{"paper_id":"333daeac-a863-4280-b30c-b02a806b438c","body_text":"32\nReproductive Health  ●  August 2025  ●  Copyright © 2025 EMJ  ●  CC BY-NC 4.0 Licence\nNew Frontiers in Endometriosis: \nImaging and Beyond\nFROM INVASIVE PROCEDURES TO \nIMAGING FIRST\nThe session opened with Mee Kristine, \nOslo University Hospital, Norway, who \ntraced the diagnostic evolution from \ninvasive laparoscopy to the current use of \nnon-invasive imaging. Today, transvaginal \nsonography (TVS) and MRI are at the heart \nof diagnosis and disease mapping, helping \nto identify the three main phenotypes \nof endometriosis: superficial peritoneal, \novarian (endometriomas), and deep \ninfiltrating endometriosis.\nTVS remains the first-line imaging modality. \nIt is widely available, cost-effective, and \nenvironmentally friendly, with excellent \ntest performance. As a dynamic tool, TVS \nallows real-time interaction with the patient, \nenabling the clinician to assess site-specific \ntenderness and gain immediate insight. \nHowever, TVS has limitations, particularly in \ndetecting peritoneal lesions and disease in \nthe lateral pelvic compartments.\nMRI serves as a valuable second-line tool \nwhen TVS is inconclusive or negative in \npatients who are symptomatic. It is also \nused preoperatively and postoperatively \nif symptoms persist. Its strengths include \nthe ability to generate multiplanar images \nand better visualisation of lateral and \nextra-pelvic disease, which are crucial \nfor identifying issues such as ureteral \ninvolvement. However, MRI lacks dynamic \ninteraction with the patient and, like TVS, \nhas limited ability to detect superficial \nperitoneal lesions.\nIMAGING AT THE CENTRE OF \nDIAGNOSIS AND MANAGEMENT\nThe 2022 ESHRE guideline1 formally \nplaced imaging at the forefront of the \nendometriosis diagnostic pathway, \nrecommending it alongside clinical \nexamination as the first-line assessment \nfor suspected endometriosis. Importantly, \na negative ultrasound does not rule out the \ncondition. In cases where empirical medical \ntherapy is ineffective or inappropriate, \nparticularly in infertility, diagnostic \nlaparoscopy may still be necessary, \nespecially to assess peritoneal involvement.\nImaging now plays a critical role \nbeyond diagnosis, guiding treatment \ndecisions, surgical planning, and long-\nterm management. It provides essential \ninformation about lesion location, size, and \ncomplexity, which informs surgical strategy, \nrisk evaluation, and the required level of \nsurgical expertise.\nAuthor: Ada Enesco, EMJ, London, UK\nCitation: EMJ Repro Health. 2025;11[1]:32-36.  \nhttps://doi.org/10.33590/emjreprohealth/NRUM8961 \nA DEDICATED session on endometriosis at the 41ˢᵗ Annual Meeting of \nthe European Society of Human Reproduction and Embryology (ESHRE) \nhighlighted the significant progress being made in both imaging technologies and \nemerging diagnostic biomarkers. Experts explored how clinical tools are evolving to \nsupport earlier, more accurate, and more patient-friendly diagnosis, redefining the way \nendometriosis is detected and managed.\nChange colours to TA area\nCongress Feature  ●  ESHRE 2025\n\nCC BY-NC 4.0 Licence  ●  Copyright © 2025 EMJ  ●  August 2025  ●  Reproductive Health\n33\nFERTILITY AND IMAGING\nEndometriosis often affects fertility, being \nlinked to reduced ovarian reserve and \nelevated oxidative stress in the pelvic \nenvironment. Here, imaging is essential: it \nsupports fertility preservation, guides egg \nretrieval in assisted reproductive technology \n(ART), and helps tailor treatment protocols \nbased on uterine and ovarian accessibility.\nA recent Australian study showed the \nimportance of early endometriosis \ndiagnosis.2 Women diagnosed with \nendometriosis after their first ART cycle \nrequired more treatment cycles, had \nhigher rates of intrauterine insemination, \nand reported lower live birth rates than \nwomen without endometriosis. In contrast, \nwomen diagnosed prior to initiating ART \nhad outcomes similar to those without \nendometriosis. This highlights how timely \nimaging and diagnosis can improve fertility \nsuccess rates.\nDIAGNOSTIC GAPS AND THE NEED \nFOR EXPERTISE \nDespite its advantages, imaging still faces \nchallenges. Both TVS and MRI have limited \nsensitivity for detecting gastrointestinal, \ndiaphragmatic, and superficial peritoneal \ndisease. Operator expertise is also a major \nfactor, though standardised protocols and \nkey signs (such as the ‘negative sliding \nsign’ or ovarian immobility) can assist in \ndetection in less specialised settings.