| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, University of Michigan Hospitals, 1500 E. Medical
Center Dr., Ann Arbor, MI 48109-0030.
2 Department of Internal Medicine, University of Michigan Hospitals, Ann Arbor,
MI 48109.
Received June 3, 2002;
accepted after revision October 21, 2002.
Presented at the annual meeting of the Radiological Society of North
America, Chicago, Illinois, November 28-December 3, 1999.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Fifty-two patients with suspected or known hepatic tumors underwent multiphasic contrast-enhanced MDCT using double arterial (early and late arterial) and venous phase acquisitions with fixed timing delays. All patients were administered 150 mL of IV contrast material at an injection rate of 4 mL/sec. Images were acquired at 20 sec for the early arterial phase, 35 sec for the late arterial phase, and 60 sec for the portal venous phase. Attenuation measurements of the celiac artery, portal vein, normal hepatic parenchyma, and the hepatic tumor were compared. Three reviewers independently and subjectively rated tumor conspicuity for each of the three phases. Ratings were compared using kappa statistics.
RESULTS. Late arterial phase images showed maximal celiac axis attenuation, whereas portal venous phase images revealed the highest portal vein and normal hepatic parenchymal attenuation. Maximal tumor-to-parenchyma differences for hypovascular tumors was superior in the portal venous phase, but we found no significant differences in maximal tumor-to-parenchyma differences for hypervascular tumors among the evaluated phases. On subjective analysis, interobserver agreement was moderate to very good for the three phases. All three reviewers graded both hypovascular and hypervascular tumor conspicuity as superior in either the late arterial phase or the portal venous phase in most patients. In only one patient was the early arterial phase graded as superior to the late arterial and portal venous phases (by two of the three reviewers).
CONCLUSION. When MDCT of the liver is performed using fixed timing delays, maximal vascular and hepatic parenchymal enhancement is achieved on either late arterial phase or portal venous phase imaging. In most patients, early arterial phase imaging does not improve tumor conspicuity by either quantitative or subjective analysis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
With the introduction of multidetector CT (MDCT), image acquisition of a larger volume of anatomy in the z axis is now possible in a shorter time (< 10 sec), with no sacrifice in image quality [19, 20]. This development has created even more possibilities for varying the timing of image acquisition after contrast material injection. MDCT has the potential to further improve tumor detection and characterization because even more dramatic hepatic tumor and hepatic parenchymal enhancement differences might now be achieved.
Preliminary studies have shown that MDCT using two sets of arterial phase imaging maximizes hepatic tumor detection rates [20, 21]. Foley et al. [20] found that when using a double arterial phase acquisition on MDCT, maximal tumor-to-parenchyma differences for hypervascular tumors were observed on the delayed arterial acquisition images (late arterial phase). In a more recent study of patients with hepatocellular carcinoma, Murakami et al. [21], using a double arterial phase acquisition on an MDCT scanner, showed that although most of the tumors had maximal conspicuity in the late arterial phase when compared with the early arterial phase, a combination of the two phases maximized the number of tumors detected. Using two arterial phases instead of one may be beneficial in hepatic tumor detection, but doing so adds more complexity, exposes the patient to more radiation, requires additional film and data-storage capacity, and requires more radiologist time for reviewing the additional images.
