Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Reimer, P.
Right arrow Articles by Saini, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Reimer, P.
Right arrow Articles by Saini, S.
(Radiology. 2000;217:152-158.)
© RSNA, 2000

Gastrointestinal Imaging

Hepatic Lesion Detection and Characterization: Value of Nonenhanced MR Imaging, Superparamagnetic Iron Oxide-enhanced MR Imaging, and Spiral CT-ROC Analysis1

Peter Reimer, MD, Nicole Jähnke, MD, Martin Fiebich, PhD, Wolfgang Schima, MD, Filip Deckers, MD, Christian Marx, MD, Nicolaus Holzknecht, MD and Sanjay Saini, MD

1 From the Department of Radiology, Städtisches Klinikum Karlsruhe, Moltkestrasse 90, D-76133 Karlsruhe, Germany (P.R.); the Department of Clinical Radiology, Westfälische Wilhelms-Universität Münster, Germany (N.J., M.F.); the Department of Radiology, University of Vienna, Austria (W.S.); the Department of Radiology, Universitair Ziekenhuis Antwerpen, Edegem, Belgium (F.D.); the Department of Radiology, Knappschafts-Krankenhaus Bochum-Langendreer-Universitätsklinik, Germany (C.M.); the Department of Diagnostic Radiology, Klinikum Grosshadern, Munich, Germany (N.H.); and the Department of Radiology, Division of Abdominal and Interventional Radiology, Massachusetts General Hospital, Boston (S.S.). Received August 12, 1999; revision requested September 28; revision received January 14, 2000; accepted January 27. Address correspondence to P.R. (e-mail: reimerp@uni-muenster.de).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
PURPOSE: To determine the accuracy for detection and characterization of focal hepatic lesions of nonenhanced, superparamagnetic iron oxide (SPIO)-enhanced, or a combination of nonenhanced and SPIO-enhanced MR imaging and contrast-enhanced spiral computed tomography (CT).

MATERIALS AND METHODS: Spiral CT and T2-weighted SPIO-enhanced (ferucarbotran-enhanced) MR imaging were performed in 35 patients within 2 weeks before surgery for malignant hepatic lesions. Only malignant lesions with histopathologic proof were considered. A total of 875 images with and 800 images without focal lesions were presented to five readers, who were asked to assess the presence and characterization of lesions by using a five-point confidence scale. Receiver operating characteristic analysis was performed.

RESULTS: Nonenhanced and SPIO-enhanced images together and SPIO-enhanced images alone yielded the best performance for lesion detection. No differences were found among all imaging techniques with regard to lesion characterization (benign vs malignant). The combined approach resulted in larger area under the ROC curve (Az = 0.9062) and accuracy (85.3%) (P < 0.02), as compared with SPIO-enhanced MR imaging (Az = 0.8667; accuracy, 73.1%).

CONCLUSION: SPIO-enhanced T2-weighted MR imaging was more accurate than nonenhanced T1-weighted and T2-weighted MR imaging and contrast-enhanced spiral CT for the detection of focal hepatic lesions. The combined analysis of nonenhanced and SPIO-enhanced images was more accurate in the characterization of focal hepatic lesions than was review of SPIO-enhanced images alone.

Index terms: Contrast media • Iron • Liver neoplasms, 761.319, 761.323, 761.33 • Liver neoplasms, MR, 761.121411, 761.121412, 761.12143 • Magnetic resonance (MR), contrast media, 761.12143


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Imaging of the liver is performed primarily to screen patients with known primary malignancies, to screen patients at high risk, or to characterize lesions detected with other modalities. The detection and characterization of focal hepatic lesions are of equal importance as regards accurate treatment planning. Metastases are the most common malignant liver tumors in Western countries, whereas hepatocellular carcinoma is predominant in African and Asian countries. Lesion characterization is of particular importance in patients with a primary malignancy, because up to 50% of small (<10–15-mm) lesions may be benign (1).

Superparamagnetic iron oxide (SPIO) such as ferumoxides has been developed as a liver-specific particulate magnetic resonance (MR) imaging contrast agent that is primarily taken up by the Kupffer cells of the liver, among other types of cells. MR imaging with commercially available SPIO has been demonstrated to be substantially more sensitive than conventional contrast agent–enhanced computed tomography (CT) (2) and nonenhanced MR imaging (3–8) in the detection of focal liver lesions. Seneterre et al (9) described SPIO-enhanced MR imaging to be at least as accurate as CT during arterial portography in helping detect focal liver lesions, and the specificity of SPIO-enhanced MR imaging was higher than that of CT during arterial portography, which was associated with a large number of false-positive results. Contrast-enhanced spiral CT currently is the most widely used CT technique for the detection and characterization of focal hepatic lesions and is associated with a smaller number of false-positive results than is CT during arterial portography (10).

