| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Neuroradiology, University Hospital Basel, Basel 4031,
Switzerland.
2 Department of Neuropathology, University Hospital Basel, Basel 4031,
Switzerland.
3 Guerbet, Zürich, Switzerland.
4 Department of Neurosurgery, University Hospital Basel, Basel 4031,
Switzerland.
Received August 15, 2004;
accepted after revision December 20, 2004.
Address correspondence to C. A. Taschner.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Nine patients with brain tumors were imaged before and 24 hr after administration of a USPIO at a dose of 2.6 mg Fe/kg. Analysis of MR images included qualitative and quantitative comparison of the USPIO and gadolinium enhancement of brain tumors. Brain surgery was performed 25-112 hr after administration of the USPIO. The histopathologic workup included iron histochemistry with diaminobenzidine (DAB)-enhanced Perls stain.
RESULTS. In seven of nine patients, USPIO-related changes of signal intensity were observed in gadolinium-enhancing brain tumors on T1- and T2*-weighted sequences. The difference in signal intensity on T1-weighted USPIO series was 40.1% ± 26.7% (mean ± SD). On T2*-weighted USPIO series, the difference in signal intensity was -33.1% ± 18.4% in solid tumor parts. Areas of suspected radiation necrosis did not enhance in three patients with prior radiation therapy. Iron histochemistry revealed the presence of iron deposits in macrophages in two patients.
CONCLUSION. USPIO agents will not replace gadolinium in the workup of patients with brain tumors. Our findings suggest that USPIO agents seem to offer complementary information and may help to differentiate between brain tumors and areas of radiation necrosis. Signal intensity changes on T2*-weighted images might be related to the blood pool properties of the agent, possibly reflecting steady-state susceptibility effects.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
MRI with ultrasmall superparamagnetic iron oxides (USPIO) is an alternative [1]. These monocrystalline particles are administered IV and are detectable at low tissue concentrations (< 60 ng Fe/mm3).
USPIO agents have been applied in various studies and have proven useful in differentiating malignant and benign lymph nodes [2-5].
Because of their pharmacokinetic characteristics with relatively long vascular persistence, USPIO agents also serve as a blood pool agent in MR angiography [6-9]. In two clinical surveys in patients with brain tumors, USPIOs have improved tumor delineation [10, 11].
The exact mechanism of USPIO enhancement in patients with brain tumors remains controversial. Several groups hypothesize that endocytosis of iron particles into tumor cells or migration of iron-loaded macrophages is responsible for enhancement of the USPIOs in brain tumors [12-16]. In addition, a superimposition of blood pool effects must be taken into account [17, 18]. Until now, only a few histopathologic data from the clinical administration of USPIOs in patients with brain tumors have been published [10, 11]. It is important, however, to determine the exact location of iron particles in the tumor and in adjacent tissues in vivo.
The aim of this study was to evaluate the characteristics of a USPIO agent in patients with brain tumors. This article discusses imaging characteristics of the agent in correlation with histopathologic data collected systematically in all patients, enhancing the understanding and interpretation of imaging changes with USPIO agents.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Contraindications to the administration of USPIO were a known allergy to dextrans or iron compounds, pregnancy, and breast-feeding. Patients with a recurrent tumor or prior brain tumor irradiation were not excluded. Histopathologic diagnoses of the patients were four glioblastomas multiforme (World Health Organization [WHO] grade IV), one gliosarcoma (WHO grade IV), one anaplastic ependymoma (WHO grade III), one oligodendroglioma (WHO grade II), one brain metastasis of a non-small cell bronchial carcinoma, and one cerebral B-cell lymphoma (Table 1).
|
Two patients (patients 1 and 2) had received interstitial radiation therapy with yttrium-90 before the neurosurgical intervention [19]. One patient was treated for tumor recurrence after open brain surgery and whole brain irradiation with 60 Gy (patient 6). The remaining six patients had newly diagnosed brain tumors. The study was approved by the ethics committee of our institution.
Contrast Agents
The contrast agent used was a pharmaceutical formulation of a USPIO with a
characteristically long plasma half-life of 21-30 hr in humans ([ferumoxtran,
AMI 227] Sinerem, Guerbet). It consists of an SPIO core surrounded by a
dextran coat to give a volume-weighted hydrodynamic diameter of 20-50 nm.
