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 Kondo, H. Articles by Nakamura, H. Search for Related Content PubMed PubMed Citation Articles by Kondo, H. Articles by Nakamura, H. Social Bookmarking                      What's this?

Preoperative Detection of Malignant Hepatic Tumors

Comparison of Combined Methods of MR Imaging with Combined Methods of CT

¹ Department of Radiology, Gifu University School of Medicine, 40 Tsukasamachi, Gifu 500-8705, Japan.
² Department of Radiology, Osaka University Medical School, Osaka 565-0871, Japan.

Received August 12, 1999; accepted after revision September 28, 1999.

 
Address correspondence to M. Kanematsu

Abstract

Top Abstract Introduction Subjects and Methods Results Discussion References

 
OBJECTIVE. We compared radiologists' performance on combined unenhanced, gadolinium-enhanced, and ferumoxides-enhanced MR imaging with their performance on helical CT during arterial portography (CTAP) and biphasic CT during hepatic arteriography (CTHA) for the preoperative detection of malignant hepatic tumors.

SUBJECTS AND METHODS. MR images and CT scans obtained in 33 patients were retrospectively analyzed. Images of the liver were reviewed on a segment-by-segment basis; a total of 261 segments with 39 hepatocellular carcinomas and 21 metastases were independently reviewed by three radiologists who were invited from outside institutions. Unenhanced and gadolinium-enhanced MR images were reviewed first, then ferumoxides-enhanced MR images were added for combined review. CTAP images and biphasic CTHA images were reviewed together.

RESULTS. Sensitivity for the detection of hepatic tumors was analogous for combined unenhanced, gadolinium-enhanced, and ferumoxides-enhanced MR images (86%) and for combined CTAP images and biphasic CTHA images (87%). Specificity was higher with MR images (95%, p < 0.01) than with CT images (91%). Radiologists' performances were improved (A_z = 0.962, p = 0.0502) by combining ferumoxides-enhanced MR images with unenhanced and gadolinium-enhanced MR images (A_z) = 0.950), and were analogous for combined unenhanced, gadolinium-enhanced, and ferumoxides-enhanced MR images and for combined CTAP images and biphasic CTHA images (A_z = 0.959).

CONCLUSION. Radiologists' performances on combined unenhanced, gadolinium-enhanced, and ferumoxides-enhanced MR imaging compared with their performances on combined helical CTAP and biphasic CTHA are analogous for the preoperative detection of malignant hepatic tumors. Such a dedicated combination of MR imaging may obviate the need for more invasive angiographically assisted helical CT for the preoperative detection of malignant hepatic tumors.

Introduction

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Researchers have described the usefulness of CT during arterial portography (CTAP) for the preoperative workup of patients with suspected liver malignancy [1,2,3]. However, CTAP has limitations detecting hepatocellular carcinoma in patients with cirrhosis, portal hypertension, and portal venous thrombosis [4], and the high incidence of pseudolesions decreases the specificity of the examination and lowers the accuracy of tumor detection [5]. To improve the accuracy of angiographically assisted CT for liver tumors, researchers combine CTAP with CT during hepatic arteriography (CTHA) [6, 7] or with biphasic CTHA [8]. However, such a dedicated technique is not widely used because of invasiveness, expense, and complexity.

Meanwhile, MR imaging, a less invasive and more feasible technique, has become available for routine examinations or as an alternative to angiographically assisted CT. Researchers describe the diagnostic usefulness of gadoliniumenhanced multiphasic dynamic MR images in addition to unenhanced MR images [9, 10]. Also, tissue-specific MR imaging with superparamagnetic iron oxide compounds has improved the focal hepatic lesion-to-liver contrast-to-noise ratio [11] and hepatic tumor detection [12]; however, whether the accuracy of ferumoxides-enhanced MR imaging exceeds that of CTAP is controversial [13, 14]. Nevertheless, by combining unenhanced, gadolinium-enhanced, and ferumoxides-enhanced MR images, the liver can be evaluated on the basis of differences in T1- and T2-relaxation times, vascularity, fat contents, and the distribution of Kupffer's cells between the liver parenchyma and malignant tumors. Although we suspect that such a dedicated combination of MR imaging improves the accuracy of detecting malignant hepatic tumors, it is unclear whether it obviates the need for more invasive angiographically assisted CT.

We compared the diagnostic accuracy of combined unenhanced T1- and T2-weighted, gadolinium-enhanced triphasic dynamic, and ferumoxides-enhanced MR imaging with combined helical CTAP and biphasic CTHA for the preoperative detection of malignant hepatic tumors. We used the receiver operating characteristic (ROC) curve to determine statistical results.

Subjects and Methods

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Patients
During the 15-month period from January 1998 to March 1999, 35 patients with symptoms suggestive of primary or metastatic malignant hepatic tumors were referred to our department for preoperative workup. Patients underwent MR imaging and angiographically assisted CT. All patients understood that the MR imaging and CT examinations were primarily for clinical diagnosis and secondarily for radiologic research. All patients provided informed consent. The study was performed in conformity with the Declaration of Helsinki [15]. Two patients were excluded from the study because the interval between MR imaging and CT was more than 2 weeks. The remaining 33 patients participated in the study (men, 23; women, 10; age range, 46-78 years; mean age, 64 years), including 19 patients with hepatocellular carcinoma in a cirrhotic liver, two with hepatocellular carcinoma in the liver with chronic viral hepatitis, eight with liver metastases from colorectal carcinoma, one with cirrhosis, and three with healthy liver. One cirrhotic patient with no tumor was examined because he had symptoms suggestive of hepatocellular carcinoma. This patient's condition was later diagnosed as confluent hepatic fibrosis. All patients with healthy livers had primary colorectal carcinoma; they underwent our liver workup protocol because they had symptoms suggestive of liver metastasis. Hepatic cirrhosis was diagnosed with definitive surgery in six patients. The remaining 14 patients' conditions were diagnosed with liver function tests or morphologic findings on sonography, CT, or MR imaging.

