| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Thomas Jefferson University, 1096 Main Bldg., 132 S.
10th St., Philadelphia, PA 19107.
2 Present address: Department of Radiology, Affiliated Hospital of North Sichuan
Medical College, 113 Culture Rd., Nanchong, Sichuan 637000, People's Republic
of China.
3 Division of Gastroenterology and Hepatology, Department of Medicine, Thomas
Jefferson University, Philadelphia, PA 19107.
4 Department of Pathology, Anatomy, and Cell Biology, Thomas Jefferson
University, Philadelphia, PA 19107.
Received July 31, 2001;
accepted after revision February 11, 2002.
Address correspondence to D. G. Mitchell.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Fifty patients with chronic active hepatitis C, who had MR imaging examinations and a related histology report from a liver biopsy obtained within 1 month of the MR imaging were chosen from our radiology database and studied retrospectively. All patients were examined over a 4-year period at a single institution to detect cirrhosis or hepatocellular carcinoma. We divided the 50 patients into the mild, moderate, or severe activity groups, according to their histology reports. Two radiologists, unaware of the histologic classifications, individually reviewed the MR images to observe the perihepatic locations, number, size (defined as the sum of the length-by-width products of the largest three nodes), and intensity of the lymph nodes relative to the spleen. The clinical records of the patients were reviewed to check the results of their liver function tests. The lymph node findings on MR imaging were compared with the histologically confirmed activity level of chronic hepatitis C.
RESULTS. Forty-four (88.0%) of 50 patients had perihepatic lymph nodes larger than 5 mm on MR images, including 64.2% (9/14) of the patients with mild activity, 96.3% (26/27) of the patients with moderate activity, and 100% (9/9) of the patients with severe activity (p = 0.0034). The average number ± the standard deviation (SD) of perihepatic lymph nodes was 2.5 ± 1.8 in patients with mild activity, 5.6 ± 2.2 in patients with moderate activity, and 8.3 ± 3.5 in patients with severe activity (p = 0.0001). The average size (± SD) of the lymph nodes was 151.0 ± 104.9 mm2 in the mild activity group, 366.8 ± 143.0 mm2 in the moderate activity group, and 488.2 ± 244.8 mm2 in the severe activity group (p = 0.0001). On fat-saturated fast spin-echo T2-weighted MR images, the average number (± SD) of hyperintense nodes was 0.17 ± 0.25 in the mild activity group, 1.7 ± 0.80 in the moderate activity group, and 2.4 ± 0.60 nodes in the severe activity group (p = 0.0001). No relationship between histologic activity and results from liver function tests was found.
CONCLUSION. MR imaging depicts perihepatic lymph nodes in most patients with chronic hepatitis C. Lymph node number, size, and hyperintensity were related to the activity of chronic hepatitis C, but the results of liver function tests were not.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Inflammatory processes in organs frequently lead to hyperplasia of regional lymph nodes. Lymph node enlargement in the hepatoduodenal ligament could reflect the inflammatory response of the host. It has been reported that sonography can reveal lymph nodes in the hepatoduodenal ligament in most patients with chronic hepatitis and help in determining total perihepatic lymph node volume corresponding to viremia and liver histology, thus enabling clinicans to predict the presence of severe inflammatory activity and to assess the response to antiviral therapy reflected by changes in the perihepatic lymph node volume [6,7,8,9,10,11].
MR imaging can not only show the anatomy of the liver and pancreas clearly [12] but can also depict enlarged abdominal lymph nodes and their locations relative to adjacent bile ducts or vascular structures [13]. MR imaging can provide better contrast between the lymph nodes and the liver than does sonography or CT, is superior to sonography for revealing deep abdominal lesions and perihepatic nodes [12,13,14], and may also be the most accurate non-invasive modality for detecting hepatocellular carcinoma [15]. We hypothesized that perihepatic lymphadenopathy on MR images could be used to grade the activity of chronic hepatitis C. In our study, we retrospectively reviewed the appearances of lymph nodes on MR imaging in patients with chronic hepatitis C, measuring perihepatic lymph node presence, number, size, and hyperintensity. We tested the relationship of these MR findings to a biopsy-derived grading of activity level. We also compared MR findings and activity level of chronic hepatitis C with the results of liver function tests among patients for whom these test results were available.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Using a protocol approved by the institutional review board for retrospective studies in the radiology department, we selected 50 patients, including 17 women and 33 men whose ages ranged from 31 to 75 years (average age ± SD, 51.3 ± 9.4 years). Most MR examinations had been performed to detect or exclude complications of hepatitis and cirrhosis, such as portal hypertension or malignancy. For most patients, these were the initial MR examinations, but when more than one examination was available for a patient, the one performed closest to the time of the biopsy was chosen for review.
