| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Osaka University Graduate School of Medicine D1, 2-2
Yamadaoka, Suita City, Osaka, 565-0871 Japan.
2 Department of Radiology, University of Pittsburgh Medical Center, 200 Lothrop
St, Pittsburgh, PA 15213.
3 General Electric Yokogawa Medical Systems, 7-127 Asahigaoka 4-chome, Hino
City, Tokyo, 191-8503 Japan.
Received October 15, 2001;
accepted after revision December 27, 2001.
Address correspondence to T. Murakami.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Forty-three patients with 87 hypervascular hepatocellular carcinomas (diameter: range, 3-80 mm; mean, 22 mm) and 19 patients with either chronic hepatitis or liver cirrhosis and without hepatocellular carcinoma who had undergone early arterial and late arterial phase imaging of the entire liver on multidetector CT were retrospectively enrolled in this study. The detector row configuration was 2.5 x 4 mm, the pitch was 6, and the scanning time was 10.5 sec for each phase. All patients received contrast medium (2 mL/kg of body weight) at a rate of 5 mL/sec; the mean scanning delay for the early arterial phase was 19.0 sec, and the mean delay for the late arterial phase was 34.5 sec. Eighty 2.5-mm-thick reconstruction images, forty 5-mm-thick reconstruction images, and twenty-six 7.5-mm-thick reconstruction images were obtained for each phase. Each image set was interpreted separately by three observers to detect hypervascular hepatocellular carcinoma by viewing images on a workstation monitor. Sensitivity, positive predictive value, and area under the receiver operating characteristic curve (Az) were calculated. We used retrospectively excellent follow-up and imaging or pathologic proof as the gold standard.
RESULTS. The mean sensitivity and positive predictive value for hypervascular hepatocellular carcinoma were 76% and 69% on 2.5-mm images, 73% and 69% on 5-mm images, and 67% and 76% on 7.5-mm images, respectively. No significant difference in sensitivity among the images was detected, except by one observer who reported a significant difference in the sensitivity between 2.5- and 7.5-mm images (p < 0.05) and between 5- and 7.5-mm images (p < 0.05). The mean Az values were 0.79, 0.80, and 0.78 for 2.5-, 5-, and 7.5-mm images, respectively. No significant difference in Az values among the images obtained with different slice thicknesses was detected.
CONCLUSION. For multidetector CT identification of hypervascular hepatocellular carcinoma, we found little or no advantage in reducing slice thickness to less than 5 mm.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Multidetector technology also enables us to obtain images of thinner slices during the same breath-hold and to retrospectively reconstruct images with different slice thicknesses obtained from the original CT data. The images of thinner slices are useful to reconstruct three-dimensional CT angiography [9,10,11,12] and have been shown to have a higher sensitivity for the detection of liver lesions [13]. The ability to detect smaller hepatocellular carcinoma may lead to earlier institution of therapy, including percutaneous ablation, and may help to better stage the disease, especially in patients being considered for liver transplantation.
However, disadvantages include the need to obtain and interpret an increased number of images. Another disadvantage of multidetector CT, especially if thin sections are used routinely, is the increased radiation exposure.
The purpose of this study was to evaluate the diagnostic impact of slice thickness and determine the optimal slice thickness in the detection of hypervascular hepatocellular carcinoma by comparing 2.5-, 5-, and 7.5-mm-thick reconstruction images.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Hypervascularity was defined as focal lesion hyperattenuation relative to the liver during at least one of the two arterial phases of CT. Proof of hypervascular hepatocellular carcinoma was confirmed using a combination of pathologic and radiologic criteria. For confirmatory imaging studies, the 43 patients with hepatocellular carcinoma and one patient without hepatocellular carcinoma underwent CT hepatic arteriography and CT during arterial portography using a technique previously described [15]. Of the 43 patients with hepatocellular carcinomas, 14 hepatocellular carcinoma nodules in 12 patients were pathologically proven after surgery or after percutaneous liver biopsy, which was performed within 1 month after double arterial phase dynamic CT. With regard to the stage of differentiation, one nodule was well differentiated, 12 nodules were moderately differentiated, and one was completely necrotic due to transcatheter hepatic arterial chemoembolization with iodized oil. The remaining 31 patients with hepatocellular carcinoma underwent angiography and transcatheter hepatic arterial chemoembolization with iodized oil. In these 31 patients, the lesions were diagnosed with certainty as hepatocellular carcinoma by assessing a combination of images on the basis of findings in the previous literature [1, 2, 16]. The diagnostic criteria for hepatocellular carcinoma on imaging were as follows: a lesion that showed hypoperfusion on CT during arterial portography; a round hyperenhancement on CT hepatic arteriography; a round, dense deposit of iodized oil on CT after transcatheter hepatic arterial chemoembolization with iodized oil; or both a definite growth and hypervascularity on arterial phase CT during follow-up.
