there is only a\nthickening or an asymmetric nodular irregularity the\ninvolvement of the USL can be difficult [ 64].\nIn a recently published meta-analysis the sensitivity and\nspecificity of MRI for the diagnosis of endometriosis of\nUSL were 85% and 80%, respectively [ 82].\nRetrocervical Area\nThe retrocervical area is a virtual extraperitoneal space\nbehind the cervix, located above the rectovaginal septum\n[21]. It is a common site of deep pelvic endometriosis.\nRetrocervical implants are often associated with USL\ninvolvement and with the retroversion of the uterus [ 83].\nDeep endometriotic lesions of the retrocervical area\nfrequently appear as ill-defined infiltrative tissue, hypoin-\ntense on T2-weighted images, extending from the posterior\nuterine serosa to the retrocervical region [ 21]. Neverthe-\nless, some lesions may contain abundant glandular com-\nponent and little fibrotic reaction, thus showing high signal\nintensity on T1-weighted images and variable signal\nintensity on T2-weighted images [ 21, 63]. The solid glan-\ndular component enhances after intravenous administration\nof contrast material [ 63].\nDel Frate et al. identified a condition they called\n‘‘hourglass-shaped’’ lesions that are found in 25% of cases\nand are due to posterior extension of a posterior forniceal\nlesion toward the anterior rectal muscularis. These lesions\nare usually larger than 3 cm, with a greater risk of exten-\nsion to the rectal wall [ 63].\nIn a recently published meta-analysis the sensitivity and\nspecificity of MRI for the diagnosis of endometriosis of the\npouch of Douglas were 89% and 94%, respectively [ 82].\nThe differential diagnosis of retrocervical lesions\nincludes peritoneal metastases from intraperitoneal malig-\nnancies (i.e., gastrointestinal and ovarian neoplasms).\nPeritoneal metastases usually show intermediate to high\nsignal intensity on T2-weighted images and, as the primary\ncancer site, high signal intensity on DWI; moreover, ascites\nand a tumor mass into the abdominal cavity may be iden-\ntified [ 84]. On the other hand, solid endometriosis shows\nlow-signal intensity on T2-weighted images [ 66].\nRectovaginal Pouch\nThe rectovaginal pouch is the anatomical region located\nbetween the posterior vaginal wall and the anterior rectal\nwall. It extends from the deepest part of the pouch of\nDouglas to the top of the perineal body [ 66]. The inferior\ntwo thirds of this space constitute the rectovaginal septum,\na thin membranous partition usually filled with fat [ 21].\nUsually rectovaginal implants represent extensions from\nretrocervical or posterior vaginal lesions.\nIn MR I, nodules of endometriosis affecting the recto-\nvaginal pouch usually appear as hypointense nodules on\nboth T1W and T2W MRI images with signal intensity\nclose to that of pelvic muscle [ 85]. Sometimes, foci of\nendometriosis may also have an abundant glandular com-\nponent and discrete fibrotic reaction. In such cases, the\nendometriotic foci are hyperintense nodules on T1W and\nfat-saturated T1W MRI images, irrespective of their\nappearance on T2W MRI images. Moreover, the solid\nglandular component shows variable enhancement after the\nintravenous administration of contrast material [ 64].