Twenty-three of our 50 patients had laboratory data that had been collected within 2 months of the MR imaging examination. The data included the alanine transaminase level, aspartate transaminase level, and alanine transaminase—aspartate transaminase ratio.
MR Imaging Technique
All examinations were performed on a 1.5-T system (General Electric Medical
Systems, Milwaukee, WI). The sequences included coronal and axial fast
spin-echo T2-weighted MR images; axial single-shot fast spin-echo T2-weighted
MR images; spoiled gradient-echo T1-weighted in- and out-of-phase MR images;
and three-dimensional enhanced spoiled gradient-echo dynamic MR images.
Coronal single-shot fast spin-echo T2-weighted MR images were obtained in two or more breath-holds with a phased array torso coil. The parameters used were TR/TE range, infinite/90-100 or 180-190; section thickness, 7 mm; intersection gap, 0; matrix, 256 x 192; signal acquired, one half; and field of view, 32 x 32 cm.
Axial single-shot fast spin-echo T2-weighted images were obtained in one or two breath-holds. The parameters were infinite/90-100 or 180-190; section thickness, 5 mm; intersection gap, 0; matrix, 256 x 160; signal acquired, one half; and field of view, 32 x 24 cm.
Axial fat-saturated fast spin-echo T2-weighted images were obtained with 2500/80; matrix, 256 x 192; section thickness, 7 mm; intersection gap, 0; signal acquired, 1; and field of view, 32 x 24 cm.
Axial two-dimensional multisection spoiled gradient-echo T1-weighted images were obtained during breath-holding. The parameters were 120/4.2 (in-phase) and 2.1 (out-of-phase); flip angle, 90°; section thickness, 8 mm; intersection gap, 0; matrix, 256 x 192; signal acquired, 1; and field of view, 32 x 24 cm.
Axial three-dimensional spoiled gradient-echo dynamic MR images were obtained with 6.1/2.1; flip angle, 15-20°; matrix, 256 x 128; signal acquired, 1; section thickness, 5 mm; overlap, 2.5 mm; and field of view, 32 x 24 cm.
Contrast material (0.1 mmol/L per kilogram of body weight of gadopentetate dimeglumine, Magnevist; Berlex Laboratories, Wayne, NJ) was IV administered by hand at a rate of approximately 2 mL/sec followed by a 20-mL saline solution flush. Imaging was initiated at a time determined using a timing bolus sequence to optimize first-pass artery enhancement.
In all patients, dynamic imaging was performed with four breath-hold sequences—before the injection (corresponding to unenhanced imaging), immediately after the injection (arterial dominant phase of enhancement), 30 sec afterward (early venous phase of enhancement), and 1 min afterward (late venous phase of enhancement). An additional delayed phase was acquired using a two-dimensional single-section spoiled gradient-echo technique with parameters of 20/1.8 and flip angle, 40°.
Image Interpretation
The original MR imaging data were loaded onto a workstation (Cannon; Sun
Microsystems, Palo Alto, CA) and evaluated (Rational Imaging software;
Intuitive Software, West Hills, CA). Two radiologists independently reviewed
the images without knowing the biopsy results or other clinical data.
