| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology, Thomas Jefferson University Hospital, 132 S 10th St, 1096 Main Bldg, Philadelphia, PA 19107 (K.I., D.G.M., T.G., P.N.K.); the Department of Radiology, Kanazawa University School of Medicine, Kanazawa, Japan (T.G.); the Department of Medicine, Jefferson Medical College, Philadelphia, Pa (H.W.L.H.); and the Department of Radiology, Yamaguchi University School of Medicine, Yamaguchi, Japan (T.F., H.A., K.H., N.M.). Received June 30, 1998; revision requested August 5; revision received September 29; accepted January 19, 1999. Address reprint requests to D.G.M. (e-mail: Donald.Mitchell@mail.tju.edu).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: In a retrospective search of MR imaging and histopathology records, 196 patients with histopathologically proved cirrhosis and with (n = 80) or without (n = 116) HCC who underwent T2-weighted conventional or fast spin-echo and gradient-echo (GRE) (echo time 
6.0 msec) imaging were identified. MR images were qualitatively and quantitatively evaluated for diffuse hepatic iron deposition and siderotic regenerative nodules to assess their correlation with the presence of HCC.
RESULTS: Hepatic parenchymal iron deposition was seen in 79 (40%) patients, and iron deposition in regenerative nodules was seen in 71 (36%) at MR imaging. The mean signal intensity ratio of GRE images in patients with hepatic iron deposition was significantly lower than that in patients without it (P < .001). The frequency of HCC in patients with iron deposition in regenerative nodules (52% [37 of 71 patients]) was significantly higher (P = .015) than that in patients without iron in regenerative nodules (34% [43 of 125 patients]).
CONCLUSION: The occurrence of HCC may be associated causally with iron deposition in regenerative nodules in patients with cirrhosis. MR imaging can enable detection of iron deposition in regenerative nodules as a possible risk factor for the development of HCC.
Index terms: Hepatitis, 761.291 • Liver, cirrhosis, 761.794 • Liver, iron content, 761.659 • Liver, MR, 761.121411, 761.121412, 761.121415, 761.12143 • Liver, nodules, 761.3198 • Liver, regeneration, 761.3198 • Liver neoplasms, diagnosis, 761.121411, 761.121412, 761.121415, 761.12143, 761.323
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Magnetic resonance (MR) imaging is an effective modality for evaluating patients with cirrhosis because of its capability to depict HCC and dysplastic nodules and to help assess the severity of cirrhosis (4–7). An additional advantage of MR imaging is its ability to help assess hepatic iron overload qualitatively and quantitatively in patients with cirrhosis or patients at risk for cirrhosis as well as in patients with ß-thalassemia major and hemochromatosis (8–11). Thus, MR imaging may play a role in the assessment of hepatic iron deposition as a possible risk factor for the development of HCC in patients with cirrhosis. To our knowledge, however, this potential role has not been evaluated. The purpose of our study was to determine whether the frequency of HCC in patients with cirrhosis is affected by hepatic iron deposition as detected at MR imaging.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Patients were excluded if (a) they had genetic hemochromatosis or Wilson disease, (b) MR images were degenerated by metallic artifacts due to surgical clips, (c) diffuse HCC was present throughout the liver, (d) MR studies could not be recovered from digital archive media, or (e) a hypervascular focal mass strongly suggestive of HCC but without histopathologic confirmation was identified on a contrast material–enhanced dynamic study. We included the last criterion to avoid having these cases categorized in the patient group without HCC, and 24 patients were excluded from this study on this basis. Consequently, this study population comprised 196 patients with cirrhosis and with HCC (n = 80) or without HCC (n = 116).
These patients had been referred for MR imaging for evaluation of the severity of cirrhosis and portal hypertension, for the preoperative evaluation of liver transplantation, or for the screening or further examination of hepatic lesions that were suspected with other imaging modalities. There were 121 men and 75 women (age range, 22–84 years; mean age, 53.2 years) with cirrhosis caused by viral infection (hepatitis B [n = 32], C [n = 93], or B and C [n = 10]), alcohol abuse (n = 19), and 
1-antitrypsin deficiency (n = 3); primary biliary cirrhosis (n = 6) and cryptogenic cirrhosis (n = 9); and undetermined causes of cirrhosis owing to insufficient clinical data (n = 24). Cirrhosis was histopathologically confirmed by means of percutaneous liver biopsy in 126 patients, liver transplantation in 27, transjugular liver biopsy in 20, surgical resection in 17, and autopsy in six. The diagnosis of HCC was proved at histopathologic examination in all 80 patients by means of percutaneous liver biopsy (n = 52), surgical resection (n = 17), liver transplantation (n = 9), and autopsy (n = 2).
