MRI reporting standard for chronic pelvic pain: consensus development

Aim: To identify radiological parameters that should be reported on gynaecological MRI in order to create a standardized assessment pro forma for reporting CPP, which may be used in clinical practice.

Chronic pelvic pain (CPP) in females is a common problem presenting a major challenge to healthcare providers. The complex multifactorial aetiology requires a multidisciplinary approach and often necessitates diagnostic laparoscopy for assessment. MRI is emerging as a potential non-invasive alternative for evaluation of CPP; however, standardization of reporting is required for it to be used in routine clinical practice. A two-generational Delphi survey with an expert panel of 28 radiologists specializing in gynaecological MRI from across the UK was used to refine a proposed reporting template for CPP.

75% response rate for the first round and 79% for the second. Following the second round, agreement was reached on the structure of the pro forma and the way in which information was sought, with overall consistency of agreement between experts deemed as fair (intraclass correlation coefficient = 0.394). This was accepted as the final version by consensus.

The standardized pro forma developed in this study will form the basis for future prospective evaluation of MRI in CPP. This template could be modified for the assessment of other benign gynaecological conditions. Advances in knowledge: Female CPP is a significant problem presenting challenges for clinicians. MRI is often used for evaluation and standardization of techniques, and reporting is required. The pro forma developed in this study will form the basis for future prospective MRI evaluation.

Chronic pelvic pain (CPP) is defined as cyclical or non-cyclical pain in the lower abdomen or pelvis of at least 6 months' duration, which limits activities of daily living, occurring continuously or intermittently.1,2 In cases where an underlying pathology has been excluded, CPP is considered “idiopathic” and may be defined as pain of uncertain or unknown structural cause that is unrelated to gynaecological organs, after exclusion of other recognizable gynaecological pathology,3 not occurring exclusively with menstruation or intercourse and not associated with pregnancy.4 It is a common and significant problem for females and presents a major challenge to healthcare providers with a prevalence of 3–4% in the UK primary care setting. Hospital episode data estimated the direct cost of healthcare for CPP at £158 million (€188 million, $237 million) with a further £24 million in indirect costs in 1992.5 CPP is the single most common indication for referral to a gynaecology clinic accounting for 20% of all outpatient appointments.6 CPP has a complex multifactorial aetiology and natural history. No effective management policy exists, with only 20–25% of patients responding to conservative treatment.7 Given that the symptoms experienced may be varied and non-specific, the assessment should be multidisciplinary, considering all pelvic organs, the musculoskeletal, neurologic and psychiatric systems. Investigations include detailed blood and urinary tests and infection screening. Laparoscopy, although invasive and expensive, is often used as a diagnostic tool with >40% of laparoscopies in females being performed for CPP. The most common laparoscopic diagnoses made in these females are endometriosis, pelvic adhesions, chronic pelvic inflammatory disease and ovarian cysts.8 However, up to 40% of laparoscopies show no pathology. Laparoscopic visual inspection has deficiencies in its performance. For example, in the diagnosis of endometriosis, a negative laparoscopy can accurately exclude endometriosis; however, a positive laparoscopy requires histological confirmation.9 In recent years, non-invasive non-ionizing radiation-based imaging investigations including detailed transabdominal and transvaginal ultrasound and MRI of the pelvis have proven useful in the evaluation of patients with CPP. MRI provides excellent anatomical information and has a sensitivity of 85% and specificity of 89% for detecting biopsy-proven endometriosis; furthermore, a recent study showed MRI had 100% accuracy for adenomyosis and 98% for endometriosis of the uterosacral ligaments.10–12 The high rate of false-positive pathology on laparoscopy and the improving performance of MRI prompt the question of whether MRI can be used to replace diagnostic laparoscopy for the diagnosis of some conditions that cause CPP. For MRI to be used in routine clinical practice for evaluation of CPP, imaging protocols and reporting need to be standardized to allow homogeneous MRI performance between centres and limit acquisition and interobserver variability. The Radiological Society of North America has suggested report templates for female pelvic MRI, but these have limited information about the location of the ovaries, the presence of tubal pathology, and bladder or bowel pathology, which may be present in conditions such as deep infiltrating endometriosis.13 Currently, there are no standardized reporting guidelines for CPP in the UK.

