DOI:10.2214/AJR.07.2940
AJR 2008; 190:1505-1511
© American Roentgen Ray Society
MDCT for Differentiation of Category T1 and T2 Malignant Lesions from Benign Gastric Ulcers
Chiao-Yun Chen1,2,
Deng-Chyang Wu3,
Yu-Ting Kuo1,2,
Chien-Hung Lee4,
Twei-Shiun Jaw1,2,
Wan-Yi Kang5 and
Jui-Sheng Hsu1,6
1 Department of Medical Imaging, Kaohsiung Medical University Hospital,
Kaohsiung, Taiwan.
2 Department of Radiology, Faculty of Medicine, College of Medicine, Kaohsiung
Medical University, Kaohsiung, Taiwan.
3 Division of Gastroenterology, Department of Internal Medicine, Kaohsiung
Medical University Hospital, and Faculty of Medicine, College of Medicine,
Kaohsiung Medical University, Kaohsiung, Taiwan.
4 Graduate Institute of Public Health, Kaohsiung Medical University, Kaohsiung,
Taiwan.
5 Department of Pathology, Kaohsiung Medical University Hospital, Kaohsiung,
Taiwan.
6 Graduate Institute of Medicine, College of Medicine, Kaohsiung Medical
University, 100 Tz-You 1st Rd., Kaohsiung, 807 Taiwan, Province of
China.
Received July 25, 2007;
accepted after revision December 3, 2007.
Address correspondence to J. S. Hsu.
Supported by a grant (Q097035) from the Kaohsiung Medical University
Research Foundation.
Abstract
OBJECTIVE. The purpose of this study was to evaluate MDCT parameters
for differentiating malignant (category T1 and T2) from benign gastric ulcers
and to evaluate the performance characteristics of these predictors with
optimal cutoff points determined in receiver operator characteristic
analysis.
SUBJECTS AND METHODS. The subjects were 26 patients with gastric
cancer (11 with T1 lesions, 15 with T2 lesions) and 26 patients with benign
gastric ulcer. MDCT and virtual gastroscopic findings were analyzed according
to four qualitative criteria: ulcer shape, base, and margin and changes in
adjacent folds. The quantitative criteria ulcer size, thickness of the gastric
wall around an ulcer, thickness of the enhanced ulcer base, and enhancement
around an ulcer were measured on multiplanar reconstruction images. We
calculated the sensitivity and specificity of each quantitative criterion.
Receiver operator characteristic analysis was used to identify cutoff points
yielding optimal sensitivity and specificity for the diagnosis of gastric
cancer.
RESULTS. On virtual gastroscopy, ulcer shape and margin and gastric
fold changes had sensitivities of 80.8%, 84.6%, and 90.9% and specificities of
76.9%, 73.1%, and 77.8%, respectively, in the diagnosis of gastric cancer. On
multiplanar reconstruction images, thickness of the enhanced ulcer base and
enhancement around the ulcer had sensitivities of 80.8% and 73.1% and
specificities of 100% and 100%.
CONCLUSION. MDCT combined with virtual gastroscopy and multiplanar
reconstruction enhances the morphologic details of gastric ulcers and is a
useful way to differentiate malignant (T1 and T2) and benign gastric
ulcers.
Keywords: benign gastric ulcer • gastric cancer • malignant gastric ulcer • MDCT • virtual gastroscopy
Introduction
Patients with benign gastric ulcer are at risk that the ulcer will
undergo malignant transformation
[1], and the symptoms of benign
gastric ulcer are similar to those of gastric cancer
[2]. Because the therapeutic
outcome of gastric cancer is related to the stage of the disease at diagnosis
[2-4],
early diagnosis of malignant ulcer therefore is crucial. Unfortunately, the
differential diagnosis of malignant and benign gastric ulcers on the basis of
macroscopic endoscopic findings can be difficult
[5,
6]. Graham et al.
[7] reported that as many as 5%
of malignant gastric ulcers may grossly appear to be benign. Therefore, it is
imperative that all such lesions be evaluated histologically
[7]. Acquisition of numerous
biopsy specimens from the ulcer margin and base can increase the sensitivity
of biopsy in the diagnosis of gastric cancer
[7]. Bytzer
[8] found that surgically
curable cases of early gastric cancer were likely to be missed owing to
reliance on appearance and histologic features alone and that follow-up
endoscopy with repeated biopsy of unhealed ulcers is essential. However,
repeated endoscopy with multiple biopsies is relatively invasive and
expensive.
