Introduction
Intra-axial brain lesion (neoplastic and infective) is a widely prevalent clinical condition that often presents as an imaging dilemma. Most such lesions are presently imaged and diagnosed using diffusion-weighted imaging, perfusion imaging, and magnetic resonance spectroscopy in addition to conventional magnetic resonance imaging (MRI) sequences.[ 1 ] Infectious brain lesions in some patients may appear similar to high grade gliomas (HGG) and brain metastases. When making treatment decisions, it is crucial to distinguish between the two.[ 2 ] Diffusion tensor imaging (DTI) is a newer way to evaluate and differentiate such lesions.[ 3 ]
DTI identifies directional differences in the water molecule diffusivity in a given voxel within the tissue using several diffusion gradients. It measures the average directional variation of water diffusivity for a specific voxel in terms of fractional anisotropy (FA) and mean diffusivity (MD), to offer information on the tissue microstructure and architecture.
Aim
To compare the neoplastic and infective brain lesions based on DTI.
Objectives
The quantitative comparison of neoplastic and infective brain lesions based on DTI values (FA and MD) in lesional and perilesional areas and qualitative comparison of white matter tract involvement in neoplastic and infective brain lesions based on fiber tracking.
Materials and Methods
We did an analytical cross-sectional study to compare the lesions using the parameters of DTI in newly diagnosed cases of gliomas, metastasis, tuberculoma, and neurocysticercosis. The study was conducted in Department of Radiodiagnosis in collaboration with Departments of Neurology and Neurosurgery of a Tertiary Health Care level hospital in Northern India, it was completed in a period of 14 months. The study was approved by Institutional Ethics Committee (Ref code: 103rd ECM II B-THESIS/P50). The participants were informed and written informed consent was taken. The sample size was 50 (25 each of neoplastic and infective category). Study population comprised -26 males and 24 females; ages 15–67, with intracranial neoplastic and infective lesions who were referred for MRI brain scan. The neoplastic category included metastasis (12 cases), high grade glioma (13 cases), infective category included tuberculomas (13 cases), and neurocysticercosis (12 cases).
Magnetic resonance imaging scanning protocol
MRI was done on a 1.5 Tesla General Electric scanner (SIGNA EXPLORER), using a 16 channel head coil. Axial T2, T1/T2 fluid attenuated inversion recovery, coronal T2, sagittal T2, susceptibility-weighted imaging, diffusion-weighted imaging (b value = 1000 mm2 /s), and DTI were obtained before Gadolinium administration. Postcontrast axial T1 and 3D BRAVO were obtained after gadolinium administration. The sensitivity-encoding, or SENSE, parallel-imaging approach was used to collect DTI data using a single-shot echo-planar imaging sequence (frequency 128, phase 128, phase field of view [FOV] - 24.00, repetition time [TR] - 8800, echo time [TE] - 86.4, FOV - 24, axial plane, number of directions- 20, number of excitations = 1) were obtained. The image matrix was 128 × 128, with a 220 × 220 mm FOV. With a spacing value of 0, transverse sections with a slice thickness of 4 mm were taken parallel to the anterior commissure–posterior commissure line. There were no gaps in any of the 50 sections that spanned the hemispheres and brainstem. The b value was 1000 mm2 /s, and the diffusion weighting was stored along 20 distinct directions. TE = 86.4 ms, TR = 8800 ms, and number of acquisitions = two.
Postprocessing
The DTI data were uploaded to a workstation (General Electric, USA) for postprocessing, and images were evaluated with READY view Software, which included automatic diffusion maps construction. Regions of interests (ROI)-based calculation of FA and MD values was done after using a motion correction algorithm to adjust for head motion and image distortion caused by eddy currents. The FA and color-coded structural diffusion tensor maps were generated. Conventional MRI sequences were evaluated for lesion location, characterization, and enhancement pattern before the diffusion maps were evaluated. On DTI postprocessed maps, three circular ROIs (2 to 10 mm2 ) were set within the lesion and in the region of perilesional edema (within 1 cm of the lesion) for the quantitative evaluation. Mean FA and MD of lesion and perilesional edema were obtained, and comparisons were made between the infective and neoplastic groups. According to the FA values, the vector maps were color coded - red (left–right), green (anterior–posterior), and blue (superior–inferior).
