HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Krinsky, G. A. Articles by Teperman, L. W. Search for Related Content PubMed PubMed Citation Articles by Krinsky, G. A. Articles by Teperman, L. W.

Hepatocellular Carcinoma and Dysplastic Nodules in Patients with Cirrhosis: Prospective Diagnosis with MR Imaging and Explantation Correlation¹

¹ From the Departments of Radiology (G.A.K., V.S.L., J.C.W., N.M.R.), Pathology (N.D.T.), and Transplant Surgery (T.D., L.W.T.,) and the Kaplan Comprehensive Cancer Center (G.A.K.), New York University Medical Center, 530 First Ave, New York, NY 10016. Received August 14, 2000; revision requested September 20; revision received October 4; accepted November 9. Supported by an RSNA Research Grant and an award from the Society of Computed Tomography and Magnetic Resonance. Address correspondence to G.A.K. (e-mail: glenn.krinsky@med.nyu.edu)

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
PURPOSE: To determine the sensitivity and specificity of magnetic resonance (MR) imaging for detection of hepatocellular carcinoma (HCC) and dysplastic nodules (DNs) by using explantation correlation in patients with cirrhosis and no known HCC.

MATERIALS AND METHODS: Seventy-one patients without a known history of HCC who underwent MR imaging and subsequent transplantation within 90 days were examined. Breath-hold turbo short inversion time inversion-recovery and/or T2-weighted turbo spin-echo MR images were obtained. Dynamic two- or three-dimensional gadolinium-enhanced gradient-echo MR images were obtained in the hepatic arterial, portal venous, and equilibrium phases. Prospective MR image interpretations were compared directly with explanted liver pathologic results.

RESULTS: Eleven (15%) of 71 patients had hepatic malignancies; MR imaging enabled diagnosis of tumor in six (54%) of 11 patients. On a lesion-by-lesion basis, MR imaging depicted 11 of 20 hepatic neoplasms, for an overall sensitivity of 55%. MR imaging depicted four (80%) of five lesions larger than 2 cm, six (50%) of 12 lesions 1–2 cm, and one (33%) of three lesions smaller than 1 cm. MR imaging depicted only nine (15%) of 59 DNs. The specificities of MR imaging for detection of HCC and DNs on a per patient basis were 60 (86%) of 70 patients and 53 (85%) of 62 patients, respectively.

CONCLUSION: MR imaging is insensitive for the diagnosis of small (<2-cm) HCCs and DNs.

Index terms: Liver, cirrhosis, 761.794 • Liver, nodules, 761.3198 • Liver, transplantation, 761.459 • Liver neoplasms, 761.31, 761.321, 761.323 • Liver neoplasms, MR, 761.121411, 761.121412, 761.121413, 761.121415, 761.12143

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Cirrhosis is a progressive diffuse process of liver fibrosis that is characterized by architectural distortion and the development of a spectrum of nodules ranging from benign regenerative nodules to dysplastic nodules to overtly malignant hepatocellular carcinoma (HCC) (1). The incidence of HCC in North America has almost doubled during the past 20 years (2), and long-term survival is poor (3). Early diagnosis and resection or transplantation, the most effective treatment for HCC, substantially increases survival (4,5). In one study (4), a 4-year recurrence-free survival rate of 85% following liver transplantation was achieved in patients with limited disease (one lesion <5 cm or up to three lesions 3 cm). It is therefore critical to detect nodules that contain HCC at an early stage both to control tumor burden while awaiting transplantation and to distinguish those patients who may have favorable long-term results with transplantation from those who will not, because cadaveric organs are in short supply.

The magnetic resonance (MR) imaging literature is limited by its lack of complete correlation between pathologic and imaging findings. Most published studies have relied on biopsy or surgical resection specimens to evaluate the accuracy of imaging detection and characterization of cirrhotic nodules, with the resulting bias being toward the positive studies. In addition, in many studies the diagnosis of HCC is known prior to MR imaging and is based on the results of other radiologic studies or markedly elevated -fetoprotein levels; these factors further compound the bias. Finally, many earlier MR studies involved imaging strategies that are no longer considered state of the art—for example, conventional T1- and T2-weighted spin-echo imaging performed with a body coil and without dynamic contrast material enhancement. It is now well established that imaging of the cirrhotic liver during the hepatic arterial phase (HAP) with computed tomography (CT) or MR increases the diagnostic yield for HCC (6–12), and, in some cases, tumors are seen only during this phase (9–12).

The purpose of our study was to assess the diagnostic sensitivity and specificity of state-of-the-art MR imaging of HCC and dysplastic nodules in the cirrhotic liver, by using thin-section whole-explant correlation following transplantation, in patients in whom the diagnosis of HCC was not known before imaging.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Subjects
From December 1995 to April 2000, 287 patients, 28 of whom had known HCC, underwent liver transplantation for cirrhosis at our institution. One hundred of these patients underwent MR imaging within 90 days (mean, 32 days) before transplantation. We selected 90 days as the longest interval between MR imaging and pathologic evaluation to minimize the chance of incorrect false-negative diagnoses of tumor due to the rapid doubling time of some HCCs (13). The reasons for incomplete capture were the interval between imaging and transplantation was longer than 90 days, the imaging study was performed at an outside institution, and the patient was moribund. Three patients were excluded because they did not receive intravenous contrast material because of poor intravenous access (n = 2) or refusal (n = 1). Patients with known HCC (n = 26) were excluded from the study to minimize positive bias. Twenty-four of these patients before MR imaging underwent transhepatic arterial chemoembolization, which can cause ischemic insult to the liver and make radiologic-pathologic correlation very difficult. Therefore, the study population consisted of 71 patients (42 men, 29 women; mean age, 50 years; age range, 23-68 years) in whom a diagnosis of neoplasm was not known at the time of MR imaging.

The causes of cirrhosis in the 71 patients are listed in Table 1. All patients, except those with primary sclerosing cholangitis, had Child-Pugh class B or C cirrhosis. Eight patients had transjugular intrahepatic portosystemic shunts at the time of MR imaging. Institutional review board approval was obtained for this study, and informed consent was obtained from all patients. During the same 52-month study period, 41 patients with HCC who were deemed poor candidates for transplantation on the basis of excessive tumor burden diagnosed at MR imaging, advanced age, and/or substantial comorbid disease were treated with palliative care.

