| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gastrointestinal Imaging |
1 From the Department of Diagnostic Radiology, Graduate School of Medical Sciences, Kumamoto University, 1-1-2 Honjyo, Kumamoto 860-8556, Japan. Received January 13, 2006; revision requested March 10; revision received April 25; accepted May 31; final version accepted August 1. Address correspondence to K.A.
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
Materials and Methods: This study received institutional review board approval; informed consent was obtained. The study included 192 patients (123 men, 69 women; mean age, 67.6 years) with known or suspected HCC who underwent dynamic CT, including HAP scanning; CT depicted 111 hypervascular HCCs in 72 patients. Scanning was performed with a 40-detector CT scanner, and bolus tracking was used to time the start of HAP imaging. Patients were randomly assigned to five protocols; HAP scanning was started at a specified interval after trigger threshold was reached: 9 seconds (protocol A), 12 seconds (protocol B), 15 seconds (protocol C), 18 seconds (protocol D), or 21 seconds (protocol E). Trigger threshold level was set at 100 HU above aortic baseline CT number. Enhancement values in the aorta and the tumor-liver contrast (TLC) were measured. Dunnett multiple comparisons were performed to compare enhancement values among the five protocols.
Results: Mean scanning time for the whole liver was 2.1 seconds. Mean enhancement value of the aorta in protocols A, B, C, D, and E were 284.3 HU ± 54.7, 293.8 HU ± 51.0, 308.7 HU ± 55.9, 291.5 HU ± 42.2, and 235.5 HU ± 51.2, respectively. Aortic enhancement was significantly lower in protocol E than in protocol A (P < .01); there was no significant difference between protocols A and B, A and C, and A and D. Mean TLCs in protocols A, B, C, D, and E were 23.4 HU ± 7.6, 35.5 HU ± 14.0, 36.2 HU ± 6.8, 47.2 HU ± 19.2, and 35.1 HU ± 15.8, respectively. A significant difference was found only between protocols A and D (P < .01).
Conclusion: Peak TLC during the HAP occurred 18 seconds after triggering.
© RSNA, 2007
Most hepatocellular carcinomas (HCCs) are hypervascular and are demonstrated as hyperattenuated lesions during the hepatic arterial phase (HAP) of hepatic dynamic computed tomography (CT) (1–7). In general, scans obtained with a smaller helical pitch provide better contrast resolution and fewer image artifacts than those obtained with a larger helical pitch (8). When a four-detector CT scanner with detector collimation of 2.5 mm, helical pitch of 0.75, and tube rotation time of 0.5 second is used to scan the liver, the scan time for the entire liver is relatively long, 10–14 seconds. Accordingly, when scanning is performed in a cephalocaudal direction with a four-detector scanner, it may be timed too early to depict lesions in the cranial portion of the liver and too late to depict those in the caudal portion.
With a 40- or 64-detector CT scanner, however, almost uniform enhancement can be achieved throughout the liver, because these scanners can theoretically scan the entire liver in less than 3 seconds with high spatial resolution. Therefore, when one is using a 40- or 64-detector scanner, it is important to choose a scan time that is adequate for depicting lesions throughout the liver. Timing HAP imaging in each patient to enable the detection of hypervascular HCCs can be difficult without automatic computer-assisted bolus tracking (9–13) or a test bolus (14,15). Moreover, the most appropriate scan time during HAP is unknown for the 40- or 64-detector scanners used for hepatic dynamic CT. Thus, the purpose of our study was to evaluate prospectively the use of bolus tracking to time HAP imaging of hypervascular HCCs with a 40-detector CT scanner.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
Patients and Tumors
Between December 2004 and September 2005, we enrolled 202 patients who met our inclusion criteria: (a) diagnosis of type B, C, or alcoholic hepatitis; (b) confirmed HCC untreated during the 3 months preceding CT examination or suspicion of space-occupying hepatic lesions based on a sonographic study or elevated levels of tumor markers (
-fetoprotein or protein induced by vitamin K absence or antagonist-II); and (c) absence of both renal failure (serum creatinine, <1.5 mg/dL) and contraindication to iodinated contrast material.
