| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Yamanashi Hospital, Kofu, Yamanashi, Japan.
2 Present address: Department of Radiology, Hyogo Medical Center for Adults,
Kitaoji 13-70, Akashi, Hyogo 673-8558, Japan.
3 Department of Radiology, School of Medicine, University of Yamanashi, Chuo
Yamanashi, Japan.
4 Present address: Department of Radiology, Yokohama Sakae kyosai Hospital,
Kanagawa, Japan.
5 Department of Pathology, Yamanashi Hospital, Kofu Yamanashi, Japan.
Received March 24, 2005;
accepted after revision December 16, 2005.
Address correspondence to S. Monzawa.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2 cm)
hepatocellular carcinomas (HCCs) in patients with liver cirrhosis. MATERIALS AND METHODS. Triphasic (arterial, portal venous, and delayed phases) dynamic CT was performed in 33 patients with 48 HCCs proven histopathologically and in 65 control subjects. Arterial, portal venous, and delayed phase images were obtained 30 seconds, 68-70 seconds, and 5 minutes after the start of contrast material injection, respectively. Three blinded observers reviewed the images independently and evaluated tumor attenuation. Diagnostic performance for the combination of phases was assessed using receiver operating characteristic (ROC) curve analysis.
RESULTS. On arterial phase images, 28 of the 48 HCCs were hyperattenuating, nine were isoattenuating, and 11 were hypoattenuating. On portal venous phase images, three tumors were hyperattenuating, 17 were isoattenuating, and 28 were hypoattenuating. On delayed phase images, five tumors were isoattenuating, and 43 were hypoattenuating. The mean sensitivity for the combination of arterial and portal venous phase imaging was 86.8%, that for the combination of arterial and delayed phase imaging was 90.3%, and that for the combination of all three phase imaging was 93.8%. The area underneath composite ROC curve (Az) for the combination of all three phase imaging (Az = 0.940) was significantly higher than that for the combination of arterial and portal venous phase imaging (Az = 0.917) and for the combination of arterial and delayed phase imaging (Az = 0.922).
CONCLUSION. Delayed phase imaging is useful for detecting small HCCs and should be included in dynamic CT examinations of patients with liver cirrhosis.
Keywords: cirrhosis • delayed phase CT • dynamic CT • hepatocellular carcinoma • liver disease • oncologic imaging • triphasic CT
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The value of adding delayed phase imaging to biphasic dynamic CT for detecting HCCs has been previously discussed. Choi et al. [6] reported that delayed phase imaging is not always necessary and that biphasic imaging consisting of the arterial and portal venous phases is adequate for the detection of hypervascular HCCs. On the other hand, several investigators reported that HCCs are often more conspicuous on delayed phase images than on portal venous phase images and that detectability and characterization are improved by adding delayed phase imaging to the biphasic CT examination [4, 5, 7, 8].
Small HCCs are often well differentiated and hypovascular, and unlike large
or advanced HCCs, small HCCs are not enhanced because of a poor arterial blood
supply, thus making them less conspicuous on arterial phase images. Therefore,
small HCCs are not as detectable as advanced HCCs on arterial phase images
[1,
5,
13-15].
Lim et al. [8] showed that
delayed phase imaging increased sensitivities in the detection of HCCs,
especially in the detection of small tumors. There are no other detailed
reports, to our knowledge, that discuss the role of delayed phase imaging in
the diagnosis of small HCCs. Thus, we undertook this study to assess the
diagnostic performance of triphasic dynamic CT and to clarify the usefulness
of delayed phase imaging in the detection and characterization of small (
2 cm) HCCs in patients with liver cirrhosis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Dynamic CT was performed using a single-detector scanner (Proceed SA, GE-YMS). Arterial and portal venous phase images were obtained using a helical scanning technique with a 5-mm beam collimation width, 1:1.4 pitch, and continuous 5-mm reconstruction; delayed phase images were obtained using an incremental (cluster) scanning technique with a 5-mm slice thickness and 2-mm gap. One hundred twenty milliliters of iohexol (300 mg I/mL) was injected at a rate of 4 mL/s, and arterial phase imaging was performed 30 seconds after the start of the contrast material injection. The delay time was 68-70 seconds for portal venous phase imaging and 5 minutes for delayed phase imaging.
A total of 339 images, including 144 images of the 33 patients with 48 HCCs and 195 images of the 65 control subjects, were reviewed. These CT images were rephotographed on hard-copy films with all identifying references, such as patient name, age, sex, and hospital record number, masked. Three blinded observers who serve mainly as gastrointestinal radiologists in daily clinical and research practice were invited from another institution to review the images. They knew that the patients had liver cirrhosis and were referred for assessment of a possible liver tumor, but they were blinded to other information, including medical history, about the patients. Three review sessions for paired arterial and portal venous phase images, paired arterial and delayed phase images, and the combination of all three phase images were performed at 2-week intervals to minimize learning bias.
