AJR 2005; 184:98-103
© American Roentgen Ray Society
CT Colonography Using 16-MDCT in the Evaluation of Colorectal Cancer
Don Jin Chung1,
Kyu Chan Huh2,
Won Jun Choi3 and
Jae Kyun Kim1
1 Department of Radiology, University of Konyang School of Medicine, 685
Gasuwon-dong, Seo-gu, Daejeon 302-718, Korea.
2 Department of Gastroenterology, University of Konyang School of Medicine,
Daejeon 302-718, Korea
3 Department of Surgery, University of Konyang School of Medicine, Daejeon
302-718, Korea.
Received January 29, 2004;
accepted after revision June 21, 2004.
Address correspondence to D. J. Chung
(bookdoo7{at}chollian.net).
Abstract
OBJECTIVE. This study evaluated CT colonography as a method to stage
colorectal cancer and detect polyps and cancers in patients with the
disease.
SUBJECTS AND METHODS. Fifty-one consecutive patients thought to have
colorectal cancer underwent CT colonography, following a colonoscopy, in both
the prone and supine positions. The transverse CT images, multiplanar
reconstruction, volume rendered, and virtual colonoscopy images, were
independently interpreted by two radiologists. Disagreements were resolved by
consensus. The diagnostic accuracy of TNM staging was calculated, and the
sensitivity of CT colonography for the detection of cancers and polyps,
compared with that of colonoscopy, was calculated using repeated colonoscopic
and surgical findings as reference standards. The technical result for
distention was also graded.
RESULTS. In the 51 patients, surgery and follow-up colonoscopy
revealed 21 colorectal cancers (one synchronous cancer) and 41 polyps. The
diagnostic accuracies of CT colonography for TNM staging were 95%, 85%, and
100% for tumor, node, and metastasis, respectively. The sensitivity of both CT
colonography and initial colonoscopy for cancer detection was 100%. The
overall sensitivities of CT colonography and initial colonoscopy for polyp
detection were 90% and 78%, respectively (p = 0.001). The
sensitivities of CT colonography for detecting polyps of 5 mm or smaller, of
6–9 mm, and of 10 mm or larger were 84%, 94%, and 100%, respectively.
The mean overall technical results for the supine and prone positions were
ranked as 2.80 (SD, ± 0.4) and 2.78 (± 0.4), respectively, but
were without statistical significance (p = 0.781).
CONCLUSION. Our preliminary data suggest that for patients with
clinical suspicion of colorectal cancer, CT colonoscopy is valuable in staging
the tumor and in detecting additional polyps or cancers in areas not evaluated
by conventional colonoscopy.
Introduction
Since its development by Vining et al.
[1] in 1994, CT colonography
has emerged as a potential screening technique
[2–5].
Several studies have shown that CT colonography depicts colorectal neoplasms
larger than 1 cm in diameter with sensitivity greater than or equal to 85%
[6–8].
Colonoscopy is currently considered the reference standard for the detection
of colorectal neoplasia in symptomatic patients and in the screening of
high-risk asymptomatic individuals
[9]. Conventional colonoscopy
has various potential limitations. First, it fails to show the entire colon in
about 5% of patients [10].
Second, it does not allow evaluation of the liver and other organs outside the
colon. Third, it has a blind area, as a colonoscope passes in only one
direction. For example, the opposite side of a colonic fold cannot be
evaluated exactly. Finally, it is invasive and uncomfortable. CT colonography
has been proposed as an alternative procedure for the examination of
colorectal cancer, because it compensates for the limitations of colonoscopy
[10]. Our principal focus in
patients with colorectal cancer is the occurrence of cases of synchronous
malignant lesions (1.5–9.0%) and polyps. Precise preoperative evaluation
of the entire colon and precise staging are mandatory in the surgical approach
and adjuvant therapies. MDCT has various advantages over single-detector CT,
including improved image staging and polyp and cancer detection
[11,
12]. However, in CT
colonography studies, the use of 16-MDCT remains to be reported. The purpose
of our study was to evaluate the usefulness of CT colonography for the staging
of colorectal cancer and the detection of polyps and cancers in patients with
the disease.