The follow-up CT examinations were performed more than 2 months later
(range, 2-31 months; mean, 14 months) for 44 patients who had undergone CT
hepatic arteriography and CT during arterial portography. The diagnosis for
the other 18 patients without hepatocellular carcinoma was confirmed by
follow-up CT more than 6 months later (range, 6-24 months; mean, 13 months)
and by an unremarkable interval change in the serum level of the tumor marker
-fetoprotein. The 14 pathologically proven hepatocellular carcinoma
nodules also satisfied the radiologic criteria.
The CT scanner (LightSpeed QX/i; General Electric Yokogawa Medical Systems, Tokyo, Japan) detector configuration was 2.5 x 4 mm in the interspaced high-speed mode, in which four interspaced helical data sets are collected from eight 1.25-mm detector rows. The high-speed mode is equivalent to a pitch of 6 with the table speed set at a 15-mm rotation. One rotation of the X-ray tube was 0.8 sec. Electric pressure of 120 kVp and current of 270 mA were used. All patients received a low-osmolarity contrast medium, iohexol (Omnipaque 300 [300 mg I/mL]; Daiichi Pharmaceutical, Tokyo, Japan) by means of a power injector (Autoenhance A-50; Nemotokyorindou, Tokyo, Japan) at a rate of 5 mL/sec through a 20-gauge plastic IV catheter placed in an antecubital vein. The volume of contrast medium delivered was 2 mL/kg of body weight. Patient weight ranged from 43 to 96 kg (mean, 60 kg); therefore, the volume of contrast medium administered ranged from 86 to 192 mL (mean, 120 mL).
The scanning delay was determined using a test bolus of 15 mL of contrast medium administered at a rate of 5 mL/sec followed by the acquisition of a series of single-detector CT scans at a low dose (120 kVp, 10 mA). The scanning location was at the dome of the liver, and the monitoring scans were acquired every 2 sec for 10-40 sec. A cursor was placed over the abdominal aorta at this level, and the time to peak aortic enhancement was used to determine the scanning delay for the early arterial phase images.
Scanning began from the dome of the liver, with the location determined by means of a scout digital radiograph, and proceeded in a caudal direction for 10.5 sec, covering a z-axis distance of 20 cm. After an interscan delay of 5.0 sec for table movement, scanning resumed from the dome of the liver in a caudal direction for 10.5 sec. These images constituted the late arterial phase images. The total acquisition time was 26 sec and was accomplished during a single breathhold. The mean scanning delay for the early arterial phase was 19.0 sec (range, 12.0-28.0 sec), whereas the mean delay for the late arterial phase was 34.5 sec (range, 27.5-43.5 sec).
Eighty reconstruction images of 2.5-mm-thick slices and 40 reconstruction images of 5-mm-thick slices were obtained of each phase from the original CT data. Twenty-six images of a 7.5-mm slice thickness were reconstructed from the data sets of 2.5-mm-thick and 1.25-mm-thick interval images (159 images for each phase) by a radiology technologist on a workstation (Advantage Windows 3.1; General Electric Medical Systems, Milwaukee, WI) using the axial multiplanar reformation technique.
