| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, UCLA, 10833 Le Conte Ave., Los Angeles, CA
90095-1721.
2
Department of Surgery, UCLA, Los Angeles, CA 90095-1721.
Received December 19, 2000;
accepted after revision April 4, 2001.
Address correspondence to S. S. Raman.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. We prospectively evaluated 25 consecutive studies in 24 patients who underwent ferumoxides-enhanced hepatic MR imaging before surgery and intraoperative sonography. Both 1.5-T scanners (13 cases) and 0.2-T scanners (12 cases) were used. Turbo spin-echo T2-weighted sequences were performed before and after the administration of ferumoxides and the images were compared. Lesions were classified as solid or nonsolid and tabulated on standard liver maps. The liver maps from MR imaging were compared with those from surgery and intraoperative sonography. For lesions greater than 1 cm, the regions of interest were measured and contrast-to-noise ratio was calculated.
RESULTS. Of 93 solid lesions found at surgery, 69 were seen on unenhanced MR imaging (sensitivity, 74.2%) and 87 were seen on ferumoxides-enhanced MR imaging (sensitivity, 93.5%) (p < 0.05). Of the seven benign lesions (five cysts, two hemangiomas) found at surgery, all were correctly identified as benign on MR imaging. Two lesions identified as solid before surgery were not found at surgery. Mean lesion contrast-to-noise ratio for the unenhanced scans was 22.9 and 34.5 (p < 0.001) for the ferumoxides-enhanced scans. Subanalysis of 1.5- and 0.2-T MR imaging revealed similar results with significant (p < 0.05) increases in sensitivity for both. The average size of the lesions missed before surgery was 0.7 cm.
CONCLUSION. Turbo spin-echo T2-weighted ferumoxides-enhanced MR imaging at either 1.5 or 0.2 T has value in preoperative liver assessment.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Over the past decade, a number of published studies and numerous abstracts have shown that ferumoxides-enhanced MR imaging has superior sensitivity and accuracy compared with unenhanced MR imaging in the detection of liver metastasis and hepatocellular carcinoma [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23]. Of published studies that have shown improved sensitivity of ferumoxides-enhanced MR imaging for hepatic lesion detection, only a few [11, 14, 24] have used the dosing regimen currently approved in North America (10 µmol/kg) and none has extended the study population to include those ineligible for traditional resection. Although early studies generally used lower field strength magnets and more recent studies used higher field strength magnets, only one study, to our knowledge, has directly compared the sensitivity of T2-weighted sequences before and after the administration of ferumoxides on both high-field and low-field magnets [9]. Furthermore, most studies have used spin-echo T2-weighted sequences; only a few published studies have examined the widely used turbo spin-echo or fast spin-echo T2-weighted sequences [9, 18, 23,24,25]. Finally, few published studies have compared the sensitivity of scans obtained after administration of ferumoxides with the sensitivity of a known standard of reference such as intraoperative sonography, surgical inspection, and palpation [15, 17, 18, 21, 23, 24].
The purpose of this study was to prospectively compare the sensitivity of ferumoxides-enhanced MR imaging for hepatic lesion detection at the 10 µmol/kg dose with both unenhanced MR imaging and the combination of intraoperative sonography, surgical inspection, and palpation. We examined results in our first 24 consecutive patients (25 studies) who were undergoing primarily intraoperative hepatic cryoablation or radiofrequency ablation, with or without resection.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
MR Imaging
All MR imaging was performed either on a 1.5-T conventional unit (Magnetom
Vision; Siemens, Erlangen, Germany) (n = 13) or a 0.2-T open unit
(Magnetom Open; Siemens) (n = 12). Patients were assigned to either
scanner by our schedulers according to availability. All patients underwent
turbo spin-echo T2-weighted axial sequences (Vision scanner: TR range/TE,
3000-6000/99; fat saturation; Open scanner: 3000-6000/102; no fat saturation),
before and after administration of ferumoxides. A phased array torso coil was
used on the 1.5-T scanner. All sequences were obtained with a slice thickness
of 8 mm, an intersection gap of 2 mm, and a field of view of 30-40 cm
depending on patient size. Four to six excitations were averaged. The matrix
size was 126-192 x 256. After unenhanced sequences were performed,
ferumoxides (Feridex, IV; Berlex Laboratories, Wayne, NJ) were administered in
100 mL of 5% glucose solution and slowly infused IV through an in-line 5-µm
filter at a dose of 10 µmol/kg over 30 min (2-4 mL/min). Patients were
scanned within 30 min after the end of ferumoxides infusion.
All MR imaging studies were prospectively read by one of the authors. The unenhanced T2-weighted turbo spin-echo images were interpreted first. The number, size, and location of lesions detected were tabulated on liver maps based on the Couinaud classification [26]. The ferumoxides-enhanced T2-weighted turbo spin-echo images were then interpreted, and the results were tabulated on separate liver maps. Lesions were classified as solid or nonsolid, the latter being either hemangiomas or cysts.
Contrast-to-noise ratios were calculated for all lesions greater than 1 cm
by placing regions of interest over the lesion and representative normal liver
to obtain quantitative measurements of their MR signals. The largest region of
interest possible over the representative portions of the lesion without
including necrotic regions was always attempted. Noise was measured by placing
a large region of interest anterior to the abdomen on the same image, and the
measurement was taken to be the standard deviation (SD) of noise.
Contrast-to-noise is defined as
 |
Intraoperative Sonography and Surgery
Patients underwent open ablation 1-60 days after the ferumoxides scan, with
a mean of 10 days after the MR examination. At surgery, the liver was
mobilized as much as possible and inspected and palpated by the surgeon.
Intraoperative sonography was performed using an SSD scanner (Toshiba Medical
Systems, Tokyo, Japan) with 7-MHz end-fire or side-fire intraoperative probes.
The liver was systematically scanned jointly by an experienced radiologist and
an oncologic surgeon, as described by Soyer et al.
[27]. Intraoperative
sonography added approximately 15 min to the procedure. All examiners were
aware of the results from preoperative MR imaging. The lesions detected on
ferumoxides-enhanced MR scans were specifically targeted for confirmation, and
screening was performed for additional lesions. In 18 cases, cryoablation was
performed alone (n = 7) or in combination with surgical resection
(n = 5), radiofrequency ablation (n = 5), or ethanol
ablation (n = 1). The remainder underwent either radiofrequency
ablation (n = 2) or surgical resection (n = 4). One patient
had disease elsewhere, that precluded local therapy. Additional solid lesions
detected at surgery were not biopsied and underwent resection or ablation. The
results of surgical inspection and intraoperative sonography with respect to
the number, size, and location of lesions were recorded on a separate liver
map. Cystic and solid lesions were differentiated on the basis of the palpated
nature of the lesions and sonographic appearances. Hemangiomas were soft,
blood-filled lesions with typical sonographic hyperechogenicity.
Statistical Analysis
The liver maps for the unenhanced and ferumoxides-enhanced MR imaging and
for the surgical inspection, palpation, and intraoperative sonography were
analyzed with respect to overall lesion number, size, and location. The number
of lesions on unenhanced and contrast-enhanced MR imaging were compared with
those of the gold standard (surgical examination and intraoperative
sonography), and the significance was assessed with the Wilcoxon's signed rank
test. Differences in the contrast-to-noise and signal-to-noise ratios were
analyzed with the Student's t test. Values for p were
considered significant at less than 0.05.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
In the 12 studies performed on the 0.2-T open magnet, 33 of 46 solid lesions were seen on unenhanced scans and 44 of 46 solid lesions were seen on ferumoxides-enhanced MR imaging. The corresponding sensitivity increased from 71.7% (33/46) to 95.7% (44/46) and the accuracy increased from 72.9% (35/48) to 95.8% (46/48) (p < 0.05) (Figs. 1A,1B,2A,2B,3A,3B).
|
|
|
|
|
|
In the 13 studies performed on the 1.5-T magnet, 36 of 47 solid lesions were seen on unenhanced MR imaging, and 43 of 47 solid lesions were seen on ferumoxides-enhanced MR imaging. The sensitivity increased from 76.6% (36/47) before the administration of ferumoxides to 91.5% (43/47) (p < 0.05) after contrast administration. The corresponding accuracy increased from 77.4% (41/53) to 88.9% (48/54) (p < 0.05). Although the increase in sensitivity for lesion detection was higher on the low-field scanner, the difference failed to reach statistical significance (p = 0.12) when compared with the increase in sensitivity in the high-field group.
In addition to the 93 solid lesions detected at surgery, seven nonsolid lesions (five cysts, two hemangiomas) were detected preoperatively and correctly classified. For these lesions, unenhanced scans were helpful because overall turbo spin-echo T2-weighted signal was high relative to solid lesions (Figs. 3A,3B and 4A,4B). We identified two lesions; both were classified as false-positives, which could not be confirmed at surgery. The first lesion, 4 mm in size, was located at the hepatic dome in a patient with metastasis from ovarian cancer (Fig. 5). Direct visualization and palpation were limited by suboptimal mobilization of the liver caused by extensive adhesions in this patient. Intraoperative sonography was compromised by a combination of incomplete mobilization, diffuse fatty liver, and overlying portal structures that attenuated the ultrasound beam in the region of interest. In the second case, a subcapsular lesion in segment VII with intermediate turbo spin-echo T2-weighted signal was located in a known area of prior wedge resection (Fig. 6). No corresponding lesion was found at surgery and the area was deemed to be a postoperative scar.
