- Email: [email protected]
Detection of hepatocellular carcinoma by ferucarbotran-enhanced magnetic resonance imaging: the efficacy of accumulation phase fat-suppressed T1-weighted imaging
Clinical Radiology (2009) 64, 22e29
Detection of hepatocellular carcinoma by ferucarbotran-enhanced magnetic resonance imaging: the efficacy of accumulation phase fat-suppressed T1-weighted imaging C.-T. Choua,b, R.-C. Chena,c,*, W.-T. Chenc, J.-M. Liic a
Department of Biomedical Imaging and Radiological Science, National Yang-Ming Medical University, Department of Radiology, Chang-Hua Christian Hospital, Erlin Branch, and cDepartment of Radiology, Taipei City Hospital, Renai Branch, Taiwan
b
Received 18 March 2008; received in revised form 18 June 2008; accepted 11 July 2008
AIM: To evaluate the effectiveness of accumulation phase, fat-suppressed, T1-weighted imaging (FS-T1WI) when detecting hepatocellular carcinoma (HCC) by ferucarbotran-enhanced magnetic resonance imaging (MRI). MATERIALS AND METHODS: Thirty patients who underwent ferucarbotran-enhanced MRI, which resulted in 35 confirmed HCCs, were included in this prospective study. Two image sets were prepared and two radiologists independently reviewed these in two reading sessions; set A was without contrast-enhanced accumulation phase FS-T1WI and set B included contrast-enhanced accumulation phase FS-T1WI. All HCCs had been confirmed by operation (n ¼ 4), by biopsy (n ¼ 28), and by follow-up study for at least 1 year (n ¼ 3). RESULTS: The contrast-to-noise ratio significantly increased from 1.2 7.5 to 12.7 7.3 with contrast-enhanced accumulation phase FS-T1WI, but was only slightly increased from 12.2 10.3 to 15.5 12.2 with contrast-enhanced T2WI (p < 0.001). The signal-to-noise ratio (SNR) was decreased with T1WI and T2WI for liver parenchyma. With T2WI, the SNR for HCCs was decreased; however, it was slightly increased with T1WI (p < 0.001). Overall, 29 HCCs were detected using set A, and 35 nodules were identified using set B, which included the contrast-enhanced accumulation phase FS-T1WI. Thus, the detection rate significantly increased using post-contrast medium accumulation phase FS-T1WI (p < 0.05). CONCLUSION: Due to the improved CNR with the post-contrast medium accumulation phase FS-T1WI, which helped to increase HCC detection, accumulation phase FS-T1WI is recommended as one of the routine protocols for inclusion in HCC detection. ª 2008 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Hepatocellular carcinoma (HCC) is the most common primary hepatic malignancy, and the detection of synchronous HCCs or intrahepatic metastasis is essential for treatment planning and prognosis.1,2 Superparamagnetic iron oxide (SPIO) is a liver-specific particulate magnetic * Guarantor and correspondent: R.-C. Chen, Department of Radiology, Taipei City Hospital, Renai Branch, No. 10, Sec 4, Renai Road, Taipei 106, Taiwan. Tel.: þ886 2 27093600x5103; fax: þ886 2 27040013. E-mail address: [email protected] (R.-C. Chen).
resonance imaging (MRI) contrast agent that is primarily taken up by the Kupffer cells of the liver and is useful for tumour detection.3 Ring enhancement of a tumour on contrast-enhanced T1-weighted imaging (T1WI) is known to be highly suggestive of malignancy when using ferumoxideenhanced MRI.4 Contrast-enhanced T1WI with fat suppression (FS) is superior to T1WI without FS when detecting small HCCS.5 Therefore, it was hypothesized that ferucarbotran-enhanced accumulation phase T1WI with FS imaging (FS-T1WI) ought to improve HCC detection. The effectiveness of this procedure for detecting HCCs has not been investigated. The purpose of the present
0009-9260/$ - see front matter ª 2008 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2008.07.010
Detection of HCC by ferucarbotran-enhanced accumulation phase FS-T1WI
study was to evaluate the effectiveness of ferucarbotran-enhanced accumulation phase FS-T1WI when used as part of a HCC-detection protocol.
Materials and methods Patients This study was approved by the institutional review board of our hospital. Written informed consent related to undergoing ferucarbotran-enhanced MRI was obtained from all patients before they were enrolled in the study. A total of 39 patients, with clinical findings of chronic liver disease or an elevated serum alpha-fetoprotein (AFP) level and those suspected of having HCC based on ultrasonography, underwent ferucarbotran-enhanced MRI as part of their diagnostic work-up. One patient refused a histological examination and was excluded from the study. In total, 38 out of 39 patients received histological examination. Eight out of the 38 patients (seven men, one woman; age range 38e73 years) were identified to have benign nodules (one inflammatory pseudotumour, one hyperplastic nodule, and six dysplastic nodules). The inflammatory pseudotumour had disappeared at the 1 year follow-up. The patient with a dysplastic nodule was lost to follow-up. The other six nodules had remained unchanged at the 1 year follow-up based on computed tomography (CT) or MRI studies. Finally, 30 of the 39 patients (20 men, 10 women; age range 28e85 years) with histologically proven HCC were included in the study and were followed up for at least 1 year. Twenty-six patients had undergone needle biopsy and four had undergone surgical resection. Among them, 25 patients had liver cirrhosis and all had chronic hepatitis (hepatitis B, 21 patients; hepatitis C, six patients; hepatitis B and C, three patients). A total of 35 HCC nodules in 30 patients were identified. Twenty-five patients had solitary HCCs and five patients had two HCCs. Among the patients who had two hepatic nodules, at least one of the lesions was biopsied and examined histologically. Thirty-two HCCs were eventually confirmed histopathologically and these consisted of 14 welldifferentiated HCCs, 15 moderately differentiated HCCs, and three poorly differentiated HCCs. When the other three hepatic lesions without histological examination were considered, two showed lipiodol retention in the tumours on follow-up CT imaging after transcatheter embolization for HCC. One of the three tumours without lipiodol retention had progressively enlarged on the 3 month and 1 year
23
follow-up CT images, and the tumour showed focal arterial enhancement on the 1 year follow-up CT image. Based on these clinical manifestations and imaging findings, these three tumours were also considered to be HCC.