In all patients, T1-weighted breath-hold spoiled gradient-echo fast low-angle shot (FLASH) images were acquired in the transverse plane with 130–220/4–5 (repetition time msec/echo time msec), a 70°–90° flip angle, a 5–8-mm section thickness, a 0%–10% intersection gap, a 128–192 x 256 matrix, and a rectangular field of view optimized to the patient’s body habitus, with the largest dimension of 30–40 cm. This resulted in a 14–26-second breath hold for 20–26 sections. The same pulse sequence with similar parameters was then performed out of phase (echo time, 2.1–2.5 msec). In five patients, image quality was poor because of the patient’s inability to follow commands or suspend respiration. Therefore, a breath-hold–independent sequence involving magnetization-prepared T1-weighted spoiled gradient-echo imaging—that is, turbo fast low-angle shot (turbo FLASH) (11.0/4.2, 300-msec inversion time, 15° flip angle)— was used. This sequence was then performed out of phase with an echo time of 2.2 msec. The imaging parameters and voxel size were similar to those used in the breath-hold spoiled gradient-echo FLASH sequence.

In 69 of 71 patients, a breath-hold echo train turbo short inversion time inversion-recovery (STIR) sequence (4,000–5,500/58 or 76 [effective], 150–165-msec inversion time, 150°–180° flip angle, echo train length of 33) was performed with an 8-mm section thickness and 2-mm intersection gap. This required two noninterleaved breath-hold acquisitions with 8–10 sections for sufficient anatomic coverage. The matrix and field of view were similar to those used to acquire the previously described T1-weighted images.

In the first 32 patients and in those who were unable to suspend respiration, a T2-weighted echo train turbo spin-echo sequence (3,600–6,000/99 [effective], 180° flip angle, echo train length of 11) was performed in the transverse plane with an 8-mm section thickness, a 2-mm intersection gap, three signals acquired, and a similar matrix and field of view.

All patients underwent dynamic gadolinium-enhanced MR imaging in the HAP, portal venous phase, and equilibrium phase after intravenous administration of 19 mL of gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Wayne, NJ) at a rate of 2 mL/sec by means of hand injection (n = 4) or an MR-compatible power injector (Spectris; MedRad, Pittsburgh, Pa) (n = 67). In addition, patients with primary sclerosing cholangitis underwent delayed imaging (10–15 minutes). A test dose of 0.5 or 1.0 mL of gadopentetate dimeglumine with transverse turbo FLASH imaging was performed to optimize HAP imaging in 67 patients (14). The time delay between the start of the injection and the start of the HAP acquisition ranged from 8–24 seconds. In the first 36 patients who were capable of breath holding, the previously described in-phase spoiled gradient-echo FLASH sequence also was performed with dynamic gadolinium-enhanced imaging.

In the first four of the 71 patients, imaging began 20 seconds after initiation of contrast material administration and was repeated at 50 and 110 seconds. In the remaining 67 patients, the first acquisition was timed according to the data from the test bolus, and an additional three or four breath-hold acquisitions were spaced 35–45 seconds apart. This allowed the patient at least 10–15 seconds of free breathing between breath holds. The second acquisition encompassed the portal venous phase, and the last acquisition was considered the equilibrium phase study. Breath holding was routinely performed at end expiration to facilitate subtraction imaging.

In five patients who were unable to suspend respiration, dynamic turbo FLASH imaging (11.0/4.2, 300-msec inversion time, 15° flip angle) was performed similarly to FLASH imaging, with five consecutive contrast material–enhanced acquisitions.

In the remaining 30 patients who were capable of breath holding, transverse three-dimensional fat-suppressed spoiled gradient-echo imaging (4.2/1.8, 12° flip angle, bandwidth of 488 Hz per pixel) was performed. The section thickness ranged from 160 to 200 mm to ensure full coverage of the liver, and interpolation in the section-select direction yielded a partition thickness of 2–3 mm with a matrix interpolated to 256 x 256. Nonenhanced acquisition was performed first, followed by an acquisition timed according to the data from the test bolus. The acquisition was then repeated twice at 45-second intervals for portal venous and equilibrium phase imaging.

Finally, in 67 patients, a flow-sensitive breath-hold gradient-echo sequence (20–200/9–20, 20°–45° flip angle) was performed, first with selective presaturation above and then with presaturation below the imaging sections to assess the patency and direction of portal venous flow. The voxel size was similar to that in the T1-weighted gradient-echo pulse sequence, but only four sections per breath hold were acquired. All patients were able to complete the examinations without untoward side effects, and all studies were of diagnostic quality.

Image Interpretation
The MR studies were prospectively reviewed in consensus by one of three attending radiologists (N.M.R., G.A.K., V.S.L.) and two fellows by using film hard-copy images from all the available pulse sequences described. The three-dimensional T1-weighted data sets—both nonenhanced and three-phase gadolinium enhanced—were evaluated at a commercially available workstation (Siemens Medical Systems). Because of the difficulty in determining the enhancement of lesions that are hyperintense on nonenhanced T1-weighted images, this data set was subtracted from the HAP acquisition. This facilitated the qualitative recognition of lesion enhancement (15). Quantitative measurements of enhancement were not performed.

The criteria for malignancy were as follows: a mass demonstrating homogeneous, heterogeneous, or ring enhancement during the HAP (excluding the classic peripheral, transient hepatic signal intensity difference [16–19]) and/or moderate hyperintensity on T2-weighted turbo spin-echo and/or turbo STIR images. Lesions that were markedly hyperintense on T2-weighted turbo spin-echo and/or turbo STIR images and not enhancing after contrast material administration were considered to be cysts, and hemangiomas were diagnosed on the basis of previously described enhancement patterns (20).

Nodules that were hyperintense and substantially larger than the background regenerative nodules at T1-weighted pulse sequences but did not demonstrate arterial phase enhancement or T2 hyperintensity were considered to be dysplastic (21,22). All other nodules were presumed to be regenerative. The size, signal intensity characteristics, enhancement patterns, lesion locations, and diagnoses were recorded in a database and in the dictated clinical report.

Pathologic Analysis
One of two pathologists (N.D.T.) cut the explanted livers sequentially into 5–8-mm sections that corresponded as closely as possible to the MR imaging planes. Dysplastic and HCC nodules were identified grossly as those that were distinct from the surrounding regenerative nodules in terms of size, texture, color, and/or degree of bulging beyond the cut surface of the liver (23). All livers were photographed, and all lesions other than ordinary regenerative nodules were sampled.

By using the diagnostic criteria from the International Working Party’s "Terminology of Nodular Hepatocellular Lesions" (23), nodules were classified as follows: regenerative nodule, dysplastic nodule-low grade, dysplastic nodule-high grade, small HCC (<2 cm), or HCC (>2 cm). There were no absolute size criteria, as described by the International Working Party, to diagnose or distinguish regenerative nodules from dysplastic nodules; a dysplastic nodule could be as small as 1 mm, and a regenerative nodule could be as large as 5 cm (23). Low-grade dysplastic nodules were defined as those showing normal architecture and cytologic features or diffuse large cell changes (23). High-grade dysplastic nodules were defined on the basis of the presence of one of the following: diffuse small cell changes, pseudogland formation, nodule-in-nodule lesions with small cell dysplasia, iron resistance in siderotic nodules, fatty change, clear cell change, or Mallory body clustering (23). HCCs were classified as well, moderately, or poorly differentiated.