Seed points -ROI were placed to reconstruct tracts of interest via fiber tracking, using current anatomic knowledge regarding tract projections. We compared tumor-affected areas to analogous tracts in the contralateral normal hemisphere. This was done both subjectively (on the maps) by comparing to the contralateral (normal) tracts and statistically by comparing these tracts’ FA and MD values.
Lesional white matter tracts were classified into displaced, edematous, infiltrated, and disrupted. The tract was classified as displaced if it displayed normal to slightly low FA (<0.5) with anomalous location and direction as a result of bulk displacement. If the tract had reduced anisotropy values (<0.22) with usual orientation but a high signal intensity on T2WI, it was deemed edematous.[ 4 ] If the tract had reduced anisotropy values (<0.2) but was still apparent on color images, it was classified as infiltrated.[ 5 ] If the tract had isotropic diffusion and could not be detected on directional color images, it was considered disrupted.[ 6 ]
Statistical analysis was done using SPSS (Statistical Package for Social Sciences) version 21.0 statistical Analysis Software (IBM Corp., Armonk, NY). The values were represented in number (%) and mean ± standard deviation (SD).
Results
Fifty cases with intra-axial brain lesions aged between 15 and 67 years (mean ± SD: 39.58 ± 15.74) were included in the study. Thirteen each of high-grade glioma and tuberculoma and 12 each of metastasis and neurocysticercosis were included. Majority of cases having neoplastic lesions were aged 31–50 years (52.0%), while a majority of cases with infectious lesions were aged 13–30 years (56.0%). On comparing statistically, higher age was found to be significantly associated with neoplastic lesions.
Perilesional FA was significantly higher in neoplastic as compared to infective group (0.17 ± 0.12 vs. 0.08 ± 0.02). Perilesional MD was significantly higher in infective group as compared to neoplastic group (1.51 ± 0.22 vs. 1.20 ± 0.33 × 10−3 mm2 /s).
Neoplastic group as compared to infective group had higher lesional FA (0.12 ± 0.13 vs. 0.10 ± 0.04) and higher lesional MD (0.98 ± 0.28 vs. 0.94 ± 0.18 × 10−3 mm2 /s). However, difference was not statistically significant [Table 1 and Graph 1 ].
Table 1: Association of diffusion tensor imaging findings with lesion type of patients enrolled in study
Graph 1: Association of DTI findings with lesion type in patients enrolled in study. DTI: Diffusion tensor imaging
Overall, the most common type of white matter tract involvement was displaced (32.0%), followed by edematous (28.0%), disrupted (20%), and infiltrated (20.0%) [Table 2 ]. In infective group, majority of cases had displaced tract involvement, followed by edematous. In neoplastic group, the most common tract involvement was disrupted, followed by edematous and infiltrated and displaced [Table 2 ].
Table 2: Association of involvement of tract with lesion type of patients enrolled in study
In cases with glioma, majority (76.9%) had tract disruption, while majority of cases in metastasis (58.3%) had edematous tract involvement. In cases with neurocysticercosis (50.0%) and tuberculoma (69.2%), the most common pattern was displacement. On comparing statistically, a significant association (P < 0.001) was found for type of tract involvement and etiology of lesions [Table 3 ].
Table 3: Association of tract involvement with etiology of lesion of patients enrolled in study
Discussion
Diffusion tensor imaging comparison between neoplastic versus infective
Mean FA values in the perilesional area of neoplastic lesions were found to be significantly higher (P < 0.001) in comparison to infective brain lesions. However, no significant difference was seen in the mean FA values in the intralesional area.
Mean MD values in the perilesional area of infective lesions were found to be significantly higher (P < 0.001) in comparison to neoplastic brain lesions. However, no significant difference was seen in the mean MD values in the intralesional area.
