Distinction between pyogenic brain abscess and necrotic brain tumour using 3-tesla MR spectroscopy, diffusion and perfusion imaging
Abstract
The purpose of this study is to compare the effectiveness of relative cerebral blood volume, apparent diffusion coefficient and spectroscopic imaging in differentiating between cerebral abscesses and necrotic tumours. In the prospective study, a 3-tesla MR unit was used to perform proton MR spectroscopy, diffusion and perfusion imaging in 20 patients with cerebral abscesses and 26 patients who had solitary brain tumours (14 high-grade gliomas and 12 metastases). We found the mean apparent diffusion coefficient value at the central cavities of the cerebral abscesses to be significantly lower than in necrotic tumours. The mean relative cerebral blood volume values of the necrotic tumour wall were statistically significantly higher than the mean relative cerebral blood volume values of the cerebral abscess wall by the Student's t-test. The proton spectra obtained revealed amino acids only in the cerebral abscesses. Although the conventional MRI characteristics of cerebral abscesses and necrotic tumours may sometimes be similar, diffusion, perfusion-weighted and spectroscopic MRI enables distinction between the two.
In vivo MR spectroscopy (MRS) techniques, including single-voxel MRS [1–3] and more advanced MRS imaging (MRSI) [4, 5], are non-invasive techniques that provide information about the metabolic characteristics of tissue in the brain. Difficulties in the diagnosis of intracranial abscesses are mainly due to the combination of non-specific clinical findings and possible similarities in the radiological appearances of necrotic gliomas, metastases and brain abscesses [6]. Conventional contrast-enhanced MRI reveals ring enhancement of a brain abscess that is similar to the ring enhancement of a necrotic high-grade glioma or metastasis. The management of these two disease entities is different and can potentially affect the clinical outcome.
Advanced MRI techniques are used to obtain metabolic information that complements the anatomical images provided by conventional MRI. Perfusion-weighted MRI provides non-invasive measurements of vascularity. Relative cerebral blood volume (rCBV) maps derived from perfusion-weighted MR images can be used to quantify areas of neovascularisation. Perfusion MRI has become an important means of characterising intracranial neoplasms [7]. In vivo MRS provides metabolic information about brain abscess [8] and tumours [9, 10].
The limited spatial and spectral resolution at the standard magnetic field of 1.5 T are overcome at a higher field strength 3-T scanner due to improved spectral resolution and increased signal-to-noise ratio (SNR). The improved spectral resolution provides separation of metabolite signals and thereby improves the ability to distinguish among tumour, normal brain tissue and non-viable tissues, including purulent material, oedema, gliosis and necrosis [11, 12]. Despite these observations, to our knowledge there are no published reports of the three diagnostic methods being performed on the same patients with pyogenic brain abscess.
The purpose of this study is to investigate the potential roles of 3-T MRS, relative cerebral blood volume and apparent diffusion coefficients (ADCs) in distinguishing between pyogenic brain abscess and necrotic tumour on the basis of differences in vascularity, water self-diffusion and metabolite levels.
Methods and materials
Patients
In the prospective study carried out between March 2006 and March 2008, a 3-T MR unit was used to perform proton MRS, diffusion imaging and conventional MRI on 26 patients with solitary brain tumours (14 high-grade gliomas and 12 metastases). 12 perfusion MR studies (8 high-grade gliomas and 4 metastases) were also performed. The 26 patients (12 men and 14 women; 25–76 years of age) with untreated solitary brain tumours underwent MRI. The presence of the high-grade glioma or metastases was histologically verified by means of either stereotactic biopsy or surgical resection.
In all of the 14 patients with high-grade gliomas, the tumour type was glioblastoma multiforme. In all 12 patients with biopsy-proven metastases, the primary tumour was a carcinoma: 9 from known primary sites (2 breast, 5 lung, 2 stomach) and 3 from an unknown primary site. 20 patients (8 women and 12 men; age range 35–81 years) presented with surgically confirmed pyogenic brain abscesses. All patients underwent MRS, diffusion and perfusion imaging. In all of these cases, purulent samples were analysed.
Informed consent for MR examination was obtained from all patients.
