| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Klinikum rechts der Isar, Technische Universität
München, Ismaninger Str. 22, D-81675 Munich, Germany.
2 Siemens Medical Solutions, Erlangen, Germany.
Received October 9, 2007;
accepted after revision May 21, 2008.
Address correspondence to M. Bruegel
(mbruegel{at}roe.med.tum.de).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Fifty-two patients with extrahepatic primary malignant tumors underwent 1.5-T MRI that included DW EPI and the following variants of T2-weighted TSE techniques: breath-hold fat-suppressed HASTE, breath-hold fat-supressed TSE, respiration-triggered fat-suppressed TSE, breath-hold STIR, and respiration-triggered STIR. Images were reviewed independently by two blinded observers who used a 5-point confidence scale to identify lesions. Results were correlated with surgical and histopathologic findings and follow-up imaging findings. The accuracy of each technique was measured with free-response receiver operating characteristic analysis.
RESULTS. A total of 118 hepatic metastatic lesions (mean diameter,
12.8 mm; range, 3–84 mm) were evaluated. Accuracy values were higher
(p < 0.001) with DW EPI (0.91–0.92) than with the
T2-weighted TSE techniques (0.47–0.67). Imaging with the HASTE sequence
(0.47–0.52) was less accurate (p < 0.05) than imaging with
the breath-hold TSE, breath-hold STIR, respiration-triggered TSE, and
respiration-triggered STIR sequences (0.59–0.67). Sensitivity was higher
(p < 0.001) with DW EPI (0.88–0.91) than with T2-weighted
TSE techniques (0.45–0.62). For small (
10 mm) metastatic lesions
only, the differences in sensitivity between DW EPI (0.85) and T2-weighted TSE
techniques (0.26–0.44) were even more pronounced.
CONCLUSION. DW EPI was more sensitive and more accurate than imaging
with T2-weighted TSE techniques. Because of the black-blood effect on vessels
and low susceptibility to motion artifacts, DW EPI was particularly useful for
the detection of small ( 10 mm) metastatic lesions.
Keywords: detection • diffusion MRI • echo-planar imaging • hepatic metastasis
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Studies [10–13] have shown that diffusion-weighted imaging (DWI) of the liver can be used for characterization of focal hepatic lesions through quantification of diffusion effects with apparent diffusion coefficient (ADC) measurements. Some reports [10, 14, 15] have suggested that owing to high lesion-to-liver contrast, DWI may be highly sensitive for the detection of focal hepatic lesions. One of the most commonly used techniques for DWI of the abdomen is breath-hold single-shot echo-planar imaging (EPI) in combination with parallel acquisition techniques [16–18]. Implementation of parallel acquisition techniques reduces acquisition time, motion artifacts, blurring due to significant T2 decay during the acquisition of relatively long echo trains, and typical EPI artifacts such as chemical shift and susceptibility artifacts [19]. Application of respiratory triggering may further improve the quality of DW EP images because high spatial resolution can be achieved with an adequate signal-to-noise ratio, enhancing ability to detect focal hepatic lesions. The purpose of our study was to assess the value of respiration-triggered DW EPI in the diagnosis of hepatic metastasis and to compare the results with those obtained with five variants of T2-weighted TSE sequences: three breath-hold and two respiration-triggered techniques.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The study group comprised 52 patients (30 men, 22 women; mean age, 56.4 years; age range, 18–75 years). A total of 118 hepatic metastatic lesions were diagnosed in 28 patients (16 patients with synchronous and 12 with metachronous hepatic metastatic lesions). The underlying primary tumors were as follows: four cases of carcinoma of the breast, two cases of bronchial carcinoma, nine cases of colorectal carcinoma, one case of duodenal carcinoma, two cases of gastric carcinoma, one case of hypernephroma, two cases of melanoma, and seven cases of neuroendocrine carcinoma. Nine of the 28 patients with hepatic metastasis had coexistent benign hepatic lesions: 11 hemangiomas and 13 cysts. Twenty patients had solely benign hepatic lesions: 46 hemangiomas and 36 cysts. The other four patients had no evidence of focal hepatic lesions.