Endoscopic sonography, with its cross-sectional imaging of the gastric
wall, is useful for determining the depth of invasion of gastric cancer. The
findings, however, are not useful for differentiating malignant from benign
gastric tumors [9]. For
example, at endoscopic sonography symmetric hypoechoic areas of peptic ulcers
resemble hypoechoic areas of cancerous invasion
[10].
In a study of the utility of double-contrast upper gastrointestinal
radiography for differentiating malignant from benign gastric ulcers, Shindoh
et al. [11] found several
pitfalls of the technique that allowed gastric cancer to be overlooked.
Treichel [12] found that flat
cancerous lesions with a diameter less than 1 cm were particularly difficult
to detect. In addition, the radiographic appearance of early gastric cancer
sometimes resembles that of benign gastric ulcer
[13]. False-negative findings
on upper gastrointestinal radio-graphs have been reported to occur in as many
as 50% of cases [14], and the
sensitivity in the diagnosis of early gastric cancer can be as low as 14%
[15]. Furthermore, lack of
cooperation by patients causes a major technical problem with this
technique.
View larger version (136K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1 —57-year-old woman with T1 malignant gastric ulcer
(circle) in gastric body. En face virtual gastroscopic view
shows uneven ulcer base, geographic ulcer shape, irregular ulcer margin,
associated gastric folds with interruption of rugae (black arrows),
and bulbous enlargement and fusion (white arrow).
|
|
Minami et al. [16]
considered CT differentiation between benign ulcer and early gastric cancer
with ulceration difficult because both entities manifest as lesions with
defects of the normal inner layer of the gastric wall. Insko et al.
[17] suggested that CT may be
useful for differentiating benign gastric lesions from potentially malignant
ones. The sensitivity for detection of malignant or potentially malignant
gastric lesions was 100%, but the specificity was only 50% in that study.
Stabile Ianora et al. [18]
found that they could not differentiate early gastric cancer from benign
gastric ulcer with single-detector helical CT.
The high speed and thin collimation of MDCT have improved temporal and
spatial resolution in the z-axis. Advances in image processing have
facilitated accurate reconstruction of gastric images. With air distention of
the stomach, 3D virtual gastroscopic images depict subtle mucosal changes in
the same way that conventional gastroscopy does
[19]. With adequate distention
of the stomach with water as a neutral contrast agent, contrast-enhanced
multiplanar reconstruction (MPR) images, which are similar to endoscopic
sonographic images, provide useful information about the gastric wall around
an ulcer. In a previous study
[20] we found marked
enhancement around the ulcers in most cases of gastric cancer. This finding
may be helpful in differentiating malignant gastric tumors from benign gastric
ulcers.
To our knowledge, no results have been published regarding the usefulness
of MDCT combined with virtual gastroscopy and MPR for differentiating
malignant from benign gastric ulcers. Previous studies have shown that
discriminating benign ulcer and early gastric cancer is difficult
[13,
17,
18,
21]. The purpose of our study
was to prospectively evaluate the use of noninvasive MDCT with virtual
gastroscopy and MPR for differentiating relatively early malignant gastric
tumors (T1 or T2) from benign gastric ulcers and to establish the test
performance characteristics, including sensitivity and specificity, of the
technique.
Subjects and Methods
The institutional review board of our institution approved the study.
Written informed consent was obtained from each patient after the purpose and
protocol of the study had been fully explained. From January 2004 to January
2006, a total of 95 patients consecutively referred for MDCT met the criteria
for enrollment in our study. The inclusion criteria were endoscopic finding
suggesting malignant gastric ulcer (n = 68), intractable gastric
ulcer not healing after 8 weeks of treatment with proton pump inhibitors
(n = 6), and endoscopic finding suggestive of benign gastric ulcer
but insufficient to exclude the possibility of malignant gastric ulcer.
Twenty-one (91%) of 23 patients who met the third criterion agreed to join the
study for further image evaluation.
All patients enrolled underwent an MDCT protocol designed specifically for
patients with gastric ulcer. All gastric ulcer patients underwent endoscopic
biopsy 2-4 days before CT. The histopathologic results were used as the
reference standard. The diagnosis of benign ulcer was confirmed with
pathologic results and follow-up findings for more than 6 months. Malignant
ulcers were confirmed with pathologic or surgical results or both. In the
patients who met the first criterion, 43 T3 or T4 gastric lesions, six T1
lesions, 12 T2 lesions, and seven benign gastric ulcers were confirmed. In the
patients who met the second criterion, two T1 and two T2 lesions and two
benign gastric ulcers were diagnosed. In the patients who met the third
criterion, there were three T1 and one T2 gastric lesions and 17 benign
gastric ulcers. To test the performance of MDCT in the diagnosis of relatively
early gastric cancer, the 43 patients with T3 or T4 gastric cancer were
excluded.