Subjects and Methods
Between October 2002 and August 2003, 51 consecutive patients (32 men and
19 women; mean age, 63 years; age range, 38–77 years) were prospectively
included in this study. Patients were at high risk for colorectal cancer and
had a history of altered bowel habits, anemia of unknown cause, abdominal
pain, positive fecal occult blood results, and hematochezia. Eligible patients
were informed of the study design and signed an institutional review
board–approved consent form on which the procedure and study were
explained. All CT scans were obtained on a 16-MDCT scanner (Mx8000 IDT,
Philips Medical Systems) after a complete or incomplete colonoscopy.
Colonoscopy was performed by a board-certified gastroenterologist. All masses
and polyps identified on colonoscopy were photographed, measured, and resected
at biopsy. Twenty-four hours before examination, each patient received a
standard bowel preparation. Patients were allowed only a clear liquid diet
until after the preparation had been completed. To prepare for the colonoscopy
and CT colonography, all patients ingested 4 L of a polyethylene glycol
electrolyte solution (Colyte, Schwarz Pharma), which was initiated at 6 pm on
the day before the examination. Once the oral lavage solution had been
administrated, each patient ingested 10 mg of bisacodyl, to reduce residual
fecal material and retained fluid, at 10 pm on the day before the examination.
Twenty milligrams of hyscine N-bromuro (Buscopan, Boehringer
Ingelheim) was subcutaneously administered before air insufflation to further
reduce bowel peristalsis and colonic spasm.
The scanning procedures were performed during a breath-hold. First,
unenhanced images were acquired with patients prone; then, contrast-enhanced
images were acquired with patients supine. The colon was insufflated by gentle
squeezing of the enema bag, using room air, until the patients stated they
were full or until 2,000 mL had been administered. The adequacy of the air
insufflation was evaluated with a CT scout view, with more air insufflated if
required. After scanning in the prone position had been completed, patients
were positioned supine. After the administration of several additional puffs
of air, the rectal tube was removed to improve patient comfort and prevent a
possible rectal lesion. CT was performed after IV injection of an iodinated
contrast agent, iopromide (Ultravist 300, Schering); 140 mL of contrast agent
was administered at 3 mL/sec. CT was performed in the portal venous phase
(start delay of 60 sec). The scanning parameters were 0.75-mm collimation, a
pitch of 1–1.3, 120 kVp, 120–160 mAs, a 512 x 512 matrix, a
0.5-sec gantry rotation, and a 1-mm reconstruction thickness with 0.7-mm
reconstruction intervals. The average acquisition time was 15 sec.
Images were processed on a PC, using commercially available software
(Rapidia, Infinitt). The processed images included sagittal and coronal 2D
reformatted, ray-sum, and virtual colonoscopy images. The evaluation began
with 2D transverse CT images, followed by a virtual colonoscopy 3D view. The
virtual colonoscopy viewing was performed both antegrade and retrograde, with
the patient both supine and prone, to avoid blind areas.
The findings for each patient were prospectively recorded. The presence,
location, size, and morphologic features of colorectal cancers and polyps were
assessed. For TNM staging, the depth of tumor invasion and the presence of
lymph nodes and metastases were evaluated. Tumor and node staging was based on
the international TNM classification. Each tumor was assigned a stage, with T1
indicating a tumor in the mucosal or submucosal layer; T2, a tumor extending
into but not beyond the muscularis propria layer; T3, a tumor extending beyond
the muscularis propria layer into perirectal tissue and fat; and T4, a tumor
directly invading other organs or structures. In CT image analysis, T1 and T2
were regarded as the same stage, because they are difficult to distinguish
from each other on imaging. T3 was defined as a tumor with dense spiculation
in the perirectal fat or a nodular or lobulated outer margin. For the node
stage, N0 indicated no regional lymph node metastasis; N1, metastasis in one
to three pericolic and perirectal lymph nodes; and N2, metastasis in four or
more pericolic and perirectal lymph nodes. The number of positive lymph nodes
was defined as the number of clustered nodes, independent of their size, or
any lymph node measuring at least 1 cm in the long axis. M0 was no evidence of
metastasis, and M1 was distant metastasis. The liver and both lower lungs were
evaluated for metastasis. The diagnostic accuracy of TNM staging was
calculated, and the sensitivity of CT colonography for the detection of
cancers and polyps was compared with that of colonoscopy. All CT images were
independently interpreted and evaluated by two abdominal radiologists who had
at least 5 years' experience with abdominal CT and, to minimize any selection
bias, were unaware of the colonoscopic findings. Differences in assessment
were resolved by consensus. With regard to polyp detection, eight cases had to
be resolved by consensus. The CT colonography results were compared with the
surgicopathologic and follow-up colonoscopic findings, which served as the
reference standard.