Verification of Pathologic Findings
Histopathologic results after biopsy or surgical resection were available
for a total of 34 metastatic lesions and three hemangiomas. In 18 patients,
the diagnosis of metastatic hepatic disease was established with biopsy of at
least one lesion; in six patients, at surgical resection; and in four patients
(all with synchronous hepatic metastatic lesions), with pathologic tracer
uptake of the lesions at PET/CT. Among the 28 patients with metastatic hepatic
disease, 17 patients started systemic chemotherapy, two patients underwent
transarterial chemoembolization, two patients were treated with stereotactic
irradiation, six patients underwent surgical resection with curative attempt,
and one patient declined therapy.
To determine the exact number and location of metastatic and benign lesions in patients who did not undergo surgery, a consensus review was performed by two experienced abdominal radiologists (16 and 21 years of experience in abdominal MRI). The consensus board reviewers were presented with histopathologic and surgical reports, if available. They also were given images obtained with all sequences included in the MRI protocol and serial cross-sectional images of all patients (25 who underwent follow-up MRI, 27 who underwent follow-up CT). Whenever new lesions were found on follow-up images, all initial MR images were carefully reevaluated. If any of the new lesions was retrospectively visible on at least one of the initial images, the lesions were included in the final study group.
For the 84 metastatic lesions without explicit histopathologic verification, diagnoses were confirmed with imaging criteria consisting of characteristic features on unenhanced and dynamic gadolinium-enhanced MR images [20, 21]. In addition, all lesions exhibited interval growth or regression (after the start of chemotherapy) in lesion diameter on serial cross-sectional images (mean follow-up interval, 5.4 months; range, 2–10 months). Diagnosis of the 54 hemangiomas without explicit histopathologic confirmation and of the 49 cysts was based on typical imaging findings on unenhanced and dynamic gadolinium-enhanced MR images [21–23] in addition to stability of all lesions on serial cross-sectional images (mean follow-up interval, 9.1 months; range, 5–16 months). To facilitate accurate lesion identification in the subsequent evaluation process, all lesions were electronically labeled, and the diagnosis, size (determined by largest diameter), and location according to Couinaud segmental anatomy of each lesion were recorded.
MRI
MRI was performed with a 1.5-T system (Magnetom Avanto, Siemens Medical
Solutions) with two anterior six-channel body phased-array coils and two
posterior spinal clusters (three channels each). The MRI protocol included
respiration-triggered DW EPI and five variants of T2-weighted TSE sequences:
breath-hold fat-suppressed HASTE, breath-hold fat-suppressed TSE, breath-hold
STIR, respiration-triggered fat-suppressed TSE, and respiration-triggered
STIR. In addition, dynamic contrast-enhanced 3D gradient-echo sequences
(volumetric interpolated breath-hold examination) were performed for all
patients. Contrast administration was a bolus injection of 0.2 mL/kg of
gadopentetate dimeglumine (Magnevist, Bayer Schering Pharma).
Image parameters for DW EPI and all T2-weighted TSE sequences are presented in Table 1. All images were obtained in the transverse plane with a section thickness of 5 mm. For shortening of the echo-train length, integrated parallel imaging techniques with generalized autocalibrating partially parallel acquisition [24] with a twofold acceleration factor were used. For respiratory triggering, prospective acquisition correction was implemented. The prospective acquisition correction technique interleaved the imaging sequence with a navigator sequence to synchronize the measurement with the patient's breathing cycle and to place the data acquisition period into the end-expiration phase.
|
The DW EPI sequence used in this study was a vendor supplied work-in-progress package. A single-shot EPI readout was preceded by a diffusion-sensitizing block consisting of two 180° radiofrequency pulses and four motion-probing gradient pulses. The gradient factors (b values) and spatial direction of the motion-probing gradients were identical for all sections acquired during one respiratory cycle and were altered only between respiratory cycles. Three mutually perpendicular spatial directions were encoded with three b values: 50, 300, and 600 s/mm2. To increase the signal-to-noise ratio, three averages were performed. Trace images were synthesized for each b value, and an ADC map was calculated from all diffusion weightings and directions.
Image Analysis
Analysis of all MR images was performed with a PACS workstation (Easy
Vision, Philips Healthcare). Patient details were removed, and the sets of
images were analyzed randomly. All MR images were evaluated independently by
two experienced radiologists (9 and 15 years of experience in abdominal MRI)
unaware of other imaging findings, the results of the consensus board review,
and the findings at surgery and histopathologic examination. To minimize bias,
the observers reviewed the images in six sessions with an interval of at least
1 week between sessions.