To develop a clinically relevant reporting pro forma for MRI assessment of CPP using the Delphi technique.

AND MATERIALS As part of the MRI to Establish Diagnosis Against Laparoscopy (MEDAL) trial, a multicentre study in the UK assessing females with unexplained CPP, an MRI assessment pro forma template was developed by the study management group and an independent radiology review committee based on their own expertise and augmented by the literature on female pelvic MRI (Appendix A). The pro forma was then reviewed by an expert panel of consultant radiologists from radiology departments across the UK who specialize in reporting gynaecological MRI. The panel reviewed the MRI assessment pro forma using the Delphi technique in conjunction with a web-based survey tool. In using the Delphi technique, it was hoped that the expert opinion would eventually overlap and achieve consensus. The Delphi technique To solicit expert opinion, a two-generational Delphi survey was used to refine the initial template report and develop a standardized MRI assessment pro forma. The Delphi method is a structured communication technique that allows one to harness and organize judgment from a panel of experts in order to achieve a consensus opinion. Features of the Delphi technique are the use of questionnaire rounds, feedback responses, opportunity for participants to modify their responses and maintenance of anonymity of responses.14 The process overcomes geographical issues and facilitates a balanced, equal opinion across experts to determine the outcome.15 We employed two rounds of the Delphi technique via a web-based survey tool to reach the final iteration of the pro forma. Expert panel To ensure that a breadth of pelvic MRI experience informed the development of the MRI pro forma, 28 consultant radiologists from various UK institutions involved in the MEDAL study were approached to form an expert panel. All the panel members had first-hand experience of using the initial MRI assessment pro forma in a clinical setting. As all the consultants were involved in the index UK-based study, no international experts were involved. The mean experience of our expert panel was 17 years, with 9.5 years reporting gynaecological MRI at a consultant level. In each round of the survey, the expert panel members scored each section of the MRI assessment pro forma using a zero (strongly disagree) to nine (strongly agree) certainty scale and free text within a web-based survey tool. The web-based survey tool was sent by email, asking the expert panel to comment specifically on: (a) the appropriateness of each section of the original MRI assessment pro forma. (b) suggested improvements for recording of imaging findings. (c) parts of the report template they felt could be omitted. (d) suggestions for any other improvements. The scores and free text feedback from the expert panel were then used by the investigators to further develop the MRI pro forma template. Anonymity was maintained between expert members and investigators. Two generic reminder emails were sent to all members of the expert panel in order to capture the maximum number of respondents. First round A 75% response rate (21 of 28 panel members) to the questionnaire was considered adequate to allow analysis of that round. Following the first round, most of the comments from the panelists were regarding the length of pro forma, difficulties with questioning style and difficulties in applying exact methodology, e.g. measuring ovarian volume, as detailed guidance notes were not provided. The suggested changes were used to modify the proposed pro forma. Second round A revised pro forma (Appendix B) was then distributed to the same expert panel who were asked to reconsider each section for appropriateness and validity. The sections were again scored on a ten-point scale, and a free text box was provided for comments. In addition, the radiologists were asked if they had responded to Round 1 of the survey. Anonymity was maintained between respondents for the second round, however, the investigators were aware of which expert panel members responded to the questionnaire in this round, allowing targeted reminders to non-respondents. One generic reminder email was sent to all panel members followed by further individual reminders to non-respondents. Ethics approval The National Institute for Health Research (NIHR) MEDAL study has ethical approval from the National Research Ethics Service (NRES) Committee East Midlands—Nottingham 1 (Ref 11/EM/0281). Data analyses Responses obtained from the electronic surveys were analysed in the same way for both the first and second round. Scores of zero to four were considered to indicate disagreement with the way information was sought on the pro forma, requiring a change in approach. A score of five was considered acceptable with minimal change required, and scores of six to nine were considered to indicate agreement with the pro forma with no change required. All free text comments arising in the first round were documented and considered in formulation of the revised pro forma. Intraexpert agreement, after the second round of the Delphi survey, was assessed by computing the intraclass correlation coefficient (ICC), assuming a two-way random effect model (items and experts). Landis criteria were used to interpret ICC agreement values: slight (ICC = 0.00–0.19); fair (ICC = 0.20–0.39); moderate (ICC = 0.40–0.59); substantial (ICC = 0.60–0.79); and almost perfect (ICC = 0.80–1.0) agreement.16 End point The end point of the study was considered to be reached when ≥90% of respondents were in agreement with each section of the pro forma (i.e. neutral or accepted, no change).

