| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Departments of Radiology (J.H.L., D.C., S.K.C., S.H.K., W.J.L., H.K.L.), Diagnostic Pathology (C.K.P.), Gastroenterology (S.W.P.), and General Surgery (Y.I.K.), Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Kangnam-ku, Seoul 135-710, Korea. Received November 8, 2000; revision requested January 9, 2001; revision received February 14; accepted March 19. Address correspondence to C.K.P. (e-mail: ckpark@smc.samsung.co.kr)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Sixty-nine histopathologically proved moderately or poorly differentiated hepatocellular carcinomas, 10 well-differentiated hepatocellular carcinomas, and 19 dysplastic nodules were retrospectively studied in 68 patients with cirrhosis who underwent ferumoxides-enhanced MR imaging. The contrast-to-noise ratio between the nodules and surrounding parenchyma was calculated at T2-weighted fast spin-echo imaging, and the difference in the number of Kupffer cells between the nodules and surrounding hepatic tissue was calculated histopathologically. The results of MR imaging and histopathologic examination were correlated.
RESULTS: All 69 moderately or poorly differentiated hepatocellular carcinomas had high contrast-to-noise ratios at MR imaging and large differences in the number of Kupffer cells. Six of the 10 well-differentiated hepatocellular carcinomas had contrast-to-noise ratios of zero or nearly zero, and five of these had little difference in the number of Kupffer cells. All 19 dysplastic nodules had contrast-to-noise ratios of zero or nearly zero, and there were virtually no differences in the number of Kupffer cells.
CONCLUSION: Hepatocellular nodule conspicuity at ferumoxides-enhanced MR imaging depends on differences in the number of Kupffer cells within a nodule and the surrounding cirrhotic liver; moderately or poorly differentiated hepatocellular carcinomas can be distinguished from well-differentiated hepatocellular carcinomas and dysplastic nodules.
Index terms: Iron • Liver, diseases, 761.323, 761.794 • Liver, MR, 761.121411, 761.121412, 761.121415, 761.12143 • Liver, nodules, 761.319 • Magnetic resonance (MR), contrast enhancement 761.12143
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
On the other hand, Kupffer cells are present in benign hepatic masses such as hepatic adenomas, focal nodular hyperplasia, and cirrhotic regenerative nodules (5,6). Kupffer cells are also known to be present in dysplastic nodules (DNs) and in well-differentiated or early hepatocellular carcinomas (HCCs) (4,5,7–9). We hypothesized that, as the Kupffer cells take up ferumoxides, the conspicuity of the hepatic nodular lesions on ferumoxides-enhanced MR images will depend on the relative difference in the number of Kupffer cells in a lesion, as compared with that in background liver. The purpose of this study was to correlate the conspicuity of HCCs and DNs on ferumoxides-enhanced MR images with the number of Kupffer cells in the lesions, as compared with that in background liver in histopathologic findings. To our knowledge, this is the first study in which ferumoxides-enhanced MR imaging has been compared with a pathomorphologic study of Kupffer cells in HCCs and DNs.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The remaining 68 patients who had HCC in cirrhotic livers and underwent hepatic resection formed the final study population. All 68 patients underwent US and triphasic (hepatic arterial, portal venous, and delayed phase) dynamic helical CT, and six of these patients underwent gadolinium-enhanced MR imaging. These 68 patients underwent segmentectomy (n = 17), lobectomy (n = 48), extended right lobectomy (n = 2), or hepatic transplantation (n = 1) after ferumoxides-enhanced MR imaging. The resectability criteria for HCC at our institution are a tumor or tumors limited to one hepatic lobe with good function (Child-Pugh class A or B) and good hepatic reserve volume at CT, regardless of the number and size of HCCs. The study was approved by the institutional review board of our hospital, and written informed consent was obtained from all patients.
A total of 98 hepatocellular nodules were resected and included in this study. There were 69 moderately or poorly differentiated HCCs, 10 well-differentiated HCCs, and 19 DNs. HCCs were graded histologically as well differentiated, moderately differentiated, or poorly differentiated in accordance with the classification of primary hepatic cancer by the Liver Cancer Study Group of Japan (10). Among the moderately or poorly differentiated HCCs, there were 35 moderately differentiated HCCs and 34 poorly differentiated HCCs. Among the DNs, there were seven high-grade DNs and 12 low-grade DNs. The DNs were classified as low- or high-grade in accordance with the criteria of the International Working Party on the terminology of nodular hepatocellular lesions (11).