Each observer recorded the presence of all suspected focal hepatic lesions by placing an electronic label and assigning one of five confidence levels using the following scale: 1, definitely not metastatic (i.e., definitely benign); 2, probably not metastatic (i.e., probably benign); 3, possibly metastatic; 4, probably metastatic; and 5, definitely metastatic. On T2-weighted TSE images, lesion scoring was based on criteria well documented in the literature [20–23]. Lesions were assigned a score of 1 if they were sharply defined with a round or lobulated contour and had homogeneously high signal intensity similar to that of cerebrospinal fluid (simple cyst or hemangioma). Irregular or ill-defined lesions with low to moderately high signal intensity or a heterogeneous signal intensity pattern were regarded as definitely metastatic and were assigned a score of 5. Lesions that did not exactly meet these criteria were given intermediate scores based on the observer's subjective rating.
For lesion evaluation with DW EPI, images with a b value of 50 s/mm2 were used to detect focal abnormalities. Subsequent readings of images with b values of 300 and 600 s/mm2 and region-of-interest measurements within lesions on ADC maps were performed to further classify lesions. Lesions were assigned a score of 1 if they were sharply defined and had an ADC greater than 2.1 x 10–3 mm2/s (hemangioma or cyst). Ill-defined lesions with an ADC less than 1.3 x 10–3 mm2/s were considered definitely metastatic and were assigned a score of 5. For lesions with an ADC in the range of 1.3–2.1 x 10–3 mm2/s, intermediate scores were allocated on the basis of the observer's subjective rating. Because pre viously published data [10–13, 18] showed large discrepancies in ADCs of focal hepatic lesions, the cutoff values were based on our experience with the DW EPI sequence implemented in our imaging protocol [25]. At the time of review, the observers were aware that sensitivity calculations would be based only on lesions assigned scores of 3–5.
Qualitative ratings of the overall image quality with each MRI sequence were recorded by the two observers, and final decisions were made in consensus. Image quality was considered excellent when the border of the liver and intrahepatic vessels were sharply defined, focal hepatic lesions were well visualized, and motion artifacts were absent or negligible; good when the margins of the liver and intrahepatic vessels were moderately well defined, lesion conspicuity was fair, or moderate motion artifacts were present; and poor when substantial motion artifacts were present.
Finally, the study coordinator performed one-to-one comparisons between lesions defined by the consensus board and those assigned scores during each of the two blinded readings. The results for each observer and each MRI sequence were documented, and all additional pseudolesions (e.g., vessels, partial volume effects) that had been allocated scores by the two blinded observers were recorded.
Statistical Analysis
The accuracy of each observer in the diagnosis of hepatic metastasis with
each MRI sequence was measured with free-response receiver operating
characteristic analysis. Calculations were per formed with jackknife
free-response receiver operating characteristic analysis software (JAFROC
version 2.2, D. P. Chakraborty). The sen sitivity of each observer with each
MRI sequence was assessed by regarding all metastatic lesions assigned
confidence levels of 3–5 as true-positive findings. The McNemar test was
used to analyze the statistical significance of any difference in per formance
between MRI sequences for an individual observer. Kappa analysis was used to
determine interobserver variability for the diagnosis of metastasis with each
MRI sequence (
0.40, poor agreement; = 0.41–0.75,
good agreement; 
0.76, excellent agreement). Calculations were
perform ed with SPSS software (version 11.5, SPSS). Two-sided p <
0.05 indicated statistical significance.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Accuracy
The results of the free-response receiver operating characteristic analysis
including the figure-of-merit values (accuracy values) for the diagnosis of
hepatic metastatic lesions for each observer and each MRI sequence are
summarized in Table 2. For both
individual observers, the DW EPI sequence was significantly more accurate than
the T2-weighted TSE sequences (p < 0.001 for all) (Fig.
