{"paper_id":"06118e09-c21e-435e-a51b-5d393ef97509","body_text":"ORIGINAL ARTICLE\nEvidence-based gynaecological practice: clinical review 3.\nThe use of imaging for pre-operative planning in deep\ninfiltrating endometriosis involving the rectum\nA. L. Nightingale & K. D. Ballard & J. T. Wright\nReceived: 29 December 2009 / Accepted: 12 June 2010 / Published online: 14 July 2010\n# Springer-V erlag 2010\nAbstract In this third review of the series of evidence-\nbased practice reviews of clinical practice, we examine the\nevidence for the use of imaging techniques in the pre-\nplanning for surgery for deep infiltrating endometriosis,\nparticularly where there is recto-sigmoid disease. As in\nprevious evidence-based practice reviews, we set a clinical\nscenario of a patient with suspected deep infiltrating\nendometriosis affecting her left utero-asacral ligament and\nwith a suspicion of recto-sigmoid disease and ask which\ndiagnostic tests would be the most accurate for diagnosis\nand pre-operative planning.\nKeywords Recto-sigmoid . Endometriosis . Imaging .\nRectal endometriosis . Deep infiltrating endometriosis .\nDisease staging . Sonography . MRI\nOur patient\nMiss A is a 32-year-old accountant with a history of severe\ndysmenorrhoea, deep dyspareunia and some dyschezia.\nDigital examination suggests she has a small nodule of\nendometriosis between her cervix and her rectum involving\nthe left utero-sacral ligament. Her clinician is concerned\nabout possible rectal involvement and the potential need for\nbowel surgery. She has been advised that it would be\nappropriate for her to undergo a magnetic resonance\nimaging scan, transvaginal and transrectal ultrasound\nexaminations to look at this area in more detail. Although\nwilling to undergo whatever examinations are completely\nnecessary, Miss A is claustrophobic and would not wish to\nundergo an enema or transrectal examination unless they\nwere going to have a real advantage to the planning of her\nsurgery. What would our advice be?\nThe question we want to ask of the literature is: in women\nwith typical clinical signs and symptoms of deep endometri-\nosis, what is the accuracy of magnetic resonance imaging\n(MRI), transvaginal sonography (TVS) and transrectal sonog-\nraphy (TRS) for the detection of and staging of deep\ninfiltrating endometriosis in the rectum and recto-sigmoid\ncolon, when compared to laparoscopy and histology?\nIntroduction\nDeep-infiltrating endometriosis is a term used to describe\nendometriotic lesions that penetrate for more than 5 mm under\nthe peritoneal surface. It infiltrates into vital structures such as\nthe bowel, bladder and ureters. For the purpose of answering\nour clinical question, we are concerned with endometriosis\nthat infiltrates into the rectal and recto-sigmoid area where\nthere is involvement of the muscularis of the bowel wall. Our\nfocus on this is simply because the surgical treatment of\nendometriosis in these areas will require some form of bowel\nexcision, and therefore appropriate patient counselling and\npre-operative preparation.\nSurgery for deep infiltrating endometriosis\nIt is now reasonably well established that deep infiltrating\nendometriosis in the posterior cul de sac is best treated\nDeclaration of interest The authors report no conflicts of interest.\nThe authors alone are responsible for the content and writing of the\npaper.\nA. L. Nightingale : K. D. Ballard : J. T. Wright ( *)\nWomen’s Health Research Unit, Postgraduate Medical School,\nUniversity of Surrey,\nGuildford, UK\ne-mail: jeremy.wright@surrey.ac.uk\nGynecol Surg (2010) 7:407 –415\nDOI 10.1007/s10397-010-0605-0\n\nsurgically [ 1, 2] and will usually involve the dissection of\nthe pelvic sidewall, utero-sacral ligaments and the recto-\nvaginal space. This dissection is within the competence of\nthe appropriately trained gynaecologist but when the rectum\nor recto-sigmoid colon is affected, appropriate involvement\nof general surgeons will be necessary. Although the\npresences of endometriosis can only be determined accu-\nrately by histology, pre-operative investigations should\nideally be able to accurately confirm, or more importantly,\nexclude rectal or recto-sigmoid involvement so that the\nwoman can receive proper counselling on the surgery she\nwill undergo, and ensure the necessary surgical team is\navailable and prepared. In this review, we look at the\naccuracy of pre-operative imaging in both confirming and\nexcluding rectal and recto-sigmoid involvement by deep\ninfiltrating endometriosis.\nAccuracy of diagnostic tests\nWhen considering the accuracy of diagnostic tests it is\nimportant to understand not only the methodology of\ndiagnostic test studies but the outcomes such as sensitivity,\nspecificity, positive and negative predictive values and\npositive and negative likelihood ratios (Table 1). With these\nmeasures, we can assess whether a test is clinically useful\nor not and then we need to decide whether the test is then\nboth cost-effective and acceptable to patients.