In patients with chronic hepatitis C, deep lymph nodes are located mainly in the hepatoduodenal ligament [6]. We divided the perihepatic locations of the lymph nodes on the MR images into four groups: periportal (at the hilum, anterior to portal vein and to the right of the proper hepatic artery), peripancreatic (in close contact with the anterior aspect of the pancreas, mainly at the body), portocaval (space), and celiac (in front of the hepatic artery and adjacent to its origin from the celiac axis). All nodes with a short-axis diameter larger than 5 mm were measured, and their locations and the number of nodes present were recorded. The largest three nodes were measured at their largest and smallest diameters; these measurements were then multiplied to produce the diameter products. We referred to the diameter products as the "size" of the nodes. The signal intensity of lymph nodes on fat-suppressed T2-weighted images was considered hyperintense if the intensity was greater than that of the spleen.
Biopsy Interpretation
Experience hepatologists performed most of the percutaneous liver biopsies
without imaging guidance using an 18-gauge spring-loaded biopsy device (ASAP
biopsy system; Boston Scientific, Boston, MA). Approximately 10% of the
biopsies were performed with sonographic guidance, usually because of a large
body habitus or the presence of a right-lobe cyst or hemangioma.
All biopsies were interpreted by the same hepatopathologist and included a qualitative grading of histologic activity using a method described by Desmet et al. [16] and a semiquantitative grading using the histologic activity index described by Knodell et al. [17]. The components of the histologic activity index included periportal necrosis with or without bridging necrosis, intralobular degeneration and focal necrosis, portal inflammation, and fibrosis [17]. Although cirrhosis of the liver was also staged on the basis of the severity of fibrosis and architectural abnormalities, this disease was not studied as part of our investigation. After retrospectively reviewing the reports that included the previously mentioned findings, we divided the patients into three groups: mild, moderate, and severe histologic activity of chronic hepatitis C. The lymph node findings on MR imaging for the three groups were compared.
Statistical Analysis
In determining the presence or absence of lymph nodes, we counted a node as
present only if both observers agreed. In determining the number, size, and
hyperintensity of the nodes, the results were given as the mean of the two
observers' values ± one standard deviation, which sometimes resulted in
obtaining a fraction of a node as a total. Agreement was measured by kappa
values for the presence or absence of nodes and by the intraclass correlation
for the continuous variables of number, size, and hyperintensity
[18]. Both of these statistics
are generally interpreted as follows: a kappa value equal to or more than 0.81
indicates very good agreement; a kappa value ranging from 0.80 to 0.61
indicates good agreement; a kappa value ranging from 0.60 to 0.41 indicates
moderate agreement; and a kappa value of less than 0.41 indicates poor
agreement.
For testing the relationship of the presence or absence of nodes and the activity level of hepatitis C, Fisher's exact test was used. For testing the relation of number, size, or hyperintensity of the nodes to the activity level of hepatitis C, the one-way analysis of variance was computed, and pairwise comparisons of the groups were made with Bonferroni's t tests. We tested the relationship of liver function variables to the presence or absence of lymph nodes with Student's t tests and to the continuous variables with Pearson product moment correlations. All analyses were computed using statistical software (SAS version 8.2 for Windows; SAS Institute, Cary, NC).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Perihepatic Lymph Node Location
Lymph nodes were distributed in the four specified locations as follows:
periportal, 44.1% (115.5/262); portocaval space, 33.6% (88/262); celiac, 17.0%
(44.5/262); and peripancreatic, 5.3% (14/262).
Reliability
Interobserver agreement between the two radiologists was generally good.
Agreement as to the presence or absence of perihepatic lymph nodes larger than
5 mm in diameter showed a kappa value of 0.64. Observations concerning the
number of nodes had a similar reliability, with an intraclass correlation of
0.62. Agreement on the size of the nodes had an intraclass correlation of
0.65. Agreement on the hyperintensity of nodes was slightly lower, with an
intraclass correlation of 0.58.
Presence of Lymph Nodes
Both radiologists noted at least one perihepatic lymph node larger than 5
mm in 88% (44/50) of the patients. The presence of the lymph nodes was related
to activity level of hepatitis C (Fisher's exact test, p = 0.0034).
