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) Alert me when this article is cited 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 Hanna, R. F. Articles by Sirlin, C. B. Search for Related Content PubMed PubMed Citation Articles by Hanna, R. F. Articles by Sirlin, C. B. Social Bookmarking                      What's this? Hotlight (NEW!) What's Hotlight?

Double-Contrast MRI for Accurate Staging of Hepatocellular Carcinoma in Patients with Cirrhosis

¹ Department of Radiology, Division of Body Imaging, University of California, San Diego, 200 W Arbor Dr., San Diego, CA 92103-8756.
² Department of Internal Medicine, University of California, San Diego, San Diego, CA.

Received May 16, 2007; accepted after revision July 16, 2007.

 
Address correspondence to C. B. Sirlin (csirlin{at}ucsd.edu).

R. F. Hanna is supported by an Annual Research Fellowship from Albert Einstein College of Medicine, New York, NY. N. Kased is supported by a University of California, San Francisco, San Francisco, CA, Quarterly Fellowship. C. B. Sirlin has received research grants from Bayer HealthCare and GE Healthcare.

FOR YOUR INFORMATION

The American Roentgen Ray Society now provides instant Web exclusive access to its annual meeting abstracts. The abstracts, featured as a supplement to the American Journal of Roentgenology, summarize the latest comprehensive and clinically important information presented at ARRS's annual meetings. The abstracts can be viewed online by visiting www.ajronline.org.

Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
OBJECTIVE. The aim of this study was to evaluate the accuracy of a double-contrast MRI protocol in staging of hepatocellular carcinoma (HCC) in patients with cirrhosis.

MATERIALS AND METHODS. This cross-sectional study was performed at a tertiary liver care center. Forty-eight patients with cirrhosis underwent double-contrast MRI for clinical care and liver transplantation. For each MRI examination, superparamagnetic iron oxide was infused, and 2D T2^*-weighted spoiled gradient-recalled echo and T2-weighted echo-train spin-echo MR images were obtained for assessment of phagocytic function. Immediately afterward, a low-molecular-weight gadolinium compound was injected, and 3D T1-weighted spoiled gradient-recalled echo images were acquired dynamically for assessment of vascularity. Two blinded radiologists independently reviewed all MR images and assigned per-lesion and perpatient cancer confidence scores to determine the American Liver Tumor Study Group tumor stage. The imaging-based cancer scores and tumor stages were correlated with pathology reports. Performance parameters were computed for imaging-based measurements.

RESULTS. Of the 48 study subjects, 25 had HCC (three, T1; 18, T2; one, T3; one, T4a; two, T4b). In total, there were 37 HCC nodules. The accuracy of MRI in prediction of pathologic tumor stage was 81-85% depending on the radiologist. Per-patient and per-lesion sensitivity in the diagnosis of HCC were 96% and 81% for one radiologist and 96% and 89% for the other.

CONCLUSION. A double-contrast MRI protocol has high accuracy in staging of HCC in patients with cirrhosis.

Keywords: hepatocellular carcinoma • liver • MRI • sensitivity • staging

Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Liver transplantation can prolong the survival of patients with early-stage hepatocellular carcinoma (HCC) [1-5]. Since 1990, however, the number of patients awaiting liver transplantation has grown 15-fold, but the number of donors has increased only three-fold [6]. Because of the large mismatch between supply and demand, the United Network for Organ Sharing (UNOS) requires that patients with HCC undergo preoperative disease staging to ensure that only optimal candidates are selected to undergo liver transplantation. Current UNOS guidelines give priority to patients with HCC in tumor stage T2 [7], as defined by the American Liver Tumor Study Group modified TNM classification [8]. Patients with stages T1, T3, and T4 HCC are not given priority.

Under current UNOS guidelines, diagnostic imaging plays a critical role in establishing eligibility and proper allocation of livers among HCC patients because tumor stage can be determined preoperatively solely with imaging; preoperative biopsy is not required [9]. Despite the critical role of diagnostic imaging in assigning prioritization for liver transplantation, standard imaging approaches, including single-contrast MRI [4, 7, 10, 11], have been shown to have low accuracy in preoperative staging (Table 1). In 2006, Freeman et al. [7] performed the largest study to that point to evaluate preoperative HCC staging. Those authors concluded that current imaging methods "for radiologic staging before liver transplantation are unacceptably inaccurate" and that "more accurate radiologic staging methodology is needed."

Because single-contrast MRI has limited staging accuracy for HCC [4, 7, 10, 11], other MRI approaches should be considered. One possible approach is to administer two contrast agents, superparamagnetic iron oxide (SPIO) and gadolinium, sequentially in the same examination [12-14]. The rationale for the double-contrast examination is that images with the two agents provide complementary biologic information: gadolinium chelates show vascularity [15-19], and SPIOs show Kupffer cell density [20-22]. In principle, assessing both biologic features may improve staging accuracy. The purpose of our study was to evaluate the accuracy of a clinical double-contrast MRI protocol in staging of HCC in patients with cirrhosis who are awaiting liver transplantation. To our knowledge, this study is the first to assess accuracy of staging of HCC with a double-contrast MRI protocol.

View larger version (18K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1 —Flow diagram depicts manner in which patients were selected. Top three boxes describe procedures completed for clinical care. Bottom five boxes describe procedures completed for research study. Inclusion criteria were double-contrast MRI performed, cirrhosis histologically confirmed, and liver explant examined after double-contrast MRI examination. Exclusion criteria were ablative therapy before study MRI and liver explant performed more than 12 months after MRI in patients in whom explant had positive results for hepatocellular carcinoma (HCC). For patients in whom explant had negative results for HCC, MRI-explant interval greater than 12 months was acceptable, and such patients were not excluded.

Top Abstract Introduction Materials and Methods Results Discussion References

Subjects
Double-contrast MRI with sequential SPIO and gadolinium administration is the routine clinical protocol for HCC surveillance at our institution. From April 2001 to August 2004, our institution performed double-contrast MRI examinations on 434 patients for clinical care. Using the selection criteria outlined in Figure 1, we selected from the 434 patients a study sample consisting of all patients with histologically proven cirrhosis and liver explant findings as the reference standard for HCC stage. Patients who underwent ablative therapy before MRI were excluded. The final study group consisted of 48 patients (34 men, 14 women; mean age, 54 years, range, 36-68 years). The cause of cirrhosis was hepatitis C in 30 patients, hepatitis B in one patient, alcohol exposure in five, hepatitis C and alcohol in five, and other causes in seven patients.

MRI Technique
Before scanning, SPIO (ferumoxides, Feridex, Bayer HealthCare) diluted in 100 mL of 5% dextrose solution was infused IV over 30 minutes (iron dose, 10 µmol/kg of body weight). MRI was performed with a 1.5-T system (Magnetom Symphony, Siemens Medical Solutions) 30-60 minutes after completion of SPIO infusion. Images were not obtained before administration of SPIO. Examinations lasted 30 minutes. Patients were imaged in the supine position with torso phased-array coils centered over the liver. Transverse breath-hold images were obtained without parallel imaging with a rectangular field of view (20-32 x 34-42 cm, depending on body habitus). Eight MRI sequences grouped into two image sets were performed.