The observers indicated the presence or absence of HCCs and assigned diagnostic confidence levels to their observations as follows: 1, definitely absent; 2, probably absent; 3, possibly present; 4, probably present; or 5, definitely present. The observers recorded the size and attenuation (compared with the liver parenchyma) of the lesions detected. Consensus readings were then performed to determine the tumor size and enhancement patterns. Differences in tumor attenuation on each phase between well-differentiated HCCs and moderately or poorly differentiated HCCs were compared using the chi-square test.
For objectivity and reproducibility of the image analysis performed in this study, the criteria for HCCs, dysplastic nodules, and arterioportal shunts were set. The criteria for HCCs were a nodule showing enhancement on arterial phase, iso- or low attenuation on portal venous phase, and low attenuation on delayed phase images [5, 7, 8]. Many well-differentiated HCCs are depicted as a nodule showing isoattenuation or low attenuation on arterial and portal venous phase images and low attenuation on delayed phase images. However, because dysplastic nodules also often show the same enhancement pattern, the distinction between these two lesions is often difficult [18]. Although there seems to be significant overlap in size between the two lesions, dysplastic nodules tend to be smaller than well-differentiated HCCs. Therefore, we chose 10 mm as a practical cutoff value, following the value used in the study by Jang et al. [19], and nodules 10 mm or larger in diameter were regarded as HCCs and those smaller than 10 mm, as dysplastic nodules.
The lesions showing enhancement on arterial phases, isoattenuation or high attenuation on portal venous phase images, and isoattenuation on delayed phase images were also regarded as HCCs. However, they needed to be distinguished from arterioportal shunts because these lesions showed a similar enhancement pattern. When the lesions showed a typical wedge shape with or without internal linear branching structures on arterial phase images, they were regarded as arterioportal shunts. When the lesions were round, they were considered HCCs; however, they were assigned a confidence level of "3, possibly present" [8, 20].
The sensitivity, specificity, and accuracy for each observer and for each combination of phases were calculated using the confidence level ratings of the images; ratings of 3 or greater were regarded as being indicative of the presence of HCCs. The two-tailed Student's t test for paired data was used to assess statistical significance in the mean values of all three observers. For each combination of phases, a binomial receiver operating characteristic (ROC) curve was fitted to each observer's confidence-rating data by using a maximum-likelihood curve-fitting algorithm. The diagnostic accuracy of each combination of phases was estimated by calculating the area underneath the ROC curve (Az). Composite ROC curves that combined the performance of all three independent observers into a single curve were obtained to rate the pooled data of the observers. The mean Az values of all three observers for each combination of phases were then compared by using the two-tailed Student's t test for paired data. For the evaluation of interobserver variability in interpreting images, kappa statistics were used to measure the degree of agreement between all pairs of observers regarding the presence or absence of small HCCs. A kappa value of up to 0.40 indicated positive but poor agreement, a kappa value of 0.41-0.75 indicated good agreement, and a kappa value of greater than 0.75 indicated excellent agreement.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
|
|
On portal venous phase images, 28 (58%) of the 48 HCCs were hypoattenuating and 17 (35%) were isoattenuating; on delayed phase images, 43 (90%) were hypoattenuating and five (10%) were isoattenuating (Figs. 1A, 1B, 1C, 2A, 2B, and 2C). There were no remarkable differences in enhancement patterns on portal venous phase images (p = 0.985) and delayed phase images (p = 0.622) between the 34 well-differentiated HCCs and 14 moderately or poorly differentiated HCCs. No tumors showed hyperattenuation on delayed phase images.
A Venn diagram (Fig. 3) shows the numbers of HCCs detected during each phase. Of the 46 HCCs depicted with triphasic dynamic CT, 26 were depicted on the images of all three phases. Two HCCs were revealed on only arterial phase imaging and three HCCs were revealed on only delayed phase imaging (Figs. 2A, 2B, and 2C). No HCCs were shown on only portal venous phase imaging.
|
|
The individual and mean Az for the combination of all three phases, paired arterial and delayed phases, and paired arterial and portal venous phases are shown in Table 3. Composite ROC curves formed on the basis of pooled data from the three observers for each combination of phases are shown in Figure 4. The combination of all three phases showed the greatest diagnostic performance for determining whether a HCC was present. Diagnostic performance of paired arterial and delayed phases was superior to that of paired arterial and portal venous phases. The mean Az for the combination of all three phases was significantly greater than that for paired arterial and delayed phases (p = 0.017) and that for paired arterial and portal venous phases (p = 0.014). The mean Az for paired arterial and delayed phases was greater than that for paired arterial and portal venous phases, but the difference was not statistically significant (p = 0.393). The kappa values obtained from all pairs of observers are summarized in Table 4. Good or excellent agreement was obtained among the observers.