Nodes were seen in 64.2% (9/14) of patients with mild activity, 96.3% (26/27)
of patients with moderate activity, and 100.0% (9/9) of patients with severe
activity. Table 1 summarizes
the results regarding nodal presence, as well as their number, size, and
hyperintensity.
|
Number of Lymph Nodes
Patients had an average of 5.2 ± 3.1 lymph nodes. The number of
nodes was related to activity level of hepatitis C (F = 17.9, p <
0.0001). Patients with mild hepatitis C activity had an average of 2.5
± 1.8 nodes, those with moderate hepatitis C activity had an average of
5.6 ± 2.2 nodes, and those with severe hepatitis C activity had an
average of 8.3 ± 3.5 nodes. Additionally, all pairwise comparisons were
significantly different (p < 0.05).
Size of Lymph Nodes
The average node size was 328.3 ± 195.1 mm2. Node size
was related to hepatitis C activity level (F = 14.5, p < 0.0001).
Average node size was 151.0 ± 104.9 mm2 in those patients
with mild activity, 366.8 ± 143.0 mm2 in those with moderate
activity, and 488.2 ± 244.8 mm2 in those with severe
activity. Pairwise comparisons showed differences between the mild and
moderate groups and the mild and severe groups (p < 0.05) but no
differences between moderate and severe groups.
Number of Hyperintense Lymph Nodes
Patients had an average of 1.4 ± 1.0 hyperintense nodes.
Hyperintense lymph nodes were related to hepatitis C activity level (F = 36.7,
p < 0.0001). There were 0.17 ± 0.25 hyperintense nodes
among patients with mild hepatitis C activity, 1.7 ± 0.80 among those
with moderate hepatitis C activity, and 2.4 ± 0.60 among those with
severe activity. All pairwise comparisons were significant (p <
0.05).
Hepatocellular Carcinoma
Among the four patients with hepatocellular carcinoma, all had chronic
hepatitis C of moderate activity. Two of the four patients had cirrhosis of
the liver as well. Three hepatocellular carcinomas were less than 3.0 cm in
diameter, and one was 4.5 cm in diameter. No significant differences were
found in the number, size, or hyperintensity of nodes between the four
patients who had hepatocellular carcinoma as well as moderately active
hepatitis C and the 23 patients who did not have hepatocellular carcinoma but
who also had moderately active hepatitis C.
Liver Function Tests
Twenty-three patients had liver function tests available that were obtained
within 2 months of MR imaging. Pearson product moment correlations between the
alanine transaminase level, aspartate transaminase level, and alanine
transaminase—aspartate transaminase ratio and the number, size, and
hyperintensity of the lymph nodes ranged from -0.07 to 0.26 and were not
significant. Analyses of variance showed no relationship between the activity
level of hepatitis C and the alanine transaminase level (F = 0.21, not
statistically significant), aspartate transaminase level (F = 0.35, not
statistically significant), or alanine transaminase—aspartate
transaminase ratio (F = 0.42, not statistically significant). Additionally,
none of the pairwise comparisons between the activity levels of chronic
hepatitis C and the alanine transaminase level, aspartate transaminase level,
or alanine transaminase—aspartate transaminase ratio was
significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In our study, we found four MR imaging signs of chronic hepatitis C related to its activity: the presence of perihepatic lymph nodes larger than 5 mm and the number, size, and hyperintensity of these nodes on fat-suppressed T2-weighted images. In our series, 88% of patients with chronic active hepatitis C had perihepatic lymph nodes larger than 5 mm, compared with a reported prevalence of enlarged perihepatic lymph nodes using sonography of from 67% to 100% [6, 7].
A relationship between inflammation and lymph node hyperplasia in liver diseases, especially chronic hepatitis C [6, 7, 22, 23] has previously been noted. Some investigators believe that routine evaluation of perihepatic lymphadenopathy is a superior method of depicting hepatic inflammatory activity than is the monitoring of alanine aminotransferase levels [10]. Enlarged lymph nodes within the dorsal portion of the hepatoduodenal ligament can usually be identified on sonography, but it may be more difficult to detect the lymph nodes in the ventral portion of the hepatoduodenal ligament because of surrounding fat and connective tissue [7]. Furthermore, it may be difficult to detect lesions smaller than 10 mm in the retroperitoneum on sonography [24].