This statistically significant difference in the FA and the MD values in between two groups could be attributed to the variation in the composition of the perilesional edema predicated on the assumption that in HGG, tumor cell infiltration spreads at least as far as the T2-weighted MRI abnormality indicates.[ 7 ]
Soni et al .[ 8 ] evaluated the mean MD values in the perilesional areas of neoplastic and infective lesions. They concluded that nonneoplastic lesions had considerably higher (P = 0.015) perilesional MD values than neoplastic lesions, as was seen in our study. Mean perilesional FA values of the neoplastic lesions were greater than infective lesions in their study, similar trend as the mean values obtained in our study; however, there was not statistically significant difference between perilesional FA values in their study, unlike our study where FA values in the perilesional area of neoplastic lesions was significantly higher in comparison to the infective brain lesions. Similar to our study, they also found no significant difference in the lesional FA and MD values.
We found that high grade glioma had a greater perilesional FA than metastases (0.27 ± 0.09 vs. 0.06 ± 0.01) and tuberculomas (0.27 ± 0.09 vs. 0.08 ± 0.02), indicating varying interplay of cellular infiltration and vasogenic edema. A study by Soni et al .[ 8 ] is done in the past to compare the two major infective and neoplastic groups found similar results.
High grade glioma versus metastasis
We found that the mean perilesional FA values of HGG and metastasis were (0.27 ± 0.09 and 0.06 ± 0.01), respectively [Table 4 ], which was significantly larger in gliomas (P < 0.001) in comparison to metastasis. Mean perilesional MD values in the HGG and metastasis lesions were 0.93 ± 0.16 × 10−3 mm2 /s and 1.5 ± 0.16 × 10−3 mm2 /s, respectively, which was significantly higher in metastases (P < 0.001). As exemplified by one case of high grade glioma as shown in Figure 1a -h and one case of brain metastases as shown in Figure 2 a -h . Lesional mean FA values in HGG and metastasis were 0.13 ± 0.17 and 0.10 ± 0.10, respectively, while lesional mean MD values were 1.08 ± 0.20 × 10−3 mm2 /s and 0.87 ± 0.31 × 10−3 mm2 /s, respectively. Statistically negligible difference was seen in mean lesional FA and MD values of the two groups.
Table 4: Association of diffusion tensor imaging findings with etiology of lesions of patients enrolled in study
Figure 1: High Grade Glioma. 54 years old Male, with High grade glioma in right frontal lobe and gangliocapsular region Axial T2 FLAIR image (a) demonstrates intra-axial Space occupying lesion with perilesional edema-as shown within orange circle. Axial diffusion weighted image (b) shows patchy restriction-as shown by red arrow. Susceptibility weighted image (c) shows blooming s/o hemorrhage-as shown by green arrow. Post contrast T1 weighted image (d) shows the lesion has heterogenous post contrast enhancement-as shown by yellow arrow. (e) and (f) images show fractional anisotropy and Mean Diffusivity images and the Fractional Anisotropy And Mean Diffusivity values in the intralesional and perilesional Region of interest. Axial Diffusion Tensor Imaging color map section (g) at the same level shows that within the unaffected left hemisphere, the intact blue fibres of the cortico-spinal tract (red arrow) are easily identifiable. Within the lesion, corresponding fibres in the involved lobe are entirely destroyed-as shown inside blue ovale. (h) Fiber tracking of both corticospinal tracts : coronal color coded Diffusion Tensor Imaging map shows intact fibres on the left side, while fibres on the involved side are difficult to identify -as demarcated by the white star, with reduction in the Fractional Anisotropy values of the adjacent visualized fibres (m – 0.105) with reciprocal increase in the Mean Diffusivity value ( m – 0.006 ) -as compared to the normal side.