\nIn normal conditions the MRI depicts the rectovaginal\nseptum as a hyperintense signal area in T1W and T2W\nimages whereas nodules of endometriosis affecting the\nrectovaginal septum usually appear as hypointense nodules\non both T1W and T2W MRI images [\n64].\nIn a recently published meta-analysis the sensitivity and\nspecificity of MRI for the diagnosis of rectovaginal septum\nendometriosis were 82% and 77%, respectively [ 82].\nRectosigmoid Colon\nAmong the bowel segments the rectosigmoid is the most\ncommonly involved by endometriosis (65.7%), followed\nby vermiform appendix, terminal ileum, cecum and\ndescending colon, in order of frequency [ 86]. Rectosigmoid\nendometriosis is often associated with other pelvic loca-\ntions and with a second intestinal lesion in 55% of cases\n[72].\nThe rectosigmoid endometriosis may cause adhesions,\nbowel strictures, or intestinal obstruction may result from\nthe inflammatory response to cyclic hemorrhage. The\nimplants are usually serosal but can sometimes involve the\nsubserosal layers and cause thickening and fibrosis of the\nmuscularis propria. Usually, an intact overlying mucosa is\npresent, since the implanted tissue only rarely invades\nthrough to the mucosa [ 21].\nTypically, endometriotic lesions infiltrating the anterior\nrectal wall have a characteristic fan-shaped configuration\n(or a pyramidal shape, with the base adhering to the rectal\nwall and the apex oriented anteriorly toward the retrocer-\nvical region). The core of the lesion shows isointense signal\ncompared to muscle on T2-weighted and T1-weighted\nCurr Radiol Rep (2022) 10:21–39 35\n123\n\nsequences and at histopathology corresponds to thickening\nand distortion of the muscularis propria and smooth muscle\nhyperplasia [ 66]. The overlying layer, hyperintense on T2-\nweighted images, at the luminal side of the bowel wall\ncorresponds to (sub)mucosal thickening, as a consequence\nof non-specific inflammation with or without infiltration of\nendometriosis [ 86]. When the longitudinal extent of the\nparietal lesion along the bowel wall is short, a pattern of\nintraluminal endophytic growth, called mushroom cap,\nmay be observed [ 87].\nWhen nodules of endometriosis are localized in the\nretroperitoneal section frequent concomitant findings are\nthe adherences that identify fibrotic tissue originating from\nthe nodules of endometriosis and involves the closest\norgans. Sometimes the adherences determine traction of the\naffected organs [88]. This is an important finding because it\ncan be considered an indirect sign of endometriosis.\nIn a recently published meta-analysis the sensitivity and\nspecificity of MRI for the diagnosis of rectosigmoid colon\nendometriosis were 83% and 88%, respectively [ 82]. MR\nimaging is useful to predict infiltration of the muscular\nlayer of the bowel with a sensitivity of 100% and speci-\nficity of 75%. On the other hand, it is of limited value in\ndiagnosing (sub)mucosal infiltration, as (sub)mucosal\nthickening may be caused by edema without infiltration of\nendometriosis. Nevertheless, extensive irregularities of the\n(sub)mucosal layer may raise suspicion of (sub)mucosal\ninvolvement [ 86]. Differential diagnosis includes rectal\ncancer and metastatic implants to the bowel [ 66].