| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Neuroradiology, University Hospital Mannheim, University of
Heidelberg, Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany.
2 Institute of Pathology, University of Heidelberg, Heidelberg, Germany.
Received May 5, 2007;
accepted after revision July 23, 2007.
WEB This is a Web exclusive article.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. From our local database, 38,000 cranial MRI scans were searched for intracranial lipomas. If available, TOF images of identified lipomas were analyzed. In addition, in vitro MRI of excised cadaveric lipomas and other fatty specimens were examined for characteristics on TOF images and the presence of chemical shift artifacts.
RESULTS. Seventeen intracranial lipomas (0.045%) were identified. Out-of-phase TOF source images available in 12 identified lipomas showed a continuous dark peripheral fringe and a hyperintense center. In vitro out-of-phase chemical shift images of excised cadaveric lipomas revealed the same consistent fringe artifact, known as "india ink," independent of size, shape, surrounding fibrous capsule, or texture of the fatty specimen. In contrast, in-phase chemical shift artifact was variable.
CONCLUSION. Small intracranial lipomas close to a cerebral artery are hyperintense on TOF MR images and could be mistaken for partially thrombosed aneurysms and associated flow-related artifact. A defining characteristic of lipomas on TOF source images results from the out-of-phase india ink artifact. This dark fringe in the periphery of the lesions is characteristic and helps avoid potential diagnostic pitfalls.
Keywords: intracranial lipoma • lipoma • neuroradiology • time-of-flight MR angiography
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Digital subtraction angiography (DSA) is the gold standard of diagnosing aneurysms; however, for screening purposes, MR angiography (MRA) is frequently used as a noninvasive, radiation-free diagnostic option [1, 5, 6]. Nonetheless, MRA and time-of-flight (TOF) angiography, in particular, are prone to artifacts that significantly limit reliable identification of aneurysms smaller than 3 mm [7-10].
Diagnostic pitfalls during evaluation of 3D TOF MRA may arise from phenomena involving high-signal structures mimicking vascular abnormalities on TOF source images or maximum-intensity-projectio n (MIP) reconstructions. Possible phenomena include subacute thrombus, complex flow artifacts, and high-signal structures such as lipomas [11-13].
Intracranial lipomas are rare and mainly asymptomatic lesions that generally do not require therapy [14]. They have been attributed to congenital malformation of the primitive meninx during development of the subarachnoid space and occur in favored locations including the interhemispheric fissure; cerebellopontine angle; and quadrigeminal, suprasellar, interpeduncular, and sylvian cisterns [15]. Characteristic findings based on very low attenuation values on CT and homogeneous high signal intensity on T1-weighted MR images usually suffice for definite diagnosis [16]. Another hallmark of lipomas on MRI results from edge artifacts at the lipid-water interface attributed to in-phase chemical shift artifact and out-of-phase india ink artifact (also known as chemical shift artifact of the second kind) [17, 18].
Despite the rather unambiguous imaging characteristics of large intracranial lipomas, it is conceivable that occasionally high-signal-intensity lesions of few millimeters close to a vessel may be more difficult to differentiate from small aneurysms. Referral of three cases with a false-positive diagnosis of saccular aneurysm of the vertebrobasilar system simulated by small lipomas on TOF MRA prompted us to retrospectively analyze the TOF images of all intracranial lipomas seen in our institution during the past 7 years. The findings were complemented by the analysis of in vitro MR images of excised lipomas. The purpose of the study was to investigate whether chemical shift artifact on TOF images is a constant characteristic that can be used to reliably differentiate small intracranial lipomas close to a vessel from small aneurysms, thus helping to avoid the diagnostic pitfall as presented.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
On independent occasions, three patients were referred to our institution from outside imaging facilities for DSA of suspected aneurysms of the vertebrobasilar system based on findings from TOF MRA. Because the initial findings were inconclusive, additional cross-sectional imaging was performed at our institution for differential diagnosis of aneurysm.