Of 33 patients, six with hepatocellular carcinoma, seven with colorectal cancer metastases, and one with suspected metastasis underwent surgery with intraoperative sonography within 2 weeks of their radiologic workup. Twelve patients' conditions were diagnosed as hepatocellular carcinoma with sonographically guided core and aspiration biopsy from at least one lesion; in these patients, other tumors with imaging findings similar to those of the histologically examined lesions were considered to be of the same disease. The histologic diagnosis of hepatocellular carcinoma was made on the basis of criteria advocated by the International Working Party on the Terminology of Chronic Hepatitis, Hepatic Allograft Rejection, and Nodular Lesions of the Liver published in 1995 [16]. In three patients with hepatocellular carcinoma and one with metastases, in whom histopathologic evidence of disease was not obtained, visible hepatic tumors were considered malignant on the basis of increased levels of serum tumor markers and tumor growth or shrinkage 5-12 months after chemotherapy.

The presence or absence of malignant hepatic tumors was determined by the consensus opinion of three radiologists based on the findings of definitive surgery with intraoperative sonography; percutaneous sonography; CTAP; biphasic CTHA; unenhanced T1- and T2-weighted, gadolinium-enhanced dynamic, and ferumoxides-enhanced MR imaging; follow-up sonography, CT, and MR imaging; serologic tests; and biopsy. Therefore, we identified 33 patients with a total of 60 lesions (36 histologically proven malignant tumors): 39 hepatocellular carcinomas (diameter, 6-53 mm; mean, 17.5 mm) and 21 metastases (diameter, 5-58 mm; mean, 18.8 mm).

Unenhanced and Gadolinium-Enhanced MR Imaging
MR imaging was performed with a superconducting magnet operating at 1.5 T (Signa Horizon; General Electric Medical Systems, Milwaukee, WI). The system provided a maximum gradient strength of 23 mT/m with a peak slew rate of 77 mT/m/msec. All MR images were obtained in the axial plane with a phased array multicoil for the body. The section thickness was 8 mm, with a 2-mm intersection gap for all pulse sequences. The unenhanced and gadolinium-enhanced MR imaging protocol comprised T1-weighted spinecho imaging, chemical shift-selective fat-suppressed (n = 31) or non—fat-suppressed (n = 2) respiratory-triggered T2-weighted fast spin-echo imaging, and gadolinium-enhanced triphasic dynamic gradient-recalled echo imaging with a fast multiplanar spoiled—gradient-recalled echo acquisition under a steady-state free precession sequence. The scanning parameters for T1-weighted spin-echo imaging included a TR/TE range of 500/8-9, a matrix of 256 x 224, a field of view of 32 x 32-29 x 29 cm, a received bandwidth of ± 16 kHz, two signals acquired, respiratory-ordered phase encoding, and an acquisition time of 6.5 min. The scanning protocol for chemical shift—selective fat-suppressed (n = 31) or non—fat-suppressed (n = 2) respiratory-triggered T2-weighted fast spin-echo imaging included an effective TR range/effective TE range of 4000-8571/77-80, a matrix of 512 x 256, an echo train length of 10-18, a 75% rectangular field of view of 32 x 24-29 x 22 cm, a received bandwidth of ± 62.5 kHz, three or four signals acquired, a 20% respiratory trigger point, a 40% trigger window, a gradient moment nulling in the frequency-encoding direction, and an acquisition time of 3.2-6.0 min. The scanning parameters for gadolinium-enhanced triphasic dynamic gradient-recalled echo imaging with a fast multiplanar spoiled—gradient-recalled echo acquisition under steady-state free precession sequence included a TR/TE range of 150/1.6-1.8, a matrix of 512 x 224, a flip angle of 110°, a 75% rectangular field of view of 32 x 24-29 x 22 cm, a received bandwidth of ± 62.5 kHz, one signal acquisition, and a breath-hold acquisition time of 26 sec. Triphasic dynamic gradient-recalled echo images were obtained before and after a manual antecubital IV bolus of 0.1 mmol/kg of gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany) flushed by 20 ml of sterile saline solution. The scan delay for triphasic dynamic gradient-recalled echo imaging was 14 sec, 60 sec, and 3 min after initiating contrast injection, representing the hepatic arterial, portal venous, and equilibrium phases, respectively. The scan delay for the hepatic arterial phase was determined 14 sec after initiating contrast material injection, so the image data obtained at approximately 27 sec were used to fill in the central K-space lines. A chemical shift—selective fat-suppression pulse was not used for dynamic gradient-recalled echo imaging because the acquisition time and number of slices were traded off.

Ferumoxides-Enhanced MR Imaging
Ferumoxides (Feridex; Tanabe Pharmaceutical, Tokyo, Japan) was administered at a dose of 10 mmol of iron per kilogram of body weight. The ferumoxides suspension was diluted in 100 ml of 5% glucose solution and was administered IV over 30 min. Patients underwent MR imaging 2-6 hr after receiving the ferumoxides solution. Ferumoxides-enhanced MR imaging was performed with the same scanner as that for unenhanced and gadolinium-enhanced MR imaging. The ferumoxides-enhanced MR imaging protocol comprised proton density—weighted spin-echo imaging, non—fat-suppressed respiratory-triggered T2-weighted fast spin-echo imaging, and breath-hold T2-weighted gradient-recalled echo imaging. The scanning parameters for proton density—weighted spin-echo imaging included a TR/TE of 2000/30, a matrix of 256 x 192, a field of view of 32 x 32-29 x 29 cm, a received bandwidth of ± 16 kHz; two signals acquired, a gradient moment nulling in the frequency-encoding direction, respiratory-ordered phase encoding, and an acquisition time of 13.8 min. The scanning parameters for non—fat-suppressed respiratory-triggered T2-weighted fast spin-echo imaging included an effective TR range/effective TE range of 2857-8571/77-80, a matrix of 512 x 256, an echo train length of 8-16, a 75% rectangular field of view of 32 x 24-29 x 22 cm, a received bandwidth of ± 62.5 kHz, three or four signals acquired, a 20% respiratory trigger point, a 40% trigger window, a gradient moment nulling in the frequency-encoding direction, and an acquisition time of 4.5-6.5 min. The scanning parameters for breath-hold T2-weighted gradient-recalled echo imaging included a TR/TE of 150/10, a matrix of 256 x 160, a flip angle of 20°, a 75% rectangular field of view of 32 x 24-29 x 22 cm, a received bandwidth of ± 31.3 kHz, one signal acquired, an acquisition time of eight locations/19 sec, and two or three data acquisitions. A chemical shift—selective fat-saturation pulse was not used for ferumoxides-enhanced MR imaging because sufficient tumor-to-liver enhancement was expected without fat suppression.

