G Model EURR-6626; No. of Pages 7
ARTICLE IN PRESS European Journal of Radiology xxx (2013) xxx–xxx
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Utility of conventional and diffusion-weighted MRI features in distinguishing benign from malignant endometrial lesions Andrea S. Kierans ∗ , Genevieve L. Bennett 1 , Mohammad Haghighi 1 , Andrew B. Rosenkrantz 1 Department of Radiology, NYU Langone Medical Center, 660 First Avenue, New York, NY 10016, United States
a r t i c l e
i n f o
Article history: Received 27 August 2013 Received in revised form 26 October 2013 Accepted 18 November 2013 Keywords: Endometrial abnormality Diffusion-weighted MRI
a b s t r a c t Purpose: To evaluate the utility of conventional MRI and diffusion-weighted imaging (DWI) in differentiating benign from malignant endometrial lesions. Methods: 52 patients with an abnormal endometrium on MRI and subsequent pathologic evaluation (35 benign, 17 malignant) were included. Two radiologists (R1, R2) independently evaluated endometrial abnormalities for characteristics on conventional MRI and DWI. Findings were assessed using unpaired t-tests, Fisher’s exact test, and multi-variate logistic regression. Results: Findings with significantly higher frequency in malignant abnormalities were: presence of irregularly marginated endometrial lesion (R1: 71% vs. 34%, R2: 94% vs. 26%), irregular endo-myometrial interface on T2WI (R1: 77% vs. 26%, R2: 94% vs. 29%), irregular endo-myometrial interface on post-contrast T1WI (R1: 82% vs. 23%, R2: 88% vs. 20%), increased signal on high b-value DWI (R1: 82% vs. 20%, R2: 94% vs. 20%), decreased ADC (R1: 88 vs. 40%, R2: 94% vs. 20%) (all p < 0.001, both readers). Endometrial thickness, presence of any focal endometrial lesion regardless of contour, diameter of endometrial lesion, endometrial heterogeneity on T2WI, decreased T2 signal, and increased endometrial enhancement, failed to show significant differences between groups (all p ≥ 0.159, both readers). At multivariate analysis, for R1, irregular endo-myometrial interface on post-contrast T1WI and increased DWI signal were significant independent predictors of malignancy (AUC = 0.89); for R2, only increased DWI signal was a significant independent predictor of malignancy (AUC = 0.87). Conclusion: Abnormal signal on DWI and irregularity of either the endo-myometrial interface or focal endometrial lesion were the most helpful MRI features in differentiating benign from malignant endometrial abnormalities. © 2013 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Evaluation of abnormalities of the endometrial cavity poses a significant diagnostic challenge for radiologists and gynecologists [1]. This challenge reflects a combination of the wide range of appearances and potentially overlapping imaging features of the normal endometrium, influenced by patient’s menopausal status and phase of menarche, and of a large spectrum of benign and malignant endometrial pathologies, including endometrial hyperplasia, submucosal fibroid, endometrial polyp, and endometrial
∗ Corresponding author at: Department of Radiology, Center for Biomedical Imaging, 660 First Avenue, NYU Langone Medical Center, New York, NY 10016, United States. Tel.: +1 212 263 0232; fax: +1 212 263 6634. E-mail addresses: [email protected] (A.S. Kierans), [email protected] (G.L. Bennett), [email protected] (M. Haghighi), [email protected] (A.B. Rosenkrantz). 1 Tel.: +1 212 263 0232; fax: +1 212 263 6634.
cancer [2]. In addition, such suspicion for endometrial pathology may be based on clinical grounds due to abnormal vaginal bleeding, or based on an abnormal appearance of the endometrium as an incidental finding on radiologic imaging performed for other indications. While endometrial biopsy and curettage serve as the mainstay for diagnosis, these procedures are invasive and not without complications, cause patient discomfort, cannot be performed or are non-diagnostic in 2–28% of attempts [3–5], and often yield inaccurate diagnoses due to sampling error [6,7]. Thus, reliable non-invasive methods are required to assist in formation of a preoperative diagnosis and appropriate triage of patients for more invasive testing such as endometrial biopsy and D&C. Given its low cost, ease of use, and proven accuracy, transvaginal ultrasonography is generally considered the primary imaging modality for initial evaluation of suspected endometrial pathology [8,9]. However, key sonographic findings of endometrial thickening and heterogeneity, as well as of the presence of a focal endometrial lesion, are non-specific and overlap between the previously noted benign and malignant disorders [10,11].
0720-048X/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejrad.2013.11.030
Please cite this article in press as: Kierans AS, et al. Utility of conventional and diffusion-weighted MRI features in distinguishing benign from malignant endometrial lesions. Eur J Radiol (2013), http://dx.doi.org/10.1016/j.ejrad.2013.11.030
G Model EURR-6626; No. of Pages 7
ARTICLE IN PRESS A.S. Kierans et al. / European Journal of Radiology xxx (2013) xxx–xxx
2
Alternatively, magnetic resonance imaging (MRI) may also play a key role in the evaluation of suspected endometrial pathology [8,12–15]. Indeed, a number of studies have identified a variety of MRI features related to endometrial morphology, signal intensity, and enhancement characteristic as useful for distinguishing between benign and malignant endometrial pathology [2,16–19]. Nonetheless, findings have been variable, if not conflicting, between studies [12]; overlap between benign and malignant cases continues to be reported [12]; and the role of conventional MRI in evaluation of suspected endometrial pathology remains unclear [18]. More recently, diffusion weighted imaging (DWI) has been applied for characterization of endometrial lesions. For instance, a number of studies [12,20–22] have demonstrated restricted diffusion in endometrial carcinoma than in other benign endometrial lesions. However, the actual clinical role of DWI in evaluating suspected endometrial pathology remains uncertain from such studies. As these studies evaluated DWI findings in isolation, the relative performance of DWI and conventional MRI findings with respect to one another in evaluation of endometrial pathology was not assessed in these studies. More important, whether DWI provides additive value to conventional MRI is unknown. Therefore, the purpose of the current investigation is to evaluate the utility of both conventional MRI and DWI in differentiation of benign from malignant endometrial abnormalities. 2. Materials and method 2.1. Patients This retrospective study was compliant with the Health Insurance Portability and Accountability Act and approved by our Institutional Review Board, with waiver of written informed consent. One author (AK, a third-year radiology resident), serving as the data coordinator for the study, searched a radiologic database to identify adult women who underwent MRI of the pelvis between January 2006 and December 2011 that reported an abnormal appearance of the endometrium, and for whom a subsequent pathologic diagnosis was available. In 29 cases, a gold standard diagnosis of hysterectomy following endometrial biopsy was available; pathologic diagnosis was achieved via biopsy in the remaining 23 cases. DWI was included in the routine female pelvis MRI protocol at our institution during this time. Initially, 69 patients were identified. Patients were then excluded for the following reasons: severe artifact on DWI (n = 1); lack of intravenous contrast (n = 3); delay of greater than six months between MRI and surgery (n = 7); and prior treatment for known endometrial carcinoma prior to MRI (n = 6). These exclusions yielded a final study cohort of 52 patients (mean age 56.4 ± 16.6 years, range 25–93 years). 2.2. MRI technique All patients underwent pelvic MRI on either a 1.5 T [MAGNETOM Avanto (n = 22), Symphony (n = 13), or Sonata (n = 7]), Siemens Healthcare, Erlangen, Germany)] or 3 T [MAGNETOM Trio (n = 10), Siemens Health Care, Erlangen, Germany] clinical systems using a torso phased-array coil. Acquisition parameters of the included sequence varied somewhat between cases given the long time period of the study, as well as use of different scanners and imaging planes and slight adjustment of parameters by the technologists based on patient size. However, representative parameters for sequences included in all examinations are as follows: multiplanar turbo-spin echo T2-weighted imaging (T2WI) angled to the plane of the uterus and endometrium [TR/TE 5000–8500/95–115 ms; slice thickness 4 mm; 30% section