MR techniques
MRI was performed with a 3-T Signa system (GE Medical Systems, Milwaukee, WI). A standard quadrature head coil (GE Medical Systems) was used for all patients. Each MR examination included pre-contrast transverse T2 and T1 weighted spin-echo imaging followed by diffusion-weighted imaging. Perfusion-weighted imaging and MRS were performed after contrast material administration. In each study, T2 weighted images were acquired using a spin-echo sequence with a repeat time to echo time (TR/TE) of 4000/100, a slice thickness of 3 mm and field of view (FOV) of 240×180. Pre-contrast T1 weighted images were then acquired with a slice thickness of 3 mm and a TR/TE of 800/9. Post-contrast axial images were obtained after perfusion-weighted images. Subsequent diffusion-weighted and spectroscopic imaging were also performed.
Perfusion-weighted imaging
A series of T2 weighted gradient-echo-planar images were obtained during the first pass of a bolus of contrast material, at a dose of 0.1 mmol kg−1 body weight. The section thickness and location of the perfusion-weighted MR data set were determined by using the axial T2 weighted images to locate the lesions. The size of the region of interest (ROI) was 2–5 mm in diameter (depending on the size of the lesion) and ROI location was based on the colour overlap maps. The rCBV values of the central region and peripheral wall of the cerebral abscess and brain tumour were calculated.
Spectroscopy
Spectroscopic data were obtained after gadopentetate dimeglumine administration. A data set was obtained from a selected volume of 18–135 cm3 (minimum 3×4×1.5 cm; maximum 10×9×1.5 cm) by using a spin-echo (SE) sequence with phase-encoding gradients in two directions, automatic shimming and Gaussian water suppression. Measurement parameters were 1500/136/2 (TR/TE/excitations), 16×16 phase-encoding steps,160×160 mm field of view,15 mm section thickness and 1024 data points. Data sets of 1.5 cm3 resolution were acquired within 12 min (6 min in the six measurements with one excitation). The rectangular spectroscopic ROI was localised by using the transverse T1 weighted 3-T MR images and the size ranged from 1×1×1 cm to 1×1×1.5 cm. Total examination time, including set-up, was approximately 40 min.
All lesions, including abscesses and tumours, exhibited a necrotic centre that was hypointense on T1 weighted images and hyperintense on T2 weighted images. The diameter of the necrotic cystic centre was at least 1 cm. The perilesional ring was regular or irregular and enhanced after an injection of gadolinium-based contrast material.
Measurements of metabolite ratios were made in ROIs in the lesion centre and in ROIs in similar contralateral regions.
Assignment of resonance peaks for the metabolites was based on previously documented 1H-MRS studies of necrotic brain tumours and brain abscesses [13].
Diffusion-weighted imaging
Diffusion-weighted imaging was performed in the transverse plane by using an SE echo-planar imaging sequence with the following parameters: TR/TE/TI (inversion time), 12 000/100/2200 ms; diffusion gradient encoding in three orthogonal directions; b = 1000 s mm−2; FOV, 20×40 cm; matrix size, 128×64 pixels; section thickness, 5 mm; section gap, 2.5 mm; and number of signals acquired, 1. After pixel-by-pixel calculation, an ADC map was obtained. The ADC values of the central region and peripheral wall of the cerebral abscess and brain tumour were then calculated by manually placing circular ROIs, guided by T2 weighted and contrast-enhanced images. Each measurement was performed by using ROI areas (average area of ROIs was 2±5 mm).
Data processing
Perfusion-weighted imaging
The MR data were processed as previously described by Knopp et al [14]. Maximal rCBV values were obtained by identifying regions of maximal perfusion from colour maps. The full calculation of relative CBV outlined in the preceding paragraphs was then applied to ROIs over these regions, expressed relative to values measured in contralateral white matter.
Spectroscopy
The in vivo MR spectroscopic data were analysed on a Sun Ultra 1 workstation (Sun Microsystems, Mountain View, CA) by using Sage IDL processing software (GE Medical Systems). The spectral data were baseline corrected, apodised, filtered and Fourier transformed. Because there were three-dimensional acquisitions, the spectra could be reconstructed to coincide with the ROI specified in three dimensions by using these programs. Metabolite values were calculated automatically from the area under each metabolite peak by using the standard commercial software program provided by the manufacturer.