Forsberg et al. [25] stated that regional lymphadenopathy is not usually found in hepatitis B surface antigen—positive hepatitis but is frequently found in primary biliary cirrhosis, sclerosing cholangitis, and immune-mediated liver diseases. The researchers found that 63.6% (21/33) of the patients with active immune-mediated liver diseases had pathologic lymph nodes detected on sonography. In most patients, the enlarged nodes were distributed along the course of lymphatic drainage of the liver and biliary ducts, which is mainly through the nodes in the hepatoduodenal ligament, running from the hilum of the liver to the first portion of the duodenum [22]. The mechanism of portal lymphadenopathy in patients with chronic hepatitis C is probably related to viral replication within the liver and the immune-mediated inflammatory response of the host [7]. Recruitment of hepatitis C virus—infected lymphocytes and macrophages into the draining lymph nodes may be a contributing mechanism [6]. In our series, enlarged perihepatic lymph nodes were mainly in the periportal (44.3%) or portocaval (33.2%) regions.
Lyttkens et al. [8] found a significant correlation between lymph node size and the level of serum alkaline phosphatase in chronic active hepatitis and between the size of the lymph nodes and the level of gammaglutamyl transpeptidase in chronic hepatitis C. A close correlation was found between lymph node size and parenchymal damage in patients with chronic hepatitis C [9]. In patients with chronic hepatitis C, sonographic assessment of lymph nodes in the hepatoduodenal ligament revealed a correlation between the total perihepatic lymph node volume and histologic activity index. Patients with histologically severe inflammation showed a larger total perihepatic lymph node volume than those with only mild to moderate inflammation [7]. Furthermore, a decreased total perihepatic lymph node volume also reflected histologic improvement in patients treated for chronic hepatitis C [11]. Dietrich et al. [11] suggested that the sonographic assessment of the perihepatic lymph node volume after treatment might replace liver biopsy. In our series, the total number of periportal and perihepatic lymph nodes, as well as their number, size, and signal intensity, were related to the activity of chronic hepatitis C (Table 1).
Our MR findings concerning the size and number of lymph nodes in patients with hepatitis C can presumably be applied to CT findings. Although CT is commonly used to reveal lymphadenopathy in patients with suspected metastatic disease, few reports have described its use in depicting lymph nodes in patients with hepatitis or cirrhosis [26,27,28,29]. In these reports, mildly enlarged lymph nodes were found in patients with various forms of cirrhosis, particularly primary biliary cirrhosis, but the findings were not correlated with the severity of disease. In our institution, MR imaging has been the preferred modality for the routine evaluation of patients with diffuse liver disease and for detection of early hepatocellular carcinoma. MR imaging is performed in all patients being considered for imminent or eventual liver transplantation, whereas CT is used more commonly for evaluating patients after transplantation. Although older MR imaging techniques had limited value for revealing small abdominal lymph nodes because of motion-induced artifacts, we are not aware of any recent literature comparing the current rapid MR imaging and CT techniques. CT has the advantages of higher spatial resolution and less dependence on cessation of motion, but MR imaging can better depict the contrast between lymph nodes and the liver, which can be helpful for distinguishing perihepatic lymph nodes from the adjacent caudate lobe. Additionally, MR imaging uses no ionizing radiation and, therefore, may be more suitable for repeated imaging of patients during long-term follow-up. Because our investigation was restricted to analysis of lymph nodes with diameters larger than 5 mm we do not believe that the lower spatial resolution of MR imaging was a substantial disadvantage.
