| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Diagnostic Radiology, Severance Hospital, Yonsei University
College of Medicine, Seodaemun-ku Shinchon-dong 134, Seoul, 120-752, South
Korea.
2 Institute of Gastroenterology and Brain Korea 21 Project for Medical Science,
Severance Hospital, Yonsei University College of Medicine, Seoul, South
Korea.
3 Department of Diagnostic Radiology, Institute of Radiological Science,
Severance Hospital, Yonsei University College of Medicine, Seoul, South
Korea.
Received February 15, 2005;
accepted after revision May 25, 2005.
Address correspondence to M.-J. Kim
(kimnex{at}yumc.yonsei.ac.kr).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Biphasic arterial phase CT examinations were performed using a 16-MDCT scanner on 198 patients (159 men and 39 women; mean age, 59 years; age range, 25-82 years) with nodular HCC. All examinations were performed after administering 120-150 mL of a nonionic contrast media (370 mg I/mL) at a rate of 3-4 mL/s. The scan delay—the interval between when the bolus-tracking program detected the threshold enhancement of 100 H in the abdominal aorta and the start of the first arterial scan—was progressively lengthened by 2-second intervals, from 10 seconds in group 1 to 20 seconds in group 6. The second arterial phase scan was started 6 seconds after the end of the early scan. A tube collimation of 1.5 mm, a table feed of 18 mm per rotation, an image thickness of 3 mm, and 3-mm increments were used. The duration of each phase scan was 4.5-8.8 seconds. Tumor-to-liver attenuation difference (TLAD) at the first (TLAD1) and second (TLAD2) arterial phase images were compared lesion by lesion. Four observers assigned subjective ratings of visual conspicuity and individual preferences for each phase in each group.
RESULTS. The mean threshold time (100 H) was 18.4 ± 3.1 seconds, and 97% of patients were within the range of 13-24 seconds. The mean TLAD1 of groups 3 to 6 and the mean TLAD2 of groups 1 to 5 were all comparable; they were also all significantly (p < 0.005) higher than the mean TLAD1 of groups 1 and 2 and the mean TLAD2 of group 6. In groups 1 and 2, the mean TLAD2 was significantly higher than the mean TLAD1 (p < 0.001); in groups 5 and 6, the mean TLAD1 was significantly higher than the mean TLAD2 (p < 0.001). In groups 3 and 4, the mean TLAD1 and TLAD2 were similar. The visual conspicuity and individual preferences were higher for the first-phase image in groups 3 to 6 and the second-phase image in groups 1 and 2.
CONCLUSION. The optimal scan window for arterial phase images in the detection of HCC seems to be approximately 14-30 seconds from the 100-H threshold.
Keywords: abdomen • contrast media • CT technique • hepatocellular carcinoma • liver • MDCT
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
If multiple HCC nodules were observed in a patient, up to five lesions from
different parts of the liver (right upper, right middle, right lower, left
upper, and left lower) were selected to eliminate cluster bias. Ultimately,
323 HCC lesions (size range, 0.4-4.9 cm; mean, 1.7 cm) were included. HCC
diagnosis was based on results from surgical resection (n = 30), on
results of percutaneous biopsy (n = 26), on elevation of tumor
markers (
-fetoprotein, protein-induced vitamin-K absence, or antagonist
II) with typical CT or MRI findings in patients with liver cirrhosis or
chronic liver disease (n = 82), or on typical CT or MRI findings
combined with typical angiographic appearance (that of a hypervascular tumor
with neovascularity) and notable iodized oil uptake in follow-up CT scans
(n = 60). Liver cirrhosis was evident in 173 patients and was a
result of hepatitis B (n = 128), hepatitis C (n = 40),
Budd-Chiari syndrome (n = 2), or alcoholism (n = 3). The
remaining 25 patients had hepatitis B (n = 21) or C (n = 4)
without cirrhosis.
CT Technique
All CT examinations were performed after the injection of 120-150 mL of
iopamidol (Iopamiro, Bracco) at a concentration of 370 mg I/mL and at a rate
of 3-4 mL/s using a power injector (EnVision CT, Medrad)
(Fig. 1). A 2 mL/kg dose of
contrast media was administered to patients whose body weight was 60-75 kg.
