| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, University of Modena and Reggio-Emilia, Policlinico,
Via del Pozzo 71, 41100 Modena, Italy.
2 Department of Gynecology, University of Modena and Reggio-Emilia, Modena,
Italy.
3 Department of Pathology, University of Modena and Reggio-Emilia, Modena,
Italy.
Received April 6, 2007;
accepted after revision September 5, 2007.
Address correspondence to P. Torricelli
(torricelli.pietro{at}unimore.it).
Abstract
 Top Abstract Introductionsubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Fifty-two patients (43 postmenopausal) with histopathologically proven endometrial carcinoma underwent preoperative 3-T MRI. The following sequences were performed: axial T1 fast spin-echo (FSE); axial, parasagittal, and paracoronal T2 FSE; paracoronal 3D T1 inversion recovery gradient-echo after contrast administration; and parasagittal fat-suppressed T1 FSE. All patients underwent a hysterectomy. The MRI findings were compared with histopathology results. The quantity and degree of artifacts were evaluated.
RESULTS. MRI performed on a 3-T unit was in agreement with histopathology in assessing the depth of invasion in 86.4% (44/52) of the patients with a mean sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of 83.5%, 93.9%, 77.8%, 92.2%, and 89.7%, respectively. Performance values were also assessed for single stages of myometrial infiltration. For the detection of an intramucosal lesion (MRI, 12/52; histopathology, 6/52), sensitivity was 100%; specificity, 86.9%; PPV, 50%; NPV, 100%; and accuracy, 88.5%. For the detection of myometrial infiltration that was less than 50% (MRI, 12/52; histopathology, 16/52), sensitivity was 62.5%; specificity, 94.4%; PPV, 83.3%; NPV, 85%; and accuracy, 84.6%. For the detection of myometrial infiltration that was greater than 50% (MRI, 28/52; histopathology, 30/52), sensitivity was 93.3%; specificity, 100%; PPV, 100%; NPV, 91.7%; and accuracy, 96.2%. The following artifacts were found: abdominal wall movement, nine patients (not affecting image quality); peristalsis, 16 patients (two deeply affecting, one affecting, and 13 scarcely affecting); magnetic susceptibility artifact, four patients (not affecting); chemical shift, 20 patients (four scarcely affecting and 16 not affecting); and dielectric effect, six patients (four deeply affecting and two affecting).
CONCLUSION. In evaluating the depth of myometrial infiltration in patients with endometrial cancer, 3-T MRI showed high diagnostic accuracy—equivalent to that of 1.5-T MRI reported in the literature. Artifacts did not significantly affect image quality.
Keywords: 3-T MRI • endometrial carcinoma • high-field-strength MRI • myometrial invasion • staging • women's imaging
|
TopAbstractIntroductionsubjects and MethodsResultsDiscussionReferences |
|---|
The aim of this study was to evaluate the diagnostic accuracy and the limits of 3-T MRI in determining the depth of myometrial infiltration in patients with endometrial cancer.
|
TopAbstractIntroductionsubjects and MethodsResultsDiscussionReferences |
|---|
All patients gave written informed consent to undergo the study protocol, and the study was approved by our institutional review board. All underwent 3-T MRI (Intera, Philips Medical Systems) using a six-channel phased-array coil (Sense Cardio, Philips). Twenty milligrams of butyl-scopolamine (Buscopan, Schering) was IV administered to all patients except when contraindicated before the examination to prevent peristalsis artifacts. All patients were asked to fast for at least 2 hours before the procedure and to void before the examination to reduce full-bladder compression of the uterus. A respiratory-gating system was applied to minimize breathing movement artifacts. In 51 of the 52 cases, a rest slab was positioned in the anterior part of the abdomen to reduce artifacts due to abdominal wall movements.
MR Sequences
Axial T1-weighted fast spin-echo (FSE) imaging (TR/TE, 574/10) was
performed from the renal hila to the pubic symphysis using the following
parameters: slice thickness, 5 mm; intersection gap, 1 mm; number of signal
averages (NSA), 2; matrix, 256 x 256; field of view (FOV), 20–25
cm, according to patient's body habitus; and voxel size, 0.98 x 0.97
x 5 mm.
