| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, St. Mary's Hospital, The Catholic University of
Korea, College of Medicine, 62, Youido-Dong, Yongdungpo-Gu, Seoul 150-010,
Korea.
2
Department of General Surgery, The Catholic University of Korea, College of
Medicine, Seoul 150-010, Korea.
3
Department of Clinical Pathology, The Catholic University of Korea, College of
Medicine, Seoul 150-010, Korea.
Received June 18, 1999;
accepted after revision November 8, 1999.
Presented at the annual meeting of the American Roentgen Ray Society, New
Orleans, May 1999.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Thirty patients with gastric carcinoma underwent pre-operative MR imaging and helical CT. MR imaging at 1.5 T was performed immediately after the intramuscular injection of scopolamine and the oral administration of water or effervescent granules. Breath-hold 2D FLASH T1-weighted images in all three planes, turbo spin-echo T2-weighted axial images, and gadolinium-enhanced fat-suppressed 2D FLASH axial images were included. Helical CT was performed 60 sec after initiation of IV contrast medium injection (2.5-3 ml/sec). Two groups of two radiologists each independently analyzed the MR and helical CT findings, and these results were compared with the pathologic findings.
RESULTS. For T staging, MR imaging accuracy was higher than that of helical CT (73.3% and 66.7%, respectively); however, the accuracies of the two methods were not significantly different from each other (McNemar test, p > 0.05). Overstaging was noted in 6.7% of cases with MR imaging and 10% with helical CT. Understaging was noted in 20% of cases with MR imaging and 23.3% with helical CT. For N staging, the accuracies of MR imaging and helical CT were 55% and 58.6%, respectively, with no statistical significance (overstaging, 10% and 6.9%; understaging, 34.5% and 34.5%, respectively).
CONCLUSION. MR imaging was comparable to helical CT in the T and N staging of gastric cancer.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Parallel advances in CT equipment and scanning techniques have reduced scanning time and decreased motion artifacts. Simultaneously, rapid IV contrast administration with an automatic power injector has improved contrast enhancement of the gastric wall and gastric cancer. Helical CT has advantages over conventional CT, including faster scanning time and fewer respiratory misregistration artifacts in a single breath-hold.
Endoscopic sonography has been reported to be the most accurate technique for the T staging of gastric cancer because it can define the five layers of the gastric wall. But this technique cannot evaluate other factors such as liver and distant lymph node metastases and peritoneal seeding [7]. As a result, endoscopic sonography is accepted as a complementary technique to CT for staging gastric cancer [7, 8]. In addition, endoscopic sonography is an invasive technique dependent on the operator and is not readily available in some hospitals, including ours.
To our knowledge, no reports have compared fast MR imaging and helical CT in the staging of gastric cancer. Therefore, we prospectively performed breath-hold two-dimensional (2D) fast low-angle shot (FLASH) T1-weighted and turbo spin-echo T2-weighted MR imaging with a phased array coil and helical CT to compare their accuracy in the staging of gastric cancer.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Pathologic T and N staging was based on the international TNM classifications [9], as follows: pT1 (tumor invades lamina propria or submucosal layer), pT2 (tumor invades muscularis propria or subserosa), pT3 (tumor penetrates the serosa without invading adjacent organs), and pT4 (tumor invades adjacent structures). Lymph nodes were likewise divided: N0 (no regional lymph node metastasis), N1 (metastases in one to six regional lymph nodes), N2 (metastases in seven to 15 regional lymph nodes), and N3 (metastases in >15 regional lymph nodes).
MR Imaging Protocol
MR imaging was performed with a superconductive magnet operating at 1.5 T
(Magnetom Vision; Siemens, Erlangen, Germany) with a phased array coil.
Patients fasted for 6 hr and ingested 500 ml of water or effervescent granules
as an oral contrast agent. In the preliminary study, we found that both
effervescent granules and water could readily distend the stomach and provide
clear contrast with the gastric lesions. Thus, we gave effervescent granules
to nine patients who underwent MR examination on the same day as surgery.
Twenty milligrams of scopolamine (Buscopan; Boehringer International,
Ingelheim, Germany) were administered intramuscularly 5 min before the
examination. Patients were placed on the MR gantry in the supine or prone
position to allow the lesion to contact the ingested water or air. When water
was used as the oral contrast agent, the patients in whom the lesion was at
the gastric cardia or fundus were imaged in the supine position, and the other
patients were imaged in the prone position, except for one patient who could
not tolerate this position. The positions were reversed when effervescent
granules were used as oral contrast agents.