Figure 2: Metastasis. 44 years old Female patient, with Carcinoma Breast displaying a metastatic lesion in right ganglio-capsular region. Axial T2 FLAIR image (a) demonstrates lesion with perilesional edema extending into fronto-parietal region-as shown by green arrow. Post contrast T1 weighted image (b) shows the lesion has peripheral post contrast enhancement-as shown by red arrow. Axial fractional anisotropy (c) and Mean diffusivity (d) images shows the Fractional Anisotropy (0.03) And Mean Diffusivity (1.2x10 -3 mm sq per sec) values in the intralesional and Fractional Anisotropy (0.08) and Mean Diffusivity (o.37 x10 -3 mm sq per sec) values in the perilesional areas. Axial Diffusion Tensor Imaging color map axial section (e) at the same level shows identifiable yet disorganized color at the site of the lesion in comparison to the contralateral normal side-as shown by yellow arrow, suggestive of infiltrated nature of the fibres. Coronal Diffusion Tensor Imaging color map section (f) shows the edema extending into the white matter region of right fronto-parietal lobe with relatively identifiable fibres in the region of the lesion-as shown within orange ovale. It is correlated in the coronal 3D BRAVO section (g) showing ring enhancing lesion in right gangliocapsular region-yellow thick arrow. Tractography image (h) shows non visualization of the fibres in the frontal region of cortico-spinal tract on right side suggestive of significant edema causing severe reduction in Fractional Anisotropy values leading to non-trackability of the fibres-as shown within blue ovale
Lu et al .[ 9 ] performed DTI on 12 patients with metastatic lesions and 12 patients with HGG. Similar to our study, they concluded that the perilesional MD of metastasis was substantially higher than the gliomas (P < 0.005), but the perilesional FA values were not significantly different which are found to be significantly different in our study.
Jiang et al .[ 10 ] evaluated FA and MD values in the peritumoral and intratumoral regions in a meta-analysis of nine studies comparing HGG and metastases using DTI. Their findings were similar to our study, they found that FA was much greater and MD was significantly lower in the peritumoral area of HGG, indicating that FA was more effective in predicting metastasis, similar to our study. However, no significant change was seen in FA and MD values within the core of the lesion. Similar findings were seen in our study.
Wang et al .[ 11 ] in their study of 49 patients with 19 metastases and 30 glioblastoma multiforme (GBM) found that FA value was considerably lower (P < 0.001) in the peritumoral region in GBM (mean – 0.32 ± 0.05) as compared to metastasis (mean– 0.41 ± 0.08). Their findings did not corroborate with our study. May be attributed to the fact their patients had more de-differentiated GBM leading to widespread tract disruption.
However, Wang et al .[ 11 ] have described, gliomas produce large amounts of extracellular matrix components, which serve as substrate for adhesion an migration of tumor cells along the edema surrounding such lesions, which is the possible reason for higher perilesional FA in gliomas in our study as compared to metastases.
Neurocysticercosis versus tuberculoma
In our study, within the infective group, further subgroup analysis of neurocysticercosis [Figure 3a -e ] and tuberculoma [Figure 4a -h ] revealed mean perilesional FA values of tuberculoma and neurocysticercosis were 0.08 ± 0.01 and 0.08 ± 0.02, respectively while mean perilesional MD values in the tuberculoma and neurocysticercosis were 1.48 ± 0.20 × 10−3 mm2 /s and 1.53 ± 0.24 × 10−3 mm2 /s, respectively. On comparing the FA and MD values in the intralesional and perilesional area among neurocysticercosis and tuberculoma, no significant difference (P > 0.05) was observed.