\nConclusions\nEndometriosis is a chronic condition that affects women\nduring the reproductive lifespan. Diagnosis of\nendometriosis is complex as it must take into account non-\npathognomonic clinical symptoms and non-specific labo-\nratory tests. The physical examination and above all the\ndifferent imaging techniques, in particular US and MRI,\nconstitute the gold standard for diagnosis. The role of the\nradiologist is fundamental both in the diagnosis of\nendometriosis, especially in the deep sites that cannot be\nseen with ultrasound, and in planning the type of thera-\npeutic approach. A multidisciplinary study is essential for\nthe management of these patients from both a clinical and\nsurgical point of view.\nAuthor Contributions All authors participated in the writing of the\npaper. All authors read and approved the final manuscript.\nFunding The authors received no financial sponsors or other\nfunding for this research.\nDeclarations\nConflict of interests The authors declare that they have no conflict of\ninterests.\nOpen Access This article is licensed under a Creative Commons\nAttribution 4.0 International License, which permits use, sharing,\nadaptation, distribution and reproduction in any medium or format, as\nlong as you give appropriate credit to the original author(s) and the\nsource, provide a link to the Creative Commons licence, and indicate\nif changes were made. The images or other third party material in this\narticle are included in the article’s Creative Commons licence, unless\nindicated otherwise in a credit line to the material. If material is not\nincluded in the article’s Creative Commons licence and your intended\nuse is not permitted by statutory regulation or exceeds the permitted\nuse, you will need to obtain permission directly from the copyright\nholder. To view a copy of this licence, visit http://creativecommons.\norg/licenses/by/4.0/.\nReferences\nPapers of particular interest, published recently, have been\nhighlighted as:\n• Of importance,\n•• Of major importance\n1. Bazot M, Bharwani N, Huchon C, et al. European society of\nurogenital radiology (ESUR) guidelines: MR imaging of pelvic\nendometriosis. Eur Radiol. 2017;27(7):2765–75. https://doi.org/\n10.1007/s00330-016-4673-z.\n2. Nisolle M, Donnez J. Peritoneal endometriosis, ovarian\nendometriosis, and adenomyotic nodules of the rectovaginal\nseptum are three different entities. Fertil Steril. 1997;68:585–96.\n3. Nisenblat V, Bossuyt PM, Farquhar C, Johnson N, Hull ML.\nImaging modalities for the non-invasive diagnosis of\nendometriosis. Cochrane Database Syst Rev.\n2016;2(2):CD009591. https://doi.org/10.1002/14651858.\nCD009591.pub2.\n4. Kinkel K, Frei KA, Balleyguier C, Chapron C. Diagnosis of\nendometriosis with imaging: a review. Eur Radiol.\n2006;16(2):285–98. https://doi.org/10.1007/s00330-005-2882-y.\n5. Chapron C, Fauconnier A, Vieira M, Barakat H, Dousset B,\nPansini V, et al. Anatomical distribution of deeply infiltrating\nendometriosis: surgical implications and proposition for a clas-\nsification. Hum Reprod. 2003;18:157–61.\n6. Pham M, Sommer C, Wessig C, et al. Magnetic resonance neu-\nrography for the diagnosis of extrapelvic sciatic endometriosis.\nFertil Steril. 2010;94(1):351.e11-351.e14. https://doi.org/10.\n1016/j.fertnstert.2009.12.046.\n7. Sampson JA. Peritoneal endometriosis due to menstrual dissem-\nination of endometrial tissue into the peritoneal cavity. Am J\nObstet Gynecol. 1927;14:442–69.