Eight colonic segments were considered to evaluate differences in overall
colonic distention between the supine and prone positions: rectum, sigmoid
colon, descending colon, splenic flexure, transverse colon, hepatic flexure,
ascending colon, and cecum. We did not use objective criteria for colonic
distention that have been defined in the literature
[13]. Instead, colonic
distention was ranked using the following three-point system: 1, suboptimal
distention in more than three segments; 2, suboptimal distention in one to two
segments; and 3, optimal distention visibility throughout. Optimal distention
was defined as colonic wall and haustral folds that were
"pencil-thin" throughout the segment. Suboptimal distention was an
inadequate distention, with a slightly thickened colonic wall and haustral
folds or lumen collapse.
Wilcoxon's signed rank test was used to evaluate the technical results for
colonic distention. For detection of colorectal cancers and polyps, the
statistical difference between the CT and initial colonoscopies was determined
with Fisher's exact test. For all tests, a p value of 0.05 or less
indicated a statistically significant difference.
Results
Cancer Location and Detection
Twenty patients underwent surgery, and 21 colorectal cancers (one
synchronous cancer) were revealed (Table
1). A right hemicolectomy was performed on eight patients, and a
lower anterior resection on 12. The mean duration between surgery and CT
colonography was 2 weeks. CT colonography was able to detect all 21 colorectal
cancers, including one synchronous cancer that was missed on an incomplete
colonoscopy. At colonoscopic examination of 11 patients with colorectal
cancer, a distal occlusion was found but colonoscopy could not explore the
colon segments proximal to the occlusion. In the remaining nine patients,
colonoscopy could explore the lumen of the whole colon. The sensitivity of CT
colonography for the detection of cancer was 100%, with no false-positive or
false-negative diagnoses of cancer (Figs.
1A,
1B,
1C, and
1D). Colonoscopy was repeated
after surgery but revealed no malignancies. The difference between CT
colonography and colonoscopy for the detection of colorectal cancer was not
statistically significant. The mean duration of image processing and
interpretation was 28 min.
View larger version (131K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1D. —52-year-old man with polypoid occlusive carcinoma in rectum.
Coronal multiplanar reconstruction image shows tumor invasion through
muscularis propria into subserosa (arrow) (T3), with no regional
lymph node (N0).
|
|
Polyp Detection
The segmental locations of the polyps are shown in
Table 1. Grouping according to
polyp diameter gave a sensitivity of 100% (6/6 for detection of polyps of 10
mm or larger); 94% (15/16) for polyps of 6–9 mm; and 84% (6/19) for
polyps of 5 mm or smaller. Forty-one polyps were found at surgery or
colonoscopy, nine of which were missed on the initial colonoscopy (Figs.
2A,
2B,
2C, and
2D) but showed no malignancy on
histologic examination after surgical and endoscopic removal. Among these,
obstruction prevented the evaluation of seven, and two were false-negatives.
The sensitivity of CT colonography, and the numbers of false-positive and
false-negative CT colonography findings, for the detection of polyps are
reported in Table 2. The
difference between CT colonography and colonoscopy for polyp detection was
statistically significant (p = 0.001). Thirteen false-positive
diagnoses (Figs. 3A,
3B, and
3C) were due to residual fecal
material. There were four false-negative diagnoses, three due to polyps
smaller than 5 mm in diameter and one due to retained fluid (Figs.
4A,
4B, and
4C).