These reconstructed images of different slice thicknesses (2.5, 5, and 7.5 mm) were interpreted separately and independently by three experienced abdominal radiologists who viewed the images on a workstation monitor. Image review consisted of three sessions. The images of the 62 patients were divided into three groups: 1 (n = 21 patients), 2 (n = 21 patients), and 3 (n = 20 patients). The first session included 2.5-mm-thick images of group 1, 5-mm-thick images of group 2, and 7.5-mm-thick images of group 3. The second session included 2.5-mm-thick images of group 2, 5-mm-thick images of group 3, and 7.5-mm-thick images of group 1. The last session included 2.5-mm-thick images of group 3, 5-mm-thick images of group 1, and 7.5-mm-thick images of group 2. During each session, the images of patients appeared in random order. Information that identified a specific patient was removed from the images on the workstation. The three review sessions were performed at 4-week intervals.
The observers knew that the patients were at risk for hepatocellular carcinoma but did not know how many, if any, hepatocellular carcinomas were present in any patient. The observers recorded the size of each focal hepatic hypervascular lesion and its location by segment number (Couinaud classification) and indicated one of the four following confidence levels for the diagnosis of a malignant tumor: 1, probably absent; 2, equivocal; 3, probably present; and 4, definitely present. After all blinded review sessions had been completed, each lesion was compared with the proven tumor burden mentioned earlier.
To assess interobserver variability, we calculated the kappa statistic for multiple observers using the nonweighted binary kappa statistic. The lesions that were among the 87 proven hepatocellular carcinoma lesions that had been assigned a confidence level of 3 or greater were considered true-positive findings. The lesions not assigned or assigned a confidence level of 1 or 2 when a lesion was actually present were considered false-negative findings. The degree of disagreement was not factored into the calculation. A kappa value of 0.01-0.20 was considered slight agreement; 0.21-0.40, fair; 0.41-0.60, moderate; 0.61-0.80, substantial; and 0.81-1.00, almost perfect.
Sensitivity and positive predictive values for each slice thickness were also calculated. Sensitivity for each slice thickness was compared by means of the McNemar test. A two-tailed p value of less than 0.05 was considered significant.
For imaging of each slice thickness, alternative—free response receiver operating characteristic (ROC) curve analysis was performed on a tumor-by-tumor basis. Although the conventional ROC method allows only one response per image, the alternative—free response ROC method allows an observer response for all the lesions present, and we analyzed all 87 lesions in this study [17]. An alternative—free response ROC curve was fitted to each observer's confidence rating using a maximum-likelihood estimation (ROCKIT 0.9B; Metz CE, Chicago, IL). The diagnostic accuracy of imaging of each phase for each observer and their composite data were estimated by calculating the area under the ROC curve (Az). Differences between the imaging techniques in terms of the mean Az values were analyzed statistically using the two-tailed Student's t test for paired data. A two-tailed p value of less than 0.05 was considered significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The detection sensitivity for tumors of three size categories (diameter:
1 cm, >1-2 cm, or >2 cm) and the positive predictive values of each
of the three observers are shown in Table
1. The total number of hypervascular hepatocellular carcinomas
detected by three radiologists ranged from 60 to 73 (mean, 66) on the images
of 2.5-mm slices, 57-71 (mean, 64) on those of 5-mm slices, 51-69 (mean, 58)
on those of 7.5-mm slices (Fig.
1). The sensitivity was 69-84% (mean, 76%) for 2.5-mm-thick
images, 66-82% (mean, 73%) for 5 mm, and 59-79% (mean, 67%) for 7.5 mm. The
sensitivity for lesions more than 1 cm in diameter was higher than that of
smaller lesions (Table 1). No
significant differences in sensitivity among the slice thicknesses were
recorded by the three observers except observer 1; this observer detected
lesions 1 cm or smaller in diameter (Figs.
2A,2B,2C
and
3A,3B,3C)
with a significantly higher sensitivity for images obtained with a 2.5- rather
than a 7.5-mm slice thickness (p = 0.0133) and with a 5- rather than
a 7.5-mm slice thickness (p = 0.0412). The mean positive predictive
value was 69% for images obtained with a 2.5-mm slice thickness, 69% for 5 mm,
and 76% for 7.5 mm. The positive predictive value for the 7.5-mm images was
slightly higher than that for the 2.5- and 5-mm images; however, no
significant difference between each slice thickness was found.
|
|
|
|
|
|
|
|
Eight small hepatocellular carcinomas (diameter: range, 3-13 mm; mean, 8.9 mm) and two large hepatocellular carcinomas (diameter: 22 and 23 mm) could be detected only on 2.5-mm images by at least one observer. Eleven small hepatocellular carcinomas (diameter: range, 5-18 mm; mean, 11 mm) and two large hepatocellular carcinomas (diameter: 22 and 29 mm) could be detected only on images obtained with a 2.5- and 5-mm slice thickness by at least one observer. No lesion was detected only on 7.5-mm images.