The purpose of our study was to compare the accuracy for the detection and characterization of focal hepatic lesions of nonenhanced MR imaging; SPIO-enhanced MR imaging with a recently developed SPIO, ferucarbotran (formerly SH U 555 A) (11,12); and contrast-enhanced spiral CT. To this end, we used a receiver operating characteristic (ROC) analysis with multiple observers from five institutions. We also studied whether a combined reading of nonenhanced and SPIO-enhanced T2-weighted MR images would improve lesion detection and characterization.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Patients
A total of 70 patients were enrolled in two phase 3 trials with ferucarbotran (Resovist; Schering, Berlin, Germany) between 1994 and 1997 at Westfälische Wilhelms-Universität (Münster, Germany). Local ethical committee approval was granted, and informed consent was obtained from each patient prior to entry into the phase 3 studies. Initially, only patients with five or fewer solid, focal, malignant or benign liver lesions diagnosed at nonenhanced MR imaging, gadolinium-enhanced MR imaging, contrast-enhanced spiral CT, or ultrasonography within 4 weeks prior to entering the study were included. Of this population, patients who had undergone contrast-enhanced spiral CT within 2 weeks prior to SPIO-enhanced MR imaging were identified, and all MR and CT images were reviewed by the study coordinator (N.J.) together with available pathology reports for the purpose of locating lesions in which histopathologic proof had been obtained. The number of analyzed lesions per patient was restricted to one to avoid bias due to inclusion of multiple lesions in individual patients.

Finally, 35 patients were enrolled for the ROC analysis, 26 with malignant and nine with benign liver tumors. There were 18 patients with liver metastases (15 with metastases from colorectal cancer and one each with metastases from breast cancer, pancreatic cancer, or extrahepatic cholangiocarcinoma) and eight patients with hepatocellular carcinoma. Results of liver surgery were available in 21 patients, and results of percutaneous biopsy were available in the remaining five patients. Lesions were determined to be benign on the basis of biopsy results (in two patients) and typical imaging findings and absence of growth during at least 2 years of follow-up. Six patients had a hemangioma, and one each had an adenoma, focal nodular hyperplasia, and abscess. A single lesion was present in 14 patients, and multiple lesions were present in 21 patients. Subsequently, a total of 35 images (one from each patient) with focal liver lesions were selected for further analysis. To ensure a comparable section position, 32 sections without focal liver lesions and any relevant artifacts were selected for comparison, to minimize for false-positive decisions.

MR Imaging
MR imaging was performed with either a 1.0-T system (Magnetom Expert; Siemens Medical Systems, Erlangen, Germany) or a 1.5-T system (Magnetom Vision, Siemens Medical Systems), both equipped with a body phased-array coil. Before injection of SPIO, transverse T1-weighted spoiled gradient-echo images (fast low-angle shot [FLASH]) and transverse double-echo T2-weighted turbo spin-echo (SE) images were obtained. Double-echo T2-weighted turbo SE imaging was again performed 20–30 minutes after administration of SPIO.

At 1.0 T, transverse T1-weighted FLASH images were acquired with the following parameters: 170/5 (repetition time msec/echo time msec), flip angle of 70°, 21-second breath hold, eight sections, and 128 x 256 (phase-by-frequency) matrix. Therefore, two or three separate acquisitions were performed to ensure adequate anatomic coverage. At 1.5 T, transverse T1-weighted FLASH images were acquired with the following parameters: 175/4.1, flip angle of 80°, 22-second breath hold, 23 sections, and 128 x 256 matrix. At both 1.0 and 1.5 T, transverse double-echo turbo SE images were obtained with the following parameters: 4,600/83, 165 (effective echo times), echo train length of nine, three signals acquired.

For all MR sequences, 17 sections were acquired within 6 minutes 31 seconds, with a 252 x 256 matrix. The section thickness was 8 mm with a 25% gap and a rectangular field of view 400 x 400 mm or smaller. The phase-encoding axis was anteroposterior, and presaturation pulses were placed above and below the imaging volume for artifact reduction. Radio-frequency transmit and receive values were set for nonenhanced images and were kept constant for enhanced images.

CT Examination
Spiral CT studies were performed with two units (SR 7000, Philips Medical Systems, Best, the Netherlands; Somatom Plus 4, Siemens Medical Systems) with minor variations in CT scanning techniques. Scanning parameters were 120 kV, 210 mA, collimation of 5–10 mm, and table increment of 5-10 mm/sec. Images were reconstructed every 5–10 mm to provide contiguous or overlapping sections, and the spiral acquisition time was 20–30 seconds, depending on liver size. Nonionic contrast medium (300 mg iodine per milliliter) was administered intravenously as a uniphasic 150-mL bolus of at a rate of 3 mL/sec. Image acquisition commenced 50–70 seconds after initiation of the contrast medium injection with a power injector in all patients. Additional arterial phase acquisitions in 16 patients suspected of having a hypervascular lesion were begun 20 seconds after the start of injection.

MR Contrast Agent: Ferucarbotran
Ferucarbotran consists of SPIO microparticles coated with carboxydextran. This contrast agent has an R1 relaxation rate of approximately 12.3 sec · mmol-1 and an R2 relaxation rate of approximately 188 sec · mmol-1 as measured in plasma at 40 MHz (37°C). Specific pharmacologic and biologic properties of the compound have been described previously (11–13). Patients received a total dose of 10 µmol iron per kilogram body weight (0.9–1.4-mL injection volume), which has been demonstrated (14,15) to be effective for the detection and characterization of focal hepatic lesions. Ferucarbotran was preloaded into a connecting intravenous tube (Connection Tubing; Clinico, Bad Hersfeld, Germany) and was manually injected through a filter with a 5-µm pore size by flushing the connecting catheter with 10 mL of saline solution within 3 seconds (injection rate, approximately 3 mL/sec).

Image Analysis
The MR sequences and the combined CT images were viewed independently from each other by five independent observers (C.M., N.H., F.D., W.S., S.S.) from five institutions. The five observers were blinded with regard to the diagnosis and clinical history. They reviewed 875 sections (175 sections per observer) with focal lesions and 800 sections (160 sections per observer) without focal liver lesions. Six hundred fifty sections displayed malignant lesions, and 225 sections displayed benign lesions. The subgroup of 875 sections with focal lesions was again reviewed for lesion characterization and to determine a specific diagnosis.