Patients were examined with a breath-hold T1-weighted 2D FLASH technique in the axial, coronal, and sagittal planes. The scanning parameters for T1-weighted axial images were TR/TE, 146.1/4.1 msec; flip angle, 80°; and one excitation. The band-width was 260 Hz; field of view, 33 cm; matrix size, 128 x 256; and the slice thickness was 8 mm with a 1.6-mm gap. The scan time was 18 sec. The parameters for the T1-weighted coronal and sagittal images were the same as those for T1-weighted axial images, except for the TR (100 for coronal and sagittal images) and field of view (45 cm for coronal images). The scan times of coronal and sagittal images were 14 and 12 sec, respectively. T2-weighted axial turbo spin-echo images were obtained with a TR/TE of 3200/138, an echo spacing of 9.2 msec, an echo train length of 29, a flip angle of 18°, one excitation, a bandwidth of 260 Hz, a field of view of 33 cm, a matrix size of 116 x 256, an 8-mm slice thickness with a gap of 1.6 mm, and a scan time of 17 sec. Sixty seconds after an IV bolus injection of 15-20 mg of gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany) at a rate of approximately 1 ml/sec through an antecubital vein, axial T1-weighted FLASH images with fat suppression were obtained (157.6/4.1; flip angle, 80°; one excitation; bandwidth, 260 Hz; field of view, 33 cm; matrix size, 128 x 256; slice thickness, 8 mm; gap, 1.6 mm) with a scan time of 19 sec. All sequences were performed in a single breath-hold. Nineteen slices were obtained for unenhanced and contrast-enhanced T1-weighted axial images, and 11 slices were obtained for coronal and sagittal T1- and T2-weighted images.
CT Protocol
Helical CT was performed using a third-generation CT scanner (Somatom Plus
4; Siemens). Each patient fasted for 6 hr before the CT examination.
Immediately before the CT scan, the patient ingested 500 ml of tap water and
stayed in a prone or supine position to allow the water to come into contact
with the gastric lesion. Unenhanced scanning was not performed. A total of 120
ml of iopromide (Ultravist; Schering) was administered IV at a flow rate of
2.5-3 ml/sec using an automatic power injector.
Contrast-enhanced helical CT scans were obtained 60 sec after initiation of the IV contrast material injection. Helical CT scanning was begun at the level of the dome of the right hemidiaphragm and extended to the iliac crest.
For every patient, an 8-mm section collimation at 8-10 mm/sec table increments was performed at 120 kVp and 150 mA during the 20-25 sec of a single breath-hold. Axial image reconstruction was performed at 8-mm intervals.
We chose a relatively large slice thickness for helical CT and MR imaging to make this parameter concordant for both techniques and to cover a wide scan range during single breath-holding because some older patients could not tolerate breath-holding for more than 25 sec with a distended stomach, especially during MR imaging. Preoperative CT and MR imaging were performed to evaluate tumor depth and to find distant metastases, such as liver or paraaortic lymphadenopathy. Hence, we used an 8-mm rather than 5-mm slice thickness.
Image Analysis
The MR and helical CT images were independently interpreted by two groups
of radiologists before surgery. Each group consisted of two radiologists, and
every decision was made by consensus of the two radiologists from each group.
MR images and helical CT scans were assessed for both depth of tumor invasion
and the presence of regional lymph node metastases. An abnormal lesion was
determined when the gastric wall showed a 6-mm or greater focal thickening or
when focal enhancement was seen in the gastric wall.
The degrees of tumor invasion seen on MR and helical CT imaging were classified as follows: MRT1 (no abnormal finding), CTT1 (no abnormal finding or presence of low-density stripe at the base of the lesion corresponding to the submucosal layer), MRT2 and CTT2 (thickened gastric wall with smooth, well-defined outer border, and absent or disrupted low-density stripe at the base on CT), MRT3 and CTT3 (nodular or irregular outer border of thickened gastric wall or perigastric infiltration), and MRT4 and CTT4 (direct extension and invasion of tumor into contiguous organ or structure).
A change in signal intensity or CT attenuation was recognized as invasion into an adjacent organ (T4) that came in contact with a gastric lesion. The obliteration of the fat plane separating the stomach from adjacent organs such as the pancreas was not regarded as evidence of stage T4 because the fat plane is frequently obliterated by inflammation or cachexia.