Image set 1 was SPIO enhanced. Three 2D spoiled gradient-recalled echo (SPGR) sequences were obtained in separate 12- to 20-second breath-holds at TEs of 2.6, 4.8, and 6.6 milliseconds. These TEs were chosen so that fat and water protons were out of phase, in phase, and out of phase, respectively, enabling assessment of intralesional and liver fat. Images with TEs of 2.6 and 4.8 milliseconds were T1 weighted; images with a TE of 6.6 milliseconds were T2^* weighted [23]. The TR was 120-220 milliseconds; flip angle, 70°; slice thickness, 8-10 mm; interslice gap, 2-3 mm; matrix size, 256 x 128; band-width, 260 Hz/pixel; acquisition time, 12-20 seconds. One 2D T2-weighted breath-hold echotrain spin-echo sequence was performed (TR/TE, 3,000-4,000/90-110; echo-train length, 29) with the same flip angle, slice thickness, slice location, matrix size, and bandwidth as the SPGR sequences. Acquisition time was 20-25 seconds.

Image set 2 was obtained without gadolinium enhancement and with dynamic gadolinium enhancement. Immediately after SPIO-enhanced images were obtained, one 3D interpolated T1-weighted fat-saturated SPGR sequence was performed before and repeated dynamically three times after bolus injection of 0.1 mmol/kg of gadopentetate dimeglumine bismethoxyethylamide (gadoversetamide, Opti-MARK, Mallinckrodt Imaging) at 2 mL/s followed by a 20-mL saline flush through a power injector (Spectris, Medrad). The scanned volume was imaged as 60-88 contiguous partitions with 3-mm nominal slice thickness and the same field of view and matrix size as image set 1. Recovery time/TE was 4.5/1.5; flip angle, 15°; bandwidth, 490 Hz/pixel. Acquisition time was 15-20 seconds per sequence. Gadolinium-enhanced images included the hepatic arterial, portal venous, and equilibrium phases.

Scanning delay for the arterial phase was chosen so that central k-space acquisition would occur 2-3 seconds after arrival of the gadolinium bolus in the hepatic arteries (to allow transit of gadolinium to arterially fed tumors) and was calculated on the basis of a 2-mL test bolus of gadolinium. Portal venous and equilibrium phase images were obtained consecutively after intervals of 25-40 seconds to allow patient hyperventilation between acquisitions. Thus, T1-weighted images were acquired approximately 20, 60, and 120 seconds after gadolinium injection. These delays were consistent with double-contrast MRI protocols previously described [12, 14].

Pathology
According to routine protocol, freshly explanted livers were sectioned by a pathologist into 8- to 10-mm-thick contiguous sections, and pathologic examination was performed for suspicious macroscopic nodules. Nodules were considered suspicious if they varied in size, color, or texture from surrounding regenerative nodules. The pathologist also reviewed the clinical radiology reports and attempted to co-localize radiographic lesions with nodules found on tissue sections. The pathologist recorded the Couinaud segment, location relative to anatomic landmarks, and size of focal nodules. Tissue samples were submitted for histologic examination and stained with H and E for lesion characterization. One hepatopathologist (15 years of postfellowship experience) reviewed the histologic findings and classified nodules in a binary manner as HCC or not HCC. According to the American Liver Tumor Study Group criteria, each liver was assigned a pathologic tumor stage (T0, T1, T2, T3, T4a, T4b), which served as the reference standard.

Blinded Retrospective MRI Review
Two abdominal MRI radiologists independently reviewed the MR images with a PACS (IMPAX, Agfa HealthCare) using 2,048 x 2,560 pixel resolution gray-scale monitors. Both radiologists had more than 5 years of experience in abdominal MRI. Radiologist 1 had 6 months of experience and radiologist 2 had 3 years of experience with the described MRI technique. The radiologists were aware that patients had cirrhosis but were unaware of all other clinical, laboratory, pathologic, and imaging findings.

The radiologists recorded the presence and Couinaud segment of focal lesions and scored them according to an ordinal five-point rating scale as follows: 1, definitely not HCC; 2, probably not HCC; 3, indeterminate for HCC; 4, probably HCC; and 5, definitely HCC. Because the focus of this study was to determine the accuracy for staging and diagnosis of HCC, no attempt was made to characterize nonmalignant hepatic nodules. The radiologists were aware that for statistical analyses, ratings would be dichotomized as HCC ( 4) or not HCC ( 3). Each radiologist's per-patient cancer rating was defined as the highest per-lesion rating.

Radiologists measured lesions with positive ratings ( 4) to the nearest millimeter (long axis) and assessed whether such lesions grossly involved intrahepatic portal or hepatic veins. According to American Liver Tumor Study Group definitions [8, 11] (Table 2), each radiologist's MRI findings were converted into a per-patient MRI tumor stage (T0, T1, T2, T3, T4a, T4b). The MRI stage was based on the size and number of lesions positively identified (rating, 4) at image review and whether gross intrahepatic vessel involvement was observed. Node and metastasis staging with MRI was not attempted.

View larger version (8K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 2 —Graph depicts degree of understaging and overstaging by each radiologist for all 48 patients with cirrhosis. MRI staging accuracy is determined by subtracting pathologic tumor stage from MRI tumor stage for each radiologist. Negative values indicate understaging with MRI; positive values indicate overstaging. For example, -2 represents understaging by two stages; 0, correct staging. *For both radiologist 1 and radiologist 2, one patient had disease understaged by four stages owing to undetected vascular invasion.

View larger version (7K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 3A —Receiver operating characteristic curves. Plots depict curves for each radiologist on per-patient (A) and per-lesion (B) basis. Per-patient and per-lesion areas under curve are shown in Table 4.

View larger version (7K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 3B —Receiver operating characteristic curves. Plots depict curves for each radiologist on per-patient (A) and per-lesion (B) basis. Per-patient and per-lesion areas under curve are shown in Table 4.

Gadolinium-enhanced T1-weighted dynamic images were reviewed for assessment of vascularity. In general, features that suggested malignancy on gadolinium-enhanced images included arterial hyperenhancement with signal intensity greater than that of liver in the hepatic arterial phase; venous washout with hypointensity relative to surrounding liver in the portal venous or equilibrium phase; heterogeneous or mosaic enhancement; or presence of discrete capsule or pseudocapsule. Fading of signal intensity to isointensity on portal venous and equilibrium phase images was generally considered benign for subcentimeter-size lesions, indeterminate for 1- to 2-cm lesions, and malignant for lesions 2 cm or larger. Lesions without enhancement and lesions with enhancement that paralleled aortic enhancement in all phases were assumed to be benign [24].

To score lesions, each radiologist reviewed the two image sets (SPIO enhanced and gadolinium enhanced) in conjunction and assigned a final rating based on the combined information. Lesions with malignant features on both image sets were assigned a high score (e.g., 5). Lesions with benign features on both image sets were assigned a low score (e.g., 1). Lesions with discordant findings (suspicious on one image set but not the other) were assigned intermediate ratings. The final score (2, 3, or 4) was determined by whether the cumulative information favored benignity or malignancy.