There were 53 men and 15 women aged 35–73 years (mean age, 51 years). Cirrhosis was classified by using the Child-Pugh system. There were 53 patients with Child-Pugh class A cirrhosis and 15 with Child-Pugh class B cirrhosis. Hepatic cirrhosis was related to hepatitis B virus in 57 patients, to hepatitis C virus in seven patients, and to alcohol in two patients and was of unknown origin in two patients.
Hepatic surgery was performed within 3 weeks following the imaging examinations, the results of which were compared with the histopathologic results in the resected hepatic specimens.
Ferumoxides-enhanced MR Imaging
All MR images were obtained by using a 1.5-T imager (Horizon; GE Medical Systems, Milwaukee, Wis) with a phased-array multicoil system. Ferumoxides solution (Feridex IV; Advanced Magnetics, Cambridge, Mass) at a dose of 15 µmol per kilogram of body weight diluted in 100 mL of 5% glucose solution was infused through a 5-µm specific filter for approximately 30 minutes. MR imaging was initiated 30 minutes after the end of the infusion.
Transverse images from five sequences were obtained with a section thickness of 6–8 mm and an intersection gap of 2 mm. The MR imaging protocol included fat-suppressed respiratory-triggered fast spin-echo imaging with two echo times (repetition time msec/effective echo time msec, 3,333–8,571/18 and 90–117; echo train length, 10–18; two signals acquired; matrix, 256 x 256; and field of view, 30–36 cm), gradient-recalled acquisition in the steady state (GRASS) (216/20; flip angle, 20°; one signal acquired), T2*-weighted fast multiplanar GRASS (130/8.4–9.5; flip angle, 30°), and breath-hold in-phase T1-weighted multiplanar fast spoiled gradient-echo (200/4.2; flip angle, 90°; matrix, 256 x 160) imaging.
Image Analysis
The MR images obtained with the five sequences after ferumoxides enhancement were evaluated independently by three gastrointestinal radiologists (J.H.L., S.H.K., W.J.L.). They knew that the patients had been referred for assessment of suspected HCCs but were not provided with any other patient information and reviewed the liver on ferumoxides-enhanced MR images for any visible HCC nodule and for DNs. Because there were no specific findings for HCC or DNs, the presence of a nodule was graded; one of five confidence levels was assigned: 1, nodule definitely absent; 2, nodule probably absent; 3, nodule possibly present; 4, nodule probably present; or 5, nodule definitely present. The sensitivities for the detection of hepatocellular nodules by the three observers and the composite data were determined by using the number of nodules assigned a score of 3 or greater (ie, possibly to definitely present).
On enlarged images (full screen) on the monitors of a picture archiving and communication system (2 K x 2 K PathSpeed Workstation; GE Medical Systems), one radiologist (S.K.C.) performed operator-defined region-of-interest (ROI) measurements of the mean signal intensity of each nodule, the hepatic parenchyma, and the background noise on the fat-suppressed respiratory-triggered T2-weighted fast spin-echo images (3,338–8,571/90–117) (12,13). To measure the hepatic signal intensity, ROIs were set in the area devoid of focal changes in signal intensity, large vessels, and prominent artifacts. For the hepatic lesions, circular or ovoid ROIs were drawn to encompass as much of the lesion as possible. Noise was measured on each image by using ROIs positioned just ventral to the right anterior abdominal wall. The areas with the most prominent ghost artifacts were not included.
ROIs were electrically drawn larger than 256 mm2 (range, 256–1,677 mm2) in the liver, 50 mm2 (range, 50–4,148 mm2) in the lesions, and 256 mm2 (range, 256–737 mm2) in the area with noise. ROIs were drawn three times in each place, and the mean values were obtained. The lesion-liver contrast-to-noise ratio (CNR) was calculated with the following formula: (SIlesion - SIliver)/SDnoise, where SI is the signal intensity. When the nodule was not visualized with any MR imaging sequence, we assumed that the nodule had the same CNR as did the surrounding cirrhotic hepatic parenchyma and recorded the CNR as zero.