Regional lymph nodes were considered to be involved by metastasis if they were larger than 8 mm in the short-axis diameter [10]. Regional lymph node involvement was based on the criteria of the N category of the international TNM classification.
Statistical Analysis
Statistical analysis using the McNemar test was performed to compare fast
MR imaging with helical CT on the basis of the accuracy in the staging of
stomach cancer and tumor detectability. Statistical significance was defined
as p greater than 0.05.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The overall accuracy of MR imaging for the T factor was 73.3% (22/30); the overall accuracy of helical CT for the T factor was 66.7% (20/30), which was less than the MR accuracy but not statistically different (Mc-Nemar test, p > 0.05). On MR imaging, the incidence of understaging was 20% (6/30), whereas that of overstaging was 6.7% (2/30). On helical CT, the incidence of understaging was 23.3% (7/30) and that of overstaging was 10% (3/30). MR imaging was concordant with CT in 80% (24/30) of cases for T staging. Both MR imaging and helical CT were correct in 60% (18/30) of cases, MR imaging alone was correct in 13.3% (4/30), helical CT alone was correct in 6.7% (2/30), and both were incorrect in 20% (6/30) (Figs. 1A,1B,1C,2A,2B,2C,3A,3B,3C).
|
|
|
|
|
|
|
|
|
The MR imaging findings were concordant with the pathologic findings for 87.5% (7/8) of pT1 tumors, 66.7% (2/3) of pT2 tumors, 78.6% (11/14) of pT3 tumors, and 40% (2/5) of pT4 tumors. One pT2 tumor was understaged as MRT1, two pT3 tumors were understaged as MRT2, and three of the pT4 tumors were understaged as MRT3. One pT1 tumor was overstaged as MRT2, and one pT3 tumor was overstaged as MRT4 (Fig. 4A,4B). The findings on helical CT were concordant with the pathologic findings in 75% (6/8) of pT1 tumors, 33.3% (1/3) of pT2 tumors, 85.7% (12/14) of pT3 tumors, and 20% (1/5) of pT4 tumors. One pT2 tumor was understaged as CTT1, two pT3 tumors were understaged as CTT2, and four pT4 tumors were understaged as CTT3. Two pT1 tumors were overstaged as CTT2, and one pT2 tumor was overstaged as a CTT3.
|
|
Of five pT4 tumors, two showed pancreatic invasion (Fig. 5A,5B). Three showed transverse mesocolon invasion and one had invasion to the gallbladder serosa. One case of pancreatic invasion was shown on both MR imaging and helical CT. Neither MR imaging nor CT revealed any of the three cases of mesocolon invasion or the one case of pancreatic invasion. The case of gallbladder invasion was depicted on MR imaging but was not definitely seen on helical CT (Fig. 2A,2B,2C).
|
|
Both MR imaging and helical CT showed the same detection rate (80%, 4/5) in the five cases of omental tumor infiltration. We missed three cases of peritoneal carcinomatosis with both techniques because no detectable peritoneal nodules or ascites were identified; however, both techniques detected the presence of omental infiltration in these cases.
One case of exophytic tumor mimicked perigastric lymphadenopathy on axial helical CT but was correctly identified on coronal MR images.
Metastasis to Regional Lymph Nodes
MR imaging correctly revealed regional lymph node involvement in 55%
(16/29) of our patients, understaged 34.5% (10/29), and overstaged 10% (3/29).
Helical CT was correct in 58.6% (17/29) of patients, understaged 34.5%
(10/29), and overstaged 6.9% (2/29). MR imaging results were concordant with
helical CT in 89.7% (26/29). No significant difference was observed between MR
imaging and helical CT in the staging of regional lymph node metastasis
(p > 0.05).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Many kinds of oral contrast agents distend the stomach for MR imaging, and these can be divided into positive (producing high signal intensity on T1- and T2-weighted images), negative (producing low signal intensity on T1- and T2-weighted images), and biphasic (producing opposite signal intensities on T1- and T2-weighted images) contrast agents [11]. In this study, we used water or effervescent granules as oral contrast agents. Water is biphasic and the simplest contrast agent. Effervescent granules (a negative contrast agent) make the lumen of the stomach dark. If the stomach is overly distended with air, a magnetic susceptibility artifact may occur. However, the use of effervescent granules still provides improved evaluation of the gastric wall and can be used to supplement or replace water for gastric distention [11].