The total dose was fixed at 150 mL for patients weighing more than 75 kg and
at 120 mL for patients weighing less than 60 kg. The injection rate was
determined by the accessibility of the venous route, but a rate of 4 mL/s was
used in most patients. The scan delay was determined using bolus-tracking
software. For bolus tracking, a series of nonhelical sequential images were
obtained 10 seconds after the administration of contrast agent. These images
were acquired with a scanning time of 0.5 seconds (360°) using a low-dose
radiation technique (120 kV, 20 mA) and a cycle time of 1.25 seconds. In each
image, an approximately 1-cm2 area of a circular region of interest
(ROI) was placed in the abdominal aorta at the level of the celiac artery; the
attenuation value (Hounsfield unit [H]) was then measured
[15]. The time at which the
bolus-tracking program detected the threshold enhancement of 100 H in the
abdominal aorta was defined as the threshold point, and the time interval
between the start of the contrast material injection and the threshold point
was defined as the threshold time. The scan delay time was defined as the time
between the threshold point and the start of the first scan. The interscan
delay time was defined as the time interval between the end of the first scan
and the start of the second scan; this was fixed at 6 seconds for every
examination because it was the lowest limit the system allowed for two scans
to be performed in the same direction. The study was conducted over 30 weeks,
divided into six 5-week intervals. During the first 5 weeks of this study, the
arterial scan was performed by starting 10 seconds after the 100-H threshold
at the aorta; the scan delay was lengthened by 2 seconds by 5-week intervals
until it reached 20 seconds. The patients were consecutively allocated into
one of the six groups: All patients who underwent the examinations during the
first five weeks of the study period were assigned to group 1; the other
groups were assigned at 5-week intervals. In each group, the scan delay varied
as follows: 10 seconds in group 1 (24 men and five women; mean age, 57.7
years; 54 HCCs), 12 seconds in group 2 (22 men and nine women; mean age, 58.8
years; 52 HCCs), 14 seconds in group 3 (30 men and five women; mean age, 59.3
years; 52 HCCs), 16 seconds in group 4 (34 men and five women; mean age, 59.5
years; 55 HCCs), 18 seconds in group 5 (29 men and eight women; mean age, 57.8
years; 58 HCCs), and 20 seconds in group 6 (20 men and seven women; mean age,
59.9 years; 52 HCCs). We used a scan delay of 10 seconds or more to obtain the
liver images after the end of the pure arterial phase.
Biphasic arterial and delayed phase scans were performed in all patients during a single breath-hold period. For all scans, a craniocaudal scan was performed with the following parameters: 0.5-second rotation, 120 kV, 300 mAs, tube collimation of 1.5 mm, table feed of 18 mm per rotation, image thickness of 3 mm, and a 3-mm increment reconstruction. The duration of each phase scan was 4.5-8.8 seconds (mean, 6.4 seconds), depending on the selected scan ranges. Therefore, the time interval between the same scan levels was 10.5-14.8 seconds (mean, 12.7 seconds). A delayed scan was performed 180 seconds after contrast medium injection. This scan was used in the diagnosis of HCC, but not in the data analysis. An unenhanced scan was also performed and was used to calculate the degree of hepatic enhancement in each phase of the scan. Both unenhanced and delayed scans were also performed with the same parameters as the biphasic scan. The scanning time of a lesion was defined as the time of the slice depicting the craniocaudal center of the lesion of interest. Meanwhile, the lesion delay time was defined as the time between the threshold point and the scanning time of a lesion. In all lesions, the scanning time of a lesion was recorded, whereas the lesion delay time was calculated using the following equation: image delay time = scan delay time + (scanning time of a lesion - scanning time of the first slice of the first arterial phase). The specific time annotated on each image by the scan system was used to calculate the scan delay and lesion delay.
Image Analysis
A circular ROI was drawn on the tumor and hepatic parenchyma by the primary
investigator to measure attenuation values. The area of the ROI in the tumor
was set to measure the homogeneous area of the lesion, which was maintained at
a size of at least 50 mm2. For lesions with necrotic components,
care was taken to measure the area only in the solid portion of the lesion. If
a lesion was poorly depicted on either of the two arterial phase images, the
location where the lesion was seen on other phase images was used. While being
sure not to include vessels, the ROI of the liver was drawn on the hepatic
parenchyma near the tumor and was maintained at a size of at least 100
mm2. The size of the ROIs of the tumor and liver were kept the same
in their respective phase images. The tumor-to-liver attenuation difference
(TLAD) on each scan was measured to obtain a lesion-by-lesion comparison of
the differences in the contrast of the HCC between the first and second
arterial scans. TLAD was defined as Hle - Hli, where
Hle is the attenuation (H) of the lesion and Hli is the
attenuation of the surrounding liver. TLAD1 was the value measured from the
first arterial phase image, and TLAD2 was the value measured from the second
arterial phase image. When TLAD1 was greater than TLAD2, the lesion was
considered to have higher contrast on the first arterial phase image, whereas
the lesion was considered to have higher contrast on the second arterial phase
image when TLAD2 was greater than TLAD1. The mean values and SDs of TLAD1 and
TLAD2 were calculated for each group (Fig.