Although several methods are available to determine optimal delay times for arterial phase imaging, no conclusive data are available as to which of these is superior [22, 23, 24]. The minibolus and automated techniques can result in a more consistent and reliable enhancement of the upper abdominal vasculature, but they have some disadvantages [23]. These drawbacks include limited availability, added cost of the automated-technique software, and increase in table time needed for review of minibolus images to calculate the time delay, thereby decreasing patient throughput. In addition, the potential exists for computer malfunctions to occur, leading to mistiming of the IV contrast bolus and, thus, suboptimal studies [24]. Use of a test bolus can also potentially decrease the conspicuity of a hepatic lesion because of the liver enhancement that is present when the diagnostic scanning begins. The timing-bolus technique also has an inherent delay before scanning begins once the target threshold is reached. These reasons led us to use fixed timing delays for a variety of examinations, including hepatic imaging, at our institution. In our study, we sought to determine which of three contrast-enhanced phases (early arterial, late arterial, or portal venous) was optimal for achieving maximal enhancement of the celiac artery, portal vein, and hepatic parenchyma, as well as which phase provided maximal tumor-to-parenchyma differences when MDCT was used with fixed timing delays.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Imaging Technique
In all patients, CT was performed using an MDCT scanner (Lightspeed QX/i;
General Electric Medical Systems, Milwaukee, WI). One hour before scanning,
patients received 320 mL of dilute barium orally (Scan C: 2.1% weight by
weight barium sulfate; Lafayette Pharmaceuticals, Yorba Linda, CA) for
gastrointestinal tract opacification. A timing bolus was not used to determine
scanning delays. Via a power injector, 150 mL of Omnipaque 300 mg I/mL
(iohexol 300 mg I/mL; Nycomed, Princeton, NJ) was injected IV at a rate of 4
mL/sec. Double arterial phase scanning consisting of early and late arterial
phase image acquisition was performed during a single breath-hold using,
respectively, 20- and 35-sec scanning delays from initiation of contrast
material injection. These timing delays were explicitly chosen so that both
arterial phases could be performed within a single breath-hold that ranged
from 30–50 sec (which could be performed in all our patients). We
elected to perform these two phases during a single breath-hold to minimize
the effect of variation on lesion conspicuity and detection. The entire liver
was scanned in a cephalad-to-caudad direction using a detector collimation of
5 mm with a table speed per rotation of 15 mm/0.8 sec, a pitch of 3 in the
scanner's HQ mode, and an image thickness of 5 mm. After a brief period of
quiet breathing, portal venous phase imaging of the entire abdomen was
performed in a cephalad-to-caudad direction during a breath-hold, using a
scanning delay of 60 sec, detector collimation of 5 mm, table speed of 15
mm/0.8 sec, pitch of 3 (HQ mode), and an image thickness of 5 mm. Using these
parameters, the combined scanning duration for both arterial phase
acquisitions was 20–24 sec and for portal venous phase, 12–15 sec.
The scanning parameters used were 120 kVp and 150–250 mAs.
Image Analysis
All images were loaded and reviewed on a workstation (Windows Advantage
3.1; General Electric Medical Systems). Regions-of-interest (ROI) measurements
ranging in area between 15 and 30 mm depending on vessel size were made by one
of the authors who did not participate in the qualitative evaluation portion
of the study. Measurements were made of the celiac axis just distal to its
takeoff from the aorta and of the portal vein in the portal hilum during each
of the three phases. Hepatic parenchymal ROI measurements were obtained from
the most homogeneous regions of the liver. Three regions 320–360 mm in
area were chosen in the cranial aspect, mid portion, and caudal aspect of the
liver, taking care to avoid regions with focal fatty change, tumor
involvement, and normal vessels. Thus, nine measurements of hepatic
attenuation values were obtained in each patient for each phase. Hepatic
parenchyma attenuation for each phase was then calculated in each patient as
the mean of the nine attenuation measurements on each scan.
The same author who made the ROI measurements of the vessels and liver performed quantitative analysis of hepatic tumor using the ROI measurements obtained on images acquired during each of the three phases of contrast enhancement. These ROI measurements encompassed most of the tumors and varied in area according to tumor size, ranging between 15 and 750 mm. When multiple liver masses were present, one of the authors who was not a reviewer for the study selected the most conspicuous or, if all were equally conspicuous, the largest hepatic tumors for ROI measurements. For each tumor, the maximal tumor-to-parenchyma difference was compared for each of the three contrast-enhanced phases. Separate analyses of maximal tumor-to-parenchyma difference for hypervascular tumors (such as hepatocellular carcinoma, hypervascular metastases from neuroendocrine tumors, hepatic adenomas, and focal nodular hyperplasia) and hypovascular tumors (such as metastases from colon and pancreatic adenocarcinoma) were also carried out.