Figure 3: 25 years old male with Neurocysticercosis showed a well-defined hyperintense lesion in right high frontal lobe on Axial T2 weighted image (a)-green arrow, with ring enhancement on post-contrast T1 weighted image (b)-red arrow. Axial (c) and coronal Diffusion Tensor Imaging color maps (d) show identifiable fibers in the region of the lesion with the same but slightly reduced hue attributed to the edema-green ovale in (c) and red ovale in (d). Coronal Tractography image (e) shows a normal identifiable corticospinal tract in the contralateral normal cerebral hemisphere whereas the fibers of the right corticospinal tract are not traceable due to the significant reduction in the Fractional Anisotropy values. Associated lateral displacement of the fibers can be seen-yellow arrow
Figure 4: Tuberculoma. MR images of an 18 years old female with multiple conglomerated tuberculomas in the left high frontal region. Axial T2 weighted image (a) shows confluent T2 hypointense lesions in the left high frontal region with moderate perilesional edema-green arrow. The lesions show peripheral enhancement on post-contrast T1 weighted images (b)-red arrow. Coronal Diffusion Tensor Imaging color map (c) shows alteration in the color of the lesion without a significant decrease in the Fractional Anisotropy value within, likely attributable to the displacement of the fibers-white arrow. Axial Diffusion Tensor Imaging color map (d) shows the reduction in the hues of the white matter tracts-yellow arrow- in the region of edema with a displaced left superior frontal-occipital tract in the axial tractography image (e)-red ovale. Coronal tractography image (f) shows left-lateral displacement of the left corticospinal tract in comparison to the normal right-sided corticospinal tract-green arrow. Fractional anisotropy (g) and the Mean Diffusivity (h) images show the Fractional Anisotropy and Mean Diffusivity values in the intralesional and perilesional areas
A similar comparison was also done by Soni et al .[ 8 ] in distinguishing 12 tuberculoma patients from 13 neurocysticercosis patients. However, similar to our study, no significant difference was seen in the intralesional and perilesional FA and MD values.
Gupta et al .[ 12 ] employed diffusion-weighted imaging in 100 patients to differentiate tuberculoma from neurocysticercosis lesions and found that the MD values of vesicular and degenerating stages of neurocysticercosis from the core substantially higher than the core of all tuberculomas and tuberculous abscesses.
Furthermore, there is no additional research that compares the DTI parameters in tuberculomas and neurocysticercosis differentiation.
High-grade glioma versus tuberculoma
We compared HGGs with tuberculomas and found a significantly increased (P < 0.001) perilesional FA in HGGs (m – 0.27 ± 0.09) when compared to tuberculomas (m – 0.08 ± 0.02) in accordance with peritumoral infiltration and lesser vasogenic edema in gliomas. Statistically significant higher value (P < 0.001) of MD perilesional edema was seen in tuberculomas (m – 1.53 × 10−3 ± 0.24 × 10−3 mm2 /s), in comparison to HGG (m – 0.93 × 10−3 ± 0.16 × 10−3 mm2 /s). The difference in lesional FA and MD was not statistically significant.
Soni et al .[ 8 ] compared 13 patients of HGGs and 12 patients with tuberculomas and similar to our study, they found that the perilesional FA value was significantly higher in HGG (P < 0.001) whereas perilesional MD values were significantly higher in tuberculomas (P < 0.001). Like our study, they also concluded that there was no significant difference in the lesional FA and MD values.
Gupta et al .[ 13 ] evaluated the FA and MD in 33 individuals with brain tuberculomas. They concluded that in comparison to normal white matter, FA values were substantially lower and MD was much higher in brain tuberculomas.
Chu et al .[ 14 ] used DTI in 12 patients of acquired immunodeficiency syndrome with brain tuberculosis and found that in the solid areas and areas with edema, FA and MD measurements had statistically significant different (P < 0.05) values when compared to the normal white matter. To date, no other original study has been documented in the literature that distinguishes HGGs from tuberculomas based on perilesional DTI parameters.
Tuberculoma versus metastasis
We compared tuberculomas with metastasis, and concluded that there was no statistically significant change (P > 0.05) in the perilesional FA and MD values, we concluded that it may be attributed to a more vasogenic character of perilesional edema in both lesions.
Soni et al .[ 8 ] compared 13 patients with tuberculoma and ten patients of metastasis for the DTI parameters and found similar results suggestive of no significant difference in the lesional and perilesional FA and MD values between the two. To date, no other DTI study has been found in the literature that uses diffusion parameters to discriminate metastasis from tuberculoma.
Toh et al .[ 15 ] used DTI to assess FA and MD values in the cystic component, thick enhancing walls, and perilesional white matter of 26 cystic metastases and 15 abscesses. They concluded that abscesses had much greater FA and lower MD in the cavity as well as in the enhancing rim of the lesion than metastasis and glioblastomas. In comparison to metastasis and cystic glioblastomas, the FA value was much lower and the MD value was significantly greater in the perilesional edema of the abscesses.
