{"paper_id":"fc7ed8ff-70bc-44bf-aeda-65b2e0830dd9","body_text":"Endometriosis is described as the presence of\nendometrial tissue in a space outside of the uterus and\nendometrial cavity. This disease affects almost 10% of\nwomen of reproductive age and is usually diagnosed with\nclinical history as most of the cases complain of chronic\npelvic pain ( 1 ,  2 ). The average interval between the start\nof symptoms and surgical diagnosis is 10.4 years ( 3 ).\nBeyond the clinical symptoms and physical examination,\nimaging is the modality for the initial assessment of these\npatients. Imaging techniques currently used to diagnose\nendometriosis are magnetic resonance imaging (MRI)\nand ultrasonography with a preference for sonography in\nrecent years ( 4 ).\nHowever, the combination of transvaginal sonography\n(TVS) and MRI is not recommended for a more accurate\ndiagnosis ( 5 ). But still, other causes such as fibroma, corpus\nluteum, cystadenoma, tubo-ovarian abscess, teratoma, and\ncarcinoma are needed to be ruled out ( 6 - 8 ). Identification\nof the endometriotic nodules and their correct localization\nenables complete lesion mapping before surgery and\nprevents unexpected plan changes in surgery ( 1 ,  6 ,  9 ,  10 ).\nDeep infiltrating endometriosis (DIE) is recognized as the most severe form of endometriosis has a complex\nclinical approach; it is described as a lesion that penetrates\n>5 mm under the peritoneal surface ( 11 ). DIE accounts\nfor 15 to 30% of all endometriosis cases of which 90%\nare characterized by chronic pelvic pain and infertility,\nand 25% are accidentally discovered during laparoscopy\nor laparotomy ( 12 ,  13 ). DIE nodules infiltrate mostly\nthe uterosacral ligaments (USL), rectosigmoid, vaginal\nfornix, rectovaginal septum, and/or bladder ( 14 ).\nIntestinal endometriosis comprises a spectrum from\nsimple adhesions between the intestine and cervix to nodular\nlesions that might involve serous membrane to the mucosa.\nThese kinds of severe involvements require simultaneous\ncooperation between the colorectal and the gynecology\nsurgeons. Due to various diameters and involvement stages,\nseveral surgical approaches have been proposed and used ( 1 ,\n 9 ). While smaller, less invasive, lesions are removed using\nstapled trans-anal resection, the larger and more invasive\nones need segmental resection ( 6 ,  15 ).\nA precise consensus on the definition and severity of\nendometriosis isn’t reached yet but the most frequently\nused classification is the American Society of Reproductive\nMedicine (ASRM) classification; however, it fails to\ncompletely represent DIE's characteristics ( 16 ,  17 ). It’s\nsuggested that TVS is 79% sensitive and 94% specific in\nthe assessment of the extent of DIE ( 2 ) meanwhile, it is\nproposed that DIE pelvic ultrasonography, which includes\nrectal and\\or vaginal ultrasonographic imaging, is more\naccurate regarding the extent and severity ( 1 ,  2 ,  6 ,  11 ).\nWe designed and conducted this cross-sectional study\nto assess the accuracy of DIE ultrasonography and to\ndo so, we compared the results with pathological and\nsurgical findings, particularly with results of rectal\ninvolvement. It’s suggested that TVS is 79% sensitive and\n94% specific in the assessment of intestinal DIE. In this\nstudy, we assessed the accuracy of DIE ultrasonography\n(rectal and\\or vaginal ultrasonography) which is thought\nto be more accurate.\n\nWe designed and conducted this cross-sectional study on\npatients with severe endometriotic symptoms who were a\ncandidate for laparoscopic surgery and their disease was\nlater confirmed histologically from December 2019 to\nDecember 2020. Our patients who were suspected of DIE\nwere assessed in regards to the following characteristics\nand variables: age, body mass index (BMI) category,\nconfirmed DIE or ovarian endometrioma (OMA), and\nthe respective location and the level of involvement.\nOur patients were 35.41 years old on average with a\nstandard deviation of 5.94. The symptoms included\npelvic pain, dysmenorrhea, dyspareunia, infertility,\nabnormal uterine bleeding (AUB), and dysphasia. The\npatients were enrolled from the laparoscopic office of\nArash hospital at Tehran university of medical sciences.\nOur exclusion criteria included the patients who were\npregnant, menopausal, or had a non-endometrial mass\nin adnexa, or other malignancies. We also excluded any\npatients who had any contraindications from the surgery.\nThe patients who were of reproductive age and had a\ntypical medical history compatible with endometriosis\nwere also assessed, and if their imaging and pathological\nfindings were consistent, they were included in the study.