Axial, parasagittal, and paracoronal T2-weighted FSE imaging (TR range/TE, 4,299–4,893/100) was performed using the following parameters: slice thickness, 4 mm; intersection gap, 1 mm; NSA, 2; matrix, 512 x 512; FOV, 20–25 cm; and voxel size, 0.98 x 0.98 x 5 mm. The parasagittal scans were obtain parallel to the long uterine axis, whereas the paracoronal scans were obtained orthogonal to the long uterine axis and parallel to the short uterine axis. The axial scans were obtained with fat saturation.
To reduce the specific absorption rate, we combined the use of long-TR sequences, parallel imaging (sensitivity encoding [SENSE]), and a refocusing flip angle.
Paracoronal 3D T1-weighted high-resolution isotropic volume examination (THRIVE) sequences were performed: TR/TE, 6.3/3; flip angle, 10°; slice thickness, 3 mm; intersection gap, over continuous slices; NSA, 1; matrix, 192 x 256; FOV, 20–25 cm; and voxel size, 0.98 x 0.97 x 5 mm. The paracoronal scans were obtained orthogonal to the long uterine axis and parallel to the short uterine axis. This sequence was used for a dynamic contrast-enhanced study for which images were acquired at four different phases relative to the IV injection of contrast agent (0.2 mL/kg of gadoterate dimeglumine [Dotarem, Guerbet]): basal (before gadolinium injection), arterial (30 seconds after injection), venous (60 seconds), and late (120 seconds). Most of the images used for diagnostic assessment were venous phase images.
Parasagittal fat-suppressed T1-weighted FSE imaging (589/10) was performed 180 seconds after gadolinium injection. This sequence was performed when cancer was suspected to have spread beyond the uterine serosa to the parametrial tissue according to previous sequence findings.
Assessment of MR Images
Two radiologists independently assessed the depth of myometrial
infiltration by evaluating both the T2-weighted sequences (axial,
parasagittal, and paracoronal) and the paracoronal contrast-enhanced 3D
T1-weighted inversion recovery sequence, and interobserver variability was
reported. In cases of discordance between the two sequences—T2-weighted
and 3D T1-weighted after contrast administration—the latter was
considered to be the more reliable sequence
[10–14].
The MRI criteria used to assess the degree of the myometrial invasion on both T2-weighted and contrast-enhanced T1-weighted sequences were described by Frei and Kinkel [25] and are reported in Table 1. MRI assessment of myometrial infiltration was performed according to the surgical–pathologic classification criteria proposed by the International Federation of Gynecology and Obstetrics (FIGO), which classifies stage I endometrial cancer as follows: IA, tumor is restricted to the endometrium without myometrial infiltration; IB, tumor infiltrates less than 50% of the thickness of the myometrium; and IC, tumor infiltrates more than 50% of the myometrium.
|
Furthermore, the presence of artifacts on images obtained of each patient was assessed. The influence of the artifacts on the image diagnostic value was graded on a scale of 1–4, as follows: 1 (not affecting image quality), artifacts did not influence diagnostic image quality; 2 (scarcely affecting), artifacts caused minor loss in quality but did not interfere with ability to assess myometrial invasion; 3 (affecting), artifacts caused a significant loss of quality and impaired myometrial invasion assessment; and 4 (deeply affecting), artifacts rendered an image uninterpretable. Images evaluated as grades 3 and 4 were not considered diagnostic. Two radiologists independently assessed the impact of artifacts on image diagnostic value, and interobserver agreement was expressed as a kappa value.
Within 1 month after MRI examination, all patients underwent hysterectomy. The whole surgical specimen was sectioned along the coronal plane orthogonal to the main uterine axis, and histologic macrosections were obtained and routinely stained with H and E. The degree of myometrial infiltration was subsequently quantified according to the FIGO criteria.
A comparative evaluation of MRI findings and histopathologic results was performed. The anatomic macrosections were considered the gold standard.
Statistical Analysis
Sensitivity, specificity, positive predictive value (PPV), negative
predictive value (NPV), and diagnostic accuracy of 3-T MRI were calculated
(Stata software, Stata Corporation). Furthermore, the impact of artifacts on
diagnosis was assessed.
The kappa statistic (Stata, release 9.0, Stata Corporation) was used to estimate interobserver agreement. The thresholds used for interpreting the results were as follows: < 0.0, poor; 0.00–0.20, slight; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, substantial; and 0.81–1.00, almost perfect.
|
TopAbstractIntroductionsubjects and MethodsResultsDiscussionReferences |
|---|
Anatomic evaluation showed intramucosal tumor in six patients (stage pIA) and myometrial infiltration of less than 50% in 16 patients (stage pIB) and greater than 50% in 30 (stage pIC).
MRI evaluation using 3 T depicted 12 cases of intramucosal tumor, 12 cases of myometrial infiltration less than 50% (Figs. 1A, 1B, 1C and 2A, 2B, 2C), and 28 cases of myometrial infiltration greater than 50% (Figs. 3A, 3B, 3C and 4A, 4B).
|
|
|
|
|
|
|
|
|
|
|
Statistical evaluations of 3-T MRI showed an overall sensitivity of 83.5% (95% CI, > 0.748 to < 0.944), specificity of 93.9% (> 0.872 to < 0.974), PPV of 77.8%, and NPV of 92.2%. The overall accuracy was 89.7%.
For single stages of myometrial invasion, the statistical values are summarized in Table 2. The 95% CIs were also calculated by stage for the following data: for an intramucosal lesion, specificity was 86.9% (95% CI, > 0.772 to < 0.967); for myometrial infiltration less than 50%, sensitivity was 62.5% (> 0.388 to < 0.862) and specificity was 94.4% (> 0.870 to < 1.019); for myometrial infiltration greater than 50%, sensitivity was 93.3% (> 0.844 to < 1.023).
|
Agreement between the two radiologists was almost perfect (
=
0.92).
Artifacts
The following artifacts were found and are summarized in
Table 3: Abdominal wall
movement artifact, peristaltic movements, magnetic susceptibility artifact,
chemical shift, and dielectric effects.
|
Agreement between the two radiologists was almost perfect ( =
0.85).
|
TopAbstractIntroductionsubjects and MethodsResultsDiscussionReferences |
|---|
In our study, a high specificity and PPV were found for 3-T MRI, as correlated with pathologic results, in evaluating the depth of myometrial infiltration. In particular, the high specificity (93.3%) and PPV (100%) obtained in detecting deep myometrial infiltration should be highlighted because, as has been described by others [6, 13, 14], deep (> 50%) myometrial infiltration is linked to lymph node positivity and is one of the most important factors affecting prognosis and therapy.
Eight of the 52 cases were understaged: in three of the eight cases, slight myometrial infiltration was identified only at histopathologic evaluation and was not detectable even when the MR images were reviewed retrospectivvely. In the remaining five cases, understaging was due to poor visualization or a lack of visualization of both the junctional zone in the T2-weighted sequences and the subendometrial enhancement in the contrast-enhanced T1-weighted sequences, likely due to the postmenopausal status of the patients and uterine atrophy. As others have reported, disruption of the junctional zone and disruption of subendometrial enhancement are the main MRI signs of myometrial infiltration, and when these signs are not clearly depicted, the diagnosis of superficial infiltration may be rather difficult and may even be impossible [27–30].
Schick [18] has stated that the image quality that can be obtained with 3 T is no higher than that achieved with state-of-the-art 1.5 T. A comparison of the results of our study performed using 3-T MRI and those reported in the most recent literature using lower-field-strength MRI did not show a clear superiority of the former (Table 4), although it is possible to conclude that the overall diagnostic accuracy of 3 T is equal to the highest values obtainable at lower magnetic field intensity [14, 27, 31].
|
However, the results of this study are lower than the sensitivity obtained by Nakao et al. [8] and the PPV by Cunha et al. [32]. The specificity, NPV, and diagnostic accuracy obtained at 3 T in this study are higher than those obtained with 1.5-T MRI in earlier studies, except the study performed by Cunha et al., which had 100% specificity and 93% accuracy.
