| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Pediatric Imaging |
1 From the Departments of Radiology (K.A.T., B.A.H., D.K.K.), Pediatrics (S.A.), and Neurology (L.A.S.), Loma Linda University Medical Center, 11234 Anderson St, Loma Linda, CA 92354; and the MRI Institute for Biomedical Research, St Louis, Mo (G.H., E.M.H.). Received February 28, 2002; revision requested April 24; final revision received October 3; accepted October 14. Address correspondence to K.A.T. (e-mail: ktong@ahs.llumc.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Seven young patients with a mean Glasgow Coma Scale score of 7 ± 4 (SD) at admission were imaged a mean of 5 days ± 3 after injury. High-spatial-resolution three-dimensional GRE imaging performed with postprocessing by using a normalized phase mask was compared with conventional GRE MR imaging. The total and mean values of lesion number and apparent hemorrhage volume load determined with both examinations were compared. Mean values were compared by using paired t test analysis. Differences were considered to be significant at P 
.05.
RESULTS: Hemorrhagic lesions were much more visible on SW MR images than on conventional GRE MR images. SW MR imaging depicted 1,038 hemorrhagic DAI lesions with an apparent total hemorrhage volume of 57,946 mm3. GRE MR imaging depicted 162 lesions with an apparent total hemorrhage volume of 28,893 mm3. SW MR imaging depicted a significantly higher mean number of lesions in all patients than did GRE MR imaging, according to results of visual (P = .004) and computer (P = .004) counting analyses. The mean hemorrhage volume load for all patients also was significantly greater (P = .014) by using SW MR imaging according to computer analysis. SW MR imaging appeared to depict much smaller hemorrhagic lesions than GRE MR imaging. The majority (59%) of individual hemorrhagic DAI lesions seen on SW MR images were small in area (<10 mm2), whereas the majority (43%) of lesions seen on GRE images were larger in area (10–20 mm2).
CONCLUSION: SW MR imaging depicts significantly more small hemorrhagic lesions than does conventional GRE MR imaging and therefore has the potential to improve diagnosis of DAI.
© RSNA, 2003
Index terms: Brain, hemorrhage, 13.434, 14.434 • Brain, injuries, 13.434, 14.434 • Brain, MR, 13.121411, 13.121412, 13.121413, 13.121416, 14.121411, 14.121412, 14.121413, 14.121416 • Magnetic resonance (MR), technology, 13.121411, 13.121412, 13.121413, 13.121416, 14.121411, 14.121412, 14.121413, 14.121416
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The ability to detect DAI by using imaging, whether the lesions are hemorrhagic or nonhemorrhagic, has substantially improved with the advent of magnetic resonance (MR) imaging (2,7,8). MR imaging has been helpful in defining patterns of injury in adults with DAI—depicting involvement predominantly in the frontal white matter, corpus callosum, brainstem, and diencephalon (2). Whether the same pattern of injury occurs in children remains uncertain.
The results of studies involving adults and children have suggested that the presence of hemorrhage in DAI lesions may portend a poor prognosis (9) compared with the absence of hemorrhage in lesions (10). The development of T2*-weighted gradient-recalled-echo (GRE) MR imaging has facilitated improved detection of intracranial hemorrhages (11,12). However, the results of neuropathologic investigations have established that DAI lesions can be detected more frequently with microscopic analysis than with conventional radiologic imaging (1).
Haacke and colleagues have designed a high-spatial-resolution three-dimensional fast low-angle shot MR imaging technique, which was first described in a study performed by Reichenbach et al (13), that is extremely sensitive to susceptibility changes and can be performed with conventional MR imaging units. This sequence was originally designed for MR venography (13–15) and involves use of the paramagnetic property of intravenous deoxyhemoglobin. At our institution, we have found that this technique is also very sensitive in the detection of extravascular blood products. Thus, the purpose of our study was to compare this high-spatial-resolution susceptibility-weighted (SW) MR imaging technique with a conventional GRE technique for the detection of DAI.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
A neurosurgeon or neurologist examined all patients prior to the MR imaging evaluation. The neurologic outcomes at discharge from the hospital were as follows: Three patients had mild and four patients moderate disability, as classified by using Glasgow Outcome Scale scores (16), which were determined by the neurologists (S.A., L.A.S.). These patients were also part of a larger ongoing study to examine the role of MR spectroscopy in predicting the outcomes of children and adolescents with various forms of acute central nervous system insult. The MR spectroscopy data are not included in this article, however, because they are the topic of a separate study. For the limited and anonymous review of these patients’ data for this study, we were not required to have formal approval or informed patient consent according to the institutional review board of Loma Linda University Medical Center.