\nSurgical treatment continues to play an \nimportant role, particularly for women with \nsevere symptoms, anatomical distortion, \nor fertility concerns. The goals of surgery \ninclude symptom relief, anatomical \nrestoration, function preservation (such  \nas bowel and ureteral integrity), and \nrecurrence prevention.\nThe latest ESHRE Guidelines support \nsurgery for endometriosis-related infertility \nin cases of minimal-to-mild disease and \novarian endometriosis, where evidence \nsuggests improved spontaneous pregnancy \nrates postoperatively.1 However, surgery \nbefore ART is not recommended for \nsuperficial or ovarian endometriosis. In  \ndeep disease, decisions should be \nindividualised based on pain severity  \nand patient preference.\nBoth TVS and MRI have limited \nsensitivity for detecting \ngastrointestinal, diaphragmatic, and \nsuperficial peritoneal disease\nESHRE 2025  ●  Congress Feature\n\n34\nReproductive Health  ●  August 2025  ●  Copyright © 2025 EMJ  ●  CC BY-NC 4.0 Licence\nSurgical technique is also evolving. For \nsuperficial peritoneal disease, the standard \napproach is shifting from ablation, which \ndestroys the lesion with heat, to excision, \nwhich involves removing the lesion along \nwith the affected peritoneum. This method \nis more complex and demands advanced \nsurgical planning, once again highlighting \nthe indispensable role of accurate imaging.\nIMAGING INNOVATIONS\nFuture directions in imaging aim to \novercome current limitations. 3D TVS \noffers better spatial visualisation and \ncould improve the detection of deep or \neven superficial lesions. Though difficult \nto identify on current imaging, superficial \nlesions may be visible if there is some fluid \nin the posterior cul-de-sac. New methods \nare under investigation to enhance the \ndetection of these subtle findings.\nIn 2024, the International Deep Endometriosis \nAnalysis (IDEA) group consensus expanded \nits guidelines to include routine evaluation \nof the parametrium, helping to improve \nidentification of lateral compartment \ndisease.3 Another exciting development is \npelveoneurosonography, which enables \nvisualisation of the sacral nerve roots and \nplexus, structures often involved in chronic \npelvic pain but difficult to assess with \ntraditional imaging.\nAdditionally, molecular imaging is showing \npromise. The University of Oxford’s DETECT \nstudy4 is evaluating 99mTc-maraciclatide, a \nradiolabelled tracer that binds to the αvβ3 \nintegrin, a protein expressed on the surface \nof endometriotic lesions. This novel agent \nmay allow non-invasive detection of early-\nstage endometriosis, a major breakthrough \nif successfully validated. The tracer offers \nthe potential for functional imaging of active \nlesions and could complement conventional \nanatomical imaging modalities.\nAI: SUPPORTING  \nDIAGNOSIS AND ACCESS\nAI is also making its way into endometriosis \ndiagnostics, offering opportunities to \nimprove early detection and overcome \nworkforce shortages. AI could be used for \ntriage, helping to identify patients who need \nfurther imaging, as well as to accelerate \ndiagnosis in adolescents or those \nexperiencing infertility.\nCongress Feature  ●  ESHRE 2025\n\nCC BY-NC 4.0 Licence  ●  Copyright © 2025 EMJ  ●  August 2025  ●  Reproductive Health\n35\nHowever, challenges remain. MRI data \nare complex, requiring substantial \ncomputational power. TVS images, by \ncontrast, are highly operator-dependent \nand variable. AI model development is also \nlimited by small and non-representative \ndatasets, a lack of validation, and \ninconsistent data quality. 5\nDespite these hurdles, AI holds long-term \npromise. It could support less experienced \nclinicians, reduce diagnostic delays, and \nstreamline patient access to expert care, \nparticularly in underserved regions.\nBEYOND IMAGING:  \nTHE RISE OF BIOMARKERS\nIn the second part of the session, Arne \nVanhie, Leuven University Fertility Centre \nand University Hospital Leuven, Belgium, \naddressed the growing interest in non-\ninvasive biomarkers as a complement, or \npotential alternative, to imaging.\nHe began by distinguishing between \nbiomarkers and diagnostic tests. While \nbiomarkers may correlate with the presence \nof disease, true diagnostic tests must \ndemonstrate measurable performance \nusing metrics such as sensitivity, specificity, \npositive predictive value (PPV), and \nnegative predictive value (NPV). Vanhie also \nhighlighted the crucial role of prevalence \nin determining a test’s utility, as lower \nprevalence dramatically reduces PPV, \nmeaning tests are more reliable in specialist \nsettings than in general practice.