Ten of the 202 patients were excluded: one patient because scan times deviated from the scanning protocol, six because of tumor thrombi in the central portal or hepatic veins, and three because numerous tumors involving the entire liver may have changed the hepatic hemodynamics. Thus, the final study population consisted of 192 patients, 123 men and 69 women aged 24–84 years (mean age, 67.6 years). For men, the age range was 24–84 years (mean, 67.0 years); for women, 43–81 years (mean, 68.7). There was no significant age difference between the male and female patients (P = .19, Student t test).
As the mean CT number of the normal liver parenchyma is about 60 HU on unenhanced images and about 80 HU during the HAP (16), we defined as hypervascular those tumors whose CT number during the HAP was 25 HU greater than that on unenhanced scans. Of the 192 patients, 72 (44 men, 28 women; age range, 46–84 years; mean age, 68.7 years) had solitary or multiple hypervascular HCC nodules (n = 111). Mean body weight for all patients was 58.9 kg ± 10.0 (standard deviation [SD]) (range, 34–80 kg).
The definitive diagnosis of hypervascular HCC was based on the following findings: histopathologic evidence after hepatic surgery, two patients; findings at needle biopsy, four patients; hypervascularity at CT arteriography or hypoattenuation at CT portography, 40 patients; or substantially increased levels of -fetoprotein or protein induced by vitamin K absence or antagonist-II with follow-up CT demonstrating no change or an increase in tumor size within 3 months, 26 patients. In our institution, CT arteriography or CT portography is performed in all patients with HCCs before the initial treatment as a routine clinical practice. Among the 40 patients who were diagnosed at CT arteriography or CT portography, small HCCs (diameter, 0.6–1.5 cm) were confirmed at follow-up CT, demonstrating an increase in tumor size within 3 months.
CT Scanning and Contrast Material Infusion Protocols
All patients were scanned with a 40-detector CT scanner (Brilliance-40; Philips Medical Systems, Cleveland, Ohio) at the following settings: rotation time, 0.5 second; beam collimation, 32 x 1.25 mm; section thickness and intersection gap, 5.0 mm; helical pitch (beam pitch), 0.781; table movement, 62.5 mm; scan field of view, 40 cm; voltage, 120 kV; and tube current, 250–300 mAs. Image reconstruction was performed in a 25–35-cm display field of view, depending on the patient's physique.
All helical studies were started at the top of the liver in a cephalocaudal direction, and unenhanced and three-phase contrast material–enhanced helical scans of the entire liver were obtained. Patients were instructed to hold their breath with tidal inspiration during scanning.
Three-phase contrast-enhanced CT scanning of the liver was performed during the hepatic arterial, portal venous, and equilibrium phases. An automatic bolus-tracking program (Bolus Pro Ultra; Philips Medical Systems) was used to time the start of HAP scanning after contrast material injection. The CT number was monitored by two radiology technologists (who had 10 and 20 years of experience with abdominal CT) at the L1 vertebral body level; the region-of-interest (ROI) cursor (0.8–2.0 cm2) was placed in the abdominal aorta. Real-time, low-dose (120 kVp, 15 mAs) serial monitoring studies began 8 seconds after the start of the contrast material injection. The trigger threshold level was set at an increase of 100 HU over the aortic baseline CT number.