First round 21 of 28 experts (75%) responded to the questionnaire in the first round. In addition to scoring each section of the pro forma (Tables 1 and 2), 17 of the 21 responding experts (81%) provided free text comments in at least one of the six sections. In total, 63 free text comments were made out of a total of 126 opportunities (50%). Table 1. | Question | n | Strongly disagree | Strongly agree | ||||||||| |---|---|---|---|---|---|---|---|---|---|---|---|---| | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | ||| | 1 | Part A: Sequences performed—Do you agree with how the information is sought? | 21 | 0% | 5% | 5% | 0% | 0% | 24% | 14% | 38% | 14% | 0% | | 2 | Part B: Uterus—Do you agree with how the information is sought? | 21 | 0% | 10% | 0% | 0% | 0% | 10% | 0% | 52% | 29% | 0% | | 3 | Part C: Ovaries—Do you agree with how the information is sought? | 21 | 0% | 5% | 0% | 0% | 5% | 5% | 10% | 43% | 19% | 14% | | 4 | Part D: Other observations 1—Do you agree with how the information is sought? | 21 | 0% | 5% | 5% | 0% | 0% | 10% | 10% | 43% | 14% | 14% | | 5 | Part E: Other observations 2—Do you agree with how the information is sought? | 21 | 0% | 5% | 0% | 0% | 0% | 5% | 14% | 43% | 19% | 14% | | 6 | Part F: Summary diagnosis—Do you agree with how the information is sought? | 21 | 0% | 5% | 0% | 0% | 0% | 10% | 10% | 29% | 24% | 24% | Table 2. | Question | Score | ||| |---|---|---|---|---| | 6 | || | Needs change | Neutral | Accepted, no change | || | 1 | Part A: Sequences performed—Do you agree with how the information is sought? | 10% | 24% | 67% | | 2 | Part B: Uterus—Do you agree with how the information is sought? | 10% | 10% | 81% | | 3 | Part C: Ovaries—Do you agree with how the information is sought? | 10% | 5% | 86% | | 4 | Part D: Other observations 1—Do you agree with how the information is sought? | 10% | 10% | 81% | | 5 | Part E: Other observations 2—Do you agree with how the information is sought? | 5% | 5% | 91% | | 6 | Part F: Summary diagnosis—Do you agree with how the information is sought? | 5% | 10% | 86% | Scores of 0–4 were considered to indicate disagreement with the way information was sought on the pro forma, requiring a change in approach. A score of five was considered acceptable with minimal change required and scores of 6–9 were considered to indicate agreement with the pro forma with no change required. The scores and comments obtained from the first round were considered in formulation of the revised CPP imaging pro forma (Appendix B). Second round 22 of 28 experts (79%) responded to the questionnaire in the second round, 4 of whom (18%) had not responded in Round 1. Scores for each section of the revised pro forma are given in Tables 3 and 4. Only one panelist scored the revised pro forma below four in one section, and the median score was seven for each of the five sections. The panelist giving Part A a score of three did not provide a reason in the corresponding free text box and the remainder of the responses to this section were in agreement with positive free text comments including “much better, short and specific comment about image quality”. Table 3. | Question | n | Strongly disagree | Strongly agree | ||||||||| |---|---|---|---|---|---|---|---|---|---|---|---|---| | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | ||| | 1 | Part A: Image quality—Do you agree with how the information is sought? | 22 | 0% | 0% | 0% | 5% | 0% | 14% | 27% | 27% | 18% | 9% | | 2 | Part B: Uterus—Do you agree with how the information is sought? | 22 | 0% | 0% | 0% | 0% | 0% | 18% | 18% | 27% | 23% | 14% | | 3 | Part C: Ovaries—Do you agree with how the information is sought? | 22 | 0% | 0% | 0% | 0% | 0% | 27% | 9% | 27% | 27% | 9% | | 4 | Part D: Other observations—Do you agree with how the information is sought? | 22 | 0% | 0% | 0% | 0% | 5% | 23% | 14% | 27% | 27% | 5% | | 5 | Part E: Summary diagnosis—Do you agree with how the information is sought? | 22 | 0% | 0% | 0% | 0% | 0% | 18% | 9% | 27% | 27% | 18% | Table 4. | Question | Score | ||| |---|---|---|---|---| | 6 | || | Needs change | Neutral | Accepted, no change | || | 1 | Part A: Sequences performed—Do you agree with how the information is sought? | 5% | 14% | 82% | | 2 | Part B: Uterus—Do you agree with how the information is sought? | 0% | 18% | 82% | | 3 | Part C: Ovaries—Do you agree with how the information is sought? | 0% | 27% | 73% | | 4 | Part D: Other observations 1—Do you agree with how the information is sought? | 5% | 23% | 73% | | 5 | Part E: Other observations 2—Do you agree with how the information is sought? | 0% | 18% | 82% | Scores of 0–4 were considered to indicate disagreement with the way information was sought on the pro forma, requiring a change in approach. A score of five was considered acceptable with minimal change required and scores of 6–9 were considered to indicate agreement with the pro forma with no change required. Once again, the opportunity for free text response was provided following scoring of each section of the pro forma. Overall consistency of agreement between experts after Round 2 of Delphi survey was deemed as fair (ICC = 0.394). In the second round, a total of 32 free text comments were made out of a possible 110 opportunities (29%). Of these, seven comments specifically mentioned that the changes had improved the section. Seven comments related to the questioning style, five felt the pro forma remained too long, while three suggested more questions. Of note, three individuals felt that more guidance should be provided for the completion of Section C of the revised pro forma relating to ovarian evaluation. There were seven responses to the final question of the survey asking for further comments and/or suggestions. Three respondents felt that the pro forma was “comprehensive” and “well constructed”. One individual pointed out that the pro forma was “excellent for the purposes of research” but that most of the measurements required would not be obtained in normal clinical practice. Only one panelist scored the revised pro forma below four in one section, and the median score was seven for each of the five sections. By consensus, this was accepted as the final round and the proposed MRI assessment pro forma as the final version. The final MRI assessment pro forma is shown in Appendix B.