View larger version (150K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A. —65-year-old man with colonic polyp that was missed on
conventional colonoscopy. Conventional colonoscopy image shows large,
lobulated mass in sigmoid colon. Colonoscope could not cross mass.
|
|
View larger version (147K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3C. —61-year-old man with false-positive findings for colonic
polyp. Narrow-window-setting transverse CT colonography image shows tiny gas
bubble at center of elevated lesion (arrow); this finding confirmed
presence of residual fecal material. No polyp was seen on colonoscopy.
|
|
View larger version (104K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4B. —59-year-old man with false-negative findings for colonic
polyp. Virtual colonoscopy (B) and transverse CT (C) images show
fluid-filled descending colon (arrows). Lesion was masked by fluid
and missed on CT colonography.
|
|
View larger version (153K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4C. —59-year-old man with false-negative findings for colonic
polyp. Virtual colonoscopy (B) and transverse CT (C) images show
fluid-filled descending colon (arrows). Lesion was masked by fluid
and missed on CT colonography.
|
|
TNM Staging
The overall diagnostic accuracies for TNM staging of colorectal cancer were
95%, 85%, and 100% for tumor, node, and metastasis, respectively
(Table 3). The tumor stage was
correctly predicted in 20 of 21 cancers, with one tumor overstaged (Figs.
5A and
5B). The node stage was
correctly predicted in 17 of 20 cancers and incorrectly predicted in three,
two of which were overstaged and one understaged. In two over-staged tumors,
the lymph nodes were larger than 1 cm (range, 1.2–1.5 cm) and were
called metastatic. All these nodes were characterized as reactive on
pathology. The metastasis stage was correctly predicted in all patients. Only
one patient had a liver metastasis, which was confirmed by biopsy. Follow-up
CT scans revealed no additional metastatic lesions.
View larger version (116K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5A. —48-year-old woman with overstaged T2 colon cancer. Axial
(A) and sagittal (B) reformatted images clearly show irregular
rectal mass with perirectal fat stranding (arrows). Initial
preoperative diagnosis was T3 lesion, but lesion proved to be T2.
|
|
View larger version (98K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5B. —48-year-old woman with overstaged T2 colon cancer. Axial
(A) and sagittal (B) reformatted images clearly show irregular
rectal mass with perirectal fat stranding (arrows). Initial
preoperative diagnosis was T3 lesion, but lesion proved to be T2.
|
|
Technical Results of CT Colonography
The whole colon was fully distended in 34 patients (83%) and 32 patients
(78%) when the patients were supine and prone, respectively. Mean overall
bowel distention when the patients were supine and prone was ranked 2.80 (SD,
± 0.4) and 2.78 (± 0.4), respectively, but the difference was
not statistically significant (p = 0.781).
Discussion
The main benefits of CT colonography in colorectal cancer are its ability
to evaluate the entire colon, to complete an unsuccessful colonoscopic
examination, and to permit TNM staging. In our study, 11 patients underwent CT
colonography for a distal occlusive carcinoma. In one of these 11 patients, a
synchronous lesion was detected on CT colonography only. Other advantages of
CT colonography over conventional colonoscopy include a shorter procedure,
less risk to the patient, no need for IV sedation, and greater precision in
localizing lesions. However, the disadvantages of CT colonography include the
need for bowel cleansing, the long viewing and interpretation times, and the
radiation dose.
MDCT has several advantages over single-detector CT, including increased
temporal and spatial resolutions, faster data acquisition, and a wider field
of view and comparable coverage times, with much thinner section collimation
[14]. The use of thinner
section collimation makes near-isotropic voxels for CT colonography. The
advantages of an isotropic image include improved rates of polyp detection and
accuracy of TNM staging because of reduced volume averaging and improved
z-axis resolution for multiplanar reformations and 3D viewing.