CT Imaging
Before angiographically assisted CT, two 3.2- to 4.0-French angiographic catheters (GHC—A; Clinical Supply, Gifu, Japan) were placed in the superior mesenteric artery and hepatic artery using the Seldinger technique through the unilateral femoral arteries with a Y-shaped sheath introducer (Twin Sheath; Medikit, Tokyo, Japan). Initially, we attempted to place a catheter in the common hepatic artery for CTHA; otherwise, we placed the catheter in the celiac artery. When the liver was perfused by two or more hepatic arteries, the artery that perfused the potentially resectable area of the liver was selected for catheter placement. The hepatic arterial catheter for CTHA was placed in the celiac (two patients), common hepatic (23 patients), replaced right hepatic (six patients), or middle hepatic (one patient) artery. In six patients with a replaced right hepatic artery, the superior mesenteric arterial catheter was inserted well beyond the hepatic arterial origin so that the contrast material did not overflow into the replaced right hepatic artery. One patient (who underwent hepatic resection 10 years earlier) had an obstructed hepatic artery and CTHA was not performed. For all patients, the catheter for CTAP was placed in the superior mesenteric artery.

First, CTAP was obtained using a CT scanner (Hi-Speed Advantage; General Electric Medical Systems, Milwaukee, WI) with a 0.8-sec helical scan capability. The helical CT images were obtained in a craniocaudal direction with 7-mm collimation, 7-mm/0.8 sec table speed, 120 kVp, and 200-220 mAs during a single breath-hold helical acquisition for 20-25 sec, depending on the liver size. We routinely supplied patients with 100% oxygen at 3-5 1/min through a nasal cannula to ensure breath-holding. Breath-holding was successful in all patients. Data acquisition started 25 sec after the initiation of a transcatheter superior mesenteric arterial automatic injection (Auto-enhance A50; Nemotokyorindo, Tokyo, Japan) of 95 ml of nonionic contrast material that contained 150 mg Iodine/ml.

Biphasic CTHA was obtained 10 min after CTAP. Data acquisition started 5 and 40 sec after the initiation of a transcatheter hepatic arterial automated injection of 30-40 ml of nonionic contrast medium that contained 150 mg Iodine/ml for the first- and secondphase CTHA, respectively. When we placed the tip of the catheter in the common, middle, or replaced right hepatic artery, we injected contrast medium at 1.5 ml/sec, and when we placed the catheter in the celiac artery, we injected contrast medium at 2.0 ml/sec. Contiguous axial images 7-mm thick with 2-mm overlap were reconstructed from the volumetric data using a 180° linear interpolation algorithm.

Some parts of the liver were insufficiently opacified on CTAP images for one patient with portal hypertension caused by severe cirrhosis and another patient with hepatocellular carcinoma thrombus in the right posterior portal vein branch. The liver was partially opacified with CTHA for six patients with a replaced right hepatic artery and one patient whose hepatic arterial catheter dislocated in the middle hepatic artery during transportation. CTHA was not performed for one patient with an obstructed hepatic artery and another patient whose catheter dislocated in the gastroduodenal artery.

Qualitative Image Analysis
Three experienced radiologists interpreted the images for our study. The radiologists knew that the patients were referred for the preoperative assessment of possible liver tumors; however, they were unaware of all other information.

For MR imaging, unenhanced T1-weighted spinecho, T2-weighted fast spin-echo, and gadolinium-enhanced triphasic dynamic gradient-recalled echo images were reviewed together, then ferumoxides-enhanced proton density—weighted spin-echo, T2-weighted fast spin-echo, and T2^*-weighted gradient-recalled echo images were added for combined review. For angiographically assisted helical CT, CTAP images and biphasic CTHA images were reviewed together. MR images and CT scans were reviewed in random order in two reading sessions separated by a 3-week interval.

For each imaging sequence, radiologists recorded the size (average of two perpendicular diameters) and site (Couinaud segment) of visible abnormalities and indicated the presence or absence of malignant tumors. To prevent mislocation of lesions, the study coordinator drew hepatic segments (based on Couinaud's numbering system [17]) on the images. Then, the radiologists assigned confidence scores (2 = probably absent, 3 = equivocal, 4 = probably present, 5 = definitely present) to each abnormality.

When a lesion was located in two or more segments, radiologists were asked to consider only the segment that was primarily involved and to assess the probability of another lesion in other segments. The radiologists assigned a score of 3 when the signal-intensity or attenuation change was subtle, illdefined, and not circular or oval. The radiologists assigned a score of 5 to well-circumscribed and circular or oval opacities with discrete signal-intensity or attenuation change. Scores of 2 and 4 were assigned on the basis of the radiologist's subjective judgment. To minimize learning bias, the radiologists were blinded to the name, age, identification number, and imaging parameters of each patient.