Diffusion-weighted imaging
Analysis of diffusion changes was performed by calculating the ADC, based on the Stejskal and Tanner equation [15], as the negative slope of the linear regression line best fitting the points for b versus ln (SI), where SI is the signal intensity from a ROI of the images acquired at the two b values (1500 s mm−2 and 1000 s mm−2). A software package written in Matlab to process the diffusion-weighted images and ADC maps were generated by performing this calculation on a pixel-by-pixel basis. Standard mean ADC values were calculated automatically and expressed in 10−3 mm2 s−1.
Statistical analysis
Relative cerebral blood volumes (rCBV; ml/100 g) from perfusion and ADC (×10−9 m2 s−1) from diffusion were measured in the centre and contrast-enhancing regions.
Apart from qualitative spectroscopic analysis, the rCBV measurements obtained from the perfusion-weighted MR data were analysed by the Student's t-test to determine the statistical difference in rCBV between pyogenic brain abscess and necrotic tumour. The Student's t-test was also used to determine if there was a statistically significant difference in ADC values between pyogenic brain abscess and necrotic tumour. A p-value of <0.05 was considered to be statistically significantly different.
Results
An axial rCBV map showed intense signal hyperintensity in the tumour wall, indicating that the tumour wall possessed a high rCBV value (Figure 1d). The mean rCBV values of the peripheral tumour wall of necrotic brain tumours were significantly higher than the mean rCBV values of the cerebral pyogenic abscess wall by the Student's t-test (Table 1).
The rCBV values in the central region of all cerebral abscesses did not differ statistically from those seen with necrotic brain tumours (Table 1).
On diffusion-weighted MR images, the central cavities of the cerebral abscesses had very low ADCs, which accounted for the signal hypointensity on ADC map images (Figure 2b). We found the mean ADC values at the central cavities of the cerebral abscesses to be significantly lower than in necrotic tumours (Table 2).
The ADCs in the wall of cerebral abscesses did not differ statistically from those seen in the peripheral portions of necrotic tumours (Table 2).
On 1H-MRS, the predominant resonance peaks (N-acetylaspartate, choline and creatine/phosphocreatine) that are usually observed in the parenchyma of the normal brain were not detectable in either tumoral or abscess necrosis. 1H-MRS of brain abscess revealed multiple additional resonance peaks. The main findings were the resonances of amino acids (valine, leucine and isoleucine; 0.9 ppm), acetate (1.9 ppm), alanine (1.5 ppm), succinate (2.4 ppm) and lactate (1.3 ppm) identified in 16 patients with abscesses who were analysed (Figure 3b). In the remaining four patients, the spectra were contaminated by surrounding tissues.
The resonances of amino acids (0.9 ppm), alanine (1.5 ppm), acetate (1.9 ppm) and succinate (2.4 ppm) were identified, respectively, in 16, 11, 7 and 2 of the 20 patients with abscesses who were analysed. Lactate was always detected (1.3 ppm), and lipids (0.9 and 1.3 ppm) were found in five patients. The resonances of lipids (at 0.9 ppm) and lactate (at 1.35 ppm) were detected in two cases of cerebral metastases. The resonances of amino acids (valine, leucine and isoleucine; 0.9 ppm), acetate (1.9 ppm), alanine (1.5 ppm) and succinate (2.4 ppm) were not detectable in tumoral necrosis.
Discussion
We found that neither visual inspection nor measuring signal intensity means from T1 and T2 weighted images could differentiate between abscesses and necrotic tumours. The conventional MRI findings in cerebral abscess are well documented. Cerebral abscesses are characterised by three distinct zones of signal abnormality.
The abscess cavity often demonstrated a long T1 and T2 value, whereas the abscess capsule often displays shortening of both the T1 and the T2 value. Peripherally, a zone of oedema is usually present surrounding the lesion. The abscess wall usually shows a ring of enhancement following contrast administration [16]. Unfortunately, these features are not present in all abscesses and may be seen in a variety of tumours. Gliomas are the most common brain tumours. On imaging studies, malignant gliomas are usually enhanced after intravenous contrast administration and show peritumoural oedema, whereas, except for pilocytic astrocytoma, low-grade gliomas usually show little to no abnormal contrast enhancement and peritumoural oedema [11]. The brain is a frequent site of haematogenous metastases from malignant tumours in other organs. Along with primary glial tumours in the brain, metastases are associated with high morbidity and mortality.