\nSensitivity is the proportion of people with the disease\nthat are correctly identified with a positive test result.\nSpecificity is the proportion of people without the disease\nwho are correctly identified with a negative test result. A\ntest that has a sensitivity of 98% and a specificity of 50% is\ngood at correctly identifying diseased patients but not so\ngood at correctly identifying true-negative patients.\nHowever, calculations of sensitivity and specificity can\nonly tell us how good the test is in general and it is difficult\nto relate these measures to an individual patient as there is\nno way of predicting whether a positive test result in a\nspecific patient is a true-positive or false-positive result and\nwhether a negative test result is a true-negative or false-\nnegative result [ 3]. For this reason, positive and negative\npredictive values were developed. The positive predictive\nvalue is the proportion of patients with a positive test result\nthat is truly positive; the negative predictive value is the\nproportion of patients with a negative test that is truly\nnegative. These measures, however, are dependent on the\nprevalence of disease within the study population and are\ntherefore not necessarily generalisable to every clinical\npopulation [ 3].\nIn order to find out what the chances are of a patient with\na positive or negative test result having or not having the\ndisease, the likelihood ratios (LRs) need to be calculated.\nThe LR incorporates both the sensitivity and specificity of\nthe test and provides an estimate of how much a positive or\nnegative test result changes the chances of having or not\nhaving the disease. So a positive LR tells you how much\nmore likely patients are to have the disease when the test is\npositive. The negative LR tells you how much less likely\npatients are to have the disease when the test is negative.\nIn general terms, a positive likelihood ratio of more than\n10 indicates that a test is good at confirming disease if the\ntest is positive and a negative likelihood ratio of less than\n0.1 indicates that the test is good at excluding disease when\nthe test is negative.\nA particular benefit with a LR is that it can be combined\nwith the prevalence of the disease (pre-test probability) to\ngive you the post-test probability for an individual patient.\nFigure 1 summarises the interpretation of LRs in relation to\nthe usefulness of the test to confirm or exclude disease and\nthe approximate change in pre-test probability of disease\n[4]. The easiest way to calculate the post-test probability is\nto use a likelihood nomogram (Fig. 2). From Fig. 2, you\nwill see that the disease prevalence (pre-test probability) is\n50%. So for patients who have not had any tests, but\npresent with the typical signs of a given disease that has a\n50% prevalence, they have a 50% chance of having the\ndisease. The positive LR for our test is 35. If you follow the\nline, you will see that this results in a post-test probability\nof 95%. This means that following a positive test result, our\npatient, has a 95% chance of having the disease. Her\nchances have increased from 50% to 95%. Similarly, if the\nnegative likelihood for our test is 0.1 and our patient had a\nnegative test result, her chances of not having the disease\nhave gone from 50% (the prevalence) to just 9%.\nYou can access an interactive nomogram and 2×2 calculator\nthat will calculate sensitivity, specificity and likelihood ratios\nwith their 95% confidence intervals from diagnostic test study\nTable 1 2×2 Table for diagnostic test studies\nTarget disorder\nPresent Absent\nDiagnostic\ntest result\nPositive True\npositive a\nFalse\npositive b\nNegative False\nnegative c\nTrue\nnegative d\nSensitivity a/(a+c)\nSpecificity d/(b+d)\nPositive predictive value a/(a+b)\nNegative predictive value d/(c+d)\nPrevalence (pre-test probability) (a+c)/(a+b+c+d)\nPositive likelihood ratio (LR+) sensitivity/(1-specificity)\nNegative likelihood ratio (LR −) (1–sensitivity)/specificity\n408 Gynecol Surg (2010) 7:407 –415\n\nresults online through the Centre for Evidence Based Medicine\nwww.cebm.net. Without the nomogram, the post-test proba-\nbility of disease can be calculated in the following way:\nPre /C0 test odds ¼ pre /C0 test probability= 1 /C0 pre /C0 test probabilityðÞ\nPost /C0 test odds ¼ pre /C0 test odds»LR\nPost /C0 test probability ¼ pre /C0 test odds= post /C0 test odds þ 1ðÞ\nð1Þ\nCritical appraisal of diagnostic test studies\nWhen critically appraising diagnostic studies, there are a\nnumber of criteria that determine whether or not there is a risk\nof bias affecting the results of the study [ 5, 6]. With a\nsignificant risk of bias, it is unlikely that the results