- Email: [email protected]
Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease
Clinical Radiology xxx (2014) e1ee6
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease S.K. Kang a, *, G. Giovanniello a, S. Kim a, c, S. Bedell b, d, J.S. Babb a, G.L. Bennett a a b
Department of Radiology, New York University Langone Medical Center, 660 First Ave, New York, NY 10016, USA New York University School of Medicine, 550 First Ave, New York, NY 10016, USA
art icl e i nformat ion Article history: Received 6 March 2014 Received in revised form 11 June 2014 Accepted 12 June 2014
AIM: To evaluate the performance of multidetector computed tomography (MDCT) in the measurement of endometrial thickness and assessment for endometrial disease. MATERIALS AND METHODS: Seventy-nine MDCT examinations, including sagittal reformats from isotropic data, were retrospectively evaluated for the presence of endometrial abnormality, endometrial thickness, and recommendation for transvaginal ultrasound (TVUS) after CT. The endometrial thickness was measured on sagittal images using two different methods, between the inner-to-inner hypoattenuating stripe, and when visible, between the outer-toouter hyperattenuating stripe. TVUS performed within 48 h of CT in premenopausal and 1 month in postmenopausal patients served as reference standard. Interobserver agreement for endometrial thickness and abnormalities was assessed using concordance correlation (CC) and kappa statistics. RESULTS: Interobserver agreement for endometrial thickness on sagittal CT images was excellent (CC 0.98), and highly accurate using the inner-to-inner measurement. For determination of abnormal thickening, the positive predictive value and negative predictive value were 67e100% and 99.5e100%. For detection of any endometrial abnormality, the positive predictive value and negative predictive value were 79e90% and 84e95%, respectively. Falsenegative missed abnormalities included small volume hydrometra, a polyp, and endometrial distortion by a fibroid. CONCLUSION: At MDCT, sagittal reformatted images provide reliable endometrial measurement using the inner-to-inner hypoattenuating stripe and are accurately categorized as normal or abnormal thickness using the same numerical criteria as at sonography. Ó 2014 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
* Guarantor and correspondent: S.K. Kang, Department of Radiology, 660 First Ave, 3rd Floor, Room 333, New York, NY 10016, USA. Tel.: þ1 617 263 0232; fax: þ1 212 263 6634. E-mail addresses: [email protected], [email protected] (S.K. Kang). c Present address: UW Medical Center Radiology, 1959 NE Pacific Street, Seattle, WA 98195, USA. d Present address: UT Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd, Dallas, TX 75390, USA. Tel.: þ1 305 322 8612.
Introduction Although pelvic ultrasound is the first-line imaging study used in evaluating the uterus and adnexa, these structures are also routinely assessed at abdominal and pelvic computed tomography (CT), performed to evaluate lower
http://dx.doi.org/10.1016/j.crad.2014.06.013 0009-9260/Ó 2014 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Kang SK, et al., Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.06.013
e2
S.K. Kang et al. / Clinical Radiology xxx (2014) e1ee6
abdominal and pelvic abnormalities.1 The CT appearance of the normal endometrium varies widely depending on patient age, phase of menstrual cycle, and anatomical position of the uterus. Depending on the orientation of the uterus, standard transaxial images may result in a falsely thickened appearance of the endometrium due to an oblique, non-orthogonal imaging plane. Uterine distortion due to abnormalities, such as fibroids or adenomyosis, may also cause a falsely thickened or even mass-like appearance of the endometrium; such variability in the CT appearance of the endometrium has previously been reported and presents potential pitfalls in distinguishing between a normal and abnormal endometrium.2,3 Given the variable appearance at CT, guidelines for establishing a normal appearance of the endometrium at CT, including thickness, have not been well-established as they are on ultrasound. Meanwhile, multidetector CT (MDCT) technology has become widely available and enables multiplanar reformations to be generated with near isotropic resolution. Reformatted images in the sagittal plane enable improved visualization of the endometrium with delineation of its contour and thickness as these images more closely resemble those obtained at sonography. The clinical performance of MDCT, including reformatted sagittal images, for endometrial assessment has not been well-evaluated, nor has a method of measurement been clearly established for the endometrium given its variable contrast-enhanced appearance at CT. Therefore, the aim of the present study was to evaluate whether a method for accurate and reliable measurement of endometrial thickness could be established at MDCT, using criteria established for TVUS, and to evaluate CT performance in detection of disease. TVUS was used as the reference standard in establishing the accuracy of CT for the measurement of endometrial thickness and in the identification and characterization of specific endometrial disease.
including last menstrual period, was recorded at the time of the ultrasound examinations. Patients were considered to be postmenopausal if at least 1 year had passed since the last menstrual period.
Imaging CT and ultrasound technique All CT examinations were performed on either a fourdetector LightSpeed Plus or 64-detector HiSpeed Advantage (GE Healthcare, Little Chalfort, United Kingdom). Images were obtained after intravenous administration of iohexol 300 (300 mg iodine/ml; Omnipaque 300, GE Healthcare) or iopromide 300 (300 mg iodine/ml; Ultravist 300, Bayer HealthCare, Montville, New Jersey, USA) administered with a single uniphasic bolus of 100e140 ml via a power injector at a rate of 2e3 ml/s. All studies were performed in the portal venous phase, with a 70e80 s delay after contrast medium injection and a pitch between 1.2 and 1.5. Images were acquired with 1.5 mm section thickness, and reviewed in the axial plane at 5 mm section thickness and in the sagittal plane reconstructed from 3 mm section thickness. All images were reviewed on a picture archiving and communication system (PACS; Sectra, Shelton, CT, USA). All patients underwent transabdominal ultrasound and TVUS utilizing Phillips IU22 or HDI5000 systems (Philips Healthcare, Amsterdam, Netherlands). Transducers for the transabdominal examinations ranged in frequency from 2.5 to 6 MHz and for the transvaginal examinations from 4 to 8 MHz. At our institution, trained sonographers perform pelvic ultrasound examinations; the interpreting radiologist then immediately reviews the study and obtains additional images as deemed necessary.