MRI MRI was performed using a 1.5 T MRI system (Gyroscan ACS-NT, Philips, Best, The Netherlands) equipped with a body phased-array coil. Unenhanced MRI was performed with T2-weighted (T2W) turbo spinecho (TSE) axial imaging [repetition time/echo time (TR/TE) 2500/90 ms; TSE factor 23; section thickness 8 mm, gap 0.8 mm] with and without FS under respiratory trigger; T2* echo planar imaging (EPI; TR/TE 500/13.8 ms) was obtained during one breath-hold. T1W dual echo imaging (TR/TE: 210/2.3 ms and 4.6 ms; section thickness 8 mm, gap 0.8 mm) and T1W with FS imaging (TR/TE 241e344/2.7 ms; flip angle, 70 ; section thickness 8 mm, gap 0.8 mm) was also performed during one or two breath-holdings depending on the liver size of the patient. Automatic shimming was applied for FS imaging to maximize magnetic field homogeneity. Flow compensation was also used. The ferucarbotran (Resovist, Schering, Berlin, Germany) consists of SPIO microparticles coated with carboxydextran and this was preloaded into 1.4 ml (>50 kg body weight) using a connecting intravenous tube; the agent was then manually injected rapidly through a filter with a 5 mm pore size (3). Dynamic T1W fast field echo (FFE) imaging (175e210/ 1.3e2.1; flip angle 80 ) was carried out before and at 18e20 s and 50e55 s after the contrast agent was injected. Approximately 10 min after intravenous injection of the SPIO agent, the contrast-enhanced images, including the T2WI with and without FS, T2*EPI were obtained in the same order as used for the unenhanced imaging acquisition. Finally, accumulation phase FS-T1WI (TR/TE 241e344/2.7 ms; flip angle 70 ; section thickness 8 mm, gap 0.8 mm) was performed.
Qualitative analysis The imaging analysis was performed at a dualscreen diagnostic workstation (GE Healthcare, Milwaukee, WI, USA). Two image sets (set A and set B) were prepared in which set A contained unenhanced T1WI with dual echo and FS-TIWI, T2WI with and without FS, T2*EPI, dynamic T1WI, contrast-enhanced T2WI with and without FS and T2* EPI images. Set B consisted of all of the images of set A plus a contrast-enhanced accumulation
24
phase FS-T1WI. Two observers experienced in abdominal MRI then independently reviewed image set A and then, 4 months later, they reviewed set B; the observers carried out the reviews separately at two different reading sessions. Both observers were blinded to the clinical information and final diagnosis. Both image sets of the whole liver were analysed using visual assessment by the two radiologists with the aim of estimating the number and location of HCC nodules based on the following criteria. A tumour presenting with the following characteristics was considered as HCC: first, hyperintense on the T2WI associated with a variety of signal intensities on T1WI; second, hyperintense on the T1WI with fatty metamorphosis detected on the dual echo T1WI; third, positive enhancement on the dynamic MRI images associated with a variety of signal intensities on T1WI, and finally, positive enhancement on the contrast-enhanced accumulation phase FS-T1WI.3,6,7 The observers recorded the HCC nodules using reading set A based on the first, second, and third diagnostic criteria. However, when using set B, all four criteria were included. Any other nodules were ignored in this study. The enhancement pattern of the HCC was visually classified into the following patterns: ring enhancement, homogeneous enhancement, heterogeneous enhancement, and no enhancement. Differences between the two reviewers were resolved by a third radiologist who joined a consensus conference and who was also blind to the clinical information.
Quantitative analysis On the unenhanced and ferucarbotran-enhanced MRI images, all HCCs were quantitatively evaluated for signal intensity (SI) and compared with normal liver parenchyma by one radiologist. Operatordefined region-of-interest SI measurements were taken directly from the monitor for the lesion and the normal liver parenchyma. As the SI of the normal liver parenchyma was measured, care was taken to avoid the major hepatic and portal vessels. The standard deviation (SD) of the background noise was measured by placing the region-ofinterest anterior to the patient near the liver parenchyma and by calculating the SD including motion artefacts. Similar regions of interest were used on the unenhanced and ferucarbotranenhanced images. Signal-to-noise ratios (SNRs) for the HCCs and the liver were calculated for all MRI sequences using the following formula: SNR ¼ SI=SDbackground
C.-T. Chou et al.
where SI is the SI of the lesion or liver parenchyma and SDbackground is the SD of the air background noise. In addition, contrast-to-noise ratios (CNRs) for the lesion to liver were computed for each sequence as follows: CNR ¼ ðSI lesion SI liverÞ=SDbackground These values are negative for hypointense lesions and positive for hyperintense lesions.
Statistical analysis The results are expressed as means SD. Statistical differences in the detection numbers for both imaging sets were compared by the McNemar test. A paired t-test was used to confirm the difference in the SNR/CNR for the unenhanced and contrastenhanced imaging. A p-value of less than 0.05 was considered statistically significant. A Kappa statistic was obtained to evaluate the lesion detection agreement between the two reading sets.8
Results Tumour detection Overall, a consensus reading was necessary for three of the 35 HCCs when analysing imaging set A, and was necessary for two HCCs when using imaging set B. In terms of HCC detection, there was good inter-observer agreement between the two readers with both set A (k ¼ 0.79) and set B (k ¼ 0.91). A total of 35 HCCs (mean size 2 1.1 cm) were found using imaging set B, but only 29 HCCs were identified using reading set A. Among the six HCC nodules (mean size 1.6 0.4 cm) only identified in set B, (Table 1), there were two that were well-differentiated (Fig. 1), one that was moderately differentiated, and the other three tumours did not have a histological diagnosis (Fig. 2). Overall, the detection of small HCCs was significantly increased with the help of the ferucarbotran-enhanced accumulation phase FS-T1WI (p < 0.05). Among the benign nodules, two nodules (one inflammatory pseudotumour and one dysplastic nodule) were considered as HCCs when identified by both imaging set A and set B. Both of the two nodules showed hyperintensity on T2WI, with arterial enhancement during the contrast-enhanced dynamic study and hyperintensity on the contrastenhanced T2WI and accumulation phase FS-T1WI. The other six nodules were considered to be benign cirrhotic nodules by the both imaging set A and set B.
Detection of HCC by ferucarbotran-enhanced accumulation phase FS-T1WI
Table 1
The additional six HCC nodules found on the imaging set B
Size (cm) Unenhanced Imaging
1 2 3 4 5 6
1.8 1 1.6 1.2 2.1 1.7
25
Contrast-enhanced Imaging
PSIL (T2WI) Imaging diagnosis
T1WI
T2WI
Dynamic T2WI enhancement
Delay FS-T1WI
Hyperintense Hyperintense Isointense Hyperintense Hyperintense Hyperintense
Hypointense Hypointense Hypointense Isointense Hypointense Hypointense
No No No No No No
Yes Yes Yes Yes Yes Yes
Hypointense Hypointense Hypointense Isointense Hypointense Hypointense
40.3% 37.0% 49.2% 65.5% 58.6% 48.8%
Set A
Set B
DN DN HN DN DN DN
HCC HCC HCC HCC HCC HCC
Pathology
Well Nonea Noneb Well Moderately Nonea
T1WI, T1-weighted imaging; T2WI, T2-weighted imaging; FS, fat saturated; PSIL, percent of signal intensity loss; DN, dysplastic nodule; HN, hyperplastic nodule; Well, well-differentiated HCC, Moderately, moderately-differentiated HCC, Yes, positive enhancement. a Lipiodol retention within the tumour was noted on the follow-up CT images after transarterial embolization treatment. b Progressively enlarged on follow-up CT studies and typical HCC enhanced pattern on 1 year follow-up CT study.