1A,
1B,
1C,
1D,
1E,
1F). In the comparison of
T2-weighted TSE sequences, none of the minor differences in figure-of-merit
values between breath-hold TSE, breath-hold STIR, respiration-triggered TSE,
and respiration-triggered STIR reached statistical significance (p
> 0.05 for all). However, the HASTE sequence was less accurate, and the
differences in figure-of-merit values compared with all other T2-weighted TSE
techniques were statistically significant for both observers (p <
0.05 for all).
|
|
|
|
|
|
|
Sensitivity
The sensitivity values for the diagnosis of all 118 hepatic metastatic
lesions with each MRI sequence and for each observer are given in
Table 3. Corresponding results
for the 66 metastatic lesions 10 mm or smaller are given in
Table 4. For both observers,
the sensitivity of DW EPI in the diagnosis of metastatic lesions of all sizes
was significantly higher than that of any of the T2-weighted TSE sequences
(p < 0.001 for all). In the comparison of T2-weighted TSE
techniques, none of the differences in sensitivity between breath-hold TSE,
breath-hold STIR, respiration-triggered TSE, and respiration-triggered STIR
imaging reached statistical significance (p > 0.05 for all). For
both observers, the HASTE sequence was significantly less sensitive than the
breath-hold TSE, respiration-triggered TSE, and respiration-triggered STIR
sequences (p < 0.05 for all). For one observer, HASTE images were
significantly less sensitive than breath-hold STIR images (p <
0.005).
|
|
When only small ( 10 mm) metastatic lesions were considered, the
differences in sensitivity between DW EPI and the T2-weighted TSE sequences
were even more pronounced than in the evaluation of larger lesions. Again,
there were no statistically significant differences between breath-hold TSE,
breath-hold STIR, respiration-triggered TSE, and respiration-triggered STIR
sequences (p > 0.05 for all). For both observers, the HASTE
sequence was significantly less sensitive than respiration-triggered TSE and
respiration-triggered STIR sequences (p < 0.05 for all). For one
observer, HASTE images were also significantly less sensitive than breath-hold
STIR images (p < 0.05).
Interobserver agreement for the diagnosis of metastatic lesions of all
sizes was good to excellent for all MRI sequences, the kappa values ranging
from 0.42 for DW EPI to 0.90 for breath-hold STIR imaging. When only small
( 10 mm) metastatic lesions were analyzed, kappa values were generally
lower, but agreement was good to excellent for all sequences.
False-Negative Findings
For the DW EPI sequence, a total of 25 false-negative interpretations (14
and 11 false-negative interpretations by observers 1 and 2, respectively) were
recorded for all patients. In 14 of the 25 cases, metastatic lesions had been
assigned low confidence levels of 1 or 2. In the other 11 cases, the lesions
had not been scored at any confidence level. The size of the unscored
metastatic lesions ranged from 4 to 8 mm.
For the T2-weighted TSE sequences, the total numbers of false-negative
interpretations (combined from both observers) were as follows: 92 with
respiration-triggered STIR, 97 with respiration-triggered TSE, 98 with
breath-hold STIR, 105 with breath-hold TSE, and 126 with HASTE. Most of the
false-negative diagnoses of metastasis had not been scored at any confidence
level (77 with breath-hold STIR, 82 with breath-hold TSE, 85 with
respiration-triggered STIR, 87 with respiration-triggered TSE, and 112 with
HASTE). Most of the unscored metastatic lesions were small ( 10 mm), but
several unscored lesions were larger than 10 mm (eight on breath-hold STIR, 13
on breath-hold TSE, 13 on respiration-triggered TSE, 15 on
respiration-triggered STIR, and 21 on HASTE images) (Fig.
2A,
2B,
2C,
2D,
2E,
2F,
2G).
|
|
|
|
|
|
|
Histopathologically Confirmed Metastasis
Thirty-four histopathologically confirmed metastatic lesions (mean size, 17
mm; range, 4–53 mm) were present in 24 patients. With the DW EPI
sequence, the numbers of correctly diagnosed metastatic lesions were 31 by
observer 1 and 32 by observer 2. Two metastatic lesions in one patient had
been mistaken for benign lesions (assigned a score of 2) by either observer;
in another patient, one metastatic lesion had not been scored by one observer.