|
|
|
|
The average size of solid lesions detected on the unenhanced scans was 2.9 cm (range, 0.5-8.2 cm). The average size of solid lesions detected only on the ferumoxides-enhanced scans was 1.1 cm (range, 0.2-2.5 cm). The average size of solid lesions detected only at surgery was 0.7 cm (range, 0.3-1.5 cm). No technical failures or adverse reactions to the contrast material occurred in our study.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In some patients who are ineligible for liver resection, a variety of intraoperative and minimally invasive thermal ablation techniques are now offered in an attempt to extend the benefits of tumor reduction surgery to a larger group [7]. To determine the relative merits of surgical resection, thermal or chemical ablation, or a combination of these techniques, the precise number, size, and location of all liver masses need to be accurately assessed preoperatively, and accurate hepatic imaging remains important. To our knowledge, the usefulness of ferumoxides-enhanced MR imaging in this emerging secondary population has not been reported.
This study differs from others in several respects. First, it prospectively studies a cohort of patients, most of whom are ineligible for traditional surgical resection, mostly undergoing open thermal ablation in combination with surgical resection. Second, the standard of reference used is intraoperative sonography, surgical inspection, and palpation. Although ferumoxides-enhanced MR imaging has been reported to increase sensitivity relative to unenhanced MR imaging, most studies have determined only the relative increase in lesion detection without providing surgical or pathologic standards of reference. Third, this study compares the performance of low-field and high-field imaging on unenhanced and ferumoxides-enhanced scans. Fourth, the study was performed using the United States Food and Drug Administration—approved dose of 10 µmol of ferumoxides per kilogram. Most studies in the English literature have also been conducted abroad using AMI 225 (Endorem; Guerbet, Aulnay-sous-Bois, France) at a dosing regimen of 15 µmol/kg [9, 10, 12, 13, 15,16,17,18,19,20,21,22,23, 25]. Finally, this study uses for comparison only the widely applicable axial turbo spin-echo T2-weighted sequence. Although various sequences have been compared, the usefulness of ferumoxides-enhanced turbo spin-echo or fast spin-echo T2-weighted sequences have been examined by only a few published studies [9, 18, 21, 23,24,25]. Because contemporary practice in the United States relies heavily on fast or turbo spin-echo T2-weighted sequences, the relevance of older studies using spin-echo T2-weighted sequences has been questioned. [28].
Overall, we found that both the low-field and high-field ferumoxides-enhanced MR imaging performed well compared with the gold standard in this nontraditional population. We found a statistically significant increase in both sensitivity and accuracy for lesion detection using turbo spin-echo T2-weighted sequences when compared with surgical examination and intraoperative sonography. The additional lesions seen on the ferumoxides-enhanced scans were considerably smaller (1.1 vs 2.9 cm) than those seen before contrast administration. The average size of the six lesions missed on MR imaging was 0.7 cm. This finding is in keeping with significant overall contrast-to-noise increases for lesions on both the 1.5- and 0.2-T scanners. Furthermore, no significant loss of specificity was noted on the ferumoxides-enhanced sequences; nonsolid lesions were accurately identified by relying mainly on the unenhanced sequences. Our results agree with previous work that relied on intraoperative sonography as a gold standard but used the 15 µmol/kg dose [15, 17, 18, 21, 23].
The average size of the missed solid lesions was 0.7 cm (0.3-1.5 cm). This finding underscores the limited ability of preoperative imaging modalities such as ferumoxides-enhanced MR imaging and CT arterial portography to detect lesions measuring less than 1 cm [15, 21, 23, 28]. Intraoperative sonography remains essential to detect lesions of less than 1 cm in preserved liver segments when surgical resection is performed.
In our series, the increase in sensitivity for solid lesion detection on the ferumoxides-enhanced MR imaging was similar on the high-field and low-field scanners. The larger numeric increase in lesion detection on the 0.2-T scanner was not statistically significant (p = 0.12) when compared with the 1.5-T scanner. Our results support those of Deckers et al. [9], who found no significant difference in lesion detection in patients who underwent both 1.5- and 0.2-T scanning. Despite the limitations of low-field open MR imaging systems, such as lower spatial resolution and lower contrast-to-noise ratio, the addition of ferumoxides increased the lesion-to-liver contrast adequately to over-come these limitations, resulting in high sensitivity for lesion detection. This finding is important because open low-field MR imaging systems are increasingly being installed in the United States for routine clinical applications because they have relatively lower cost and configurations preferable for claustrophobic or large patients.