MR Imaging Technique
MR imaging was performed with a 1.5-T unit (Signa, GE Medical Systems, Milwaukee, Wis; Magnetom H-15 or Vision, Siemens, Erlangen, Germany) with the use of either a whole-body coil (n = 41) or a phased-array torso coil (n = 155). Because the purpose of our study was to assess hepatic iron deposition, T2-weighted SE or fast SE images and GRE images (echo time 6.0 msec) were reviewed in all patients. These sequences have been relatively sensitive to the paramagnetic effects of hepatic iron among the MR images obtained at our routine abdominal examination.
Previous reports (9,10,12) showed a statistically significant correlation between the hepatic iron concentration revealed at MR imaging with T2-weighted or T2*-weighted sequences and the hepatic iron concentration value measured at biopsy, and they indicated that MR imaging with these sequences can be used to accurately quantify the amount of hepatic iron. T2-weighted imaging included conventional SE (1,500–3,000/50–100 [repetition time msec/echo time msec]), respiratory-triggered fast SE (3,000–7,500/91–144 [effective echo time]) with or without fat suppression, or breath-hold fast SE (2,500–4,200/70–138 [effective echo time]) sequences.
GRE images were obtained with a repetition time of 22–150 msec, an echo time of 6.0–9.9 msec, and a flip angle of 20°–60°. In addition, in-phase GRE (80–210/4.2, 90° flip angle), contrast-enhanced GRE (80–170/2.2–4.8, 70°–90° flip angle), and phase-contrast GRE (30–39/9.9, 30° flip angle) images with or without contrast enhancement were also reviewed as supplemental sequences, if available, although GRE images with shorter echo times generally are less sensitive in the detection of hepatic iron compared with T2-weighted SE and GRE (echo time 6.0 msec) sequences.
Contrast-enhanced GRE images were obtained after the administration of 0.1 mmol of gadopentetate dimeglumine (Magnevist, Berlex Laboratory, Wayne, NJ, or Japan-Schering, Osaka, Japan) per kilogram of body weight, but these images were not used to evaluate hepatic iron. Other imaging parameters included a 256 or 512 x 128–256 imaging matrix, usually with use of a rectangular field of view to reduce the number of phase-encoding views, and a 5–10-mm section thickness with a section gap of 2 mm or less.
Image Interpretation
MR images of each patient were evaluated on a computer monitor (SPARC 20; Sun Microsystems, Palo Alto, Calif). Three representative images per sequence obtained at different levels (eg, upper, middle, and lower level) of the liver were given to the readers (K.I., D.G.M., T.G.). Consequently, the total image number was six to 15 per patient—in most patients, nine or 12 images. The images in which HCC was seen were not included because knowledge of the presence of HCC might bias the readers' evaluation for hepatic iron deposition. Thus, the readers were blinded with regard to the presence of HCC in each patient during their reading. The images obtained in patients with HCC or without HCC were intermixed randomly and presented to the readers.
MR images were qualitatively evaluated for (a) hepatic parenchymal iron deposition and (b) iron deposition in regenerative nodules to assess the correlation with the presence of HCC. Hepatic parenchymal iron deposition was considered present when the signal intensity of the liver diffusely decreased and was scored with use of the following four-point severity scale: "0" was defined as absent if the signal intensity of the liver was greater than that of skeletal muscle; "1," as mild if the signal intensity of the liver was equal to or slightly less than that of skeletal muscle; "2," as moderate if the signal intensity of the liver was moderately less than that of skeletal muscle; and "3," as severe if the signal intensity of the liver was markedly less than that of skeletal muscle. We used skeletal muscle as a reference tissue according to the findings of prior studies (10,11,13) because iron tends not to accumulate in skeletal muscle in patients with iron overload. Iron deposition in regenerative nodules was scored "0" for absent, "1" for probably present, and "2" for definitely present. Siderotic regenerative nodules were defined as small nodules with low signal intensity compared with surrounding hepatic parenchyma on GRE images (14,15).
Images were reviewed retrospectively and independently by three radiologists (K.I., D.G.M., T.G.) experienced in the interpretation of MR images of the liver. When there was disagreement in their opinions regarding the absence or presence of hepatic iron deposition, a majority opinion was used as the final decision. When hepatic iron deposition was determined to be present, a mean score of the three observers' readings was used for further analysis.
For the quantitative evaluation, region-of-interest analysis of the images was performed by a single observer (K.I.). Region-of-interest measurements were made of the liver and the right paraspinous muscles by using the T2-weighted SE and GRE (echo time 6.0 msec) images; the same location on the images was used for the measurements. Signal intensities were measured on the posterior part of the right hepatic lobe adjacent to the right paraspinous muscles because the signal intensity of the central portion of the liver often decreased on MR images obtained with a phased-array torso coil. Vessels were excluded from the region of interest as much as possible. The liver-to-muscle signal intensity ratio was calculated from region-of-interest measurements. The signal intensity ratio was defined as the signal intensity of the liver divided by the signal intensity of the paraspinous muscle.