Patients were randomly assigned to five scan protocols; HAP scanning was started at a specified interval after the trigger threshold was reached: 9 seconds (protocol A, n = 38), 12 seconds (protocol B, n = 39), 15 seconds (protocol C, n = 39), 18 seconds (protocol D, n = 38), or 21 seconds (protocol E, n = 38). During the real-time, low-dose serial monitoring studies, the patients were instructed to take shallow, regular breaths. Among the five groups there were no significant differences in age or weight (P = .712, and .349, respectively, by using one-way analysis of variance), sex distribution (P = .299 by using the 
2 test), the number of patients with HCCs, or the size of HCC nodules (Table 1).
|
Quantitative Analysis
In all patients, we recorded the mean time for reaching the trigger threshold, and we calculated the scanning time for the entire liver. One radiologist (S.S.), who had 4 years of experience with liver CT and was blinded to the protocol used, measured the mean attenuation values of the abdominal aorta, hepatic parenchyma, and portal vein with a circular ROI cursor. An attempt was made to maintain a constant ROI area of approximately 1 cm2; the range of ROI areas was 0.5–1.0 cm2. Aortic attenuation values were determined on three consecutive images at the level of the main portal vein during unenhanced and HAP scanning. The measured attenuation values obtained for each phase were averaged. Contrast enhancement in the abdominal aorta during the HAP was calculated as the absolute difference in aortic attenuation (in Hounsfield units) between unenhanced and HAP scans.
Hepatic attenuation was measured in three separate areas (the left lobe of the liver and anterior and posterior segments of the right lobe) on images obtained at the level of the main portal vein during unenhanced and HAP studies. The attenuation values for each phase were averaged. An attempt was made to maintain a constant ROI area of approximately 2 cm2; the range of ROI areas was 0.8–2.0 cm2. Visible blood vessels, bile ducts, and artifacts were carefully excluded from ROI measurements in the hepatic parenchyma. Contrast enhancement in the parenchyma was calculated as the absolute difference in the attenuation value of the liver (in Hounsfield units) between unenhanced and HAP scans.
We also measured attenuation values in the main portal vein. An attempt was made to maintain a constant ROI area of approximately 0.5 cm2; the range of ROI areas was 0.3–0.5 cm2. Contrast enhancement in the main portal vein during the HAP was assessed as described for the hepatic parenchyma.
The conspicuity of a hepatic tumor can be expressed by the attenuation difference between it and the hepatic parenchyma, the so-called tumor-liver contrast (TLC) (20). We defined the TLC as the result obtained by subtracting the parenchymal attenuation value from the attenuation value for the hepatic tumor. In the 72 patients with hypervascular HCC, the same radiologist (S.S.) also determined the TLC for each phase of contrast-enhanced scanning. We also excluded tumors larger than 5.0 cm because they may have altered hepatic hemodynamics.
Tumor attenuation was assessed in the most enhanced portion of the tumor. An attempt was made to maintain an ROI area of approximately 0.5 cm2; the range of ROI areas was 0.3–0.5 cm2. To obtain the parenchymal attenuation value used for TLC calculations, we measured the normal hepatic parenchyma at least 1 cm from the edge of the tumor to nullify the risk of encountering fibrosis. An attempt was made to maintain a constant ROI area of approximately 2 cm2; the range of ROI areas was 0.8–2.0 cm2. In patients with fewer than three tumors, we calculated the mean TLC for all tumors; in those with three or more tumors, we used the averaged TLC of the three largest tumors.
To investigate the homogeneity of hepatic parenchymal enhancement during HAP, we measured the attenuation gradient between the cranial and caudal portion of the liver. We defined the attenuation gradient of the hepatic parenchyma as the value obtained by subtracting the parenchymal attenuation value of the cranial portion of the liver from that of the caudal portion of the liver. Parenchymal attenuation values in the cranial portion were measured in the right lobe of the liver on the three consecutive sections of the most cranial portion of the liver, and these values were averaged. Parenchymal attenuation values in the caudal portion were measured in the right lobe of the liver on the three consecutive sections of the most caudal portion of the liver, and these values were also averaged. An attempt was made to maintain a constant ROI area of approximately 2 cm2. Visible blood vessels, bile ducts, hepatic tumors, and artifacts were carefully excluded from ROI measurements in the hepatic parenchyma.