Statistical Analysis
The sensitivity of each imaging sequence to detect liver tumors was determined using the number of segments assigned a score of 3 or greater. The specificity of each imaging sequence to detect liver segments without tumor was determined using the number of segments assigned a score of 1 or 2. Radiologists identified 56 segments with malignant tumors and 205 segments without malignant tumors. The sensitivity of each imaging sequence to detect malignant tumors 10 mm or smaller was determined using the number of tumors assigned a score of 3 or greater. Radiologists identified 20 tumors with a size of 10 mm or smaller. Sensitivities and specificities were determined using the McNemar test.

The ROC curve analysis was conducted according to Couinaud's segment-by-segment method because the chief determinant of hepatic resectability is the accurate identification of the segments to be resected and because our objective was to compare the radiologists' performances in detecting focal hepatic tumors using different imaging techniques.

We analyzed the confidence-level data for malignant hepatic tumors assigned to a total of 261 liver segments harboring 39 hepatocellular carcinomas and 21 metastatic lesions. ROC analyses were determined for all patients, a subset of 22 patients comprising 21 with hepatocellular carcinomas and one with cirrhosis, and another subset of 11 patients comprising eight with metastases and three with healthy livers.

For each imaging technique, a binomial ROC curve was fitted to each radiologist's confidence rating using a maximum-likelihood estimation with an ROC analysis program (ROCKIT 0.9.1B; Metz CE, Chicago, IL) [18]. We calculated the diagnostic accuracy (the area under the ROC curve [A_z]) for each imaging sequence, for each radiologist, and for the composite data. Differences between the ROC curves were determined using a univariate z score test.

To assess interobserver variability in interpreting images, the kappa statistic for interobserver agreement was used. We used nonweighted kappa statistics with binary data defined by the presence (definitely present, probably present, equivocal) or absence (probably absent, definitely absent) of malignant tumors in liver segments. The degree of disagreement was not factored into the calculation. A kappa value of 0.40 or less indicated positive but poor agreement, a value of 0.41-0.75 indicated good agreement, and a value greater than 0.75 indicated excellent agreement.

Results

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Table 1 summarizes the sensitivities and specificities of each imaging technique to detect the presence and absence of liver tumors. We found no differences in the sensitivity to detect liver tumors (Figs. 1A,1B,1C,1D,1E and 2A,2B,2C,2D,2E,2F). For one radiologist and the composite data, the specificity to detect liver segments without tumor was significantly (p < 0.05) higher with combined unenhanced, gadoliniumenhanced, and ferumoxides-enhanced MR images than with unenhanced and gadolinium-enhanced MR images. For two radiologists and the composite data, the specificity to detect liver segments without tumor was significantly (p < 0.01) higher with combined unenhanced, gadolinium-enhanced, and ferumoxides-enhanced MR images than with combined CTAP images and biphasic CTHA images. We found no differences in the sensitivity to detect malignant tumors 10 mm or smaller (Table 2).

View larger version (134K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1A. —73-year-old man with poorly differentiated hepatocellular carcinoma that was histologically diagnosed after core and aspiration biopsy. CT during arterial portography image (A), first-phase CT hepatic arteriogram (B), breath-hold gadolinium-enhanced hepatic arterial phase gradient-recalled echo MR image (TR/TE, 150/1.6) (C), ferumoxides-enhanced respiratory-triggered fast spin-echo MR image (5714/80) (D), and ferumoxides-enhanced breath-hold T2-weighted gradient-recalled echo MR image (150/10) (E) show poorly differentiated hepatocellular carcinoma (arrows). Note that lesion conspicuity on D is comparable with that on A.

View larger version (125K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1B. —73-year-old man with poorly differentiated hepatocellular carcinoma that was histologically diagnosed after core and aspiration biopsy. CT during arterial portography image (A), first-phase CT hepatic arteriogram (B), breath-hold gadolinium-enhanced hepatic arterial phase gradient-recalled echo MR image (TR/TE, 150/1.6) (C), ferumoxides-enhanced respiratory-triggered fast spin-echo MR image (5714/80) (D), and ferumoxides-enhanced breath-hold T2-weighted gradient-recalled echo MR image (150/10) (E) show poorly differentiated hepatocellular carcinoma (arrows). Note that lesion conspicuity on D is comparable with that on A.

View larger version (140K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1C. —73-year-old man with poorly differentiated hepatocellular carcinoma that was histologically diagnosed after core and aspiration biopsy. CT during arterial portography image (A), first-phase CT hepatic arteriogram (B), breath-hold gadolinium-enhanced hepatic arterial phase gradient-recalled echo MR image (TR/TE, 150/1.6) (C), ferumoxides-enhanced respiratory-triggered fast spin-echo MR image (5714/80) (D), and ferumoxides-enhanced breath-hold T2-weighted gradient-recalled echo MR image (150/10) (E) show poorly differentiated hepatocellular carcinoma (arrows). Note that lesion conspicuity on D is comparable with that on A.

View larger version (130K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1D. —73-year-old man with poorly differentiated hepatocellular carcinoma that was histologically diagnosed after core and aspiration biopsy. CT during arterial portography image (A), first-phase CT hepatic arteriogram (B), breath-hold gadolinium-enhanced hepatic arterial phase gradient-recalled echo MR image (TR/TE, 150/1.6) (C), ferumoxides-enhanced respiratory-triggered fast spin-echo MR image (5714/80) (D), and ferumoxides-enhanced breath-hold T2-weighted gradient-recalled echo MR image (150/10) (E) show poorly differentiated hepatocellular carcinoma (arrows). Note that lesion conspicuity on D is comparable with that on A.

View larger version (126K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 1E. —73-year-old man with poorly differentiated hepatocellular carcinoma that was histologically diagnosed after core and aspiration biopsy. CT during arterial portography image (A), first-phase CT hepatic arteriogram (B), breath-hold gadolinium-enhanced hepatic arterial phase gradient-recalled echo MR image (TR/TE, 150/1.6) (C), ferumoxides-enhanced respiratory-triggered fast spin-echo MR image (5714/80) (D), and ferumoxides-enhanced breath-hold T2-weighted gradient-recalled echo MR image (150/10) (E) show poorly differentiated hepatocellular carcinoma (arrows). Note that lesion conspicuity on D is comparable with that on A.