MR Imaging
When medically stable, the patients were transported to the MR imaging unit (Magnetom Vision; Siemens Medical Solutions, Iselin, NJ). They were monitored during the MR imaging examinations by intensive care unit and radiology department personnel, as previously reported (17). All MR imaging examinations were performed by using a circularly polarized head coil and a conventional 1.5-T whole-body MR imaging system (Magnetom Vision).
Routine MR imaging included transverse dual spin-echo (repetition time, 2,500 msec; echo time, 22/80 msec; one signal acquired, 5-mm-thick sections), sagittal T1-weighted spin-echo (550/22 [repetition time msec/echo time msec], four signals acquired, 5-mm-thick sections), transverse fluid-attenuated inversion-recovery (9,000/110, one signal acquired, 5-mm-thick sections), and transverse two-dimensional GRE (fast imaging with steady-state precession, 500/18, 15° flip angle, bandwidth of 78 Hz per pixel, two signals acquired, 4–5-mm-thick sections) sequences. The findings on these MR images were used for comparison with the information available on the SW MR images.
SW MR imaging consisted of a strongly SW low-bandwidth (78 Hz/pixel) three-dimensional fast low-angle shot sequence (57/40, 20° flip angle) that was first-order flow compensated in all three orthogonal directions. Thirty-two 2-mm partitions were acquired by using a rectangular field of view (5/8 of 256 mm) and a matrix size of 160 x 512, which resulted in a voxel size of 1.0 x 0.5 x 2.0 mm3. With the SW MR imaging sequence, the majority of the cerebral hemispheres and the posterior fossa were imaged in an acquisition time of approximately 9.5 minutes.
SW MR images were created by using the magnitude and phase images, as previously explained in the Reichenbach et al (13) article. The phase image was low pass filtered (by using a 32 x 128 exclusion of low-spatial-frequency information) to remove much of the brain’s low-spatial-frequency background static-field variation. A phase mask was created by setting all positive-phase values (between 0° and 180°) to unity and normalizing the negative-phase values ranging from 0° to -180° to a gray scale of values ranging linearly from 1 to 0, respectively. This normalized phase mask was multiplied four times against the original magnitude image and yielded images that maximized the negative signal intensities of the regions containing deoxygenated blood and increased the contrast between regions containing deoxygenated blood and the surrounding tissue. Finally, a minimum intensity projection over two sections was performed to display the processed data by using contiguous 4-mm-thick sections in the transverse plane. The underlying contrast mechanism is associated with the magnetic susceptibility difference between oxygenated and deoxygenated hemoglobin, which leads to a phase difference between regions containing deoxygenated blood and the surrounding tissues and concomitantly to signal cancellation.
This technique was originally used to enhance the visibility of venous structures containing deoxyhemoglobin and was therefore previously described as high-spatial-resolution blood oxygen level–dependent venography (14,15). In the setting of traumatic brain injury, SW MR imaging is primarily focused on the phase dispersion caused by the presence of extravascular deoxyhemoglobin and methemoglobin rather than by the presence of intravenous deoxyhemoglobin.