Qualitative Evaluation
To investigate the relationship between subjectively determined tumor conspicuity and TLC during the HAP, two radiologists (K.A. and Y.N.) who had 20 and 10 years of experience with liver CT, respectively, performed independent visual assessments of the conspicuity of the hypervascular HCCs. They used a three-point scale in which grade 1 indicated poor conspicuity; grade 2, fair conspicuity; and grade 3, excellent conspicuity. After their independent evaluation, they assigned a grade by consensus. In patients with two or more tumors, they assessed the conspicuity of the largest tumor.
All images were reviewed in random order on a cathode-ray tube monitor with a spatial resolution of 1600 x 1200 (RadiForce R22; Nanao, Ishikawa, Japan) by means of our picture archiving and communications system (Image VINS Pro, version 3.01; Yokogawa Electric, Tokyo, Japan). The preset window level (50 HU) and window width (300 HU) could be changed at will.
Statistical Analysis
The significance difference between protocols in the mean time required to reach the trigger threshold, the mean time required to scan the liver, and the attenuation gradients of the liver parenchyma were tested with one-way analysis of variance. Contrast enhancement values in the aorta, hepatic parenchyma, and portal vein are reported as means ± SDs, as are the TLC values and the attenuation gradients of the hepatic parenchyma.
The Dunnett method was used to perform multiple comparisons among the five scan protocols for aortic, hepatic, and portal vein enhancement values and TLC values; values obtained with protocol A served as the controls. To assess the interobserver variability in the visual analysis of tumor conspicuity, we applied the 
test of concordance to measure the degree of agreement between the two radiologists.
We used t statistics to estimate the 95% confidence interval (CI) of the population mean for TLC during the HAP for each visual grade. P values of less than .05 were considered to indicate statistically significant differences. Statistical analysis was performed with a statistical software package (SPSS, version 11.0; SPSS, Chicago, Ill).
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
|
|
|
|
|
= 0.65). The mean TLCs were 18.0 HU ± 5.0 (95% CI: 15.5 HU, 20.4 HU), 30.3 HU ± 4.6 (28.3 HU, 32.3 HU), and 50.3 HU ± 16.0 HU (44.3 HU, 56.3 HU) for grades 1, 2, and 3, respectively (Table 4). Representative cases for each protocol are shown in Figure 6.
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
In the study by Kim et al (16) of single-level serial hepatic CT, the mean increase in parenchymal attenuation in the liver was about 1.0 HU/sec in the period 0–50 seconds after the start of contrast material injection protocols, with an injection rate of 3.0–5.0 mL/sec. Thus, the shorter scanning time for the liver may achieve more homogenous enhancement throughout the liver. Accordingly, with rapid scanning performed by using a 40-detector CT scanner, the ability to depict the lesions is almost constant throughout the liver. However, stricter scan timing is required with a 40-detector CT scanner than with a four- or 16-detector scanner.
In our visual evaluation of tumor conspicuity, the lower limit of the 95% CI for TLC was 44.3 HU for grade 3. Given these results, we regarded 44.3 HU as the lower limit for an excellent depiction of hypervascular HCCs, and 28.3 HU as the lower limit for a fairly good depiction. Only in protocol D did the mean TLC value exceed 44.3 HU; in all other protocols it was lower than 40 HU. Therefore, we concluded that the scan timing in protocol D—that is, 18 seconds after triggering—was optimal for the depiction of hypervascular HCCs.
In our study, there was a 3-second delay between aortic peak enhancement and the peak TLC. Theoretically, for a given aortic arrival time, the time from aortic arrival to aortic peak enhancement is consistent with the duration of the injection (21–23). Therefore, it is possible to predict the peak aortic enhancement time by measuring the aortic arrival time of a test bolus. The injection rate does not affect the time required for the material to travel from the abdominal aorta to the hepatic artery, because the lungs are located between the vein receiving the contrast material injection and the abdominal aorta from which the hepatic artery originates. Therefore, when the delay between the aortic peak enhancement and the peak TLC is known, it is theoretically possible to determine the optimal scan time for hypervascular HCCs. Given our results, we concluded that the optimal time to scan these tumors is 3 seconds after the aortic peak enhancement time. Peak TLC lags several seconds behind peak aortic enhancement because of differences in arterial vascularity between the tumor and the liver. Contrast material diffusion from the intravascular space into the interstitium may also contribute to the time delay. The TLC in hepatic CT reflects the difference between contrast material accumulation in the tumor, both intravascular and interstitial, and contrast material accumulation in the hepatic sinusoids and the normal hepatic interstitium.