\n8. Chapron C, Marcellin L, Borghese B, Santulli P. Rethinking\nmechanisms, diagnosis and management of endometriosis. Nat\nRev Endocrinol. 2019;15(11):666–82. https://doi.org/10.1038/\ns41574-019-0245-z.\n9. Vercellini P, Vigano P, Somigliana E, Fedele L. Endometriosis:\npathogenesis and treatment. Nat Rev Endocrinol.\n2014;10:261–75.\n10. Chapron C, Chopin N, Borghese B, et al. Deeply infiltrating\nendometriosis: pathogenetic implications of the anatomical dis-\ntribution. Hum Reprod. 2006;21(7):1839–45. https://doi.org/10.\n1093/humrep/del079.\n36 Curr Radiol Rep (2022) 10:21–39\n123\n\n11. BuckLouis GM, et al. Bisphenol A and phthalates and\nendometriosis: the endometriosis: natural history, diagnosis and\noutcomes study. Fertil Steril. 2013;100:162-169.e1-2.\n12. Stefansson H, et al. Genetic factors contribute to the risk of\ndeveloping endometriosis. Hum Reprod. 2002;17:555–9.\n13. •• Bazot M, Daraı¨ E. Diagnosis of deep endometriosis: clinical\nexamination, ultrasonography, magnetic resonance imaging, and\nother techniques. Fertil Steril. 2017;108(6):886–94. https://doi.\norg/10.1016/j.fertnstert.2017.10.026 For the explanation of the\nimaging features of the different localization of endometriosis .\n14. Fauconnier A, Chapron C. Endometriosis and pelvic pain: epi-\ndemiological evidence of the relationship and implications. Hum\nReprod Update. 2005;11:595–606.\n15. Ballard K, Lane H, Hudelist G, Banerjee S, Wright J. Can specific\npain symptoms help in the diagnosis of endometriosis? A cohort\nstudy of women with chronic pelvic pain. Fertil Steril.\n2009;94:20–7.\n16. Koninckx PR, Meuleman C, Oosterlynck D, Cornillie FJ. Diag-\nnosis of deep endometriosis by clinical examination during\nmenstruation and plasma CA-125 concentration. Fertil Steril.\n1996;65:280–7.\n17. Abrao MS, Goncalves MO, Dias JA Jr, Podgaec S, Chamie LP,\nBlasbalg R. Comparison between clinical examination,\ntransvaginal sonography and magnetic resonance imaging for the\ndiagnosis of deep endometriosis. Hum Reprod. 2007;22:3092–7.\n18.\n••Hindman N, VanBuren W. Imaging spectrum of endometriosis\n(endometriomas to deep infiltrative endometriosis). Radiol Clin\nNorth Am. 2020;58(2):275–89. https://doi.org/10.1016/j.rcl.2019.\n11.001 For the explanation of the imaging features of the dif-\nferent localization of endometriosis .\n19 Schenken RS. Delayed diagnosis of endometriosis. Fertil Steril.\n2006;86(5):1305–6 ( Discussion: 1317 ).\n20. Bean E, Naftalin J, Jurkovic D. How to assess the ureters during\npelvic ultrasound. Ultrasound Obstet Gynecol.\n2019;53(6):729–33.\n21. Coutinho A, Bittencourt LK, Pires CE, et al. MR imaging in deep\npelvic endometriosis: a pictorial essay. Radiographics.\n2011;31:549–67.\n22. Exacoustos C, Zupi E, Piccione E. Ultrasound imaging for\novarian and deep infiltrating endometriosis. Semin Reprod Med.\n2017;35(1):5–24. https://doi.org/10.1055/s-0036-1597127.\n23. Maccagnano C, Pellucchi F, Rocchini L, et al. Diagnosis and\ntreatment of bladder endometriosis: state of the art. Urol Int.\n2012;89:249–58.\n24. Pateman K, Holland TK, Knez J, et al. Should a detailed ultra-\nsound examination of the complete urinary tract be routinely\nperformed in women with suspected pelvic endometriosis? Hum\nReprod. 2015;30:2802–7.\n25. Crispi CP, de Souza CA, Oliveira MA, Dibi RP, Cardeman L,\nSato H, et al. Endometriosis of the round ligament of the uterus.