Image interpretation Endometrial assessment at CT
Materials and methods Patient population The institutional review board approved this HIPAA compliant, retrospective study with waiver of informed consent. A computerized search of the radiology database at New York University Langone Medical Center from January 2008 to January 2011 was performed to identify adult female patients who underwent both intravenous (IV) contrast-enhanced abdominopelvic MDCT and TVUS. Patients were excluded if they had undergone hysterectomy. The presence of an intrauterine device was not exclusionary. In order to ensure that physiological changes of the endometrium would not account for differences in imaging appearance between the two modalities, only premenopausal women in whom the CT and ultrasound examinations were performed within 48 h of each other were included in the study cohort. For postmenopausal patients, the maximum time interval between the two studies was 1 month. Patient information regarding menstrual status,
Three readers (S.K., G.G., G.B.) with abdominal imaging fellowship training and 15, 16, and 18 years of experience respectively, independently evaluated the appearance of the endometrium utilizing axial and sagittal CT images. Readers were provided with the age and menstrual status of the patient but no other clinical information. Readers determined the endometrial appearance at CT, specifically indicating the presence of a single hypoattenuating central stripe representing the endometrium, versus a bilaminar appearance with hypoattenuating layer and definable outer enhancing stripe. The endometrial thickness was then recorded on sagittal images using two methods: (1) the measured thickness including only the hypoattenuating component (inner to inner), and (2) when visible, the measured thickness between the outer margins of the stripe, including the outer enhancing layer (Fig 1). Readers indicated whether the endometrium was abnormally thickened for patient age and menstrual status using the inner-to-inner measurements, as all patients were expected to demonstrate the hypoattenuating stripe but not necessarily the variable bilaminar appearance. Readers also indicated whether a specific abnormality was detected,
Please cite this article in press as: Kang SK, et al., Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.06.013
S.K. Kang et al. / Clinical Radiology xxx (2014) e1ee6
e3
Figure 1 Measurement of the endometrium on sagittal CT images. (a) The thickest inner-to-inner measurement of the endometrium (white bracket) was obtained on sagittal images for each patient. Outer-to-outer measurements (black bracket) of the endometrium were also recorded when visible, and may represent the subendometrial myometrium as an additional intermediate density between the endometrium and myometrium in some patients. (b) TVUS provided the reference measurement of endometrial thickness.
including focal masses, such as endometrial polyp, submucosal fibroid, carcinoma, intracavitary fluid, or endometrial distortion by a myometrial abnormality. Given the lack of established criteria for normal endometrial thickness at CT, the same criteria for thickening using pelvic ultrasound criteria were applied.4e6 Readers also attempted to distinguish polyps from other focal masses using criteria of a round, well-circumscribed, homogeneous lesion. Finally, the readers also indicated whether they would recommend pelvic ultrasound following CT to further assess apparent endometrial abnormalities.
Endometrial assessment at ultrasound After completion of the CT readings, one reader (G.B.) with 15 years of experience in abdominal imaging reviewed all pelvic ultrasound examinations and recorded TVUS measurements of endometrial thickness and any abnormal endometrial findings. Sonographic measurements of endometrial thickness and assessment of abnormal thickening for age and menstrual status and endometrial abnormalities were performed using previously published criteria; normal thickness was defined as 16 mm for premenopausal women and 5 mm in postmenopausal women.4e6
Statistical methods Interobserver agreement was assessed using simple kappa coefficients for the binary assessments and concordance correlation (CC) coefficients for the measurements of endometrial thickness. For CC, the agreement was considered excellent for values greater than 0.9, strong for values between 0.8 and 0.9, moderate for values from 0.7 to 0.8, and weak if less than 0.7. For kappa values, agreement was considered strong if greater than 0.6, moderate if between 0.41-0.60, and poor if less than or equal to 0.40. Results were pooled over the three distinct reader pairs, and the data from the three readers were analysed separately. All statistical tests were conducted at the two-sided 5% significance level using SAS 9.3 (SAS Institute, Cary, NC, USA).
Results The patient cohort consisted of 79 women, with mean patient age of 39.2 years; (range 18e80 years); 61 (77%) were premenopausal and 18 (23%) were postmenopausal. No patients were receiving hormone-replacement therapy or tamoxifen. The time interval between CT and ultrasound was 3 8 days (mean standard deviation) for all patients, 11.6 13.4 days for postmenopausal women and 0.36 0.63 days for premenopausal women. The most common indication for CT was abdominal pain, given for 31 of 79 patients, followed by right lower quadrant pain (28 patients). Other provided indications included left lower quadrant pain (n ¼ 9), abdominal mass (n ¼ 5), abdominal distention (n ¼ 2), trauma (n ¼ 1), painless jaundice (n ¼ 1), urinary obstruction (n ¼ 1), and left flank pain (n ¼ 1). The most common indication for ultrasound was pelvic pain, for 27 patients. Other indications were adnexal or pelvic mass (n ¼ 23), possible tubo-ovarian abscess or pelvic inflammatory disease (n ¼ 10), abdominal pain (n ¼ 7), right lower quadrant pain with vaginal bleeding (n ¼ 3), postmenopausal bleeding (n ¼ 3), endometrial cancer (n ¼ 2), cervical mass (n ¼ 1), uterine anomaly (n ¼ 1), fibroid (n ¼ 1), and bladder mass (n ¼ 1). The mean thickness of the endometrium on pelvic ultrasound was 9.5 12.4 mm, and 9 4.6 mm in premenopausal women and 10.9 23.7 mm in postmenopausal women. Twenty-two of 79 patients (72.1%) had a bilaminar endometrial appearance at CT as determined by at least one reader, with inner hypoattenuating and outer hyperattenuating stripes.
Interobserver agreement For measurement of endometrial thickness, the CC was 0.98 when readers used the inner-to-inner measure of thickness on sagittal CT images, indicating excellent agreement. Outer-to-outer margin measures of endometrial
Please cite this article in press as: Kang SK, et al., Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.06.013
e4
S.K. Kang et al. / Clinical Radiology xxx (2014) e1ee6
thickness demonstrated moderate agreement (CC 0.75 using sagittal images). Inter-reader agreement was assessed for diagnosis of normal versus abnormal endometrium and for specific endometrial abnormalities or uterine abnormalities affecting the endometrium. Kappa values for determination of a normal endometrium at CT was 0.65, showing strong agreement among the three readers. For diagnosis of a thickened endometrium, presence of fluid within the endometrial cavity, and submucosal fibroids, the kappa value showed moderate agreement (0.53, 0.43, and 0.56 respectively). Agreement among readers for presence of a focal mass was poor at CT with kappa value of 0.35.