Statistical Analyses
A two-tailed Student t test was performed to compare the size differences between HCCs detected and HCCs not seen and the size differences between dysplastic nodules detected and dysplastic nodules not seen. A P value of less than .05 was used to define statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Hepatic Neoplasms
Eleven (15%) of 71 patients—10 with HCC and one with cholangiocarcinoma—had hepatic malignancies, and MR imaging facilitated the diagnosis of tumor in six (54%) of 11 patients. One patient had six lesions, two had three lesions, and eight had one lesion. On a lesion-by-lesion basis, MR imaging depicted 11 of 20 malignant hepatic neoplasms—19 HCCs and one cholangiocarcinoma—for an overall sensitivity of 55%. Lesion size ranged from 0.6 to 7.5 cm (mean, 1.8 cm). The mean size of the lesions detected (2.2 cm) was larger than that of the lesions missed (1.3 cm), although this difference was not statistically significant (P = .17). MR imaging depicted four (80%) of five lesions larger than 2 cm, six (50%) of 12 lesions 1–2 cm, and one (33%) of three lesions smaller than 1 cm.

In the subset of 30 patients who underwent thin-section (2–3-mm) three-dimensional fat-suppressed T1-weighted imaging, both with and without contrast material, only two patients had hepatic neoplasms—cholangiocarcinoma (Fig 1) in one patient and HCC in the other—and both were detected prospectively by using MR imaging.

View larger version (131K): [in this window] [in a new window] [Download PPT slide]   Figure 1a. Cholangiocarcinoma in a 47-year-old man with primary sclerosing cholangitis. (a) Transverse breath-hold turbo STIR MR image (4,000/76) of the liver shows a hyperintense mass (arrows) in the caudate lobe. (b) Transverse breath-hold three-dimensional fat-suppressed T1-weighted HAP MR image (4.2/1.9) of the liver shows a poorly defined hypovascular mass (long arrows) and a peripheral wedge-shaped area of enhancement (short arrow) in the right lobe. (c) Transverse portal venous phase MR image shows a hypointense mass (arrows) and normal liver parenchyma at the site of arterial phase enhancement, consistent with a transient hepatic signal intensity difference. (d) MR image obtained 10 minutes after c shows the mass (arrows) is now hyperintense, characteristic of intrahepatic cholangiocarcinoma. (e) Corresponding transverse section from the explanted liver demonstrates the neoplasm (arrows) in the caudate lobe.

View larger version (125K): [in this window] [in a new window] [Download PPT slide]   Figure 1b. Cholangiocarcinoma in a 47-year-old man with primary sclerosing cholangitis. (a) Transverse breath-hold turbo STIR MR image (4,000/76) of the liver shows a hyperintense mass (arrows) in the caudate lobe. (b) Transverse breath-hold three-dimensional fat-suppressed T1-weighted HAP MR image (4.2/1.9) of the liver shows a poorly defined hypovascular mass (long arrows) and a peripheral wedge-shaped area of enhancement (short arrow) in the right lobe. (c) Transverse portal venous phase MR image shows a hypointense mass (arrows) and normal liver parenchyma at the site of arterial phase enhancement, consistent with a transient hepatic signal intensity difference. (d) MR image obtained 10 minutes after c shows the mass (arrows) is now hyperintense, characteristic of intrahepatic cholangiocarcinoma. (e) Corresponding transverse section from the explanted liver demonstrates the neoplasm (arrows) in the caudate lobe.

View larger version (124K): [in this window] [in a new window] [Download PPT slide]   Figure 1c. Cholangiocarcinoma in a 47-year-old man with primary sclerosing cholangitis. (a) Transverse breath-hold turbo STIR MR image (4,000/76) of the liver shows a hyperintense mass (arrows) in the caudate lobe. (b) Transverse breath-hold three-dimensional fat-suppressed T1-weighted HAP MR image (4.2/1.9) of the liver shows a poorly defined hypovascular mass (long arrows) and a peripheral wedge-shaped area of enhancement (short arrow) in the right lobe. (c) Transverse portal venous phase MR image shows a hypointense mass (arrows) and normal liver parenchyma at the site of arterial phase enhancement, consistent with a transient hepatic signal intensity difference. (d) MR image obtained 10 minutes after c shows the mass (arrows) is now hyperintense, characteristic of intrahepatic cholangiocarcinoma. (e) Corresponding transverse section from the explanted liver demonstrates the neoplasm (arrows) in the caudate lobe.

View larger version (130K): [in this window] [in a new window] [Download PPT slide]   Figure 1d. Cholangiocarcinoma in a 47-year-old man with primary sclerosing cholangitis. (a) Transverse breath-hold turbo STIR MR image (4,000/76) of the liver shows a hyperintense mass (arrows) in the caudate lobe. (b) Transverse breath-hold three-dimensional fat-suppressed T1-weighted HAP MR image (4.2/1.9) of the liver shows a poorly defined hypovascular mass (long arrows) and a peripheral wedge-shaped area of enhancement (short arrow) in the right lobe. (c) Transverse portal venous phase MR image shows a hypointense mass (arrows) and normal liver parenchyma at the site of arterial phase enhancement, consistent with a transient hepatic signal intensity difference. (d) MR image obtained 10 minutes after c shows the mass (arrows) is now hyperintense, characteristic of intrahepatic cholangiocarcinoma. (e) Corresponding transverse section from the explanted liver demonstrates the neoplasm (arrows) in the caudate lobe.

View larger version (93K): [in this window] [in a new window] [Download PPT slide]   Figure 1e. Cholangiocarcinoma in a 47-year-old man with primary sclerosing cholangitis. (a) Transverse breath-hold turbo STIR MR image (4,000/76) of the liver shows a hyperintense mass (arrows) in the caudate lobe. (b) Transverse breath-hold three-dimensional fat-suppressed T1-weighted HAP MR image (4.2/1.9) of the liver shows a poorly defined hypovascular mass (long arrows) and a peripheral wedge-shaped area of enhancement (short arrow) in the right lobe. (c) Transverse portal venous phase MR image shows a hypointense mass (arrows) and normal liver parenchyma at the site of arterial phase enhancement, consistent with a transient hepatic signal intensity difference. (d) MR image obtained 10 minutes after c shows the mass (arrows) is now hyperintense, characteristic of intrahepatic cholangiocarcinoma. (e) Corresponding transverse section from the explanted liver demonstrates the neoplasm (arrows) in the caudate lobe.