of the\nstudy are accurate. The following points should be consid-\nered when reading studies of diagnostic test accuracy that\nrelate to our clinical scenario:\n& Were the diagnostic tests (MRI, TVS, TRS) conducted\non all women with clinical signs and symptoms of deep\ninfiltrating endometriosis or were patients randomly\nselected to be in the study? It is important that\nrecruitment of people to a diagnostic study is either\ndone consecutively or randomly so that there is no bias\nin the sample. In our clinical scenario, if patients are\nrecruited only on the basis of a high suspicion of deep\ninfiltrating disease, the tests (MRI, TVS, TRS) are not\nbeing rigorously examined. Moreover, the number of\nnegative test results (both true and false negative) will\nbe very low, leading to wide confidence intervals and\nimprecise results which are difficult to interpret.\n& Did all participants receive both the tests being studied\n(MRI, TVS, TRS) and the gold standard test which the\nstudy tests are being compared to? For our clinical\nscenario, we would want the gold standard test to be\nsurgical diagnosis with histology.\n& If, as often happens when the gold standard test is\nsurgical, a significant proportion of patients do not have\na laparoscopy to confirm the presence or absence of\nendometriosis, the true negatives in the MRI, TVS and\nTRS will be overestimated and the false negatives will\nbe underestimated.\n& Were the investigators who performed the study test\nblinded to the clinical information? In our clinical\nscenario, if the investigator became aware that a patient\nhad all the typical signs of deep endometriosis, s/he\nmight be more vigilant in looking for evidence of\ndisease when carrying out the MRI/TVS/TRS.\n& Was there independent blind comparison between the\nstudy test results and the gold standard? In our clinical\nscenario, if the surgeon was aware of the results of the\nMRI/TVS/TRS, s/he might be more or less vigilant in\nlooking for evidence of the disease.\n& Was the sample size sufficiently large enough to\nproduce precise estimates of the sensitivity, specificity\nand LRs? Diagnostic test studies require sample size\ncalculations, generally based on the expected sensitivity\nand specificity. Unfortunately, few studies are based on\nLikelihood Ratio\n1 2 5 10 0.2 0.1 0.5 0 ∞\nExcludes disease Confirms disease \n0 +30%+15% +45% -15%-30% -45% \nApproximate % change to pre-test probability Fig. 1 Interpretation of likeli-\nhood ratios [ 3, 4]\nX\nA\nB\nFig. 2 Likelihood nomogram (adapted from Centre for Evidence Based\nMedicine www.cebm.net). For example, if our test has a LR+ of 35 and\nLR− of 0.1 and our patient ’s pre-test probability of 50% ( X), a positive\ntest will lead to a post-test probability (A) of approximately 95.5% and a\nnegative test a post-test probability ( B) of approximately 9%\nGynecol Surg (2010) 7:407 –415 409\n\na sample size calculation, resulting in small, underpow-\nered studies, with wide confidence intervals.\n& Are the results of the study applicable to my patients?\nIt is important to consider whether similar results\nwould be achieved if applied to the patients in your\nhospital. In our clinical scenario, we would want to\nlook at the both the test equipment being used in the\nresearch and the experience of the person carrying out\nthe MRI/TVS/TRS. If the research study used imaging\nequipment that was not available in your own hospital,\nyou may expect to get less reliable results in your own\npatients. On the whole, published studies draw on the\nwork of clinicians with a specialist interest in the\ndiagnostic test, and therefore, it may be that the person\nconducting the MRI/TVS/TRS is highly skilled in\nperforming these tests.\nThis would mean that when performed in your own\nhospital, these tests may not perform as well as they did in\nthe research study. Moreover, clinicians with an interest in\nthe use of ultrasound for the diagnosis of endometriosis are\nmore likely to identify disease than those without experi-\nence in the use of ultrasound for this condition; gynaecol-\nogists specialising in laparoscopic surgery for deep\ninfiltrating disease are more likely to identify disease than\ngeneral gynaecologists and those surgeons are also more\nlikely to identify disease during digital examination. The\ngeneralisability of the results of diagnostic test studies,\ntherefore, will very much depend on whether the environ-\nment in which the patient is being treated is similar in terms\nof clinical experience with the disease to that in which the\nstudy was conducted.\nWhat question do you want to ask of the literature?