\nJ Minim Invasive Gynecol. 2012;19:46–51.\n26 Hudelist G, Ballard K, English J, et al. Transvaginal sonography\nvs. clinical examination in the preoperative diagnosis of deep\ninfiltrating endometriosis. Ultrasound Obstet Gynecol.\n2011;37:480–7.\n27.\n• Moro F, Leombroni M, Testa AC. Ultrasound Imaging in\nEndometriosis. Obstet Gynecol Clin North Am.\n2019;46(4):643–59. https://doi.org/10.1016/j.ogc.2019.07.004.\nFor the description of the role of US and MRI in the diagnosis of\nendometriosis\n28. Carfagna P, De Cicco NC, De Cicco NA, et al. Role of\ntransvaginal ultrasound in evaluation of ureteral involvement in\ndeep infiltrating endometriosis. Ultrasound Obstet Gynecol.\n2018;51:550–5.\n29. Maccagnano C, Pellucchi F, Rocchini L, et al. Ureteral\nendometriosis: proposal for a diagnostic and therapeutic algo-\nrithm with a review of the literature. Urol Int. 2013;91(1):1–9.\n30. Knabben L, Imboden S, Fellmann B, Nirgianakis K, Kuhn A,\nMueller MD. Urinary tract endometriosis in patients with deep\ninfiltrating endometriosis: prevalence, symptoms, management,\nand proposal for a new clinical classification. Fertil Steril.\n2015;103(1):147–52.\n31. Noventa M, Saccardi C, Litta P, Vitagliano A, D’Antona D,\nAbdulrahim B, et al. Ultrasound techniques in the diagnosis of\ndeep pelvic endometriosis: algorithm based on a systematic\nreview and meta-analysis. Fertil Steril. 2015;104:366-83.e2.\n32. Holland TK, Cutner A, Saridogan E, Mavrelos D, Pateman K,\nJurkovic D. Ultrasound mapping of pelvic endometriosis: does\nthe location and number of lesions affect the diagnostic accuracy?\nA multicentre diagnostic accuracy study. BMC Womens Health.\n2013;13:43.\n33. Van den Bosch T, Dueholm M, Leone FP, et al. Terms, defini-\ntions and measurements to describe sonographic features of\nmyometrium and uterine masses: a consensus opinion from the\nMorphological Uterus Sonographic Assessment (MUSA) group.\nUltrasound Obstet Gynecol. 2015;46:284–98.\n34. Guerriero S, Condous G, Bosch T, et al. Systematic approach to\nsonographic evaluation of the pelvis in women with suspected\nendometriosis, including terms, definitions and measurements: a\nconsensus opinion from the International Deep Endometriosis\nAnalysis (IDEA) group. Ultrasound Obstet Gynecol.\n2016;48:318–32.\n35. Van den Bosch T, de Bruijn AM, de Leeuw RA, et al. A sono-\ngraphic classification and reporting system for diagnosing ade-\nnomyosis. Ultrasound Obstet Gynecol. 2019;53(5):576–82.\n36. Reinhold C, Atri M, Mehio A, et al. Diffuse uterine adenomyosis:\nmorphologic criteria and diagnostic accuracy of endovaginal\nsonography. Radiology. 1995;197:609–14.\n37. Kepkep K, Tuncay YA, Goynumer G, et al. Transvaginal\nsonography in the diagnosis of adenomyosis: which findings are\nmost accurate? Ultrasound Obstet Gynecol. 2007;30:341–5.\n38. Di Donato N, Bertoldo V, Montanari G, Zannoni L, Caprara G,\nSeracchioli R. Question mark form of uterus: a simple sono-\ngraphic sign associated with the presence of adenomyosis.\nUltrasound Obstet Gynecol. 2015;46(1):126–7.\n39. Dueholm M. Transvaginal ultrasound for diagnosis of adeno-\nmyosis: a review. Best Pract Res Clin Obstet Gynaecol.\n2006;20(4):569–82.\n40. Meredith SM, Sanchez-Ramos L, Kaunitz AM. Diagnostic\naccuracy of transvaginal sonography for the diagnosis of adeno-\nmyosis: systematic review and metaanalysis. Am J Obstet\nGynecol. 