View larger version (150K): [in this window] [in a new window] [Download PPT slide]   Figure 2a. Hypervascular HCCs, with one lesion seen only during the HAP, in a 65-year-old man with cryptogenic cirrhosis. (a) Transverse nonenhanced magnetization-prepared T1-weighted gradient-echo MR image (11.0/4.2) shows a 2.5-cm hyperintense mass (arrow). (b) Transverse HAP MR image shows an enhancing mass (short arrow) and a hyperintense larger lesion (long arrow) similar to that seen in a. (c) Nonenhanced subtracted HAP MR image shows unequivocal enhancement of the larger lesion (arrow). The smaller lesion is not seen because a slightly cranial section was used for subtraction.

View larger version (144K): [in this window] [in a new window] [Download PPT slide]   Figure 2b. Hypervascular HCCs, with one lesion seen only during the HAP, in a 65-year-old man with cryptogenic cirrhosis. (a) Transverse nonenhanced magnetization-prepared T1-weighted gradient-echo MR image (11.0/4.2) shows a 2.5-cm hyperintense mass (arrow). (b) Transverse HAP MR image shows an enhancing mass (short arrow) and a hyperintense larger lesion (long arrow) similar to that seen in a. (c) Nonenhanced subtracted HAP MR image shows unequivocal enhancement of the larger lesion (arrow). The smaller lesion is not seen because a slightly cranial section was used for subtraction.

View larger version (150K): [in this window] [in a new window] [Download PPT slide]   Figure 2c. Hypervascular HCCs, with one lesion seen only during the HAP, in a 65-year-old man with cryptogenic cirrhosis. (a) Transverse nonenhanced magnetization-prepared T1-weighted gradient-echo MR image (11.0/4.2) shows a 2.5-cm hyperintense mass (arrow). (b) Transverse HAP MR image shows an enhancing mass (short arrow) and a hyperintense larger lesion (long arrow) similar to that seen in a. (c) Nonenhanced subtracted HAP MR image shows unequivocal enhancement of the larger lesion (arrow). The smaller lesion is not seen because a slightly cranial section was used for subtraction.

View larger version (179K): [in this window] [in a new window] [Download PPT slide]   Figure 3a. Dysplastic nodule that was enhancing during the HAP and incorrectly diagnosed as HCC in a 61-year-old woman with cryptogenic cirrhosis. (a) Transverse T2-weighted turbo spin-echo MR image (4,200/99) shows a hypointense mass (arrow) in the inferior right lobe of the liver. The mass (arrow in c) is hyperintense on the T1-weighted breath-hold (b) in-phase (190.0/4.0) and (c) out-of-phase (200.0/2.5) FLASH images. (d) Transverse HAP image (190.0/4.0) shows the enhancing mass (arrow). Subtracted MR image findings (not shown) confirmed the enhancement. (e) Portal venous phase MR image shows the mass (arrow) is isointense relative to the liver parenchyma. (f) Corresponding transverse sections from the explanted liver demonstrate the mass (arrows), which was pathologically proved to be a high-grade dysplastic nodule.

View larger version (178K): [in this window] [in a new window] [Download PPT slide]   Figure 3b. Dysplastic nodule that was enhancing during the HAP and incorrectly diagnosed as HCC in a 61-year-old woman with cryptogenic cirrhosis. (a) Transverse T2-weighted turbo spin-echo MR image (4,200/99) shows a hypointense mass (arrow) in the inferior right lobe of the liver. The mass (arrow in c) is hyperintense on the T1-weighted breath-hold (b) in-phase (190.0/4.0) and (c) out-of-phase (200.0/2.5) FLASH images. (d) Transverse HAP image (190.0/4.0) shows the enhancing mass (arrow). Subtracted MR image findings (not shown) confirmed the enhancement. (e) Portal venous phase MR image shows the mass (arrow) is isointense relative to the liver parenchyma. (f) Corresponding transverse sections from the explanted liver demonstrate the mass (arrows), which was pathologically proved to be a high-grade dysplastic nodule.

View larger version (198K): [in this window] [in a new window] [Download PPT slide]   Figure 3c. Dysplastic nodule that was enhancing during the HAP and incorrectly diagnosed as HCC in a 61-year-old woman with cryptogenic cirrhosis. (a) Transverse T2-weighted turbo spin-echo MR image (4,200/99) shows a hypointense mass (arrow) in the inferior right lobe of the liver. The mass (arrow in c) is hyperintense on the T1-weighted breath-hold (b) in-phase (190.0/4.0) and (c) out-of-phase (200.0/2.5) FLASH images. (d) Transverse HAP image (190.0/4.0) shows the enhancing mass (arrow). Subtracted MR image findings (not shown) confirmed the enhancement. (e) Portal venous phase MR image shows the mass (arrow) is isointense relative to the liver parenchyma. (f) Corresponding transverse sections from the explanted liver demonstrate the mass (arrows), which was pathologically proved to be a high-grade dysplastic nodule.

View larger version (189K): [in this window] [in a new window] [Download PPT slide]   Figure 3d. Dysplastic nodule that was enhancing during the HAP and incorrectly diagnosed as HCC in a 61-year-old woman with cryptogenic cirrhosis. (a) Transverse T2-weighted turbo spin-echo MR image (4,200/99) shows a hypointense mass (arrow) in the inferior right lobe of the liver. The mass (arrow in c) is hyperintense on the T1-weighted breath-hold (b) in-phase (190.0/4.0) and (c) out-of-phase (200.0/2.5) FLASH images. (d) Transverse HAP image (190.0/4.0) shows the enhancing mass (arrow). Subtracted MR image findings (not shown) confirmed the enhancement. (e) Portal venous phase MR image shows the mass (arrow) is isointense relative to the liver parenchyma. (f) Corresponding transverse sections from the explanted liver demonstrate the mass (arrows), which was pathologically proved to be a high-grade dysplastic nodule.

View larger version (182K): [in this window] [in a new window] [Download PPT slide]   Figure 3e. Dysplastic nodule that was enhancing during the HAP and incorrectly diagnosed as HCC in a 61-year-old woman with cryptogenic cirrhosis. (a) Transverse T2-weighted turbo spin-echo MR image (4,200/99) shows a hypointense mass (arrow) in the inferior right lobe of the liver. The mass (arrow in c) is hyperintense on the T1-weighted breath-hold (b) in-phase (190.0/4.0) and (c) out-of-phase (200.0/2.5) FLASH images. (d) Transverse HAP image (190.0/4.0) shows the enhancing mass (arrow). Subtracted MR image findings (not shown) confirmed the enhancement. (e) Portal venous phase MR image shows the mass (arrow) is isointense relative to the liver parenchyma. (f) Corresponding transverse sections from the explanted liver demonstrate the mass (arrows), which was pathologically proved to be a high-grade dysplastic nodule.