\nIt is likely that this lady has deep infiltrating endometriosis\nand in order for her surgery to be appropriately planned\nwith respect to any rectal or recto-sigmoid involvement, it\nwould be useful to have some further information\nregarding the location of any endometriotic nodules that\nare present. The question we would want to ask of the\nliterature is:\nIn women with typical clinical signs and symptoms of\ndeep endometriosis, what is the accuracy of MRI,\nTVS and TRS for the detection of and staging of deep\ninfiltrating endometriosis in the rectum and recto-\nsigmoid colon, when compared to laparoscopy and\nhistology?\nIn diagnostic test studies, the comparison, or reference,\ntest would be that which is considered to be the gold\nstandard. For the diagnosis of deep infiltrating endometri-\nosis, the current gold standard would be the definitive\nsurgical laparoscopy when the full extent of the disease will\nhave been dissected out and endometriosis confirmed by\nhistology.\nDiagnostic laparoscopy relies on visual inspection and\npalpation and may under-estimate the extent of the disease.\nConsequently, any estimate of accuracy will depend not\nonly on the skill of the clinician conducting the ultrasound\nexamination or interpreting the MRI or ultrasound images,\nbut also on the skill and experience of the clinician\nundertaking the laparoscopy.\nDeveloping the search strategy and results\nfrom the search strategy\nIn reviewing the literature, it is important to look only at the\nrecent literature as both the quality of images and expertise\nhave improved dramatically over the last 6 or 7 years. For\nthis reason, we have searched the literature from 2004\nonwards bearing in mind the time lapse between the start of\na study and its publication, this strategy would identify\nstudies conducted from 2002.\nA detailed review on sou rces of literature and on\ndeveloping search strategies has been described in our first\ntwo papers in this clinical review series [ 7, 8]. The first type\nof paper we would search for would be for systematic\nreviews of the diagnostic accuracy of MRI, TVS or TRS for\nthe diagnosis of endometriosis. The Cochrane Library now\ninclude systematic reviews with meta-analysis of diagnostic\ntest accuracy and therefore we searched the Cochrane\nLibrary using the term “Endometrio*” and the MeSH term\n‘Endometriosis ’. There were no completed systematic\nreviews of the diagnostic accuracy of MRI or ultrasound\nin the detection of endometriosis.\nWe then searched Medline f or relevant systematic\nreviews 2004 –2009 using the following search terms:\n(Endometriosis (MeSH) and (Exp Ultrasonography\n(MeSH)) or Exp Endosonography (MeSH) or Exp Magnet-\nic Resonance Imaging (MeSH) or Exp V agina/ultrasonog-\nraphy or Exp Rectum/ultrasonography or Exp Anal Canal/\nultrasonography or Endosonograph* or Ultrasound or\nUltrasonography) or Endometriosis/ultrasonography, radio-\nnuclide imaging, radiography. We used a clinical query to\nrefine the search to literature reviews. The clinical queries\nare search filters and are set up to be highly sensitive,\nhighly specific or the best balance of the two. We used the\nclinical query ‘Reviews–Best Balance ’ for our search. The\nsearch produced 85 results none of which were useful in\nanswering our question. The only review containing data on\ndiagnostic accuracy of imaging for endometriosis that\nspecifically reviewed bowel involvement [ 9] was not a\nsystematic literature review, did not present LRs for the\n410 Gynecol Surg (2010) 7:407 –415\n\ntests and the included studies for this section of the review\nwere published between 1991 and 2003.\nA search of Embase using the following terms: (Exp\nEndometriosis and (Exp Transrectal Ultrasonography or\nExp Transvaginal echography or Exp Echography or Exp\nNuclear Magnetic Resonance Imaging or Ultrasound or\nUltrasonography or Sonography or Endosonography)) and\n(Meta-analysis or Systematic Literature Reviews or Re-\nview) limited to 2004 –2009 produced seven results, none of\nwhich were useful in answering our clinical question.\nAs the answer to our question could not be found from a\nsystematic review of the diagnostic test accuracy literature,\nwe searched Medline and Embase using the same search\nterms for primary studies. In Medline, the search was limited\nto 2004 –2009 and we used the clinical query ‘Diagnosis–\nBest Balance ’ to identify studies of diagnostic accuracy. In\nEmbase, the search was combined with the following terms:\n(Sensitivity or Specificity or Diagnostic Accuracy).\nDescription of selected papers\nThe search of Medline produced 80 hits and the search of\nEmbase produced 494 hits. After an initial screen of the\ntitles and abstracts of the hits identified by the search, 29\npapers were considered to be potentially useful and, of\nthose, nine reported on the accuracy of imaging for rectal or\nrecto-sigmoid involvement in deep infiltrating endometri-\nosis [ 10–18], were deemed to be of sufficiently high\nmethodological quality (as defined above) and provided\ndata that could be used to construct 2×2 tables to calculate\nsensitivity, specificity, LR+ and LR − with 95% confidence\nintervals or reported these outcomes with confidence\nintervals within the paper. One further paper was identified\nfrom a publication alert after completion of the search [ 19].