Furthermore, no significant relation was found in the intralesional and perilesional FA values of neurocysticercosis and metastasis. No studies are done to date comparing the role of diffusion tensor images in the differentiation of neurocysticercosis from metastatic lesions.
Qualitative assessment by color-coded diffusion tensor imaging maps and fiber tracking
We evaluated the white matter tracts involvement in patients of both neoplastic and infective groups and compared them with the corresponding normal tracts in the contralateral half of the brain.
The track was labeled as displaced if it demonstrated normal or slightly reduced FA (<0.5) with anomalous location or direction [Figure 4d ]. The tract was termed edematous if it showed lowered anisotropy (<0.22) with normal orientation but a high signal intensity on T2WI as [Figure 3f ].[ 4 ] It was classed as infiltrated if the tract had decreased anisotropy (<0.2) but could still be detected on color maps [Figure 2e ].[ 5 ] and disrupted if it had isotropic diffusion but could not be seen on directional color maps [Figure 1g ].[ 6 ] These cut-offs were decided in the corroboration of previous studies.[ 4–6 ]
Using this method, we characterized the involvement of the adjoining white matter fibers in the two groups into four categories, namely edematous, displaced, infiltrated, and destroyed. Of the total number of lesions displacement of the tracts was the most common pattern in a total of 18 cases (32%) followed by edematous in 16 (28%) disrupted in 13 cases (20%) and infiltrated in a total of 3 cases (6%). A combination of the patterns of involvement of white matter tracts was seen in most of the cases with edema being the most overlapping pattern of involvement. As vasogenic edema was seen in association of most of the lesions, the lesions with the other pattern of tract involvement in addition to the edema were evaluated in the study. Displacement of the tracts was the most common pattern of involvement in the infective lesion (60%), whereas disruption of tracts within the lesion was the most common pattern in neoplastic (40%).
The patients in the neoplastic and infectious groups were compared in our study. With a P < 0.001, there was a significant difference between these groups, with the benign group having a higher prevalence of displacement of the tracts (60%) and the malignant group having a higher prevalence of the disruption (40%). In contrast, there was no significant difference between the two groups when it came to edema and infiltration, with a P > 0.05 (found in both groups). On comparing statistically, a significant association was found between the etiology of the lesion and tract involvement. Our results were corroborative with the results of Ibrahim et al .[ 3 ] who performed DTI in 32 patients with intracranial neoplasm for their characterization and preoperative assessment. They concluded that the occurrence of tract displacement was substantially greater (P < 0.05) in the benign group than in the malignant group, whereas the occurrence of disruption was significantly higher (P < 0.05) in the malignant group than in the benign group. Field et al .[ 16 ] came to the same conclusion that in both benign and malignant tumors, displacement and edema patterns were found. Gliomas were the only ones that infiltrated adjoining white matter. White matter tracts adjacent to only malignant tumors were destroyed.
We studied lesional and perilesional changes in various groups and sub-groups, both quantitatively (FA and MD values) and qualitatively (by fiber tracking) and have found significant differences among various groups (Neoplastic vs. Infective) and subgroups (HGGs vs. Metastasis, HGGs vs. Tuberculoma and HGGs vs. Neurocysticercosis).
Conclusion
We conclude that DTI can be used to differentiate quantitatively and qualitatively among these groups and subgroups; particularly as a useful adjunct in diagnostic dilemma cases. Thus, in addition to regular magnetic resonance sequences, DTI sequences may aid in the differentiation of brain lesions of various etiologies. This might be extremely helpful in determining therapy options and enhancing the patient’s quality of life and prognostication.
Limitations
The number of patients included in the trial is not very big. A bigger sample size might be beneficial in determining the role of DTI in distinguishing brain lesions from various causes. Heterogeneity of the lesion may impact the assessment of DTI measures. We tried to minimize this issue by using three ROIs in different regions and using the average of the three in the study, however, the discrepancy due to heterogeneity cannot be completely eliminated.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References
1. Al-Okaili RN, Krejza J, Woo JH, Wolf RL, O'Rourke DM, Judy KD, et al. Intraaxial brain masses:MR imaging-based diagnostic strategy –Initial experience. Radiology 2007;243:539–50.