Qualitative analysis The signal intensity of the HCC was compared to the surrounding normal parenchyma, and the results are displayed in Table 2. According to the enhancement pattern, 29 of 35 HCCs were homogeneous,
three showed heterogeneous enhancement on the accumulation phase FS-T1WI, and three were isointense relative to the liver parenchyma. The three HCCs with heterogeneous enhancement on the accumulation phase FS-T1WI were also heterogeneous on the contrast-enhanced T2WI.
Figure 1 A well-differentiated HCC (arrow) proven by biopsy in a 71-year-old woman. (a) The unenhanced T2WI image shows a hypointense nodule beside a hepatic cyst on S8 of the liver. (b) The unenhanced FS-T1WI image shows a hyperintense lesion in the same area. (c) The tumour is obscured in the contrast-enhanced T2WI image. (d) Positive enhancement of the tumour is clearly depicted on the contrast-enhanced accumulation phase FS-T1WI image.
26
C.-T. Chou et al.
Figure 2 A nodule (arrow) at S5/8 of the liver was considered to be HCC in a 68-year-old man. (a) The unenhanced FS-T2WI image shows a hypointense nodule in S5/8 of liver. (b) The unenhanced FS-T1WI shows no obvious tumour in the same area. (c) Persistent hypointensity of the tumour is depicted in the contrast-enhanced FS-T2WI image. (d) A hyperintense tumour is clearly depicted on the accumulation phase FS-T1WI image. (e) The arterial phase of the 1 year follow-up CT reveals an enlargement of the nodule (from 1.6 to 2.2 cm) and focal arterial enhancement with a nodule-in-nodule appearance (arrowhead).
Quantitative analysis The SNRs and CNRs were quantitatively analysed using T1WI and T2WI before and after ferucarbotran administration. The difference in SNR/CNR for unenhanced and contrast-enhanced imaging and the altered SNR/CNR between T1WI and T2WI are shown on Table 3. The SNRs of the HCCs showed a slight increase in mean value on FS-T1WI, but a decrease on the T2WI (p < 0.001). The HCC-to-liver CNRs were
calculated and these showed a significant increase in mean value on the FS-T1WI, but only a slight increase on the T2WI. (p < 0.001; Fig. 3)
Discussion It is extremely important to detect small, multifocal HCCs because this will affect the decision to proceed to surgical resection or transplantation,
Detection of HCC by ferucarbotran-enhanced accumulation phase FS-T1WI
27
Table 2 Numbers of HCC with variable signal intensity relative to normal liver parenchyma on pre- and post-Ferucarbotran contrast-enhanced T1- and T2-weighted images
T1WI
T2WI
Hyperintense Hypointense Isointense Hyperintense Hypointense Isointense
Unenhanced Imaging
Contrast-enhanced Imaging
12 18 5 27 5 3
32 0 3 29 4 2
which is based on the numbers and sizes of the lesions.9 SPIO-enhanced MRI of the liver has been intensely investigated and it is now widely accepted that SPIO is able to increase the detection rate of focal lesions using T2WI.10e12 However, several authors have encountered difficulties when detecting HCC due to the presence of Kupffer cells in the tumour.6,13,14 Lee et al.15 described how SPIO-enhanced T1WI was useful for lesion characterization only, but not lesion detection. In the present results, hyperintensity on the contrast-enhanced accumulation phase FS-T1WI was readily able to identify six additional HCC nodules compared with contrast-enhanced T2WI. One of the six HCCs showed up as isointense on T1WI, was hypointense on T2WI, and was considered to be a cirrhotic nodule on imaging set A; however, this tumour showed homogeneous enhancement on the delay FS-T1WI and was subsequently considered to be HCC. (Fig. 2) SPIO particles have both high T1 and T2 relaxivities, and both circulating and clustered particles may, therefore, theoretically contribute to either the positive or negative enhancement of liver tissue.16 There are three possible mechanisms by which HCC may present with hyperintensity on an accumulation phase FS-T1WI. The first reason is tumour enhancement by the T1 effect of the residual
Figure 3 (a) The changes in the SNR of HCCs and normal liver parenchyma on the T1WI and T2WI images before and after administration of ferucarbotran are presented. (b) Elevation of the HCC-to-liver CNRs on the T1WI images was significantly greater than with the T2WI images (p < 0.001).
circulating SPIO particles.17,18 The second is ferucarbotran uptake by the Kupffer cells and as a result the clustered particles produce a T2* effect that might influence the SI on T1WI. In the present results, there was signal loss by the liver parenchyma and this was greater than with the HCC nodules, which caused a dark background on the accumulation phase FS-T1WI. The third is that the FS used for the T1WI in the present study might provide a higher lesion-to-liver contrast,5 and
Table 3 The signal-to-noise ratios (SNR) of HCCs/liver parenchyma and the contrast-to-noise ratios (CNR) of HCCs for unenhanced and contrast-enhanced T1W/T2W images T1WI
SNR (mean SD)
CNR (mean SD) a b
p < 0.05. p > 0.05.
T2WI
unenhanced
Contrast enhanced
Contrast enhanced e unenhanced difference
unenhanced
Contrast enhanced
Contrast enhanced e unenhanced difference
HCC 35.3 9.6
36.3 11
1 9.8b
31.6 11.2
24.6 12.3
7 8.1a
Liver 36.6 11.3 23.6 7.7
12.9 9.9a
19.4 7.5
9.1 4.4
10.4 5.6a
HCC 1.2 7.5
14 10.4a
12.2 10.3
15.5 12.2
3.4 7.1a
12.7 7.3
T1WIpost-pree T2WIpost-pre difference
8 12.4 P < 0.001 2.6 10.2 p > 0.05 10.6 11.5 p < 0.001
28
therefore, some HCCs became more conspicuous on the contrast-enhanced accumulation phase FST1WI than that on the contrast-enhanced T2WI. Two benign nodules presented with hyperintensity on the accumulation phase FS-T1WI and both nodules also presented hyperintensity on the T2WI and with arterial enhancement during the dynamic studies. Both nodules were considered to be HCCs using both imaging set A and set B. Furthermore, three HCCs were isointense and were not found on the contrast-enhanced accumulation phase FS-T1WI in the present study, but were well depicted on the contrast-enhanced T2WI. Although the accumulation phase FS-T1WI was effective for the detection of HCCs, the delay FS-T1WI is not able to completely replace the contrast-enhanced T2WI when carrying out ferucarbotran-enhanced MRI and the probability of benign nodules presenting with hyperintensity on the accumulation phase FS-T1WI should be kept in mind. The mean value of the HCC-to-liver CNRs showed a significant increase on the mean value FS-T1WI. In contrast, the CNRs were only slightly increased on the T2WI. Thus, the improvement in CNR for the tumours is more prominent using T1WI rather than T2WI after ferucarbotran administration. This may explain the better detection rate for small HCCs on the contrast-enhanced accumulation phase FS-T1WI (Fig. 3b). Peritumoural ring enhancement has been described as an important sign in distinguishing benign from malignant lesions when using SPIOenhanced MRI.19,20 The mechanism of HCC ring enhancement on SPIO-enhanced T1WI has been discussed by Tanimoto et al.13 The frequency of HCC presenting with ring enhancement on contrast-enhanced T1WI images varies from 10% in the study by Lee et al.15 to 72% in the study by Tanimoto et al.3,13 When hepatic tumours presented with hyperintensity, ring enhancement was not readily visualized.20 Ring enhancement tends to occur with larger tumours rather than small tumours.13 In the present study, 82% of the HCCs were homogeneous and 9% demonstrated heterogeneous enhancement on the accumulation phase FS-T1WI; it was notable that the mean size of the HCCs was small (2 1.1 cm). The small size might explain why none of the 35 HCCs in the present study showed ring enhancement. The scan time for accumulation phase FS-T1WI with breath-holding is short (18e21 s for one breath-holding and 28e31 s for two breathholding) and can be performed immediately after the contrast-enhanced T2WI. It is very convenient to obtain an accumulation phase T1WI without
C.-T. Chou et al.