The size distribution among benign and malignant lesions is shown in Table 1. We presented the readers with single sections with focal lesions (based on available histopathologic findings), and we tried to avoid presenting images with partial volume effects from lesions or organs for sections without lesions. In the clinical setting, readings are performed not only with sections obtained with a single pulse sequence but with those obtained with all pulse sequences used. Therefore, a section-based analysis does not completely reflect the clinical situation but will, however, show the differences between the different techniques (16,17). Therefore, double-echo T2-weighted turbo SE MR images were presented as follows: nonenhanced images alone, contrast-enhanced images alone, or combined enhanced and nonenhanced images. Data from both echoes of the double-echo T2-weighted turbo SE sequence were presented together.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Size of Lesions
 
Images were randomized and presented at two sessions, with patient data masked. Each observer recorded the presence and segmental location of lesions according to the classification scheme of Couinaud. They assigned to each finding a confidence rating by using a five-point scale: score of 1 for definitely not a lesion, 2 for probably not a lesion, 3 for indeterminate, 4 for a probable lesion, and 5 for a definite lesion. Distributions of confidence scores are shown in Table 2. Observers were asked to further classify detected lesions (those with a confidence rating of 4 or 5) as benign or malignant by assigning a confidence rating with the use of a five-point scale: score of 1 for definitely benign, 2 for probably benign, 3 for indeterminate, 4 for probably malignant, and 5 for definitely malignant. Subsequently, observers were asked to name the tumor type, based on their characterization as benign or malignant according to established criteria in the literature. In general, lesions that showed no enhancement on SPIO-enhanced T2-weighted double-echo turbo SE images were considered to be malignant or cystic (high signal intensity on nonenhanced T2-weighted turbo SE images); lesions that showed enhancement on SPIO-enhanced T2-weighted double-echo turbo SE images were considered to be benign (18–22). All observes were trained with five test cases not included in the analysis and were aware that sensitivity calculations would be determined only for those lesions assigned a confidence rating of 4 or 5.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Distribution of Confidence Ratings for ROC Analysis of Lesion Detection
 
ROC curves for each observer and for each study were calculated by using the ROCFIT program (C. Metz, University of Chicago, Ill), which plotted the true-positive fraction against the likelihood of a false-positive image (16,23–25). The area under each ROC curve (Az) was used to indicate the overall performance of imaging techniques and observers (26,27). The sensitivity for each observer and each technique was also assessed for only those lesions assigned a confidence rating of 1 or 2 versus 4 or 5. For each MR study, the mean Az value and the mean sensitivity for all five observers were calculated, and the significance of any difference was assessed by using analysis of variance (28).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
DISCUSSION
 REFERENCES
 
The Az values for lesion detection for each observer and each technique are shown in Table 3 and Figure 1a. Because interobserver variability was not significant for any of the imaging techniques, the data for the five readers were pooled. The mean sensitivities and specificities for each technique are shown in Table 4. The mean Az values for SPIO-enhanced T2-weighted double-echo turbo SE MR images alone (Az = 0.9890) and for nonenhanced and SPIO-enhanced T2-weighted double-echo turbo SE images combined (Az = 0.9903) were significantly greater (P < .01) than those for contrast-enhanced spiral CT images (Az = 0.8968), T1-weighted FLASH MR images (Az = 0.9308), and nonenhanced T2-weighted double-echo turbo SE MR images alone (Az = 0.9332). The latter three techniques showed no significant differences with regard to lesion detection. The sensitivity for lesion detection was also significantly higher (P < .01) for SPIO-enhanced T2-weighted double-echo turbo SE MR images alone (95.4%) and for both nonenhanced and SPIO-enhanced T2-weighted double-echo turbo SE MR images (97.1%). There were no significant differences in the specificity values among the three techniques.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Az Values for Each Observer and Imaging Technique
 


View larger version (36K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a. ROC curves for lesion (a) detection and (b) characterization calculated with means from all five observers for contrast-enhanced (CE) spiral CT scans and T1-weighted FLASH, nonenhanced T2-weighted turbo SE (T2-TSE pre), SPIO-enhanced T2-weighted turbo SE (T2-TSE post), and combined nonenhanced and SPIO-enhanced T2-weighted turbo SE (T2-TSE comb) MR images. (a) SPIO-enhanced T2-weighted turbo SE MR images alone and nonenhanced and SPIO-enhanced T2-weighted turbo SE MR images combined resulted in the best performance for detection of focal hepatic lesions. (b) Nonenhanced and SPIO-enhanced T2-weighted turbo SE MR images combined resulted in the best performance for characterization of focal hepatic lesions.

 


View larger version (43K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b. ROC curves for lesion (a) detection and (b) characterization calculated with means from all five observers for contrast-enhanced (CE) spiral CT scans and T1-weighted FLASH, nonenhanced T2-weighted turbo SE (T2-TSE pre), SPIO-enhanced T2-weighted turbo SE (T2-TSE post), and combined nonenhanced and SPIO-enhanced T2-weighted turbo SE (T2-TSE comb) MR images. (a) SPIO-enhanced T2-weighted turbo SE MR images alone and nonenhanced and SPIO-enhanced T2-weighted turbo SE MR images combined resulted in the best performance for detection of focal hepatic lesions. (b) Nonenhanced and SPIO-enhanced T2-weighted turbo SE MR images combined resulted in the best performance for characterization of focal hepatic lesions.