1). The number of lesions showing a higher contrast on the first
or second arterial phase scans was counted in each group. The ROI of the liver
was also measured on the unenhanced CT scan, and the degree of the liver
enhancement (Lipost - Lipre, where Lipost is
the liver attenuation from the contrast-enhanced image and Lipre is
the liver attenuation from the unenhanced image) was calculated.
To perform the qualitative analysis, four observers retrospectively and randomly reviewed all of the cases without any knowledge of the scan delay times. The radiologists had 4-12 years of experience in abdominal imaging. The first and second arterial phase images were compared side by side at a PACS workstation (PathSpeed version 8.1 or Centricity version 1.0, GE Healthcare). The images were reviewed at the soft-tissue setting of both a moderate window width and window level (450 H and 60 H, respectively) and a narrow window width and window level (300 H and 50 H, respectively); the interpreters were then free to adjust the settings according to their preferences. The observers assigned the following subjective ratings of visual conspicuity: 1, the lesion was more conspicuous on the first arterial phase image; 2, the lesion showed a similar conspicuity on the two phase images; or 3, the lesion was more conspicuous on the second arterial phase image. The observers also recorded their individual preferences according to which phase image they would select if only one of the two phase images could be obtained. They also recorded the number of the lesions that were seen on each phase image. The final number of these lesions was determined in conference by the consensus of the four interpreters.
Statistical Analysis
Among the six groups, the sex of the patients was statistically compared
using the Kruskal-Wallis test. A one-way analysis of variance was performed to
compare patients' ages, lesion sizes, and threshold times. Pearson's
correlation coefficient was calculated to determine the relationship between
threshold time and age. An independent sample Student's t test was
performed to compare the difference in threshold time between men and women.
The time interval between the thresholds of 50 H and 100 H in the abdominal
aorta was calculated. To determine the effect of previous transarterial
embolization, an independent sample Student's t test was performed to
compare the differences in threshold time, tumor and liver attenuation in each
phase scan, and TLAD values in each phase between those who had previously
undergone transarterial embolization and those who had not.
A one-way analysis of variance was performed to compare TLAD1 and TLAD2 values among the six groups. Post-hoc multiple comparisons were performed using the Bonferroni or Tamhane methods according to the homogeneity of the variance test results. In each group, TLAD1 and TLAD2 values were compared using a paired Student's t test. The McNemar test was performed to compare the detection rates between the first and the second arterial phases.
Pearson's correlation coefficients were calculated to compare the attenuation values of the liver and tumor and the TLAD at the early- and late-phase images (TLAD1 vs TLAD2). Pearson's correlation coefficients were also calculated to determine the correlation between the lesion delay time and the attenuation of the liver, the tumor, and the TLAD at each phase. A p value of less than 0.05 was considered significant in all tests.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
The attenuation values for the liver gradually increased from groups 1 to 6 in both the first and second arterial phase images (Figs. 2A, 2B, and 2C). The TLADs showed a wide range in each group for both the first and second arterial phase images. The mean TLAD1 of group 1 was significantly lower than those in groups 3 to 6 (p < 0.05), and that of group 2 was significantly lower than those of groups 4 and 5 (p < 0.05). The mean TLAD2 was relatively constant from groups 1 to 5, but that of group 6 was significantly lower than that of groups 1 to 5 (p < 0.05). The mean TLAD1 of groups 3 to 6 and the mean TLAD2 of groups 1 to 5 were comparable.
|
|
|
|
|
|
|
|
|
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Using 4-MDCT, several authors have studied different settings for double arterial phase CT [7-11]. Because of hardware limitations in the 4-MDCT scanner, one of the early or late arterial phase CT images was deemed to be too early or too late for optimally depicting hypervascular HCCs in studies that used double arterial CT. Using a 16-MDCT scanner, the scanning time of the entire liver could be decreased to 4-6 seconds. Therefore, optimizing the scan window has become even more critical.