MDCT Techniques
CT examinations were performed with a 16-MDCT scanner (LightSpeed H16, GE
Healthcare) on patients who had fasted for at least 8 hours. For gastric
distention, patients ingested 6 g of gas-producing crystals with 10 mL of
water a short time before unenhanced CT and virtual gastroscopy. Patients with
insufficient air distention of the stomach were given an additional 3 g of
gas-producing crystals. Unenhanced upper abdominal CT scans from the
diaphragmatic domes to 2 cm below the lower margin of the air-distended
gastric body were obtained at 16 x 1.25 mm collimation, 27.5 mm/s table
speed, 250-300 mAs, and 120 kVp. In three cases in which a great deal of
residual fluid covered the stomach, the patient shifted to the other side, and
additional scanning was performed. All procedures were performed under the
guidance of an experienced radiologist.
Immediately after unenhanced CT and while still on the CT table, each
patient drank 800-1,000 mL of tap water, which served as a neutral gastric
contrast agent for contrast-enhanced CT. A non ionic iodine contrast agent
(100 mL of iopromide, Ultravist, Bayer HealthCare) was administered through
the antecubital vein at 3 mL/s with a 20-gauge needle and an automatic
dual-head injector (LF Opti Vantage). All CT acquisitions were performed
during the portal venous phase (70 seconds), and scanning ranged from the
diaphragmatic domes to the iliac crest. On a work station (AW 4.1, GE
Healthcare) we re constructed raw data sets at 1.25-mm slice thick ness and
0.625-mm reconstruction intervals for virtual gastroscopic (air-filled
unenhanced images) and MPR images.
Image Analysis
Virtual gastroscopic images—Virtual gastroscopic imaging was
performed by an experienced abdominal radiologist blinded to endoscopic
results, lesion size, and macroscopic features. This observer independently
evaluated CT images on the workstation with a navigator tool for virtual
gastroscopic images. One en face view, two profile views, and four
oblique views around 30-45° of each ulcer were obtained in all cases.
Virtual gastroscopic images were independently interpreted by two independent
abdominal radiologists. Because endoscopic criteria for benign and malignant
gastric ulcers had been well established
[12,
13], we followed the criteria
used by most endoscopists for virtual gastroscopic images. The following
findings were taken to suggest malignant gastric ulcer (Figs.
1 and
2): virtual gastroscopic
features of gastric ulcer with an irregular, angulated, or geographic shape;
uneven base; irregular or asymmetric edges surrounding the ulcer; disruption
of the gastric folds reaching the crater edge, clubbing of folds, or fold
fusion; or a combination of these findings. In contrast, benign gastric ulcers
(Figs. 3 and
4) had a smooth, regular shape;
an even base; sharply demarcated or regular rounded edges; con verging gastric
folds with smooth tapering and radiation; or a combination of these findings.
Differences in assessment were resolved by consensus.
View larger version (104K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2 —72-year-old woman with T2 malignant gastric ulcer
(arrows) in gastric angle. En face virtual gastroscopic view
shows even ulcer base, oval ulcer shape, irregular ulcer margin, and no
associated gastric folds around ulcer.
|
|
View larger version (141K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3 —68-year-old man with benign gastric ulcer (circle)
in gastric body. En face virtual gastroscopic view shows even ulcer
base, regular triangular ulcer shape, regular ulcer margin, and associated
regular gastric folds terminating at ulcer margin (arrows). N =
nasogastric tube.
|
|
View larger version (109K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4 —58-year-old man with benign gastric ulcer (arrows)
in gastric antrum. En face virtual gastroscopic view shows even ulcer
base, oval ulcer shape, regular ulcer margin, and no associated gastric folds
around ulcer.
|
|
View larger version (156K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5B —56-year-old man with T1 malignant gastric ulcer. Reformatted
paraaxial image in plane of line 1 (A) shows ulcer (circle) in
gastric antrum with periulcer gastric wall thickening, marked periulcer
enhancement, high attenuation at ulcer base, and preserved low attenuation of
outer submucosal layer.
|
|
MPR images—MPR images (3-mm slice thickness) of the ulcers
were obtained by the same radiologist who obtained the virtual gastroscopic
images. To choose the optimal plane of the ulcers and to avoid partial volume
effects on MPR images, two vertical planes around the ulcer were used for
unenhanced and contrast-enhanced imaging (Figs.
5A,
5B and
5C).