additional body positioning or hardware preparation. In the present study, six additional tumours were detected by adding accumulation phase FS-T1WI. One limitation of the present study was that not all HCCs (32/35) were proven histologically. This is a known shortcoming in most comparative studies.21,22 The major problem is that ultrasound or CT-guided biopsy is impossible for nodules that are only visible on an MRI image. Furthermore, it may not be practical to biopsy all the hepatic nodules in patients with multiple HCC nodules. In conclusion, due to the better CNR on the contrast-enhanced accumulation phase FS-T1WI, which increased HCC detection, accumulation phase FS-T1WI is recommended as part of the routine protocol for HCC detection.
References 1. Szklaruk J, Silverman PM, Charnsangavej C. Imaging in the diagnosis, staging, treatment and surveillance of hepatocellular carcinoma. AJR Am J Roentgenol 2003;180: 441e54. 2. Noone TC, Semmelka RC, Chaney DM, et al. Abdominal imaging studies; comparison of diagnostic accuracies resulting from ultrasound, computed tomography, and magnetic resonance imaging in the same individual. Magn Reson Imaging 2004;22:19e24. 3. Tanimoto A, Kuribayashi S. Application of superparamagnetic iron oxide to imaging of hepatocellular carcinoma. Eur J Radiol 2006;58:200e16. 4. Kanematsu M, Matsuo M, Yamada Y, et al. Perilesional hyperintense rim of malignant hepatic tumours on ferumoxide-enhanced T1-weighted gradient-echo MR images: correlation between MR imaging and histopathologic findings. J Magn Reson Imaging 2003;18:40e8. 5. Koushima Y, Ebara M, Fukuda H, et al. Small hepatocellular carcinoma: assessment with T1-weighted spin-echo magnetic resonance imaging with and without fat suppression. Eur J Radiol 2002;41:34e41. 6. Imai Y, Murakami T, Yoshida S, et al. Superparamagnetic iron oxide-enhanced magnetic resonance images of hepatocellular carcinoma: correlation with histological grading. Hepatology 2000;32:205e12. 7. Efremidis SC, Hytiroglou P, Matsui O. Enhancement patterns and signal-intensity characteristics of small hepatocellular carcinoma in cirrhosis: pathologic basis and diagnostic challenges. Eur Radiol 2007;17:2969e82. 8. Landis JR, Koch GG. The measurement of observer agreement for categrorical data. Biometrics 1977;33:159e74. 9. Mazzaferro V, Regalia E, Doci R, et al. Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis. N Engl J Med 1996;334:693e9. 10. Reimer P, Jahnke N, Fiebich M, et al. Hepatic lesion detection and characterization: value of nonenhanced MR imaging, superparamagnetic iron oxide-enhanced MR imaging, and spiral CT-ROC analysis. Radiology 2000;217:152e8. 11. Poeckler-Schoeniger C, Koepke J, Gueckel F, et al. MRI with superparamagnetic iron oxide: efficacy in the detection and characterization of focal hepatic lesions. Magn Reson Imaging 1999;17:383e92.
Detection of HCC by ferucarbotran-enhanced accumulation phase FS-T1WI
12. Ward J, Robinson PJ, Guthrie JA, et al. Liver metastases in candidates for hepatic resection: comparison of helical CT and gadolinium- and SPIO-enhanced MR imaging. Radiology 2005;237:170e80. 13. Tanimoto A, Wakabayashi G, Shinmoto H, et al. The mechanism of ring enhancement in hepatocellular carcinoma on superparamagnetic iron oxide-enhanced T1-weighted images: an investigation into peritumoral Kupffer cells. J Magn Reson Imaging 2005;21:230e6. 14. Macarini L, Marini S, Murrone M, et al. Alterations in hepatic uptake and distribution of organ-specific superparamagnetic MRI contrast media: clinical findings and classification according to pathogenesis. Radiol Med 2004;108:92e106. 15. Lee JM, Kim CS, Youk JH, et al. Characterization of focal liver lesions with superparamagnetic iron oxide-enhanced MR imaging: value of distributional phase T1-weighted imaging. Korean J Radiol 2003;4:9e18. 16. Oswald P, Clement O, Chambon C, et al. Liver positive enhancement after injection of superparamagnetic nanoparticles: respective role of circulating and uptaken particles. Magn Reson Imaging 1997;15:1025e31.
29
17. Hamm B, Staks T, Taupitz M. SHU 555A. A new superparamagnetic iron oxide contrast agent for magnetic resonance imaging. Invest Radiol 1994;29(Suppl. 2):S87e9. 18. Hamm B, Staks T, Taupitz M, et al. Contrast-enhanced MR imaging of liver and spleen: first experience in humans with a new superparamagnetic iron oxide. J Magn Reson Imaging 1994;4:659e68. 19. Kim JH, Kim MJ, Suh SH, et al. Characterization of focal hepatic lesions with ferumoxides-enhanced MR imaging: utility of T1-weighted spoiled gradient recalled echo images using different echo times. J Magn Reson Imaging 2002;15: 573e83. 20. Van Beers BE, Gallez B, Pringot J. Contrast-enhanced MR imaging of the liver. Radiology 1997;203:297e306. 21. Torzilli G, Minagawa M, Takayama T, et al. Accurate preoperative evaluation of liver mass lesions without fine-needle biopsy. Hepatology 1999;30:889e93. 22. Kim YK, Kwak HS, Kim CS, et al. Hepatocellular carcinoma in patients with chronic liver disease: comparison of SPIO-enhanced MR imaging and 16-detector row CT. Radiology 2006;238:531e41.