View larger version (141K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 2A. —49-year-old man with surgically resected liver metastasis from colonic carcinoma. CT during arterial portography image (A), second-phase CT hepatic arteriogram (B), breath-hold gadolinium-enhanced portal venous phase gradient-recalled echo MR image (TR/TE, 150/1.6) (C), ferumoxides-enhanced proton density—weighted spin-echo MR image (2000/80) (D), and ferumoxides-enhanced respiratory-triggered fast spin-echo MR image (5454/80) (E) show liver metastasis (arrows) from colonic carcinoma. First-phase CT hepatic arteriogram and unenhanced MR images (not shown) failed to depict tumor. Note that lesion conspicuity is better on A than on C-E.

View larger version (143K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 2B. —49-year-old man with surgically resected liver metastasis from colonic carcinoma. CT during arterial portography image (A), second-phase CT hepatic arteriogram (B), breath-hold gadolinium-enhanced portal venous phase gradient-recalled echo MR image (TR/TE, 150/1.6) (C), ferumoxides-enhanced proton density—weighted spin-echo MR image (2000/80) (D), and ferumoxides-enhanced respiratory-triggered fast spin-echo MR image (5454/80) (E) show liver metastasis (arrows) from colonic carcinoma. First-phase CT hepatic arteriogram and unenhanced MR images (not shown) failed to depict tumor. Note that lesion conspicuity is better on A than on C-E.

View larger version (144K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 2C. —49-year-old man with surgically resected liver metastasis from colonic carcinoma. CT during arterial portography image (A), second-phase CT hepatic arteriogram (B), breath-hold gadolinium-enhanced portal venous phase gradient-recalled echo MR image (TR/TE, 150/1.6) (C), ferumoxides-enhanced proton density—weighted spin-echo MR image (2000/80) (D), and ferumoxides-enhanced respiratory-triggered fast spin-echo MR image (5454/80) (E) show liver metastasis (arrows) from colonic carcinoma. First-phase CT hepatic arteriogram and unenhanced MR images (not shown) failed to depict tumor. Note that lesion conspicuity is better on A than on C-E.

View larger version (140K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 2D. —49-year-old man with surgically resected liver metastasis from colonic carcinoma. CT during arterial portography image (A), second-phase CT hepatic arteriogram (B), breath-hold gadolinium-enhanced portal venous phase gradient-recalled echo MR image (TR/TE, 150/1.6) (C), ferumoxides-enhanced proton density—weighted spin-echo MR image (2000/80) (D), and ferumoxides-enhanced respiratory-triggered fast spin-echo MR image (5454/80) (E) show liver metastasis (arrows) from colonic carcinoma. First-phase CT hepatic arteriogram and unenhanced MR images (not shown) failed to depict tumor. Note that lesion conspicuity is better on A than on C-E.

View larger version (140K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 2E. —49-year-old man with surgically resected liver metastasis from colonic carcinoma. CT during arterial portography image (A), second-phase CT hepatic arteriogram (B), breath-hold gadolinium-enhanced portal venous phase gradient-recalled echo MR image (TR/TE, 150/1.6) (C), ferumoxides-enhanced proton density—weighted spin-echo MR image (2000/80) (D), and ferumoxides-enhanced respiratory-triggered fast spin-echo MR image (5454/80) (E) show liver metastasis (arrows) from colonic carcinoma. First-phase CT hepatic arteriogram and unenhanced MR images (not shown) failed to depict tumor. Note that lesion conspicuity is better on A than on C-E.

View larger version (108K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 2F. —49-year-old man with surgically resected liver metastasis from colonic carcinoma. Photograph of resected specimen shows liver metastasis from colonic carcinoma (size, 5 x 7 mm).

View this table: [in this window] [in a new window]   TABLE 2 Sensitivity for Tumor Detection of 20 Small Tumors (10 mm) with Different Imaging Sequences

Table 3 summarizes the A_z values for the three imaging techniques to detect malignant hepatic tumors. We determined the composite ROC curves for the radiologists in 33 patients, 22 patients with hepatocellular carcinoma or cirrhosis, and 11 patients with metastasis or healthy liver (Figs. 3,4,5). The diagnostic accuracy with combined unenhanced, gadolinium-enhanced, and ferumoxides-enhanced MR images (A_z = 0.962, p = 0.0502) was marginally higher than with unenhanced and gadolinium-enhanced MR images (A_z = 0.950). No significant difference in A_z value was found for combined unenhanced, gadolinium-enhanced, and ferumoxides-enhanced MR images and combined CTAP images and biphasic CTHA images (A_z = 0.959), or for unenhanced and gadolinium-enhanced MR images and combined CTAP images and biphasic CTHA images. MR imaging was more accurate than CT for patients with liver metastasis; however the difference in A_z values showed no statistical significance.

View larger version (15K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 3. —Receiver operating characteristic curves (generated from composite data) show confidence of three observers for detection of malignant hepatic tumors in 33 patients with unenhanced and gadolinium-enhanced MR imaging () (Az = 0.950); combined unenhanced, gadolinium-enhanced, and ferumoxides-enhanced MR imaging () (Az = 0.962); and combined helical CT during arterial portography and biphasic CT during hepatic arteriography ([UNK]) (Az = 0.959).

View larger version (16K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 4. —Receiver operating characteristic curves (generated from composite data) show confidence of three observers for detection of malignant hepatic tumors in 22 patients with hepatocellular carcinoma (n = 21) or cirrhosis (n = 1) with unenhanced and gadolinium-enhanced MR imaging () (Az = 0.949); combined unenhanced, gadolinium-enhanced, and ferumoxides-enhanced MR imaging () (Az = 0.953); and combined helical CT during arterial portography and biphasic CT during hepatic arteriography ([UNK]) (Az = 0.959).

View larger version (14K): [in this window] [in a new window] [as a PowerPoint slide]   Fig. 5. —Receiver operating characteristic curves (generated from composite data) show confidence of three observers for detection of malignant hepatic tumors in 11 patients with metastases (n = 8) or healthy liver (n = 3) with unenhanced and gadolinium-enhanced MR imaging () (Az = 0.944); combined unenhanced, gadolinium-enhanced, and ferumoxides-enhanced MR imaging () (Az = 0.977); and combined helical CT during arterial portography and biphasic CT during hepatic arteriography ([UNK]) (Az = 0.952).

The kappa values among the three radiologists were 0.72, 0.75, and 0.65 for unenhanced and gadolinium-enhanced MR imaging; combined unenhanced, gadolinium-enhanced, and ferumoxides-enhanced MR imaging; and combined CTAP and biphasic CTHA, respectively. The radiologists were in good agreement for the presence or absence of malignant tumors in specific liver segments.

Discussion

Top Abstract Introduction Subjects and Methods Results Discussion References

 
A study by Senéterre et al. [13] found ferumoxides-enhanced MR imaging more accurate than CTAP in detecting liver tumors because MR imaging had a higher specificity (the techniques had comparable sensitivities). In our study, the accuracy and sensitivity to detect liver tumors were analogous for combined MR imaging and combined CTAP and biphasic CTHA. The specificity for detecting liver segments without tumor was greater with MR imaging than with CT. The specificity of MR imaging may be high because unenhanced and gadolinium-enhanced MR images helped distinguish nonsolid benign hepatic lesions or peripheral hepatic vessels from small malignant hepatic tumors, and ferumoxides-enhanced MR images that were free of tumor-mimicking false-positive findings helped distinguish pseudolesions seen on gadolinium-enhanced MR images [19, 20]. When considering invasiveness, patients' burden, and medical expenses, the accuracy and specificity of MR imaging support its use in the preoperative workup of patients with liver tumors and the planning of hepatic surgery. The cost-benefit issue is important when comparing CT and MR imaging. In Japan, the medical expenses for MR imaging with gadolinium and ferumoxides contrast enhancement and angiographically assisted CT were estimated at $730 and $890, respectively. However, these costs may vary from country to country because of differences in health care systems.

CTAP reveals small hepatic tumors with good conspicuity [1,2,3]. The sensitivity of superparamagnetic iron oxide—enhanced MR imaging to detect hepatic lesions smaller than 1 cm is as high as 56% [21]. In our study, although the size of hepatic tumors (10 mm) was small, combined CTAP and biphasic CTHA detected more lesions than combined MR imaging. Nevertheless, the accuracy determined with ROC curve analysis was analogous for MR images and CT scans. There are three explanations for this contradiction: first, loss of image quality with CTAP by hepatofugal portal flow caused by portal hypertension in cirrhosis; second, major arterioportal shunting caused by hepatocellular carcinoma or previous liver biopsy; or third, lobar perfusion abnormality caused by laminar flow in the portal veins [22]. The image quality of CTHA was degraded by the uneven distribution of contrast medium, partial opacification of hepatic parenchyma (resulting from replaced or anomalous hepatic arteries), or technical failures. Under such conditions, small tumors may be overlooked on angiographically assisted CT. On the other hand, degradation of image quality caused by benign perfusion abnormality, technical failure, or anatomical variation is less frequent with gadolinium-enhanced dynamic MR imaging and is rare with ferumoxides-enhanced MR imaging. The stable image quality of MR imaging probably resulted in the analogous detectability of tumors with CT.

Researchers have compared the pulse sequences for superparamagnetic iron oxide—enhanced MR imaging; however, their results depended on the strength of magnetic fields [11, 23]. In our study, we performed ferumoxides-enhanced MR imaging using a combination of proton density—weighted spin-echo MR imaging, expecting good tumor-to-liver contrast [11, 23, 24]; respiratory-triggered T2-weighted fast spin-echo MR imaging, expecting superior image quality and spatial resolution [11, 25]; and T2-weighted gradient-recalled echo MR imaging, expecting improved tumor detectability and high image quality because of breath-hold data acquisition [12, 24]. We believe that ferumoxides-enhanced MR imaging with these pulse sequences was the optimal technique, and the total examination time (less than 30 min) was clinically acceptable.

Our study had some limitations. Only 14 patients underwent surgery with intraoperative sonography; the remaining 16 patients with malignant hepatic tumors did not. Although it is ideal that histologic proof be obtained from all lesions by definitive surgery, most lesions are confirmed with intraoperative sonography. Surgical proof is becoming unnecessary, especially in patients with hepatocellular carcinoma because small lesions are often treated with trans-catheter arterial chemoembolization and percutaneous ablation that are frequently used for patients with cirrhosis, whose functional reserve is severely impaired.

MR images were obtained with an 8-mm slice thickness and a 2-mm interslice gap with or without breath-hold, but helical CT images were obtained with a 7-mm slice thickness and a 2-mm overlap. These differences in imaging parameters may obscure the correlation of segmental anatomy with different imaging techniques. We excluded histologically proven or suspected dysplastic nodules in cirrhosis from standard lesions because the clinical significance of these nodules is still controversial. However, angiographically assisted CT is not sensitive for dysplastic nodules in cirrhosis and some well-differentiated hepatocellular carcinomas because these hepatic nodules are often isovascular or slightly hypovascular to the surrounding liver parenchyma [26, 27]. Unenhanced MR imaging can detect nodules in cirrhosis [28, 29]. Other potential limitations of this study include reading-order bias and recall bias. These biases were minimal because the radiologists reviewed images in a random order, the interval between the reading sessions was 3 weeks, and the tumors evaluated were relatively small (mean, 18 mm).