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
During arterial and portal venous phases of hepatic dynamic CT, tumor vascularity and blood supply of the liver play a greater role in determining tumor-liver parenchyma contrast. During arterial phase, the liver parenchyma is enhanced minimally because of the low dependency on arterial blood supply. Tumors are enhanced to variable degrees according to their vascularity; hypervascular tumors are enhanced maximally and show the greatest contrast. On the other hand, hypovascular tumors are minimally enhanced because of poor arterial blood supply and they are often isoattenuating and inconspicuous on arterial phase images [24]. During portal venous phase, the liver parenchyma has peak enhancement because of the enormous volume of portal venous blood supply to the liver. Hypovascular tumors remain unenhanced or minimally enhanced and are depicted as a hypoattenuating area with the greatest contrast, whereas hypervascular tumors often show increased attenuation similar to the attenuation of enhanced liver parenchyma and decreased conspicuity [4, 6-8, 24].
During delayed phase (e.g., 3-5 minutes after injection), equilibrium is attained and the concentration of contrast material in the vascular space and in the interstitial space becomes almost equal. On reaching this equilibrium, the increase in tissue attenuation by the distribution of contrast material depends chiefly on the volume of interstitial space because it is about three times as large as the vascular space in most tissues. Therefore, enhancement observed on delayed phase images reflects the relative volume of the interstitial space in tissues [22]. Most tumors show hypoattenuation on delayed phase images because they have increased cellularity or, conversely speaking, decreased interstitial space than the liver parenchyma. Hemangiomas have a disproportionately large vascular space and lesions with rich fibrous tissues have a large interstitial space; these are depicted as a hyperattenuating area on delayed phase images [23].
The detection of small ( 2 cm) HCCs is important because they can be
treated effectively by means of local ablation therapy or surgical removal,
and when treated, the prognosis for those patients is better than that of
patients with HCCs larger than 2 cm
[8]. Small HCCs may include
tumors with variable differentiation grades from well to poorly
differentiated. In hepatocarcinogenesis, well-differentiated HCCs are
considered to be an early form of HCCs and to develop later into moderately or
poorly differentiated tumors with increasing malignancy.
Well-differentiated HCCs are usually small; in fact, most are less than 2 cm. Their vascular supply and vascularity are variable and dependent on the grade of malignancy. Tumors with low-grade malignancy (or borderline lesions) receive poor arterial blood supply and frequently receive portal venous blood supply in addition to arterial blood supply. As the grade of malignancy increases, tumor neovascularity develops and arterial blood supply is increased in tumors [25, 26]. In our study, 44% of well-differentiated HCCs were hyperattenuating, 26% were isoattenuating, and 29% were hypoattenuating on arterial phase images. This variety of enhancement seen among the tumors reflects differences in vascular supply and development of neovascularity and might suggest that the tumors have a different grade of malignancy.
Moderately or poorly differentiated HCCs are considered an advanced form, and they are almost solely supplied by the hepatic artery because the portal tracts in the tumors disappear [25, 26]. As shown in our data that 93% were hyperattenuating, most moderately or poorly differentiated HCCs are hypervascular and receive large amounts of arterial blood flow and show hyperattenuation on arterial phase images, although some cases receive poor arterial supply and show iso- or hypoattenuation on arterial phase images.
On portal venous phase images, 58% of the 48 tumors showed hypoattenuation and 35% showed isoattenuation; the well- and moderately to poorly differentiated HCCs showed similar percentages. Even in the hyperattenuating tumors, tumor conspicuity was decreased during this phase compared with the arterial phase. Kim et al. [7] reported that attenuation values were increased in many HCCs about as much as in the liver parenchyma. This tumor enhancement occurring on portal venous phase images may be due to consecutive supply of contrast material via the hepatic arteries continuing after the arterial phase [1]; to residuals of contrast material in the tumors, mainly fed during the arterial phase [7]; or to dual blood supply of tumors by both arterial and portal venous blood flow [13].
On delayed phase images, 88% of well-differentiated HCCs and 93% of moderately or poorly differentiated HCCs showed hypoattenuation. The cause of hypoattenuation seen in HCCs on delayed phase images has been discussed. Several authors have proposed that washout of contrast material from tumors is the reason for hypoattenuation on delayed phase images [5, 8]. Also, delayed enhancement of the hepatic parenchyma due to portal hypertension or uptake by hepatocytes was suggested as a possible cause [5]. However, in HCCs, even those of well-differentiated grade, tumor cells proliferate more densely in a trabecular or solid pattern in tumor tissues than in surrounding liver parenchyma [27, 28]. This increased cellularity is thought to be a primary reason for this CT finding.