Detection of hepatocellular carcinoma by ferucarbotran-enhanced magnetic resonance imaging: the efficacy of accumulation phase fat-suppressed T1-weighted imaging C.-T. Choua,b, R.-C. Chena,c,*, W.-T. Chenc, J.-M. Liic a
Department of Biomedical Imaging and Radiological Science, National Yang-Ming Medical University, Department of Radiology, Chang-Hua Christian Hospital, Erlin Branch, and cDepartment of Radiology, Taipei City Hospital, Renai Branch, Taiwan
b
Received 18 March 2008; received in revised form 18 June 2008; accepted 11 July 2008
AIM: To evaluate the effectiveness of accumulation phase, fat-suppressed, T1-weighted imaging (FS-T1WI) when detecting hepatocellular carcinoma (HCC) by ferucarbotran-enhanced magnetic resonance imaging (MRI). MATERIALS AND METHODS: Thirty patients who underwent ferucarbotran-enhanced MRI, which resulted in 35 confirmed HCCs, were included in this prospective study. Two image sets were prepared and two radiologists independently reviewed these in two reading sessions; set A was without contrast-enhanced accumulation phase FS-T1WI and set B included contrast-enhanced accumulation phase FS-T1WI. All HCCs had been confirmed by operation (n ¼ 4), by biopsy (n ¼ 28), and by follow-up study for at least 1 year (n ¼ 3). RESULTS: The contrast-to-noise ratio significantly increased from 1.2 7.5 to 12.7 7.3 with contrast-enhanced accumulation phase FS-T1WI, but was only slightly increased from 12.2 10.3 to 15.5 12.2 with contrast-enhanced T2WI (p < 0.001). The signal-to-noise ratio (SNR) was decreased with T1WI and T2WI for liver parenchyma. With T2WI, the SNR for HCCs was decreased; however, it was slightly increased with T1WI (p < 0.001). Overall, 29 HCCs were detected using set A, and 35 nodules were identified using set B, which included the contrast-enhanced accumulation phase FS-T1WI. Thus, the detection rate significantly increased using post-contrast medium accumulation phase FS-T1WI (p < 0.05). CONCLUSION: Due to the improved CNR with the post-contrast medium accumulation phase FS-T1WI, which helped to increase HCC detection, accumulation phase FS-T1WI is recommended as one of the routine protocols for inclusion in HCC detection. ª 2008 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Hepatocellular carcinoma (HCC) is the most common primary hepatic malignancy, and the detection of synchronous HCCs or intrahepatic metastasis is essential for treatment planning and prognosis.1,2 Superparamagnetic iron oxide (SPIO) is a liver-specific particulate magnetic * Guarantor and correspondent: R.-C. Chen, Department of Radiology, Taipei City Hospital, Renai Branch, No. 10, Sec 4, Renai Road, Taipei 106, Taiwan. Tel.: þ886 2 27093600x5103; fax: þ886 2 27040013. E-mail address: [email protected] (R.-C. Chen).
resonance imaging (MRI) contrast agent that is primarily taken up by the Kupffer cells of the liver and is useful for tumour detection.3 Ring enhancement of a tumour on contrast-enhanced T1-weighted imaging (T1WI) is known to be highly suggestive of malignancy when using ferumoxideenhanced MRI.4 Contrast-enhanced T1WI with fat suppression (FS) is superior to T1WI without FS when detecting small HCCS.5 Therefore, it was hypothesized that ferucarbotran-enhanced accumulation phase T1WI with FS imaging (FS-T1WI) ought to improve HCC detection. The effectiveness of this procedure for detecting HCCs has not been investigated. The purpose of the present
0009-9260/$ - see front matter ª 2008 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2008.07.010
Detection of HCC by ferucarbotran-enhanced accumulation phase FS-T1WI
study was to evaluate the effectiveness of ferucarbotran-enhanced accumulation phase FS-T1WI when used as part of a HCC-detection protocol.
Materials and methods Patients This study was approved by the institutional review board of our hospital. Written informed consent related to undergoing ferucarbotran-enhanced MRI was obtained from all patients before they were enrolled in the study. A total of 39 patients, with clinical findings of chronic liver disease or an elevated serum alpha-fetoprotein (AFP) level and those suspected of having HCC based on ultrasonography, underwent ferucarbotran-enhanced MRI as part of their diagnostic work-up. One patient refused a histological examination and was excluded from the study. In total, 38 out of 39 patients received histological examination. Eight out of the 38 patients (seven men, one woman; age range 38e73 years) were identified to have benign nodules (one inflammatory pseudotumour, one hyperplastic nodule, and six dysplastic nodules). The inflammatory pseudotumour had disappeared at the 1 year follow-up. The patient with a dysplastic nodule was lost to follow-up. The other six nodules had remained unchanged at the 1 year follow-up based on computed tomography (CT) or MRI studies. Finally, 30 of the 39 patients (20 men, 10 women; age range 28e85 years) with histologically proven HCC were included in the study and were followed up for at least 1 year. Twenty-six patients had undergone needle biopsy and four had undergone surgical resection. Among them, 25 patients had liver cirrhosis and all had chronic hepatitis (hepatitis B, 21 patients; hepatitis C, six patients; hepatitis B and C, three patients). A total of 35 HCC nodules in 30 patients were identified. Twenty-five patients had solitary HCCs and five patients had two HCCs. Among the patients who had two hepatic nodules, at least one of the lesions was biopsied and examined histologically. Thirty-two HCCs were eventually confirmed histopathologically and these consisted of 14 welldifferentiated HCCs, 15 moderately differentiated HCCs, and three poorly differentiated HCCs. When the other three hepatic lesions without histological examination were considered, two showed lipiodol retention in the tumours on follow-up CT imaging after transcatheter embolization for HCC. One of the three tumours without lipiodol retention had progressively enlarged on the 3 month and 1 year
23
follow-up CT images, and the tumour showed focal arterial enhancement on the 1 year follow-up CT image. Based on these clinical manifestations and imaging findings, these three tumours were also considered to be HCC.