A disadvantage of our study was that we performed unenhanced and gadolinium-enhanced MR imaging and ferumoxides-enhanced MR imaging on 2 days, increasing our patients' burden and medical costs. MR imaging can be performed at an outpatient clinic, but angiographically assisted CT usually requires hospital admission. In the future, other liver-specific contrast agents such as iron oxide/carboxydextran sol (SH U 555 A) [30], gadoxetic acid, disodium (Gd-EOB-DTPA) [31], or gadobenate dimeglumine (Bracco, Milan, Italy) [32], which can be used for dynamic and static contrast-enhanced MR imaging, are expected to decrease patients' burden and medical costs, while maintaining high diagnostic accuracy.

In conclusion, the specificity for detecting liver tumors is increased by combining ferumoxides-enhanced MR imaging with unenhanced T1-weighted, T2-weighted, and gadolinium-enhanced triphasic dynamic MR imaging. The specificity is increased even more with combined helical CTAP and biphasic CTHA. Radiologists' performances are analogous for the preoperative detection of liver tumors with combined MR images and combined angiographically assisted CT. Such a dedicated combination of MR imaging may obviate the need for more invasive angiographically assisted helical CT. Our results may support the perspective that MR imaging can replace angiographically assisted CT in the preoperative workup of patients with suspected liver tumors.

References

Top Abstract Introduction Subjects and Methods Results Discussion References

 

1. Matsui O, Kadoya M, Suzuki M, et al. Dynamic sequential computed tomography during arterial portography in the detection of hepatic neoplasms. Radiology 1983;146: 721 -727[Abstract/Free Full Text]
2. Nelson RC, Chezmar JL, Sugarbaker PH, Bernardino ME. Hepatic tumors: comparison of CT during arterial portography, delayed CT, and MR imaging for preoperative evaluation. Radiology 1989;172: 27 -34[Abstract/Free Full Text]
3. Soyer P, Levesque M, Elias D, Zeitoun G, Roche A. Detection of liver metastases from colorectal cancer: comparison of intraoperative US and CT during arterial portography. Radiology 1992;183: 541 -544[Abstract/Free Full Text]
4. Kanematsu M, Oliver JH III, Carr B, Baron RL. Hepatocellular carcinoma: the role of helical biphasic contrast-enhanced CT versus CT during arterial portography. Radiology 1997;205: 75 -80[Abstract/Free Full Text]
5. Bluemke DA, Soyer P, Fishman EK. Nontumorous low-attenuation defects in the liver on helical CT during arterial portography: frequency, location, and appearance. AJR 1995;164: 1141 -1145[Abstract/Free Full Text]
6. Chezmar JL, Bernardino ME, Kaufman SH, Nelson RC. Combined CT arterial portography and CT hepatic angiography for evaluation of the hepatic resection candidate. Radiology 1993;189: 407 -410[Abstract/Free Full Text]
7. Kanematsu M, Hoshi H, Imaeda T, Murakami T, Inaba Y, Nakamura H. Detection and characterization of hepatic tumors: value of combined helical CT hepatic arteriography and CT during arterial portography. AJR 1997;168: 1193 -1198[Abstract/Free Full Text]
8. Inoue E, Fujita M, Hosomi N, et al. Double phase CT arteriography of the whole liver in the evaluation of hepatic tumors. J Comput Assist Tomogr 1998;22: 64 -68[Medline]
9. Low RN, Francis IR, Herfkens RJ, et al. Fast multiplanar spoiled gradient-recalled imaging of the liver: pulse sequence optimization and comparison with spin-echo MR imaging. AJR 1993;160: 501 -509[Abstract/Free Full Text]
10. 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]
11. Schwartz LH, Seltzer SE, Adams DF, et al. Effects of superparamagnetic iron oxide (AMI-25) on liver and spleen imaging using spin-echo and fast spin-echo magnetic resonance pulse sequences. Invest Radiol 1994;29: 21 -23
12. Reimer P, Rummeny EJ, Daldrup HE, et al. Clinical results with Resovist: a phase 2 clinical trial. Radiology 1995;195: 489 -496[Abstract/Free Full Text]
13. Senéterre E, Taourel P, Bouvier Y, et al. Detection of hepatic metastasis: ferumoxides-enhanced MR imaging versus unenhanced MR imaging and CT during arterial portography. Radiology 1996;200: 785 -792[Abstract/Free Full Text]
14. Sharf J, Hoffmann V, Roeren TK, Kauffmann GW. MR imaging of liver neoplasms with superparamagnetic iron oxides versus CTAP: histopathologically controlled study (abstr). Radiology 1996;201(P): 320[Free Full Text]
15. Declaration of Helsinki. Recommendations guiding doctors in clinical research: adopted by the World Medical Association in 1964. Wis Med J 1967;66: 25 -26[Medline]
16. Terminology of nodular hepatocellular lesions: International Working Party. Hepatology 1995;22: 983 -993[Medline]
17. Couinaud C. Le foie: études anatomiques et chirurgicales. Paris: Masson, 1957
18. Rockette HE, Gur D, Metz CE. The use of continuous and discrete confidence judgments in receiver operator characteristic studies of diagnostic imaging techniques. Invest Radiol 1992;27: 169 -172[Medline]
19. Itai Y, Matsui O. Blood flow and liver imaging. Radiology 1997;202: 306 -314[Free Full Text]
20. Kanematsu M, Hoshi H, Imaeda T, et al. Nonpathological focal enhancements on spiral CT hepatic angiography. Abdom Imaging 1997;22: 55 -59[Medline]
21. Hagspiel KD, Neidl KF, Eichenberger AC, Weder W, Marincek B. Detection of liver metastases: comparison of superparamagnetic iron oxide-enhanced and unenhanced MR imaging at 1.5 T with dynamic CT, intraoperative US, and percutaneous US. Radiology 1995;196: 471 -478[Abstract/Free Full Text]
22. Little AF, Baron RL, Peterson MS, et al. Optimizing CT portography: a prospective comparison of injection into the splenic versus superior mesenteric artery. Radiology 1994;193: 651 -655[Abstract/Free Full Text]
23. Yamamoto H, Yamashita Y, Yoshimatsu S, et al. Hepatocellular carcinoma in cirrhotic livers: detection with unenhanced and iron oxide-enhanced MR imaging. Radiology 1995;195: 106 -112[Abstract/Free Full Text]
24. Ward J, Naik KS, Guthrie JA, Wilson D, Robinson PJ. Hepatic lesion detection: comparison of MR imaging after the administration of superparamagnetic iron oxide with dual-phase CT by using alternative—free response receiver operating characteristic analysis. Radiology 1999;210: 459 -466[Abstract/Free Full Text]
25. Kanematsu M, Hoshi H, Murakami T, et al. Focal hepatic lesion detection: comparison of four T2-weighted MR imaging pulse sequences. Radiology 1998;206: 167 -175[Abstract/Free Full Text]
26. 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]
27. Takayasu K, Muramatsu Y, Furukawa H, et al. Early hepatocellular carcinoma: appearance at CT during arterial portography and CT arteriography with pathologic correlation. Radiology 1995;194: 101 -105[Abstract/Free Full Text]
28. Matsui O, Kadoya M, Kameyama, 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]
29. Kanematsu M, Hoshi H, Yamada T, et al. Small hepatic nodules in cirrhosis: ultrasonographic, CT, and MR imaging findings. Abdom Imaging 1999;24: 47 -55[Medline]
30. Shamsi K, Balzer T, Saini S, et al. Superparamagnetic iron oxide particles (SH U 555 A): evaluation of efficacy in three doses for hepatic MR imaging. Radiology 1998;206: 365 -371[Abstract/Free Full Text]
31. Vogl TJ, Kummel S, Hammerstingl R, et al. Liver tumors: comparison of MR imaging with Gd-EOB-DTPA and Gd-DTPA. Radiology 1996;200: 59 -67[Abstract/Free Full Text]
32. Caudana R, Morana G, Pirovano GP, et al. Focal malignant hepatic lesions: MR imaging enhanced with gadolinium benzyloxypropionictetra-acetate (BOPTA)—preliminary results of phase II clinical application. Radiology 1996;199: 513 -520[Abstract/Free Full Text]