We developed modified criteria
[14,
15] for differentiating
malignant ulcers (Figs. 5A,
5B and
5C) from benign ulcers (Figs.
6A and
6B) on contrast-enhanced MPR
images. Our criteria were focused on enhancement patterns and thickness of the
gastric wall around the ulcer. To determine the best predictors of an accurate
differential diagnosis between benign and malignant ulcers, we quantified
ulcer size (maximum diameter of the ulcer in the MPR image), thickness of the
gastric wall around the ulcer (maximum length of the ulcer margin vertical to
the gastric serosal margin), enhancement of the ulcer base (attenuation of
gastric wall, which is vertical to the ulcer base, at the ulcer base), and
enhancement around the ulcer on MPR images (maximal difference in periulcer
attenuation between enhanced and unenhanced images) in each case (Figs.
7,
8,
9). To obtain these
measurements, a fifth observer, an abdominal radiologist, selected the optimal
MPR image of the ulcer from the images (3-mm slice thickness) of the ulcers
obtained by the second radiologist. The measurements were made three times,
and the mean of the three measurements was used.
View larger version (180K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6A —66-year-old woman with benign gastric ulcer. Paracoronal
reformatted image shows gastric ulcer (arrows) in gastric body with
mild periulcer gastric wall thickening, no increased periulcer enhancement,
and only linear sharp high attenuation at ulcer base.
|
|
View larger version (58K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 7 —43-year-old woman with malignant gastric ulcer. Multiplanar
reformation shows maximum diameter of ulcer (line D), periulcer wall
thickening (line T), and thickening of enhanced ulcer base (line
B).
|
|
View larger version (71K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 8 —66-year-old woman with benign gastric ulcer. Multiplanar
reformation shows maximum diameter of ulcer (line D), periulcer wall
thickening (line T), and discernible thickness of enhanced ulcer base
(line B).
|
|
View larger version (57K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 9 —48-year-old man with malignant gastric ulcer. Multiplanar
reformation shows three optimally sized regions of interest (circles)
placed on right, left, and inferior portions of ulcer to measure attenuation
around ulcer.
|
|
Statistical Analysis
In all cases, 95% CIs for sensitivity and specificity were calculated to
show variability. Sensitivity and specificity were calculated with qualitative
criteria for the diagnosis of malignant gastric ulcer on virtual gastroscopic
images. Statistical differences between malignant and benign ulcers on virtual
gastroscopic images were analyzed with chi-square and Fisher's exact tests.
The final classification of ulcers as benign or malignant was proved with
pathologic evidence. Receiver operating characteristic curve methods were used
to identify cutoff points for each quantitative criterion. This procedure
maximized the likelihood of correct identification of malignant gastric
ulcers. The cutoff value corresponding to the optimal diagnostic accuracy
(defined as the highest sum of the values for sensitivity and specificity) and
the area under the curve of the receiver operating characteristic plot were
determined for each quantitative criterion obtained with imaging. Significant
differences were inferred for a two-tailed p < 0.05. Statistical
analysis was conducted with the program Stata/SE 9.1 (Stata) for Windows
(Mircrosoft).
Results
The study population consisted of 26 men and 26 women (mean age, 59 years;
range, 31-84 years). Eleven patients had T1 and 15 had T2 gastric lesions, and
26 patients had benign gastric ulcers. In all cases of benign ulcers the
histologic findings were eroded or defective layers of the gastric mucosa with
variable degrees of inflammation or granulomatous tissue infiltration in the
lamina propria. According to standard gastric cancer classification
[22], there were 11
differentiated gastric tumors (one papillary, 10 tubular adenocarcinomas) and
15 undifferentiated gastric tumors (nine poorly differentiated
adenocarcinomas, six signet ring cell carcinomas) (Figs.
10A,
10B and
10C). All ulcer lesions
(52/52) were correctly detected on virtual gastroscopic images.
View larger version (119K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 10A —51-year-old man with signet ring cell carcinoma. En
face (A) and profile (B) virtual gastroscopic images of
gastric ulcer (circle) in gastric angle show even ulcer base,
irregular ulcer shape, and irregular ulcer margin but no associated periulcer
gastric folds. Findings indicate typical malignant gastric ulcer.
|
|
View larger version (125K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 10B —51-year-old man with signet ring cell carcinoma. En
face (A) and profile (B) virtual gastroscopic images of
gastric ulcer (circle) in gastric angle show even ulcer base,
irregular ulcer shape, and irregular ulcer margin but no associated periulcer
gastric folds. Findings indicate typical malignant gastric ulcer.
|
|
View larger version (127K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 10C —51-year-old man with signet ring cell carcinoma. Axial
multiplanar reformatted CT image shows ulcer (arrow) with periulcer
gastric wall thickening, no increased periulcer enhancement, and only linear
sharp high attenuation of ulcer base.
|
|
Qualitative Criteria for Differentiating Malignant and Benign Gastric Ulcers on Virtual Gastroscopic Images
Table 1 shows the
sensitivity and specificity of each virtual gastroscopic imaging criterion in
the detection of malignant gastric ulcers in all patients. In the detection of
malignant ulcers, the sensitivity and specificity values were ulcer shape,
80.8% and 76.9%; ulcer margin, 84.6% and 73.1%; associated changes in gastric
folds, 90.9% and 77.8%; and ulcer base, 53.9% and 76.9%. In the 32 patients
without periulcer gastric folds, the ulcers were located in the antrum
(n = 26) or the gastric angle (n = 6). There were
significant differences between malignant and benign ulcers for each of these
variables (p < 0.05).
View this table:
[in this window]
[in a new window]
|
TABLE 1: Sensitivity and Specificity of Qualitative Imaging Criteria for
Differentiating Malignant and Benign Gastric Ulcers on Virtual Gastroscopic
and Multiplanar Reformatted Images of All Patients (n = 52)
|
|
Quantitative Criteria for Differentiating Benign from Malignant Gastric Ulcers on MPR Images
We calculated sensitivity and specificity values for the utility of each
quantitative criterion in accurate differentiation of malignant and benign
ulcers. Cutoff values were determined with the maximum sum of sensitivity and
specificity (Table 2). The
sensitivity and specificity for thickness of the enhanced ulcer base (cutoff
value, 0.27 cm) were 80.8% and 100%. The sensitivity and specificity for
enhancement around the ulcer (cutoff value, 66 H) were 73.1% and 100%. Ulcer
size (cutoff value, 2.28 cm) and thickness of the gastric wall around the
ulcer (cutoff value, 1.32 cm) did not allow statistically significant
differentiation between malignant and benign gastric ulcers (chi-square value,
< 3.84; p > 0.05).
View this table:
[in this window]
[in a new window]
|
TABLE 2: Sensitivity and Specificity of Quantitative Imaging Criteria for
Detection of Malignancy of Gastric Ulcers on Multiplanar Reformatted Images of
All Patients (n = 52)
|
|
Discussion
Our findings suggest that MDCT with 2D MPR and 3D virtual gastroscopic
imaging can yield comprehensive information about gastric ulcers. Patients
with gastric ulcer conventionally undergo three
examinations—gastroscopy, endoscopic sonography, and CT—for the
same purpose [23]. In our
study, MDCT had results comparable with those of conventional gastroscopy in
differentiation of malignant and benign gastric ulcers. The sensitivity was
80.8-90.9% and the specificity 73.1-77.8% on virtual gastroscopic images and
the sensitivity 73.1-80.8% and the specificity 100% on MPR images. Our results
are comparable with those of optical gastroscopy, which has a sensitivity of
76-84% and a specificity of 90-95% in the diagnosis of malignant gastric ulcer
[24-26].
More important is that the high specificity on MPR images may help avoid delay
in the treatment of patients with gastric cancer and thus improve their
survival rate. In addition, all cases of gastric cancer in our study were in a
relatively early stage (T1 or T2 lesions), which is usually more challenging
for accurate diagnosis with any imaging technique currently available.
Like conventional gastroscopy, virtual gastroscopy is useful in the
detection and evaluation of gastric ulcer. Compared with conventional
gastroscopy, virtual gastroscopy can depict most abnormal endoluminal lesions
without a limited field of view or blind areas
[19]. The results in our study
show that ulcer shape and margin on virtual gastroscopic images are accurate
differentiating features of most malignant and benign gastric ulcers. In some
cases, however, acute benign ulcers with severe periulcer edema mimic
malignant ulcers in terms of ulcer shape and ulcer margin. Conversely, some
malignant ulcers mimic benign ulcers in shape and margin owing to a small size
and the presence of minimal periulcer edema. Associated change in periulcer
gastric folds was a useful criterion. Nevertheless, in a distended stomach,
ulcers in the antrum and angle are always free of folds. The disadvantage of
virtual gastroscopy is its lack of color change at the ulcer base, which can
be clearly visualized at conventional gastroscopy. This factor may explain why
an uneven ulcer base on virtual gastroscopic images was insufficiently
sensitive for differentiating malignant from benign gastric ulcers in our
study.