Accuracy of endometrial thickness measurements determined at CT There was excellent CC (0.91) between all inner-to-inner measurements of endometrial thickness determined at CT and those determined at sonography (Fig 2), whereas the corresponding outer-to-outer measurements of thickness had a relatively weak CC with sonography (0.51). For all three readers, the mean difference in measurements obtained on sagittal CT images versus sonography was less than 3 mm. Only one postmenopausal woman had underestimated thickness at CT compared with ultrasound by more than 2 mm (10 mm versus 14 mm), still correctly interpreted as abnormally thickened at CT.
CT assessment of specific endometrial abnormalities Twenty-two of 79 patients (20.2%) had a total of 40 abnormalities of the endometrium at TVUS (Table 1). The positive predictive value and negative predictive value for the presence of any abnormality at CT were 79e90% and 84e95%, respectively. There were zero false-positive reads for the presence of any endometrial abnormality involving all three readers. False-negative reads by at least two of the three readers comprised endometrial fluid, distortion by a fibroid, an endometrial polyp, and one case of endometrial thickening in a premenopausal woman where the
Table 1 Overall diagnostic accuracy using multidetector computed tomography (CT) images to detect endometrial abnormality, relative to pelvic ultrasound. Attribute Any abnormality
Reader No. cases Accuracy at CT of abnormality
1 2 3 Endometrial cancer 1 2 3 Distorted endometrium 1 2 3 Submucosal fibroid 1 2 3 Fluid-filled endometrial cavity 1 2 3 Focal mass (including cancer) 1 2 3 Endometrial polyp 1 2 3 Endometrial thickening 1 2 3
40 e e 2 e e 9 e e 2 e e 7 e e 8 e e 7 e e 7 e e
90.7% 83.8% 82.9% 98.7% 98.6% 97.4% 98.7% 93.2% 93.4% 98.7% 95.9% 97.4% 88% 89.2% 90.8% 96% 93.2% 90.8% 94.7% 90.5% 92.1% 92% 97.3% 94.7%
(68/75) (62/74) (63/76) (74/75) (73/74) (74/76) (74/75) (69/74) (71/76) (74/75) (71/74) (74/76) (66/75) (66/74) (69/76) (72/75) (69/74) (69/76) (71/75) (67/74) (70/76) (69/75) (72/74) (72/76)
endometrium measured 13 mm at CT and 19 mm at sonography. No false-negative interpretations of thickened endometrium or endometrial mass occurred in postmenopausal women. For diagnosis of an abnormally thickened endometrium, the accuracy of sagittal CT images was 92e94.7%, and positive predictive value and negative predictive value of thickening on sagittal CT images were 66.7e100% and 99.5e100%, respectively. Accuracy in detection of specific abnormalities aside from thickening, including focal masses such as submucosal fibroids, polyps, and fluid-filled endometrial cavity, was variable among the three readers, but no worse than 88% for
Figure 2 An 80-year-old woman with abdominal pain. (a) Endometrial thickening as determined on sagittal CT images, with thickness of 11 mm (arrow). (b) TVUS demonstrated thickened endometrium measuring 12 mm (arrow); endometrial hyperplasia was diagnosed at histopathology. Please cite this article in press as: Kang SK, et al., Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.06.013
S.K. Kang et al. / Clinical Radiology xxx (2014) e1ee6
any reader interpreting any specific abnormality (Fig 3). Endometrial cancers were detected with greater than 97% accuracy, although the present data included a small number of malignancies (Table 1).
Recommendations for additional imaging The rate of recommendations for additional imaging with pelvic ultrasound for all three readers was 12.7e30.4% of patients after review of CT images (compared with 20.2% of patients demonstrating abnormality at pelvic sonography).
Discussion Pelvic ultrasound, and in particular TVUS, is the imaging method of choice for assessment of the endometrium, with endometrial thickness a well-established predictor of significant endometrial disease.4e6 MRI may serve as a problem-solving tool if the endometrium requires further assessment after ultrasound, given well-delineated softtissue contrast in uterine tissues.7,8 In contrast, assessment of the endometrium at CT is generally viewed as limited4 due to several factors that restrict visualization or predispose to interpretive errors: poor soft-tissue contrast at CT relative to ultrasound and MRI, limited endometrial visualization in the standard transaxial plane due to variations in uterine position and/or distortion by myometrial abnormalities, and variable patterns of normal uterine contrast enhancement that depend on patient age and menstrual status.9,10 Despite the presence of such limitations, the uterus must be assessed on all abdominopelvic CT images, the performance of MDCT in assessing the endometrium has not been well-evaluated quantitatively, and criteria for discriminating a normal from abnormal endometrium at CT have not been well established. In the present study, the measurement of endometrial thickness on sagittal reformatted CT images was accurate and reproducible compared with TVUS. The endometrium may therefore be accurately categorized as being of normal
e5
or abnormal thickness using the same numerical criteria as at sonography. Furthermore, when the endometrium appears with an inner and outer, hyperattenuating stripe, the inner-to-inner measurement including only the hypoattenuating component provides an accurate and reproducible measure of thickness. The outer enhancing layer may correspond to the inner myometrium (junctional zone) seen at MRI, and should not be included in the CT measurement of endometrial thickness. With high positive predictive value for the presence of endometrial abnormality, MDCT detection of an abnormality may be relied upon to determine the need for further clinical evaluation without concern for a high false-positive rate. Although CT demonstrated high accuracy for determining the presence of endometrial abnormality, its clinical performance for differentiating among specific endometrial abnormality types, thereby eliminating the need for additional imaging evaluation with pelvic sonography, requires larger studies. Readers in the present study recommended pelvic sonography after CT as they would have done in actual clinical practice, with their recommendation rates likely reflect uncertainty in the accuracy of CT findings in the endometrium; however, the proportion of patients with endometrial abnormalities at CT were in a similar range to the proportion established at sonography. In the present study, most of the endometrial abnormalities detected at CT represented incidental imaging findings as most studies were not performed for indications related to localized endometrial disease (i.e. vaginal bleeding), and our study does not support use of CT for primary evaluation of the endometrium. By moving toward guidelines for measurement and interpretation at CT, however, the endometrium may be more uniformly and completely assessed at CT. Quantitative and qualitative use of sagittal reformatted images may serve to help establish such interpretive criteria, while requiring no additional patient exposure to ionizing radiation and minimal associated increase in interpretation time. A single study previously evaluated the efficacy of contrast-enhanced CT in endometrial assessment, with