Results of calculating the Az values for each phase are shown in Table 2. The mean Az values were 0.79 for images obtained with a 2.5-mm slice thickness, 0.80 for 5 mm, and 0.78 for 7.5 mm. There was no significant difference in the Az values between each slice thickness for all observers.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Using a single-detector helical scanner with a very thin collimation has the disadvantage of delivering an increased radiation dose to the patient. Using a multidetector scanner to image thin sections has the disadvantages of a significantly increased radiation burden and of the creation of a very large set of images that must be viewed and possibly filmed. Therefore, it is important to document the optimal collimation for detection of focal liver lesions.
Weg et al. [13] used a dual-detector CT scanner to study the usefulness of thin collimation to evaluate small (diameter < 10 mm) liver lesions, primarily hypovascular metastases. These researchers detected more lesions on images obtained with a 2.5-mm collimation than on those obtained with a 5-, 7.5-, or 10-mm collimation. Using multidetector CT, other investigators have described findings regarding the detection rate for focal hepatic lesions that conflict in terms of optimal slice thickness [23, 24].
The value of a relatively thin (4-7 mm) collimation has been well
documented for detection of most hypovascular or hypoattenuating focal liver
lesions. We focused our investigation on the detection sensitivity of
hypervascular or hyperattenuating lesions using double arterial phase CT of
hypervascular hepatocellular carcinoma as our model. Only one of the three
observers in our study had a significantly improved detection rate when
viewing thin sections (2.5- vs 7.5-mm collimation and 5- versus 7.5-mm
collimation). There was no significant difference of detection sensitivity for
2.5- versus 5-mm sections by all observers. Although we did not document
increased sensitivity or positive predictive value for 5- versus 7.5-mm
sections, we are reluctant to recommend routine use of 7.5-mm sections because
the resulting partial volume artifact may obscure the size of the lesion and
render the attenuation measurement of small lesions inaccurate. Because
observer 1 had a significantly improved rate of detection especially for small
lesions (1 cm in diameter), we suggest the use of at least a 5-mm
collimation for detecting hypervascular lesions.
The use of very thin sections (e.g., 1.25 mm) was not evaluated in our study, but seems a poor choice for the evaluation of focal hepatic lesions because of the decreased signal-to-noise ratio, a more prolonged scanning time necessary to cover the entire liver, and significantly increased radiation burden. The 5-mm-collimation images seem to provide the optimal combination of adequate signal-to-noise ratio and spatial and contrast resolution.
Urban et al. [25] have shown that overlapping reconstructions detect more lesions than contiguous reconstructions. These researchers compared an 8- versus a 4-mm reconstruction interval with an 8-mm collimation and detected 7% more lesions with the thinner sections. The increased detection rate was primarily for small lesions (diameter, <2 cm). In our study, we evaluated only contiguous nonoverlapping reconstructions and showed that reducing slice thickness to less than 5 mm offers little or no advantage in detection. We think that using a 5-mm slice thickness with overlapping reconstructions could yield better images than using a very thin slice thickness without overlapping reconstructions, and we believe that further study may be necessary.
We did not evaluate portal venous phase images in this study, because hypervascular tumors are usually not detected well in this phase. Nevertheless, we believe that portal venous phase images should be obtained routinely in any CT evaluation for known or suspected primary or metastatic tumor. Images obtained in the portal venous phase are usually optimal for additional characterization of liver tumors, which can improve differential diagnosis between hepatocellular carcinoma and other malignant and benign tumors, and are useful for the visualization of vascular anatomy and pathology.