Three reviewers performed qualitative analysis of tumor conspicuity for the three phases of IV contrast enhancement. The tumors selected for this analysis were the same as those chosen for obtaining quantitative ROI measurements. Tumor conspicuity was graded on a 4-point scale (0 = not seen, 1 = poorly seen, 2 = moderately well seen, 3 = very well seen).
Statistical Analysis
Quantitative analysis.—Because more than one attenuation
measurement was obtained for each patient, repeated measures analysis of
variance was used for evaluating the effect of phase on attenuation for celiac
axis, portal vein, liver, and tumor-to-parenchyma differences
[25]. Tumors were analyzed as
two separate groups: 22 hypervascular neoplasms (hepatocellular carcinoma;
metastatic neuroendocrine tumors including carcinoid, islet cell tumor,
gastrinoma; and focal nodular hyperplasia) and 30 hypovascular neoplasms
(metastases from colon, pancreas, lung, gastrointestinal stromal, breast,
melanoma, uterus, ovary, intrahepatic cholangiocarcinoma, large cavernous
hemangiomas, and a hypovascular adenoma). The three metastases from
gastrointestinal tumors were necrotic, and the solitary case of metastasis
from melanoma was hypovascular on qualitative analysis. We therefore included
them in the hypovascular group. When the phase effect was found to be
significant, multiple comparisons were made using the Tukey-Kramer test. The
rating for phase effect on tumor conspicuity was analyzed using the general
estimating equations approach for repeated ordinal data
[26]. Statistical analysis was
performed using the SAS software (Statistical Analysis System, Cary, NC)
[25]. Power calculations were
performed for assessing the variance among the means of the three phases using
nQuery Advisor 4.0 (Statistical Solutions, Boston, MA)
[27].
Tumor conspicuity reviewer analysis.—For each of the phases of imaging, four kappa statistics were computed for interobserver variability with accompanying 95% confidence intervals (CI). For the pair-wise comparisons, weighted kappa statistics were used.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Portal Vein Enhancement
The greatest average attenuation of the portal vein was observed on the
portal venous phase. This attenuation was significantly superior to that
observed in the late arterial phase (p < 0.0001) (Figs.
1A,
1B and
1C).
|
|
|
Normal Hepatic Parenchymal Enhancement
Hepatic parenchymal attenuation was greatest in the portal venous phase and
was significantly higher (p < 0.0001) than that of each of the
other phases. Although the hepatic attenuation in the late arterial phase was
significantly greater than that in the early arterial phase, the difference
was not as great (Table 1).
|
Quantitative Analysis: Tumor-to-Parenchyma Differences
For the subgroup of hypervascular tumors, although maximal
tumor-to-parenchyma differences were greater during the late arterial and
portal venous phases than during the early arterial phase, the differences
were not statistically significant (Table
2). In comparison, in the subgroup of hypovascular tumors, maximal
tumor-to-parenchyma differences were significantly greater during the late
arterial and portal venous phases than during the early arterial phase.
However, no significant difference was found between the late arterial and
portal venous phases (Table
2).
|
The available data suggests that a univariate two-group repeated measures analysis of variance will have 99% power to detect a variance among the tumor vascularity group marginal means of 4648.434, will have 99% power to detect a variance of 6557.350 among the means of the three phases (early arterial, late arterial, and portal venous), and will have 99% power to detect an interaction between tumor vascularity groups and phases with a variance of 3220.540. These outcomes assume that the between-groups error term is 35.22; the within-groups error term is 21.93; and the measure of "sphericity" of the covariance matrix, epsilon, is 0.79 (its estimate, the Greenhouse-Geisser correction, has an expected bias of approximately g1 /[2n – 2] where g1 is –0.28) when the significance level is 0.050 and the sample size in each of the two groups is 23. Power calculations were performed using nQuery Advisor 4.0 software.