Figure 3 A 44-year-old woman with abdominal pain. (a) Sagittal CT images demonstrated an intramural uterine fibroid (white arrow). In addition, a smaller mass was also apparent (black arrow) with a substantial submucosal component in relation to the endometrium (white bracket). (b) TVUS demonstrated a submucosal fibroid (white arrow) and normal endometrium (black arrow). Please cite this article in press as: Kang SK, et al., Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.06.013
e6
S.K. Kang et al. / Clinical Radiology xxx (2014) e1ee6
limited use of sagittal reformatted images in qualitative evaluation of endometrial thickness; however, the authors did not assess accuracy of measured endometrial thickness at CT.11 The study described the CT appearance of the endometrium but predated the widespread availability of MDCT and utilized only axial images in the majority of cases.11 As expected, given the anatomical position of the uterus, use of axial images only to assess the endometrium resulted in overall low positive predictive value for thickening (66%), and low sensitivity for detection of thickening with a high false-negative rate of 47% (23 of 49 patients)11 in contrast to the findings of the present study of 99.5e100% negative predictive value with inclusion of reformatted sagittal images. Multiplanar reformatted images are now widely available as nearly isotropic datasets using current MDCT technology, and have been shown to improve reader confidence and accuracy in multiple abdominopelvic imaging applications: location of transition points in small bowel obstruction, identification of an inflamed appendix, and in diagnosis of several other acute abdominal conditions.12e14 Improved visualization of the endometrium on sagittal reformatted views has been described3 and the added value of sagittal images in staging of endometrial cancer at MDCT demonstrated.15 Limitations of the present study include the lack of histopathological correlation. However, the aim of the present study was to compare CT with sonography, which is considered the reference standard in imaging evaluation of the endometrium. Second, the time frame of the ultrasound examinations was within 2 days of CT in premenopausal women and while changes in endometrial thickening of approximately 1e2 mm are possible in this time frame, sagittal CT measurements of thickening did not significantly underestimate the reference standard, and did not result in missed cases of thickened endometrium in the present study. Third, given the relatively small number of cases included, the qualitative appearance of the endometrium at CT was not correlated with menstrual status or phase of menstrual cycle, as has been done for ultrasound. Also variations in homogeneity, attenuation, or enhancement of the endometrium were not assessed, which would require a larger cohort of patients imaged at different phases of the menstrual cycle. Similarly, the effect of bolus timing of injected contrast medium on the variability of endometrial attenuation was not examined. Despite the presence of such factors that can contribute to the variable appearance of the endometrium at CT, endometrial measurement on sagittal images results in highly accurate measurements of the endometrial thickness, specifically using the inner-to-inner measurement, and strong performance in determination of abnormal thickening. These findings are generalizable for endometrial assessment in the portal venous timing (70e80 s) of contrast enhancement at abdominopelvic CT, without the restriction of requiring knowledge of the phase of the menstrual cycle in a particular patient.
Fourth, the sagittal images included in the study were automatically generated by the CT scanner in a preset sagittal plane. The accuracy of CT for determination of endometrial thickness may improve further if images were reconstructed that were truly sagittal with respect to the uterus and endometrium, particularly in cases of oblique deviation; however, the overall yield in doing so would not be substantial with accuracy already 92% or more for all readers in the present study in the sagittal plane. Finally, the readers were all experienced radiologists with many years of practice dedicated to abdominal imaging. If less experienced readers, including trainees, were included, it is possible that reliability and accuracy of CT interpretations may be less robust. In conclusion, MDCT provides accurate and reliable measurement of the endometrial thickness and detection of endometrial abnormalities. The high accuracy of CT measurements of endometrial thickness on sagittal images compared with TVUS enable clinical interpretation using the same imaging metrics as at sonography. If characterization of the endometrium remains indeterminate, TVUS remains the gold standard.
References 1. Swart JE, Fishman EK. Gynecologic pathology on multidetector CT: a pictorial review. Emerg Radiol 2008;15:383e9. 2. Lim PS, Nazarian LN, Wechsler RJ, et al. The endometrium on routine contrast-enhanced CT in asymptomatic postmenopausal women: avoiding errors in interpretation. Clin Imaging 2002;26:325e9. 3. Yitta S, Hecht EM, Slywotzky CM, et al. Added value of multiplanar reformation in the multidetector CT evaluation of the female pelvis: a pictorial review. Radiographics 2009;29:1987e2003. 4. Nalaboff KM, Pellerito JS, Ben-Levi E. Imaging the endometrium: disease and normal variants. RadioGraphics 2001;21:1409e24. 5. Bennett GL, Andreotti RF, Lee SI, et al. ACR Appropriateness Criteria on abnormal vaginal bleeding. J Am Coll Radiol 2011;8:460e8. 6. Goldstein RB, Bree RL, Benson CB, et al. Evaluation of the woman with postmenopausal bleeding: Society of Radiologists in UltrasoundSponsored Consensus Conference statement. J Ultrasound Med 2001;20:1025e36. 7. Hricak H. MRI of the female pelvis: a review. AJR Am J Roentgenol 1986;146:1115e22. 8. Hardesty LA, Sumkin JH, Hakim C, et al. The ability of helical CT to preoperatively stage endometrial carcinoma. AJR Am J Roentgenol 2001;176:603e6. 9. Kaur H, Loyer EM, Minami M, et al. Patterns of uterine enhancement with helical CT. Eur J Radiol 1998;28:250e5. 10. Yitta S, Hecht EM, Mausner EV, et al. Normal or abnormal? Demystifying uterine and cervical contrast enhancement at multidetector CT. RadioGraphics 2011;31:647e61. 11. Grossman J, Ricci ZJ, Rozenblit A, et al. Efficacy of contrast-enhanced CT in assessing the endometrium. AJR Am J Roentgenol 2008;191:664e9. 12. Hodel J, Zins M, Desmottes L, et al. Location of the transition zone in CT of small-bowel obstruction: added value of multiplanar reformations. Abdom Imaging 2009;34:35e41. 13. Zangos S, Steenburg SD, Philllips KD, et al. Acute abdomen: added diagnostic value of coronal reformations with 64-slice multidetector row computed tomography. Acad Radiol 2007;14:19e27. 14. Neville AM, Paulson EK. MDCT of acute appendicitis: value of coronal reformations. Abdom Imaging 2009;34:42e8. 15. Tsili AC, Tsampoulas C, Dalkalitsis N, et al. Local staging of endometrial carcinoma: role of multidetector CT. Eur Radiol 2008;18:1043e8.