MRI MRI was performed using a 1.5 T MRI system (Gyroscan ACS-NT, Philips, Best, The Netherlands) equipped with a body phased-array coil. Unenhanced MRI was performed with T2-weighted (T2W) turbo spinecho (TSE) axial imaging [repetition time/echo time (TR/TE) 2500/90 ms; TSE factor 23; section thickness 8 mm, gap 0.8 mm] with and without FS under respiratory trigger; T2* echo planar imaging (EPI; TR/TE 500/13.8 ms) was obtained during one breath-hold. T1W dual echo imaging (TR/TE: 210/2.3 ms and 4.6 ms; section thickness 8 mm, gap 0.8 mm) and T1W with FS imaging (TR/TE 241e344/2.7 ms; flip angle, 70 ; section thickness 8 mm, gap 0.8 mm) was also performed during one or two breath-holdings depending on the liver size of the patient. Automatic shimming was applied for FS imaging to maximize magnetic field homogeneity. Flow compensation was also used. The ferucarbotran (Resovist, Schering, Berlin, Germany) consists of SPIO microparticles coated with carboxydextran and this was preloaded into 1.4 ml (>50 kg body weight) using a connecting intravenous tube; the agent was then manually injected rapidly through a filter with a 5 mm pore size (3). Dynamic T1W fast field echo (FFE) imaging (175e210/ 1.3e2.1; flip angle 80 ) was carried out before and at 18e20 s and 50e55 s after the contrast agent was injected. Approximately 10 min after intravenous injection of the SPIO agent, the contrast-enhanced images, including the T2WI with and without FS, T2*EPI were obtained in the same order as used for the unenhanced imaging acquisition. Finally, accumulation phase FS-T1WI (TR/TE 241e344/2.7 ms; flip angle 70 ; section thickness 8 mm, gap 0.8 mm) was performed.
Qualitative analysis The imaging analysis was performed at a dualscreen diagnostic workstation (GE Healthcare, Milwaukee, WI, USA). Two image sets (set A and set B) were prepared in which set A contained unenhanced T1WI with dual echo and FS-TIWI, T2WI with and without FS, T2*EPI, dynamic T1WI, contrast-enhanced T2WI with and without FS and T2* EPI images. Set B consisted of all of the images of set A plus a contrast-enhanced accumulation
24
phase FS-T1WI. Two observers experienced in abdominal MRI then independently reviewed image set A and then, 4 months later, they reviewed set B; the observers carried out the reviews separately at two different reading sessions. Both observers were blinded to the clinical information and final diagnosis. Both image sets of the whole liver were analysed using visual assessment by the two radiologists with the aim of estimating the number and location of HCC nodules based on the following criteria. A tumour presenting with the following characteristics was considered as HCC: first, hyperintense on the T2WI associated with a variety of signal intensities on T1WI; second, hyperintense on the T1WI with fatty metamorphosis detected on the dual echo T1WI; third, positive enhancement on the dynamic MRI images associated with a variety of signal intensities on T1WI, and finally, positive enhancement on the contrast-enhanced accumulation phase FS-T1WI.3,6,7 The observers recorded the HCC nodules using reading set A based on the first, second, and third diagnostic criteria. However, when using set B, all four criteria were included. Any other nodules were ignored in this study. The enhancement pattern of the HCC was visually classified into the following patterns: ring enhancement, homogeneous enhancement, heterogeneous enhancement, and no enhancement. Differences between the two reviewers were resolved by a third radiologist who joined a consensus conference and who was also blind to the clinical information.
Quantitative analysis On the unenhanced and ferucarbotran-enhanced MRI images, all HCCs were quantitatively evaluated for signal intensity (SI) and compared with normal liver parenchyma by one radiologist. Operatordefined region-of-interest SI measurements were taken directly from the monitor for the lesion and the normal liver parenchyma. As the SI of the normal liver parenchyma was measured, care was taken to avoid the major hepatic and portal vessels. The standard deviation (SD) of the background noise was measured by placing the region-ofinterest anterior to the patient near the liver parenchyma and by calculating the SD including motion artefacts. Similar regions of interest were used on the unenhanced and ferucarbotranenhanced images. Signal-to-noise ratios (SNRs) for the HCCs and the liver were calculated for all MRI sequences using the following formula: SNR ¼ SI=SDbackground
C.-T. Chou et al.
where SI is the SI of the lesion or liver parenchyma and SDbackground is the SD of the air background noise. In addition, contrast-to-noise ratios (CNRs) for the lesion to liver were computed for each sequence as follows: CNR ¼ ðSI lesion SI liverÞ=SDbackground These values are negative for hypointense lesions and positive for hyperintense lesions.
Statistical analysis The results are expressed as means SD. Statistical differences in the detection numbers for both imaging sets were compared by the McNemar test. A paired t-test was used to confirm the difference in the SNR/CNR for the unenhanced and contrastenhanced imaging. A p-value of less than 0.05 was considered statistically significant. A Kappa statistic was obtained to evaluate the lesion detection agreement between the two reading sets.8
Results Tumour detection Overall, a consensus reading was necessary for three of the 35 HCCs when analysing imaging set A, and was necessary for two HCCs when using imaging set B. In terms of HCC detection, there was good inter-observer agreement between the two readers with both set A (k ¼ 0.79) and set B (k ¼ 0.91). A total of 35 HCCs (mean size 2 1.1 cm) were found using imaging set B, but only 29 HCCs were identified using reading set A. Among the six HCC nodules (mean size 1.6 0.4 cm) only identified in set B, (Table 1), there were two that were well-differentiated (Fig. 1), one that was moderately differentiated, and the other three tumours did not have a histological diagnosis (Fig. 2). Overall, the detection of small HCCs was significantly increased with the help of the ferucarbotran-enhanced accumulation phase FS-T1WI (p < 0.05). Among the benign nodules, two nodules (one inflammatory pseudotumour and one dysplastic nodule) were considered as HCCs when identified by both imaging set A and set B. Both of the two nodules showed hyperintensity on T2WI, with arterial enhancement during the contrast-enhanced dynamic study and hyperintensity on the contrastenhanced T2WI and accumulation phase FS-T1WI. The other six nodules were considered to be benign cirrhotic nodules by the both imaging set A and set B.
Detection of HCC by ferucarbotran-enhanced accumulation phase FS-T1WI
Table 1
The additional six HCC nodules found on the imaging set B
Size (cm) Unenhanced Imaging
1 2 3 4 5 6
1.8 1 1.6 1.2 2.1 1.7
25
Contrast-enhanced Imaging
PSIL (T2WI) Imaging diagnosis
T1WI
T2WI
Dynamic T2WI enhancement
Delay FS-T1WI
Hyperintense Hyperintense Isointense Hyperintense Hyperintense Hyperintense
Hypointense Hypointense Hypointense Isointense Hypointense Hypointense
No No No No No No
Yes Yes Yes Yes Yes Yes
Hypointense Hypointense Hypointense Isointense Hypointense Hypointense
40.3% 37.0% 49.2% 65.5% 58.6% 48.8%
Set A
Set B
DN DN HN DN DN DN
HCC HCC HCC HCC HCC HCC
Pathology
Well Nonea Noneb Well Moderately Nonea
T1WI, T1-weighted imaging; T2WI, T2-weighted imaging; FS, fat saturated; PSIL, percent of signal intensity loss; DN, dysplastic nodule; HN, hyperplastic nodule; Well, well-differentiated HCC, Moderately, moderately-differentiated HCC, Yes, positive enhancement. a Lipiodol retention within the tumour was noted on the follow-up CT images after transarterial embolization treatment. b Progressively enlarged on follow-up CT studies and typical HCC enhanced pattern on 1 year follow-up CT study.