Compatible with Dukes' classification, TNM staging adds greater precision to
the identification of prognostic subgroups. Although comparisons between 16-
and 4-MDCT for CT colonography remain to be studied, 16-MDCT theoretically may
yield greatly improved polyp detection rates, decreased false-positive
findings, and more accurate TNM staging than does 4-MDCT. In our study,
correlation with pathologic TNM stage was good, and sensitivities for the
detection of polyps with a diameter of less than 5 mm, of 6–9 mm, and of
more than 10 mm were 84%, 94%, and 100%, respectively. These results were
higher than in previously reported studies using 4-MDCT
[3,
11,
14]. Two additional reasons
were thought responsible for the greater sensitivity of CT colonography over
that of conventional colonoscopy for the detection of polyps. First, CT
colonography could be used to evaluate the proximal colon beyond an
obstructive lesion, whereas the colonoscope was unable to pass the
obstruction. Second, CT colonography has no blind areas, such as the opposite
side of the colonic fold, which colonoscopy cannot depict. In our study, the
colonoscopic results were false-negative for seven polyps because of an
occlusive carcinoma, with two polyps located near the colorectal cancer. These
conditions were thought to be the main factors causing the false-negative
results for polyps in patients with colorectal cancer. Also, 16-MDCT can
reduce respiratory artifacts by reducing scanning time.
For polyp detection, interpretation technique is important. Review of
transverse images as the primary interpretation technique is difficult and
takes a long time for detection of small polyps. Smaller polyps might be
quickly and easily detected through evaluation of virtual colonoscopy images.
Therefore, transverse images and complete virtual colonoscopy images in both
the antegrade and retrograde directions, with the patient both supine and
prone, were used to increase the detection rate of small polyps. Although
diagnostic accuracy for polyp detection was not compared between 2D and 3D
interpretations, several polyps that were initially missed on the transverse
image were later detected on the endoluminal image. No instance of the reverse
occurred.
In this study, thinner-section MDCT was found to reduce false-positive
results. The major causes of false-positive findings are residual fecal
material, retained fluid, suboptimal bowel distention, and respiratory
artifacts [15]. However, the
combination of virtual colonoscopic images, transverse views, and multiplanar
images was helpful in distinguishing residual fecal material from a colonic
polyp. Internal heterogeneity was better depicted by the transverse image, and
external morphologic features were better depicted by the virtual colonoscopy
view. Morrin et al. [16]
reported that contrast-enhanced CT colonography increased diagnostic
confidence about medium-sized polyps (diameter, 5–9 mm). Residual fecal
material can also be digitally subtracted from the images by dietary fecal
tagging [17]. Although
polyethylene glycol is highly effective at cleansing the bowel, it often
leaves retained fluid in the colon. Bisacodyl decreases the residual fecal
material and retained fluid. Retained fluid and collapsed segments have been
markedly reduced by the combination of supine and prone imaging. No
statistically significant difference in mean overall bowel distention was
found between the supine and prone positions.
Our study has some limitations. First, in the TNM classification, a
sensitivity of 100% was reported for detection of cancer. However, this
percentage was inaccurate and potentially misleading. Eleven patients had
obstructive lesions, and these were easy to detect. In such cases, the
sensitivity is in localization of the lesion, not its detection. Second, the
imaging criteria for TNM staging and the criteria for determining
lymphadenopathy are unclear. Differentiation of T1 from T2 is also difficult.
The number of patients with positive lymph nodes and metastasis was
insufficient to determine the accuracy of CT colonography and should be
addressed to evaluate colon cancer. In our study, only one patient had
metastases. Third, no objective criteria were used for bowel distention.
Taylor et al. [13] graded
colonic distention as partial collapse (grade 1), suboptimal distention (grade
2), or optimal distention (grade 3) according to the thickness of the colonic
wall and haustral fold. Optimal distention was defined as a colonic wall that
was pencil-thin throughout the segment, with haustral folds less than 2 mm
thick throughout their length.
In conclusion, CT colonography using 16-MDCT correlated well with
pathologic TNM stage and was more sensitive than colonoscopy for the detection
of cancers and polyps. Our preliminary data suggest that for patients in whom
colorectal cancer is clinically suspected, CT colonoscopy would be valuable
for staging the tumor and detecting polyps or cancers in areas not evaluated
by conventional colonoscopy. This possibility should be confirmed by a larger
study.
References
- Vining DJ, Gelfand DW, Bechthold RE, Scharling ES, Grishaw EK,
Shifrin RY. Technical feasibility of colon imaging with helical CT and virtual
reality. (abstr) AJR1994; 162[suppl]:104
- Gluecker TM, Johnson CD, Harmsen WS, et al. Colorectal cancer
screening with CT colonography, colonoscopy, and double contrast barium enema
examination: prospective assessment of patient perceptions and preferences.