\nFour studies reported the diagnostic accuracy of MRI [ 10,\n12, 13, 19]; five studies reported the diagnostic accuracy of\nTRS [ 11, 13, 14, 18, 19] and seven studies reported on the\ndiagnostic accuracy of TVS [ 10, 11, 15–19]. All of the\npapers investigated the accuracy of TVS, TRS or MRI for\ndetecting rectal and/or recto-sigmoid endometriosis, al-\nthough many of the papers used these two terms inter-\nchangeably. Seven of the papers reported data specifically\nfor recto-sigmoid disease and six of the papers reported\ndata specifically for rectal disease. One paper reported data\nfor ‘intestinal endometriosis ’.\nThe pre-test probability (prevalence) of rectal or recto-\nsigmoid involvement was between 24% and 83% with most\nstudies being between 40% and 60%. All of the studies\ncompared the results of imaging to a surgical diagnosis with\nhistological confirmation of deep infiltrating endometriosis.\nAll but two (Chapron, Bazot 2009) of the studies were\nprospective in design and most used consecutive recruit-\nment of patients and all studies were conducted in specialist\ncentres for pelvic pain with experience in imaging in\ngynaecology. The studies by Bazot [ 11] and Chapron [ 13]\nwere conducted using women referred for treatment of\nknown deep infiltrating endometriosis, Delpy [ 14], Menada\n[17] and Piketty [ 18] included women with clinically\nsuspected deep infiltrating endometriosis and Abrao [ 10],\nChamie [ 12], Guerriero [ 15], Hudelist [ 16], and Bazot\n(2009) [ 19] included all women with clinically suspected\nendometriosis. Table 2 is a summary of the relevant studies\nidentified from the search strategy.\nThe level of blinding in studies was variable but in the\nmajority of studies, clinicians interpreting the results of the\nimaging examinations were aware of a potential diagnosis\nof endometriosis but were blinded to the results of physical\nexamination and any other imaging used. The exception to\nthis was in the study by Hudelist where the combination of\nper-vaginal (PV) examination and TVS was being investi-\ngated. The same examiner conducted the PV examination\nand TVS ultrasound which is possibly the study that used\nmethodology closest to usual clinical practice.\nWe calculated the pre-test probability (prevalence) of\nrectal or recto-sigmoid deep infiltrating endometriosis,\nsensitivity, specificity, pos itive and negative likelihood\nratios and the post-test probability for each study population\nbased on a positive and negative test with 95% confidence\nintervals from data given in each paper (Table 2).\nGenerally, the very low numbers of false-positive and\nfalse-negative results meant that the confidence intervals\nwere wide for the estimates of the LRs.\nTransvaginal ultrasound\nTVS techniques varied between studies but the majority used\ntransvaginal probes operating at 5-9 MHz. Most of the studies\nused TVS alone however, Hudelist et al. [16] combined TVS\nwith PV examination and Menada et al. [ 17]c o m b i n e dw a t e r\ncontrast in the rectum with TVS. Positive LRs for TVS\nindicate that it is a good to excellent test for confirming the\npresence of rectal involvement in deep infiltrating endome-\ntriosis with the higher positive LRs from studies using\ntransvaginal probes operating at 5-9 MHz. Six out of the\neight studies demonstrated extremely high LR+, with four\nstudies showing infinite LR+. TVS can therefore be\nconsidered to be an excellent test for detecting rectal\nendometriosis. Negative LRs were very low, with six out\nof the eight studies reporting a LR − of 0.1 or less. TVS can\ntherefore be considered to be an excellent test for excluding\nthe presence of rectal involvement in deep infiltrating\nendometriosis. The lowest negative LRs were from studies\nusing transvaginal probes operating at 5-9 MHz rather than\nthose operating at lower frequencies. Overall, TVS seems to be\na useful imaging technique to confirm and exclude the presence\nGynecol Surg (2010) 7:407 –415 411\n\nTable 2 Summary of relevant studies\nAuthor\n[reference]\nStudy\nperiod\nStudy\npopulation\nAge TVS TRS MRI Blinding\nAbrao [ 10] 2004-\n2006\n104 Consecutive\nwomen with\nclinically suspected\nendometriosis\nMean 33.8 years\n(SD 6.1)\nHDI 5000 ultrasound\nscanner with 5-9MHz\ntransducer within\n3months before\nsurgery. Rectal\nenema used\nN/A 1.5T scanner with a Torso\nphase array coil.\nContrast agent gadolinium\n0.2mmol/kg. No bowel\npreparation used\nTVS carried out blinded\nto clinical data. MRI\nradiologist blinded\nto clinical data and\nTVS results\nBazot [ 11] 2000 –\n2004\n81 Consecutive women\nreferred for surgical\nmanagement of DIE\nMedian 31.9\nyears\nUltramark HDI 5000 or\nSiemens Elegra\nultrasound machine.