 

View this table:
[in this window]
[in a new window]

  		TABLE 4. Sensitivity and Specificity for Lesion Detection
 
The ROC curves (Az values) for lesion characterization are shown in Figure 1b. Interobserver variability was not significant for all techniques. The mean accuracy for each technique is shown in Table 5. Reading of nonenhanced and SPIO-enhanced T2-weighted double-echo turbo SE images together resulted in a mean Az value (Az = 0.9062) and accuracy (85.3%) that were significantly higher (P < .02) than those for all other techniques. There were no significant differences in these measures among readings of contrast-enhanced spiral CT images (mean Az value, 0.8056; accuracy, 69.5%), nonenhanced T2-weighted double-echo turbo SE MR images (mean Az value, 0.8405; accuracy, 70.1%), and SPIO-enhanced T2-weighted double-echo turbo SE MR images (mean Az value, 0.8667; accuracy, 73.1%). Readings of T1-weighted FLASH MR images resulted in a significantly lower Az value and accuracy (P < .05) than did the T2-weighted MR techniques. When a specific diagnosis was requested, the accuracy decreased to between 59.7% and 71.7% (Table 5).


View this table:
[in this window]
[in a new window]

  		TABLE 5. Accuracy for Characterization of Detected Lesions as Benign or Malignant and for Diagnosis of Tumor Type
 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
REFERENCES
 
In the present study, we evaluated the performance of nonenhanced MR imaging, SPIO-enhanced MR imaging, and contrast-enhanced spiral CT for the detection and characterization of focal liver lesions. In addition, we studied whether a combined approach of reading nonenhanced and SPIO-enhanced MR images might offer advantages over the acquisition of SPIO-enhanced images alone for the detection and characterization of focal liver lesions.

The highest accuracies for detection of focal liver lesions were achieved in our study with reading of SPIO-enhanced T2-weighted MR images alone and with the combined approach of readings of nonenhanced and SPIO-enhanced T2-weighted MR images together (Tables 4). However, the differences among the five imaging techniques were not significant. Both, the combined reading approach and SPIO-enhanced image reading approach provided the best performance for lesion detection (Figs 1–3). Our results are within the range of those in published reports (9,22,29) on the accuracy of SPIO-enhanced MR imaging for lesion detection; however, the 97.1% sensitivity is likely higher than that encountered in clinical practice due to selection of lesions that had been proved at histopathologic analysis. There were no significant differences among spiral CT, nonenhanced T2-weighted MR, and SPIO-enhanced T2-weighted MR images with regard to the characterization of detected lesions as benign or malignant (Table 5). However, the combined approach of presenting nonenhanced and enhanced T2-weighted MR images together resulted in a significantly higher sensitivity. Use of nonenhanced and SPIO-enhanced MR images together resulted in significantly more accurate differentiation of benign from malignant lesions than did reading spiral CT images, nonenhanced T2-weighted MR images, or SPIO-enhanced T2-weighted MR images alone. The determination of the diagnosis of specific tumor types, performed after the clinically more relevant characterization as benign or malignant, was not improved significantly (Table 5).



View larger version (140K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2a. Liver metastasis from colorectal carcinoma. A 12-mm-diameter metastasis was proved at surgery to be present in segment 8 of the cranial right lobe of the liver. (a) Transverse contrast-enhanced spiral CT scan shows the ring-enhancing lesion (arrowhead). (b) Transverse T1-weighted FLASH MR image (175/4.1, 80° flip angle) demonstrates a moderately hypointense lesion (arrowhead). (c) Transverse nonenhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates a moderately hyperintense lesion (arrowhead). (d) Transverse SPIO-enhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates substantially improved lesion conspicuity and clearly depicts the lesion (arrowhead). All readers rated the lesion as a definite lesion and a definite metastasis when comparing nonenhanced and SPIO-enhanced T2-weighted MR images.

 


View larger version (176K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2b. Liver metastasis from colorectal carcinoma. A 12-mm-diameter metastasis was proved at surgery to be present in segment 8 of the cranial right lobe of the liver. (a) Transverse contrast-enhanced spiral CT scan shows the ring-enhancing lesion (arrowhead). (b) Transverse T1-weighted FLASH MR image (175/4.1, 80° flip angle) demonstrates a moderately hypointense lesion (arrowhead). (c) Transverse nonenhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates a moderately hyperintense lesion (arrowhead). (d) Transverse SPIO-enhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates substantially improved lesion conspicuity and clearly depicts the lesion (arrowhead). All readers rated the lesion as a definite lesion and a definite metastasis when comparing nonenhanced and SPIO-enhanced T2-weighted MR images.

 


View larger version (172K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2c. Liver metastasis from colorectal carcinoma. A 12-mm-diameter metastasis was proved at surgery to be present in segment 8 of the cranial right lobe of the liver. (a) Transverse contrast-enhanced spiral CT scan shows the ring-enhancing lesion (arrowhead). (b) Transverse T1-weighted FLASH MR image (175/4.1, 80° flip angle) demonstrates a moderately hypointense lesion (arrowhead). (c) Transverse nonenhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates a moderately hyperintense lesion (arrowhead). (d) Transverse SPIO-enhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates substantially improved lesion conspicuity and clearly depicts the lesion (arrowhead). All readers rated the lesion as a definite lesion and a definite metastasis when comparing nonenhanced and SPIO-enhanced T2-weighted MR images.