Over the past decade, pelvic MRI has emerged as an invaluable tool in the clinical assessment of females with gynaecological malignancies, benign gynaecological conditions such as fibroids and increasingly in CPP. For these applications, the performance of MRI differs in the literature because of variations in image acquisition, image interpretation and how findings are conveyed to clinicians within the MRI report. The clinical report provided for an imaging study needs to communicate both positive and relevant negative findings in a format that is easily accessible to the referring clinician. Studies have shown that 90% of clinicians and 60% of radiologists prefer structured, itemized reports rather than reports in prose.17 Reports structured in a thorough, yet standardized way help to minimize interpretation error and promote better patient care. Our study aimed to set standards for MRI reporting in patients with unexplained CPP. The ideal method for setting standards involves a comprehensive and independent process that encourages broad participation of expert panel members; objectively considers all opinions; and is subject to critical and iterative comment. In this study, a standardized MRI reporting pro forma was developed based on a review performed by 28 UK radiology consultants using a two-stage Delphi consensus technique. The use of a web-based questionnaire allowed the expert panel the opportunity to comment on style, length and ease of use of the pro forma for routine clinical practice. The merit and validity of the Delphi method has been called in to question;18 however, it is an important tool in situations where no conclusive evidence is available and offers the opportunity to pool knowledge and insights of experts in a field. We acknowledge the possibility that a panel composed of a different group of experts may identify a different end result. However, this exercise has refined the initially proposed pro forma and we believe can act as a strong basis for further development. The strengths of this study were use of a large number of radiology consultants with expertize reporting gynaecological MRI, and who were familiar with the causes of CPP. Pooling of expert opinions allowed creation of a refined pro forma with only two-stages of Delphi survey. It is acknowledged that using a broader group of radiologists may have minimized any bias and avoided a narrow clinical viewpoint but would also have created less agreement and the need for several more stages to the Delphi survey. Asking general gynaecologists to evaluate this reporting pro forma may have been beneficial to assess its value, but as the purpose of the study was to standardize an MRI assessment tool from a radiological point, this was not performed. Unfortunately, agreement could only be calculated on the second round of the Delphi survey since the responses from the first round were anonymous. Additional information that may be useful to a finalized reporting tool would be recording of the last menstrual period in order to calculate the time in the menstrual cycle when MRI was performed. This would allow correlation of the endometrial and junctional zone volume in the follicular and luteal phases.19 MRI was usually performed without intravenous contrast medium to reflect standard clinical practice. Our proposed reporting pro forma for MRI in CPP provides a clear consistent template but also the opportunity for the reporting radiologist to provide additional comment and interpretation, taking into account the clinical and radiological context of a study. This structured template also allows quick and easy comparison between studies for an individual patient and also within a specified patient population, allowing long-term clinical and surgical validation. The proposed pro forma could also be modified to allow evaluation for other benign gynaecological indications.