Statistical Analysis
The 
2 test was used to compare the difference of the frequency of HCC between the patients with and the patients without hepatic iron deposition. Comparisons of the signal intensity ratios were performed with an unpaired two-tailed Student t test between subgroups with and subgroups without hepatic iron deposition or between those with and those without HCC. Data were presented as the mean ± SD. A P value of less than .05 was considered to indicate a statistically significant difference. To assess interobserver variability in the reading, 
values were calculated for the three readers (16). The level of agreement was defined as follows: no agreement, < 0; poor agreement, = 0.01–0.40; good agreement, = 0.41–0.75; and excellent agreement, = 0.76–1.00.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
values for MR imaging findings showed good agreement (0.50–0.72). At visual inspection, the signal intensity of the liver was equal to or less than that of skeletal muscle in 79 (40%) of the 196 patients, which indicates the presence of hepatic parenchymal iron deposition (Figs 1, 2). The mean visual score in these patients was 1.22 ± 0.67. Iron deposition in regenerative nodules was seen in 71 (36%) of the 196 patients, with a mean visual score of 1.42 ± 0.50 (Fig 3). Iron deposition in both the hepatic parenchyma and regenerative nodules was seen in 43 (22%) of the 196 patients. In the quantitative analysis, the 79 patients with visible hepatic parenchymal iron deposition had significantly lower mean signal intensity ratios than did the 117 patients without it on both T2-weighted SE (0.97 ± 0.32 vs 1.85 ± 0.97, P < .001) and GRE images (0.76 ± 0.29 vs 1.25 ± 0.42, P < .001). The 71 patients with siderotic regenerative nodules on GRE images had significantly lower mean signal intensity ratios than did the 125 patients without them (0.82 ± 0.39 vs 1.13 ± 0.45, P < .001). There was no significant difference in the mean signal intensity ratio between the 80 patients with HCC and the 116 patients without HCC (Table 1).
|
|
|
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Siderotic regenerative nodules are seen frequently in cirrhotic livers at MR imaging (14,15). In the current study, iron deposition in regenerative nodules was seen in 71 (36%) patients at visual inspection. The mean signal intensity ratio in these 71 patients was significantly lower than that in 125 patients without siderotic regenerative nodules on GRE images, probably owing to the inclusion of microscopic siderotic nodules within measured regions.
Our methodology did not include histopathologic proof that decreased signal intensity is caused by hepatic iron deposition. However, similar imaging techniques have been used to detect hepatic iron overload and siderotic regenerative nodules, and results have been confirmed with quantitative histopathologic measurements (9,10,15). Because of little attention to hepatic iron deposition in many histopathology reports, limited availability of histopathologic specimens, and the small number of iron stains and quantitative tissue iron studies of hepatic specimens, direct correlation between findings in MR studies and at histopathologic examination regarding the degree and distribution of hepatic iron deposition was difficult and was not performed in this study.
Excessive tissue iron is potentially toxic and mutagenic and can lead to malignant transformation, and the liver is especially vulnerable (18). Authors of prior reports (19–21) have shown an association between HCC and the histopathologic findings of hepatic iron excess, and they have indicated that hepatic iron overload is a risk factor for developing HCC, even in the absence of genetic hemochromatosis. In the current study, the frequency of HCC in patients with siderotic regenerative nodules visible on MR images was significantly higher than that in patients without iron in regenerative nodules.
Our results suggest that the occurrence of HCC may be causally related to iron concentration within regenerative nodules that are large enough to be visible on MR images and that MR imaging can depict these siderotic regenerative nodules as a marker of HCC development. However, depiction of these siderotic regenerative nodules is not a mandatory process leading to the development of HCC, because there were some patients with HCC but without visible siderotic regenerative nodules. In our analysis of subgroups (n = 71) with visible iron deposition in regenerative nodules (Table 3), there was no significant difference in the signal intensity ratio between patients with HCC (n = 37) and those without HCC (n = 34). Therefore, it seems that the severity of diffuse hepatic parenchymal iron deposition is unlikely to affect the prevalence of HCC among patients with visible siderotic regenerative nodules.
Prior reports offer two explanations as to how increased hepatic iron may contribute to hepatocarcinogenesis. One possible direct mechanism is that free cellular iron may induce mutations by generating reactive oxygen species. A number of DNA changes are produced by reactive oxygen species (22). An indirect possibility is that excess hepatic iron may facilitate the persistence of chronic hepatitis B and C viral infections, both of which are major risk factors for the development of HCC (21).