The use of magnetic field strength equal to or higher than 3 T is limited in comparison with 1.5-T MRI by artifacts that may, in particular, affect abdominal and pelvic evaluations. The most commonly described artifacts are field inhomogeneity; greater magnetic susceptibility; and the dielectric effect, which is caused by reduced radiofrequency penetration power [18, 20, 22, 26].
In this study, the artifacts related to the high field significantly affected the diagnostic value of the images in nine of the 52 cases. The most significant artifacts proved to be the dielectric artifacts, which were observed in six obese patients. These artifacts, which are due to insufficient radiofrequency penetration, have been widely described [33, 34] and can be corrected by placing a gel pad on the anterior abdominal wall.
The chemical shift phenomenon is also strictly related to field strength doubles from 1.5 to 3 T [22]. In our series, chemical shift artifacts were found on the anterior and posterior surfaces of the uterus in the sagittal sequences and in the lateral bladder wall in the axial sequences, but these artifacts did not affect image quality because visualization of the zonal anatomy of the uterus was not impaired.
Magnetic susceptibility artifacts double from 1.5 to 3 T and limit the use of 3-T MRI in patients with metallic prostheses or devices [35]. In our series, these artifacts were found in four women with hip prostheses, but the artifacts did not affect diagnosis because these artifacts did not reach the uterus. Other investigators did not find a significant difference with regard to susceptibility artifacts between 1.5- and 3-T MRI of the liver [34].
The artifacts related to pelvic wall and intestinal movements can usually be reduced by positioning a rest slab on the anterior abdominal wall or by using antiperistaltic agents, respectively. In the present study, these artifacts affected the diagnostic value of the images in only three of the 52 cases. On the contrary, von Falkenhausen et al. [34] found a significant difference in the number and severity of these artifacts in evaluating the liver at 1.5 and 3 T. These artifacts are probably more common and severe in the upper abdomen than in the pelvis.
The quantity of radiofrequency deposition in human tissue was not evaluated in this study. However, sequence limitations related to specific absorption rate were not found because we used sequences with a prolonged TR and with a reduced refocusing flip angle (130°) in the T2-weighted sequences, as suggested by Morakkabati-Spitz et al. [24]. Moreover, we used parallel imaging, which reduces the quantity of radiofrequency deposition.
The results of the present study do not justify a preferential use of 3 T for the evaluation of myometrial infiltration in endometrial cancer patients, especially in light of technical, organizational, and cost issues associated with the use of this technique.
The routine use of 3-T MRI for endometrial cancer staging can be considered only if additional studies directly comparing pelvic 3- and 1.5-T MRI show a significant superiority of the former.
The results of our study suggest that 3-T MRI has a high diagnostic accuracy for endometrial cancer staging without the excessive detrimental artifacts typical of high magnetic fields. The diagnostic accuracy is equivalent to the highest accuracy values reported in the literature using 1.5-T MRI.
To prove the superiority of 3 T for endometrial cancer staging, conducting both examinations on the same cohort of patients would be necessary, with statistical validation in a larger patient population.
|
TopAbstractIntroductionsubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Hori, T. Kim, T. Murakami, I. Imaoka, H. Onishi, K. Tomoda, T. Tsutsui, T. Enomoto, T. Kimura, and H. Nakamura Uterine Cervical Carcinoma: Preoperative Staging with 3.0-T MR Imaging--Comparison with 1.5-T MR Imaging Radiology, April 1, 2009; 251(1): 96 - 104. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Alcazar, R. Galvan, S. Albela, S. Martinez, J. Pahisa, M. Jurado, and G. Lopez-Garcia Assessing Myometrial Infiltration by Endometrial Cancer: Uterine Virtual Navigation with Three-dimensional US Radiology, March 1, 2009; 250(3): 776 - 783. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