 


View larger version (172K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2d. Liver metastasis from colorectal carcinoma. A 12-mm-diameter metastasis was proved at surgery to be present in segment 8 of the cranial right lobe of the liver. (a) Transverse contrast-enhanced spiral CT scan shows the ring-enhancing lesion (arrowhead). (b) Transverse T1-weighted FLASH MR image (175/4.1, 80° flip angle) demonstrates a moderately hypointense lesion (arrowhead). (c) Transverse nonenhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates a moderately hyperintense lesion (arrowhead). (d) Transverse SPIO-enhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates substantially improved lesion conspicuity and clearly depicts the lesion (arrowhead). All readers rated the lesion as a definite lesion and a definite metastasis when comparing nonenhanced and SPIO-enhanced T2-weighted MR images.

 


View larger version (195K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3a. Surgically proved 3-cm-diameter hepatocellular carcinoma in segments 5 and 6 of the caudal right lobe of the liver. Transverse (a) contrast-enhanced CT and (b) T1-weighted FLASH MR (175/4.1, 80° flip angle) images do not demonstrate the lesion. (c) Transverse nonenhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates poor lesion conspicuity (arrowhead). (d) Transverse SPIO-enhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates moderately improved lesion conspicuity (arrowhead). All readers rated the lesion as a definite lesion and probable hepatocellular carcinoma when comparing nonenhanced and SPIO-enhanced T2-weighted MR images.

 


View larger version (189K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3b. Surgically proved 3-cm-diameter hepatocellular carcinoma in segments 5 and 6 of the caudal right lobe of the liver. Transverse (a) contrast-enhanced CT and (b) T1-weighted FLASH MR (175/4.1, 80° flip angle) images do not demonstrate the lesion. (c) Transverse nonenhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates poor lesion conspicuity (arrowhead). (d) Transverse SPIO-enhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates moderately improved lesion conspicuity (arrowhead). All readers rated the lesion as a definite lesion and probable hepatocellular carcinoma when comparing nonenhanced and SPIO-enhanced T2-weighted MR images.

 


View larger version (199K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3c. Surgically proved 3-cm-diameter hepatocellular carcinoma in segments 5 and 6 of the caudal right lobe of the liver. Transverse (a) contrast-enhanced CT and (b) T1-weighted FLASH MR (175/4.1, 80° flip angle) images do not demonstrate the lesion. (c) Transverse nonenhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates poor lesion conspicuity (arrowhead). (d) Transverse SPIO-enhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates moderately improved lesion conspicuity (arrowhead). All readers rated the lesion as a definite lesion and probable hepatocellular carcinoma when comparing nonenhanced and SPIO-enhanced T2-weighted MR images.

 


View larger version (202K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3d. Surgically proved 3-cm-diameter hepatocellular carcinoma in segments 5 and 6 of the caudal right lobe of the liver. Transverse (a) contrast-enhanced CT and (b) T1-weighted FLASH MR (175/4.1, 80° flip angle) images do not demonstrate the lesion. (c) Transverse nonenhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates poor lesion conspicuity (arrowhead). (d) Transverse SPIO-enhanced T2-weighted turbo SE MR image (4,600/83 [effective]) demonstrates moderately improved lesion conspicuity (arrowhead). All readers rated the lesion as a definite lesion and probable hepatocellular carcinoma when comparing nonenhanced and SPIO-enhanced T2-weighted MR images.

 
Technical improvements in CT and MR imaging, including the acquisition of thinner sections, reduced acquisition time, and better signal-to-noise ratios, may lead to better conspicuity of small lesions. Although both CT and MR imaging were performed with equipment that was state of the art at the time of the study, fewer small (<1-cm) lesions were present than larger lesions. Several authors (3,6,8) have described the difficulty of distinguishing small lesions from peripheral vessels and interference from susceptibility effects with small lesions, which are more conspicuous after enhancement with SPIO. Currently, contrast-enhanced CT or gadolinium-enhanced MR imaging are best suited to facilitate the differentiation of small lesions from vessels; however, dynamic MR imaging during injection of SPIO may help overcome this problem in the future (30,31). The availability of nonenhanced MR images for the combined reading may explain the superior performance for lesion detection and the minimized risk of a false-positive diagnosis of a vessel as a lesion. The combination also improved characterization, because intratumoral signal intensity changes in lesions that contained cells of the reticuloendothelial system (eg, focal nodular hyperplasia) or in hemangiomas could be taken into consideration (32).