Lyophilized USPIO was reconstituted in 100 mL of a 0.9% saline solution and
administered IV through a microfilter by slow drip infusion at a dose of 2.6
mg Fe/kg over 30 min, with a slow flow rate of 2 mL/min during the first 10
min followed by an increased rate of 4 mL/min. During the administration,
patients were monitored closely, and no adverse events were noted in any
patients. Minor back pain was the adverse event most frequently reported in
other clinical trials. This pain probably is related to the speed of infusion,
which in our case was reduced during the first 10 min.
In addition, gadolinium-enhanced T1-weighted sequences were acquired after the unenhanced series with the IV administration of 0.1 mmol/kg of body weight gadopentetate dimeglumine (Magnevist, Schering).
MRI
Cerebral MRI was performed on a 1.5-T super-conducting system (Magnetom
Vision, Siemens Medical Solutions). A standard circularly polarized head coil
was used for all imaging procedures. The imaging protocol consisted of
T1-weighted spinecho, T2-weighted turbo spin-echo, and T2*-weighted
gradient-refocused echo (GRE) sequences in axial slices covering the entire
CNS.
In the first sitting, T1-weighted, T2-weighted, and T2*-weighted images were obtained. After the IV administration of gadopentetate dimeglumine, T1-weighted imaging was repeated. In a second sitting 2 hr after the administration of gadolinium, the USPIO agent was administered IV. At least 21 hr (mean, 24.8 hr; range, 21-30 hr) after administration of the USPIO agent, T1-weighted imaging and T2*-weighted imaging (T1- and T2*-weighted USPIO) were repeated in identical spatial orientation.
For axial T1-weighted spin-echo images, a TR/TE of 560/17 was used, and up to 24 slices with a 5-mm thickness without gap were obtained. A field of view of 173 x 230 mm and an acquisition matrix of 173 x 230 were chosen. Axial T2*-weighted GRE images were obtained using 997/24, a flip angle of 15°, and two excitations. The field of view was 173 x 230 with an acquisition matrix of 205 x 256, and the slice thickness was 5 mm without gap. For axial T2-weighted turbo spin-echo images, we used 3,800/90 at a slice thickness of 5 mm. The field of view was 173 x 230 with an acquisition matrix of 190 x 256.
Image Analysis
Qualitative analysis—The total number of enhancing brain
lesions as visualized on T1-weighted gadolinium-enhanced, T1-weighted
USPIO-enhanced, and T2*-weighted USPIO-enhanced images was
determined by two senior neuroradiologists in consensus and recorded for each
patient. Qualitative analysis included determination of tumor location and
detection of concomitant disorders such as hemorrhage and tumor necrosis.
Necrosis was assumed in those cases in which histopathology revealed the
presence of necrosis and imaging characteristics of necrosis were present. The
use of a neurosurgical navigation system would have been mandatory to
accurately link imaging and histopathology. Because of the setting of our
study, this was not possible.
|
Quantitative analysis—Quantitative analysis included the calculation of tumor volume and signal-to-noise ratio of enhancing brain tumors. To determine tumor volumes, the relevant cross-sectional diameters of the enhancing lesions were measured in seven patients on T1-weighted gadolinium, T1-weighted USPIO, and T2*-weighted USPIO images in the three cardinal planes using the caliper tool on the workstation. Tumor volumes were calculated using the ellipsoid formula and assuming an ellipsoid with three unequal axes [20].
Quantitative assessment of contrast enhancement in brain tumors was
performed as followed: Signal-to-noise ratios were determined in enhancing
brain tumors on T1-weighted gadolinium, T1-weighted USPIO, and
T2*-weighted USPIO images. To standardize our measurements, 3-5
regions of interest (ROIs) were placed on T1-weighted gadolinium images at the
level of the greatest transverse diameter and within the largest enhancing
portion of the tumor (SIpost, where SI = signal intensity). ROIs
were placed in corresponding locations on T1-weighted USPIO images and on
T2*-weighted USPIO images when contrast enhancement was present
(SIpost). In addition, signal intensities were determined in
corresponding areas on unenhanced (SIpre) T1- and T2-weighted
images. The sizes of the ROIs were predetermined by the size of the enhancing
tumor portion. The size of the ROI for the measurement of the background noise
was limited to 10 mm in diameter. Normalized signal intensities
(signal-to-noise ratio) were calculated by dividing the measured signal
intensity in the ROI by the SD of the background noise. Relative contrast
enhancement—that is, difference in signal intensity of brain
tumors—was calculated as
 |
where SIpre and SIpost denote the signal-to-noise values before and after the administration of gadolinium or USPIO.