The results of our study showed that the first arterial scans in groups 3 to 6 (scan delay, 14-20 seconds) and the second scans in groups 1 to 5 (scan delay, 23-30 seconds) provided comparable TLAD values, subjective lesion conspicuity, and lesion detection rates that were favorable compared with the earlier or later scans. From these results, the optimal scan delay for arterial phase images seems to be approximately 14 to 30 seconds from the 100-H threshold. Because the individual variation of the 100-H threshold times ranged from 10 to 29 seconds in our study, the optimal total scan delay for patients with a 10-second 100-H threshold would be 24-40 seconds, whereas for patients with a 29-second 100-H threshold the optimal total scan delay would be 41-59 seconds. Therefore, a fixed scan window that would meet the optimal scan range for all patients cannot be determined. However, because the 100-H threshold time of 97% patients is within 13-24 seconds, a scan window that is 38 (24 + 14) to 43 (13 + 30) seconds from the start of contrast agent injection may be within the optimal range of scan delay. This temporal window for hypervascular tumor detection is comparable to those suggested by Lee et al. [16] and Frederick et al. [17] using a single-detector CT. Although our data suggest that a fixed scan delay may be justified in many patients, using a bolus-tracking technique should be recommended in patients with cardiovascular problems. Nonetheless, there is still a possibility that the use of a bolus-tracking technique allows more accurate planning of the scan. At this time, we cannot be sure whether the two different techniques—using a bolus-tracking or using a fixed scan delay—shows a significant difference in detection sensitivity and lesion conspicuity. Therefore, further studies comparing these two techniques are warranted.
There are several differences in study design and results between our study and the study of Murakami et al. [18], in which middle arterial scans obtained with a fixed 30-second delay showed higher sensitivity than 20-second or 40-second delay scans after administration of 100 mL of 300 mg I/mL nonionic contrast medium at 4 mL/s. Compared with our study, they used a fixed scan delay and a smaller amount of contrast medium. Also, the injection of the contrast agent in their study was finished within 25 seconds. In our study, the contrast agent injection period was 30-37.5 seconds because 120-150 mL of contrast medium was administered. Although there was a significant difference in HCC detection sensitivity between groups in the Murakami et al. study, the detection rates were not remarkably different between the several groups in our study. In our study, the higher dose of contrast agent and the optimization of scan delay using the bolus-tracking method might have improved overall HCC detection sensitivity and decreased its difference between the groups.
Our data suggest that the first-phase images in groups 3 and 4 showed hypervascular HCCs at the upslope of TLAD, whereas the second arterial images were at the downslope of TLAD after their peak values. In these groups, the TLAD values were generally higher at the second arterial phase image, but the subjective ratings of the lesion conspicuity and the interpreter preference were generally higher for the first arterial phase image. This may be explained by the fact that the liver and the healthy vasculature are typically less enhanced in the first-phase image; this is favored by interpreters in the visual perception of the true lesions. However, because the tumor was not yet fully enhanced, the TLAD values were not different between the early and the late scans. With a scan delay of 14-20 seconds, a tumor may be less enhanced than its peak values, but the enhancement of background liver or venous structure can be minimized. Using this scan delay time, liver enhancement is usually within the range of 10-20 H, or 20-40% when compared with the unenhanced CT. Meanwhile, a longer scan delay of 20-30 seconds may provide higher enhancement of the tumor, but also higher enhancement of the liver and the portal veins.
Sandstede et al. [15] suggested that a delay of 10 seconds after the 75-H threshold was the optimal scan delay for the arterial phase after the injection of 120 mL of contrast material at 3 mL/s. In their study, the optimal arterial phase was determined by calculating the value at which 20-30% of the maximum hepatic enhancement occurred. Our data suggest that liver enhancement of 20-60% (relative to the unenhanced images) may be acceptable for the detection of hypervascular tumors if one can achieve tumor enhancement higher than hepatic enhancement. Some authors have suggested 20 H of hepatic enhancement to be the end of the arterial phase [17, 19]. In our study, the first arterial phase images in groups 4 through 6 were comparable to the range at which the liver enhancement was between 10 and 20 H. Furthermore, the second arterial phase image from groups 1 through 4 showed hepatic enhancement of 20-30 H.
The optimal temporal window suggested in this study may not be long enough for a single-detector CT unit to scan the entire liver. Therefore, by using a total scan delay of 20-30 seconds, which is recommended for the depiction of hypervascular tumors in some reports [20, 21], the upper slice of the arterial phase image may be too early or the lower slice of the images may be too late. Using a 4-MDCT unit, it is possible to obtain an entire image of the liver within this window. If double arterial phase examinations are performed using a 4-MDCT unit, the early arterial phase images may be too early or the late arterial phase images may be too late [6, 22, 23]. Therefore, in our opinion, double arterial CT examination may not be warranted when we can optimize the scan delay using a bolus-tracking method. By shortening the scanning time with a 16-MDCT unit, double arterial phase images can be obtained with a similar level of lesion conspicuity.