Qualitative Analysis
We used kappa statistics to determine interobserver variability for tumor
conspicuity. The level of overall interobserver agreement for all three phases
for the three reviewers ranged from 0.41 to 0.80 (moderate to good), with
interobserver agreement being best in the portal venous phase, ranging from
0.767 to 0.791 (Table 3).
|
On the basis of qualitative analysis, the overall tumor conspicuity score for all 52 hepatic tumors as determined by all three reviewers was generally graded as superior on the late arterial or portal venous phase but not on the early arterial phase images. Reviewers 1, 2, and 3 graded tumor conspicuity as being superior in the late arterial phase images for 15, 17, and 15 tumors, respectively. Portal venous phase images were graded as being superior for 12, 19, and 18 tumors. In comparison, the three reviewers graded early arterial phase images as being superior for only one, zero, and four tumors, respectively.
For 11 (50%) of 22 hypervascular tumors, late arterial phase images were thought to be superior to early arterial and portal venous phase images for tumor conspicuity by at least two of three reviewers (Figs. 2A, 2B, 2C and 3A, 3B, 3C). In seven tumors (32%), late arterial and portal venous phase images were graded as equivalent but superior to early arterial phase images by at least two reviewers, and in the remaining four tumors (18%), portal venous phase images were rated as superior. In no patient were early arterial phase images thought to be superior by any reviewer (Table 4).
|
|
|
|
|
|
|
For 12 of 30 hypovascular tumors, portal venous phase images were thought to be superior to late arterial phase and early arterial phase images by at least two reviewers (Figs. 4A, 4B and 4C). Tumor conspicuity was rated as equivalent on late arterial and portal venous phase images but superior to early arterial phase images by at least two reviewers for 10 of the 18 remaining tumors. In six others, late arterial phase images were rated as superior to both portal venous and early arterial phase images. Conspicuity was rated as equivalent on all phases in one patient. In only one instance—in a patient with metastasis in a liver with fatty changes—were early arterial phase images rated superior to late arterial and portal venous phase images (Table 4).
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MDCT provides substantial improvement in volume coverage over single-detector CT. More rapid image acquisition allows greater separation of arterial and venous phases, thus facilitating multiphasic acquisition [20]. This improvement does not sacrifice image quality: a recent study has shown that images of comparable diagnostic quality were achieved with three times the volume coverage with MDCT than with single-detector CT [19].
Several methods have been used to determine the optimal timing of arterial phase image acquisition during helical CT. These include the use of fixed timing delays, as used in this study; a timing bolus with an initial minibolus of iodinated contrast material; and semiautomated computer bolus-tracking techniques, such as SmartPrep (General Electric Medical Systems) and Care Bolus (Siemens Medical Systems, Forkheim, Germany).
Although timing bolus or bolus tracking methods may seem to offer the unique advantage of tailoring the CT technique to each patient, a number of studies have shown that in clinical practice, these techniques are not consistently superior to that of fixed timing delays. Platt et al. [22] have shown that the test injection computed time delay does not correlate well to the actual time to peak aortic enhancement in patients undergoing CT angiography. Additionally, in a study specifically evaluating ways to achieve optimal arterial phase hepatic imaging, Sandstede et al. [23] reported that optimal hepatic arterial phase imaging was obtained in only 41% of patients with the use of an automated bolustracking technique. Conversely, Macari et al. [24] found that using empiric fixed time delays resulted in adequate qualitative and quantitative enhancement of the aorta and iliac arteries. Because no consensus exists that either the bolus-tracking or bolus-tracking method is superior to the use of fixed timing delays for determining the optimal scanning delay times for achieving arterial phase imaging, we elected to use fixed timing delays for our study. We found that this technique was easier for our technologists to use and, thus, to be most applicable to our busy clinical practice.