gap; field of view (FOV) 250–275 mm with 80–100% rectangular FOV; matrix 512 × 200–210; 2 signal averages; receiver bandwidth 195–200 Hz/voxel, parallel imaging factor 2], axial inand-opposed-phase gradient-echo T1-weighted imaging (T1WI) [TR 150–240; TE 2.0–2.2/5.0–5.3 at 1.5 T, 1.5/4.4 ms at 3 T; flip angle (FA) 80◦ ; slice thickness 6–8 mm; 20% section gap; FOV 350–360 mm with 80% rectangular FOV; matrix 256 × 180–220; 1 signal average; receiver bandwidth 380–930 Hz/voxel; parallel imaging factor 2], and dynamic 3D fat-suppressed spoiled gradientecho T1WI (TR/TE 3.2–4.2/1.2–2.0 ms; FA 10–15◦ ; slice thickness 1.4–3.0 mm; 20% section gap; FOV 220–275 mm with 100% rectangular FOV; matrix 256 × 130–146; 1 signal average; receiver bandwidth 490–530 Hz/voxel; parallel imaging factor 2) performed before and at multiple time points following intravenous administration of 0.1 mmol/kg of gadopentetate dimeglumine (Magnevist, Bayer HealthCare). All examinations also included a fat-suppressed single-shot echo-planar DWI sequence of the pelvis performed using tridirectional motion-probing gradients and b-values of 0, 50, and 500 s/mm2 (n = 13) or 0, 500, and 1000 s/mm2 (n = 39), with inline reconstruction of the apparent diffusion coefficient (ADC) map and the following parameters: TR/TE 2100–2500/76–82 ms; slice thickness 6–8 mm; FOV 350 mm with 75–80% rectangular FOV; matrix 144 × 192; 3 signal averages; receiver bandwidth 1300 Hz/voxel. DWI was performed in the sagittal plane in cases ordered for suspected endometrial or other uterine pathology and in the axial plane in cases ordered for suspected pathology of the ovaries or other extra-uterine pelvic structures [23]. 2.3. Qualitative analysis Images were reviewed independently by two radiologists fellowship-trained in abdominal imaging (GB with 16 years of experience, including dedicated expertise in women’s imaging, and MH with 2 years of experience in body MRI), both blinded to histologic findings. For each patient, the readers recorded the presence or absence of each of the following MRI features: focal endometrial lesion regardless of contour, focal endometrial lesion with irregular margins, irregular endo-myometrial interface on T2WI, irregular endo-myometrial interface on post-contrast T1WI, endometrial heterogeneity on T2WI, decreased T2 signal of the endometrial abnormality, increased T1 signal of the endometrial abnormality relative to normal endometrium; increased enhancement of the endometrial abnormality on early post-contrast T1WI, increased enhancement of the endometrial abnormality on late post-contrast T1WI, increased signal of the endometrial abnormality on DWI when compared to the normal visualized endometrium using the highest obtained b-value, and decreased signal of the endometrial abnormality on the ADC map when compared to the normal visualized endometrium. Increased DWI and decreased ADC signal was determined as compared to normal outer myometrium. In addition, the observers measured the maximal endometrial thickness on sagittal T2-weighted images as well as the maximal diameter of the focal endometrial abnormality in any plane in cases scored as positive for presence of a focal endometrial lesion. 2.4. Statistical analysis Fisher’s exact test was used to assess for significant differences in the frequency of each binary imaging feature between benign and malignant endometrial abnormalities. Unpaired t-tests were used to assess for significant differences in quantitative features between these two groups. Multivariate logistic regression with stepwise variable selection was used to identify for each reader those features showing significant differences at univariate analysis that represented significant independent predictors of endometrial malignancy. For the binary factors, the percentage of endometrial
Please cite this article in press as: Kierans AS, et al. Utility of conventional and diffusion-weighted MRI features in distinguishing benign from malignant endometrial lesions. Eur J Radiol (2013), http://dx.doi.org/10.1016/j.ejrad.2013.11.030
G Model EURR-6626; No. of Pages 7
ARTICLE IN PRESS A.S. Kierans et al. / European Journal of Radiology xxx (2013) xxx–xxx
3
Table 1 Comparison of MRI features between benign and malignant endometrial lesions for both readers. MRI feature
Focal endometrial lesion regardless of contour Focal endometrial lesion with irregular margins Irregular endo-myometrial interface on T2WI Irregular endo-myometrial interface on post-contrast T1WI Endometrial heterogeneity Decreased T2 signal of endometrial abnormality Increased T1 signal of endometrial abnormality Increased enhancement of endometrial abnormality on early post-contrast T1WI images Increased enhancement of endometrial abnormality on late post-contrast T1WI images Increased signal of endometrial abnormality on DWI using the highest obtained b-value Decreased signal of endometrial abnormality on ADC map
Reader 1
Reader 2
Benign
Malignant
p
Benign
Malignant
p
71.4% (25/35) 34.3% (12/35) 25.7% (9/35) 22.9% (8/35)
70.6% (12/17) 70.6% (12/17) 76.5% (13/17) 82.4% (14/17)
1.000 0.019
ARTICLE IN PRESS European Journal of Radiology xxx (2013) xxx–xxx