\n5-9MHz transducer.\nNo bowel preparation\nused\nOlympus GF UM 20\nEcho endoscope,\n7.5 and 12MHz.\nSonographers informed\nof women ’s clinical\nhistory and symptoms\nbut blinded to physical\nexamination and\nprevious imaging.\nDifferent physicians\nperformed TVS and TRS\nBazot [ 19] 2000 –\n2005\nRetrospective study of\n92 consecutive women\nwith clinically\nsuspected pelvic\nendometriosis\nMedian 31.8\nyears\nUltramark HDI 5000\nor Siements Elegra\nultrasound machine,\n5-9MHz transducer.\nNo bowel preparation\nused\nOlympus GF UM 20\nEcho endoscope,\n7.5 and 12MHz\nprobe\n1.5T scanner. Bowel\npreparation given.\nContrast agent\ngadolinium\nAll examinations conducted\nby different physicians\nwith knowledge of clinical\nhistory and symptoms\nbut blind to results of\nphysical exam and other\nimaging\nChamié [ 12] 2005-\n2007\n92 Women with a\nhistory and clinical\nexamination\nconsistent with\nendometriosis\nMean 33\nyears\nN/A N/A GE Signa 1.5T scanner.\nContrast agent gadolinium.\nNo bowel preparation used\nMRI images interpreted\nindependently by\n2 radiologists blinded\nto patient history\nChapron [ 13] Retrospective study of 81\nconsecutive patients with\nhistologically proven DIE.\nMRI and transrectal\nultrasound given prior\nto planned surgery\nMean 31.0 (SD6.7) N/A Olympus GF-UM20 scope\nultrasound machine\nwith 7.5 and\n12MHz probes\n1.5T Tesla Unit with\na phased-array\ncoil. No contrast\naged used\nPatients already had a\ndiagnosis of DIE but the\nultrasonographer and\nradiologist were blind\nto clinical information\nwhen they interpreted the\nresults of the tests\nDelpy [ 14] 1998 –\n2003\n31 Women with suspected\nrectovaginal\nendometriosis based\non clinical symptoms\nand/or abnormal clinical\nexamination\nMean 31.5\nyears\nN/A (for rectal infiltration) 7.5MHz radial-scanning\nminiprobe (Fujinon)\nfitted with a\ndistal balloon\nN/A Blind to precise clinical\nfindings but with the\nknowledge of suspected\nendometriosis. Surgery\nconducted with full\nknowledge of imaging\nresults\nGuerriero [ 15] 2005-\n2007\n88 Consecutive women\nwith clinically suspected endometriosis\nMean 33\nyears (SD 5)\n1Week before surgery using\nTechnos MPX with\n6.5-7.0MHz transducer.\nPaid special attention to\ntender areas. No rectal enema\nused.\nN/A N/A Not reported\nHudelist [ 16] 2007 –\n2008\n200 Women with\nclinically suspected\nendometriosis\nMedian 33\nyears\nLogic 9 or Accuvix XQ\nultrasound machine\n5-9MHz transducer\ncombined with\nbimanual PV\nexamination within\nPV examination performed\nfirst followed by TVS\nby the same examiner\n412 Gynecol Surg (2010) 7:407 –415\n\nof deep infiltrating endometriosis involving the rectum or\nrecto-sigmoid colon (Table 3). In a clinic population where\nthe pre-test probability of disease is 50%, we would expect the\npost-test probability given a positive test to be between 90%\nand >99%. In the same population, a negative test would reduce\nthe probability of disease to between 5% and <1%.\nTransrectal ultrasound\nThere were five studies [ 11, 13, 14, 18] investigating the\ndiagnostic accuracy of TRS for rectal involvement in deep\ninfiltrating endometriosis. In three of the studies, 7.5 MHz\nand 12 MHz probes were used and in one study only the\n7.5 MHz probe was used [ 14]. Positive LRs were calculated\nfor four studies and for three of them, TRS was a good test\nto confirm the presence of disease [ 11, 13, 19] all of which\nused 7.5 and 12 MHz probes.\nIn the study by Delpy [ 14], the LR+ was lower and this\ncorresponded to the use of a probe operating at a lower\nfrequency (7.5 MHz). The negative LRs generally indicated\nthat TRS is a reasonable test for excluding disease. The\nnegative LRs did not seem to be related to the frequency of\nprobe used or to any other consistent methodological\ndifference between the studies (Table3). In a clinic population\nwhere the pre-test probability of disease is 50%, we would\nexpect the post-test probability of disease given a positive test\nto be between 65% and 93%. In the same population, a\nnegative test would reduce the probability of disease to\nbetween 9% and 3%. Although TRS is a good test for\nexcluding disease and a reasonable test for confirming\ndisease, it does not seem to perform better than TVS, requires\na rectal enema and might be more uncomfortable than TVS.