Our choice of pulse sequences requires explanation. We included nonenhanced MR images because such images are helpful in distinguishing small lesions from vessels, in the characterization of lesions, and in the distinction of the liver from adjacent structures. Nonenhanced T1-weighted breath-hold FLASH MR images are useful for the detection of lesions, and acquisition of such images adds little to the overall examination time. High-quality images are obtained with the newer technique, available at 1.5 T, of covering the liver within a single breath hold free of misregistration and motion artifacts. We did not repeat this sequence after administration of SPIO because, at a field strength of 1.0 T, the decrease in signal intensity of normal liver produced by SPIO is less than that at 1.5 T. On SPIO-enhanced T1-weighted FLASH MR images obtained at 1.0 T, liver parenchyma is typically isointense to focal lesions. For T2-weighted MR imaging we used a double-echo turbo SE sequence with a relatively short echo train length and no fat suppression; this sequence is suitable for both lesion detection and lesion characterization. Turbo SE MR imaging is performed with a short acquisition time, which results in fewer motion artifacts than occur with conventional SE imaging. We repeated the turbo SE sequence with SPIO enhancement to compare the same technique and to prevent technique-based changes. In a recent study (29), turbo SE MR imaging was superior to fat-suppressed T2-weighted SE and gadolinium-enhanced T1-weighted spoiled gradient-echo imaging for the detection of focal liver lesions. Schwartz et al (33) reported that turbo SE imaging is superior to conventional SE and that no additional information is provided by applying fat suppression. Gaa et al (34) pointed out that fat suppression clearly improves the quality of T2-weighted images and minimizes respiratory artifacts from high-signal-intensity fat. We did not perform a comparison with gadolinium-enhanced MR imaging, which is a routine technique for abdominal MR imaging. This comparison should be addressed in future studies (35).

We decided to use rigorous inclusion and exclusion criteria without consensus readings, to minimize the risk of false-positive lesions being presented to the readers. A higher certainty for the absence of lesions can be achieved by focusing on patients scheduled to undergo liver transplantation. According to the present data, SPIO-enhanced MR imaging was more sensitive than nonenhanced MR imaging and contrast-enhanced spiral CT with regard to the detection and characterization of focal hepatic lesions. The review of nonenhanced and SPIO-enhanced MR images combined provided better results than did the review of SPIO-enhanced MR images alone. Therefore, we believe that the acquisition of nonenhanced MR images should be mandatory. Our data also justify the continued development of specific contrast agents to help improve clinical performance with hepatic MR imaging. Further studies should also focus on improvements in the accuracy for detection and characterization of small (<1-cm) lesions.

In conclusion, the use of a combination of nonenhanced and SPIO-enhanced T2-weighted MR images marginally improved the detection and substantially improved the characterization of focal hepatic lesions and resulted in higher diagnostic accuracy than did contrast-enhanced spiral CT or SPIO-enhanced T2-weighted MR imaging alone. Longer examination and analysis times and higher costs offset this advantage.


   FOOTNOTES
 
Abbreviations: Az = area under the ROC curve, FLASH = fast low-angle shot, ROC = receiver operating characteristic, SE = spin echo, SPIO = superparamagnetic iron oxide

Author contributions: Guarantor of integrity of entire study, P.R.; study concepts, P.R.; study design, P.R., M.F.; definition of intellectual content, P.R., N.J., M.F.; literature research, P.R.; clinical studies, P.R.; data acquisition, N.J.; data analysis, W.S., F.D., C.M., N.H., S.S.; statistical analysis, N.J., M.F.; manuscript preparation, P.R.; manuscript editing, N.J., M.F.; manuscript review, W.S., F.D., C.M., N.H., S.S.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 

  1. Karhunen PJ. Benign hepatic tumors and tumor like conditions in men. J Clin Pathol 1986; 39:183-189.[Abstract/Free Full Text]
  2. Hagspiel KD, Neidel KFW, Eichenberger AC, Weder W, Marincek B. Detection of liver metastases: comparison of superparamagnetic iron-oxide–enhanced MR imaging at 1.5 T with dynamic CT, intraoperative US, and percutaneous US. Radiology 1995; 196:471-478.[Abstract/Free Full Text]
  3. Stark DD, Weissleder R, Elizondo G, et al. Superparamagnetic iron oxide: clinical application as a contrast agent for MR imaging of the liver. Radiology 1988; 168:297-301.[Abstract/Free Full Text]
  4. Fretz CJ, Elizondo G, Weissleder R, Hahn PF, Stark DD, Ferrucci JT. Superparamagnetic iron oxide–enhanced MR imaging: pulse sequence optimization for detection of liver cancer. Radiology 1989; 172:393-397.[Abstract/Free Full Text]
  5. Ferrucci JT, Stark DD. Iron oxide-enhanced MR imaging of the liver and spleen: review of the first 5 years. AJR Am J Roentgenol 1990; 155:943-950.[Abstract/Free Full Text]
  6. Winter TC, III, Freeny PC, Nghiem HV, et al. MR imaging with i.v. superparamagnetic iron oxide: efficacy in the detection of focal hepatic lesions. AJR Am J Roentgenol 1993; 161:1191-1198.[Abstract/Free Full Text]
  7. Bellin MF, Zaim S, Auberton E, et al. Liver metastases: safety and efficacy of detection with superparamagnetic iron oxide in MR imaging. Radiology 1994; 193:657-663.[Abstract/Free Full Text]
  8. Ros PR, Freeny PC, Harms SE, et al. Hepatic MR Imaging with ferumoxides: a multicenter clinical trial of the safety and efficacy in the detection of focal hepatic lesions. Radiology 1995; 196:481-488.[Abstract/Free Full Text]
  9. 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. Radiology 1996; 200:785-792.[Abstract/Free Full Text]
  10. Soyer P, Lacheheb M, Levesque M. False positive diagnosis based on CT portography: correlation with pathological findings. AJR Am J Roentgenol 1992; 160:285-289.[Abstract/Free Full Text]
  11. Hamm B, Staks T, Taupitz M, et al. Contrast enhanced MR imaging of liver and spleen: first experience in humans with a new superparamagnetic iron oxide. J Magn Reson Imaging 1994; 4:659-668.[Medline]
  12. Hamm B, Staks T, Taupitz M. A new superparamagnetic iron oxide contrast agent for magnetic resonance imaging. Invest Radiol 1994; 29(suppl 2):S87-S89.
  13. Reimer P, Müller M, Marx C, et al. T1 effects of a bolus-injectable superparamagnetic iron oxide, SH U 555 A: dependence on field strength and plasma concentration—preliminary clinical experience with dynamic T1-weighted MR imaging. Radiology 1998; 209:831-836.[Abstract/Free Full Text]
  14. 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]
  15. Vogl TJ, Hammerstingl R, Schwarz W, et al. Magnetic resonance imaging of focal liver lesions: comparison of the superparamagnetic iron oxide Resovist versus gadolinium-DTPA in the same patient. Invest Radiol 1996; 31:696-708.[Medline]
  16. Metz CE. Basic principles of ROC analysis. Semin Nucl Med 1978; 8:283-298.[Medline]
  17. Metz CE. Some practical issues of experimental design and data analysis in radiological ROC studies. Invest Radiol 1989; 24:234-245.[Medline]
  18. Vogl TJ, Hammerstingl R, Schwarz W, et al. Superparamagnetic iron oxide–enhanced versus gadolinium-enhanced MR imaging for differential diagnosis of focal liver lesions. Radiology 1996; 198:881-887.[Abstract/Free Full Text]
  19. Mitchell DG. Hepatic imaging: techniques and unique applications of magnetic resonance imaging. Magn Reson Q 1993; 9:84-112.[Medline]
  20. Larson RE, Semelka RC. Magnetic resonance imaging of the liver. Top Magn Reson Imaging 1995; 7:71-81.[Medline]
  21. Reimer P, Tombach B. Hepatic MRI with SPIO: detection and characterization of focal liver lesions. Eur Radiol 1998; 8:1198-1204.[Medline]
  22. Ward J, Naik KS, Guthrie JA, Wilson D, Robinson PJA. 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]
  23. Metz CE. ROC methodology in radiological imaging. Invest Radiol 1986; 21:720-733.[Medline]
  24. Dorfman DD, Berbaum KS. Degeneracy and discrete receiver operating characteristic rating data. Acad Radiol 1995; 2:907-215.[Medline]
  25. Chakraborty DP, Winter LHL. Free response methodology: alternate analysis and a new observer—performance experiment. Radiology 1990; 174:873-881.[Abstract/Free Full Text]
  26. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982; 143:29-36.[Abstract/Free Full Text]
  27. Ransohoff DF, Feinstein AR. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests. N Engl J Med 1978; 299:926-930.[Abstract]
  28. Wassertheil-Smoller S. Biostatistics and epidemiology: a primer for health professionals 2nd ed. New York, NY: Springer-Verlag, 1995; 185.
  29. Blakeborough A, Ward J, Wilson D, et al. Hepatic lesion detection at MR imaging: a comparative study with four sequences. Radiology 1997; 203:759-765.[Abstract/Free Full Text]
  30. Fretz CJ, Stark DD, Metz CE, et al. Detection of hepatic metastases: comparison of contrast-enhanced CT, unenhanced MR imaging, and iron oxide-enhanced MR imaging. AJR Am J Roentgenol 1990; 155:763-770.[Abstract/Free Full Text]
  31. Mergo PJ, Helmberger T, Nicolas AI, Ros PR. Ring enhancement in ultrasmall superparamagnetic iron oxide imaging: a potential new sign for characterization of liver lesions. AJR Am J Roentgenol 1996; 166:379-384.[Abstract/Free Full Text]
  32. Hahn PF, Stark DD, Elizondo G, Weissleder R, Saini S, Ferrucci JT. Clinical application of superparamagnetic iron oxide to MR imaging of tissue perfusion in vascular liver tumors. Radiology 1990; 174:361-366.[Abstract/Free Full Text]
  33. Schwartz LH, Seltzer SE, Tempany CM, et al. Prospective comparison of T2-weighted fast spin-echo, with and without fat suppression, and conventional spin-echo pulse sequences in the upper abdomen. Radiology 1993; 189:411-416.[Abstract/Free Full Text]
  34. Gaa J, Hatabu H, Jenkins RL, Finn JP, Edelman RR. Liver masses: replacement of conventional T2-weighted spin-echo MR imaging with breath-hold MR imaging. Radiology 1996; 200:459-464.[Abstract/Free Full Text]
  35. Semelka RC, Canca WG, Marcos HB, Mauro MA. Liver metastases: comparison of current MR techniques and spiral CT during arterial portography for detection in 20 surgically staged cases. Radiology 1999; 213:86-91.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 JCOHome page
J. Ward, J. A. Guthrie, M. B. Sheridan, S. Boyes, J. T. Smith, D. Wilson, J. I. Wyatt, D. Treanor, and P. J. Robinson
Sinusoidal Obstructive Syndrome Diagnosed With Superparamagnetic Iron Oxide-Enhanced Magnetic Resonance Imaging in Patients With Chemotherapy-Treated Colorectal Liver Metastases
J. Clin. Oncol., September 10, 2008; 26(26): 4304 - 4310.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
X. Yang
Nano- and Microparticle-based Imaging of Cardiovascular Interventions: Overview
Radiology, May 1, 2007; 243(2): 340 - 347.
[Abstract] [Full Text] [PDF]