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Utility of conventional and diffusion-weighted MRI features in distinguishing benign from malignant endometrial lesions Andrea S. Kierans ∗ , Genevieve L. Bennett 1 , Mohammad Haghighi 1 , Andrew B. Rosenkrantz 1 Department of Radiology, NYU Langone Medical Center, 660 First Avenue, New York, NY 10016, United States
a r t i c l e
i n f o
Article history: Received 27 August 2013 Received in revised form 26 October 2013 Accepted 18 November 2013 Keywords: Endometrial abnormality Diffusion-weighted MRI
a b s t r a c t Purpose: To evaluate the utility of conventional MRI and diffusion-weighted imaging (DWI) in differentiating benign from malignant endometrial lesions. Methods: 52 patients with an abnormal endometrium on MRI and subsequent pathologic evaluation (35 benign, 17 malignant) were included. Two radiologists (R1, R2) independently evaluated endometrial abnormalities for characteristics on conventional MRI and DWI. Findings were assessed using unpaired t-tests, Fisher’s exact test, and multi-variate logistic regression. Results: Findings with significantly higher frequency in malignant abnormalities were: presence of irregularly marginated endometrial lesion (R1: 71% vs. 34%, R2: 94% vs. 26%), irregular endo-myometrial interface on T2WI (R1: 77% vs. 26%, R2: 94% vs. 29%), irregular endo-myometrial interface on post-contrast T1WI (R1: 82% vs. 23%, R2: 88% vs. 20%), increased signal on high b-value DWI (R1: 82% vs. 20%, R2: 94% vs. 20%), decreased ADC (R1: 88 vs. 40%, R2: 94% vs. 20%) (all p < 0.001, both readers). Endometrial thickness, presence of any focal endometrial lesion regardless of contour, diameter of endometrial lesion, endometrial heterogeneity on T2WI, decreased T2 signal, and increased endometrial enhancement, failed to show significant differences between groups (all p ≥ 0.159, both readers). At multivariate analysis, for R1, irregular endo-myometrial interface on post-contrast T1WI and increased DWI signal were significant independent predictors of malignancy (AUC = 0.89); for R2, only increased DWI signal was a significant independent predictor of malignancy (AUC = 0.87). Conclusion: Abnormal signal on DWI and irregularity of either the endo-myometrial interface or focal endometrial lesion were the most helpful MRI features in differentiating benign from malignant endometrial abnormalities. © 2013 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Evaluation of abnormalities of the endometrial cavity poses a significant diagnostic challenge for radiologists and gynecologists [1]. This challenge reflects a combination of the wide range of appearances and potentially overlapping imaging features of the normal endometrium, influenced by patient’s menopausal status and phase of menarche, and of a large spectrum of benign and malignant endometrial pathologies, including endometrial hyperplasia, submucosal fibroid, endometrial polyp, and endometrial
∗ Corresponding author at: Department of Radiology, Center for Biomedical Imaging, 660 First Avenue, NYU Langone Medical Center, New York, NY 10016, United States. Tel.: +1 212 263 0232; fax: +1 212 263 6634. E-mail addresses: [email protected] (A.S. Kierans), [email protected] (G.L. Bennett), [email protected] (M. Haghighi), [email protected] (A.B. Rosenkrantz). 1 Tel.: +1 212 263 0232; fax: +1 212 263 6634.
cancer [2]. In addition, such suspicion for endometrial pathology may be based on clinical grounds due to abnormal vaginal bleeding, or based on an abnormal appearance of the endometrium as an incidental finding on radiologic imaging performed for other indications. While endometrial biopsy and curettage serve as the mainstay for diagnosis, these procedures are invasive and not without complications, cause patient discomfort, cannot be performed or are non-diagnostic in 2–28% of attempts [3–5], and often yield inaccurate diagnoses due to sampling error [6,7]. Thus, reliable non-invasive methods are required to assist in formation of a preoperative diagnosis and appropriate triage of patients for more invasive testing such as endometrial biopsy and D&C. Given its low cost, ease of use, and proven accuracy, transvaginal ultrasonography is generally considered the primary imaging modality for initial evaluation of suspected endometrial pathology [8,9]. However, key sonographic findings of endometrial thickening and heterogeneity, as well as of the presence of a focal endometrial lesion, are non-specific and overlap between the previously noted benign and malignant disorders [10,11].