\nMagnetic Resonance Imaging\nTwo of the studies investigating the diagnostic accuracy of\nMRI for rectal involvement in deep infiltrating endometri-\nosis were prospectively conducted and included women\nwith symptoms that lead to a clinical suspicion of\nendometriosis [ 10, 12] whilst the other two were retrospec-\ntive [13, 19]. Chapron et al. [ 13] included only women with\nhistologically proven deep infiltrating endometriosis whilst\nBazot et al. [ 19] included women with clinically suspected\nendometriosis. Three of the studies [ 12, 13, 19] used\ngadolinium as a contrast agent for the MRI whereas Abaro\net al. [ 10] used no contrast agent. The positive LRs ranged\nfrom 12.0 to 41.7 indicating that MRI is a very good to\nexcellent test for confirming the presence of disease. The\nnegative LR from was between 0.1 and 0.2 indicating that\nMRI is a very good test for excluding the presence of deep\ninfiltrating rectal endometriosis (Table 3). In a clinic\npopulation where the pre-test probability of disease is\n50%, we would expect the post-test probability of disease\nTable 2 (continued)\nAuthor\n[reference]\nStudy\nperiod\nStudy\npopulation\nAge TVS TRS MRI Blinding\n2months of surgery.\nNo rectal enema used\nMenada [ 17] 2006 –\n2007\n90 Women with\nclinically suspected\nrectovaginal\nendometriosis\nMedian 32years Siemens Sonoline Antares\nultrasound machine.\n3.6-8.0MHz multifrequency\ntransducer\nGiven on day of surgery.\nBowel preparation\ngiven on the day\nbefore surgery. Saline\nsolution injected into\nthe rectum during\nthe procedure\nN/A N/A TVS carried out\nindependently by 2 ultrasonographers\nwith the knowledge\nof clinically suspected\ndisease by blind to\nany other clinical\ninformation\nPiketty [ 18] 2005 –\n2007\n134 Women with\nclinically suspected DIE\nMean 32.1\nyears (SD 5)\nToshiba ultrasound machine.\n5-9MHz transducer. No\nrectal enema used\nOlympus UM 160 following\nrectal enema. 5, 7.5\nand 12MHz frequencies\nused\nN/A Examiners told DIE was\nsuspected but were not\ngiven information\non clinical findings\nor other imaging\nfindings\nGynecol Surg (2010) 7:407 –415 413\n\nTable 3 Summary of study results: imaging of rectal or recto-sigmoid involvement in deep infiltrating endometriosis\nStudy Pre-test\nprobabilitya\n(n)\nSensitivity%\n(95% CI)\nSpecificity%\n(95% CI)\nLR+\n(95% CI)\nLR−\n(95% CI)\nPost-test probability\nfor study population:\npositive test (%)\nPost-test probability\nfor study population:\nnegative test (%)\nTransvaginal Ultrasound\nAbrao [ 10] 52% (54/104) 98 (95,100) 100 (100,100) ∞ 0.02 (0.00, 0.1) >99 2\nMenada [ 17] TVS 83% (75/90) 56.5 92.5 7.57 0.47 97 60\nMenada [ 17] RWC-TVSb,c 83% (75/90) 95.7 100.0 ∞ 0.04 >99 15\nGuerriero [ 15] 44% (39/88) 67 (55,73) 92 (84, 100) 8.2 (3.1,21.4) 0.4 (0.2,0.6) 87 22\nBazot [ 11] 67% (54/81) 93 (86,100) 100 (100,100) ∞ 0.07 (0.03,0.2) >99 12\nHudelist [ 16] 24% (48/200) 96 (90,100) 98 (96, 100) 48.6 (15.8,149.1) 0.04 (0.01,0.2) 94 1\nPiketty [ 18] 56% (75/134) 91 (84,97) 97 (92,100) 26.3 (6.7, 102.8) 0.1 (0.05, 0.2) 97 11\nBazot [ 19] 68% (63/92) 94 (88,100) 100 (100,100) ∞ 0.06 (0.02, 0.2) >99 11\nTrans-rectal ultrasound\nChapron [ 13] 42% (34/81) 97 (91,100) 89 (81,98) 9.1 (4.0,20.9) 0.03 (0.00, 0.2) 87 2\nDelpy [ 14] 40% (12/40) 92 (76,100) 67 (45,88) 2.8 (1.4,5.0) 0.1 (0.02,0.8) 65 8\nBazot [ 11] 67% (54/81) 89 (81,97) 93 (83,100) 12.0 (3.2,45.7) 0.1 (0.06,0.3) 96 20\nPiketty [ 18] 56% (75/134) 96 (92,100) 100 (100,100) ∞ 0.04 (0.01,0.1) >99 5\nBazot [ 19] 68% (63.92) 89 (81,97) 93 (84,100) 12.9 (3.4,49.2) 0.1 (0.06,0.2) 96 20\nMagnetic Resonance Imaging (MRI)\nChapron [ 13] 42% (34/81) 76 (62,91) 98 (94,100) 35.9 (5.1,252.1) 0.2 (0.1,0.4) 98 15\nAbrao [ 10] 52% (54/104) 83 (73,93) 98 (94,100) 41.7 (6.0,291.1) 0.2 (0.09, 0.3) 98 16\nChamie [ 12] 54% (50/92) 86 (76,96) 93 (85,100) 12.0 (4.0,36.0) 0.2 (0.08, 0.3) 93 15\nBazot [ 19] 68% (63.92) 87 (79,96) 93 (84,100) 12.7 (3.3,48.4) 0,1 (0.07,0.3) 96 23\na Pre-test probability is the prevalence of deep infiltrating endometriosis involving the rectum or rectosigmoid colon.\nb Figures published in the paper given, no data were available to construct a 2×2 table in order to calculate 95% confidence intervals\nc RWC-TVS, water contrast in the rectum combined with TVS\n414 Gynecol Surg (2010) 7:407 –415\n\ngiven a positive test to be between 94% and 97%. In the\nsame population, a negative test would reduce the probability\nof disease to approximately 9-18%. MRI is therefore a useful\ntest to both confirm and exclude the presence of disease,\nalthough it does not appear to perform as well as TVS.\nFrom our three tests, we can conclude that TRS is\nless accurate than TVS at confirming the presence of\ndisease, but is equally accurate at excluding the presence\nof disease. TVS is also more slightly accurate than MRI\nin both confirming and excluding disease. However,\nMRI is significantly more expensive, and for a claustro-\nphobic patient, considerably more traumatic than TVS,\nand it would therefore not seem to be beneficial to\ninclude MRI in the pre-operative assessment for this\nparticular patient.\nHow would you present the evidence to Miss A?