For fast MR imaging, we performed breath-hold T1-weighted 2D FLASH, T2-weighted turbo spin-echo, and contrast-enhanced T1-weighted FLASH sequences with fat suppression. In most cases, we staged the tumor with two or three planes (axial, coronal, and sagittal) of unenhanced T1-weighted images only. We used the fat suppression technique in contrast-enhanced MR imaging to make the contrast between enhanced gastric tumor and perigastric fat more conspicuous. However, contrast-enhanced T1-weighted images showed more artifacts and lower visual contrast than unenhanced T1-weighted images (Figs. 1A,1B,1C and 2A,2B,2C). The T2-weighted images showed little difference in signal intensity between tumor and normal gastric wall and did not show an advantage for tissue characterization. Coronal and sagittal images were useful for evaluation of extraluminal out-growth and omental infiltration by tumor. MR images, even with breath-hold fast MR imaging, were not completely free from motion artifacts, and sometimes a ghost artifact from aortic pulsation hid a lesion at the gastric angle.
A recent MR study proved that under experimental conditions, up to five layers of the gastric wall can be differentiated on MR imaging as they can in endoscopic sonography [12]. But this result was from an ex vivo study, and we could not visualize the normal gastric wall to such a degree. We think that further development and study for faster, proper MR imaging sequences and planes will be necessary for such an experimental result to become clinically applicable and to overcome motion artifacts.
CT is the most frequently used imaging technique for the staging of gastric
cancer. Recent studies [13,
14] using dynamic or helical
CT reported that the normal gastric wall frequently showed a two- or
three-layer pattern that was interrupted by a tumor; thus, more accurate
staging of the T factor could be expected. In addition, the use of thin
collimation (
5 mm) could improve the depiction of small lesions and make
it possible to obtain multiplanar reformatted images. However, it still is
difficult for helical CT to differentiate between a pT2 tumor and a pT1 tumor
with massive submucosal invasion because the hypoattenuating stripe
corresponding to the intact submucosal layer could be obliterated in the
latter [14]. We also
experienced a case of a pT2 tumor with minimal invasion of the muscularis
propria layer, which showed focal wall thickening with preservation of the
hypoattenuating stripe (Fig.
3A,3B,3C).
In this study, advanced gastric cancer was easily detectable on both MR imaging and helical CT; however, the incidence of detectability of early gastric cancer was low on both techniques. MR imaging was slightly more accurate than helical CT in the T staging of gastric cancer, but this difference was not statistically significant (p > 0.05). MR imaging and helical CT showed understaging more frequently than overstaging because of the frequent understaging of pT4 tumors. The low concordance rates of pT4 tumors on MR imaging and helical CT were partly caused by the fact that patients with evident CTT4 or MRT4 tumors did not undergo surgery and therefore were not included in this study. Also, invasion into the transverse mesocolon was difficult to detect. Under normal conditions, the transverse mesocolon can be identified on CT scans and MR images as the fatty plane extending from the pancreas to the transverse colon. However, cachexia and a distended stomach frequently efface this fatty plane in patients with advanced gastric cancer, thereby hindering proper evaluation. For pancreatic invasion, the accuracy was the same for both techniques; however, the number of cases of pancreatic invasion (n = 2) was too small to analyze in this study.
Another group of researchers [15] suggested the superiority of MR imaging over CT because the difference in signal intensity of the stomach and pancreas could make the detection of tumor invasion into the pancreas easier on MR images. We propose that further study is needed to evaluate the usefulness of MR imaging for this purpose. Early or miliary peritoneal carcinomatosis without ascites was difficult to detect preoperatively, but omental infiltration was easily detected on both MR imaging and helical CT in most (80%) of our patients.
We assumed that MR multiplanar imaging would be more accurate in nodal staging than helical CT with only axial images, but with a relative lack of experience and unfamiliarity with MR anatomy in coronal and sagittal images for lymph node detection, we could not prove the superiority of MR imaging over helical CT in nodal staging. We think the detectability of lymph node metastases would be improved by radiologists having more experience with interpretation of coronal and sagittal MR images. The detection rate of lymph node metastasis was low for both techniques. The low detection rate was due to frequent microscopic nodal invasion and the fact that reactive or inflammatory nodal enlargement could not be differentiated from metastatic nodal enlargement on MR imaging or helical CT. An accurate assessment of metastatic lymph nodes was difficult when they were grouped together or attached to a gastric mass. Though an exophytic tumor growth can sometimes mimic a regional lymph node metastasis on axial helical CT, coronal or sagittal MR imaging can easily differentiate them.