2009;201(1):107.e1-107.e6.\n41. Luciano DE, Exacoustos C, Albrecht L, et al. Three-dimensional\nultrasound in diagnosis of adenomyosis: histologic correlation\nwith ultrasound targeted biopsies of the uterus. J Minim Invasive\nGynecol. 2013;20(6):803–10.\n42. Patel MD, Feldstein VA, Chen DC, et al. Endometriomas: diag-\nnostic performance of US. Radiology. 1999;210(3):739–45.\n43. Jones LP, Morgan MA, Chauhan A. The sonographic spectrum of\npelvic endometriosis: pearls, pitfalls, and mimics. Ultrasound Q.\n2019;35(4):355–75.\n44. Guerriero S, Van Calster B, Somigliana E, et al. Age-related\ndifferences in the sonographic characteristics of endometriomas.\nHum Reprod. 2016;31:1723–31.\n45. Van Holsbeke C, Zhang J, Van Belle V, et al. Acoustic streaming\ncannot discriminate reliably between endometriomas and other\ntypes of adnexal lesion: a multicenter study of 633 adnexal\nmasses. Ultrasound Obstet Gynecol. 2010;35(3):349–53.\nCurr Radiol Rep (2022) 10:21–39 37\n123\n\n46. Van Holsbeke C, Van Calster B, Bourne T, et al. External vali-\ndation of diagnostic models to estimate the riskof malignancy in\nadnexal masses. Clin Cancer Res. 2012;18(3):815–25.\n47 Mascilini F, Moruzzi C, Giansiracusa C, et al. Imaging in gyne-\ncological disease. 10: Clinical and ultrasound characteristics of\ndecidualized endometriomas surgically removed during preg-\nnancy. Ultrasound Obstet Gynecol. 2014;44(3):354–60.\n48. Timor-Tritsch IE, Lerner JP, Monteagudo A, Murphy KE, Heller\nDS. Transvaginal sonographic markers of tubal inflammatory\ndisease. Ultrasound Obstet Gynecol. 1998;12(1):56–66.\n49. Rezvani M, Shaaban AM. Fallopian tube disease in the non-\npregnant patient. Radiographics. 2011;31(2):527–48.\n50. Chamie´ LP, Ribeiro DMFR, Tiferes DA, Macedo Neto AC,\nSerafini PC. Atypical sites of deeply infiltrative endometriosis:\nclinical characteristics and imaging findings. Radiographics.\n2018;38(1):309–28. https://doi.org/10.1148/rg.2018170093.\n51. Guerriero S, Alca ´zar JL, Ajossa S, Pilloni M, Melis GB. Three-\ndimensional sonographic characteristics of deep endometriosis.\nJ Ultrasound Med. 2009;28(8):1061–6.\n52. Exacoustos C, Manganaro L, Zupi E. Imaging for the evaluation\nof endometriosis and adenomyosis. Best Pract Res Clin Obstet\nGynaecol. 2014;28(5):655–81.\n53. Guerriero S, Ajossa S, Minguez JA, Jurado M, Mais V, Melis GB,\net al. Accuracy of transvaginal ultrasound for diagnosis of deep\nendometriosis in uterosacral ligaments, rectovaginal septum,\nvagina and bladder: systematic review and meta-analysis. Ultra-\nsound Obstet Gynecol. 2015;46:534–45.\n54. Squifflet J, Feger C, Donnez J. Diagnosis and imaging of ade-\nnomyotic disease of the retroperitoneal space. Gynecol Obstet\nInvest. 2002;54:43–51.\n55. Reid S, Lu C, Hardy N, et al. Office gel sonovaginography for the\nprediction of posterior deep infiltrating endometriosis: a multi-\ncenter prospective observational study. Ultrasound Obstet\nGynecol. 2014;44(6):710–8.\n56. Nisenblat V, Prentice L, Bossuyt PM, Farquhar C, Hull ML,\nJohnson N. Combination of the non-invasive tests for the diag-\nnosis of endometriosis. Cochrane Database Syst Rev. 2016.\nhttps://doi.org/10.1002/14651858.CD012281.\n57. Guerriero S, Ajossa S, Gerada M, et al. ‘ ‘Tenderness-guided’ ’\ntransvaginal ultrasonography: a new method for the detection of\ndeep endometriosis in patients with chronic pelvic pain. Fertil\nSteril. 