A neuroradiologist (K.A.T.) and a pediatric neurologist (S.A.) reviewed the MR images in consensus on a workstation (DS3000, Impax version 4.1; Agfa-Gevaert, Kontich, Belgium) that is routinely used to view images. Hemorrhagic lesions were defined as hypointense foci that were not compatible with vascular, bone, or artifactual structures on conventional GRE and SW MR images. If there was doubt as to the etiology of any foci, the lesions were not considered to be hemorrhagic lesions. Hemorrhagic shearing lesions of the entire cranium were counted individually with both the GRE and the SW MR imaging sequence. Nonhemorrhagic shearing lesions were defined as hyperintense lesions that were visible primarily on T2-weighted or fluid-attenuated inversion-recovery MR images. The hemorrhagic lesions depicted by GRE and SW MR imaging were also categorized according to individual brain regions: frontal white matter; frontal gray matter; corpus callosum; parietotemporo-occipital white matter; parietotemporo-occipital gray matter; basal ganglia, internal capsule, and thalami; and brainstem and cerebellum.
Because hemorrhagic lesions were variable in shape, the MR images were then analyzed (by K.A.T.) with a computer software program (Image Pro Plus; Media Cybernetics, Silver Springs, Md) to automatically trace the outline of the lesions by using selected minimum signal intensity threshold levels. This program then automatically counted and calculated the areas of the lesions. The area of each lesion was then multiplied by the effective section thickness of the image (4 or 5 mm for GRE images, 4 mm for SW MR images) for determination of the total volume of the hemorrhagic DAI load per patient. The area range of individual lesions was also divided into 10-mm2 increments and plotted against frequency distribution. The actual sizes of the hemorrhages were probably exaggerated owing to blooming effects. However, because our intention was to compare the visibility differences between the two MR imaging examinations, the approximated areas and volumes are reported as "apparent" values.
Data Analysis
The number of lesions, as counted by using visual and computer methods, in each of the seven patients was determined by using both the GRE sequence and the SW sequence. In addition, the apparent hemorrhage volume in each patient, which was calculated by using the computer method, was determined for both techniques. The total values for all seven patients (ie, number of lesions and lesion volume), as well as the mean differences in total values between the two imaging methods, were compared. The mean differences in total values were compared by using paired t test analysis. Differences were considered significant at P .05.
We dichotomized the patient groups into severe (Glasgow Coma Scale [GCS] score 8) and nonsevere (GCS score > 8) injury categories according to standard definitions used by trauma specialists and groups such as the American Academy of Neurological Surgeons and the Brain Injury Foundation. By using the computer-generated GRE and SW MR imaging measurements, we determined the mean lesion number and mean hemorrhage load for the patient group with severe initial traumatic brain injury—that is, a GCS score of less than or equal to 8 and compared them with the mean lesion number and mean hemorrhage load for the patient group with an initial GCS score higher than 8. The mean values were compared by using Student t test analysis. Differences were considered significant at P .05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
8) had a greater number of lesions and a greater hemorrhage load according to both GRE and SW MR imaging data (Fig 6); however, the difference was not statistically significant in this small patient sample.
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Most blood products, including deoxyhemoglobin, methemoglobin, and hemosiderin, are paramagnetic and thus make it possible to exploit susceptibility effects and make hemorrhages visible on MR images that accentuate signal intensity loss from rapid spin dephasing. In the acute and early subacute phases, these effects largely occur from deoxyhemoglobin and methemoglobin forms of hemoglobin. As spins encounter different fields, they precess at different rates and cause signal intensity loss on T2*-weighted (ie, GRE) images. Sensitivity to the magnetic susceptibility effects of hemorrhage increases as one progresses from fast spin-echo to routine spin-echo to GRE techniques, from T1 to T2 to T2* weighting, from short to long echo times, and from lower to higher magnetic field strengths (18).
The development of the SW MR imaging technique has further improved the ability to detect hemorrhage, as compared with the degree of ability to detect hemorrhage with conventional GRE MR sequences. There are some difficulties in directly comparing the two MR imaging methods due to the inherent unique properties of SW MR image postprocessing. However, several factors should be noted: The parameters used for conventional GRE MR imaging often differ between institutions, and it is possible that a longer echo time or a different flip angle would improve the visibility of susceptibility foci, although probably at a cost of an overall decrease in signal intensity. Differences in section thickness also might affect comparisons of the two MR imaging techniques, although the postprocessed SW MR images obtained in this study were effectively 4 mm thick (after minimum signal intensity projection reconstruction over two sections), as compared with the 4- or 5-mm section thickness used for the GRE sequence. Thinner sections could be acquired in the GRE sequence at a time cost, resulting in an acquisition time similar to that with the SW MR imaging technique. Although our conventional GRE parameters could be optimized further, the described comparison was meant to be applicable to the general MR imaging community.