Qualitative analysis The signal intensity of the HCC was compared to the surrounding normal parenchyma, and the results are displayed in Table 2. According to the enhancement pattern, 29 of 35 HCCs were homogeneous,
three showed heterogeneous enhancement on the accumulation phase FS-T1WI, and three were isointense relative to the liver parenchyma. The three HCCs with heterogeneous enhancement on the accumulation phase FS-T1WI were also heterogeneous on the contrast-enhanced T2WI.
Figure 1 A well-differentiated HCC (arrow) proven by biopsy in a 71-year-old woman. (a) The unenhanced T2WI image shows a hypointense nodule beside a hepatic cyst on S8 of the liver. (b) The unenhanced FS-T1WI image shows a hyperintense lesion in the same area. (c) The tumour is obscured in the contrast-enhanced T2WI image. (d) Positive enhancement of the tumour is clearly depicted on the contrast-enhanced accumulation phase FS-T1WI image.
26
C.-T. Chou et al.
Figure 2 A nodule (arrow) at S5/8 of the liver was considered to be HCC in a 68-year-old man. (a) The unenhanced FS-T2WI image shows a hypointense nodule in S5/8 of liver. (b) The unenhanced FS-T1WI shows no obvious tumour in the same area. (c) Persistent hypointensity of the tumour is depicted in the contrast-enhanced FS-T2WI image. (d) A hyperintense tumour is clearly depicted on the accumulation phase FS-T1WI image. (e) The arterial phase of the 1 year follow-up CT reveals an enlargement of the nodule (from 1.6 to 2.2 cm) and focal arterial enhancement with a nodule-in-nodule appearance (arrowhead).
Quantitative analysis The SNRs and CNRs were quantitatively analysed using T1WI and T2WI before and after ferucarbotran administration. The difference in SNR/CNR for unenhanced and contrast-enhanced imaging and the altered SNR/CNR between T1WI and T2WI are shown on Table 3. The SNRs of the HCCs showed a slight increase in mean value on FS-T1WI, but a decrease on the T2WI (p < 0.001). The HCC-to-liver CNRs were
calculated and these showed a significant increase in mean value on the FS-T1WI, but only a slight increase on the T2WI. (p < 0.001; Fig. 3)
Discussion It is extremely important to detect small, multifocal HCCs because this will affect the decision to proceed to surgical resection or transplantation,
Detection of HCC by ferucarbotran-enhanced accumulation phase FS-T1WI
27
Table 2 Numbers of HCC with variable signal intensity relative to normal liver parenchyma on pre- and post-Ferucarbotran contrast-enhanced T1- and T2-weighted images
T1WI
T2WI
Hyperintense Hypointense Isointense Hyperintense Hypointense Isointense
Unenhanced Imaging
Contrast-enhanced Imaging
12 18 5 27 5 3
32 0 3 29 4 2
which is based on the numbers and sizes of the lesions.9 SPIO-enhanced MRI of the liver has been intensely investigated and it is now widely accepted that SPIO is able to increase the detection rate of focal lesions using T2WI.10e12 However, several authors have encountered difficulties when detecting HCC due to the presence of Kupffer cells in the tumour.6,13,14 Lee et al.15 described how SPIO-enhanced T1WI was useful for lesion characterization only, but not lesion detection. In the present results, hyperintensity on the contrast-enhanced accumulation phase FS-T1WI was readily able to identify six additional HCC nodules compared with contrast-enhanced T2WI. One of the six HCCs showed up as isointense on T1WI, was hypointense on T2WI, and was considered to be a cirrhotic nodule on imaging set A; however, this tumour showed homogeneous enhancement on the delay FS-T1WI and was subsequently considered to be HCC. (Fig. 2) SPIO particles have both high T1 and T2 relaxivities, and both circulating and clustered particles may, therefore, theoretically contribute to either the positive or negative enhancement of liver tissue.16 There are three possible mechanisms by which HCC may present with hyperintensity on an accumulation phase FS-T1WI. The first reason is tumour enhancement by the T1 effect of the residual
Figure 3 (a) The changes in the SNR of HCCs and normal liver parenchyma on the T1WI and T2WI images before and after administration of ferucarbotran are presented. (b) Elevation of the HCC-to-liver CNRs on the T1WI images was significantly greater than with the T2WI images (p < 0.001).
circulating SPIO particles.17,18 The second is ferucarbotran uptake by the Kupffer cells and as a result the clustered particles produce a T2* effect that might influence the SI on T1WI. In the present results, there was signal loss by the liver parenchyma and this was greater than with the HCC nodules, which caused a dark background on the accumulation phase FS-T1WI. The third is that the FS used for the T1WI in the present study might provide a higher lesion-to-liver contrast,5 and
Table 3 The signal-to-noise ratios (SNR) of HCCs/liver parenchyma and the contrast-to-noise ratios (CNR) of HCCs for unenhanced and contrast-enhanced T1W/T2W images T1WI
SNR (mean SD)
CNR (mean SD) a b
p < 0.05. p > 0.05.
T2WI
unenhanced
Contrast enhanced
Contrast enhanced e unenhanced difference
unenhanced
Contrast enhanced
Contrast enhanced e unenhanced difference
HCC 35.3 9.6
36.3 11
1 9.8b
31.6 11.2
24.6 12.3
7 8.1a
Liver 36.6 11.3 23.6 7.7
12.9 9.9a
19.4 7.5
9.1 4.4
10.4 5.6a
HCC 1.2 7.5
14 10.4a
12.2 10.3
15.5 12.2
3.4 7.1a
12.7 7.3
T1WIpost-pree T2WIpost-pre difference
8 12.4 P < 0.001 2.6 10.2 p > 0.05 10.6 11.5 p < 0.001
28
therefore, some HCCs became more conspicuous on the contrast-enhanced accumulation phase FST1WI than that on the contrast-enhanced T2WI. Two benign nodules presented with hyperintensity on the accumulation phase FS-T1WI and both nodules also presented hyperintensity on the T2WI and with arterial enhancement during the dynamic studies. Both nodules were considered to be HCCs using both imaging set A and set B. Furthermore, three HCCs were isointense and were not found on the contrast-enhanced accumulation phase FS-T1WI in the present study, but were well depicted on the contrast-enhanced T2WI. Although the accumulation phase FS-T1WI was effective for the detection of HCCs, the delay FS-T1WI is not able to completely replace the contrast-enhanced T2WI when carrying out ferucarbotran-enhanced MRI and the probability of benign nodules presenting with hyperintensity on the accumulation phase FS-T1WI should be kept in mind. The mean value of the HCC-to-liver CNRs showed a significant increase on the mean value FS-T1WI. In contrast, the CNRs were only slightly increased on the T2WI. Thus, the improvement in CNR for the tumours is more prominent using T1WI rather than T2WI after ferucarbotran administration. This may explain the better detection rate for small HCCs on the contrast-enhanced accumulation phase FS-T1WI (Fig. 3b). Peritumoural ring enhancement has been described as an important sign in distinguishing benign from malignant lesions when using SPIOenhanced MRI.19,20 The mechanism of HCC ring enhancement on SPIO-enhanced T1WI has been discussed by Tanimoto et al.13 The frequency of HCC presenting with ring enhancement on contrast-enhanced T1WI images varies from 10% in the study by Lee et al.15 to 72% in the study by Tanimoto et al.3,13 When hepatic tumours presented with hyperintensity, ring enhancement was not readily visualized.20 Ring enhancement tends to occur with larger tumours rather than small tumours.13 In the present study, 82% of the HCCs were homogeneous and 9% demonstrated heterogeneous enhancement on the accumulation phase FS-T1WI; it was notable that the mean size of the HCCs was small (2 1.1 cm). The small size might explain why none of the 35 HCCs in the present study showed ring enhancement. The scan time for accumulation phase FS-T1WI with breath-holding is short (18e21 s for one breath-holding and 28e31 s for two breathholding) and can be performed immediately after the contrast-enhanced T2WI. It is very convenient to obtain an accumulation phase T1WI without
C.-T. Chou et al.