Unblinded Retrospective MRI Review
After completion of the blinded review, MR images were re-reviewed in conjunction with pathology reports and with any additional cross-sectional imaging studies of the liver. Lesions detected on the initial, blinded MRI review were classified as true-positive if the lesions matched pathologically confirmed HCCs in location (Couinaud segment and relation to anatomic landmarks). Lesions were classified as false-positive if they did not match pathologically confirmed HCCs in location, either because the patient had no HCC or because pathologically confirmed HCCs were in a different location. False-negative MRI findings were defined as pathologically confirmed HCCs that did not have a corresponding lesion detected at the blinded MRI review in the same location. MRI features of true-positive, false-positive, and, if visible at unblinded review, false-negative lesions were recorded. An effort was made to identify the cause of false-positive findings.

Statistical Analysis
For each radiologist, the total accuracy of MRI stage for prediction of pathologic stage was calculated. Total accuracy was defined as the number of patients in which MRI stage equaled pathologic stage divided by the number of all patients. Ninety-five percent CIs were computed by recalculation of each radiologist's staging accuracy on 1,500 resampled patient data sets. Each resampled data set was stratified according to HCC and non-HCC groups to maintain constant proportions in the resampling process [25]. A subanalysis of staging accuracy was performed for all patients who underwent transplantation within 90 days of imaging. The frequency and degree of understaging and overstaging were determined.

The sensitivity and positive predictive value (PPV) of MRI for identification of patients with disease in pathologic stage T2 were assessed. For this analysis, the stage scores (MRI and pathologic) were dichotomized as stage 2 or not stage 2. Sensitivity was defined as the number of patients with stage T2 disease accurately identified on MRI divided by the number of patients with pathologic stage T2 disease. A subanalysis was performed for patients with an MRI-transplant interval of 90 days or less.

For assessment of diagnostic performance, receiver operating characteristic curves were generated for each radiologist on a per-patient and per-lesion basis with the radiologists' five-point ordinal ratings. The area under each receiver operating characteristic curve was computed. With the dichotomized lesion ratings (< 4, no HCC; 4, HCC), the following diagnostic performance parameters were calculated for each radiologist: perpatient sensitivity, per-patient specificity, per-lesion sensitivity, and per-lesion PPV. Intraclass correlation coefficients were computed to assess inter-radiologist agreement for overall staging accuracy and diagnostic accuracy for HCC on a per-patient and a per-lesion basis. Because the number of potential lesions in a liver is unlimited, the number of true-negative lesions could not be meaningfully determined, and per-lesion specificity was not calculated. Descriptive summaries of true-positive, false-positive, and false-negative lesion reports were generated.

Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Subjects
Of the 48 study subjects, 23 had no HCC (stage T0), and 25 had HCC (three, stage T1; 18, T2; one, T3; one, T4a; two, T4b). In total, there were 37 HCC nodules (median size, 23 mm; range, 7-74 mm). The median MRI-transplantation interval was 69 days (range, 7-284 days) for patients with HCC-positive results and 60 days (range, 7-698 days) for patients with HCC-negative results. Thirty patients had an MRI-transplantation interval of 90 days or less, and 42 had an MRI-transplantation interval of 180 days or less.

Staging Accuracy
The total accuracy of MRI in prediction of pathologic tumor stage was 81% for radiologist 1 and 85% for radiologist 2. A subanalysis of patients with MRI-transplantation intervals of less than 90 days yielded similar results (Table 3). The staging sensitivity and PPV for stage T2 disease were 83% and 88%, respectively, for radiologist 1 and 94% and 90% for radiologist 2. For patients with an MRI-transplantation interval of 90 days or less, the staging sensitivity and PPV for stage T2 disease were 75% and 90%, respectively, for radiologist 1 and 90% and 92% for radiologist 2.

Understaging and Overstaging
In three patients disease was overstaged, in three cases by radiologist 1, in two cases by radiologist 2, and in two cases by both radiologists (Fig. 2). The three patients with overstaged disease had a total of seven false-positive lesions (mean size, 12 mm; range, 7-19 mm). In eight patients, disease was understaged, in six cases by radiologist 1, in five cases by radiologist 2, and in four cases by both radiologists. The eight patients with understaged disease had a total of seven false-negative lesions (mean size, 15 mm; range, 7-20 mm). Two of the patients with understaged disease (imaged 35 and 177 days before transplantation) were correctly classified as having HCC but had undetected vascular invasion discovered at analysis of the explant. The invasion was not visible at unblinded review.

Diagnostic Accuracy
Per-patient and per-lesion receiver operating characteristic curves were generated (Fig. 3A, 3B); areas under the curve ranged from 0.90 to 0.98 depending on radiologist and analysis (Table 4). Per-patient sensitivity and specificity ranged from 91% to 96%; per-lesion sensitivity and PPV ranged from 81% to 89%. Results were similar in the subset of patients with an MRI-transplantation interval of 90 days or less.

Agreement Between Radiologists
The intraclass correlation coefficient between the two radiologists for assigning MRI tumor stage was 92%. For diagnosis of HCC on a per-patient and per-lesion basis, the correlation coefficients were 94% and 82%, respectively.

True-Positive Lesion Reports
Of 37 HCCs, 33 true-positive lesion reports were made (29 by both radiologists). Twenty-six HCCs (mean size, 27 mm; range, 8-74 mm) had concordant findings (high signal intensity on SPIO-enhanced images, arterial hypervascularity on dynamic gadolinium-enhanced images) (Fig. 4A, 4B, 4C, 4D, 4E, 4F, 4G, 4H). Seven HCCs had discordant findings: two (11 and 24 mm) were hypervascular but on SPIO-enhanced images had signal intensity similar to that of liver; five (mean size, 24 mm; range, 15-44 mm) had high signal intensity on SPIO-enhanced images but became only minimally enhanced with gadolinium (Fig. 5A, 5B, 5C, 5D, 5E, 5F, 5G, 5H).

View larger version (100K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4A —51-year-old man with hypervascular hepatocellular carcinoma (HCC) in pathologic stage T2. Two nodules measuring 23 (black arrows) and 24 (white arrows) mm are present in segments V and III of liver. Dynamic 3D T1-weighted transverse fat-saturated spoiled gradient-recalled echo (SPGR) MR images before gadolinium administration (A) and during hepatic arterial (B), portal venous (C), and equilibrium (D) phases after gadolinium administration show both nodules are enhanced in arterial phase (B) and wash out to become hypointense relative to surrounding liver in equilibrium phase (D). Degree of arterial phase enhancement is greater for segment III than segment V nodule.

View larger version (123K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4B —51-year-old man with hypervascular hepatocellular carcinoma (HCC) in pathologic stage T2. Two nodules measuring 23 (black arrows) and 24 (white arrows) mm are present in segments V and III of liver. Dynamic 3D T1-weighted transverse fat-saturated spoiled gradient-recalled echo (SPGR) MR images before gadolinium administration (A) and during hepatic arterial (B), portal venous (C), and equilibrium (D) phases after gadolinium administration show both nodules are enhanced in arterial phase (B) and wash out to become hypointense relative to surrounding liver in equilibrium phase (D). Degree of arterial phase enhancement is greater for segment III than segment V nodule.