View larger version (181K): [in this window] [in a new window] [Download PPT slide]   Figure 3f. Dysplastic nodule that was enhancing during the HAP and incorrectly diagnosed as HCC in a 61-year-old woman with cryptogenic cirrhosis. (a) Transverse T2-weighted turbo spin-echo MR image (4,200/99) shows a hypointense mass (arrow) in the inferior right lobe of the liver. The mass (arrow in c) is hyperintense on the T1-weighted breath-hold (b) in-phase (190.0/4.0) and (c) out-of-phase (200.0/2.5) FLASH images. (d) Transverse HAP image (190.0/4.0) shows the enhancing mass (arrow). Subtracted MR image findings (not shown) confirmed the enhancement. (e) Portal venous phase MR image shows the mass (arrow) is isointense relative to the liver parenchyma. (f) Corresponding transverse sections from the explanted liver demonstrate the mass (arrows), which was pathologically proved to be a high-grade dysplastic nodule.

View larger version (142K): [in this window] [in a new window] [Download PPT slide]   Figure 4. Mass that was enhancing during the HAP and incorrectly diagnosed as HCC in a 61-year-old man with cirrhosis due to primary sclerosing cholangitis. Transverse breath-hold three-dimensional fat-suppressed T1-weighted HAP MR image (4.2/1.3) shows a 0.5-cm homogeneously enhancing mass (arrow). No corresponding abnormality was seen at either gross or microscopic evaluation.

View larger version (168K): [in this window] [in a new window] [Download PPT slide]   Figure 5a. Dysplastic nodule correctly diagnosed at MR imaging in a 41-year-old man with cirrhosis due to hepatitis C and ethanol abuse. (a) Transverse breath-hold T1-weighted in-phase FLASH image (190.0/4.1) shows a 1-cm hyperintense mass (arrow). (b) On the transverse breath-hold turbo STIR MR image (4,000/76), the mass (arrow) is hypointense. Gross and histologic examination results (not shown) revealed a low-grade dysplastic nodule.

View larger version (166K): [in this window] [in a new window] [Download PPT slide]   Figure 5b. Dysplastic nodule correctly diagnosed at MR imaging in a 41-year-old man with cirrhosis due to hepatitis C and ethanol abuse. (a) Transverse breath-hold T1-weighted in-phase FLASH image (190.0/4.1) shows a 1-cm hyperintense mass (arrow). (b) On the transverse breath-hold turbo STIR MR image (4,000/76), the mass (arrow) is hypointense. Gross and histologic examination results (not shown) revealed a low-grade dysplastic nodule.

View larger version (166K): [in this window] [in a new window] [Download PPT slide]   Figure 6a. Regenerative nodule incorrectly diagnosed as dysplastic at MR imaging in a 52-year-old woman with cirrhosis due to hepatitis C. On the transverse breath-hold T1-weighted (a) in-phase (154.0/4.1) and (b) out-of-phase (160.0/2.3) FLASH images, the mass (arrow) is hyperintense. (c) On the transverse breath-hold turbo STIR MR image (4,000/76), the mass (arrow) is hypointense. Gross and histologic examination results (not shown) revealed a regenerative nodule.

View larger version (179K): [in this window] [in a new window] [Download PPT slide]   Figure 6b. Regenerative nodule incorrectly diagnosed as dysplastic at MR imaging in a 52-year-old woman with cirrhosis due to hepatitis C. On the transverse breath-hold T1-weighted (a) in-phase (154.0/4.1) and (b) out-of-phase (160.0/2.3) FLASH images, the mass (arrow) is hyperintense. (c) On the transverse breath-hold turbo STIR MR image (4,000/76), the mass (arrow) is hypointense. Gross and histologic examination results (not shown) revealed a regenerative nodule.

View larger version (173K): [in this window] [in a new window] [Download PPT slide]   Figure 6c. Regenerative nodule incorrectly diagnosed as dysplastic at MR imaging in a 52-year-old woman with cirrhosis due to hepatitis C. On the transverse breath-hold T1-weighted (a) in-phase (154.0/4.1) and (b) out-of-phase (160.0/2.3) FLASH images, the mass (arrow) is hyperintense. (c) On the transverse breath-hold turbo STIR MR image (4,000/76), the mass (arrow) is hypointense. Gross and histologic examination results (not shown) revealed a regenerative nodule.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
State-of-the-art hepatic MR imaging techniques, including those involving the use of a phased-array coil, T2-weighted echo train fast spin-echo or STIR pulse sequences, and dynamic gadolinium-enhanced breath-hold imaging, are widely used for the noninvasive detection and staging of cirrhotic liver malignancies (24–26). In our study, because whole explanted liver specimens were available within 90 days after MR imaging in all cases, we could determine both the sensitivity and specificity of hepatic MR imaging for the diagnosis of hepatic malignancies and dysplastic nodules. Despite the use of these techniques, we achieved a sensitivity of only 53% (10 of 19 lesions) for the detection of HCC and of 55% (11 of 20 lesions) for the detection of hepatic malignancies (including one cholangiocarcinoma); these values are much lower than those previously reported in the MR imaging literature (75%–100%) (9,25–28). This is likely due in part to the small size of the lesions missed (mean, 1.3 cm), which reflects a cohort of patients in whom the diagnosis of hepatic malignancy was not known prior to MR imaging, and the ability to sample the entire liver at pathologic analysis. Our results are similar to those of a smaller retrospective MR imaging study (29) with explantation correlation, in which 13 (62%) of 21 malignant lesions were detected.

The sensitivity for the detection of hepatic malignancies (55%) that we achieved is comparable also to those in two explantation studies involving the use of CT after intraarterial administration of iodized oil: The sensitivities were 53% (30) and 58% (31), which are slightly lower than the 68% sensitivity achieved in a nonhelical contrast-enhanced CT study (32) with 200 consecutive patients who underwent transplantation. In a more recent explantation study (33) involving three-phase helical CT, a sensitivity of 71% for HCC detection was reported. However, compared with our study, in most prior explantation studies, the diagnosis of HCC was known or clinically suspected in some patients.