additional body positioning or hardware preparation. In the present study, six additional tumours were detected by adding accumulation phase FS-T1WI. One limitation of the present study was that not all HCCs (32/35) were proven histologically. This is a known shortcoming in most comparative studies.21,22 The major problem is that ultrasound or CT-guided biopsy is impossible for nodules that are only visible on an MRI image. Furthermore, it may not be practical to biopsy all the hepatic nodules in patients with multiple HCC nodules. In conclusion, due to the better CNR on the contrast-enhanced accumulation phase FS-T1WI, which increased HCC detection, accumulation phase FS-T1WI is recommended as part of the routine protocol for HCC detection.
References 1. Szklaruk J, Silverman PM, Charnsangavej C. Imaging in the diagnosis, staging, treatment and surveillance of hepatocellular carcinoma. AJR Am J Roentgenol 2003;180: 441e54. 2. Noone TC, Semmelka RC, Chaney DM, et al. Abdominal imaging studies; comparison of diagnostic accuracies resulting from ultrasound, computed tomography, and magnetic resonance imaging in the same individual. Magn Reson Imaging 2004;22:19e24. 3. Tanimoto A, Kuribayashi S. Application of superparamagnetic iron oxide to imaging of hepatocellular carcinoma. Eur J Radiol 2006;58:200e16. 4. Kanematsu M, Matsuo M, Yamada Y, et al. Perilesional hyperintense rim of malignant hepatic tumours on ferumoxide-enhanced T1-weighted gradient-echo MR images: correlation between MR imaging and histopathologic findings. J Magn Reson Imaging 2003;18:40e8. 5. Koushima Y, Ebara M, Fukuda H, et al. Small hepatocellular carcinoma: assessment with T1-weighted spin-echo magnetic resonance imaging with and without fat suppression. Eur J Radiol 2002;41:34e41. 6. Imai Y, Murakami T, Yoshida S, et al. Superparamagnetic iron oxide-enhanced magnetic resonance images of hepatocellular carcinoma: correlation with histological grading. Hepatology 2000;32:205e12. 7. Efremidis SC, Hytiroglou P, Matsui O. Enhancement patterns and signal-intensity characteristics of small hepatocellular carcinoma in cirrhosis: pathologic basis and diagnostic challenges. Eur Radiol 2007;17:2969e82. 8. Landis JR, Koch GG. The measurement of observer agreement for categrorical data. Biometrics 1977;33:159e74. 9. Mazzaferro V, Regalia E, Doci R, et al. Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis. N Engl J Med 1996;334:693e9. 10. Reimer P, Jahnke N, Fiebich M, et al. Hepatic lesion detection and characterization: value of nonenhanced MR imaging, superparamagnetic iron oxide-enhanced MR imaging, and spiral CT-ROC analysis. Radiology 2000;217:152e8. 11. Poeckler-Schoeniger C, Koepke J, Gueckel F, et al. MRI with superparamagnetic iron oxide: efficacy in the detection and characterization of focal hepatic lesions. Magn Reson Imaging 1999;17:383e92.
Detection of HCC by ferucarbotran-enhanced accumulation phase FS-T1WI
12. Ward J, Robinson PJ, Guthrie JA, et al. Liver metastases in candidates for hepatic resection: comparison of helical CT and gadolinium- and SPIO-enhanced MR imaging. Radiology 2005;237:170e80. 13. Tanimoto A, Wakabayashi G, Shinmoto H, et al. The mechanism of ring enhancement in hepatocellular carcinoma on superparamagnetic iron oxide-enhanced T1-weighted images: an investigation into peritumoral Kupffer cells. J Magn Reson Imaging 2005;21:230e6. 14. Macarini L, Marini S, Murrone M, et al. Alterations in hepatic uptake and distribution of organ-specific superparamagnetic MRI contrast media: clinical findings and classification according to pathogenesis. Radiol Med 2004;108:92e106. 15. Lee JM, Kim CS, Youk JH, et al. Characterization of focal liver lesions with superparamagnetic iron oxide-enhanced MR imaging: value of distributional phase T1-weighted imaging. Korean J Radiol 2003;4:9e18. 16. Oswald P, Clement O, Chambon C, et al. Liver positive enhancement after injection of superparamagnetic nanoparticles: respective role of circulating and uptaken particles. Magn Reson Imaging 1997;15:1025e31.
29
17. Hamm B, Staks T, Taupitz M. SHU 555A. A new superparamagnetic iron oxide contrast agent for magnetic resonance imaging. Invest Radiol 1994;29(Suppl. 2):S87e9. 18. Hamm B, Staks T, Taupitz M, et al. Contrast-enhanced MR imaging of liver and spleen: first experience in humans with a new superparamagnetic iron oxide. J Magn Reson Imaging 1994;4:659e68. 19. Kim JH, Kim MJ, Suh SH, et al. Characterization of focal hepatic lesions with ferumoxides-enhanced MR imaging: utility of T1-weighted spoiled gradient recalled echo images using different echo times. J Magn Reson Imaging 2002;15: 573e83. 20. Van Beers BE, Gallez B, Pringot J. Contrast-enhanced MR imaging of the liver. Radiology 1997;203:297e306. 21. Torzilli G, Minagawa M, Takayama T, et al. Accurate preoperative evaluation of liver mass lesions without fine-needle biopsy. Hepatology 1999;30:889e93. 22. Kim YK, Kwak HS, Kim CS, et al. Hepatocellular carcinoma in patients with chronic liver disease: comparison of SPIO-enhanced MR imaging and 16-detector row CT. Radiology 2006;238:531e41.