Pathomorphologic Study of Kupffer Cells
The surgically resected specimens were serially sectioned with 5–10-mm thickness in a transverse or coronal plane, depending on the location of a nodule. A total of 98 nodules were identified and their diameters were measured. All nodules, including the surrounding cirrhotic liver, were fixed in 10% formalin, embedded in paraffin, cut into 4-µm sections, and stained with hematoxylin-eosin. Histopathologic diagnosis in the hepatic nodules was made by means of consensus of at least two pathologists (including C.K.P.) who specialized in hepatic pathology. The pathologists were informed that the resected specimens were suspected of containing HCC, but detailed radiologic information on the location, number, and size of the hepatic nodules was not given. In the nodules with histologic differentiation, the most dominant component was chosen.
Immunohistochemical staining of CD68 was performed by using a routine avidin-biotin-peroxidase complex technique with an antigen retrieval method. The primary antibody was mouse monoclonal antibody to human macrophage (Anti-CD68; Dako, Glostrup, Denmark) at a dilution of 1:300. Diaminobenzidine was used as a chromogen substrate, and sections were counterstained with hematoxylin-eosin. Among the cells with a positive reaction to the antihuman macrophage antibody, those located in the sinusoids of surrounding nonneoplastic tissue or blood spaces in neoplastic tissues and those having a stellate or spindle shape were regarded as Kupffer cells (5). One pathologist (C.K.P.) counted the Kupffer cells in the hepatocellular nodule and surrounding nonneoplastic liver in two high-power fields (x400) in which Kupffer cells were present at a relatively uniform distribution and then calculated the average. The difference in the number of Kupffer cells between the hepatocellular nodule and the surrounding nonneoplastic liver was calculated by subtracting the number of Kupffer cells in the surrounding liver from that in the nodule.
Correlation of Images and Difference in the Number of Kupffer Cells and Statistical Analysis
The study coordinator (S.K.C.) correlated the nodules on MR images and the histopathologically confirmed HCCs and DNs. Benign lesions, including hemangiomas and cysts classified by the three observers, as hepatocellular nodules were excluded.
Interobserver agreement for evaluating the MR images was assessed with the 
statistic. A value of more than 0.80 was considered to indicate excellent agreement (14). The CNRs and differences in the number of Kupffer cells of DNs, well-differentiated HCCs, and moderately or poorly differentiated HCCs were compared by performing the Scheffé multiple comparisons test with ranking. The relationship between the CNRs and the differences in the number of Kupffer cells between the nodules and surrounding cirrhotic hepatic parenchyma was evaluated with Spearman correlation analysis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
values for the three observers showed excellent agreement (observer 1 vs 2, 0.826; observer 1 vs 3, 0.924; and observer 2 vs 3, 0.900).
The differences in lesion CNRs between HCCs and DNs on the T2-weighted fast spin-echo ferumoxides-enhanced MR images are shown in the Table. On the ferumoxides-enhanced MR images, 69 moderately or poorly differentiated HCCs were depicted as high-signal-intensity masses (Fig 1); the average CNR was 25.8. Six nodules were depicted as having slightly high signal intensity (5 < CNR < 15). At pathomorphologic examination, there was a considerably lower number of Kupffer cells than in the surrounding cirrhotic liver parenchyma (
10) in all but one moderately or poorly differentiated HCC.
|
|
|
|
10) in the number of Kupffer cells.
|
|
|
|
|
|
|
|
Figure 5 shows the CNRs of the hepatocellular nodules (ie, DNs, well-differentiated HCCs, and moderately or poorly differentiated HCCs), and Figure 6 shows the differences in the number of Kupffer cells between the nodules and the surrounding cirrhotic hepatic parenchyma. The CNRs of the moderately or poorly differentiated HCCs were significantly higher than those of the DNs and well-differentiated HCCs (P < .05, Scheffé multiple comparisons test with ranking). The difference in CNR between the well-differentiated HCCs and DNs was not statistically significant (P > .05, Scheffé multiple comparisons test with ranking). The differences in the number of Kupffer cells between the nodules and the surrounding hepatic parenchyma of the moderately or poorly differentiated HCCs were significantly larger than those of the DNs and well-differentiated HCCs (P < .05, Scheffé multiple comparisons test with ranking). The difference in the number of Kupffer cells in the well-differentiated HCCs was slightly larger than that in the DNs but was not statistically significant (P > .05, Scheffé multiple comparisons test with ranking).