The most important disadvantage of the SW MR imaging technique is probably the substantially longer acquisition time of 9.5 minutes, as compared with the 3.5 minutes required for GRE MR image acquisition. This longer time could potentially increase the risk of motion artifact. However, our patient population was sedated for the entire MR imaging examination (which also included MR spectroscopy). It is likely that most patients who undergo MR imaging early after substantial head injury require some form of sedation for the length of the examination. The acquisition time required for the sequence, although not substantially greater than that required for many high-spatial-resolution sequences performed at various academic centers, could also be problematic in terms of throughput in the general MR imaging community. With this in mind, there are current plans to decrease the acquisition time by incorporating echo-planar MR imaging techniques.
The presence of intracranial susceptibility artifacts related to bone structures, extracranial debris, or devices occasionally interfered with the visualization of parenchymal hemorrhage, but usually these artifacts were not difficult to distinguish from hemorrhage. The artifacts were in typical locations (eg, anterior skull base or inferior temporal regions), or they sometimes appeared as amorphous regions with less hypointensity than the hemorrhagic lesions, which were typically smaller, more well-defined, and markedly hypointense. Blooming artifact of the calvarium did cause problems in visualizing subtle hypointense extraaxial hemorrhage, which was difficult to delineate from bone.
Isointense or hyperintense subdural or epidural hemorrhages were easily recognized. Mild acute subarachnoid hemorrhage, which had a hypointense linear appearance within the sulci, could also be difficult to distinguish from normal veins, unless the brain surface was displaced from the calvarium by other extraaxial fluid collections. The presence of extraaxial hemorrhage was infrequent in this patient population and not specifically evaluated.
Although there were no age-matched control subjects in this study, anecdotal experience with SW MR imaging in pediatric or young adult patients without a history of trauma has not involved the encountering of hypointense foci similar to those observed in the posttraumatic population described herein. This may not be the case in older adults, because we have sporadically seen small hypointense foci that may have been related to microangiopathy. Theoretically, it is possible that some small hypointense foci may represent microinfarctions associated with deoxygenated venous blood. However, intravascular deoxyhemoglobin would probably follow a linear pattern that is consistent with vascular architecture rather than punctate discontinuous foci. We have observed increased linear structures conforming to small veins at the margins of acute infarctions in other patients who were examined with SW MR imaging. These structures are thought to be caused by increased oxygen extraction in the penumbra that results in higher levels of intravenous deoxyhemoglobin.
In this study we estimated the volume of hemorrhagic lesions, assuming that the volumes of hemorrhage were proportional to the apparent volumes of the blooming artifact known to occur with GRE MR imaging. However, it is known that the size of susceptibility artifact created by the presence of hemorrhage is determined by two competing effects that are controlled by imaging parameters. For a given spatial resolution and increasing echo time, the signal intensity loss increases and begins to extend beyond the source of the signal intensity loss itself (ie, blooming). Counterbalancing this blooming effect is the reduction in signal intensity loss when the echo time is held fixed and the spatial resolution is increased (ie, the voxel size decreases), leading to less phase dispersion across the voxel and less signal intensity loss. We cannot presume that an SW MR imaging examination involving the use of a 40-msec echo time will have a larger blooming effect than a conventional two-dimensional GRE MR imaging examination that involves the use of an 18-msec echo time, because SW MR imaging is initially performed with a thinner section (2 mm, as compared with 4 or 5 mm) and a higher in-plane spatial resolution (0.5 mm vs 1.0 mm in the read direction).