\nTo meet the diverse needs in endometriosis \ndiagnosis, he identified three types of \ndiagnostic tests, each with a different \naim, population, and expected outcome. A \nreferral test aims to triage patients who are \nsymptomatic more effectively for imaging, \nand must prioritise high sensitivity to avoid \nmissed cases. A replacement test could \neventually obviate the need for laparoscopy \nin patients with negative imaging, requiring \na high NPV or PPV, and should reduce \nhealthcare costs. A ‘red flag’ test would \nidentify patients likely to have deep disease \nand ensure they are referred to expert \ncentres. This type of test must offer a high \nPPV and contribute to increased detection \nof deep endometriosis.\nPROMISING BIOMARKER RESEARCH\nRecent advances have yielded some \nexciting candidates for the diagnosis of \nendometriosis. One area of promise lies in \nsalivary microRNAs. A 2022 study involving \n153 patients with various disease stages \nused a random forest model based on 109 \nsalivary microRNAs.6 Interim data from \na multicentre validation study are highly \nencouraging, showing 96.2% sensitivity, \n95.1% specificity, and a PPV of 95.1%.7 \nThese results suggest real potential for \nclinical application, although full validation is \nstill ongoing.\nAnother line of research has identified \nplasma protein biomarkers, including \nproteins involved in the coagulation \ncascade, complement system, and protein-\nlipid complexes.8 One model, trained \nspecifically to detect Stage III–IV disease, \nshowed excellent sensitivity and specificity \nand could be a promising candidate for a \nred flag test. Even when applied across all \n3D TVS offers better spatial \nvisualisation and could improve \nthe detection of deep or even \nsuperficial lesions\nInterim data from a multicentre validation \nstudy are highly encouraging, showing\n96.2 \nsensitivity%\n95.1 \nspecificity\n95.1 \nPPV\nESHRE 2025  ●  Congress Feature\n\n36\nReproductive Health  ●  August 2025  ●  Copyright © 2025 EMJ  ●  CC BY-NC 4.0 Licence\ndisease stages, the model performed well, \nwith 87% sensitivity and 72% specificity \nfor early-stage (Stage I) endometriosis. As \nwith the salivary test, further independent \nvalidation is needed.\nA FUTURE WITHIN REACH\nThe ESHRE 2025 session made it clear \nthat the future of endometriosis diagnosis \nlies in integrated, non-invasive, and \npersonalised care. Imaging has become \ncentral not just for diagnosis but also for \nsurgical planning, fertility management, and \ndisease monitoring. Molecular imaging and \nAI are adding new layers of insight, while \nbiomarkers are approaching  \nclinical readiness. \nThough challenges remain in validation, \nstandardisation, and access, the \ncombination of advanced imaging, AI, \nand biomarkers offers a path toward \nearlier detection, fewer diagnostic delays, \nimproved surgical outcomes, and better \nquality of life for patients. As Vanhie \nconcluded, “Are we there yet? Not quite, but \nwe may be closer than ever.”\nReferences\n1. Becker CM et al.; ESHRE \nEndometriosis Guideline Group. ESHRE \nguideline: endometriosis. Hum Reprod \nOpen. 2022;2022(2):hoac009.\n2. Moss KM et al. Delayed diagnosis of \nendometriosis disadvantages women \nin ART: a retrospective population \nlinked data study. Hum Reprod. \n2021;36(12):3074-82.\n3. Guerriero S et al. Addendum to \nconsensus opinion from International \nDeep Endometriosis Analysis (IDEA) \ngroup: sonographic evaluation of \nthe parametrium. Ultrasound Obstet \nGynecol. 2024;64(20):275-80.\n4. University of Oxford. DETECT \n(Detecting Endometriosis inTEgrins \nUsing teChneTium-99m Imaging \nStudy). NCT05623332. https://\nclinicaltrials.gov/study/NCT05623332.\n5. Dungate B et al. Assessing the \nutility of artificial intelligence in \nendometriosis: promises and \npitfalls. Womens Health (Lond). \n2024;20:17455057241248121. \n6. Bendifallah et al. Salivary \nmicroRNA signature for diagnosis \nof endometriosis. J Clin Med. \n2022;11(3):612.\n7. Validation of a salivary miRNA \nsignature of endometriosis \n– interim data. NEJM Evid. \n2023;2(7):EVIDoa2200282.\n8. Schoeman EM et al. Identification \nof plasma protein biomarkers \nfor endometriosis and the \ndevelopment of statistical models \nfor disease diagnosis. Hum Reprod. \n2025;40(2):270-9.\nCongress Feature  ●  ESHRE 2025","source_license":"CC0","license_restricted":false}