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
In our study, the conspicuity of nodular lesions in the liver correlated well with the difference in the number of Kupffer cells between the hepatic parenchyma and the nodular lesions. When there was a large difference, the lesion stood out prominently, and when the difference was small or not significant, the nodular lesions were faint or not depicted. Thus, all moderately or poorly differentiated HCCs (69 of 69) had high signal intensity, and the differences in the number of Kupffer cells between the lesion and surrounding cirrhotic hepatic parenchyma were large. All DNs (19 of 19) contained nearly the same number of Kupffer cells as did the surrounding cirrhotic hepatic parenchyma and were not depicted on the T2-weighted MR images.
Seven of the 10 well-differentiated HCCs had nearly the same number of Kupffer cells as did the surrounding hepatic parenchyma; five of these seven were not depicted on the T2-weighted MR images (Fig 3), and two were visible as faintly high signal intensity. These two were well-differentiated HCCs in which the masses of HCC were well depicted as areas of high signal intensity; in these two cases, there were high CNRs and large differences in the numbers of Kupffer cells between the tumor and hepatic parenchyma (Fig 2). In the remaining case, there was a zero CNR but a large difference in the number of Kupffer cells. Therefore, in three cases, there was no explainable cause of the mismatch between the CNR and the difference in the number of Kupffer cells. It may be postulated that Kupffer cells may have been present, but their phagocytic function may have been impaired so that the mere presence of Kupffer cells may not have correlated with the conspicuity of a well-differentiated HCC on the ferumoxides-enhanced MR images. Some factor other than Kupffer cells may have been responsible, so that the high signal intensity of a nodule per se on precontrast MR images may persist on ferumoxides-enhanced MR images.
In some cases, the conspicuity of a nodule was better on the T2*-weighted images than on the T2-weighted fast spin-echo images (Fig 4). These findings made no difference in the moderately or poorly differentiated HCCs because these HCC nodules were generally clearly visible with any imaging sequence. However, there were two well-differentiated HCCs and one DN in which the nodules were visible on the T2*-weighted MR images but not on the T2-weighted fast spin-echo images.
The efficacy of ferumoxides-enhanced MR imaging combined with the most recent optimized sequences, such as gradient-echo, has been evaluated (3,8,19). Gradient-echo pulse sequences can be used to recruit the contributions of local field inhomogeneities (intravoxel dephasing) to T2* relaxations by a better contrast (21,22). Thus, in some studies (3,8,19), protocols of MR imaging with ferumoxides enhancement have included gradient-echo sequences. The use of appropriate gradient-echo pulse sequences may lead to a high level of sensitivity (13), which will help in the detection and characterization of well-differentiated HCCs or DNs at ferumoxides-enhanced MR imaging.
There may be well-differentiated HCCs or DNs that contain more Kupffer cells than does the surrounding parenchyma (3,5) and are depicted as having a slightly lower signal intensity than does the surrounding liver, standing out as a nodule. In the current study, two of the 10 well-differentiated HCCs and six of the 19 DNs had more Kupffer cells did than the surrounding cirrhotic hepatic parenchyma, although the differences were small and conspicuity on MR images was not sufficient for detection as a nodule with low signal intensity.
There were some limitations in our study. We included only patients who had cirrhosis and were suspected of having HCC on the basis of imaging findings. We did not know what the CNR was in other tumors or its relationship with the number of Kupffer cells. Our aim was to correlate the conspicuity of hepatocellular nodules at ferumoxides-enhanced MR imaging and the number of Kupffer cells in patients with cirrhosis, not to perform an efficiency study of a diagnostic modality.
The histologic characteristics of hepatocellular nodules may not be uniform. Sometimes histologic differentiation is present within a tumor, and, therefore, the signal-to-noise ratio of a portion of a nodule may not represent the whole nodule, and the number of Kupffer cells in a portion of a tumor may not represent the whole tumor. We measured the signal-to-noise ratio from ROIs covering the largest area of nodules on ferumoxides-enhanced MR images and counted the number of Kupffer cells from two areas in which Kupffer cells were at a relatively uniform density. Counting Kupffer cells in two areas and calculating an average may be sufficient. To perform a pilot study, we counted Kupffer cells in five representative areas in 10 cases, and the results were almost the same as those of calculating averages in two areas.
