{"paper_id":"9f934bd6-b718-4168-a0fb-cffac98c8d9e","body_text":"Magnetic resonance imaging (MRI) is an increasingly popular imaging method in medical\npractice. It allows the acquisition of multiplanar images, with high resolution,\nwithout exposure to radiation, and offers the option of using a paramagnetic\ncontrast agent (gadolinium). In oncology, MRI can provide morphological information\nsuch as size, contours, number of lesions, edema, necrosis, relationship to adjacent\nstructures, physiological alterations, and cellular metabolism, allowing a more\ncomplete evaluation in terms of disease distribution and activity. MRI has been\nincreasingly indicated for the management of cancer patients in Brazil ( 1 - 11 ) .\nMRI applied to gynecology offers additional information on the anatomy of the female\npelvis in comparison with other imaging modalities, such as ultrasound and computed\ntomography (CT). Therefore, MRI is the method of choice for the diagnosis and\nstaging, as well as for the evaluation of the treatment response and the detection\nof relapse after treatment, of gynecologic neoplasms. Non-oncologic indications for\nMRI of the female pelvis include inconclusive pelvic ultrasound examination,\nevaluation of postoperative complications, pelvic pain, malformations of the vagina\nor uterus, and pelvic floor defect ( 12 - 17 ) . The American\nCollege of Radiology (ACR) compiled these indications into an educational guide and\nguideline for clinical decision-making in medical practice ( 18 ) .\nTo our knowledge, there have been no nationwide surveys on the main indications for\nMRI of the female pelvis in Brazil. Therefore, it is essential to describe the\nexperience of a national referral center for cancer in the use of pelvic MRI for the\nmanagement of patients with gynecologic tumors in order to design future research\nprojects, develop more coherent flowcharts, and devise study protocols that are more\nfocused on specific indications.\nThe objective of this study was to evaluate the indications for MRI of the female\npelvis at a referral center for cancer, in comparison with the criteria proposed by\nthe ACR.\n\nThis was a retrospective, single-center, descriptive study, carried out through a\nreview of medical charts and imaging reports. The study was approved by the research\nethics committee of the institution. The imaging department of the institution\nconducts approximately 1400 examinations per month. Of those, approximately 10% are\nMRI studies of the female pelvis. We evaluated female patients who underwent MRI of\nthe pelvis between January 2013 and June 2014 at a referral center for cancer. Cases\nin which the physician order did not list an indication for the examination were\nexcluded, as were those for which there was no electronic medical record available\nat the institution.\nFor all of the patients included, an electronic questionnaire was completed. The\nquestionnaire was designed to collect data regarding demographic characteristics\n(age and gender), the MRI protocol (intravenous or intravaginal administration of\ncontrast medium), and clinical status, as well as a detailed description of the\nindication for the examination. In patients who had a confirmed diagnosis or were\nunder high clinical suspicion of malignant neoplasm, the examinations were\nclassified as being indicated for oncological purposes (oncologic indication). The\nexaminations in patients who had no such previous diagnosis and presented low\nclinical suspicion for malignancy, the examinations were classified as being\nindicated for non-oncological purposes (non-oncologic indication). In cases with an\noncologic indication, the primary tumor site and the reason for the examination\n(diagnosis, staging, response evaluation, or posttreatment follow-up) were\nevaluated.\nThe indications for the examination were further divided, according to the ACR\ncriteria ( 18 ) , into the\nfollowing groups: detection and staging of gynecologic neoplasms; evaluation of\npelvic pain or a pelvic mass; identification of congenital anomalies; determination\nof the number, location, and type of fibroids; detection of pelvic floor defects;\ndetection and staging of other nongynecologic pelvic tumors; assessment of the\nrecurrence of pelvic tumors; evaluation of postoperative complications;\ndetermination of arterial or venous anatomy and patency; identification and staging\nof soft tissue sarcomas; identification of the source of lower abdominal pain in\npregnant women; assessment of fetal or placental abnormalities; identification of\ninflammatory bowel disease and its complications; and planning of guidance for\nminimally invasive surgery and brachytherapy.\nThe information collected via the electronic questionnaire was exported to a\nMicrosoft Excel-based database. Data were processed with the SPSS Statistics\nsoftware package, version 20.0 (IBM Corp.; Armonk, NY, USA). We calculated\ndescriptive statistics, adopting the usual measures of central tendency and\ndispersion for numerical variables, as well as absolute and relative frequencies for\ncategorical variables.\n\nA total of 1060 MRI scans of the female pelvis were included during the study period.\nThe mean age of the patients was 52.6 ± 14.8 years (range, 8-90 years). Of\nthe 1060 patients, 693 (65.4%) were perimenopausal or postmenopausal. A history of\ncancer was noted in 678 patients (63.9%). The cancer was gynecologic in 313 patients\n(29.5%) and nongynecologic in 365 (34.4%).\nThe most common specialties of the requesting physicians were gynecology and\nobstetrics ( n  = 418; 39.4%); clinical oncology ( n \n= 313; 29.5%); urology ( n  = 58; 5.5%); abdominal surgery\n( n  = 57; 5.4%); and pelvic surgery ( n  = 18;\n1.7%). The remaining 196 MRIs (18.6%) were requested by physicians from other\nspecialties.\nOf the 1060 patients evaluated, 594 (56.0%) were asymptomatic at the time of the MRI.\nTherefore, 466 patients (44.0%) had clinical complaints. The most commonly reported\nsymptoms were pelvic pain, in 122 (11.5%), and bleeding, in 104 (9.8%). Pelvic\nultrasound prior to MRI was reported in 425 patients (40.1%), and the ultrasound had\nrevealed alterations in 368 (34.7%), the most common finding being adnexal mass,\nwhich had been observed in 153 (40.8%). Intravenous contrast was used in almost all\nof the examinations (92.8%), whereas intravaginal contrast was used in only\n4.4%.\nOf the 1060 patients evaluated, 813 (76.7%) presented an appropriate indication for\nMRI, according to the ACR criteria. The main indications observed were evaluation of\ntumor recurrence after resection, in 275 (25.9%), as shown in  Figure 1 ; detection and staging of gynecologic neoplasms, in 247\n(23.3%), as shown in  Figure 2 ; and evaluation\nof pelvic pain or a pelvic mass, in 181 (17.1%), as shown in  Figure 3 .  Table 1 \ndescribes the frequency of all indications according to the ACR criteria.\nFigure 1 MRI scan requested for the assessment of pelvic tumor recurrence in a\npatient with cervical cancer after cone biopsy. T2-weighted sagittal\nsection showing the uterus in anteversion, midline, with signs of\nsurgical manipulation of the colon, and a diffuse reduction of T2 signal\nintensity.\nMRI scan requested for the assessment of pelvic tumor recurrence in a\npatient with cervical cancer after cone biopsy. T2-weighted sagittal\nsection showing the uterus in anteversion, midline, with signs of\nsurgical manipulation of the colon, and a diffuse reduction of T2 signal\nintensity.\nFigure 2 MRI scan requested for the detection/staging of a gynecologic neoplasm\n(cervical mass).  A:  Post-contrast T1-weighted sagittal\nsection showing an irregular infiltrative lesion restricted to the\ncervix.  B:  Axial diffusion-weighted image showing diffusion\nrestriction.\nMRI scan requested for the detection/staging of a gynecologic neoplasm\n(cervical mass).  A:  Post-contrast T1-weighted sagittal\nsection showing an irregular infiltrative lesion restricted to the\ncervix.  B:  Axial diffusion-weighted image showing diffusion\nrestriction.\nFigure 3 MRI scan requested for the diagnosis of an adnexal mass.\nContrast-enhanced coronal T2-weighted slice ( A ) and\nsagittal T1-weighted slice ( B ), showing a heterogeneous\ninfiltrative lesion occupying the endometrial cavity, with invasion to\n> 50% of the myometrial thickness (thick arrow) and solid-cystic mass\nin the left ovary (thin arrow), consistent with endometrial\nadenocarcinoma with metastasis to an ovary.\nMRI scan requested for the diagnosis of an adnexal mass.\nContrast-enhanced coronal T2-weighted slice ( A ) and\nsagittal T1-weighted slice ( B ), showing a heterogeneous\ninfiltrative lesion occupying the endometrial cavity, with invasion to\n> 50% of the myometrial thickness (thick arrow) and solid-cystic mass\nin the left ovary (thin arrow), consistent with endometrial\nadenocarcinoma with metastasis to an ovary.\nFrequency of indications for MRI of the female pelvis according to the\ncriteria established the ACR ( n  = 1060).\nThe indications for pelvic MRI examinations were categorized as oncologic in 763\npatients (72.0%) and as nononcologic in 297 (28.0%). Of the non-oncologic\nindications, the most common were adnexal masses, in 89 patients (29.2%),\nleiomyomas, in 70 (23.0%), and adenomyosis or endometriosis, in 46 (15.1%). Among\nthe oncologic indications, the most common primary tumors were ovarian cancer, in\n145 patients (19.7%), cervical cancer, in 110 (15.0%), and endometrial cancer, in 74\n(10.1%). Also among the oncologic indications, the examination was performed for\nfollow-up/evaluation of post-treatment relapse, in 421 patients (55.2%); for\ndiagnosis, in 176 (23.1%); for staging, in 94 (12.3%); for evaluation of the\ntreatment response, in 46 (6.0%); or for other purposes, in 26 (3.4%).  Table 2  describes the indications for MRI in\npatients with the most common primary tumors evaluated in our sample.\nFrequency of the MRI examination of the female pelvis in the patients with\nthe most common types of gynecologic cancer ( n  = 324).