View larger version (140K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4C —51-year-old man with hypervascular hepatocellular carcinoma (HCC) in pathologic stage T2. Two nodules measuring 23 (black arrows) and 24 (white arrows) mm are present in segments V and III of liver. Dynamic 3D T1-weighted transverse fat-saturated spoiled gradient-recalled echo (SPGR) MR images before gadolinium administration (A) and during hepatic arterial (B), portal venous (C), and equilibrium (D) phases after gadolinium administration show both nodules are enhanced in arterial phase (B) and wash out to become hypointense relative to surrounding liver in equilibrium phase (D). Degree of arterial phase enhancement is greater for segment III than segment V nodule.

View larger version (134K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4D —51-year-old man with hypervascular hepatocellular carcinoma (HCC) in pathologic stage T2. Two nodules measuring 23 (black arrows) and 24 (white arrows) mm are present in segments V and III of liver. Dynamic 3D T1-weighted transverse fat-saturated spoiled gradient-recalled echo (SPGR) MR images before gadolinium administration (A) and during hepatic arterial (B), portal venous (C), and equilibrium (D) phases after gadolinium administration show both nodules are enhanced in arterial phase (B) and wash out to become hypointense relative to surrounding liver in equilibrium phase (D). Degree of arterial phase enhancement is greater for segment III than segment V nodule.

View larger version (140K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4E —51-year-old man with hypervascular hepatocellular carcinoma (HCC) in pathologic stage T2. Two nodules measuring 23 (black arrows) and 24 (white arrows) mm are present in segments V and III of liver. Superparamagnetic iron oxide-enhanced 2D SPGR images obtained before administration of gadolinium with TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echo-train spin-echo image with TE of 90 milliseconds (H) show both lesions have high signal intensity suggestive of phagocyte depletion. Each radiologist correctly diagnosed both HCC nodules and staged HCC burden.

View larger version (129K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4F —51-year-old man with hypervascular hepatocellular carcinoma (HCC) in pathologic stage T2. Two nodules measuring 23 (black arrows) and 24 (white arrows) mm are present in segments V and III of liver. Superparamagnetic iron oxide-enhanced 2D SPGR images obtained before administration of gadolinium with TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echo-train spin-echo image with TE of 90 milliseconds (H) show both lesions have high signal intensity suggestive of phagocyte depletion. Each radiologist correctly diagnosed both HCC nodules and staged HCC burden.

View larger version (136K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4G —51-year-old man with hypervascular hepatocellular carcinoma (HCC) in pathologic stage T2. Two nodules measuring 23 (black arrows) and 24 (white arrows) mm are present in segments V and III of liver. Superparamagnetic iron oxide-enhanced 2D SPGR images obtained before administration of gadolinium with TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echo-train spin-echo image with TE of 90 milliseconds (H) show both lesions have high signal intensity suggestive of phagocyte depletion. Each radiologist correctly diagnosed both HCC nodules and staged HCC burden.

View larger version (146K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4H —51-year-old man with hypervascular hepatocellular carcinoma (HCC) in pathologic stage T2. Two nodules measuring 23 (black arrows) and 24 (white arrows) mm are present in segments V and III of liver. Superparamagnetic iron oxide-enhanced 2D SPGR images obtained before administration of gadolinium with TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echo-train spin-echo image with TE of 90 milliseconds (H) show both lesions have high signal intensity suggestive of phagocyte depletion. Each radiologist correctly diagnosed both HCC nodules and staged HCC burden.

View larger version (94K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 5A —53-year-old woman with pathologic tumor stage T2 hypovascular hepatocellular carcinoma (one 23-mm nodule in segment II). Both radiologists detected lesion and correctly assigned MRI tumor stage of T2. Asterisk indicates ascites. Dynamic 3D T1-weighted transverse fat-saturated spoiled gradient-recalled echo (SPGR) MR image before gadolinium administration shows high signal intensity facilitating recognition of nodule (arrow).

View larger version (107K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 5B —53-year-old woman with pathologic tumor stage T2 hypovascular hepatocellular carcinoma (one 23-mm nodule in segment II). Both radiologists detected lesion and correctly assigned MRI tumor stage of T2. Asterisk indicates ascites. Dynamic 3D T1-weighted transverse fat-saturated SPGR MR images in hepatic arterial (B), portal venous (C), and equilibrium (D) phases after gadolinium administration show poor visibility of hepatocellular carcinoma nodule (arrows). Minimal enhancement of nodule is evident in B.

View larger version (114K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 5C —53-year-old woman with pathologic tumor stage T2 hypovascular hepatocellular carcinoma (one 23-mm nodule in segment II). Both radiologists detected lesion and correctly assigned MRI tumor stage of T2. Asterisk indicates ascites. Dynamic 3D T1-weighted transverse fat-saturated SPGR MR images in hepatic arterial (B), portal venous (C), and equilibrium (D) phases after gadolinium administration show poor visibility of hepatocellular carcinoma nodule (arrows). Minimal enhancement of nodule is evident in B.

View larger version (113K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 5D —53-year-old woman with pathologic tumor stage T2 hypovascular hepatocellular carcinoma (one 23-mm nodule in segment II). Both radiologists detected lesion and correctly assigned MRI tumor stage of T2. Asterisk indicates ascites. Dynamic 3D T1-weighted transverse fat-saturated SPGR MR images in hepatic arterial (B), portal venous (C), and equilibrium (D) phases after gadolinium administration show poor visibility of hepatocellular carcinoma nodule (arrows). Minimal enhancement of nodule is evident in B.

View larger version (121K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 5E —53-year-old woman with pathologic tumor stage T2 hypovascular hepatocellular carcinoma (one 23-mm nodule in segment II). Both radiologists detected lesion and correctly assigned MRI tumor stage of T2. Asterisk indicates ascites. Superparamagnetic iron oxide-enhanced 2D SPGR transverse MR images obtained before administration of gadolinium at TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echo-train spin-echo image at TE of 90 milliseconds (H) depict nodule (arrows) owing to high signal intensity. Image quality is limited by patient's obesity and marked ascites (asterisks).

View larger version (124K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 5F —53-year-old woman with pathologic tumor stage T2 hypovascular hepatocellular carcinoma (one 23-mm nodule in segment II). Both radiologists detected lesion and correctly assigned MRI tumor stage of T2. Asterisk indicates ascites. Superparamagnetic iron oxide-enhanced 2D SPGR transverse MR images obtained before administration of gadolinium at TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echo-train spin-echo image at TE of 90 milliseconds (H) depict nodule (arrows) owing to high signal intensity. Image quality is limited by patient's obesity and marked ascites (asterisks).

View larger version (130K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 5G —53-year-old woman with pathologic tumor stage T2 hypovascular hepatocellular carcinoma (one 23-mm nodule in segment II). Both radiologists detected lesion and correctly assigned MRI tumor stage of T2. Asterisk indicates ascites. Superparamagnetic iron oxide-enhanced 2D SPGR transverse MR images obtained before administration of gadolinium at TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echo-train spin-echo image at TE of 90 milliseconds (H) depict nodule (arrows) owing to high signal intensity. Image quality is limited by patient's obesity and marked ascites (asterisks).

View larger version (132K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 5H —53-year-old woman with pathologic tumor stage T2 hypovascular hepatocellular carcinoma (one 23-mm nodule in segment II). Both radiologists detected lesion and correctly assigned MRI tumor stage of T2. Asterisk indicates ascites. Superparamagnetic iron oxide-enhanced 2D SPGR transverse MR images obtained before administration of gadolinium at TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echo-train spin-echo image at TE of 90 milliseconds (H) depict nodule (arrows) owing to high signal intensity. Image quality is limited by patient's obesity and marked ascites (asterisks).

False-Positive Lesion Reports
Eleven false-positive lesion reports were made (three by both radiologists). At unblinded review, three falsely reported lesions (5, 5, and 18 mm) corresponded to vessels mistaken for nodules, and three (10, 11, and 15 mm) were attributed to focal areas of fibrosis. Another three falsely reported lesions (7, 11, and 19 mm) corresponded in size and location to histologically confirmed benign regenerative nodules identified at gross pathologic examination. The cause of the last two MRI-detected lesions (7 and 9 mm) was not established. The 7-mm lesion had similar signal intensity to two histologically proven HCC nodules elsewhere in the liver. Both lesions without an established cause were conservatively classified false-positive.

Of the eight false-positive lesions that corresponded to parenchymal findings (as opposed to vessels), four (mean size, 11 mm; range 7-19 mm) had concordant findings on the SPIO-enhanced and gadolinium-enhanced image sets. These findings were focal abnormalities with high signal intensity on gadolinium-enhanced arterial phase images and on SPIO-enhanced images. Four lesions had discordant findings. One discordant lesion (11 mm) had high signal intensity on SPIO-enhanced images but did not become enhanced with gadolinium in the arterial phase. Two discordant lesions (10 and 15 mm) had arterial phase hypervascularity on dynamic gadolinium-enhanced images but isointensity relative to background liver on SPIO-enhanced images. The last discordant lesion (9 mm) had subtle venous washout on dynamic gadolinium-enhanced images and hypointensity on SPIO-enhanced images (Fig. 6A, 6B, 6C, 6D, 6E, 6F, 6G, 6H).

View larger version (76K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 6A —54-year-old man with cirrhosis and false-positive lesion report. Radiologists reached different decisions on final interpretation. One radiologist scored nodule benign; one radiologist scored nodule malignant. Pathology report mentioned only regenerative nodules in liver segment, and no atypical nodules were found. Malignant interpretation by one radiologist was classified as false-positive lesion report. Dynamic 3D T1-weighted transverse fat-saturated spoiled gradient-recalled echo (SPGR) image before gadolinium administration shows 9-mm nodule (arrow) in segment II has higher signal intensity than liver.

View larger version (84K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 6B —54-year-old man with cirrhosis and false-positive lesion report. Radiologists reached different decisions on final interpretation. One radiologist scored nodule benign; one radiologist scored nodule malignant. Pathology report mentioned only regenerative nodules in liver segment, and no atypical nodules were found. Malignant interpretation by one radiologist was classified as false-positive lesion report. Dynamic 3D T1-weighted transverse fat-saturated SPGR images in hepatic arterial phase (B), portal venous phase (C), and equilibrium phase (D) after administration of gadolinium show nodule (arrows) does not become enhanced in arterial phase (B) but washes out relative to liver on delayed images (C and D).

View larger version (97K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 6C —54-year-old man with cirrhosis and false-positive lesion report. Radiologists reached different decisions on final interpretation. One radiologist scored nodule benign; one radiologist scored nodule malignant. Pathology report mentioned only regenerative nodules in liver segment, and no atypical nodules were found. Malignant interpretation by one radiologist was classified as false-positive lesion report. Dynamic 3D T1-weighted transverse fat-saturated SPGR images in hepatic arterial phase (B), portal venous phase (C), and equilibrium phase (D) after administration of gadolinium show nodule (arrows) does not become enhanced in arterial phase (B) but washes out relative to liver on delayed images (C and D).

View larger version (97K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 6D —54-year-old man with cirrhosis and false-positive lesion report. Radiologists reached different decisions on final interpretation. One radiologist scored nodule benign; one radiologist scored nodule malignant. Pathology report mentioned only regenerative nodules in liver segment, and no atypical nodules were found. Malignant interpretation by one radiologist was classified as false-positive lesion report. Dynamic 3D T1-weighted transverse fat-saturated SPGR images in hepatic arterial phase (B), portal venous phase (C), and equilibrium phase (D) after administration of gadolinium show nodule (arrows) does not become enhanced in arterial phase (B) but washes out relative to liver on delayed images (C and D).

View larger version (108K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 6E —54-year-old man with cirrhosis and false-positive lesion report. Radiologists reached different decisions on final interpretation. One radiologist scored nodule benign; one radiologist scored nodule malignant. Pathology report mentioned only regenerative nodules in liver segment, and no atypical nodules were found. Malignant interpretation by one radiologist was classified as false-positive lesion report. Superparamagnetic iron oxide-enhanced 2D SPGR transverse MR images obtained before administration of gadolinium with TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echo-train spin-echo image with TE of 90 milliseconds (H) show nodule (arrows) with low signal intensity relative to liver as TE increases, suggesting elevated phagocytic function.

View larger version (110K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 6F —54-year-old man with cirrhosis and false-positive lesion report. Radiologists reached different decisions on final interpretation. One radiologist scored nodule benign; one radiologist scored nodule malignant. Pathology report mentioned only regenerative nodules in liver segment, and no atypical nodules were found. Malignant interpretation by one radiologist was classified as false-positive lesion report. Superparamagnetic iron oxide-enhanced 2D SPGR transverse MR images obtained before administration of gadolinium with TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echo-train spin-echo image with TE of 90 milliseconds (H) show nodule (arrows) with low signal intensity relative to liver as TE increases, suggesting elevated phagocytic function.

View larger version (121K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 6G —54-year-old man with cirrhosis and false-positive lesion report. Radiologists reached different decisions on final interpretation. One radiologist scored nodule benign; one radiologist scored nodule malignant. Pathology report mentioned only regenerative nodules in liver segment, and no atypical nodules were found. Malignant interpretation by one radiologist was classified as false-positive lesion report. Superparamagnetic iron oxide-enhanced 2D SPGR transverse MR images obtained before administration of gadolinium with TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echo-train spin-echo image with TE of 90 milliseconds (H) show nodule (arrows) with low signal intensity relative to liver as TE increases, suggesting elevated phagocytic function.

View larger version (111K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 6H —54-year-old man with cirrhosis and false-positive lesion report. Radiologists reached different decisions on final interpretation. One radiologist scored nodule benign; one radiologist scored nodule malignant. Pathology report mentioned only regenerative nodules in liver segment, and no atypical nodules were found. Malignant interpretation by one radiologist was classified as false-positive lesion report. Superparamagnetic iron oxide-enhanced 2D SPGR transverse MR images obtained before administration of gadolinium with TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echo-train spin-echo image with TE of 90 milliseconds (H) show nodule (arrows) with low signal intensity relative to liver as TE increases, suggesting elevated phagocytic function.

View larger version (111K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 7A —49-year-old man with pathologic tumor stage T2 disease (two nodules measuring 8 [not shown] and 15 mm [arrows]) and false-negative lesion report. One radiologist detected hepatocellular carcinoma on superparamagnetic iron oxide (SPIO)-enhanced images and correctly assigned stage T2; other radiologist did not detect lesion on any images (error of observation). Dynamic 3D T1-weighted transverse fat-saturated spoiled gradient-recalled echo (SPGR) images before gadolinium administration (A) and during hepatic arterial (B), portal venous (C), and equilibrium (D) phases after administration of gadolinium show poor visibility of nodules. In B, 15-mm nodule (arrow) in segment VI has ill-defined gadolinium enhancement.

View larger version (130K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 7B —49-year-old man with pathologic tumor stage T2 disease (two nodules measuring 8 [not shown] and 15 mm [arrows]) and false-negative lesion report. One radiologist detected hepatocellular carcinoma on superparamagnetic iron oxide (SPIO)-enhanced images and correctly assigned stage T2; other radiologist did not detect lesion on any images (error of observation). Dynamic 3D T1-weighted transverse fat-saturated spoiled gradient-recalled echo (SPGR) images before gadolinium administration (A) and during hepatic arterial (B), portal venous (C), and equilibrium (D) phases after administration of gadolinium show poor visibility of nodules. In B, 15-mm nodule (arrow) in segment VI has ill-defined gadolinium enhancement.

View larger version (131K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 7C —49-year-old man with pathologic tumor stage T2 disease (two nodules measuring 8 [not shown] and 15 mm [arrows]) and false-negative lesion report. One radiologist detected hepatocellular carcinoma on superparamagnetic iron oxide (SPIO)-enhanced images and correctly assigned stage T2; other radiologist did not detect lesion on any images (error of observation). Dynamic 3D T1-weighted transverse fat-saturated spoiled gradient-recalled echo (SPGR) images before gadolinium administration (A) and during hepatic arterial (B), portal venous (C), and equilibrium (D) phases after administration of gadolinium show poor visibility of nodules. In B, 15-mm nodule (arrow) in segment VI has ill-defined gadolinium enhancement.

View larger version (137K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 7D —49-year-old man with pathologic tumor stage T2 disease (two nodules measuring 8 [not shown] and 15 mm [arrows]) and false-negative lesion report. One radiologist detected hepatocellular carcinoma on superparamagnetic iron oxide (SPIO)-enhanced images and correctly assigned stage T2; other radiologist did not detect lesion on any images (error of observation). Dynamic 3D T1-weighted transverse fat-saturated spoiled gradient-recalled echo (SPGR) images before gadolinium administration (A) and during hepatic arterial (B), portal venous (C), and equilibrium (D) phases after administration of gadolinium show poor visibility of nodules. In B, 15-mm nodule (arrow) in segment VI has ill-defined gadolinium enhancement.

View larger version (150K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 7E —49-year-old man with pathologic tumor stage T2 disease (two nodules measuring 8 [not shown] and 15 mm [arrows]) and false-negative lesion report. One radiologist detected hepatocellular carcinoma on superparamagnetic iron oxide (SPIO)-enhanced images and correctly assigned stage T2; other radiologist did not detect lesion on any images (error of observation). SPIO-enhanced 2D SPGR transverse MR images with TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echotrain spin-echo image with TE of 90 milliseconds (H) show hepatocellular carcinoma (arrows) visible as sharply circumscribed area of high signal intensity.

View larger version (148K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 7F —49-year-old man with pathologic tumor stage T2 disease (two nodules measuring 8 [not shown] and 15 mm [arrows]) and false-negative lesion report. One radiologist detected hepatocellular carcinoma on superparamagnetic iron oxide (SPIO)-enhanced images and correctly assigned stage T2; other radiologist did not detect lesion on any images (error of observation). SPIO-enhanced 2D SPGR transverse MR images with TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echotrain spin-echo image with TE of 90 milliseconds (H) show hepatocellular carcinoma (arrows) visible as sharply circumscribed area of high signal intensity.

View larger version (147K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 7G —49-year-old man with pathologic tumor stage T2 disease (two nodules measuring 8 [not shown] and 15 mm [arrows]) and false-negative lesion report. One radiologist detected hepatocellular carcinoma on superparamagnetic iron oxide (SPIO)-enhanced images and correctly assigned stage T2; other radiologist did not detect lesion on any images (error of observation). SPIO-enhanced 2D SPGR transverse MR images with TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echotrain spin-echo image with TE of 90 milliseconds (H) show hepatocellular carcinoma (arrows) visible as sharply circumscribed area of high signal intensity.

View larger version (131K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 7H —49-year-old man with pathologic tumor stage T2 disease (two nodules measuring 8 [not shown] and 15 mm [arrows]) and false-negative lesion report. One radiologist detected hepatocellular carcinoma on superparamagnetic iron oxide (SPIO)-enhanced images and correctly assigned stage T2; other radiologist did not detect lesion on any images (error of observation). SPIO-enhanced 2D SPGR transverse MR images with TE of 2.6 milliseconds (E), 4.8 milliseconds (F), and 6.6 milliseconds (T2*-weighted) (G) and T2-weighted 2D echotrain spin-echo image with TE of 90 milliseconds (H) show hepatocellular carcinoma (arrows) visible as sharply circumscribed area of high signal intensity.

Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
Diagnostic imaging plays a critical role in pretransplantation staging of HCC. It can expedite transplantation in patients with stage T2 tumors and help remove from the waiting list patients with a tumor burden that exceeds transplantation guidelines. Despite this critical role, study results [7, 26] indicate that current imaging techniques are inaccurate for HCC staging.

In eight studies (Table 1) investigators evaluated imaging accuracy in preoperative HCC staging using the liver explant findings as the sole reference standard. Using single-detector helical CT, Valls et al. [27] found the highest accuracy among these studies. The radiologists in that study used a binary staging system (stage T 2 vs T > 2), which, although appropriate for that group's institutional transplantation protocol, is not applicable in the United States, where the utility of imaging in differentiation of stage T2 disease from all other stages is clinically important. In addition, the radiologists in the study by Valls et al. interpreted the images in consensus, which may limit the ability to generalize the findings to standard clinical practice. In comparison, another group [4] using a similar single-detector helical CT method reported a staging accuracy of only 58%. Other studies [10, 11, 26] also show low staging accuracy (35-57%) of MRI and MDCT.

In our study, a double-contrast MRI protocol performed in a clinical setting had 81-85% staging accuracy depending on the radiologist. This accuracy is higher than that in all previous studies of which we are aware in which a nonbinary staging classification system (Table 1) was used. The frequency of understaging and overstaging was low, and the sensitivity for identification of stage T2 disease was high. Thus, the double-contrast technique had high accuracy in correct identification of patients who, according to current guidelines, are most likely to benefit from transplantation.

The administration of contrast agents narrows the differential diagnosis of HCC versus nonmalignant cirrhosis-associated nodules but, as emphasized by Bhartia et al. [12], imaging features of benign and malignant entities may overlap when either SPIO or gadolinium is used as a sole contrast agent. Double-contrast MRI is performed in an attempt to overcome this difficulty by use of two contrast agents. SPIO-enhanced images can be used to characterize HCCs with atypical or indeterminate gadolinium enhancement features and to detect hypovascular HCCs missed on gadolinium-enhanced dynamic acquisitions. Likewise, gadolinium-enhanced dynamic acquisitions can be used to detect malignant HCCs with preserved phagocytic function. Thus, the complementary information provided by SPIOs and gadolinium chelates helps to characterize otherwise nonspecific nodules and to increase diagnostic confidence for lesions with concordant findings.