|
Biopsy and Histopathology
Tumor specimens were collected systematically in all nine patients. Surgery
took place an average of 49.5 ± 26.4 hr (range, 25-112 hr) after the
administration of the USPIO. Open brain surgery with total or subtotal tumor
removal was performed in seven patients (patients 1-7). Two patients (patients
8 and 9) underwent CT-guided stereotactic biopsies in various tumor areas.
Formalin-fixed, paraffin-embedded tumor tissue was evaluated using conventional histology and immunohistochemistry. Brain tumors were diagnosed and graded according to the WHO classification of tumors of the CNS [21]. To assess the presence of macrophages, 4-µm sections were subsequently subjected to histoimmunologic analysis of macrophage content using a CD68 antibody (Novocastra Laboratories).
To document the presence of iron particles, iron histochemistry was performed using diaminobenzidine (DAB)-enhanced Perls stain. For this process, tissue sections were rehydrated in deionized water and incubated with a 7% solution of potassium ferrocyanide in 3% HCl for 60 min. Sections were then rinsed and counterstained with 1% neutral red. Tissue sections were washed in phosphate-buffered saline (PBS) and then incubated in 4% potassium ferrocyanide in 4% HCl for 1 hr. After rinsing in PBS, tissue sections were incubated with inactivated DAB containing 1% nickel chloride for 15 min. This was followed by a 15-min incubation with activated DAB and counterstaining with neutral red.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Side-by-side analysis revealed that the density in gadolinium-enhancing brain lesions appeared to be more homogeneous and the enhancing lesions appeared more clearly compared with the hazy pattern of enhancement obtained with the USPIO (Figs. 1, 2, and 3A). All USPIO-enhancing lesions also enhanced on T1-weighted gadolinium images. No USPIO-related changes of signal intensity were observed beyond the margins of gadolinium enhancement. In three patients who had undergone prior radiation therapy (patients 1, 2, and 6; Figs. 1 and 3A), large gadolinium-enhancing areas appeared but did not enhance on T1-weighted USPIO images. This finding was exemplified in a patient with a gliosarcoma previously treated with interstitial radiation therapy (Fig. 1). The gadolinium-enhancing medial rim of the right frontoparietal brain lesion paralleling the course of the catheter used for interstitial radiation therapy did not enhance with USPIO on T1-weighted images.
T1-weighted gadolinium-enhanced versus T2*-weighted USPIO-enhanced—On T2*-weighted USPIO images in seven of nine patients, a total of 11 enhancing brain tumors appeared with a decrease in signal intensity. In addition to the patient with oligodendroglioma (patient 9), the tumor in a patient with cerebral B-cell lymphoma (patient 8, Fig. 4) did not enhance on T2*-weighted USPIO images.
|
|
|
Side-by-side analysis revealed that demarcation of tumor margins was more clear on T2*-weighted USPIO images than on T1-weighted USPIO images. Gadolinium-enhancing areas of presumed radiation necrosis did not enhance on T1- or T2*-weighted USPIO images (patient 1, Fig. 1).
Tumor Imaging: Quantitative Analysis
For the exact quantification of signal intensity changes, we determined
normalized signal intensities in enhancing lesions on T1-weighted gadolinium
and T1-weighted USPIO images. The overall results are outlined in
Figure 5. On T1-weighted
gadolinium images, the normalized signal intensity showed a mean increase of
52.3% ± 23.4% in the enhancing tumor rind and 53.7% ± 22.1% in
enhancing solid tumor parts. On T1-weighted USPIO images, the normalized
signal intensity increased 40.1% ± 26.7% in the enhancing tumor rind
and 47.2% ± 31.9% in enhancing solid tumor parts. The decrease of the
normalized signal intensity was less pronounced on T2*-weighted
USPIO images. In the enhancing tumor rind, the decrease in signal intensity
was 33.1% ± 18.4%, and in enhancing solid tumor parts it measured 28.4%
± 16.4%. In the paired two-tailed Student's t test, the
changes of signal intensity on T1-weighted gadolinium, T1-weighted USPIO, and
T2*-weighted USPIO images were statistically significant
(p < 0.015).