As does endoscopic sonography, MPR imaging provides useful information
about the gastric wall, including stratification
[16], horizontal extension,
and depth of tumor invasion
[16,
20]. In our study, the
periulcer enhancement pattern was a good indicator for differentiating
malignant from benign gastric ulcers. To maximize the sensitivity of detecting
gastric cancer by depiction of its neovascularity, we selected maximum CT
attenuation from three periulcer regions during the portal venous phase. This
technique may minimize the volume-averaging effects caused by heterogeneous
tumor components. In accordance with previous reports
[20,
27,
28], our study showed that
most malignant gastric ulcers have significantly enhanced tumor parts in the
portal venous phase. Benign gastric ulcers, in contrast, exhibited no
significant enhancement on contrast-enhanced MPR images
[18,
27].
Signet ring cell carcinoma (Figs.
10A,
10B and
10C) is a subtype of
undifferentiated gastric cancer. The hypovascularity of this tumor has been
reported [29,
30]. In our study, four of six
malignant gastric ulcers of signet ring cell carcinoma had no strong
enhancement in the tumor areas on contrast-enhanced images. The incidence of
neovascularity of gastric signet ring cell carcinomas may be low because this
tumor is characteristically of the scirrhous subtype
[31] and because a predominant
component (> 50%) of isolated carcinoma cells contains mucin
[32]. These factors may
explain why the criteria enhanced ulcer base and enhancement around the ulcer
on MPR images are inadequate for differentiating the malignant gastric ulcer
of signet ring cell carcinoma from benign gastric ulcer. On the other hand,
ulcer shape and margin on virtual gastroscopic images are proper indicators
for differentiating most malignant gastric ulcers of signet ring cell
carcinomas from benign gastric ulcers. Therefore, in lesions with
malignant-appearing virtual gastroscopic morphologic features but no obvious
periulcer enhancement, signet ring cell carcinoma should be considered.
In our study, ulcer size and periulcer wall thickening showed no
significant differences between malignant and benign gastric ulcers, probably
because periulcer wall thickening can be identified in all benign and
malignant ulcers. An unexpected finding was that malignant ulcers did not
produce more gastric wall thickening than did benign ulcers. An explanation
may be that that only subjects with T1 and T2 gastric lesions rather than more
advanced adenocarcinomas were enrolled in the study. This limitation can be
overcome by use of two additional quantitative criteria on MPR images. For
example, in benign ulcers, thickened walls are due to edema, which always has
a normal mural enhancement pattern and preservation of wall stratification. On
the other hand, periulcer wall thickening together with strong enhancement of
tumor parts and loss of normal wall stratification is observed in malignant
ulcers [18].
On the basis of our findings, we concluded that MDCT combined with virtual
gastroscopy and MPR is a promising strategy that combines the features of
endoscopic viewing and multiplanar cross-sectional imaging
[22]. The technique can be a
powerful tool for noninvasive evaluation of both endoluminal morphologic
changes and intraluminal and extraluminal information on gastric ulcers.
Optical gastroscopy can be difficult to perform on patients who have
difficulty swallowing. Virtual gastroscopy is a good alternative diagnostic
tool under these circumstances. Furthermore, MDCT with virtual gastroscopy and
MPR can assist endoscopists in planning repeated gastroscopic biopsy and
should increase the probability of finding malignancy within a gastric ulcer.
Findings with this protocol may raise suspicion of malignancy in gastric
ulcers with negative biopsy results and suggest repetition of endoscopy sooner
than would occur with the conventional approach. In our study, one patient had
initial negative histologic results for malignancy, but gastric cancer was
suspected at gastroscopy and MDCT. Biopsy repeated 3 days after MDCT showed
early gastric cancer. On the other hand, two patients had gastroscopic
findings suggesting malignant ulcers, but findings at histologic examination
of a biopsy specimen and MDCT were normal. Biopsies repeated 14 days after
MDCT and at 6- and 11-month follow-up evaluations confirmed the finding of
benign gastric ulcer.
Our study had limitations. First, all CT scans were obtained 2-4 days after
endoscopic biopsy, which may raise concern about local inflammation and
confounding CT findings. Results of animal studies
[33-35],
however, have shown that epithelial restitution of injured gastric mucosa can
happen within minutes to hours. Therefore, effects on CT images should be
limited. Nevertheless, this limitation would not change our results
significantly because patients with both malignant and benign gastric ulcers
underwent the same protocol. Prospective randomized studies comparing MDCT and
gastroscopy are needed to validate the use of our protocol. Second, there was
a potential bias due to patient population. Although blinded to the results of
endoscopic examinations and histopathologic analyses, the readers were aware
of the presence of an ulcer. Thus the ability of MDCT to depict gastric ulcers
might have been overestimated. However, the aim of this study was to
investigate the utility of MDCT in differentiating malignant from benign
gastric ulcers rather than to investigate the detectability of gastric ulcers
with MDCT. Third, the virtual gastroscopic reconstruction used in our study
was relatively time-consuming. It demanded that well-trained radiologists
process a relatively large number of CT images, which may limit wider clinical
use.