When very thin sections on multidetector CT are used to evaluate hypervascular lesions, more lesions could possibility be detected including false-positive lesions such as small hemangiomas and shunts. These findings are difficult to characterize or biopsy. Moreover, an improved sensitivity for the detection of hypervascular lesions leads to all sorts of problems—an increased need for follow-up, performance of unnecessary invasive procedures, increased patient and clinician anxiety, and so on. However, as we mentioned earlier, dynamic studies including arterial and portal venous phase imaging will help in the differential diagnosis of these lesions.
The other limitation of our study could be the lack of histologic proof for every lesion that we believed to represent hepatocellular carcinoma. However, all lesions had several confirmatory studies such as CT hepatic arteriography, CT during arterial portography, CT after arterial infusion of iodized oil, and follow-up CT. Each of these studies, especially in combination, has been found to detect hypervascular hepatocellular carcinoma with an accuracy approaching 100%. Moreover, we were able to follow the course of most lesions over time and in response to therapy, especially transcatheter arterial chemoembolization.
In summary, we found no improvement in detection sensitivity of hypervascular hepatocellular carcinoma using a very thin collimation and believe that a 5-mm collimation is optimal for evaluation of focal hypervascular hepatic lesions.
|
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]
|
|
|
|
|
|
B. Kim, K. H. Lee, K. J. Kim, R. Mantiuk, H.-r. Kim, and Y. H. Kim Artifacts in Slab Average-Intensity-Projection Images Reformatted from JPEG 2000 Compressed Thin-Section Abdominal CT Data Sets Am. J. Roentgenol., June 1, 2008; 190(6): W342 - W350. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Ishigami, K. Yoshimitsu, H. Irie, T. Tajima, Y. Asayama, M. Hirakawa, and H. Honda Accessory Left Gastric Artery from Left Hepatic Artery Shown on MDCT and Conventional Angiography: Correlation with CT Hepatic Arteriography Am. J. Roentgenol., October 1, 2006; 187(4): 1002 - 1009. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Onishi, T. Murakami, T. Kim, M. Hori, R. Iannaccone, M. Kuwabara, H. Abe, S. Nakata, K. Osuga, K. Tomoda, et al. Hepatic Metastases: Detection with Multi-Detector Row CT, SPIO-enhanced MR Imaging, and Both Techniques Combined Radiology, April 1, 2006; 239(1): 131 - 138. [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]
|
|
|
|
|
|
D. A. Bluemke, D. Sahani, M. Amendola, T. Balzer, J. Breuer, J. J. Brown, D. D. Casalino, P. L. Davis, I. R. Francis, G. Krinsky, et al. Efficacy and Safety of MR Imaging with Liver-specific Contrast Agent: U.S. Multicenter Phase III Study Radiology, October 1, 2005; 237(1): 89 - 98. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. H. Kim, D. Choi, S. H. Kim, J. H. Lim, W. J. Lee, M. J. Kim, H. K. Lim, and S. J. Lee Ferucarbotran-Enhanced MRI Versus Triple-Phase MDCT for the Preoperative Detection of Hepatocellular Carcinoma Am. J. Roentgenol., April 1, 2005; 184(4): 1069 - 1076. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Valls, M. Cos, J. Figueras, E. Andia, E. Ramos, A. Sanchez, T. Serrano, and J. Torras Pretransplantation Diagnosis and Staging of Hepatocellular Carcinoma in Patients with Cirrhosis: Value of Dual-Phase Helical CT Am. J. Roentgenol., April 1, 2004; 182(4): 1011 - 1017. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. E. Steingruber, A. Mallouhi, B. V. Czermak, P. Waldenberger, E. Gassner, F. Offner, A. Chemelli, A. Koenigsrainer, W. Vogel, and W. R. Jaschke Pretransplantation Evaluation of the Cirrhotic Liver with Explantation Correlation: Accuracy of CT Arterioportography and Digital Subtraction Hepatic Angiography in Revealing Hepatocellular Carcinoma Am. J. Roentgenol., July 1, 2003; 181(1): 99 - 108. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Noguchi, T. Murakami, T. Kim, M. Hori, K. Osuga, S. Kawata, S. Kumano, A. Okada, T. Sugiura, and H. Nakamura Detection of Hepatocellular Carcinoma: Comparison of Dynamic MR Imaging with Dynamic Double Arterial Phase Helical CT Am. J. Roentgenol., February 1, 2003; 180(2): 455 - 460. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