The approach to imaging diagnosis, treatment and follow-up is different between the cerebral abscesses and necrotic tumours. Because conventional MRI is limited in its ability to distinguish brain abscesses from necrotic tumours, new and more advanced techniques are urgently needed. Diffusion-weighted imaging has occasionally been used to diagnose cerebral abscesses. Desprechins et al [17] showed that it is possible to reach a differential diagnosis based on the high signal intensity in diffusion imaging and a strongly reduced apparent diffusion coefficient. Noguchi et al [18], investigating two pyogenic brain abscesses and six high-grade gliomas, reported a significant difference between the ADC values. Because of the low incidence of abscesses, the preliminary studies involved only small numbers of patients. However, we were able to use a larger sample, which allowed us to detect a statistically significant difference between the groups.
Until recently, absolute ADC measured in intracranial tumours had been rarely reported. Brunberg et al [19] reported 40 patients with cerebral gliomas, using a motion-insensitive line-scanning SE sequence measuring ADC along the three Cartesian axes. Lu et al [20] confirmed statistically significant changes between the mean diffusivity (MD) and fractional anisotropy (FA) values of normal-appearing white matter and those of the peritumoural T2 signal intensity abnormality. Surrounding both gliomas and metastatic lesions, there was an increase in MD and a decrease in FA, which are best explained by increased extracellular bulk water.
The cystic or necrotic portion of the brain tumour has a low viscosity and free molecular diffusion and thus a high ADC. In contrast, the central cavity of an abscess is filled with pus, which is a yellowish brown viscous fluid that consists of inflammatory cells, bacteria, necrotic tissue and proteinaceous exuded plasma with a very high viscosity and cellularity [21], and thus resulting in markedly decreased ADC. This finding may help to distinguish between abscesses and necrotic tumours pre-operatively.
In this study, it was shown that there was no significant difference in signal intensity on diffusion-weighted images between the cerebral abscess wall and the brain tumour wall. The hypointensity of the abscess wall is probably due to the increase in extracellular fluid in the capsular wall as a result of inflammation. On the other hand, the peripheral portion (or wall) of the cystic or necrotic brain tumour has a higher ADC, which suggests increased intracellular and extracellular water fractions owing to more fluid production [22].
In our prospective study we found that the cerebral abscess wall possessed a decreased rCBV owing to the relatively poor vascularity of the capsular wall. In contrast, the cystic or necrotic tumour wall possessed an elevated rCBV as a result of the microvessel angiogenesis of the tumour tissue. We verify that these findings are consistent with known pathophysiological findings.
Cerebral abscesses have been shown to have large amounts of mature collagen and decreased vascularity [23], which may serve as the explanation for the lower rCBV ratios. In high-grade gliomas, the enhancing portion of the tumour demonstrates breakdown of the blood–brain barrier. Concerning the relationship between the degree of contrast enhancement on post-contrast images and of vascularity on perfusion-weighted images in brain tumours, in a previous report by Sugahara et al [24] all glioblastoma multiforme tumours showed high vascularity as well as prominent contrast enhancement. Other authors have also reported that high CBV seems to correlate with high-grade tumours [25]. Our cases are all high-grade gliomas and cerebral metastases with elevated rCBV values in the enhancing portions guided by T1 weighted post-enhanced images.
In many articles on in vivo MRS, the voxel size is 8 cm3 or larger [26–29]. The 3-T data obtained in the present study suggest that this degree of spatial resolution may not be adequate for characterising heterogeneous tumour beds. Because of the improved spectral resolution and gain in SNR, we used a higher resolution than the standard magnetic field of 1.5 T. The twofold improvement in SNR at 3 T compared with that at 1.5 T allowed the use of a 10 mm section thickness (rather than the usual 15–20 mm) and an in-plane resolution as small as 0.75 (32×32 phase-encoding matrix), while maintaining an SNR of 5–6 for creatine.