2007;88:1293–7.\n58. Valenzano Menada M, Remorgida V, Abbamonte LH, Nicoletti\nA, Ragni N, Ferrero S. Does transvaginal ultrasonography com-\nbined with water-contrast in the rectum aid in the diagnosis of\nrectovaginal endometriosis infiltrating the bowel? Hum Reprod.\n2008;23(5):1069–75.\n59. Reid S, Lu C, Casikar I, Reid G, Abbott J, Cario G, et al. Pre-\ndiction of pouch of Douglas obliteration in women with suspected\nendometriosis using a new real-time dynamic transvaginal\nultrasound technique: the sliding sign. Ultrasound Obstet Gyne-\ncol. 2013;41:685–91.\n60. Rossi L, Palazzo L, Yazbeck C, et al. Can rectal endoscopic\nsonography be used to predict infiltration depth in patients with\ndeep infiltrating endometriosis of the rectum? Ultrasound Obstet\nGynecol. 2014;43(3):322–7.\n61. Guerriero S, Saba L, Ajossa S, Peddes C, Angiolucci M, Perni-\nciano M, et al. Three-dimensional ultrasonography in the diag-\nnosis of deep endometriosis. Hum Reprod. 2014;29:1189–98.\n62. Guerriero S, Spiga S, Ajossa S, et al. Role of imaging in the\nmanagement of endometriosis. Minerva Ginecol.\n2013;65:143–66.\n63. Del Frate C, Girometti R, Pittino M, et al. Deep retroperitoneal\npelvic endometriosis: MR imaging appearance with laparoscopic\ncorrelation. Radiographics. 2006;26:1705–18.\n64. Saba L, Sulcis R, Melis GB, et al. Endometriosis: the role of\nmagnetic resonance imaging. Acta Radiol. 2015;56(3):355–67.\nhttps://doi.org/10.1177/0284185114526086.\n65. Nakai A, Togashi K, Kosaka K, et al. Uterine peristalsis: com-\nparison of transvaginal ultrasound and two different sequences of\ncine MR imaging. J Magn Reson Imaging. 2004;20(3):463–9.\n66.\n• Foti PV, Farina R, Palmucci S, et al. Endometriosis: clinical\nfeatures, MR imaging findings and pathologic correlation.\nInsights Imaging. 2018;9(2):149–72. https://doi.org/10.1007/\ns13244-017-0591-0. For the description of the role of US and\nMRI in the diagnosis of endometriosis\n67. Bis KG, Vrachliotis TG, Agrawal R, Shetty AN, Maximovich A,\nHricak H. Pelvic endometriosis: MR imaging spectrum with\nlaparoscopic correlation and diagnostic pitfalls. Radiographics.\n1997;17(3):639–55.\n68. Woodward PJ, Sohaey R, Mezzetti TP Jr. Endometriosis: radio-\nlogic-pathologic correlation. Radiographics.\n2001;21(1):193–216.\n69. Liang CC, Tsai CC, Chen TC, et al. Management of perineal\nendometriosis. Int J Gynaecol Obstet. 1996;53:261–5.\n70. Donnez J, Nisolle M, Squifflet J. Ureteral endometriosis: a\ncomplication of rectovaginal endometriotic (adenomyotic) nod-\nules. Fertil Steril. 2002;77:32–7.\n71. Seracchioli R, Raimondo D, Di Donato N, et al. Histological\nevaluation of ureteral involvement in women with deep infil-\ntrating endometriosis: analysis of a large series. Hum Reprod.\n2015;30(4):833–9. https://doi.org/10.1093/humrep/deu360.\n72. Chamie´ LP, Blasbalg R, Pereira RM, Warmbrand G, Serafini PC.\nFindings of pelvic endometriosis at transvaginal US, MR imag-\ning, and laparoscopy. Radiographics. 2011;31(4):E77–100.\nhttps://doi.org/10.1148/rg.314105193.\n73. Ghezzi F, Raio L, Cromi A, et al. ‘ ‘kissing ovaries’ ’: a sono-\ngraphic sign of moderate to severe endometriosis. Fertil Steril.\n2005;83(1):143–7.\n74. Togashi K, Nishimura K, Kimura I. Endometrial cysts: diagnosis\nwith MR imaging. Radiology. 