Among the 48 study patients, disease was understaged by one or both radiologists in eight patients. Explant analysis showed two of these patients had vascular invasion. Although the invasion was not detected by either radiologist and was not visible at unblinded review, it should not be assumed that double-contrast imaging has low sensitivity in the detection of vascular invasion. Patients with vascular invasion detected on preoperative imaging do not undergo liver transplantation; thus, the use of the explant findings as the reference standard eliminated from our study patients in whom findings at double-contrast MRI examinations were positive for invasion.

Of the seven missed HCCs (false-negative results), five were detectable at unblinded review, suggesting that further improvements in diagnostic performance may be possible. Three of the five misses were errors of observation, and two were hypovascular nodules that had high signal intensity on SPIO-enhanced images but were surrounded by and partially obscured by fibrosis of high signal intensity. Burrel et al. [4] reported that on SPIO-enhanced imaging, intense fibrosis can cause false-positive lesion reports because the signal intensity alterations of fibrosis can mimic those of nodules. In this study, we observed that intense fibrosis can cause false-negative lesion reports by obscuring nodules. We speculate that increasing the spatial resolution of SPIO-enhanced images will lead to better delineation of the reticulated pattern of fibrosis and improve the visibility of nodules surrounded by dense fibrosis. Two false-negative lesions were invisible in all sequences and presumably had vascularity and phagocytic function similar to those of the rest of the liver. Visualization of such lesions is likely to remain difficult on double-contrast imaging regardless of future technical refinements.

Of the 11 false-positive lesion reports, three corresponded to blood vessels rather than to parenchymal nodules. Misinterpretation of the vessels as nodules was probably due in part to the low spatial imaging resolution used in the imaging protocol. In addition, use of interslice gaps for 2D images made it difficult to delineate the out-of-plane trajectory of vessels. Three false-positive lesion reports corresponded to regenerative nodules and three to focal areas of fibrosis. Further investigation is needed to identify imaging features that allow reliable differentiation of benign from malignant causes of altered signal intensity on MRI. The cause of two false-positive lesion reports was not established. One of these lesions (7-mm nodule) had signal-intensity characteristics identical to those of two coincident HCCs on all images. This finding raises the possibility that the lesion was an HCC that went undetected at pathologic examination because of sampling error with the manual sectioning technique.

Despite our promising results, the use of double-contrast MRI remains controversial [24]. Compared with single-agent approaches, the double-contrast method is more complex logistically, is more expensive, and, because it exposes the patient to two drugs, poses greater theoretic risk [28]. Furthermore, only five of the 33 true-positive lesions in our study were detected mainly on the basis of SPIO-enhanced imaging findings. Thus, even modest improvement in the performance of dynamic gadolinium-enhanced imaging may enable high staging accuracy without the use of a second agent.

A limitation of our study was that we did not directly compare double-versus single-contrast approaches to HCC staging. A prospective clinical trial is needed for this comparison. Our study also was limited by its retrospective design, which may have introduced positive selection bias. The absence of unenhanced images obtained before administration of SPIO was an additional limitation. Such images were eliminated from our imaging protocol in 2000, before the period defined for this study, because unenhanced imaging increased the logistic difficulty [13, 29] of the MRI examination and anecdotally added little diagnostic benefit to evaluation of patients with cirrhosis. Although some authors [30] have reported that unenhanced images help decrease false-positive lesion reports due to small vessels on SPIO-enhanced images, we and others [13, 29] had a low number of false-positive findings due to vessels even in the absence of unenhanced images. A further limitation was the relatively long MRI-transplantation interval in some patients with HCC. This factor did not meaningfully alter our results, however, because similarly high accuracy was found in the subanalysis of 30 patients with an MRI-transplantation interval of 90 days or less.

A technical limitation of our imaging protocol was that equilibrium phase images were acquired approximately 120 seconds after gadolinium injection. Reports [15, 28, 31] indicate that scan delays of 180-240 seconds are superior because they aid assessment of venous washout and characterization of nodule vascularity. Thus, lengthening the scan delay of the equilibrium phase acquisition may further improve performance. Nevertheless, the scan delays we report are similar to those used in other double-contrast studies [12, 14] and facilitate comparison of diagnostic accuracy.

A theoretic limitation of our imaging protocol was that in addition to T2^* shortening, iron oxides also cause modest T1 shortening; thus, their administration may reduce the efficacy of subsequent gadolinium injection on dynamic T1-weighted sequences. Kim et al. [24], for example, observed that SPIO accumulation in background liver impaired assessment of dynamic lesion enhancement. Our study was not intended to assess whether SPIO should be given before or after gadolinium, only to show that a protocol in which both agents are administered can achieve high accuracy in staging and diagnosis. Nevertheless, it is possible that giving SPIO after gadolinium would have yielded even higher staging and diagnostic accuracy, but giving SPIO after gadolinium usually requires two scanning sessions and is more time-consuming than the protocol used in our study [13].

The SPIO-enhanced data set obtained before gadolinium administration had technical limitations, including relatively low spatial resolution and use of interslice gaps. These factors likely contributed to occasional difficulty in co-localizing small nodules in different sequences, identifying nodules buried in dense fibrosis, and differentiating small nodules from end-on vessels. Increasing spatial resolution and removing interslice gaps may overcome some of these difficulties.

We found that a clinical double-contrast MRI protocol has high accuracy for staging of HCC in a population with cirrhosis. The staging accuracy was higher than that reported elsewhere with other imaging techniques with explant pathologic findings as the reference. Because we did not compare gadolinium-enhanced with double-contrast MRI, we do not advocate the double-contrast technique as being superior to single-agent techniques and believe that only in a prospective clinical trial should double-contrast and gadolinium-only imaging be compared. This study, in conjunction with previous studies of double-contrast imaging, provides preliminary data to plan and support a future clinical trial. Although our study had limitations, it showed, contrary to the conclusions of previous HCC staging studies [7, 10], that a clinical MRI protocol can have high accuracy for staging of HCC. It is likely that the performance of this and other imaging techniques will continue to improve as technology advances and as our understanding of imaging findings progresses.

Acknowledgments
 
We thank Tanya Wolfson, University of California, San Diego, Department of Family and Preventive Medicine, for her statistical advice and contribution.