|
|
Intracellular iron deposits were detected with DAB-enhanced Perls stain in two of nine patients. These iron deposits were located in macrophages. Iron histochemistry did not reveal any iron particles in the tumor interstitium or in tumor cells. In a patient with anaplastic ependymoma, iron-loaded macrophages were found in the border zone between normal white matter and tumor tissue (patient 6, Fig. 3B). In another patient with glioblastoma multiforme (patient 2), few iron deposits were detected in intratumoral macrophages. In the five remaining patients, who had shown distinct imaging changes on T1- and T2*-weighted USPIO images, no intracellular or interstitial iron deposits were detected. In five of nine patients (patients 1-4 and 6), the histopathologic workup of tumor specimens revealed the presence of intra- or peritumoral necrosis. The exact spatial correlation between MRI findings and histopathologic specimens was not feasible because of the clinical setup of our study.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Contrast-enhanced MRI in the management of patients with brain tumors is supposed to help define tumor margins. Yet gadolinium enhancement in brain tumors is variable and quite nonspecific, and any of a variety of inflammatory and infectious conditions show similar patterns of enhancement. A thorough side-by-side analysis of our series revealed that USPIO and gadolinium enhancement on T1-weighted images, in most cases, occurred in identical anatomic locations. Yet the tumor volumes differed considerably between the gadolinium and USPIO series. In addition, the demarcation of tumor margins on T2*-weighted images seemed biased by blooming artifacts because of the susceptibility effects of the USPIO agent. Therefore, we do not expect USPIO agents to replace gadolinium in the workup of patients with brain tumors.
On the basis of our observations in a preliminary study, we suggest that MRI with USPIO might provide additional information of a more specific nature. The lack of enhancement on T1- and T2*-weighted USPIO images in areas of presumed radiation necrosis in three patients with prior interstitial radiation therapy or whole brain irradiation suggests that USPIO might help to differentiate between gadolinium-enhancing brain tumors and areas of radiation necrosis. Improved delineation of vascular structures on T2*-weighted USPIO images and the visualization of areas of suspected neovascularization seem to corroborate findings from animal studies that USPIO agents help to image tumor microvascularization [22]. It is a short-coming of our study that we were not able to provide an accurate link between imaging findings and histopathologic data. The use of a neurosurgical navigation system would have been necessary to offer this correlation. Necrosis was assumed in those cases in which histopathology revealed the presence of necrosis and imaging characteristics of necrosis were present.
Unfortunately, we could not correlate our findings with MR perfusion studies because those techniques were not available at our institution at that time. These important data will have to be integrated into future studies.
Mode of Action
The potential of USPIO particles to act as a contrast agent in brain tumors
has been shown in various experimental and two clinical studies. Previous
results from studies in different animal models suggest local endocytosis of
iron particles by tumor cells
[12-14].
Evidence from experimental studies shows that iron oxides are internalized
into microglia and tumor cells
[16]. Dousset et al.
[15] have proposed an
alternative mode of action. They showed accumulation of iron particles in
lesion-centered macrophages in an animal model for experimental autoimmune
encephalomyelitis [15]. In
addition, change of signal intensity after USPIO administration is thought to
be related to a permeability effect, with extravasation of iron particles into
the peritumoral interstitium through a disrupted blood-brain barrier
[14].
Systematic histopathologic workup of tumor specimens in our series did not provide additional evidence that imaging changes with USPIO are related to intratumoral or interstitial iron deposits. Only in two of nine patients were iron deposits detected, and these were located in macrophages. It remains unclear whether these iron deposits are related to the USPIO agent. At least one patient revealed massive intratumoral hemorrhage. We cannot rule out that the described macrophages are siderophages in the vicinity of tumoral bleeding.
The apparent absence of iron particles in five of seven patients from our series presenting with distinct imaging changes on T1- and T2*-weighted USPIO sequences needs further consideration.