Radiology2003; 227:378
-384[Abstract/Free Full Text]
- Macari M, Bini EJ, Xue X, et al. Colorectal neoplasm: prospective
comparison of thin-section low-dose multi-detector row CT colonography and
conventional colonoscopy for detection. Radiology2002; 224:383
-392[Abstract/Free Full Text]
- Macari M, Milano A, Lavelle M, Berman P, Megibow AJ. Comparison of
time-efficient CT colonography with two-dimensional and three-dimensional
colonic evaluation for detecting colorectal polyps.
AJR 2000;174:1543
-1549[Abstract/Free Full Text]
- Ferrucci JT. Colon cancer screening with virtual colonoscopy:
promise, polyps, politics. AJR2001; 177:975
-988[Free Full Text]
- Hara AK, Johnson CD, Reed JE, et al. Detection of colorectal polyps
with CT colonography: initial assessment of sensitivity and specificity.
Radiology1997; 205:59
-65[Abstract/Free Full Text]
- Yee J, Akerkar GA, Hung RK, et al. Colorectal neoplasia of CT
colonography for detection in 300 patients. Radiology2001; 219:685
-691[Abstract/Free Full Text]
- Fenlon HM, Nunes DP, Schroy PC, et al. A comparison of virtual and
conventional colonoscopy for the detection of colorectal polyps. N
Engl J Med 1999;341:1496
-1503[Abstract/Free Full Text]
- Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer
screening: clinical guidelines and rationale.
Gastroenterology1997; 112:594
-642[Medline]
- Rex DK, Cutler CS, Lemmel GT, et al. Colonoscopic miss rates of
adenomas determined by back-to-back colonoscopies.
Gastroenterology1997; 112:24
-28[Medline]
- Hara AK, Johnson CD, MacCarty RL, Welch TJ, McCollough CH, Harmsen
WS. CT colonography: single- versus multi-detector row imaging.
Radiology2001; 219:461
-465[Abstract/Free Full Text]
- Neri E, Giusti P, Battolla L, et al. Colorectal cancer: role of CT
colonography in preoperative evaluation after incomplete colonoscopy.
Radiology 2002;223
: 612-619
- Taylor SA, Halligan S, Goh V, et al. Optimizing colonic distention
of multi-detector row CT colonography: effect of hyoscine butylbromide and
rectal balloon catheter. Radiology2003; 229:99
-108[Abstract/Free Full Text]
- Rydberg J, Buckwalter KA, Caldemeyer KS. Multisection CT: scanning
techniques and clinical applications. RadioGraphics2000; 20:1787
-1806[Abstract/Free Full Text]
- Fletcher JG, Johnson CD, Welch TJ, et al. Optimization of CT
colonography technique: prospective trial in 180 patients.
Radiology 2000;216
: 704-711[Abstract/Free Full Text]
- Morrin MM, Farrell RJ, Kruskal JB, Reynolds K, McGee JB,
Raptopoulos V. Utility of intravenously administered contrast material at CT
colonography. Radiology2000; 217:765
-771[Abstract/Free Full Text]
- Lefere PA, Gryspeerdt SS, Dewyspelaere J, Baeklandt M Van Holsbeeck
BG. Dietary fecal tagging as cleansing method before CT colonography: initial
results—polyp detection and patient acceptance.
Radiology2002; 224:393
-403[Abstract/Free Full Text]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
|
|
 
M. Minagawa, J. Yamamoto, S. Miwa, Y. Sakamoto, N. Kokudo, T. Kosuge, S.-i. Miyagawa, and M. Makuuchi
Selection Criteria for Simultaneous Resection in Patients With Synchronous Liver Metastasis
Arch Surg,
October 1, 2006;
141(10):
1006 - 1012.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
B. P. Mulhall, G. R. Veerappan, and J. L. Jackson
Meta-Analysis: Computed Tomographic Colonography
Ann Intern Med,
April 19, 2005;
142(8):
635 - 650.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