References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Wong LL. Current status of liver transplantation for hepatocellular cancer. Am J Surg 2002;183 : 309-316[CrossRef][Medline]
2. Yu EW, Chie WC, Chen TH. Does screening or surveillance for primary hepatocellular carcinoma with ultrasonography improve the prognosis of patients? Cancer J 2004;10 : 317-325[Medline]
3. Bruix J, Sherman M. Management of hepatocellular carcinoma. Hepatology 2005;42 : 1208-1236[CrossRef][Medline]
4. Burrel M, Llovet JM, Ayuso C, et al. MRI angiography is superior to helical CT for detection of HCC prior to liver transplantation: an explant correlation. Hepatology 2003;38 : 1034-1042[CrossRef][Medline]
5. 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]
6. Trotter JF. Selection of donors and recipients for living donor liver transplantation. Liver Transpl2000; 6:S52 -S58[CrossRef][Medline]
7. Freeman RB, Mithoefer A, Ruthazer R, et al. Optimizing staging for hepatocellular carcinoma before liver transplantation: a retrospective analysis of the UNOS/OPTN database. Liver Transpl2006; 12:1504 -1511[CrossRef][Medline]
8. Befeler AS, Hayashi PH, Di Bisceglie AM. Liver transplantation for hepatocellular carcinoma. Gastroenterology2005; 128:1752 -1764
9. Organ Procurement and Transplantation Network. Policy 3.6: Organ distribution—allocation of livers. Revised December 14, 2006. www.optn.org/policiesAndBylaws/policies.asp. Accessed December 28, 2006
10. Libbrecht L, Bielen D, Verslype C, et al. Focal lesions in cirrhotic explant livers: pathological evaluation and accuracy of pretransplantation imaging examinations. Liver Transpl2002; 8:749 -761[CrossRef][Medline]
11. Bharat A, Brown DB, Crippin JS, et al. Pre-liver transplantation locoregional adjuvant therapy for hepatocellular carcinoma as a strategy to improve longterm survival. J Am Coll Surg2006; 203:411 -420[CrossRef][Medline]
12. Bhartia B, Ward J, Guthrie JA, Robinson PJ. Hepatocellular carcinoma in cirrhotic livers: double-contrast thin-section MR imaging with pathologic correlation of explanted tissue. AJR2003; 180:577 -584[Abstract/Free Full Text]
13. Qayyum A, Thoeni RF, Coakley FV, Lu Y, Guay JP, Ferrell LD. Detection of hepatocellular carcinoma by ferumoxides-enhanced MR imaging in cirrhosis: incremental value of dynamic gadolinium-enhancement. J Magn Reson Imaging 2006; 23:17 -22[CrossRef][Medline]
14. Ward J, Guthrie JA, Scott DJ, et al. Hepatocellular carcinoma in the cirrhotic liver: double-contrast MR imaging for diagnosis. Radiology 2000;216 : 154-162[Abstract/Free Full Text]
15. Lutz AM, Willmann JK, Goepfert K, Marincek B, Weishaupt D. Hepatocellular carcinoma in cirrhosis: enhancement patterns at dynamic gadolinium- and superparamagnetic iron oxide-enhanced T1-weighted MR imaging. Radiology 2005;237 : 520-528[Abstract/Free Full Text]
16. Glockner JF, Forauer AR, Solomon H, Varma CR, Perman WH. Three-dimensional gadolinium-enhanced MR angiography of vascular complications after liver transplantation. AJR 2000;174 : 1447-1453[Abstract/Free Full Text]
17. Yamashita Y, Hatanaka Y, Yamamoto H, et al. Differential diagnosis of focal liver lesions: role of spin-echo and contrast-enhanced dynamic MR imaging. Radiology 1994;193 : 59-65[Abstract/Free Full Text]
18. Ohtomo K, Itai Y, Yoshikawa K, et al. Hepatic tumors: dynamic MR imaging. Radiology 1987;163 : 27-31[Abstract/Free Full Text]
19. Yoshida H, Itai Y, Ohtomo K, Kokubo T, Minami M, Yashiro N. Small hepatocellular carcinoma and cavernous hemangioma: differentiation with dynamic FLASH MR imaging with Gd-DTPA. Radiology1989; 171:339 -342[Abstract/Free Full Text]
20. Imai Y, Murakami T, Yoshida S, et al. Superparamagnetic iron oxide-enhanced magnetic resonance images of hepatocellular carcinoma: correlation with histological grading. Hepatology2000; 32:205 -212[CrossRef][Medline]
21. Lim JH, Choi D, Cho SK, et al. Conspicuity of hepatocellular nodular lesions in cirrhotic livers at ferumoxides-enhanced MR imaging: importance of Kupffer cell number. Radiology2001; 220:669 -676[Abstract/Free Full Text]
22. Tanaka M, Nakashima O, Wada Y, Kage M, Kojiro M. Pathomorphological study of Kupffer cells in hepatocellular carcinoma and hyperplastic nodular lesions in the liver. Hepatology 1996;24 : 807-812[CrossRef][Medline]
23. Matsuo M, Kanematsu M, Itoh K, et al. Detection of malignant hepatic tumors with ferumoxides-enhanced MRI: comparison of five gradient-recalled echo sequences with different TEs. AJR 2004; 182:235 -242[Abstract/Free Full Text]
24. Kim MJ, Kim JH, Chung JJ, Park MS, Lim JS, Oh YT. Focal hepatic lesions: detection and characterization with combination gadolinium- and superparamagnetic iron oxide-enhanced MR imaging. Radiology 2003;228 : 719-726[Abstract/Free Full Text]
25. Efron B. Better bootstrap confidence intervals. J Am Stat Assoc 1987; 82:171 -200[CrossRef]
26. Shah SA, Tan JC, McGilvray ID, et al. Accuracy of staging as a predictor for recurrence after liver transplantation for hepatocellular carcinoma. Transplantation 2006;81 : 1633-1639[CrossRef][Medline]
27. Valls C, Cos M, Figueras J, et al. Pretransplantation diagnosis and staging of hepatocellular carcinoma in patients with cirrhosis: value of dual-phase helical CT. AJR 2004;182 : 1011-1017[Abstract/Free Full Text]
28. Kim YK, Kim CS, Lee YH, Kwak HS, Lee JM. Comparison of superparamagnetic iron oxide-enhanced and gadobenate dimeglumine-enhanced dynamic MRI for detection of small hepatocellular carcinomas. AJR 2004; 182:1217 -1223[Abstract/Free Full Text]
29. Kim SK, Kim SH, Lee WJ, et al. Preoperative detection of hepatocellular carcinoma: ferumoxides-enhanced versus mangafodipir trisodium-enhanced MR imaging. AJR 2002;179 : 741-750[Abstract/Free Full Text]
30. Seneterre E, Taourel P, Bouvier Y, et al. Detection of hepatic metastases: ferumoxides-enhanced MR imaging versus unenhanced MR imaging and CT during arterial portography. Radiology1996; 200:785 -792[Abstract/Free Full Text]
31. Kanematsu M, Semelka RC, Matsuo M, et al. Gadolinium-enhanced MR imaging of the liver: optimizing imaging delay for hepatic arterial and portal venous phases: a prospective randomized study in patients with chronic liver damage. Radiology 2002;225 : 407-415[Abstract/Free Full Text]
32. Kaihara S, Kiuchi T, Ueda M, et al. Living-donor liver transplantation for hepatocellular carcinoma. Transplantation 2003;75 : S37-S40[Medline]
33. Baccarani U, Adani GL, Avellini C, et al. Comparison of clinical and pathological staging and long-term results of liver transplantation for hepatocellular carcinoma in a single transplant center. Transplant Proc 2006; 38:1111 -1113[CrossRef][Medline]

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				B. Guiu, R. Loffroy, and J.-P. Cercueil Double-Contrast MRI in Patients with Cirrhosis Am. J. Roentgenol., August 1, 2008; 191(2): W78 - W78. [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Alert me when this article is cited 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 Hanna, R. F. Articles by Sirlin, C. B. Search for Related Content PubMed PubMed Citation Articles by Hanna, R. F. Articles by Sirlin, C. B. Social Bookmarking                      What's this? Hotlight (NEW!) What's Hotlight?