Our data do not suggest that using a double arterial phase examination significantly increases the sensitivity of MDCT for the detection of hypervascular HCCs. Our data showed that tumors strongly enhanced in the first arterial phase image also tended to remain strongly enhanced in the second arterial phase image. This suggests that lesion conspicuity may depend more on the intrinsic characteristics of the tumor itself (such as vascularity or histologic type) than on the optimal scan delay. Therefore, even if dual arterial phase images are obtained, a relatively hypovascular lesion may still be difficult to see. As shown in the comparison of detection sensitivities between the two arterial phase images, nearly all lesions that were seen in one of the arterial phase images were also seen in the other arterial phase image. Currently, hepatic CT examinations in our institute for the evaluation of HCC include a single late arterial phase image (20 seconds from the 100-H threshold), portal or hepatic venous phase images (70-second delay from the start of contrast material injection), and delayed equilibrium phase images (180-second delay).
One of the limitations of this study is that most of the lesions were not pathologically confirmed. This study used strict criteria to exclude pseudotumors, such as an arterioportal shunt or a transient attenuation difference, based on information from a combination of findings (such as delayed phase CT images, MRI or angiography, prior and follow-up CT, and tumor markers). Although the inclusion of non-HCC hypervascular lesions, such as hypervascular dysplastic nodules, or the exclusion of true HCC lesions that might be not detected may have been possible, we believe that this would not greatly alter our results because these were based on the comparison of the biphasic scans for a large number of clinically or pathologically confirmed lesions.
To obtain both of the two phase images in the late arterial or portal inflow phase, we began the first scan at least 10 seconds after the 100-H threshold point in our study. We used at least a 10-second delay because of reports suggesting that the mean duration of the pure arterial phase was 8-12 seconds [14, 19, 24, 25]. We used a 100-H threshold in this study, whereas some authors used 50 or 75 H to define the aortic arrival of contrast material [9, 15]. Our data showed that the difference in the threshold time between 50 H and 100 H was less than 2 seconds. Therefore, we may adjust the scan delay according to this result. From the results of our study, we recommend using 100 H as a threshold point and starting the arterial scan 20 seconds after the threshold. Because the mean and median values of the 100-H threshold fall at 18 seconds, this protocol results in approximately a 38-second delay in most cases. Recently, we launched a study comparing this technique with a fixed scan delay of 38 seconds.
Approximately half of our study population had a history of TACE for an HCC lesion in another part of the liver. However, our results showed no significant difference in the threshold time and the tumor-to-lesion contrast between the TACE and non-TACE groups. Furthermore, we excluded HCC lesions in the same segment of the previous TACE lesions. Therefore, we believe that the inclusion of previous TACE patients would not affect the results of this study.
It is possible that a different injection rate or iodine concentration may have affected the threshold time or the optimal scan delay. Therefore, further studies assessing the effect of iodine concentration or injection rate will be necessary to define the optimal scan delay for each concentration and injection protocol more precisely. We have launched a study that compares lesion conspicuity and sensitivity according to the dose and injection rate using a scan delay determined in this study.
In conclusion, our results suggest that the optimal scan window for arterial phase images in the detection of HCC seems to be approximately 14 to 30 seconds from the 100-H threshold. For the margin of safety, we recommend a 20-second delay from the 100-H threshold as optimal scan timing for 16-MDCT. However, it should be adjusted according to injection protocol and the available scan system.
|
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:
|
|
 |
|
  X. Ma, A. E. Samir, N.-S. Holalkere, and D. V. Sahani Optimal Arterial Phase Imaging for Detection of Hypervascular Hepatocellular Carcinoma Determined by Continuous Image Capture on 16-MDCT Am. J. Roentgenol., September 1, 2008; 191(3): 772 - 777. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
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]
|
|
|
|
 |
|
 L. L. Chu, B. N. Joe, A. C. A. Westphalen, E. M. Webb, F. V. Coakley, and B. M. Yeh Patient-specific Time to Peak Abdominal Organ Enhancement Varies with Time to Peak Aortic Enhancement at MR Imaging Radiology, December 1, 2007; 245(3): 779 - 787. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