Baron and Oliver (34) screened 1,000 patients with cirrhosis and without known HCC, 195 of whom underwent transplantation, by using predominantly dual-phase helical CT. HCC was present in 32 (16%) of 195 patients and detected in 19 (59%) of the 32 patients. These data are similar to those in our study and representative of findings in a population of patients with late-stage cirrhosis who undergo transplantation. However, on a lesion-by-lesion basis, CT depicted only 23 (36%) of 63 HCCs in the study by Baron and Oliver (34). The poor sensitivity was likely due in part to the fact that 52 (82%) of the 63 HCCs were 2 cm or smaller. Detection of these lesions is of great clinical importance, because the tumor doubling time can be as short as 90 days (13) and the wait for a cadaveric liver can be as long as 2 years. Therefore, if a 2-cm HCC is not detected and the patient undergoes imaging on a yearly basis, there is a possibility that the lesion may be 8 cm by the time it is diagnosed; this size may preclude transplantation at some centers. The detection of small HCCs and subsequent nonsurgical treatment may control the tumor burden while the patient awaits cadaveric or living donor liver transplantation.

The results of previous MR imaging studies (7–9,12,24,25) have demonstrated the importance of imaging during the HAP, especially for detection of small HCCs, because these may be occult at other pulse sequences and on portal venous and equilibrium phase images. Our study results confirm these findings: Eight HCCs demonstrated uniform enhancement during the arterial phase; one (0.7 cm) of these lesions was seen during only the HAP (Fig 2). To optimize HAP imaging, we performed a test bolus examination in 67 patients, because in some patients, empiric delays can result in suboptimal studies owing to variations in circulation times (14).

In our study, a large number of cases (15 lesions in 10 patients) were false-positive for neoplasm. Three patients had four regions of poorly marginated confluent hepatic fibrosis that could not be confidently distinguished from infiltrative neoplasms. In a large CT explantation study, Miller et al (32) reported similar problems in distinguishing confluent hepatic fibrosis from infiltrative neoplasms. In that study, only two of the 10 ill-defined infiltrating lesions were due to neoplasms. Although the MR imaging features of confluent hepatic fibrosis have been established (35), there remains substantial overlap in the signal intensity and enhancement characteristics of confluent hepatic fibrosis, as compared with infiltrative neoplasm, so a definitive diagnosis of fibrosis may be difficult in many cases. One of the two true-positive infiltrating neoplasms demonstrated irregular ring enhancement, and the other encased the portal vein and inferior vena cava.

Four high-grade dysplastic nodules were detected but incorrectly diagnosed as HCC owing to the presence of arterial phase enhancement. This finding confirms that in a previous case report (36)—that is, these lesions can demonstrate arterial phase enhancement at MR imaging and CT—and thus provides further support that criteria other than arterial phase enhancement should be used to distinguish dysplastic nodules from small HCCs. In the remaining seven lesions that were false-positive for malignant neoplasms at MR imaging, no pathologic diagnosis could be obtained despite meticulous correlation of the findings of MR imaging and pathologic analysis of the explanted specimens. Four of these lesions were seen during only the HAP and appeared as well-circumscribed masses measuring 0.5–2.1 cm. Any of these lesions could have represented a nodular form of a transient hepatic signal intensity difference due to arterioportal shunts (16). This belief is supported by the fact that ordinary regenerative nodules have not been demonstrated previously to have an arterial blood supply at angiography (37), CT hepatic arteriography (38), or triphasic spiral CT (39).

Dysplastic nodules are neoplastic premalignant nodules (40,41) that are found in 15%–28% of all explanted livers with cirrhosis (42–44). The development of HCC within a dysplastic nodule in as little as 4 months has been documented in longitudinal studies (45). In our study, the overall sensitivity of MR imaging for the detection of dysplastic nodules was poor on both a patient-by-patient (three of 18 [17%]) and a lesion-by-lesion (nine of 59 [15%]) basis. The 15% sensitivity for dysplastic nodule detection is much lower than the 60% sensitivity reported by Horigome et al (46). However, that series did not have explantation correlation and relied on focal resection and 21-gauge fine-needle biopsy findings for pathologic confirmation; these factors probably led to the higher sensitivities. By using the previously published criteria of T1 hyperintensity and T2 hypointensity (21,22), only nine dysplastic nodules were identified in our study. In fact, a lesion that met these signal intensity criteria was more likely to be a regenerative nodule (n = 15) than a dysplastic nodule. Inclusion of arterial phase enhancement as a criterion for dysplastic nodule would have increased the sensitivity to 22% (13 of 59 lesions) at the expense of decreased specificity, because the HCC lesions demonstrated this pattern more frequently.

Earls et al (22) studied explanted livers with high-spatial-resolution MR imaging (256 x 512 matrix, 20–26-cm field of view) in a retrospective nonblinded fashion and achieved a sensitivity of 96% (24 of 25 lesions) for the detection of dysplastic nodules with explantation correlation. However, in vivo MR imaging is subject to patient motion and flow artifacts that may impede the detection of these nodules. Despite the use of thin-section (2–3-mm) three-dimensional fat-suppressed contrast-enhanced T1-weighted imaging in a subset of 30 patients, none of the 31 dysplastic nodules was identified in our series. This implies that increased spatial resolution may not be sufficient to augment the detection of dysplastic nodules in vivo. Further research to determine the optimal imaging strategy for detecting and characterizing dysplastic nodules is needed. Detection of dysplastic nodules in patients without HCC is important because these lesions are premalignant (40,41) and a marker for hepatocarcinogenesis occurring elsewhere in the liver; thus, patients with dysplastic nodules have a higher risk of developing HCC than do patients without dysplastic nodules (43). Therefore, it may be prudent to perform imaging in these patients with greater frequency compared with that in patients with cirrhosis without dysplastic nodules.

In our study, four of the 10 pathologically proved HCCs detected at MR imaging were well-circumscribed masses that were mild to moderately hyperintense at turbo T2-weighted and/or STIR imaging. No dysplastic nodule or regenerative nodule had these signal intensity characteristics. In fact, to our knowledge, only one dysplastic nodule has been reported to be hyperintense at T2-weighted pulse sequences (46). It has also been reported that regenerative nodules in patients with cirrhosis due to Budd-Chiari syndrome (47) and in those who have undergone infarction (48) also may be hyperintense at T2-weighted pulse sequences. In our study, all of the well-circumscribed lesions that were markedly hyperintense at turbo T2-weighted and/or STIR pulse sequences were either cysts or hemangiomas (20).

Despite the fact that four HCCs and the nine dysplastic nodules detected were hyperintense at T1-weighted gradient-echo sequences both in and out of phase, only one lesion—an HCC—proved to be steatotic. In our study, four of five HCCs that were hyperintense on T1-weighted gradient-echo images were well differentiated at pathologic analysis; these findings confirmed those of Ebara et al (49)—that is, increased signal intensity on T1-weighted images correlates with a more well differentiated histologic grade than does isointensity or hypointensity.