In addition to noting the number and size of lymph nodes, we also qualitatively evaluated the signal intensity of lymph nodes, a unique application of MR imaging. We found that the signal intensity of the perihepatic lymph nodes on fat-suppressed fast spin-echo T2-weighted images increased with the activity of chronic hepatitis C (Figs. 1A,1B,2,3). We measured only the signal intensity of the largest three nodes, because the size of the perihepatic lymph nodes correlated with the activity of the inflammation [7,8,9]; thus, the largest nodes might have the most inflammation. Inaccurate measurements due to partial volume effects are also less likely to occur with large nodes. The hyperintensity of those nodes may result from the swelling of lymph nodes in patients with chronic active hepatitis C [19] and its associated increased water retention. Although we did not measure signal intensity quantitatively, such measurements may have been helpful.
|
|
|
|
Several reports have shown that liver function test results do not correlate with perihepatic lymph nodes in patients with chronic hepatitis C [6, 7, 10]. Although slightly less than half of the patients in our series had liver function tests that had been obtained within 2 months of the MR imaging and biopsy examinations, the lack of a significant relationship between the results of the liver function tests on one hand and the histologic activity, number of lymph nodes, or size of the nodes on MR images on the other hand is consistent with these prior observations. This suggests that MR imaging may be superior to liver function tests in evaluating the activity of chronic hepatitis C.
In our series, four patients had hepatocellular carcinoma. In a study of 660 consecutive patients with hepatocellular carcinoma who underwent autopsy, the researchers found no metastasis from tumors smaller than 3 cm in diameter. Metastatic lymph node sites were peripancreatic, 13.6%; hepatic hilar, 13.5%; and perigastric, 10.8% [30]. Of the four hepatocellular carcinomas in our patients, three had diameters smaller than 3 cm, and one had a diameter of 4.5 cm. Among these patients, the perihepatic lymph nodes were mainly located at the porta hepatis and portocaval space, which are the same nodal locations as those in the patients without hepatocellular carcinoma who had the same activity level of hepatitis C. Therefore, we believe that the perihepatic lymph nodes in these four patients resulted from chronic active hepatitis rather than from metastases from hepatocellular carcinoma.
In our study, 27 of 50 patients had cirrhosis of the liver. However, in clinical practice, patients with earlier stages of the disease are the most important population. A prospective study with a larger number of patients with precirrhotic disease is needed. It would also be useful to compare findings among patients with other liver diseases, such as hepatitis B and alcohol-induced liver disease. Finally, a future study should evaluate patients before and after treatment with antiviral agents, with imaging results correlated with improvement or progression of the clinical disease, histologically determined severity of the disease, and viral load.
Other limitations of our study include the selection of our patients, who all were referred to MR imaging for various reasons. Our patient population is likely to differ from patients referred specifically for evaluation of the activity level of hepatitis. An MR imaging examination that focused on that factor might have been conducted differently, perhaps reducing the field of view to cover the liver and perihepatic tissues exclusively, thereby increasing spatial resolution. Slightly less than half of the patients had liver function tests that were obtained within 2 months of the MR imaging and histologic data, preventing a more substantial analysis of liver function tests. It would also have been interesting to obtain data using modern CT techniques. A prospective study correlating MR imaging with biopsy results at the time that treatment decisions are made would be useful.
In summary, the presence, number, size, and intensity on T2-weighted images of perihepatic lymph nodes are related to the activity level of chronic hepatitis C. MR imaging may be superior to liver function tests for evaluating the activity of chronic hepatitis C and may reduce the need for repeated liver biopsies.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Yenarkarn, R. F. Thoeni, and D. Hanks Case 107: Lymphoma of the Mesentery Radiology, February 1, 2007; 242(2): 628 - 631. [Full Text] [PDF]
|
|
|
|
|
|
F. Bilaj, W. B. Hyslop, H. Rivero, Z. Firat, G. Vaidean, R. Shrestha, J. T. Woosley, and R. C. Semelka MR Imaging Findings in Autoimmune Hepatitis: Correlation with Clinical Staging Radiology, September 1, 2005; 236(3): 896 - 902. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Papakonstantinou, T. G. Maris, S. Kostaridou, V. Ladis, A. Vasiliadou, and N. C. Gourtsoyiannis Abdominal Lymphadenopathy in {beta}-Thalassemia: MRI Features and Correlation with Liver Iron Overload and Posttransfusion Chronic Hepatitis C Am. J. Roentgenol., July 1, 2005; 185(1): 219 - 224. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