0720-048X/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejrad.2013.11.030
Please cite this article in press as: Kierans AS, et al. Utility of conventional and diffusion-weighted MRI features in distinguishing benign from malignant endometrial lesions. Eur J Radiol (2013), http://dx.doi.org/10.1016/j.ejrad.2013.11.030
G Model EURR-6626; No. of Pages 7
ARTICLE IN PRESS A.S. Kierans et al. / European Journal of Radiology xxx (2013) xxx–xxx
2
Alternatively, magnetic resonance imaging (MRI) may also play a key role in the evaluation of suspected endometrial pathology [8,12–15]. Indeed, a number of studies have identified a variety of MRI features related to endometrial morphology, signal intensity, and enhancement characteristic as useful for distinguishing between benign and malignant endometrial pathology [2,16–19]. Nonetheless, findings have been variable, if not conflicting, between studies [12]; overlap between benign and malignant cases continues to be reported [12]; and the role of conventional MRI in evaluation of suspected endometrial pathology remains unclear [18]. More recently, diffusion weighted imaging (DWI) has been applied for characterization of endometrial lesions. For instance, a number of studies [12,20–22] have demonstrated restricted diffusion in endometrial carcinoma than in other benign endometrial lesions. However, the actual clinical role of DWI in evaluating suspected endometrial pathology remains uncertain from such studies. As these studies evaluated DWI findings in isolation, the relative performance of DWI and conventional MRI findings with respect to one another in evaluation of endometrial pathology was not assessed in these studies. More important, whether DWI provides additive value to conventional MRI is unknown. Therefore, the purpose of the current investigation is to evaluate the utility of both conventional MRI and DWI in differentiation of benign from malignant endometrial abnormalities. 2. Materials and method 2.1. Patients This retrospective study was compliant with the Health Insurance Portability and Accountability Act and approved by our Institutional Review Board, with waiver of written informed consent. One author (AK, a third-year radiology resident), serving as the data coordinator for the study, searched a radiologic database to identify adult women who underwent MRI of the pelvis between January 2006 and December 2011 that reported an abnormal appearance of the endometrium, and for whom a subsequent pathologic diagnosis was available. In 29 cases, a gold standard diagnosis of hysterectomy following endometrial biopsy was available; pathologic diagnosis was achieved via biopsy in the remaining 23 cases. DWI was included in the routine female pelvis MRI protocol at our institution during this time. Initially, 69 patients were identified. Patients were then excluded for the following reasons: severe artifact on DWI (n = 1); lack of intravenous contrast (n = 3); delay of greater than six months between MRI and surgery (n = 7); and prior treatment for known endometrial carcinoma prior to MRI (n = 6). These exclusions yielded a final study cohort of 52 patients (mean age 56.4 ± 16.6 years, range 25–93 years). 2.2. MRI technique All patients underwent pelvic MRI on either a 1.5 T [MAGNETOM Avanto (n = 22), Symphony (n = 13), or Sonata (n = 7]), Siemens Healthcare, Erlangen, Germany)] or 3 T [MAGNETOM Trio (n = 10), Siemens Health Care, Erlangen, Germany] clinical systems using a torso phased-array coil. Acquisition parameters of the included sequence varied somewhat between cases given the long time period of the study, as well as use of different scanners and imaging planes and slight adjustment of parameters by the technologists based on patient size. However, representative parameters for sequences included in all examinations are as follows: multiplanar turbo-spin echo T2-weighted imaging (T2WI) angled to the plane of the uterus and endometrium [TR/TE 5000–8500/95–115 ms; slice thickness 4 mm; 30% section