This study had recognized limitations. Because the prospective interpretation of lesions was performed by using consensus reading, interobserver variability could not be evaluated. Furthermore, although the MR imaging criteria used to define dysplastic nodule were developed from previous studies (21,22), one of which had explantation correlation (22), they were not perfect. As noted previously, four dysplastic nodules in two patients were incorrectly interpreted as HCC on the basis of arterial phase enhancement. However, HAP enhancement of dysplastic nodules at MR imaging is rare: To our knowledge, the first such case report was published in 1998 (35), 3 years after this study began. In addition, we were unable to diagnose any of the 11 siderotic iron-retentive dysplastic nodules because we did not consider hypointensity at all pulse sequences as a criterion for dysplastic nodule. However, siderotic regenerative nodules may be difficult to distinguish from siderotic dysplastic nodules at MR imaging. Furthermore, we did not prospectively assess for the presence of a capsule. Although this finding is quite specific for HCC, capsules rarely are seen in small HCCs in non-Asian populations. We were unable to compare the accuracy of thin-section (2–3-mm) three-dimensional fat-suppressed contrast-enhanced T1-weighted imaging, which was performed in 30 patients, with that of dynamic two-dimensional imaging, because the three-dimensional imaging group had only two neoplasms. Finally, the overall number of pathologically proved HCCs in this study was small (n = 19). However, this was a direct result of excluding patients with known HCC, who often have a much greater tumor burden.

Although we used gadopentetate dimeglumine as the contrast material in this study, other agents have been used to detect and characterize hepatic masses in the cirrhotic liver. Although it was beyond the scope of this article to review these other agents and the studies performed with them, the combination of reticuloendothelial agents (ie, ferumoxides) and extracellular gadolinium chelates shows promise (50,51) by improving contrast at T2-weighted imaging while still enabling dynamic imaging. The use of ferumoxide-enhanced imaging alone, however, has been shown to be inferior to breath-hold gadolinium-enhanced imaging (26).

In conclusion, MR imaging is insensitive for detection of small (<2-cm) HCCs and dysplastic nodules, but it may be useful in detecting HCCs larger than 2 cm. Multiinstitutional prospective trials with much larger patient cohorts are needed to further validate the results of this study.

	FOOTNOTES

 
² Current address: Department of Radiology, Beth Israel Deaconess Medical Center, Boston, Mass

Abbreviations: FLASH = fast low-angle shot, HAP = hepatic arterial phase, HCC = hepatocellular carcinoma, STIR = short inversion time inversion recovery