Furthermore, we evaluated the widely used signal-averaged turbo spin-echo T2-weighted sequences at the United States Food and Drug Administration—approved dose of 10 µmol/kg. Although turbo spin-echo T2-weighted imaging is theoretically less sensitive [25] to the T2 shortening effect of ferumoxides, the standard signal-averaged turbo spin-echo T2-weighted sequence is a robust, time-efficient, and widely used routine for liver MR imaging [9, 24, 28]. Our results are in keeping with previous work using turbo spin-echo T2-weighted sequences at the 15 µmol/kg dose to show a significant improvement in lesion detection after ferumoxides imaging [9, 18, 21, 22, 25].
One limitation of ferumoxides-enhanced scans has been difficulty differentiating blood vessels, hemangiomas, and cysts from true lesions [8, 13, 15, 16, 25]. Nevertheless, we accurately and prospectively identified and classified nonsolid lesions, primarily on the basis of the marked lesion hyperintensity on the unenhanced turbo spin-echo T2-weighted scans. Seven nonsolid lesions were identified preoperatively. All seven lesions were confirmed with intraoperative sonography. Five lesions were sonographically proven to be simple cysts. Two lesions were uniformly hyperechoic sonographically and changed shape with compression, consistent with hemangiomas. These results emphasize that benign lesions are difficult to characterize on ferumoxides-enhanced scans. Unenhanced T2-weighted images or gadopentetate dimeglumine—enhanced images are necessary to accurately diagnose such lesions.
This study has three major limitations. First, the true denominator of all hepatic lesions was not established, because the liver was not resected and pathologic validity could not be provided. However, the ferumoxides-enhanced scan performed well when compared with the currently accepted imaging gold standard, intraoperative sonography. Two false-positive lesions were found in this study. The false-positive 4-mm lesion at the hepatic dome in segment VII was not detected on intraoperative sonography, surgical inspection, or palpation. This fact underscores the limitations of our gold standard, which has reported sensitivities between 80% and 96% [15, 27]. In the other false-positive lesion, persistent wedge-shaped high signal in an area of previous wedge resection could not be excluded as recurrent tumor. At intraoperative inspection, palpation, and sonography, no corresponding lesion was detected in this presumed region of scar (Figs. 5 and 6).
Second, the study was not truly blinded because reviewers, before interpretation of the ferumoxides-enhanced scan, knew the results of the unenhanced T2-weighted scan. We tried to closely approximate clinical practice conditions and therefore designed a prospective study. We decided that a blinded review of prospective MR images compared with operative results was more important than a blinded review between enhanced and unenhanced images. Also, because the MR imaging results were immediately needed for clinical, preoperative decision making, the same observer could not be realistically made unaware in interpreting both unenhanced and contrast-enhanced images. Although these biases result in favor of the contrast-enhanced scan, there was still a significant increase in the number of lesions detected on the ferumoxides-enhanced scan that could not be identified retrospectively with high confidence. Also, results of the ferumoxides-enhanced MR imaging were known before intraoperative sonography, resulting in intrinsic bias. Nevertheless, every effort was made to survey the entire liver by sonography regardless of the MR imaging findings.
Third, the presence of liver cirrhosis in the three patients with
hepatocellular carcinoma somewhat limited the gold standard. In these three
cases, additional correlation with postoperative 
-fetoprotein levels
and imaging was performed. Postoperative -fetoprotein levels declined
in all three patients and follow-up imaging did not detect additional
lesions.
Two lesser limitations may have relevance. First, our study had a small patient sample size, which likely resulted in an overestimate of the true accuracy and specificity at 0.2 T. Although we showed an overall increase in solid lesion detection with high sensitivity, comparison between lesions of differing histology could not be accomplished because the sample size was inadequate. However, given our rigorous criteria of prospective preoperative MR interpretation and subsequent intraoperative surgical and sonographic confirmation for all lesions, a total of 89 solid lesions were individually proven. Statistical significance was achieved in all the comparisons we set out to accomplish. In addition, the criteria used in our study may not apply to some cystic metastases. No cystic metastases occurred in our study cohort.
In conclusion, hepatic MR imaging with 10 µmol/kg of ferumoxides at both 1.5 and 0.2 T is highly sensitive and accurate for detecting liver lesions and correlates well with intraoperative sonography in a nontraditional population.
Acknowledgments
We thank Julia Fendrick for her invaluable assistance in editing the
manuscript.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  D. A. Bluemke, D. Sahani, M. Amendola, T. Balzer, J. Breuer, J. J. Brown, D. D. Casalino, P. L. Davis, I. R. Francis, G. Krinsky, et al. Efficacy and Safety of MR Imaging with Liver-specific Contrast Agent: U.S. Multicenter Phase III Study Radiology, October 1, 2005; 237(1): 89 - 98. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. E. J. Pijl, J. Doornbos, M. N. J. M. Wasser, H. C. van Houwelingen, R. A. E. M. Tollenaar, and J. L. Bloem Quantitative Analysis of Focal Masses at MR Imaging: A Plea for Standardization Radiology, June 1, 2004; 231(3): 737 - 744. [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]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