\nAll patients included in the study provided informed\nconsent. In this study we considered pathologically\napproved surgical results as our gold standard; thus, all\nour data was compared and tested with surgical findings\nconfirmed by pathology. All patients were assessed by\nboth the attending professor and the fellowship trainees,\nand all data relating to endometriosis such as pelvic\npain, dysmenorrhea, dyspareunia, infertility, and AUB\ndysphasia were collected.\nAll the features and data gathered from ultrasonographic\nimaging along with surgical and pathological findings\nwere collected, recorded, and analyzed. The patient’s\nintestinal involvement was scored from 0 to 3 (0 being no\ninvolvement and 3 being full mucosal involvement). Other\nanatomical sites and areas such as adnexa, cul de sac,\nUSLs, and the salpinx were also assessed and compared.\nWe also gathered general body statistics of the patients\nand assessed the accuracy using the aforementioned data.\nBased on the assumption from previous studies that DIE\nultrasonography is up to 96% sensitive we calculated that\nour minimum cases should include 70 patients (Cochrane’s\nsample size formula). In total, 109 cases were chosen for\nthe study, and the data were analyzed using IBM’s SPSS\nv26 software (IBM, USA). Our primary goal was to\nassess the sensitivity, specificity, and positive predictive\nvalue of DIE ultrasonographic examination particularly\nin the extent of intestinal involvement. We also used cross\ntabulation and chi-square tests to assess the significance\nof the tests.\nThis study was ethically approved by the Ethical\nCommittee of the Tehran University of Medical Sciences\n(IR.TUMS.MEDICINE.REC.1399.936) and all patients\nhad signed informed consent forms.\n\nIn total, 150 cases with symptoms were chosen and 109\ncases had either DIE or OMA, and 41 were not chosen\ndue to no findings in ultrasonography. As reported\nby the pathology laboratory, there were 97 cases of\npathologically confirmed ovarian endometrioma, 42\ncases had intestinal involvement, 56 had uterosacral\nDIE, 19 cases had uterus adenomyosis and 9 cases were\ndiagnosed with myoma. We also asked the patients to\nevaluate and score their symptoms from 0 to 10 and on\naverage; The main symptoms that patients complained\nof were pelvic pain (80.3%), dysmenorrhea (85.3%),\ndyspareunia (48.6%), dysphasia (43.1%), AUB\n(29.4%) and infertility (29.4%) respectively ( Fig .1 ).\nThe symptoms were scored as follows; scored as the\nfollowing dysmenorrhea at 6.74, dyspareunia at 3.36,\nand dysphasia at 2.72 respectively.\nProportion of each symptom felt by patients.\nIn regards to the accuracy of ultrasonography imaging\nin the diagnosis of intestinal DIE, which was our primary\noutcome, we found that ultrasonographic imaging\nperformed excellently in overall diagnosis since it was\n97.6% sensitive and 73.8% specific. However, laparoscopic\nevaluation was far more diagnostic (97.6% sensitive and\n97.2% specific). As for the levels of involvement in the\nintestine, we compared the ultrasonographic imaging\nfindings with pathologic results and the results showed\nlower accuracy; 55.6, 50.0, 66.7% sensitive, and 72.0,\n85.6, 91.5% specific for serous membrane, muscular layer\nand mucus membrane respectively. The average BMI\nwas 24.7 and most of the cases were in the normal range\n(46.3%) (Tables 1 ,  2 ,  Fig .2 ).\nDiagnostic accuracy of DIE ultrasonography and laparoscopy in diagnosis of DIE and endometrial\nlesions\nDIE US; Ultrasonographic imaging, PPV; Positive predictive value, NPV; Negative predictive value, and OMA; Ovarian endometrioma.\nDiagnostic accuracy of ultrasonographic imaging in regards to intestinal level of involvement\nPPV; Positive predictive value, NPV; Negative predictive value, and BMI; Body mass index.\nROC curve diagrams showcasing the accuracy of DIE sonography and laparoscopy in regards to\nintestinal involvement. ROC; Receiver operating characteristic and DIE; Deep\ninfiltrating endometriosis.\nWe also assessed the effect of obesity and weight on\nUS imaging; we compared the results of US imaging of\nthe intestine in 4 BMI brackets as follows: underweight\n(BMI<18.5), normal range (18.5 to 24.9), overweight (25\nto 29.9), and obese (BMI>30). The results showed that\nboth sensitivity and specificity were negatively affected.\nThese results were statistically significant except for the\nunderweight BMI bracket, which we believe was due to\nthe small sample size ( Table 3 ).\nThe effect of body mass composition on accuracy of ultrasonographic imaging\nPPV; Positive predictive value and NPV; Negative predictive value.