In summary, the number of Kupffer cells in moderately or poorly differentiated HCCs is profoundly decreased, and these nodules present with high signal intensity against the background low-signal-intensity hepatic parenchyma. The number of Kupffer cells in a majority of well-differentiated HCCs and most DNs is similar to that in the surrounding cirrhotic parenchyma or is slightly increased; the signal intensity in these lesions on ferumoxides-enhanced MR images is identical to or slightly lower than that in the liver. As a result, such nodules are mostly undetected. Therefore, the conspicuity of hepatocellular nodules in cirrhotic liver on ferumoxides-enhanced MR images depends on the difference in the number of Kupffer cells within a nodule and the surrounding cirrhotic hepatic parenchyma, and moderately or poorly differentiated HCCs can be distinguished from well-differentiated HCCs and DNs.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, J.H.L.; study concepts, J.H.L., C.K.P.; study design, J.H.L., D.C., C.K.P.; literature research, J.H.L., D.C., C.K.P.; clinical studies, J.H.L., D.C., S.H.K., W.J.L., S.W.P., Y.I.K.; data acquisition, D.C., S.K.C., C.K.P.; data analysis/interpretation, D.C., S.H.K., C.K.P.; statistical analysis, D.C., S.H.K.; manuscript preparation, J.H.L., D.C.; manuscript definition of intellectual content, J.H.L., S.H.K., H.K.L.; manuscript editing, J.H.L., D.C., H.K.L.; manuscript revision/review, J.H.L., H.K.L.; manuscript final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. F. Hanna, D. A. Aguirre, N. Kased, S. C. Emery, M. R. Peterson, and C. B. Sirlin Cirrhosis-associated Hepatocellular Nodules: Correlation of Histopathologic and MR Imaging Features RadioGraphics, May 1, 2008; 28(3): 747 - 769. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. F. Hanna, N. Kased, S. W. Kwan, A. C. Gamst, A. C. Santosa, T. Hassanein, and C. B. Sirlin Double-Contrast MRI for Accurate Staging of Hepatocellular Carcinoma in Patients with Cirrhosis Am. J. Roentgenol., January 1, 2008; 190(1): 47 - 57. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Choi, H. K. Lim, J.-W. Joh, S.-J. Kim, M. J. Kim, H. Rhim, Y.-s. Kim, B. C. Yoo, S. W. Paik, and C. K. Park Combined Hepatectomy and Radiofrequency Ablation for Multifocal Hepatocellular Carcinomas: Long-term Follow-up Results and Prognostic Factors Ann. Surg. Oncol., December 1, 2007; 14(12): 3510 - 3518. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. H. Lee, S. H. Kim, M.-Y. Cho, K. Y. Shim, and M. S. Kim Focal Nodular Hyperplasia-Like Nodules in Alcoholic Liver Cirrhosis: Radiologic-Pathologic Correlation Am. J. Roentgenol., May 1, 2007; 188(5): W459 - W463. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M.-J. Kim, J. H. Kim, J. Y. Choi, S. H. Park, J.-J. Chung, K. W. Kim, and D. G. Mitchell Optimal TE for SPIO-enhanced gradient-recalled echo MRI for the detection of focal hepatic lesions. Am. J. Roentgenol., September 1, 2006; 187(3): W255 - W266. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C-S Li, R-C Chen, H-Y Tu, L-S Shih, T-A Zhang, J-M Lii, W-T Chen, S-J Duh, and L-C Chiang Imaging well-differentiated hepatocellular carcinoma with dynamic triple-phase helical computed tomography Br. J. Radiol., August 1, 2006; 79(944): 659 - 665. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. J. Pijl, J. Doornbos, M. N. J. M. Wasser, H. C. van Houwelingen, R. A. E. M. Tollenaar, and J. L. Bloem Quantitative Analysis of Focal Masses at MR Imaging: A Plea for Standardization Radiology, June 1, 2004; 231(3): 737 - 744. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. K. Kang, J. H. Lim, S. H. Kim, D. Choi, H. K. Lim, W. J. Lee, and S. J. Lee Preoperative Depiction of Hepatocellular Carcinoma: Ferumoxides-enhanced MR Imaging versus Triple-Phase Helical CT Radiology, January 1, 2003; 226(1): 79 - 85. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. K. Kim, S. H. Kim, W. J. Lee, H. Kim, J. W. Seo, D. Choi, H. K. Lim, S. J. Lee, and J. H. Lim Preoperative Detection of Hepatocellular Carcinoma: Ferumoxides-Enhanced Versus Mangafodipir Trisodium--Enhanced MR Imaging Am. J. Roentgenol., September 1, 2002; 179(3): 741 - 750. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY |