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				W. P. Hosch, S. M. Schmidt, S. Plaza, C. Dechow, J. Schmidt, S. Ley, G. W. Kauffmann, and J. Hansmann Comparison of CT during arterial portography and MR during arterial portography in the detection of liver metastases. Am. J. Roentgenol., June 1, 2006; 186(6): 1502 - 1511. [Abstract] [Full Text] [PDF]

				C. Lejeune, M. J. Bismuth, T. Conroy, C. Zanni, P. Bey, L. Bedenne, J. Faivre, P. Arveux, and F. Guillemin Use of a Decision Analysis Model to Assess the Cost-Effectiveness of 18F-FDG PET in the Management of Metachronous Liver Metastases of Colorectal Cancer J. Nucl. Med., December 1, 2005; 46(12): 2020 - 2028. [Abstract] [Full Text] [PDF]

				J A Guthrie Cirrhosis and focal liver lesions: MRI findings Imaging, September 1, 2004; 16(4): 351 - 363. [Abstract] [Full Text] [PDF]

				M.-J. Kim, J. H. Kim, J.-J. Chung, M. S. Park, J. S. Lim, and Y. T. Oh Focal Hepatic Lesions: Detection and Characterization with Combination Gadolinium- and Superparamagnetic Iron Oxide-enhanced MR Imaging Radiology, September 1, 2003; 228(3): 719 - 726. [Abstract] [Full Text] [PDF]

				D. A. Bluemke, T. M. Weber, D. Rubin, E. E. de Lange, R. Semelka, R. D. Redvanly, J. Chezmar, E. Outwater, R. Carlos, S. Saini, et al. Hepatic MR Imaging with Ferumoxides: Multicenter Study of Safety and Effectiveness of Direct Injection Protocol Radiology, August 1, 2003; 228(2): 457 - 464. [Abstract] [Full Text] [PDF]

				H. C. Kim, T. K. Kim, K.-B. Sung, H.-K. Yoon, P. N. Kim, H. K. Ha, A. Y. Kim, H. J. Kim, and M.-G. Lee Preoperative Evaluation of Hepatocellular Carcinoma: Combined Use of CT with Arterial Portography and Hepatic Arteriography Am. J. Roentgenol., June 1, 2003; 180(6): 1593 - 1599. [Abstract] [Full Text] [PDF]

				C. D. Frate, C. Zuiani, V. Londero, G. Como, M. Bazzocchi, K. Mortele, and P. R. Ros Comparing Levovist-Enhanced Pulse Inversion Harmonic Imaging and Ferumoxides-Enhanced MR Imaging of Hepatic Metastases Am. J. Roentgenol., May 1, 2003; 180(5): 1339 - 1346. [Abstract] [Full Text] [PDF]

				K. Kinkel, Y. Lu, M. Both, R. S. Warren, and R. F. Thoeni Detection of Hepatic Metastases from Cancers of the Gastrointestinal Tract by Using Noninvasive Imaging Methods (US, CT, MR Imaging, PET): A Meta-Analysis Radiology, September 1, 2002; 224(3): 748 - 756. [Abstract] [Full Text] [PDF]

C. H. Coulam, F. P. Chan, and K. C. P. Li Can a Multiphasic Contrast-Enhanced Three-Dimensional Fast Spoiled Gradient-Recalled Echo Sequence Be Sufficient for Liver MR Imaging? Am. J. Roentgenol., February 1, 2002; 178(2): 335 - 341. [Abstract] [Full Text] [PDF]