Our study had several possible limitations. First, we set the trigger threshold for starting the scanning at the point where we obtained an increase of 100 HU over the baseline aortic CT number; the mean time required to reach the trigger threshold was 19.1 seconds ± 3.4. That interval may vary among injection protocols with different injection rates, contrast material doses, and trigger threshold levels, and aortic peak enhancement may therefore occur at an interval other than 15 seconds after triggering. Second, to account for the statistical problem of clustering (multiple tumors in patients), we averaged TLC values in patients with multiple HCC nodules. Potential interaction or dependency effects may be missed by using such an approach.
In conclusion, we believe that the use of computer-assisted bolus tracking or a test bolus is important to determine the appropriate scan time during the HAP when a 40- or 64-detector CT scanner is used for hepatic dynamic CT. The peak TLC during the HAP occurred 18 seconds after triggering when a 40-detector CT was used and the trigger threshold level was set at an increase of 100 HU over the aortic baseline CT number. In the future, it will be important to clarify how much time elapses between attainment of the trigger threshold for bolus tracking and aortic peak enhancement with various injection protocols.
|
ADVANCES IN KNOWLEDGE |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
|
FOOTNOTES |
|---|
Abbreviations: CI = confidence interval • HAP = hepatic arterial phase • HCC = hepatocellular carcinoma • ROI = region of interest • SD = standard deviation • TLC = tumor-liver contrast
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, K.A., Y.Y.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, S.S., K.A., Y.N., Y.F.; clinical studies, S.S., K.A., Y.N., T.N., D.L., M.H., Y.F., S.M.; statistical analysis, K.A.; and manuscript editing, S.S., K.A., Y.Y.
|
References |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
1.5 cm) malignant hepatic neoplasms. AJR Am J Roentgenol 1995;164:879–884.This article has been cited by other articles:
|
|
 |
|
  D. Marin, R. C. Nelson, E. Samei, E. K. Paulson, L. M. Ho, D. T. Boll, D. M. DeLong, T. T. Yoshizumi, and S. T. Schindera Hypervascular Liver Tumors: Low Tube Voltage, High Tube Current Multidetector CT during Late Hepatic Arterial Phase for Detection--Initial Clinical Experience Radiology, June 1, 2009; 251(3): 771 - 779. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Marin, M. Di Martino, A. Guerrisi, G. De Filippis, M. Rossi, S. Ginanni Corradini, R. Masciangelo, C. Catalano, and R. Passariello Hepatocellular Carcinoma in Patients with Cirrhosis: Qualitative Comparison of Gadobenate Dimeglumine-enhanced MR Imaging and Multiphasic 64-Section CT Radiology, April 1, 2009; 251(1): 85 - 95. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Marin, C. Catalano, G. De Filippis, M. Di Martino, A. Guerrisi, M. Rossi, and R. Passariello Detection of Hepatocellular Carcinoma in Patients with Cirrhosis: Added Value of Coronal Reformations from Isotropic Voxels with 64-MDCT Am. J. Roentgenol., January 1, 2009; 192(1): 180 - 187. [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]
|
|
|
|
|
|
M. Kanematsu, S. Goshima, H. Kondo, K. T. Bae, K. Awai, and Y. Yamashita How Should We Optimize Bolus Tracking with Multidetector CT of the Abdomen? Radiology, February 1, 2008; 246(2): 643 - 645. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