Histopathologic observations further support a role for iron-accumulative regenerative nodules in carcinogenesis. Authors of previous reports (23–25) have described the emergence of hyperplastic or malignant foci with atypical hepatocytes within large regenerative nodules, which suggests incipient neoplastic hepatocellular lesions. In one autopsy series (26), the frequency of hyperplastic foci within large regenerative nodules in patients with cirrhosis was higher in patients with iron-accumulative nodules versus those without iron-accumulative nodules (57% [20 of 35] vs 31% [20 of 64]). Iron concentration within regenerative nodules may expose hyperplastic or dysplastic foci to the tumor-stimulating effects of iron. Malignant foci within iron-accumulative regenerative nodules can show a nodule-within-nodule appearance on MR images (27) because these malignant foci do not contain iron (28). Therefore, MR imaging may be used to follow up patients with cirrhosis with iron-accumulative regenerative nodules for the early detection of malignant foci.
It is unclear why a subset of regenerative nodules in some patients with cirrhosis tends to accumulate iron. Because hepatocellular iron uptake is largely mediated by transferrin receptor proteins (29), these proteins may be abnormal in iron-accumulative regenerative nodules. The iron in regenerative nodules has been shown to disappear with the development of iron deficiency anemia (30), which suggests that iron deposition in regenerative nodules can be reversible and that the intralesional iron is metabolically available for redistribution for hematopoiesis. Therefore, phlebotomy may reduce iron deposition in regenerative nodules and potentially decrease the risk of HCC. MR imaging could help evaluate the changes of iron in regenerative nodules during the clinical course as well as after phlebotomy.
Our study has some limitations. First, the MR imaging parameters and sequences differed between patients. Recent technical improvements have allowed the use of more efficient pulse sequences, including fast SE and GRE imaging with a short echo time. Fast SE sequences are less sensitive than conventional SE sequences to magnetic field heterogeneities because of the use of repeated 180° refocusing pulses, which probably results in overlooking mild hepatic iron overload. GRE sequences with an echo time longer than 15 msec are considered preferable for detecting slight hepatic iron (12). More standardized pulse sequences optimized for iron sensitivity may have resulted in the detection of more cases of mild hepatic iron deposition. However, it was not practical to include these less efficient sequences in the current routine abdominal protocol.
Second, we could not exclude the possibility of small HCCs, if they were not seen on the images, being categorized in the group of patients without HCC. It is not possible to reliably exclude this possibility without histopathologic investigation of the whole hepatic specimen. However, these limitations do not reduce the validity of the correlation we observed between visible siderotic regenerative nodules and the frequency of HCC.
In conclusion, the occurrence of HCC may be associated causally with iron deposition in large regenerative nodules in patients with cirrhosis, and MR imaging is a useful modality for assessing iron deposition in these regenerative nodules as a possible risk factor for the development of HCC.
|
Footnotes |
|---|
Author contributions: Guarantors of integrity of entire study, K.I., D.G.M.; study concepts, K.I., D.G.M., H.W.L.H.; study design, K.I.; definition of intellectual content, K.I.; literature research, K.I., P.N.K.; clinical studies, D.G.M., H.W.L.H., T.F., H.A., K.H., N.M.; data acquisition, K.I., D.G.M., T.G.; data analysis, K.I., D.G.M.; statistical analysis, K.I.; manuscript preparation, K.I.; manuscript editing, D.G.M.; manuscript review, D.G.M., T.G., H.W.L.H.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. M. Willatt, H. K. Hussain, S. Adusumilli, and J. A. Marrero MR Imaging of Hepatocellular Carcinoma in the Cirrhotic Liver: Challenges and Controversies Radiology, May 1, 2008; 247(2): 311 - 330. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. A. Krinsky and G. Israel Nondysplastic Nodules That Are Hyperintense on T1-Weighted Gradient-Echo MR Imaging: Frequency in Cirrhotic Patients Undergoing Transplantation Am. J. Roentgenol., April 1, 2003; 180(4): 1023 - 1027. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. L. Baron and M. S. Peterson From the RSNA Refresher Courses: Screening the Cirrhotic Liver for Hepatocellular Carcinoma with CT and MR Imaging: Opportunities and Pitfalls RadioGraphics, October 1, 2001; 21(90001): S117 - 132. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. A. Krinsky, V. S. Lee, M. T. Nguyen, N. M. Rofsky, N. D. Theise, G. R. Morgan, L. W. Teperman, and J. C. Weinreb Siderotic Nodules in the Cirrhotic Liver at MR Imaging with Explant Correlation: No Increased Frequency of Dysplastic Nodules and Hepatocellular Carcinoma Radiology, January 1, 2001; 218(1): 47 - 53. [Abstract] [Full Text]
|
|
|
|
|
|
H. Okazaki, K. Ito, T. Fujita, S. Koike, K. Takano, and N. Matsunaga Discrimination of Alcoholic from Virus-Induced Cirrhosis on MR Imaging Am. J. Roentgenol., December 1, 2000; 175(6): 1677 - 1681. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