Please cite this article in press as: Kang SK, et al., Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.06.013
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease S.K. Kang a, *, G. Giovanniello a, S. Kim a, c, S. Bedell b, d, J.S. Babb a, G.L. Bennett a a b
Department of Radiology, New York University Langone Medical Center, 660 First Ave, New York, NY 10016, USA New York University School of Medicine, 550 First Ave, New York, NY 10016, USA
art icl e i nformat ion Article history: Received 6 March 2014 Received in revised form 11 June 2014 Accepted 12 June 2014
AIM: To evaluate the performance of multidetector computed tomography (MDCT) in the measurement of endometrial thickness and assessment for endometrial disease. MATERIALS AND METHODS: Seventy-nine MDCT examinations, including sagittal reformats from isotropic data, were retrospectively evaluated for the presence of endometrial abnormality, endometrial thickness, and recommendation for transvaginal ultrasound (TVUS) after CT. The endometrial thickness was measured on sagittal images using two different methods, between the inner-to-inner hypoattenuating stripe, and when visible, between the outer-toouter hyperattenuating stripe. TVUS performed within 48 h of CT in premenopausal and 1 month in postmenopausal patients served as reference standard. Interobserver agreement for endometrial thickness and abnormalities was assessed using concordance correlation (CC) and kappa statistics. RESULTS: Interobserver agreement for endometrial thickness on sagittal CT images was excellent (CC 0.98), and highly accurate using the inner-to-inner measurement. For determination of abnormal thickening, the positive predictive value and negative predictive value were 67e100% and 99.5e100%. For detection of any endometrial abnormality, the positive predictive value and negative predictive value were 79e90% and 84e95%, respectively. Falsenegative missed abnormalities included small volume hydrometra, a polyp, and endometrial distortion by a fibroid. CONCLUSION: At MDCT, sagittal reformatted images provide reliable endometrial measurement using the inner-to-inner hypoattenuating stripe and are accurately categorized as normal or abnormal thickness using the same numerical criteria as at sonography. Ó 2014 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
* Guarantor and correspondent: S.K. Kang, Department of Radiology, 660 First Ave, 3rd Floor, Room 333, New York, NY 10016, USA. Tel.: þ1 617 263 0232; fax: þ1 212 263 6634. E-mail addresses: [email protected], [email protected] (S.K. Kang). c Present address: UW Medical Center Radiology, 1959 NE Pacific Street, Seattle, WA 98195, USA. d Present address: UT Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd, Dallas, TX 75390, USA. Tel.: þ1 305 322 8612.
Introduction Although pelvic ultrasound is the first-line imaging study used in evaluating the uterus and adnexa, these structures are also routinely assessed at abdominal and pelvic computed tomography (CT), performed to evaluate lower
http://dx.doi.org/10.1016/j.crad.2014.06.013 0009-9260/Ó 2014 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Kang SK, et al., Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.06.013
e2
S.K. Kang et al. / Clinical Radiology xxx (2014) e1ee6
abdominal and pelvic abnormalities.1 The CT appearance of the normal endometrium varies widely depending on patient age, phase of menstrual cycle, and anatomical position of the uterus. Depending on the orientation of the uterus, standard transaxial images may result in a falsely thickened appearance of the endometrium due to an oblique, non-orthogonal imaging plane. Uterine distortion due to abnormalities, such as fibroids or adenomyosis, may also cause a falsely thickened or even mass-like appearance of the endometrium; such variability in the CT appearance of the endometrium has previously been reported and presents potential pitfalls in distinguishing between a normal and abnormal endometrium.2,3 Given the variable appearance at CT, guidelines for establishing a normal appearance of the endometrium at CT, including thickness, have not been well-established as they are on ultrasound. Meanwhile, multidetector CT (MDCT) technology has become widely available and enables multiplanar reformations to be generated with near isotropic resolution. Reformatted images in the sagittal plane enable improved visualization of the endometrium with delineation of its contour and thickness as these images more closely resemble those obtained at sonography. The clinical performance of MDCT, including reformatted sagittal images, for endometrial assessment has not been well-evaluated, nor has a method of measurement been clearly established for the endometrium given its variable contrast-enhanced appearance at CT. Therefore, the aim of the present study was to evaluate whether a method for accurate and reliable measurement of endometrial thickness could be established at MDCT, using criteria established for TVUS, and to evaluate CT performance in detection of disease. TVUS was used as the reference standard in establishing the accuracy of CT for the measurement of endometrial thickness and in the identification and characterization of specific endometrial disease.
including last menstrual period, was recorded at the time of the ultrasound examinations. Patients were considered to be postmenopausal if at least 1 year had passed since the last menstrual period.
Imaging CT and ultrasound technique All CT examinations were performed on either a fourdetector LightSpeed Plus or 64-detector HiSpeed Advantage (GE Healthcare, Little Chalfort, United Kingdom). Images were obtained after intravenous administration of iohexol 300 (300 mg iodine/ml; Omnipaque 300, GE Healthcare) or iopromide 300 (300 mg iodine/ml; Ultravist 300, Bayer HealthCare, Montville, New Jersey, USA) administered with a single uniphasic bolus of 100e140 ml via a power injector at a rate of 2e3 ml/s. All studies were performed in the portal venous phase, with a 70e80 s delay after contrast medium injection and a pitch between 1.2 and 1.5. Images were acquired with 1.5 mm section thickness, and reviewed in the axial plane at 5 mm section thickness and in the sagittal plane reconstructed from 3 mm section thickness. All images were reviewed on a picture archiving and communication system (PACS; Sectra, Shelton, CT, USA). All patients underwent transabdominal ultrasound and TVUS utilizing Phillips IU22 or HDI5000 systems (Philips Healthcare, Amsterdam, Netherlands). Transducers for the transabdominal examinations ranged in frequency from 2.5 to 6 MHz and for the transvaginal examinations from 4 to 8 MHz. At our institution, trained sonographers perform pelvic ultrasound examinations; the interpreting radiologist then immediately reviews the study and obtains additional images as deemed necessary.
Image interpretation Endometrial assessment at CT
Materials and methods Patient population The institutional review board approved this HIPAA compliant, retrospective study with waiver of informed consent. A computerized search of the radiology database at New York University Langone Medical Center from January 2008 to January 2011 was performed to identify adult female patients who underwent both intravenous (IV) contrast-enhanced abdominopelvic MDCT and TVUS. Patients were excluded if they had undergone hysterectomy. The presence of an intrauterine device was not exclusionary. In order to ensure that physiological changes of the endometrium would not account for differences in imaging appearance between the two modalities, only premenopausal women in whom the CT and ultrasound examinations were performed within 48 h of each other were included in the study cohort. For postmenopausal patients, the maximum time interval between the two studies was 1 month. Patient information regarding menstrual status,
Three readers (S.K., G.G., G.B.) with abdominal imaging fellowship training and 15, 16, and 18 years of experience respectively, independently evaluated the appearance of the endometrium utilizing axial and sagittal CT images. Readers were provided with the age and menstrual status of the patient but no other clinical information. Readers determined the endometrial appearance at CT, specifically indicating the presence of a single hypoattenuating central stripe representing the endometrium, versus a bilaminar appearance with hypoattenuating layer and definable outer enhancing stripe. The endometrial thickness was then recorded on sagittal images using two methods: (1) the measured thickness including only the hypoattenuating component (inner to inner), and (2) when visible, the measured thickness between the outer margins of the stripe, including the outer enhancing layer (Fig 1). Readers indicated whether the endometrium was abnormally thickened for patient age and menstrual status using the inner-to-inner measurements, as all patients were expected to demonstrate the hypoattenuating stripe but not necessarily the variable bilaminar appearance. Readers also indicated whether a specific abnormality was detected,
Please cite this article in press as: Kang SK, et al., Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.06.013
S.K. Kang et al. / Clinical Radiology xxx (2014) e1ee6
e3
Figure 1 Measurement of the endometrium on sagittal CT images. (a) The thickest inner-to-inner measurement of the endometrium (white bracket) was obtained on sagittal images for each patient. Outer-to-outer measurements (black bracket) of the endometrium were also recorded when visible, and may represent the subendometrial myometrium as an additional intermediate density between the endometrium and myometrium in some patients. (b) TVUS provided the reference measurement of endometrial thickness.
including focal masses, such as endometrial polyp, submucosal fibroid, carcinoma, intracavitary fluid, or endometrial distortion by a myometrial abnormality. Given the lack of established criteria for normal endometrial thickness at CT, the same criteria for thickening using pelvic ultrasound criteria were applied.4e6 Readers also attempted to distinguish polyps from other focal masses using criteria of a round, well-circumscribed, homogeneous lesion. Finally, the readers also indicated whether they would recommend pelvic ultrasound following CT to further assess apparent endometrial abnormalities.
Endometrial assessment at ultrasound After completion of the CT readings, one reader (G.B.) with 15 years of experience in abdominal imaging reviewed all pelvic ultrasound examinations and recorded TVUS measurements of endometrial thickness and any abnormal endometrial findings. Sonographic measurements of endometrial thickness and assessment of abnormal thickening for age and menstrual status and endometrial abnormalities were performed using previously published criteria; normal thickness was defined as 16 mm for premenopausal women and 5 mm in postmenopausal women.4e6
Statistical methods Interobserver agreement was assessed using simple kappa coefficients for the binary assessments and concordance correlation (CC) coefficients for the measurements of endometrial thickness. For CC, the agreement was considered excellent for values greater than 0.9, strong for values between 0.8 and 0.9, moderate for values from 0.7 to 0.8, and weak if less than 0.7. For kappa values, agreement was considered strong if greater than 0.6, moderate if between 0.41-0.60, and poor if less than or equal to 0.40. Results were pooled over the three distinct reader pairs, and the data from the three readers were analysed separately. All statistical tests were conducted at the two-sided 5% significance level using SAS 9.3 (SAS Institute, Cary, NC, USA).
Results The patient cohort consisted of 79 women, with mean patient age of 39.2 years; (range 18e80 years); 61 (77%) were premenopausal and 18 (23%) were postmenopausal. No patients were receiving hormone-replacement therapy or tamoxifen. The time interval between CT and ultrasound was 3 8 days (mean standard deviation) for all patients, 11.6 13.4 days for postmenopausal women and 0.36 0.63 days for premenopausal women. The most common indication for CT was abdominal pain, given for 31 of 79 patients, followed by right lower quadrant pain (28 patients). Other provided indications included left lower quadrant pain (n ¼ 9), abdominal mass (n ¼ 5), abdominal distention (n ¼ 2), trauma (n ¼ 1), painless jaundice (n ¼ 1), urinary obstruction (n ¼ 1), and left flank pain (n ¼ 1). The most common indication for ultrasound was pelvic pain, for 27 patients. Other indications were adnexal or pelvic mass (n ¼ 23), possible tubo-ovarian abscess or pelvic inflammatory disease (n ¼ 10), abdominal pain (n ¼ 7), right lower quadrant pain with vaginal bleeding (n ¼ 3), postmenopausal bleeding (n ¼ 3), endometrial cancer (n ¼ 2), cervical mass (n ¼ 1), uterine anomaly (n ¼ 1), fibroid (n ¼ 1), and bladder mass (n ¼ 1). The mean thickness of the endometrium on pelvic ultrasound was 9.5 12.4 mm, and 9 4.6 mm in premenopausal women and 10.9 23.7 mm in postmenopausal women. Twenty-two of 79 patients (72.1%) had a bilaminar endometrial appearance at CT as determined by at least one reader, with inner hypoattenuating and outer hyperattenuating stripes.
Interobserver agreement For measurement of endometrial thickness, the CC was 0.98 when readers used the inner-to-inner measure of thickness on sagittal CT images, indicating excellent agreement. Outer-to-outer margin measures of endometrial
Please cite this article in press as: Kang SK, et al., Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.06.013
e4
S.K. Kang et al. / Clinical Radiology xxx (2014) e1ee6
thickness demonstrated moderate agreement (CC 0.75 using sagittal images). Inter-reader agreement was assessed for diagnosis of normal versus abnormal endometrium and for specific endometrial abnormalities or uterine abnormalities affecting the endometrium. Kappa values for determination of a normal endometrium at CT was 0.65, showing strong agreement among the three readers. For diagnosis of a thickened endometrium, presence of fluid within the endometrial cavity, and submucosal fibroids, the kappa value showed moderate agreement (0.53, 0.43, and 0.56 respectively). Agreement among readers for presence of a focal mass was poor at CT with kappa value of 0.35.