In our investigation using MDCT, we chose to use a fixed scanning delay of 20 sec for early arterial, 35 sec for late arterial, and 60 sec for portal venous phase imaging. When we used this timing and a contrast agent injection rate of 4 mL/sec, maximal celiac axis attenuation occurred in the late arterial phase—rather than in the early arterial phase, as has been seen in studies using single-detector scanners [1, 2]. This finding can be explained by the more rapid scanning of MDCT. Given the larger volume of anatomic coverage achieved in a shorter time when our protocol is used, only approximately 80% of the total IV contrast material dose (120 of the 150 mL in our study) is injected by the end of the early arterial phase. Therefore, the remaining 20% of the total dose of IV contrast material is still being injected at the beginning of the late arterial phase. Our results also differ from those observed in the more recent studies by Foley et al. [20] and Murakami et al. [21], who also used MDCT scanners and found that when using a double arterial acquisition, maximal arterial enhancement occurred in the early arterial phase. This difference is likely explained by the fact that in those studies, a timing bolus was used to individually optimize the early arterial phase, whereas in our study, a fixed scanning delay was used.
We also found that statistically significant greater opacification of the portal vein occurred during the portal venous phase. These findings are similar to the results of Foley et al. [20].
Tublin et al. [28] have recently shown that enhancement of the normal liver is dependent on the total amount and rate of contrast material injected, as well as on the timing of the scan acquisition. These authors showed that both peak enhancement and time to peak enhancement are directly related to the rate of contrast material injection. In their study, at an injection rate of 2.5 mL/sec, peak enhancement of the liver was reached in 87 sec, in comparison to a peak enhancement at 63 sec with an injection rate of 5 mL/sec. Thus, in our study, in which a total of 150 mL of nonionic iodinated contrast material was injected at a rate of 4 mL/sec in all patients, it would be expected that using a scanning delay of 60 sec for portal venous phase image acquisition and a mean scanning duration of 14.5 sec would result in portal venous phase images showing the highest liver attenuation. This observation is similar to that reported by Foley et al. [20] who found that liver parenchymal enhancement tended to progressively increase from early arterial phase to portal venous phase images.
Our results with MDCT concerning conspicuity of hypervascular tumors also differ from those observed on earlier studies using single detector scanners, in that in our study tumor conspicuity for hypervascular tumors was qualitatively superior on the late arterial phase images rather than on the early arterial phase images [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14]. Two of three reviewers rated the late arterial phase to be superior to early arterial and portal venous phase images in 11 (50%) of 22 patients and, in another seven patients (32%), superior to early arterial phase and equivalent to portal venous phase images. In 4 (18%) of 22 patients, portal venous phase images were judged to be superior to late arterial phase images. In no instances were hypervascular tumors judged to be best seen on early arterial phase images. Our findings with respect to conspicuity of the subgroup of hypervascular liver tumors are similar to those of recent maximal tumor-to-parenchyma differences reported by Foley et al. [20]. In that study, maximal tumor-to-liver contrast difference occurred in the second pass (late arterial phase) for patients both with and without cirrhosis. A study by Murakami et al. [21] showed that although the sensitivity for detection of hepatocellular carcinoma was superior in the late arterial phase, a combination of both late and early arterial phase imaging resulted in improved detection; more tumors were detected when both early and late arterial phase images were analyzed in combination than when they were analyzed individually. We did not find this to be true in our study, which included a smaller number of hepatocellular carcinomas, but the differences may in part be explained by the use of a timing bolus to optimize early arterial phase images in the study by Murakami et al. When we used quantitative analysis, we found no significant difference in tumor-to-parenchyma differences for the three phases in the hypervascular tumor group in our study. However, inclusion of larger numbers of patients might have yielded different results.
Also in our series, for 22 of 30 hypovascular tumors, portal venous phase images were rated as superior or equivalent to late arterial phase images, but superior to early arterial phase images. In only one instance were early arterial phase images rated as superior to late arterial and portal venous phase images.
One limitation of our study is its small sample size. We also did not tailor the IV contrast dose on the basis of patient weight, but instead chose a standard dose of 150 mL for all patients. Although prior studies have shown that aortic and hepatic enhancement can be adversely influenced if patient weight is not factored into account when determining the total dose and rate of IV contrast material administered, few institutions tailor contrast dose to patient weight [22, 29].