\n\nThe results of the present study demonstrate that, for the majority of the pelvic MRI\nscans evaluated in the present study, the indication for the examination was\nappropriate according to the ACR criteria. In our sample of female patients treated\nat a referral center for cancer, the main indication for MRI was the evaluation of\npost-treatment tumor recurrence, followed by the diagnosis and staging of\ngynecologic tumors.\nMRI is widely accepted as a reliable imaging technique for the diagnosis, staging,\nand post-treatment follow-up of patients with gynecologic neoplasms. It plays a\nstrategic role in the initial assessment, as well as in treatment planning, and is\nmore effective than are other imaging techniques, especially in evaluating the\ninvasion of adjacent structures and lymph node involvement ( 19 - 23 ) .\nIn addition to the morphological evaluation, with highcontrast resolution for\nevaluation of soft tissues, MRI makes it possible to obtain functional information\nthat is associated with histological and prognostic factors. The functional\nevaluation is already part of the standard protocol for MRI of the female pelvis,\nthe most widely used methods being dynamic post-contrast evaluation and\ndiffusion-weighted imaging. Those methods contribute to more accurate\ncharacterization of the lesions, staging (especially in the evaluation of small\nperitoneal implants and extrauterine masses), evaluation of treatment responses, and\ndifferentiation between post-treatment changes and tumor recurrence ( 24 , 25 ) .\nAlthough the evaluation of post-treatment tumor recurrence was the most common\nindication in the present study, it is worth noting that there is no evidence to\nsupport the use of MRI or other routine imaging techniques for posttreatment\nfollow-up. In general, imaging examinations should be reserved for patients with\nalterations on the physical examination or with elevated or symptomatic levels of\nthe tumor marker CA-125 ( 26 ) .\nHowever, such examinations are routinely requested in clinical practice. In a survey\nof United Kingdom oncology gynecologists, 43% of the respondents reported using MRI\nin cervical cancer patients, whereas 14% reported using the technique in ovarian\ncancer patients ( 27 ) . MRI is also\nfrequently requested for the evaluation of recurrence of other nongynecologic pelvic\ntumors, especially in patients with colorectal cancer ( 28 ) .\nIn our sample, the most common non-oncologic indication was the evaluation of an\nadnexal mass. Although ultrasound is the method of choice in such cases, MRI can\nprovide additional information when the ultrasound findings are inconclusive. MRI\nallows better characterization of the soft tissues of adnexal masses, with an\naccuracy of 91-93% for differentiating between benign and malignant\nlesions ( 29 - 31 ) .\nOther common non-oncologic indications include the assessment of leiomyomas or\nadenomyosis/endometriosis. MRI is the most accurate imaging technique for the\nidentification and evaluation of leiomyomas, providing information about the\nlocation, vascularization, and degeneration, thus facilitating the treatment\nplanning ( 32 ) . In the study\nof adenomyosis, MRI has a reported diagnostic accuracy of 85% ( 33 , 34 ) . For patients with suspected endometriosis, MRI helps\nassess the extent of disease, especially in the detection of deep implants, which\nare difficult to characterize by other methods ( 35 , 36 ) .\nThe results of this study have a direct impact on clinical practice. Knowledge of the\nmain indications for pelvic MRI examination allows the radiologist to follow\nappropriate examination protocols for each patient and to develop the skills needed\nin order to provide the necessary information for the proper management of each\ncase. For example, according to the findings of this study, it is fundamental that\nthe radiologist who analyzes this type of examination has knowledge of the current\nstaging, as well as the recurrence patterns of the gynecologic tumors. However,\nthese findings should be considered in the context of certain limitations. This was\na retrospective study, based on the review of medical records and physician\nrequests, the main limitations of which are due to incomplete data recording. In\naddition, because our results depict the routine at a referral center for cancer,\nthey cannot be generalized to centers that serve other populations.\nIn conclusion, the majority of the MRI examinations evaluated presented an\nappropriate indication according to the ACR criteria. The main indications were the\ninvestigation of local recurrence after surgical treatment of pelvic neoplasms and\nthe diagnosis/staging of gynecologic tumors, which is consistent with the routine at\na cancer center. Future studies may evaluate the relationship between the indication\nfor the test and its results for each specific disease, contributing to a better\nrationalization of the use of MRI in Brazil.","source_license":"CC-BY-4.0","license_restricted":false}