Author contributions: Guarantor of integrity of entire study, G.A.K.; study concepts, G.A.K., N.M.R.; study design, G.A.K.; literature research, G.A.K.; clinical studies, all authors; data acquisition, G.A.K., V.S.L.; data analysis/interpretation, G.A.K., V.S.L., N.M.R., J.C.W.; statistical analysis, V.S.L.; manuscript preparation, definition of intellectual content, editing, revision/review, and final version approval, all authors.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1. Brown JJ, Naylor MJ, Yagan N. Imaging of hepatic cirrhosis. Radiology 1997; 202:1-16.[Free Full Text]
2. El-Serag HB, Mason AC. Rising incidence of hepatocellular carcinoma in the United States. N Engl J Med 1999; 340:745-750.[Abstract/Free Full Text]
3. Ince N, Wands JR. The increasing incidence of hepatocellular carcinoma. N Engl J Med 1999; 340:798-799.[Free Full Text]
4. 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; 334:693-699.[Abstract/Free Full Text]
5. Figueras J, Jaurrieta E, Valls C, et al. Resection or transplantation for hepatocellular carcinoma in cirrhotic patients: outcomes based on indicated treatment strategy. J Am Coll Surg 2000; 190:580-587.[Medline]
6. Ito K, Choji T, Nakada T, Nakanishi T, Kurokawa F, Okita K. Multislice dynamic MRI of hepatic tumors. J Comput Assist Tomogr 1993; 17:390-396.[Medline]
7. Oi H, Murakami T, Kim T, Matsushita M, Kishimoto H, Nakamura H. Dynamic MR imaging and early phase helical CT for detecting intrahepatic metastases of hepatocellular carcinoma. AJR Am J Roentgenol 1996; 166:369-374.[Abstract/Free Full Text]
8. Yamashita Y, Mitsuzaki K, Yi T, et al. Small hepatocellular carcinoma in patients with chronic liver damage: prospective comparison of detection with dynamic MR imaging and helical CT of the whole liver. Radiology 1996; 200:79-84.[Abstract/Free Full Text]
9. Peterson MS, Baron RL, Murakami T. Hepatic malignancies: usefulness of acquisition of multiple arterial and portal venous phase images at dynamic gadolinium-enhanced MR imaging. Radiology 1996; 201:337-345.[Abstract/Free Full Text]
10. Baron RL, Oliver JH, III, Dodd GD, III, Nalesnik M, Holbert BL, Carr BI. Hepatocellular carcinoma: evaluation with biphasic, contrast-enhanced, helical CT. Radiology 1996; 199:505-511.[Abstract/Free Full Text]
11. Oliver JH, III, Baron R, Federle MP, Rockette HE, Jr. Detecting hepatocellular carcinoma: value of unenhanced or arterial phase CT imaging or both used in conjunction with conventional portal venous phase contrast-enhanced CT imaging. AJR Am J Roentgenol 1996; 167:71-77.[Abstract/Free Full Text]
12. Kelekis NL, Semelka RC, Worawattanakul S, et al. Hepatocellular carcinoma in North America: a multiinstitutional study of appearance on T1-weighted, T2-weighted, and serial gadolinium-enhanced gradient-echo images. AJR Am J Roentgenol 1998; 170:1005-1013.[Abstract/Free Full Text]
13. Sheu JC, Sung JL, Chen DS. Growth rate of asymptomatic HCC and its clinical implications. Gastroenterology 1985; 89:257-266.
14. Earls J, Rofsky NR, DeCorato D, Krinsky G, Weinreb JC. Arterial-phase dynamic gadolinium-enhanced MR imaging: optimization with a test examination and a power injector. Radiology 1997; 202:268-273.[Abstract/Free Full Text]
15. Suto Y, Caner BE, Tamagawa Y, et al. Subtracted synthetic images in Gd-DTPA enhanced MR. J Comput Assist Tomogr 1989; 13:925-928.[Medline]
16. Itai Y, Matsui O. Blood flow and liver imaging. Radiology 1997; 202:306-314.[Free Full Text]
17. Matsui O, Kadoya M, Yoshikawa J, et al. Aberrant gastric venous drainage in cirrhotic livers: imaging findings in focal areas of liver parenchyma. Radiology 1995; 197:345-349.[Abstract/Free Full Text]
18. Nelson RC, McDermott VG, Paulson EK. Aberrant venous drainage to the liver: imaging implications. Radiology 1995; 197:338-340.[Free Full Text]
19. Ito K, Choji T, Fujita T, Matsumoto T, Nakada T, Nakanishi T. Early-enhancing pseudolesion in medial segment of left hepatic lobe detected with multisection dynamic MR. Radiology 1993; 187:695-699.[Abstract/Free Full Text]
20. Semelka RC, Brown ED, Ascher SM, et al. Hepatic hemangiomas: a multiinstitutional study of appearance on T2- weighted and serial gadolinium-enhanced gradient-echo MR images. Radiology 1994; 192:401-406.[Abstract/Free Full Text]
21. Matsui O, Kadoya M, Kameyama T, et al. Adenomatous hyperplastic nodules in the cirrhotic liver: differentiation from hepatocellular carcinoma with MR imaging. Radiology 1989; 173:123-126.[Abstract/Free Full Text]
22. Earls J, Thiese N, Weinreb JC, et al. Dysplastic nodules and hepatocellular carcinoma: thin-section MR imaging of explanted cirrhotic livers with pathologic correlation. Radiology 1996; 201:207-214.[Abstract/Free Full Text]
23. International Working Party. Terminology of nodular hepatocellular lesions: International Working Party. Hepatology 1995; 22:983-993.[Medline]
24. Yu JS, Kim KW, Kim EK, Lee JT, Yoo HS. Contrast enhancement of small hepatocellular carcinoma: usefulness of three successive early image acquisitions during multiphase dynamic MR imaging. AJR Am J Roentgenol 1999; 173:597-604.[Abstract/Free Full Text]
25. Fujita T, Ito K, Honjo K, et al. Detection of hepatocellular carcinoma: comparison of T2-weighted breath-hold fast spin-echo sequences and high resolution dynamic MR imaging with a phased-array body coil. J Magn Reson Imaging 1999; 9:274-279.[Medline]
26. Tang Y, Yamashita Y, Arakawa A, et al. Detection of hepatocellular carcinoma arising in cirrhotic livers: comparison of gadolinium- and ferumoxides-enhanced MR imaging. AJR Am J Roentgenol 1999; 172:1547-1544.[Abstract/Free Full Text]
27. Winter TC, Takayasu K, Muramatsu Y, et al. Early advanced hepatocellular carcinoma: evaluation of CT and MR appearance with pathologic correlation. Radiology 1994; 192:379-387.[Abstract/Free Full Text]
28. Kanematsu M, Hoshi H, Murakami T, et al. Detection of hepatocellular carcinoma in patients with cirrhosis: MR imaging versus angiographically assisted helical CT. AJR Am J Roentgenol 1997; 169:1507-1515.[Abstract/Free Full Text]
29. Born M, Layer G, Kreft B, Schwarz N, Schild H. MRI, CT and CT arterial portography in the diagnosis of malignant liver tumors in liver cirrhosis. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 1998; 168:567-572.[Medline]
30. Taourel PG, Pageaux GP, Coste V, et al. Small hepatocellular carcinoma in patients undergoing liver transplantation: detection with CT after injection of iodized oil. Radiology 1995; 197:377-380.[Abstract/Free Full Text]
31. Spreafico C, Marchiano A, Mazzaferro V, et al. Hepatocellular carcinoma in patients who undergo liver transplantation: sensitivity of CT with iodized oil. Radiology 1997; 203:457-460.[Abstract/Free Full Text]
32. Miller WJ, Baron RL, Dodd GD, III, Federle MP. Malignancies in patients with cirrhosis: CT sensitivity and specificity in 200 consecutive transplant patients. Radiology 1994; 193:645-650.[Abstract/Free Full Text]
33. Lim JH, Chan KK, Lee WJ, et al. Detection of hepatocellular carcinoma and dysplastic nodules in cirrhotic livers: accuracy of helical CT in transplant patients. AJR Am J Roentgenol 2000; 175:693-698.[Abstract/Free Full Text]
34. Baron RL, Oliver JH, III. Utility of helical biphasic contrast-enhanced CT for screening cirrhotic patients for hepatocellular carcinoma: analysis of a large transplant population (abstr). Radiology 1997; 205(P):287.[Medline]
35. Ohtomo K, Baron RL, Dodd GD, III, Federle MP, Ohtomo Y, Confer SR. Confluent hepatic fibrosis in advanced cirrhosis: evaluation with MR imaging. Radiology 1993; 189:871-874.[Abstract/Free Full Text]
36. Krinsky G, Thiese N, Rofsky NM, Mizrachi H, Teperman L, Weinreb JC. Dysplastic nodules in cirrhotic liver: arterial phase enhancement at CT and MR imaging—a case report. Radiology 1998; 209:461-464.[Abstract/Free Full Text]
37. Matsui O, Kadoya M, Kameyama T, et al. Benign and malignant nodules in cirrhotic livers: distinction based on blood supply. Radiology 1991; 178:493-497.[Abstract/Free Full Text]
38. Lim JH, Kim EY, Lee WJ, et al. Regenerative nodules in liver cirrhosis: findings at CT during arterial portography and CT hepatic arteriography with histopathologic correlation. Radiology 1999; 210:451-458.[Abstract/Free Full Text]
39. Baron RL, Marsh W, Jr, Oliver JH, III, Confer S, Hunt LE, Peterson MS. Screening cirrhosis for hepatocellular carcinoma (HCC) with helical contrast CT: specificity (abstr). Radiology 1997; 213(P):143.
40. Takayama T, Makuuchi M, Hirohashi S, et al. Malignant transformation of adenomatous hyperplasia to hepatocellular carcinoma. Lancet 1990; 336:1150-1153.[Medline]
41. Sakamoto M, Hirohashi S, Shimosato Y. Early stages of multistep hepatocarcinogenesis: adenomatous hyperplasia and early hepatocellular carcinoma. Hum Pathol 1991; 22:172-178.[Medline]
42. Hytiroglou P, Theise ND, Schwartz M, Mor E, Miller C, Thung S. Macroregenerative nodules in a series of adult cirrhotic liver explants: issues of classification and nomenclature. Hepatology 1995; 21:703-708.[Medline]
43. Theise ND, Schwartz M, Miller C, et al. Macroregenerative nodules and hepatocellular carcinoma in 44 sequential adult explants with cirrhosis. Hepatology 1992; 16:949-955.[Medline]
44. Ferrell LD, Wright T, Lake J, et al. Incidence and diagnostic features of macroregenerative nodules vs small hepatocellular carcinomas in cirrhotic livers. Hepatology 1992; 16:1372-1381.[Medline]
45. Sadek AG, Mitchell DG, Siegelman ES, et al. Early hepatocellular