Home page
 RadioGraphicsHome page
T. J. Kim, K. H. Lee, Y. H. Kim, S. W. Sung, S. Jheon, S.-k. Cho, and K. W. Lee
Postoperative Imaging of Esophageal Cancer: What Chest Radiologists Need to Know
RadioGraphics, March 1, 2007; 27(2): 409 - 429.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
M.-J. Kim, J. H. Kim, J. Y. Choi, S. H. Park, J.-J. Chung, K. W. Kim, and D. G. Mitchell
Optimal TE for SPIO-enhanced gradient-recalled echo MRI for the detection of focal hepatic lesions.
Am. J. Roentgenol., September 1, 2006; 187(3): W255 - W266.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
J. Schnorr, S. Wagner, C. Abramjuk, R. Drees, T. Schink, E. A. Schellenberger, H. Pilgrimm, B. Hamm, and M. Taupitz
Focal Liver Lesions: SPIO-, Gadolinium-, and Ferucarbotran-enhanced Dynamic T1-weighted and Delayed T2-weighted MR Imaging in Rabbits
Radiology, July 1, 2006; 240(1): 90 - 100.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
H. J. Kim, K. W. Kim, J. H. Byun, H. J. Won, Y. M. Shin, P. N. Kim, M.-S. Lee, and M.-G. Lee
Comparison of mangafodipir trisodium- and ferucarbotran-enhanced MRI for detection and characterization of hepatic metastases in colorectal cancer patients.
Am. J. Roentgenol., April 1, 2006; 186(4): 1059 - 1066.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
H. Onishi, T. Murakami, T. Kim, M. Hori, R. Iannaccone, M. Kuwabara, H. Abe, S. Nakata, K. Osuga, K. Tomoda, et al.
Hepatic Metastases: Detection with Multi-Detector Row CT, SPIO-enhanced MR Imaging, and Both Techniques Combined
Radiology, April 1, 2006; 239(1): 131 - 138.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
Y. K. Kim, H. S. Kwak, C. S. Kim, G. H. Chung, Y. M. Han, and J. M. Lee
Hepatocellular Carcinoma in Patients with Chronic Liver Disease: Comparison of SPIO-enhanced MR Imaging and 16-Detector Row CT
Radiology, December 21, 2005; (2005) 2381042193.
[Abstract] [Full Text]

Home page
 RadiologyHome page
A. M. Lutz, J. K. Willmann, K. Goepfert, B. Marincek, and D. Weishaupt
Hepatocellular Carcinoma in Cirrhosis: Enhancement Patterns at Dynamic Gadolinium- and Superparamagnetic Iron Oxide-enhanced T1-weighted MR Imaging
Radiology, November 1, 2005; 237(2): 520 - 528.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
J. Ward, P. J. Robinson, J. A. Guthrie, S. Downing, D. Wilson, J. P. A. Lodge, K. R. Prasad, G. J. Toogood, and J. I. Wyatt
Liver Metastases in Candidates for Hepatic Resection: Comparison of Helical CT and Gadolinium- and SPIO-enhanced MR Imaging
Radiology, October 1, 2005; 237(1): 170 - 180.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
S. H. Kim, D. Choi, S. H. Kim, J. H. Lim, W. J. Lee, M. J. Kim, H. K. Lim, and S. J. Lee
Ferucarbotran-Enhanced MRI Versus Triple-Phase MDCT for the Preoperative Detection of Hepatocellular Carcinoma
Am. J. Roentgenol., April 1, 2005; 184(4): 1069 - 1076.
[Abstract] [Full Text] [PDF]

Home page
 Med Decis MakingHome page
M. Mazumdar
Group Sequential Design for Comparative Diagnostic Accuracy Studies: Implications and Guidelines for Practitioners
Med Decis Making, October 1, 2004; 24(5): 525 - 533.
[Abstract] [PDF]

Home page
 Am. J. Roentgenol.Home page
K. W. Kim, A. Y. Kim, T. K. Kim, S. H. Park, H. J. Kim, Y. K. Lee, M.-S. Park, H. K. Ha, P. N. Kim, J. C. Kim, et al.
Small (l.e. 2cm) Hepatic Lesions in Colorectal Cancer Patients: Detection and Characterization on Mangafodipir Trisodium-Enhanced MRI
Am. J. Roentgenol., May 1, 2004; 182(5): 1233 - 1240.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
M. Matsuo, M. Kanematsu, K. Itoh, T. Murakami, Y. Maetani, H. Kondo, S. Goshima, N. Kako, H. Hoshi, J. Konishi, et al.
Detection of Malignant Hepatic Tumors with Ferumoxides-Enhanced MRI: Comparison of Five Gradient-Recalled Echo Sequences with Different TEs
Am. J. Roentgenol., January 1, 2004; 182(1): 235 - 242.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
S. Kumano, T. Murakami, T. Kim, M. Hori, A. Okada, T. Sugiura, Y. Noguchi, S. Kawata, K. Tomoda, and H. Nakamura
Using Superparamagnetic Iron Oxide-Enhanced MRI to Differentiate Metastatic Hepatic Tumors and Nonsolid Benign Lesions
Am. J. Roentgenol., November 1, 2003; 181(5): 1335 - 1339.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
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]

Home page
 RadiologyHome page
J. Ward, J. A. Guthrie, D. Wilson, P. Arnold, J. P. Lodge, G. J. Toogood, J. I. Wyatt, and P. J. Robinson
Colorectal Hepatic Metastases: Detection with SPIO-enhanced Breath-hold MR Imaging—Comparison of Optimized Sequences
Radiology, September 1, 2003; 228(3): 709 - 718.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
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,