Findings from several studies have suggested that in vivo H-MRS, a non-invasive examination, might contribute to the establishment of the differential diagnosis between brain tumours and abscesses [30–32]. The main characteristic features of pyogenic abscesses were the resonances of amino acids (valine, leucine and isoleucine; 0.9 ppm), acetate (1.9 ppm), alanine (1.5 ppm), lactate (1.3 ppm) and succinate (2.4 ppm). The specific spectrum of the abscess cavity (Figure 3b) shows elevation of acetate, succinate and some amino acids, as well as lactate, and this spectrum appears to be significantly different from the spectra of cystic or necrotic brain tumours. Our results agree with those of previous studies [32–34].
Increases in lactate, acetate and succinate presumably originate from the enhanced glycolysis and fermentation of the infecting micro-organisms [32–34]. Amino acids such as valine and leucine are known to be the end products of proteolysis by enzymes released by neutrophils in pus [32–34]. Lactate and lipids are non-specific metabolites produced by anaerobic glycolysis and necrotic tissue in brain abscesses. Both lactate and lipid peaks can also be observed in necrotic tumours, and are therefore not useful markers of abscess [33]. Alanine has been detected in the spectra of meningiomas [35]. Although the occurrence of acetate appears relatively high (35%, 7 out of 20), succinate was detected in only 2 out of 20 patients (10%). Lactate and alanine lack specificity for the distinction of brain abscesses, but they are reliably present: 100% (20 out of 20) for lactate and 55% (11 out of 20) for alanine.
Some workers report that normal brain peaks for N-acetylaspartate, choline and creatine were found on MRS of brain abscesses due to the contamination by surrounding brain tissue [36]. We saw no normal brain peaks in 16 of 20 analysed purulent samples. This may be because we always used small ROI with a voxel size of 1 cm3 to prevent such contamination.
Practically, it may be difficult to perform all these three techniques in addition to conventional images in each patient because it would be too time-consuming. However, given the fact that diffusion-weighted images have become routine protocol for most brain MRI studies, we suggest reviewing diffusion-weighted images for every patient with this clinical condition if MRS and perfusion-weighted sequencing are not feasible.
In conclusion, spectroscopic, diffusion and perfusion-weighted MRI are advanced MR techniques that are used to add important physiological and metabolic information to that obtained with conventional MRI. This study demonstrates that diffusion (ADC values of central regions), perfusion-weighted (rCBV values of peripheral regions) and spectroscopic MR measurements can be used to demonstrate differences in cerebral abscesses and necrotic tumours.
(a) Localising image from post-contrast axial T1 weighted MR in a 54-year-old woman with surgically proven pyogenic brain abscess. (b) The relative cerebral blood volume (rCBV) in the abscess wall (region of interest (ROI) 2) is reduced to 32% (relative to values measured in contralateral white matter). (c) Localising image from axial T1 weighted images in a 46-year-old man with pathologically proven glioblastoma multiforme. (d) There is elevation of rCBV in the peripheral wall (ROI 2) to 143% (relative to values measured in contralateral white matter). (a) Localising image from post-contrast axial T1 weighted MR in a 54-year-old woman with surgically proven pyogenic brain abscess. (b) Axial apparent diffusion coefficient (ADC) map shows signal hypointensity in the abscess cavity owing to relatively restricted diffusion (region of interest (ROI) 1). (c) Localising image from axial T1 weighted images in a 46-year-old man with pathologically proven glioblastoma multiforme. (d) Axial ADC map shows signal hyperintensity in the necrotic cavity (ROI 1) due to relatively free diffusion. (a) Localising image from post-contrast axial T1 weighted MR in a 54-year-old woman with surgically proven pyogenic brain abscess with a voxel in the necrotic centre. (b) In vivo1H spectra (1500/136). Note the absence of peaks for N-acetylaspartate, choline and creatine, and the presence of peaks for acetate (AC), alanine (ALA), lactate (LAC, inverted peak at 1.3 ppm with a 136 ms echo time) and succinate (SUC). The multiplet of amino acids (AA) is large and can be differentiated from lipids by its inversion at 136 ms. (c) Localising image from axial T1 weighted images in a 46-year-old man with pathologically proven glioblastoma multiforme with a voxel in the necrotic centre. (d) In vivo1H spectra (1500/136) from the necrotic centre of the tumour. Note the absence of peaks for N-acetylaspartate, choline, creatine and the multiplet of amino acids (AA).
 |
 |
References
1 Kaminogo M, Ishimaru H, Morikawa M, Oki M, Ushijima R, Tani M, et al. Diagnostic potential of short echo time MR spectroscopy of gliomas with single-voxel and point-resolved spatially localized proton spectroscopy of brain. Neuroradiology 2001;43:353–63.