1991;180:73–8.\n75. Onbas O, Kantarci M, Alper F, et al. Nodular endometriosis:\ndynamic MR imaging. Abdom Imaging. 2007;32:451–6.\n76. Balaban M, Idilman IS, Toprak H, Unal O, Ipek A, Kocakoc E.\nThe utility of diffusion-weighted magnetic resonance imaging in\ndifferentiation of endometriomas from hemorrhagic ovarian\ncysts. Clin Imaging. 2015;39(5):830–3. https://doi.org/10.1016/j.\nclinimag.2015.05.003.\n77. Foti PV, Ognibene N, Spadola S, et al. Non-neoplastic diseases of\nthe fallopian tube: MR imaging with emphasis on diffusion-\nweighted imaging. Insights Imaging. 2016;7(3):311–27. https://\ndoi.org/10.1007/s13244-016-0484-7.\n78. Siegelman ES, Oliver ER. MR imaging of endometriosis: ten\nimaging pearls. Radiographics. 2012;32(6):1675–91. https://doi.\norg/10.1148/rg.326125518.\n79. Tham WP, Busmanis I, Tan WC, Kwek JW. Polypoid\nendometriosis of post vaginal fornix: utility of MRI imaging of\npelvis with diffusion weighted imaging for diagnosis. Med J\nMalaysia. 2016;71(3):144–6.\n80 Gui B, Valentini AL, Ninivaggi V, Marino M, Iacobucci M,\nBonomo L. Deep pelvic endometriosis: don’t forget round liga-\nments. Review of anatomy, clinical characteristics, and MR\nimaging features. Abdom Imaging. 2014;39(3):622–32. https://\ndoi.org/10.1007/s00261-014-0091-3.\n81. Bazot M, Gasner A, Ballester M, Daraı ¨ E. Value of thin-section\noblique axial T2-weighted magnetic resonance images to assess\nuterosacral ligament endometriosis. Hum Reprod.\n2011;26(2):346–53. https://doi.org/10.1093/humrep/deq336.\n82. Medeiros LR, Rosa MI, Silva BR, et al. Accuracy of magnetic\nresonance in deeply infiltrating endometriosis: a systematic\n38 Curr Radiol Rep (2022) 10:21–39\n123\n\nreview and meta-analysis. Arch Gynecol Obstet.\n2015;291(3):611–21. https://doi.org/10.1007/s00404-014-3470-7.\n83. Chamie´ LP, Blasbalg R, Gonc ¸alves MO, Carvalho FM, Abra ˜o\nMS, de Oliveira IS. Accuracy of magnetic resonance imaging for\ndiagnosis and preoperative assessment of deeply infiltrating\nendometriosis. Int J Gynaecol Obstet. 2009;106(3):198–201.\nhttps://doi.org/10.1016/j.ijgo.2009.04.013.\n84. Namimoto T, Awai K, Nakaura T, Yanaga Y, Hirai T, Yamashita\nY. Role of diffusion-weighted imaging in the diagnosis of\ngynecological diseases. Eur Radiol. 2009;19(3):745–60. https://\ndoi.org/10.1007/s00330-008-1185-5.\n85. Caramella T, Novellas S, Fournol M, et al. Deep pelvic\nendometriosis: MR imaging features. J Radiol. 2008;89:473–9.\n86. Busard MP, van der Houwen LE, Bleeker MC, et al. Deep\ninfiltrating endometriosis of the bowel: MR imaging as a method\nto predict muscular invasion. Abdom Imaging.\n2012;37(4):549–57. https://doi.org/10.1007/s00261-011-9790-1.\n87. Yoon JH, Choi D, Jang KT, et al. Deep rectosigmoid\nendometriosis: B mushroom cap ^ sign on T2-weighted MR\nimaging. Abdom Imaging. 2010;35(6):726–31.\n88. Balleyguier C, Chapron C, Dubuisson JB, et al. Comparison of\nmagnetic resonance imaging and transvaginal ultrasonography in\ndiagnosing bladder endometriosis. J Am Assoc Gynecol Lapar-\nosc. 2002. https://doi.org/10.1016/s1074-3804(05)60099-0.\nPublisher’s Note Springer Nature remains neutral with regard to\njurisdictional claims in published maps and institutional affiliations.\nCurr Radiol Rep (2022) 10:21–39 39\n123","source_license":"CC0","license_restricted":false}