Accuracy of endometrial thickness measurements determined at CT There was excellent CC (0.91) between all inner-to-inner measurements of endometrial thickness determined at CT and those determined at sonography (Fig 2), whereas the corresponding outer-to-outer measurements of thickness had a relatively weak CC with sonography (0.51). For all three readers, the mean difference in measurements obtained on sagittal CT images versus sonography was less than 3 mm. Only one postmenopausal woman had underestimated thickness at CT compared with ultrasound by more than 2 mm (10 mm versus 14 mm), still correctly interpreted as abnormally thickened at CT.
CT assessment of specific endometrial abnormalities Twenty-two of 79 patients (20.2%) had a total of 40 abnormalities of the endometrium at TVUS (Table 1). The positive predictive value and negative predictive value for the presence of any abnormality at CT were 79e90% and 84e95%, respectively. There were zero false-positive reads for the presence of any endometrial abnormality involving all three readers. False-negative reads by at least two of the three readers comprised endometrial fluid, distortion by a fibroid, an endometrial polyp, and one case of endometrial thickening in a premenopausal woman where the
Table 1 Overall diagnostic accuracy using multidetector computed tomography (CT) images to detect endometrial abnormality, relative to pelvic ultrasound. Attribute Any abnormality
Reader No. cases Accuracy at CT of abnormality
1 2 3 Endometrial cancer 1 2 3 Distorted endometrium 1 2 3 Submucosal fibroid 1 2 3 Fluid-filled endometrial cavity 1 2 3 Focal mass (including cancer) 1 2 3 Endometrial polyp 1 2 3 Endometrial thickening 1 2 3
40 e e 2 e e 9 e e 2 e e 7 e e 8 e e 7 e e 7 e e
90.7% 83.8% 82.9% 98.7% 98.6% 97.4% 98.7% 93.2% 93.4% 98.7% 95.9% 97.4% 88% 89.2% 90.8% 96% 93.2% 90.8% 94.7% 90.5% 92.1% 92% 97.3% 94.7%
(68/75) (62/74) (63/76) (74/75) (73/74) (74/76) (74/75) (69/74) (71/76) (74/75) (71/74) (74/76) (66/75) (66/74) (69/76) (72/75) (69/74) (69/76) (71/75) (67/74) (70/76) (69/75) (72/74) (72/76)
endometrium measured 13 mm at CT and 19 mm at sonography. No false-negative interpretations of thickened endometrium or endometrial mass occurred in postmenopausal women. For diagnosis of an abnormally thickened endometrium, the accuracy of sagittal CT images was 92e94.7%, and positive predictive value and negative predictive value of thickening on sagittal CT images were 66.7e100% and 99.5e100%, respectively. Accuracy in detection of specific abnormalities aside from thickening, including focal masses such as submucosal fibroids, polyps, and fluid-filled endometrial cavity, was variable among the three readers, but no worse than 88% for
Figure 2 An 80-year-old woman with abdominal pain. (a) Endometrial thickening as determined on sagittal CT images, with thickness of 11 mm (arrow). (b) TVUS demonstrated thickened endometrium measuring 12 mm (arrow); endometrial hyperplasia was diagnosed at histopathology. Please cite this article in press as: Kang SK, et al., Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.06.013
S.K. Kang et al. / Clinical Radiology xxx (2014) e1ee6
any reader interpreting any specific abnormality (Fig 3). Endometrial cancers were detected with greater than 97% accuracy, although the present data included a small number of malignancies (Table 1).
Recommendations for additional imaging The rate of recommendations for additional imaging with pelvic ultrasound for all three readers was 12.7e30.4% of patients after review of CT images (compared with 20.2% of patients demonstrating abnormality at pelvic sonography).
Discussion Pelvic ultrasound, and in particular TVUS, is the imaging method of choice for assessment of the endometrium, with endometrial thickness a well-established predictor of significant endometrial disease.4e6 MRI may serve as a problem-solving tool if the endometrium requires further assessment after ultrasound, given well-delineated softtissue contrast in uterine tissues.7,8 In contrast, assessment of the endometrium at CT is generally viewed as limited4 due to several factors that restrict visualization or predispose to interpretive errors: poor soft-tissue contrast at CT relative to ultrasound and MRI, limited endometrial visualization in the standard transaxial plane due to variations in uterine position and/or distortion by myometrial abnormalities, and variable patterns of normal uterine contrast enhancement that depend on patient age and menstrual status.9,10 Despite the presence of such limitations, the uterus must be assessed on all abdominopelvic CT images, the performance of MDCT in assessing the endometrium has not been well-evaluated quantitatively, and criteria for discriminating a normal from abnormal endometrium at CT have not been well established. In the present study, the measurement of endometrial thickness on sagittal reformatted CT images was accurate and reproducible compared with TVUS. The endometrium may therefore be accurately categorized as being of normal
e5
or abnormal thickness using the same numerical criteria as at sonography. Furthermore, when the endometrium appears with an inner and outer, hyperattenuating stripe, the inner-to-inner measurement including only the hypoattenuating component provides an accurate and reproducible measure of thickness. The outer enhancing layer may correspond to the inner myometrium (junctional zone) seen at MRI, and should not be included in the CT measurement of endometrial thickness. With high positive predictive value for the presence of endometrial abnormality, MDCT detection of an abnormality may be relied upon to determine the need for further clinical evaluation without concern for a high false-positive rate. Although CT demonstrated high accuracy for determining the presence of endometrial abnormality, its clinical performance for differentiating among specific endometrial abnormality types, thereby eliminating the need for additional imaging evaluation with pelvic sonography, requires larger studies. Readers in the present study recommended pelvic sonography after CT as they would have done in actual clinical practice, with their recommendation rates likely reflect uncertainty in the accuracy of CT findings in the endometrium; however, the proportion of patients with endometrial abnormalities at CT were in a similar range to the proportion established at sonography. In the present study, most of the endometrial abnormalities detected at CT represented incidental imaging findings as most studies were not performed for indications related to localized endometrial disease (i.e. vaginal bleeding), and our study does not support use of CT for primary evaluation of the endometrium. By moving toward guidelines for measurement and interpretation at CT, however, the endometrium may be more uniformly and completely assessed at CT. Quantitative and qualitative use of sagittal reformatted images may serve to help establish such interpretive criteria, while requiring no additional patient exposure to ionizing radiation and minimal associated increase in interpretation time. A single study previously evaluated the efficacy of contrast-enhanced CT in endometrial assessment, with