For the T2-weighted TSE sequences, the total numbers of metastatic lesions
correctly diagnosed by observers 1 and 2 were as follows: 22 and 21 with
HASTE, 24 and 25 with breath-hold TSE, 25 and 25 with respiration-triggered
TSE, 25 and 26 with breath-hold STIR, and 26 and 26 with respiration-triggered
STIR. At least six metastatic lesions in three patients had not been scored by
either observer with any of the T2-weighted TSE sequences. In addition, at
least two metastatic lesions had been mistaken for benign lesions (assigned a
score of 1 or 2) by either observer with each T2-weighted technique.
Although the sensitivity of DW EPI (0.91 and 0.94 for observers 1 and 2) was higher than that of any of the T2-weighted TSE sequences (maximum 0.76 for observer 2 with breath-hold STIR and for either observer with respiration-triggered STIR), the differences did not reach statistical significance except for the HASTE sequence (0.65 and 0.62 for observers 1 and 2, p < 0.05).
Image Quality
In our study group of 52 patients, the overall quality of DW EPI images
with a b value of 50 s/mm2 was rated excellent in 45 cases and good
in seven cases. For breath-hold TSE and breath-hold STIR images, the ratings
were as follows: excellent in 19 and 22 cases, good in 30 and 27 cases, and
poor in three cases each. With respiration-triggered TSE and
respiration-triggered STIR sequences, image quality was judged excellent in 23
cases each, good in 26 and 25 cases, and poor in three and four cases. The
quality of HASTE images was rated good in all cases.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Use of navigator-controlled respiratory triggering and generalized autocalibrating partially parallel acquisition enabled us to acquire high-quality DW images with b values of 50, 300, and 600s/mm2 during an acquisition time of approximately 5 minutes. The overall quality of DW EP images obtained with a b value of 50 s/mm2 was rated good or excellent in all cases. Occasional limitations in image quality resulted from susceptibility artifacts at the periphery of the liver adjacent to air in the lungs or gastrointestinal tract and from signal intensity alterations in the left lobe of the liver due to cardiac pulsation. Although identical navigator-controlled triggering techniques (prospective acquisition correction) were applied to DW EPI, respiration-triggered TSE, and respiration-triggered STIR sequences, respiratory motion artifacts were less frequently encountered on DW EP images. The latter may be explained by the single-shot EPI readout, which diminishes motion artifacts when respiratory triggering is suboptimal in patients with a fast or irregular breathing cycle.
For the diagnosis of 118 hepatic metastatic lesions, including 66 lesions 1 cm or smaller, the sensitivity of DW EPI was 0.88–0.91. In the study by Nasu et al. [14], who evaluated 40 pathologically proved hepatic metastatic lesions, including 18 lesions 1 cm or smaller, combined reading of unenhanced T1- and T2-weighted images and DW EP images acquired with a b value of 500 s/mm2 resulted in a sensitivity of 0.78–0.88. Although high-b-value images facilitate differentiation of metastatic lesions (which retain high signal intensity because of restricted diffusion of extracellular water molecules in a cell-rich environment) from hemangiomas and cysts (which fade out at increasing b values owing to their high fluid content), we prefer low-b-value images for initial lesion detection. According to our experience and that of other investigators [28], visualization of hepatic lesions on high b-value-images often is hampered by suboptimal signal-to-noise ratios and more pronounced artifacts.
With T2-weighted TSE techniques, sensitivity in the diagnosis of hepatic metastatic lesions has ranged from 0.45 to 0.62. Like Huang et al. [29], who reported that T2-weighted breath-hold fat-suppressed fast-recovery fast spin-echo images and respiration-triggered fat-suppressed fast spin-echo images were comparable for detection and characterization of focal hepatic lesions, we did not observe a significant difference in accuracy between respiration-triggered TSE and breath-hold TSE images. Contrarily, other investigators [8, 30] found respiration-triggered T2-weighted TSE and fast spin-echo techniques more accurate than corresponding breath-hold techniques in the detection of solid hepatic lesions. However, as with all studies comparing MRI pulse sequences, the results strongly depend on the choice of individual MRI parameters.