Additionally, we may be seeing smaller punctate hemorrhages depicted by SW MR imaging owing to another phenomenon that is related to signal intensity loss in small vessels or other small objects. For a given spatial resolution, we have found that for venous blood the optimal signal intensity contrast occurs when the vein occupies roughly one quarter of a voxel. For blood products this could be a much smaller fraction, which means that an object much smaller than a voxel could have a dramatic appearance, even in a single voxel. This may explain why a greater number of small objects are visible with this higher spatial resolution technique (19).
The numbers and volumes of hemorrhages may have been overestimated given that the connectivity of hemorrhages across several adjacent images was not accounted for. This could be further evaluated by using a computer software program that could help determine whether lesions on adjacent sections comprise one large lesion. This type of software could also help distinguish small vessels from artifacts created by hemorrhage, although in the present study most vessels could be distinguished visually from punctate hemorrhages. However, for the purposes of this study, the degree of error would be similar for both the GRE and the SW MR images, because the same method of counting and measuring was used with both examinations.
According to pathologic study results (20), about 50% of DAI lesions are located in the deep white matter or the corticomedullary junction (ie, gray matter–white matter interface) of the frontal and temporal lobes, and in most series multiple lesions varying in size from 5 to 15 mm are detected. The corpus callosum, and particularly the splenium of the corpus callosum, is the second most commonly affected area. DAI can also be seen in the corona radiata and to a lesser extent in the rostral brain stem and cerebellum (20). Although the locations of DAI lesions in the current study were similar to the locations of lesions previously reported in the radiologic literature (2,21–23), we observed a greater number and volume of hemorrhagic lesions in the brain stem and cerebellum in this sample population. There are probably several factors involved in this phenomenon. MR imaging has greater capability to depict the posterior fossa structures without the beam hardening artifact of computed tomography.
Alternatively, we may have examined the data on a unique population of patients—that is, those with a greater occurrence of brainstem and cerebellar lesions owing to the mechanism of injury. Finally, the SW MR imaging sequence may inherently have greater capability to depict hemorrhages in the posterior fossa owing to the postprocessing technique, which involves filtering algorithms to reduce any bone artifact.
The majority of DAI lesions are reported to be nonhemorrhagic in the pathologic literature. In an earlier radiologic study by Gentry and colleagues (2), about 20% of lesions showed evidence of bleeding at MR imaging. In our study, the majority of lesions detected were hemorrhagic. This is probably because of the improved sensitivity of SW MR imaging for the detection of small hemorrhages that previously could not be visualized. We believe that it is interesting that some of the nonhemorrhagic lesions also were visible as hyperintense areas on the SW MR images owing to underlying T2 effects contributed by the magnitude information. Further studies with pathologic correlation are warranted to determine the true percentage of hemorrhagic versus nonhemorrhagic lesions and thus the accuracy of SW MR imaging in distinguishing between the two.
In conclusion, SW MR imaging depicts significantly more hemorrhages than conventional GRE MR imaging, regardless of whether the lesions are detected by using a visual (P = .004) or a computer (P = .004) counting method, and can help improve the diagnosis of DAI. In the present study, SW MR imaging, as compared with conventional GRE imaging, depicted four to six times more lesions and a twofold greater apparent volume of hemorrhagic DAI lesions. A greater number and volume of hemorrhagic lesions were observed in the brain stem and cerebellum in this sample population compared with the amount of hemorrhage described in previous reports. An association between severity of hemorrhagic DAI and initial GCS score was suggested but not statistically significant in this small patient cohort. With a larger series, it may be possible to identify a statistically significant correlation between extent of hemorrhage and both GCS score and clinical outcome. Our findings suggest that SW MR imaging yields additional neurologic imaging information that can improve the evaluation, treatment, and management of patients with traumatic brain injury and suspected DAI.