MRI was performed on a 1.5-T unit (Magnetom Sonata or Avanto, Siemens Medical Solutions), including TOF MRA (TR/TE, 37/7; flip angle, 25°; rectangular field of view, 17 x 23 cm; matrix, 384 x 187; slice thickness, 0.8 mm; 4 slabs) and standard spin-echo axial and sagittal T1-weighted imaging with and without fat suppression and gadolinium-based contrast medium (560/17; matrix, 256 x 173; sagittal slice thickness, 3.0 mm; axial slice thickness, 5.0 mm). Cranial CT scans were obtained using a Volume Zoom scanner (Siemens) with the slice thickness set to 4 and 5 mm for the posterior fossa and cerebrum scans, respectively (120 kV, 350 or 400 mAs). Two patients were examined by DSA (Axiom Artis Ax FA/FB, biplanar, Siemens).
For retrospective analysis, images were evaluated pertaining to a local database query on all reported lipomas of cranial MRI scans from January 2000 to March 2007. A total of 38,000 cranial MRI scans were searched for the keyword "lipoma" in the final report. All examinations included standard axial T1-weighted spin-echo (TR range/TE range, 430-540/10-12; slice thickness, 5 mm), FLAIR (8,000-8,500/100-132; slice thickness, 5 mm), and T2-weighted turbo spin-echo (3,900-5,420/100-102; slice thickness, 5 mm) sequences. Images were analyzed by two independent neuroradiologists with 3 and 15 years' experience in interpreting cranial MRI examinations.
The following inclusion criteria were applied to confirm the reported diagnosis of lipoma: first, high signal intensity of the lesion on T1-weighted images matching the signal intensity of subcutaneous fat and T2-weighted isointensity to fat; and, second, signal loss of the lesion on additional T1-weighted fat-suppressed images or hypodensity of -20 H or less on CT images. Identified cranial lipomas were characterized according to location; size; concomitant anomalies; and, if available, TOF appearance.
For analysis of edge artifacts on TOF images, in vitro MRI was performed on excised cadaveric lipomas in isotonic saline solution; the lipomas were between 5 and 40 mm in diameter. The saline solution was titrated with gadolinium to obtain a ratio of signal intensities of the lipoma specimen to the surrounding solution of 8:1 on T1-weighted spin-echo images (TR/TE, 500/10; slice thickness, 5 mm; matrix, 256 x 224) to approximate the signal ratio of intracranial lipomas to adjacent liquor. Analogously, the solution was titrated to a signal ratio of 3:1 to simulate the signal of lipomas adjacent to brain tissue. TOF images were acquired with standard parameters (TR, 37 ms; flip angle, 25°; field of view, 17 x 23 cm; matrix, 384 x 187; slice thickness, 0.8 mm); out-of-phase (TE, 7 ms) and in-phase (TE, 5 ms) TOF images were obtained. Additional standard sequences included fat-suppressed T1-weighted spin-echo, T2-weighted turbo spin-echo, and FLAIR, as specified earlier. To compare the behavior of edge artifacts of excised lipomas with regular fatty tissue, samples from appendices epiploicae of the colon were imaged likewise. To examine whether a fibrous capsule alters the appearance of edge artifact, two specimens were excised from a large block of lipoma and therefore lacked a fibrous capsule during imaging.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
|
|
A search of 38,000 cranial MRI scans obtained between 2000 and 2007 applying the criteria as specified revealed 17 intracranial lipomas including the three cases described earlier. Detailed characteristics are given in Table 1. The sites of occurrence in order of prevalence were interhemispheric (n =5, 29%) and suprasellar or interpeduncular (n = 4, 24%), followed by quadrigeminal cistern (n = 3, 18%), epitentorial or occipital (n = 2, 12%), cerebellopontine angle (n =1, 6%), chiasmatic cistern (n = 1, 6%), and temporal pole (n = 1, 6%). The median size was 4 mm excluding two large callosal lipomas of 48 x 8 and 57 x 11 mm.
In 12 cases, out-of-phase TOF source images were available. In 11 of these,
the lesions presented with a uniform appearance composed of a small continuous
fringe of very low signal and a center of moderate to high signal intensity.
In one case, a very small lesion (
1.7 mm in diameter; case 14,
Table 1) presented as a spot of
signal void on the TOF image comparable to the intensity of fringe artifact.