2 Luyten PR, Marien AJ, Heindel W, van Gerwen PH, Herholz K, den Hollander, et al. Metabolic imaging of patients with intracranial tumors: H-1 MR spectroscopic imaging and PET. Radiology 1990;176:791–9.
3 Alger JR, Frank JA, Bizzi A, Fulham MJ, DeSouza BX, Duhaney Mo, et al. Metabolism of human gliomas: assessment with H-1 MR spectroscopy and F-18 fluorodeoxyglucose PET. Radiology 1990;177:633–41.
4 Nelson SJ, Huhn S, Vigneron DB, Day MR, Wald LL, Prados M, et al. Volume MRI and MRSI techniques for the quantitation of treatment response in brain tumors. Magn Reson Imaging 1997;7:1146–52.
5 Vigneron DB, Bollen A, McDermott M, Lawrence W, Mark D, Roland H, et al. Three-dimensional magnetic resonance spectroscopic imaging of histologically confirmed brain tumors. Magn Reson Imaging 2001;19:89–101.
6 Lau DW, Klein NC, Cunha BA. Brain abscess mimicking brain tumor. Heart Lung 1989;18:634–7.
7 Henry RG, Vigneron DB, Fischbein NJ, Grant PE, Day MR, Noworolski SM, et al. Comparison of relative cerebral blood volume and proton spectroscopy in patients with treated gliomas. AJNR Am J Neuroradiol 2000;21:357–66.
8 Harada M, Tanouchi M, Miyoshi H, Nishitani H, Kannuki S. Brain abscess observed by localized proton magnetic resonance spectroscopy. Magn Reson Imaging 1994;12:1269–74.
9 McBride DQ, Miller BL, Nikas DL, Buchthai S, Chang L, Chiang F, et al. Analysis of brain tumors using 1H magnetic resonance spectroscopy. Surg Neurol 1995;44:137–44.
10 Shimuzu H, Kumabe T, Tominaga T, Kayama T, Hara K, Ono Y, et al. Non-invasive evaluation of malignant brain tumors with proton MR spectroscopy. AJNR Am J Neuroradiol 1996;17:737–47.
11 Knopp EA, Cha S, Johnson G, Mazumdar A, Golfinos G, Zagzag D, et al. Glial neoplasms: dynamic contrast enhanced T2-weighted MR imaging. Radiology 1999;211:791–8.
12 Tien RD, Felsberg GJ, Friedman H, Brown M, McFall J. MR imaging of high-grade cerebral gliomas: value of diffusion-weighted echoplanar pulse sequences. AJR Am J Roentgenol 1994;162:671–7.
13 Demaerel P, Van Hecke P, Van Oostende S, Baert AL, Jaeken J, Declercq P, et al. Bacterial metabolism shown by magnetic resonance spectroscopy. Lancet 1994;344:1234–5.
14 Knopp EA, Cha S, Johnson G, Mazumdar A, Golfinos JG, Zagzag D, et al. Glial neoplasms: dynamic contrast enhanced T2-weighted MR imaging. Radiology 1999;211:791–8.
15 Stejskal EO, Tanner JE. Spin diffusion measurement: spin echoes in the presence of a time-dependent field gradient. J Chem Physics 1965;42:288–92.
16 Haimes AB, Zimmerman RD, Morgello S, Weingarten K, Becker RD, Jennis R, et al. MR imaging of brain abscess. AJR Am J Roentgenol 1989;152:1073–85.
17 Desprechins B, Stadnik T, Koerts G, Shabana W, Breucq C, Osteaux M. Use of diffusion-weighted MR Imaging in differential diagnosis between intracerebral necrotic tumors and cerebral abscesses. AJNR Am J Neuroradiol 1999;20:1252–7.
18 Noguchi K, Watanabe N, Nagayoshi T, Kanazawa T, Shimizu M, Seto H. Role of diffusion-weighted echo-planar MRI in distinguishing between brain abscess and tumor: a preliminary report.Neuroradiology 1999; 171–4.