Although unlike optical endoscopy, MDCT does not show color change or
provide the opportunity to perform biopsy, virtual gastroscopy does provide an
excellent overview of mucosal change with less restriction than with
conventional gastroscopy and allows the operator to measure gastric ulcers
with accuracy. MPR images provide useful information about the gastric wall
around an ulcer. Furthermore, MPR images provide information not only about
intraluminal and extraluminal gastric lesions but also about conditions
outside the stomach. These ancillary findings are useful in cancer staging and
in differential diagnosis, although they were beyond the scope of this study.
In addition, MDCT is less invasive than optical endoscopy. Although CT has not
been a routine examination for the detection of gastric ulcer, with continued
advances in CT scanners and computer technology, MDCT may play an increasingly
important role in the detection of malignant gastric ulcers in patients at
high risk and in preoperative cancer staging. Whether use of the technique
will translate into improved overall survival among patients with gastric
cancer or even cost saving is not clear. Randomized double-blind studies with
larger patient groups are needed to validate the role of MDCT in the diagnosis
of gastric cancer.
Our study showed that MDCT with virtual gastroscopy and MPR shows useful
morphologic details and mural enhancement patterns of gastric ulcers, which
are valuable for differentiating malignant from benign gastric ulcers.
Furthermore, thickening of an enhanced ulcer base is the most important
feature in differentiating a malignant ulcer from a benign ulcer. In lesions
with a malignant appearance on virtual gastroscopy but without obvious
periulcer enhancement or thickening of an enhanced ulcer base, signet ring
cell carcinoma should be considered.
References
- Hansson LE, Nyren O, Hsing AW, et al. The risk of stomach cancer in
patients with gastric or duodenal ulcer disease. N Engl J
Med 1996; 335:242
-249[Abstract/Free Full Text]
- Everett SM, Axon AT. Early gastric cancer in Europe.
Gut 1997; 41:142
-150[Abstract/Free Full Text]
- Sigon R, Canzonieri V, Rossi C. Early gastric cancer: a
single-institution experience on 60 cases. Suppl
Tumori 2003; 2:S23
-S26[Medline]
- Eckardt VF, Giessler W, Kanzler G, Remmele W, Bernhard G. Clinical
and morphological characteristics of early gastric cancer: a case-control
study. Gastroenterology 1990;98
: 708-714[Medline]
- Rollag A, Jacobsen CD. Gastric ulcer and risk of cancer: a
five-year follow-up study. Acta Med Scand1984; 216:105
-109[Medline]
- Haukland HH, Johnson JA, Eide JT. Carcinoma diagnosed in excised
gastric ulcers. Acta Chir Scand 1981;147
: 439-443[Medline]
- Graham DY, Schwartz JT, Cain GD, Gyorkey F. Prospective evaluation
of biopsy number in the diagnosis of esophageal and gastric carcinoma.
Gastroenterology 1982;82
: 228-231[Medline]
- Bytzer P. Endoscopic follow-up study of gastric ulcer to detect
malignancy: is it worthwhile? Scand J Gastroenterol1991; 26:1193
-1199[CrossRef][Medline]
- Maruta S, Tsukamoto Y, Niwa Y, Goto H, Hase S, Yoshikane H.
Endoscopic ultrasonography for assessing the horizontal extent of invasive
gastric carcinoma. Am J Gastroenterol1993; 88:555
-559[Medline]
- Nomura N, Goto H, Niwa Y, Arisawa T, Hirooka Y, Hayakawa T.
Usefulness of contrast-enhanced EUS in the diagnosis of upper GI tract
diseases. Gastrointest Endosc 1999;50
: 555-560[CrossRef][Medline]
- Shindoh N, Nakagawa T, Ozaki Y, Kyogoku S, Sumi Y, Katayama H.
Overlooked gastric carcinoma: pitfalls in upper gastrointestinal radiology.