A limitation of our study is the thick-slice collimation for helical CT to match that of MR imaging. The depiction of small lesions could be improved by thinner collimation, and the apparent advantage of the multiplanar capability of MR imaging would be relatively eliminated with the use of thinner slices on helical CT and multiplanar reformatted images. We assume that further study will be necessary to compare the advantages of thinner collimation helical CT with the multiplanar imaging capability of MR imaging. We are now using a slice thickness of 5 mm for gastric evaluation at our institution and can achieve full coverage of the liver and stomach in a single breath-hold in most patients.
Another limitation of this study is that helical CT and MR imaging were not performed simultaneously; MR imaging was done on average 11.9 days after helical CT. In our opinion, however, this interval did not affect the degree of tumor invasion.
In summary, no statistically significant difference was seen between fast MR imaging and helical CT with respect to the accuracy of staging gastric cancer and tumor detectability (p > 0.05). Breath-hold fast MR imaging was comparable to helical CT for the staging of gastric cancer. Compared with helical CT, fast MR imaging has some limitations (i.e., inevitable motion artifacts or nondifferentiation of the gastric wall in in vivo studies). Nonetheless, we conclude that MR imaging could be a useful alternative to helical CT because of its high resolution of soft tissue, its multiplanar imaging capability, and its lack of ionizing radiation. In addition, MR imaging could be useful for patients who cannot undergo CT because of renal impairment, pregnancy, or hypersensitivity to CT contrast materials.
Acknowledgments
We thank Bonnie Hami (Department of Radiology, University Hospitals of
Cleveland, Cleveland, OH) for her assistance in the linguistic improvement of
the manuscript.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  R. M. Kwee and T. C. Kwee Imaging in Local Staging of Gastric Cancer: A Systematic Review J. Clin. Oncol., May 20, 2007; 25(15): 2107 - 2116. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Shimizu, K. Ito, N. Matsunaga, A. Shimizu, and Y. Kawakami Diagnosis of Gastric Cancer with MDCT Using the Water-Filling Method and Multiplanar Reconstruction: CT-Histologic Correlation Am. J. Roentgenol., November 1, 2005; 185(5): 1152 - 1158. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. J. Kim, A. Y. Kim, S. T. Oh, J.-S. Kim, K. W. Kim, P. N. Kim, M.-G. Lee, and H. K. Ha Gastric Cancer Staging at Multi-Detector Row CT Gastrography: Comparison of Transverse and Volumetric CT Scanning Radiology, September 1, 2005; 236(3): 879 - 885. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W Ajaj, S C Goehde, N Papanikolaou, G Holtmann, S G Ruehm, J F Debatin, and T C Lauenstein Real time high resolution magnetic resonance imaging for the assessment of gastric motility disorders Gut, September 1, 2004; 53(9): 1256 - 1261. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Torabi, S. L. Aquino, and M. G. Harisinghani Current Concepts in Lymph Node Imaging J. Nucl. Med., September 1, 2004; 45(9): 1509 - 1518. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. C. Lauenstein, F. M. Vogt, C. U. Herborn, A. DeGreiff, J. F. Debatin, and G. Holtmann Time-Resolved Three-Dimensional MR Imaging of Gastric Emptying Modified by IV Administration of Erythromycin Am. J. Roentgenol., May 1, 2003; 180(5): 1305 - 1310. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M.-S. Park, J.-S. Yu, M.-J. Kim, S.-W. Yoon, S. H. Kim, T. W. Noh, K.-H. Lee, J. T. Lee, H. S. Yoo, and K. W. Kim Mucinous versus Nonmucinous Gastric Carcinoma: Differentiation with Helical CT Radiology, March 21, 2002; (2002) 2232010905. [Abstract] [Full Text]
|
|
|
|
|
|
K.-M. Sohn, J. M. Lee, S.-Y. Lee, B.-Y. Ahn, S.-M. Park, and K.-M. Kim Author's Correction Am. J. Roentgenol., August 1, 2000; 175(2): 556 - 556. [Full Text] [PDF]
|
|
|
|
|
|
M.-S. Park, J.-S. Yu, M.-J. Kim, S.-W. Yoon, S. H. Kim, T. W. Noh, K.-H. Lee, J. T. Lee, H. S. Yoo, and K. W. Kim Mucinous versus Nonmucinous Gastric Carcinoma: Differentiation with Helical CT Radiology, May 1, 2002; 223(2): 540 - 546. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