The Delphi technique helped establish the consensus-based priorities for an MRI assessment pro forma and may help guide Radiologists and Gynaecologists when reviewing patients with CPP symptoms. Communication via a web-based survey tool and email was a cost-effective, time-saving exercise that overcame geographical boundaries. The present study may form part of the evidence base for future work in developing a standardized MRI assessment pro forma as part of the management of females with CPP. CONFLICTS OF INTEREST All of the authorship team, excluding N Bharwani and J Zamora, are members of the MEDAL study management committee. This project will be published in full by the NIHR Health Technology Assessment (HTA). The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the [Health Technology Assessment (HTA) Programme], NIHR, NHS or the Department of Health. Acknowledgments ACKNOWLEDGMENTS Consultant radiologists involved in the evaluation of the standardized reporting pro forma are all investigators for the MEDAL study. The MEDAL study (http://www.controlled-trials.com/ISRCTN13028601/medal) aims to compare pelvic MRI against laparoscopy to establish a diagnosis in CPP. We are grateful to the MEDAL study management group and the consultants who participated in the qualitative survey and consensus. Appendix A Appendix B Contributor Information Nishat Bharwani, Email: [email protected], [email protected]. Seema A Tirlapur, Email: [email protected]. Moji Balogun, Email: [email protected]. Lee Priest, Email: [email protected]. Khalid S Khan, Email: [email protected]. Javier Zamora, Email: [email protected]. Anju Sahdev, Email: [email protected]. FUNDING This project was funded by the National Institute for Health Research [Health Technology Assessment (HTA) Programme] (project number 09/22/50).