In summary, although we found that late arterial and portal venous phase images had better MDCT results than early arterial phase images for hypervascular liver tumors, this finding was not statistically significant. For hypovascular liver tumors, late arterial and portal venous phase images had statistically significant superior maximal tumor-to-parenchyma differences compared with early arterial phase images. On subjective analysis of tumor conspicuity, late arterial phase images were also thought to be superior to early arterial phase images for most (18/22) hypervascular tumors. In 22 of 30 hypovascular tumors, portal venous phase images were judged to be superior or equivalent to late arterial phase images, but superior to early arterial phase images. Because maximal celiac arterial and portal venous opacification is obtained in the late arterial and portal venous phases, respectively, and because use of early arterial phase images does not contribute to tumor detection in most patients, we no longer routinely perform early arterial phase imaging in patients with known or suspected hepatic neoplasms. At our institution, patients referred for CT with possible or known hepatic tumors undergo MDCT using only late arterial and portal venous phase imaging. The early arterial phase images are acquired only if CT angiography is required to depict hepatic arterial anatomy before surgery. Although the arterial enhancement is superior in the late arterial phase, some venous contamination occurs, which makes the early arterial phase more suitable for CT angiography.
Acknowledgments
We thank Vanessa Brazeau for help with manuscript preparation, Robert Combs
for the photography, and the technologists of the Computed Tomography Division
for protocol development and application.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1.5 cm) malignant hepatic neoplasms.
AJR
1995;164:879
–884
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  Y. Yanaga, K. Awai, T. Nakaura, T. Namimoto, S. Oda, Y. Funama, and Y. Yamashita Optimal Contrast Dose for Depiction of Hypervascular Hepatocellular Carcinoma at Dynamic CT Using 64-MDCT Am. J. Roentgenol., April 1, 2008; 190(4): 1003 - 1009. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Nakaura, K. Awai, Y. Yanaga, Y. Nakayama, S. Oda, Y. Funama, and Y. Yamashita Detection of Early Enhancement of Hypervascular Hepatocellular Carcinoma Using Single Breath-Hold 3D Pixel Shift Dynamic Subtraction MDCT Am. J. Roentgenol., January 1, 2008; 190(1): W13 - W18. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Ronzoni, D. Artioli, R. Scardina, L. Battistig, E. Minola, S. Sironi, and A. Vanzulli Role of MDCT in the Diagnosis of Hepatocellular Carcinoma in Patients with Cirrhosis Undergoing Orthotopic Liver Transplantation Am. J. Roentgenol., October 1, 2007; 189(4): 792 - 798. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M.-J. Kim, J. Y. Choi, J. S. Lim, J. Y. Kim, J. H. Kim, Y. T. Oh, E. H. Yoo, J. J. Chung, and K. W. Kim Optimal scan window for detection of hypervascular hepatocellular carcinomas during MDCT examination. Am. J. Roentgenol., July 1, 2006; 187(1): 198 - 206. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. T. Johnson and E. K. Fishman IV Contrast Selection for MDCT: Current Thoughts and Practice Am. J. Roentgenol., February 1, 2006; 186(2): 406 - 415. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. K. Kim, C. S. Kim, G. H. Chung, Y.-M. Han, S. Y. Lee, S. B. Chon, and J. M. Lee Comparison of Gadobenate Dimeglumine-Enhanced Dynamic MRI and 16-MDCT for the Detection of Hepatocellular Carcinoma Am. J. Roentgenol., January 1, 2006; 186(1): 149 - 157. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Awai, M. Inoue, Y. Yagyu, M. Watanabe, T. Sano, S. Nin, R. Koike, Y. Nishimura, and Y. Yamashita Moderate versus High Concentration of Contrast Material for Aortic and Hepatic Enhancement and Tumor-to-Liver Contrast at Multi-Detector Row CT Radiology, December 1, 2004; 233(3): 682 - 688. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. V. Sahani and S. P. Kalva Imaging the Liver Oncologist, July 1, 2004; 9(4): 385 - 397. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