Radiology 2000;217
: 409-414[Abstract/Free Full Text]
- Treichel J. Double contrast radiography of the stomach: technique
and results in early gastric cancer [in French]. J
Radiol 1979; 60:299
-306[Medline]
- Maruyama M, Baba Y. Gastric carcinoma. Radiol Clin North
Am 1994; 32:1233
-1252[Medline]
- Dooley CP, Larson AW, Stace NH, et al. Double-contrast barium meal
and upper gastrointestinal endoscopy: a comparative study. Ann
Intern Med 1984; 101:538
-545[Abstract/Free Full Text]
- Longo WE, Zucker KA, Zdon MJ, Modlin IM. Detection of early gastric
cancer in an aggressive endoscopy unit. Am Surg1989; 55:100
-104[Medline]
- Minami M, Kawauchi N, Itai Y, Niki T, Sasaki Y. Gastric tumors:
radiologic-pathologic correlation and accuracy of T staging with dynamic CT.
Radiology 1992;185
: 173-178[Abstract/Free Full Text]
- Insko EK, Levine MS, Birnbaum BA, Jacobs JE. Benign and malignant
lesions of the stomach: evaluation of CT criteria for differentiation.
Radiology 2003;228
: 166-171[Abstract/Free Full Text]
- Stabile Ianora AA, Wolowiec A, Francioso G, Scardapane A, Rotondo
A, Angelelli G. Benign and malignant gastric ulcer: CT findings.
Radiol Med (Torino) 2001;102
: 32-36[Medline]
- Inamoto K, Kouzai K, Ueeda T, Marukawa T. CT virtual endoscopy of
the stomach: comparison study with gastric fiberscopy. Abdom
Imaging 2005; 30:473
-479[CrossRef][Medline]
- Chen CY, Hsu JS, Wu DC, et al. Gastric cancer: preoperative local
staging with 3D multi-detector row CT: correlation with surgical and
histopathologic results. Radiology 2007;242
: 472-482[Abstract/Free Full Text]
- Yalamarthi S, Witherspoon P, McCole D, Auld CD. Missed diagnoses in
patients with upper gastrointestinal cancers.
Endoscopy 2004;36
: 874-879[CrossRef][Medline]
- Moayyedi P, Axon AT. Endoscopy and gastric ulcers.
Endoscopy 1995;27
: 689-693[Medline]
- Kim JH, Eun HW, Goo DE, Shim CS, Auh YH. Imaging of various gastric
lesions with 2D MPR and CT gastrography performed with multidetector CT.
RadioGraphics 2006;26
: 1101-1116[Abstract/Free Full Text]
- Gurin NN, Logunov KV. Effectiveness of differential diagnosis of
benign and malignant ulcers at endoscopy [in Russian]. Ter
Arkh 1998; 70:37
-40[Medline]
- Maniatis AG, Eisen GM, Brazer SR. Endoscopic discrimination of
gastric ulcers. J Clin Gastroenterol1997; 24:203
-206[CrossRef][Medline]
- Bustamante M, Devesa F, Borghol A, Ortuno J, Ferrando MJ. Accuracy
of the initial endoscopic diagnosis in the discrimination of gastric ulcers:
is endoscopic follow-up study always needed? J Clin
Gastroenterol 2002; 35:25
-28[CrossRef][Medline]
- Efsen F, Fischerman K. Angiography in gastric tumours.
Acta Radiol Diagn (Stockh) 1974;15
: 193-197[Medline]
- Reuter SR, Redman HC, Miller WJ, Hoskins PA. Gastric angiography.
Radiology 1970;94
: 271-276[Medline]
- Adachi Y, Mori M, Enjoji M, Sugimachi K. Microvascular architecture
of early gastric carcinoma: microvascular-histopathologic correlates.
Cancer 1993; 72:32
-36[CrossRef][Medline]
- Kumano S, Murakami T, Kim T, et al. T staging of gastric cancer:
role of multi-detector row CT. Radiology2005; 237:961
-966[Abstract/Free Full Text]
- Japanese Gastric Cancer Association. Japanese classification of
gastric carcinoma, 2nd English edition. Gastric Cancer1998; 1:10
-24[Medline]
- Kim DY, Park YK, Joo JK, et al. Clinicopathological characteristics
of signet ring cell carcinoma of the stomach. ANZ J
Surg 2004; 74:1060
-1064[CrossRef][Medline]
- Terano A, Sakata-Horie K, Shimada T, et al. The role of cellular
migration in the repair process of gastric epithelial cells. Life
Sci 2001; 69:3083
-3089[CrossRef][Medline]
- Lacy ER. Rapid epithelial restitution in the stomach: an updated
perspective. Scand J Gastroenterol 1995;210
[suppl]:6
-8
- Ito S, Lacy ER, Rutten MJ, Critchlow J, Silen W. Rapid repair of
injured gastric mucosa. Scand J Gastroenterol1984; 101[suppl]:87
-95
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