- 1.Latthe P, Latthe M, Say L, Gülmezoglu M, Khan KS. WHO systematic review of prevalence of chronic pelvic pain: a neglected reproductive health morbidity. BMC Public Health 2006; 6: 177. doi: 10.1186/1471-2458-6-177 [DOI] [PMC free article] [PubMed] [Google Scholar] - 2.Daniels JP, Khan KS. Chronic pelvic pain in women. BMJ 2010; 341: c4834. doi: 10.1136/bmj.c4834 [DOI] [PubMed] [Google Scholar] - 3.Tirlapur SA, Priest L, Daniels JP, Khan KS; MEDAL Study Management Group. How do we define the term idiopathic? Curr Opin Obstet Gynecol 2013; 25: 468–73. doi: 10.1097/GCO.0000000000000025 [DOI] [PubMed] [Google Scholar] - 4.Royal College of Obstetricians and Gynaecologists. Initial management of chronic pelvic pain. London, UK: RCOG Press; 2012. pp. 1–16. [Google Scholar] - 5.Aslam N, Harrison G, Khan K, Patwardhan S. Visceral hyperalgesia in chronic pelvic pain. BJOG 2009; 116: 1551–5. doi: 10.1111/j.1471-0528.2009.02305.x [DOI] [PubMed] [Google Scholar] - 6.Howard FM. The role of laparoscopy in chronic pelvic pain: promise and pitfalls. Obstet Gynecol Surv 1993; 48: 357–87. doi: 10.1097/00006254-199306000-00001 [DOI] [PubMed] [Google Scholar] - 7.Henzl MR. Dysmenorrhoea: achievements and challenge. Sex Med Today 1985; 9: 8–12. [Google Scholar] - 8.Howard FM. The role of laparoscopy as a diagnostic tool in chronic pelvic pain. Baillieres Best Pract Res Clin Obstet Gynaecol 2000; 14: 467–94. doi: 10.1053/beog.1999.0086 [DOI] [PubMed] [Google Scholar] - 9.Wykes CB, Clark TJ, Khan KS. Accuracy of laparoscopy in the diagnosis of endometriosis: a systematic quantitative review. BJOG 2004; 111: 1204–12. doi: 10.1111/j.1471-0528.2004.00433.x [DOI] [PubMed] [Google Scholar] - 10.Bazot M, Bornier C, Dubernard G, Roseau G, Cortez A, Daraï E. Accuracy of magnetic resonance imaging and rectal endoscopic sonography for the prediction of location of deep pelvic endometriosis. Hum Reprod 2007; 22: 1457–63. doi: 10.1093/humrep/dem008 [DOI] [PubMed] [Google Scholar] - 11.Stratton P, Winkel C, Premkumar A, Chow C, Wilson J, Hearns-Stokes R, et al. Diagnostic accuracy of laparoscopy, magnetic resonance imaging, and histopathologic examination for the detection of endometriosis. Fertil Steril 2003; 79: 1078–85. doi: 10.1016/S0015-0282(03)00155-9 [DOI] [PubMed] [Google Scholar] - 12.Di Paola V, Manfredi R, Castelli F, Negrelli R, Mehrabi S, Pozzi Mucelli R. Detection and localization of deep endometriosis by means of MRI and correlation with the ENZIAN score. Eur J Radiol 2015; 84: 568–74. doi: 10.1016/j.ejrad.2014.12.017 [DOI] [PubMed] [Google Scholar] - 13.RSNA Radiology Reporting Templates [Internet]. [Updated 10 December 2012; accessed 4 November 2015.] Available from: http://www.radreport.org/template/0000259 - 14.McKenna HP. The Delphi technique: a worthwhile research approach for nursing? J Adv Nurs 1994; 19: 1221–5. doi: 10.1111/j.1365-2648.1994.tb01207.x [DOI] [PubMed] [Google Scholar] - 15.Fink A, Kosecoff J, Chassin M, Brook RH. Consensus methods: characteristics and guidelines for use. Am J Public Health 1984; 74: 979–83. doi: 10.2105/AJPH.74.9.979 [DOI] [PMC free article] [PubMed] [Google Scholar] - 16.Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33: 159–74. doi: 10.2307/2529310 [DOI] [PubMed] [Google Scholar] - 17.Naik SS, Hanbridge A, Wilson SR. Radiology reports: examining radiologist and clinician preferences regarding style and content. AJR Am J Roentgenol 2001; 176: 591–8. doi: 10.2214/ajr.176.3.1760591 [DOI] [PubMed] [Google Scholar] - 18.Sackman H. Delphi critique. Boston, MA: Lexington Books; 1975. [Google Scholar] - 19.Hoad CL, Raine-Fenning NJ, Fulford J, Campbell BK, Johnson IR, Gowland PA. Uterine tissue development in healthy women during the normal menstrual cycle and investigations with magnetic resonance imaging. Am J Obstet Gynecol 2005; 192: 648–54. doi: 10.1016/j.ajog.2004.07.032 [DOI] [PubMed] [Google Scholar]