gap; field of view (FOV) 250–275 mm with 80–100% rectangular FOV; matrix 512 × 200–210; 2 signal averages; receiver bandwidth 195–200 Hz/voxel, parallel imaging factor 2], axial inand-opposed-phase gradient-echo T1-weighted imaging (T1WI) [TR 150–240; TE 2.0–2.2/5.0–5.3 at 1.5 T, 1.5/4.4 ms at 3 T; flip angle (FA) 80◦ ; slice thickness 6–8 mm; 20% section gap; FOV 350–360 mm with 80% rectangular FOV; matrix 256 × 180–220; 1 signal average; receiver bandwidth 380–930 Hz/voxel; parallel imaging factor 2], and dynamic 3D fat-suppressed spoiled gradientecho T1WI (TR/TE 3.2–4.2/1.2–2.0 ms; FA 10–15◦ ; slice thickness 1.4–3.0 mm; 20% section gap; FOV 220–275 mm with 100% rectangular FOV; matrix 256 × 130–146; 1 signal average; receiver bandwidth 490–530 Hz/voxel; parallel imaging factor 2) performed before and at multiple time points following intravenous administration of 0.1 mmol/kg of gadopentetate dimeglumine (Magnevist, Bayer HealthCare). All examinations also included a fat-suppressed single-shot echo-planar DWI sequence of the pelvis performed using tridirectional motion-probing gradients and b-values of 0, 50, and 500 s/mm2 (n = 13) or 0, 500, and 1000 s/mm2 (n = 39), with inline reconstruction of the apparent diffusion coefficient (ADC) map and the following parameters: TR/TE 2100–2500/76–82 ms; slice thickness 6–8 mm; FOV 350 mm with 75–80% rectangular FOV; matrix 144 × 192; 3 signal averages; receiver bandwidth 1300 Hz/voxel. DWI was performed in the sagittal plane in cases ordered for suspected endometrial or other uterine pathology and in the axial plane in cases ordered for suspected pathology of the ovaries or other extra-uterine pelvic structures [23]. 2.3. Qualitative analysis Images were reviewed independently by two radiologists fellowship-trained in abdominal imaging (GB with 16 years of experience, including dedicated expertise in women’s imaging, and MH with 2 years of experience in body MRI), both blinded to histologic findings. For each patient, the readers recorded the presence or absence of each of the following MRI features: focal endometrial lesion regardless of contour, focal endometrial lesion with irregular margins, irregular endo-myometrial interface on T2WI, irregular endo-myometrial interface on post-contrast T1WI, endometrial heterogeneity on T2WI, decreased T2 signal of the endometrial abnormality, increased T1 signal of the endometrial abnormality relative to normal endometrium; increased enhancement of the endometrial abnormality on early post-contrast T1WI, increased enhancement of the endometrial abnormality on late post-contrast T1WI, increased signal of the endometrial abnormality on DWI when compared to the normal visualized endometrium using the highest obtained b-value, and decreased signal of the endometrial abnormality on the ADC map when compared to the normal visualized endometrium. Increased DWI and decreased ADC signal was determined as compared to normal outer myometrium. In addition, the observers measured the maximal endometrial thickness on sagittal T2-weighted images as well as the maximal diameter of the focal endometrial abnormality in any plane in cases scored as positive for presence of a focal endometrial lesion. 2.4. Statistical analysis Fisher’s exact test was used to assess for significant differences in the frequency of each binary imaging feature between benign and malignant endometrial abnormalities. Unpaired t-tests were used to assess for significant differences in quantitative features between these two groups. Multivariate logistic regression with stepwise variable selection was used to identify for each reader those features showing significant differences at univariate analysis that represented significant independent predictors of endometrial malignancy. For the binary factors, the percentage of endometrial
Please cite this article in press as: Kierans AS, et al. Utility of conventional and diffusion-weighted MRI features in distinguishing benign from malignant endometrial lesions. Eur J Radiol (2013), http://dx.doi.org/10.1016/j.ejrad.2013.11.030
G Model EURR-6626; No. of Pages 7
ARTICLE IN PRESS A.S. Kierans et al. / European Journal of Radiology xxx (2013) xxx–xxx
3
Table 1 Comparison of MRI features between benign and malignant endometrial lesions for both readers. MRI feature
Focal endometrial lesion regardless of contour Focal endometrial lesion with irregular margins Irregular endo-myometrial interface on T2WI Irregular endo-myometrial interface on post-contrast T1WI Endometrial heterogeneity Decreased T2 signal of endometrial abnormality Increased T1 signal of endometrial abnormality Increased enhancement of endometrial abnormality on early post-contrast T1WI images Increased enhancement of endometrial abnormality on late post-contrast T1WI images Increased signal of endometrial abnormality on DWI using the highest obtained b-value Decreased signal of endometrial abnormality on ADC map
Reader 1
Reader 2
Benign
Malignant
p
Benign
Malignant
p
71.4% (25/35) 34.3% (12/35) 25.7% (9/35) 22.9% (8/35)
70.6% (12/17) 70.6% (12/17) 76.5% (13/17) 82.4% (14/17)
1.000 0.019