Fat suppression in the center of the lesions was incomplete, thus resulting in
signal intensities similar to those of blood flow or thrombus.
In vitro MRI of spherical encapsulated lipomas showed a consistent dark fringe artifact on out-of-phase TOF images (Fig. 4A). The fatty samples from appendices epiploicae (Fig. 4B) and the lipoma specimens missing a fibrous capsule exhibited the same type of dark fringe. This specific artifact, known as "india ink" artifact, was not present on the analogous in-phase TOF images, so only chemical shift artifact remained (Figs. 4C and 4D). The intensity of in-phase chemical shift artifacts was variable, depending on the shape of the excised specimen and the granularity of the border, and was difficult to discern for specimens 5 mm or smaller. Neither india ink nor chemical shift artifacts were affected by the varying signal intensity of the surrounding saline solution.
|
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Small, inherently T1 hyperintense structures in the vicinity of the vertebrobasilar system may pose a diagnostic pitfall because T1 hyperintensity is caused not only by numerous possible lesions, including lipomas and aneurysms, but also by normal anatomy, such as the posterior lobe of the pituitary gland [12, 13]. Common localizations of intracranial lipomas include the basal subarachnoid spaces as a result of maldifferentiated meninx [16]. Although cross-sectional imaging of lipomas is generally very characteristic, lesions smaller than 5 mm localized close to a vertebrobasilar artery may be misinterpreted as aneurysms on TOF MRA if additional T1 fat-suppressed series or CT images are not available for definite diagnosis.
The cases of our study document that small intracranial lipomas have a constant appearance, defined by a consistently uniform and unbroken dark boundary, on the source images of TOF MR angiograms. The in vitro scans showed that the hypointense rim and hyperintense center are exclusively caused by the fatty composition of the specimen independent of change in structure. The "boundary effect" is caused by fat-water phase cancellation in the out-of-phase (TE, 7 ms at 1.5 T) TOF MRA sequence, well known from abdominal imaging as india ink artifact [18]. The value of this edge artifact for differential diagnosis has been emphasized in a recent study in which investigators found that india ink was indicative of all 23 studied angiomyolipomas [23].
Small aneurysms are unlikely to exhibit this characteristic continuous dark edge for several reasons. First, the imaging effect of phase cancellation is physically impossible for aneurysms. Second, a theoretically possible signal void due to calcifications, usually observed only in large aneurysms, would not present this uniformly regular and unbroken boundary.
Edge artifact by in-phase chemical shift alone is not suitable for definite distinction of lipoma from small aneurysm due to inconsistency and variability. The inconsistency of this artifact was shown on the in vitro TOF MR images and primarily varied with the shape of the specimen (Figs. 4C and 4D). In-phase chemical shift results from positional misregistration of the fat signal along the frequency-encoding axis at the lipid-water interface because the precession frequency of water protons is higher than that of lipid protons [24]. Truwit and Barkovich [15] saw chemical shift in only 29 of 42 lipoma MR studies. In-phase chemical shift artifact of fatty lesions depends not only on the orientation of the lipid-water interface and plane of section [17], in particular, but also on field strength [24] and receiver bandwidth [25] and may be entirely absent.
The prevalence of intracranial lipomas detected on cranial MRI in the patient population of our hospital was 0.045% (17/38,000), which is slightly below the range reported in the literature (0.06-0.30%) [26]. The observed median size of 4 mm seems less than most case reports in the literature, whereas sites of predilection are in accordance with reported data [14-16]. Evidently, intracranial lipomas are rare; however, incidental findings are subject to change with rapidly increasing utilization of MRI [27], which makes the reported differential pitfall between small intracranial aneurysms and lipomas more likely.
In conclusion, in some instances, the T1 hyperintensity of a small intracranial lipoma close to a vessel may be difficult to differentiate from aneurysms smaller than 5 mm on TOF MRA. The out-of-phase india ink artifact, a dark fringe surrounding the lesion that is consistently uniform and unbroken, is a characteristic feature of lipomas on TOF source images. In the presence of this artifact, other defining imaging techniques (e.g., T1 fat-suppressed imaging or CT) should be performed before invasive DSA to exclude small aneurysms.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