19 Brunberg JA, Chenevert TL, Mckeever PE, Ross DA, Junck LR, Muraszko KM, et al. In vivo MR determination of water diffusion coefficients and diffusion anisotropy: correlation with structural alteration in gliomas of cerebral hemispheres. AJNR Am J Neuroradiol 1995;16:361–71.
20 Lu S, Ahn D, Johnson G, Cha S. Peritumoral diffusion tensor imaging of high-grade gliomas and metastatic brain tumors. Am J Neuroradiol 2003;24:937–41.
21 Winn WC Jr, Kissane JM. Bacterial disease. In: Damjanov I, Linder J, editors. Anderson's pathology, 10th edn. St Louis, MO: Mosby, 1996;747–842.
22 Steen RG. Edema and tumor perfusion: characterization by quantitative proton MR imaging. AJR Am J Roentgenol 1992;158:259–64.
23 Britt RH, Enzmann DR. Neuropathological and computerized tomographic findings in experimental brain abscess. J Neurosurg 1981;55:590–603.
24 Sugahara T, Korogi Y, Kochi M, Ikushima I, Hirai , Okuda T, et al. Correlation of MR imaging-determined cerebral blood volume maps with histologic and angiographic determination of vascularity of gliomas. AJR Am J Roentgenol 1998;171:1479–86.
25 Aronen HJ, Gazit IE, Louis DN, Buchbinder BR, Pardo FS, Weisskoff RM, et al. Cerebral blood volume maps of gliomas: comparison with tumor grade and histologic findings. Radiology 1994;191:41–51.
26 Ikehira H, Miyamoto T, Yasukawa T, Obata T, Katoh H, Koga M, et al. Differences in metabolic and morphological reactions after radiation therapy: proton NMR spectroscopy and imaging of patients with intracranial tumors. Radiat Med 1995;13:199–204.
27 Usenius T, Usenius J, Tenhunen M, Vainio P, Johansson P, Soimakallio S, et al. Radiation-induced changes in human brain metabolites as studied by 1H nuclear magnetic resonance spectroscopy in vivo. Int J Radiat Oncol Biol Phys 1995;33:719–24.
28 Walecki J, Sokol M, Pieniazek P, Maciejewski B, Tarnawski B, Krupska T, et al. Role of short TE 1HMR spectroscopy in monitoring of post-operation irradiated patients. Eur J Radiol 1999;30:154–61.
29 Taylor JS, Langston JW, Reddick WE, Kingsley P, Ogg R, Pui M, et al. Clinical value of proton magnetic resonance spectroscopy for differentiating recurrent or residual brain tumor from delayed cerebral necrosis. Int J Radiat Oncol Biol Phys 1996;36:1251–61.
30 Harada M, Tanouchi M, Miyoshi H, Nishitani H, Kannuki S. Brain abscess observed by localized proton magnetic resonance spectroscopy. Magn Reson Imaging 1994;12:1269–74.
31 Remy C, Grand S, Lai ES, Belle V, Hoffmann D, Berger F, et al. 1H MRS of human brain abscesses in vivo and in vitro. Magn Reson Med 1995;34:508–14.
32 Poptani H, Gupta RK, Jain VK, Pandey R. Cystic intracranial mass lesions: possible role of in vivo MR spectroscopy. Magn Reson Imaging 1995;13:1019–29.
33 Grand S, Lai ES, Esteve F, Rubin C; Hoffmann D, Rémy C, et al. In vivo 1H MRS of brain abscesses versus necrotic brain tumors. Neurology 1996;47:846–8.
34 Kim SH, Chang KH, Song IC, Ham MH, Kim HC, Kang HS, et al. Brain abscess and brain tumor: discrimination with in vivo H-1 MR spectroscopy. Radiology 1997;204:239–45.
35 Ott D, Hennig J, Ernst T. Human brain tumors: assessment with in vivo proton MR spectroscopy. Radiology 1993;186:745–52.
36 Dev R, Gupta RK, Poptani H, Roy R, Sharma S, Husain M, et al. Role of in vivo proton magnetic resonance spectroscopy in the diagnosis and management of brain abscesses. Neurosurgery 1998;42:37–43.