The use of parallel imaging techniques may have been especially advantageous for our breath-hold TSE sequences, because it limited breath-hold times to 15 seconds, keeping respiratory motion artifacts low. Consequently, the overall quality of breath-hold TSE and breath-hold STIR images was good or excellent in most cases and was comparable with that of the corresponding respiration-triggered TSE images. According to a previous report [31], T2-weighted breath-hold fat-suppressed fast-recovery fast spin-echo images had better lesion clarity and higher overall image quality than did corresponding HASTE images. Similarly, we found breath-hold TSE images (and all other T2-weighted TSE images) more accurate than HASTE images in the diagnosis of hepatic metastasis. The latter were almost free of motion artifacts, but overall image quality suffered from blurring. Comparing the TSE sequences with chemical shift-selective fat suppression (breath-hold TSE and respiration-triggered TSE) and the corresponding inversion recovery techniques (breath-hold STIR and respiration-triggered STIR), we did not observe significant differences in accuracy in the diagnosis of hepatic metastasis. Contrarily, Gaa et al. [32] reported that the conspicuity of hepatic lesions was higher with inversion recovery fast spin-echo images than with fat-suppressed fast spin-echo images. Nearly one half of the lesions evaluated by Gaa et al., however, were nonsolid lesions, which specifically may have benefited from the additive effects of T1 and T2 weighting at inversion recovery imaging.
Most (14 of 25) of the false-negative observations with the DW EPI sequence in our study arose from misclassification of lesions (i.e., metastatic lesions were assigned scores of 1 or 2), whereas only 11 metastatic lesions were completely missed. Fourteen of the 16 false-positive observations, however, were made because hemangiomas had been mistaken for metastatic lesions. At retrospective review, these misclassified hemangiomas were found to have relatively low (< 1.6 x 10–3 mm2/s) ADCs. The difficulties in differentiating hemangiomas from metastatic lesions on the basis of ADC are well-known from previous studies [10–13, 25]. Although the mean ADCs of metastatic lesions and hemangiomas have been found to differ significantly, a certain degree of overlap of ADCs of individual lesions also has been observed. All in all, our results indicate that limitations of the DW EPI sequence used are related predominantly to lesion characterization rather than to lesion detectability.
With T2-weighted TSE sequences, most false-negative diagnoses of metastasis
had not been scored with any confidence level. Although several missed lesions
were identified at retrospective review, initial lesion detectability was
hampered by low lesion-to-liver contrast and the interfering high signal
intensity from intrahepatic vessels. On the other hand, pseudolesions
originating from the signal intensity of intrahepatic vessels were responsible
for several false-positive observations on T2-weighted TSE images. Similar
misinterpretations did not occur with the DW EPI sequence because of the
coincident black-blood effect on vessels from diffusion-sensitizing gradients.
The positive influence of the black-blood effect, particularly on the
detection of small ( 10 mm) hepatic metastatic lesions, has also been
discussed by Coenegrachts et al.
[28].
A potential limitation of our study was the relatively small number of lesions with histopathologic confirmation. However, a thorough review of all MRI sequences with follow-up cross-sectional imaging was performed by two experienced readers, so the probability of misclassification of lesions is low. Although some small lesions might have been missed with all imaging techniques used during the consensus review, the validity of our comparison of MRI sequences should not be essentially affected. According to our results, it is unlikely that such small lesions would have been particularly detected on T2-weighted TSE images while being missed on DW EP images.
In conclusion, respiration-triggered DW EPI was more accurate and more sensitive in the diagnosis of hepatic metastasis than were T2-weighted TSE techniques. Combining DW EPI with respiratory triggering and generalized autocalibrating partially parallel acquisition resulted in high overall image quality and low susceptibility to respiratory motion artifacts. The black-blood effect on vessels also contributed to the high conspicuity of even very small lesions. Our preliminary results indicate that the respiration-triggered DW EPI technique may provide useful information in the preoperative evaluation of patients with suspected hepatic metastasis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  L. Mannelli, S. Kim, C. H. Hajdu, J. S. Babb, T. W. I. Clark, and B. Taouli Assessment of Tumor Necrosis of Hepatocellular Carcinoma After Chemoembolization: Diffusion-Weighted and Contrast-Enhanced MRI With Histopathologic Correlation of the Explanted Liver Am. J. Roentgenol., October 1, 2009; 193(4): 1044 - 1052. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Qayyum Diffusion-weighted Imaging in the Abdomen and Pelvis: Concepts and Applications RadioGraphics, October 1, 2009; 29(6): 1797 - 1810. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