Figure 3 A 44-year-old woman with abdominal pain. (a) Sagittal CT images demonstrated an intramural uterine fibroid (white arrow). In addition, a smaller mass was also apparent (black arrow) with a substantial submucosal component in relation to the endometrium (white bracket). (b) TVUS demonstrated a submucosal fibroid (white arrow) and normal endometrium (black arrow). Please cite this article in press as: Kang SK, et al., Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.06.013
e6
S.K. Kang et al. / Clinical Radiology xxx (2014) e1ee6
limited use of sagittal reformatted images in qualitative evaluation of endometrial thickness; however, the authors did not assess accuracy of measured endometrial thickness at CT.11 The study described the CT appearance of the endometrium but predated the widespread availability of MDCT and utilized only axial images in the majority of cases.11 As expected, given the anatomical position of the uterus, use of axial images only to assess the endometrium resulted in overall low positive predictive value for thickening (66%), and low sensitivity for detection of thickening with a high false-negative rate of 47% (23 of 49 patients)11 in contrast to the findings of the present study of 99.5e100% negative predictive value with inclusion of reformatted sagittal images. Multiplanar reformatted images are now widely available as nearly isotropic datasets using current MDCT technology, and have been shown to improve reader confidence and accuracy in multiple abdominopelvic imaging applications: location of transition points in small bowel obstruction, identification of an inflamed appendix, and in diagnosis of several other acute abdominal conditions.12e14 Improved visualization of the endometrium on sagittal reformatted views has been described3 and the added value of sagittal images in staging of endometrial cancer at MDCT demonstrated.15 Limitations of the present study include the lack of histopathological correlation. However, the aim of the present study was to compare CT with sonography, which is considered the reference standard in imaging evaluation of the endometrium. Second, the time frame of the ultrasound examinations was within 2 days of CT in premenopausal women and while changes in endometrial thickening of approximately 1e2 mm are possible in this time frame, sagittal CT measurements of thickening did not significantly underestimate the reference standard, and did not result in missed cases of thickened endometrium in the present study. Third, given the relatively small number of cases included, the qualitative appearance of the endometrium at CT was not correlated with menstrual status or phase of menstrual cycle, as has been done for ultrasound. Also variations in homogeneity, attenuation, or enhancement of the endometrium were not assessed, which would require a larger cohort of patients imaged at different phases of the menstrual cycle. Similarly, the effect of bolus timing of injected contrast medium on the variability of endometrial attenuation was not examined. Despite the presence of such factors that can contribute to the variable appearance of the endometrium at CT, endometrial measurement on sagittal images results in highly accurate measurements of the endometrial thickness, specifically using the inner-to-inner measurement, and strong performance in determination of abnormal thickening. These findings are generalizable for endometrial assessment in the portal venous timing (70e80 s) of contrast enhancement at abdominopelvic CT, without the restriction of requiring knowledge of the phase of the menstrual cycle in a particular patient.
Fourth, the sagittal images included in the study were automatically generated by the CT scanner in a preset sagittal plane. The accuracy of CT for determination of endometrial thickness may improve further if images were reconstructed that were truly sagittal with respect to the uterus and endometrium, particularly in cases of oblique deviation; however, the overall yield in doing so would not be substantial with accuracy already 92% or more for all readers in the present study in the sagittal plane. Finally, the readers were all experienced radiologists with many years of practice dedicated to abdominal imaging. If less experienced readers, including trainees, were included, it is possible that reliability and accuracy of CT interpretations may be less robust. In conclusion, MDCT provides accurate and reliable measurement of the endometrial thickness and detection of endometrial abnormalities. The high accuracy of CT measurements of endometrial thickness on sagittal images compared with TVUS enable clinical interpretation using the same imaging metrics as at sonography. If characterization of the endometrium remains indeterminate, TVUS remains the gold standard.
References 1. Swart JE, Fishman EK. Gynecologic pathology on multidetector CT: a pictorial review. Emerg Radiol 2008;15:383e9. 2. Lim PS, Nazarian LN, Wechsler RJ, et al. The endometrium on routine contrast-enhanced CT in asymptomatic postmenopausal women: avoiding errors in interpretation. Clin Imaging 2002;26:325e9. 3. Yitta S, Hecht EM, Slywotzky CM, et al. Added value of multiplanar reformation in the multidetector CT evaluation of the female pelvis: a pictorial review. Radiographics 2009;29:1987e2003. 4. Nalaboff KM, Pellerito JS, Ben-Levi E. Imaging the endometrium: disease and normal variants. RadioGraphics 2001;21:1409e24. 5. Bennett GL, Andreotti RF, Lee SI, et al. ACR Appropriateness Criteria on abnormal vaginal bleeding. J Am Coll Radiol 2011;8:460e8. 6. Goldstein RB, Bree RL, Benson CB, et al. Evaluation of the woman with postmenopausal bleeding: Society of Radiologists in UltrasoundSponsored Consensus Conference statement. J Ultrasound Med 2001;20:1025e36. 7. Hricak H. MRI of the female pelvis: a review. AJR Am J Roentgenol 1986;146:1115e22. 8. Hardesty LA, Sumkin JH, Hakim C, et al. The ability of helical CT to preoperatively stage endometrial carcinoma. AJR Am J Roentgenol 2001;176:603e6. 9. Kaur H, Loyer EM, Minami M, et al. Patterns of uterine enhancement with helical CT. Eur J Radiol 1998;28:250e5. 10. Yitta S, Hecht EM, Mausner EV, et al. Normal or abnormal? Demystifying uterine and cervical contrast enhancement at multidetector CT. RadioGraphics 2011;31:647e61. 11. Grossman J, Ricci ZJ, Rozenblit A, et al. Efficacy of contrast-enhanced CT in assessing the endometrium. AJR Am J Roentgenol 2008;191:664e9. 12. Hodel J, Zins M, Desmottes L, et al. Location of the transition zone in CT of small-bowel obstruction: added value of multiplanar reformations. Abdom Imaging 2009;34:35e41. 13. Zangos S, Steenburg SD, Philllips KD, et al. Acute abdomen: added diagnostic value of coronal reformations with 64-slice multidetector row computed tomography. Acad Radiol 2007;14:19e27. 14. Neville AM, Paulson EK. MDCT of acute appendicitis: value of coronal reformations. Abdom Imaging 2009;34:42e8. 15. Tsili AC, Tsampoulas C, Dalkalitsis N, et al. Local staging of endometrial carcinoma: role of multidetector CT. Eur Radiol 2008;18:1043e8.
Please cite this article in press as: Kang SK, et al., Performance of multidetector CT in the evaluation of the endometrium: Measurement of endometrial thickness and detection of disease, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.06.013