2. Huisman TA. Tumor-like lesions of the brain. Cancer Imaging 2009;9:S10–3.
3. Ibrahim AS, Gomaa M, Sakr H, Elzaher YA. Role of diffusion tensor imaging in characterization and preoperative planning of brain neoplasms. Egypt J Radiol Nucl Med 2013;2:297–307.
4. Min ZG, Niu C, Rana N, Ji HM, Zhang M. Differentiation of pure vasogenic edema and tumor-infiltrated edema in patients with peritumoral edema by analyzing the relationship of axial and radial diffusivities on 3.0T MRI. Clin Neurol Neurosurg 2013;115:1366–70.
5. Min Z, Rana N, Niu C, Ji H, Zhang M. Does diffusion tensor tractography of the corticospinal tract correctly reflect motor function?. Med Princ Pract 2014;23:174–6.
6. Zhang P, Liu B. Differentiation among glioblastomas, primary cerebral lymphomas, and solitary brain metastases using diffusion-weighted imaging and diffusion tensor imaging:A PRISMA-compliant meta-analysis. ACS Chem Neurosci 2020;11:477–83.
7. Kelly PJ, Daumas-Duport C, Scheithauer BW, Kall BA, Kispert DB. Stereotactic histologic correlations of computed tomography and magnetic resonance imaging-defined abnormalities in patients with glial neoplasms. Mayo Clin Proc 1987;62:450–9.
8. Soni N, Srindharan K, Kumar S, Bhaisora KS, Kalita J, Mehrotra A, et al. Application of diffusion tensor imaging in brain lesions:A comparative study of neoplastic and non-neoplastic brain lesions. Neurol India 2018;66:1667–71.
9. Lu S, Ahn D, Johnson G, Cha S. Peritumoral diffusion tensor imaging of high-grade gliomas and metastatic brain tumors. AJNR Am J Neuroradiol 2003;24:937–41.
10. Jiang R, Du FZ, He C, Gu M, Ke ZW, Li JH. The value of diffusion tensor imaging in differentiating high-grade gliomas from brain metastases:A systematic review and meta-analysis. PLoS One 2014;9:e112550.
11. Wang W, Steward CE, Desmond PM. Diffusion tensor imaging in glioblastoma multiforme and brain metastases:The role of p, q, L, and fractional anisotropy. AJNR Am J Neuroradiol 2009;30:203–8.
12. Gupta RK, Prakash M, Mishra AM, Husain M, Prasad KN, Husain N. Role of diffusion weighted imaging in differentiation of intracranial tuberculoma and tuberculous abscess from cysticercus granulomas –A report of more than 100 lesions. Eur J Radiol 2005;55:384–92.
13. Gupta RK, Haris M, Husain N, Saksena S, Husain M, Rathore RK. DTI derived indices correlate with immunohistochemistry obtained matrix metalloproteinase (MMP-9) expression in cellular fraction of brain tuberculoma. J Neurol Sci 2008;275:78–85.
14. Chu X, Zhao L, Ma J, Wang Y, Jia W, Yang Y, et al. Application of diffusion tensor imaging in AIDS patients with brain opportunistic diseases:A comparative study of tuberculosis and toxoplasmosis. Radiol Infect Dis 2015;2:11–5.
15. Toh CH, Wei KC, Ng SH, Wan YL, Lin CP, Castillo M. Differentiation of brain abscesses from necrotic glioblastomas and cystic metastatic brain tumors with diffusion tensor imaging. AJNR Am J Neuroradiol 2011;32:1646–51.
16. Field AS, Alexander AL, Wu YC, Hasan KM, Witwer B, Badie B. Diffusion tensor eigenvector directional color imaging patterns in the evaluation of cerebral white matter tracts altered by tumor. J Magn Reson Imaging 2004;20:555–62.