A possible explanation would be that DAB-enhanced Perls stain is not sufficiently sensitive to detect iron particles at low tissue concentrations. Yet DAB-enhanced Perls stain was successfully used for the detection of iron particles in clinical [10, 11] and various experimental [14-16] studies. The use of more sensitive methods, such as electron microscopy, should be considered in future studies. Furthermore, a sampling error could be responsible for our results. With the use of modern neurosurgical techniques such as cavitational ultrasonic surgical aspirator (CUSA), the tumor is aspirated rather than resected. In these cases, not all of the removed tumor might have reached the pathology department. However, it seems important to analyze the border zone between unaffected brain tissue and tumor interstitium because iron-loaded macrophages seem to be located predominantly in this area [15].
We suggest that an alternative mode of action be considered. USPIO agents have been widely used for MR angiography. Because of their large structure, with hydrodynamic diameters of 20-50 nm, they remain exclusively intravascular. Normal vessels and tumor vessels may thus be clearly visible on sensitive sequences for up to 20-30 hr after the administration of contrast material [22]. The blood pool properties of the USPIO agent might be responsible for the reported changes of signal intensities in our study.
The assessment of tumor microvasculature with USPIO agents has been reported in several other studies [18, 23, 24]. Le Duc et al. [23] recently published an experimental study with the same agent (ferumoxtran) in rats with gliomas. Fifteen minutes after the administration of the USPIO agent, peritumoral hypointense signals on T2-weighted sequences were observed. On histologic sections, these areas corresponded to highly vascularized regions overlapping the external part of the tumor. In that study, the authors concluded that T2-weighted steady-state susceptibility contrast can be used for studying intracerebral tumor vascularization.
Our observations support this theory: In our series, tumors that usually are highly perfused (4 glioblastoma, 1 gliosarcoma, 1 anaplastic ependymoma, and 1 metastasis) showed a distinct drop in signal intensity on T2*-weighted images after the administration of the USPIO agent. Anatomic structures with a high perfusion rate, such as the choroid plexus, changed signal intensity correspondingly on T2*-weighted USPIO images. Tumors known to be less perfused (1 cerebral B-cell lymphoma and 1 oligodendroglioma) did not show any changes in signal intensity on T2*-weighted sequences after the administration of the USPIO agent [25]. The decreased vascularity in areas of radiation necrosis with a subsequently decreased blood pool effect might help to explain why we did not observe USPIO enhancement in areas of suspected radiation necrosis (Fig. 1).
On the basis of these findings, we expect a mixed mode of action to be responsible for the reported image changes: The signal changes on T1-weighted USPIO images, which reach a maximum about 24 hr after administration of the contrast agent [10, 11], are most likely caused by diffusion of iron nanoparticles through a disrupted blood-brain barrier, intracellular uptake by tumor cells, and peritumoral macrophages. DAB-enhanced Perls stain is probably not sufficiently sensitive to reveal the presence of interstitial iron particles at low tissue concentrations. In addition, we might have missed iron-loaded macrophages, which seem to be located predominantly at the hyperpermeable tumor-brain interface, because of a sampling error [15].
The decrease of signal intensity on T2*-weighted GRE sequences is dose-dependent, probably reflecting two effects: First, signal intensity decreases because of an intracellular iron particle accumulation related to macrophage activity and migration within the tumor tissue. Second, the blood pool properties of the agent seem to support an additional mode of action, probably reflecting steady-state susceptibility effects.
Conclusion
USPIO agents will not replace gadolinium in the workup of patients with
brain tumors. USPIO agents seem to offer complementary information. This study
suggests that USPIO might help to differentiate between brain tumors and areas
of radiation necrosis. In addition, USPIO seems to enhance the delineation of
neovascularization.
Detection of iron particles was restricted to two patients, although imaging changes were present in seven of nine patients. We suggest that the blood pool properties of the agent need to be considered because they seem to support an additional mode of action on T2*-weighted sequences, probably reflecting tumor microvascularization.
Further clinical studies, including functional brain imaging with diffusion, perfusion, and MR spectroscopy, are necessary to investigate whether the additional costs of a USPIO agent are justified by improved diagnostic accuracy when compared with gadolinium-based MRI.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