\nWe also assessed the accuracy with respect to OMA and\nthe DIE that infiltrates cul de sac and USLs. The results\nindicated that ultrasonographic imaging was 99.0%\nsensitive and 84.6% specific. The data of accuracy show\nthat the examination for cul de sac was 100% sensitive,\nand 50.8% specific while it was 96.4% sensitive and\n59.1% specific in assessing USLs. Although imaging is\nquite sensitive, it can be inaccurate regarding cul de sac\nand USL assessment since their positive predictive value\nwas 25.0 % and 58% respectively; however, the results\nfor ovarian assessment showed 92.3% PPV ( Table 1 ).\n\nIn our study, we identified that overall diagnostic\naccuracy was 97.6% sensitive and 73.8% specific.\nHowever laparoscopic evaluation was found to be far\nmore accurate (97.6% sensitive and 97.2% specific). DIE\nwas also found to lack accuracy in regard to the extent\nof involvement. It was also not accurate with respect to\nassessing cul de sac and USL. Imaging has always been\nan important tool in both the diagnosis and surgical\napproach to endometriosis. A thorough evaluation can\nhelp diagnosis and the entire approach and planning.\nThus, it’s of utmost importance that the data pertaining\nto the lesion is both accurate and reproducible, therefore\nwe aimed to assess DIE ultrasonographic imaging as a\ncomplementary and multi-perspective imaging approach.\nDIE pelvic ultrasonography consists of vaginal and\\or\nrectal US imaging ( 18 ). In a study conducted by S. Alborzi\net al. ( 19 ), it was stated that ultrasonographic imaging\n(transvaginal or transrectal) is as accurate as MRI in the\ndetection of lesions.\nIn our study, the diagnostic accuracy of ultrasonographic\nimaging in the identification of intestinal lesions,\nwhich was our primary outcome, was almost as high\nas laparoscopic evaluation. Therefore, we suggest that,\nin the overall diagnosis of DIE in the intestine, this\nprocedure could be useful. In a multicenter prospective\nand retrospective cohort study conducted in the royal\ncollege of obstetrics and gynecology the accuracy of\nthe preoperative ultrasound‐based endometriosis staging\nsystem (UBESS) regarding the complexity of surgery was\nassessed; this study showed that US-based imaging can be\nutilized to plan the surgery ( 16 ).\nWe also assessed the accuracy of DIE US imaging\nwith respect to other pelvic cavities and sites; in regards\nto ovarian endometrioma, we concluded that DIE\nultrasonography can be a very efficient and accurate tool\n(99.0% sensitive and 84.6% specific) and as manifested\nby several other studies such as the study conducted by\nHolland et al. ( 16 ) can distinguish between different\npathologies. Their study showed that TVS is an accurate\nassessment tool for the severity of pelvic endometriosis\nand the results are mostly in accord with laparoscopic\nfindings. Meanwhile, we also studied the accuracy of\nDIE ultrasonographic imaging in the diagnosis of lesions\nlocated at USLs and cul de sac and concluded that even\nthough sensitivity for these lesions was high (100% and\n96.4% sensitive for cul de sac and USLs respectively) the\ntests can be inaccurate as their PPV and specificity were\nlow.\nThere were some limitations in our study that reduced the\ndiagnostic accuracy of ultrasonography in DIE patients.\nWe believe that ultrasound imaging accuracy could be\nhampered as poor bowel preparation can limit ultrasound\nwave penetration. On the other hand, the procedure itself\n(TRUS) is painful. These two can both limit the time\nrequired for investigation. Another reason that has led to\nlower accuracy could be the fact that linear nodules could\nbe missed during laparoscopic surgery, particularly in cul\nde sac. we lacked sufficient samples for specific groups\nsuch as the patients with obese body composition.\nIn regards to the body composition of the subjects, we\nconcluded that with higher BMI values the efficacy of US\nimaging plummets. As described by Bushberg et al. ( 20 ),\ndue to fat impedance, 94% of the original sound wave is\nattenuated particularly in patients with more than 8 cm\nof subcutaneous fat before it even reaches the peritoneal\ncavity; hence, this phenomenon affects the acuity of\nultrasonographic imaging.\n\nOur study showed that while DIE pelvic ultrasonographic\nimaging can be a helpful paraclinical tool in the\nassessment and diagnosis of DIE and endometriosis in\ngeneral and particularly with adnexal and bowel lesions,\nit can have some shortcomings with respect to cul de sac\nand USLs. We also suggest that in overweight patients\nthese procedures should be performed more meticulously\nand probably in conjunction with other imaging methods\nsuch as MRI.","source_license":"CC0","license_restricted":false}