\nMiss A should be counselled that although it is likely that\nshe has endometriosis, this will need to be confirmed\nhistologically at surgery. Nevertheless, since she has a high\nprobability of having endometriosis, in order to prepare her\nadequately for surgery, it is important to determine whether\nshe may have the disease in her bowel. She should be told\nthat the most useful test would be a transvaginal ultrasound\ncarried out in a specialist centre by an appropriately trained\nspecialist. If the ultrasound examination suggests that there\nis rectal involvement, she has a high likelihood of requiring\nsome bowel intervention and needs to be counselled\nappropriately. However, she needs to be warned that no\ntests are 100% accurate and therefore, even if she does have\na negative test, there remains a possibility that her bowel is\ninvolved and that she is operated on in a facility that has the\nappropriately trained surgeons to carry out the surgery\nrequired without the need to abandon surgery and re-\narrange it subsequently.\nReferences\n1. Kennedy S, Bergqvist A, Chapron C, D ’Hooghe T, Dunselman G,\nGreb R, Hummeshoj L, Prentice A, Saridogan E (2005) ESHRE\nguidelines for the treatment of endometriosis. Hum Reprod 20\n(10):2698–2704\n2. Royal College of Obstetricians and Gynaecologists Green Top\nGuideline No. 24 (2006) RCOG Press: London\n3. Attia J (2003) Moving beyond sensitivity and specificity: using\nlikelihood ratios to help interpret diagnostic tests. Aust Prescr 26\n(5):111–113\n4. McGee S (2002) Simplifying likelihood ratios. J Gen Intern Med\n17(8):647–650\n5. Jaesche R, Guyatt DH, Sackett DL (1994) Users' guides to the\nmedical literature IV . How to use and article about a diagnostic\ntest. JAMA 271(5):389 –391\n6. Public Health Resource Unit (2006) Critical skills appraisal\nprogramme: diagnostic studies. Critical Skills Appraisal\nProgramme. www.phru.nhs.uk/Pages/PHD/CASP .htm\n7. Wright JT, Nightingale AL, Ballard KD (2007) Evidence-based\ngynecological practice —clinical review 1. Management of ovar-\nian endometriomas. Gynecol Surg 4:275 –280\n8. Khazali S, Dimitriou E, Wright JT (2009) Evidence-based\ngynecological practice —clinical review 2. Surgery for pain in\nendometriosis. What is the evidence? Gynecol Surg 6(4):357 –365\n9. Kinkel K, Frei KA, Balleyguier C, Chapron C (2006)\nDiagnosis of endometriosis with imaging: a review. Eur Radiol\n16:285–298\n10. Abrão MS, Gonçalves MOC, Dias JA Jr, Podgaec S, Chamie LP ,\nBlasbalg R (2007) Comparison between clinical examination,\ntransvaginal sonography and magnetic resonance imaging for the\ndiagnosis of deep endometriosis. Hum Reprod 22(12):3092 –3097\n11. Bazot M, Malzy P , Cortez A, Roseau G, Amouyal P , Darai E\n(2007) Accuracy of transvaginal sonography and rectal endoscop-\nic sonography in the diagnosis of deep infiltrating endometriosis.\nUltrasound Obstet Gynecol 30:994 –1001\n12. Chamié LP , Blasbalg T, Gonçalves MOC, Carvalho FM, Abrão MS,\nde Oliveria IS (2009) Accuracy of magnetic resonance imaging for\ndiagnosis and preoperative assessment of deeply infiltrating endo-\nmetriosis. Int J Gynecol Obstet Suppl 106:198–201\n13. Chapron C, Vieira M, Chopin N, Balleyguier C, Barakat H,\nDumontier I, Roseau G, Fauconnier A, Foulot H, Dousset B\n(2004) Accuracy of transvaginal sonography and rectal endoscop-\nic sonography in the diagnosis of deep infiltrating endometriosis.\nUltrasound Obstet Gynecol 24:175 –179\n14. Delpy R, Marthet M, Gasmi M, Berhah S, Shojai R, Desjeux A,\nBoubli L, Grimaud J-C (2005) V alue of endorectal ultrasonogra-\nphy for diagnosing rectovaginal septal endometriosis infiltrating\nthe rectum. Endoscopy 37(4):357 –361\n15. Guerriero S, Ajossa S, Gerada M, Virgilio B, Angioni S, Melis\nGB (2008) Diagnostic value of transvaginal ‘tenderness-guided’\nultrasonography for the prediction of location of deep endometri-\nosis. Hum Reprod 23(11):2452 –2457\n16. Hudelist G, Oberwinkler KH, Singer CF, Tuttlies F, Rauter G,\nRittrer O, Keckstein J (2009) C ombination of transvaginal\nsonography and clinical examination for preoperative diagnosis\nof pelvic endometriosis. Hum Reprod 24(5):1018 –1024\n17. Menada MV , Remorgida V , Abbamote LH, Nicoletti A, Ragni N,\nFerrero S (2008) Does transvaginal ultrasonography combined\nwith water-contrast in the rectum aid in the diagnosis of\nrectovaginal endometriosis infiltrating the bowel? Hum Reprod\n23(5):1069–1075\n18. Piketty M, Chopin N, Dousett B, Millicher-Bellaishce AE, Roseau\nG, Leconte M, Borghese B, Chapron C (2009) Preoperative work-\nup for patients with deeply infiltrating endometriosis: transvaginal\nultrasonography must definitely be the first-line imaging exami-\nnation. Hum Reprod 24(3):602 –607\n19. Bazot M, Lafont C, Rouzier R, Roseau G, Thomassin-Naggara I,\nDarai E (2009) Diagnostic accuracy of physical examination,\ntransvaginal sonography, rectal endoscopic sonography, and\nmagnetic resonance imaging to diagnose deep infiltrating endo-\nmetriosis. Fertil Steril 92(6):1825 –\n1833\nGynecol Surg (2010) 7:407 –415 415","source_license":"CC0","license_restricted":false}