Diagnostic performance of paired arterial and delayed phase imaging was superior to that of paired arterial and portal venous phase imaging for the detection of small HCCs. Delayed phase imaging showed more tumors than portal venous phase imaging. There were no tumors depicted only on portal venous phase images, but three tumors were detected only on delayed phase images. These results may account for the differences in diagnostic performance.
In our study, the combination of all three phases showed the greatest diagnostic performance. Choi et al. [6] reported that the combination of arterial and portal venous phase is adequate for the detection of HCCs and concluded that delayed phase imaging can be omitted to decrease scanning time and radiation hazards. However, their subjects consisted of patients with hypervascular HCCs. Lim et al. [8] compared triphasic dynamic CT with biphasic dynamic CT with the combination of arterial and portal venous phase imaging for tumor detectability in a large number of HCCs with a variable size including hypovascular tumors. In their study, although delayed imaging did not increase sensitivity for the detection of large HCCs measuring 2 cm or greater, small HCCs less than 2 cm were usually more conspicuous on delayed phase images than on portal venous phase images; some were detected only on delayed phase images, and triphasic dynamic CT was found to be superior to biphasic dynamic CT for the detection of small HCCs.
In our results, the contribution of portal venous phase imaging in the combination of all three phase imaging cannot be fully explained. The possible reason for this might be related to the fact that the portal veins and hepatic veins in the liver are depicted clearly on portal venous phase images because they are opacified maximally. Viewing portal venous phase images together might help an interpreting radiologist distinguish small tumors from blood vessels and might increase confidence levels [7, 8].
Delayed phase imaging may have advantages for distinguishing HCCs from attenuation abnormalities due to arterioportal shunts. Arterioportal shunts can cause a hyperattenuating area mimicking HCCs on arterial phase images in cirrhotic livers. Although large lesions with typical findings—including a wedge shape and a hyperattenuating linear branching structure representing early opacification of portal veins—can be readily diagnosed, small lesions are difficult to distinguish from small hypervascular HCCs because they often show a nodular shape and lack a branching structure. Because the lesions usually show isoattenuation or slight hyperattenuation on portal venous phase images [20], portal venous phase imaging is not helpful for the distinction. The finding of arterioportal shunts on delayed phase images has not been well described; however, it is probable that arterioportal shunts will show isoattenuation [8]. Therefore, delayed phase imaging may play an important role in distinguishing hypervascular HCCs from arterioportal shunts. When a lesion shows hypoattenuation on delayed phase images, it is unlikely to be due to arterioportal shunts. Delayed phase imaging necessitates increased radiation exposure to patients and additional scanning time. However, these costs can be offset by the benefits that result from reliable distinction from arterioportal shunts and increased detection of hypovascular tumors.
Distinction of well-differentiated HCCs from moderately or poorly differentiated HCCs is important. The biologic behavior of well-differentiated HCCs is uncertain; however, they are thought to have relatively low malignancy potential, rarely invading into blood vessels and metastasizing. Therefore, the tumors can be treated before they spread to the liver or distant organs [29-31]. Triphasic dynamic CT may not be useful for this purpose. Hypervascular well-differentiated HCCs making up 44% of well-differentiated HCCs may be difficult to distinguish from moderately or poorly differentiated HCCs because there is a considerable overlap in the enhancement pattern between the two. Also, the distinction of hypovascular well-differentiated HCCs from dysplastic nodules seems to be difficult because both lesions show a similar enhancement pattern [18]. Percutaneous biopsy may ultimately be needed if definitive histology is required.
This study has several potential limitations. We used incremental scanning for delayed phase imaging because of limitations in X-ray tube heating. Single-detector helical scanning used for arterial and portal venous phase imaging is thought to be inferior to incremental scanning in z-axis spatial resolution at the same collimation thickness. Thus, differences in scanning techniques might result in differences in the detection of small HCCs. The dose of contrast material, injection speed of contrast material administration, and scanning timing for each phase were not individualized. Instead, these parameters were fixed in our study, which might be suboptimal. Tumors detected on sonography and proven at biopsy were included in this study. We could not study tumors not detected on sonography because histologic confirmation was not obtained for most of them. The selection criteria for our study population might have caused some bias in our study.
In conclusion, delayed phase imaging of dynamic CT has superior capabilities in the depiction of small HCCs and allows detection or confirmation of tumors that would not be detected or would be difficult to detect using arterial and portal venous phase imaging. Therefore, we recommend that delayed phase imaging should be included in multiphasic dynamic CT for detecting HCCs in patients with liver cirrhosis.
|
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:
|
|
 |
|
  M A Silva, B Hegab, C Hyde, B Guo, J A C Buckels, and D F Mirza Needle track seeding following biopsy of liver lesions in the diagnosis of hepatocellular cancer: a systematic review and meta-analysis Gut, November 1, 2008; 57(11): 1592 - 1596. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) Among Observers Regarding the Presence of 48 Small
(