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, K.A.T.; study concepts, D.K.K., E.M.H., G.H.; study design, K.A.T., S.A.; literature research, K.A.T., S.A., L.A.S.; clinical studies, K.A.T., S.A., L.A.S.; data acquisition, K.A.T., S.A., G.H.; data analysis/interpretation, K.A.T., S.A., B.A.H.; statistical analysis, K.A.T., S.A.; manuscript preparation and definition of intellectual content, K.A.T., S.A., B.A.H.; manuscript editing, K.A.T.; manuscript revision/review and final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Kakeda, Y. Korogi, K. Kamada, N. Ohnari, J. Moriya, T. Sato, M. Kitajima, H. Hasnine, and N. Hirata Signal Intensity of the Motor Cortex on Phase-Weighted Imaging at 3T AJNR Am. J. Neuroradiol., June 1, 2008; 29(6): 1171 - 1175. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Deistung, A. Rauscher, J. Sedlacik, S. Witoszynskyj, and J. R. Reichenbach Informatics in Radiology: GUIBOLD: A Graphical User Interface for Image Reconstruction and Data Analysis in Susceptibility-weighted MR Imaging RadioGraphics, May 1, 2008; 28(3): 639 - 651. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Ding, Q. Jiang, L. Li, L. Zhang, Z. G. Zhang, K. A. Ledbetter, L. Gollapalli, S. Panda, Q. Li, J. R. Ewing, et al. Angiogenesis Detected After Embolic Stroke in Rat Brain Using Magnetic Resonance T2*WI Stroke, May 1, 2008; 39(5): 1563 - 1568. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Sidaros, A. W. Engberg, K. Sidaros, M. G. Liptrot, M. Herning, P. Petersen, O. B. Paulson, T. L. Jernigan, and E. Rostrup Diffusion tensor imaging during recovery from severe traumatic brain injury and relation to clinical outcome: a longitudinal study Brain, February 1, 2008; 131(2): 559 - 572. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K.A. Tong, S. Ashwal, A. Obenaus, J.P. Nickerson, D. Kido, and E.M. Haacke Susceptibility-Weighted MR Imaging: A Review of Clinical Applications in Children AJNR Am. J. Neuroradiol., January 1, 2008; 29(1): 9 - 17. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.L. Harder, K.M. Hopp, H. Ward, H. Neglio, J. Gitlin, and D. Kido Mineralization of the Deep Gray Matter with Age: A Retrospective Review with Susceptibility-Weighted MR Imaging AJNR Am. J. Neuroradiol., January 1, 2008; 29(1): 176 - 183. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E.M. Haacke, Z.S. DelProposto, S. Chaturvedi, V. Sehgal, M. Tenzer, J. Neelavalli, and D. Kido Imaging Cerebral Amyloid Angiopathy with Susceptibility-Weighted Imaging AJNR Am. J. Neuroradiol., February 1, 2007; 28(2): 316 - 317. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. H. Taber, D. L. Warden, and R. A. Hurley Blast-Related Traumatic Brain Injury: What Is Known? J Neuropsychiatry Clin Neurosci, May 1, 2006; 18(2): 141 - 145. [Full Text] [PDF]
|
|
|
|
|
|
E. Giugni, U. Sabatini, G. E. Hagberg, R. Formisano, and A. Castriota-Scanderbeg Fast Detection of Diffuse Axonal Damage in Severe Traumatic Brain Injury: Comparison of Gradient-Recalled Echo and Turbo Proton Echo-Planar Spectroscopic Imaging MRI Sequences AJNR Am. J. Neuroradiol., May 1, 2005; 26(5): 1140 - 1148. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. A. Holshouser, K. A. Tong, and S. Ashwal Proton MR Spectroscopic Imaging Depicts Diffuse Axonal Injury in Children with Traumatic Brain Injury AJNR Am. J. Neuroradiol., May 1, 2005; 26(5): 1276 - 1285. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Wintermark, G. van Melle, P. Schnyder, J.-P. Revelly, F. Porchet, L. Regli, R. Meuli, P. Maeder, and R. Chiolero Admission Perfusion CT: Prognostic Value in Patients with Severe Head Trauma Radiology, July 1, 2004; 232(1): 211 - 220. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. A. Hurley, J. C. McGowan, K. Arfanakis, and K. H. Taber Traumatic Axonal Injury: Novel Insights Into Evolution and Identification J Neuropsychiatry Clin Neurosci, February 1, 2004; 16(1): 1 - 7. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
